From nobody Sun Nov 10 21:34:25 2024 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass(p=none dis=none) header.from=linaro.org ARC-Seal: i=1; a=rsa-sha256; t=1711631046; cv=none; d=zohomail.com; s=zohoarc; b=XtVOw9hv2VJWXPhYEKh9zBC/FtQGRZwfTA5IdrcILnKBEAq7WlStD8pZ8X2mapa5gXHCThinliTE5mGDL5qcqhxqbUYurcqXfWYU+BrW7LeABe7b+FVZUsG+7kqvMF8ZCM3r66lHI39J+3mK4oUrqKtZD+eSA3Jx0gPwCu3d2Ck= ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com; s=zohoarc; t=1711631046; h=Content-Type:Content-Transfer-Encoding:Cc:Cc:Date:Date:From:From:In-Reply-To:List-Subscribe:List-Post:List-Id:List-Archive:List-Help:List-Unsubscribe:MIME-Version:Message-ID:References:Sender:Subject:Subject:To:To:Message-Id:Reply-To; bh=NTMG8MJ3mXAAw5DE5lYuXiIXFlAbdlMp3fkaaBvZgAs=; b=cdDbybud1U1fSrIB1Nlz/wDoHPfgpnULmB8tkYaDdsTyVlRqWayTWsI5qWcfp78jqFikdc8/39kkNg3VvBT5PBY1pX4r/cGAF6fLLNKrERJGz60d5hqTP+ozVjQE+XT9Obk8IgT40dK1Qai51+sOZCykC9RKlNC+SUGrym8/+ts= ARC-Authentication-Results: i=1; mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass header.from= (p=none dis=none) Return-Path: Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 1711631046159553.7609779717455; Thu, 28 Mar 2024 06:04:06 -0700 (PDT) Received: from localhost ([::1] helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1rppPn-0000HE-Cc; Thu, 28 Mar 2024 09:03:31 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1rppPf-00009I-0j for qemu-devel@nongnu.org; Thu, 28 Mar 2024 09:03:23 -0400 Received: from mail-lf1-x12c.google.com ([2a00:1450:4864:20::12c]) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1rppPY-00087f-Px for qemu-devel@nongnu.org; Thu, 28 Mar 2024 09:03:22 -0400 Received: by mail-lf1-x12c.google.com with SMTP id 2adb3069b0e04-513cf9bacf1so1093607e87.0 for ; Thu, 28 Mar 2024 06:03:16 -0700 (PDT) Received: from m1x-phil.lan (pas38-h02-176-184-5-52.dsl.sta.abo.bbox.fr. [176.184.5.52]) by smtp.gmail.com with ESMTPSA id re26-20020a170906d8da00b00a474ef94fddsm734372ejb.70.2024.03.28.06.03.04 (version=TLS1_3 cipher=TLS_CHACHA20_POLY1305_SHA256 bits=256/256); Thu, 28 Mar 2024 06:03:08 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; t=1711630994; x=1712235794; darn=nongnu.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=NTMG8MJ3mXAAw5DE5lYuXiIXFlAbdlMp3fkaaBvZgAs=; b=C6H6naHM0uydiasyAFcDdfVXixrvBDEyMQ0qsQMUWfv5aMB6oz2JZWswO8z/BZx6TB rZqbz3tt69EpwaCiTxr+k32lblQRiounDX/sG5OzbhoRNyCFSlgTRgqcSnaM+115j+Rc HzWNKuUE9IHQZknMG4gevRl8ddTSJngKFE4RF8YTCaiaZ95y1AdfRB/CXdTXevdlM3X2 sW9AZq2p3hlDljWIbGU67Nkr1uzb0JN5U2iwKfOqfvkFUhU/tf6BzAYI1s9tlesykA+V Q4RgD9djxO3VJ+wBGum1ls3HkCnmCJQEYT3Ehi766JWJprRas85yJzwtls2e0O34d4bB 06+A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1711630994; x=1712235794; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=NTMG8MJ3mXAAw5DE5lYuXiIXFlAbdlMp3fkaaBvZgAs=; b=Xl6eDIVlCZEYAKAfWvbWDepQEF8boo1BjkL+dnpkzFLAQ6NV51RoiLIytUBBVwuKPt zopdAMTXN8aEFYp2e+Ef4rIfuTOX9ND6rgt/LDttzdr//wGBQTz4H1Xf4O0rPuOGCOhO vDUhdG6ZLTHnWDaQPcfPBjtWetTjpT5ltWL6w8bJmRhBli8aZz9b6Wdn/EFYLMbIASD8 loNbL4KWd4kZsX/u7KL8l97iWq/lt+x9OqLiPMaZRV+qfogUA5RP3fsms3lrycGOVP4G +Lry6JvLVV1sW0MgKt6pG8inJknKj8kaiWQ/7aIbex1Z7ZCi/Wm4fJKV4CJzUSvE5SKF P/7g== X-Gm-Message-State: AOJu0YwxJdiRJW0KSW1VRUSTTwLX+GWxqJ0Hq1Krs01ZvT/Y86Ld4S0l ZMb+BDXyZXWtb1uXLS+QpNVor0BwX4jZ/1aOg2P/ltP7ciKzdZjg1wm4dIhgycOhZhX3RAtBEKq t X-Google-Smtp-Source: AGHT+IEAQGrH+6b+vfnLl4TcLgvkrXI+FCnYlzGW0uvOsv8K2toT3xa05aP9iIEQLKl47sQ5Ijyuow== X-Received: by 2002:ac2:5f8e:0:b0:515:a5b1:1dd0 with SMTP id r14-20020ac25f8e000000b00515a5b11dd0mr1646227lfe.55.1711630989406; Thu, 28 Mar 2024 06:03:09 -0700 (PDT) From: =?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?= To: qemu-devel@nongnu.org Cc: Yuval Shaia , Kevin Wolf , Prasanna Kumar Kalever , Fabiano Rosas , Cornelia Huck , Michael Roth , Li Zhijian , Prasanna Kumar Kalever , Peter Xu , integration@gluster.org, Paolo Bonzini , qemu-block@nongnu.org, =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= , devel@lists.libvirt.org, Hanna Reitz , "Michael S. Tsirkin" , Thomas Huth , Eric Blake , =?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?= , Marcel Apfelbaum , Song Gao , "Dr. David Alan Gilbert" , Eduardo Habkost , Yanan Wang , =?UTF-8?q?Marc-Andr=C3=A9=20Lureau?= , Markus Armbruster , =?UTF-8?q?Alex=20Benn=C3=A9e?= , Wainer dos Santos Moschetta , Beraldo Leal Subject: [PATCH-for-9.1 v2 1/3] hw/rdma: Remove pvrdma device and rdmacm-mux helper Date: Thu, 28 Mar 2024 14:02:53 +0100 Message-ID: <20240328130255.52257-2-philmd@linaro.org> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20240328130255.52257-1-philmd@linaro.org> References: <20240328130255.52257-1-philmd@linaro.org> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Received-SPF: pass client-ip=2a00:1450:4864:20::12c; envelope-from=philmd@linaro.org; helo=mail-lf1-x12c.google.com X-Spam_score_int: -20 X-Spam_score: -2.1 X-Spam_bar: -- X-Spam_report: (-2.1 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, RCVD_IN_DNSWL_NONE=-0.0001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=unavailable autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: qemu-devel-bounces+importer=patchew.org@nongnu.org X-ZohoMail-DKIM: pass (identity @linaro.org) X-ZM-MESSAGEID: 1711631047939100001 The whole RDMA subsystem was deprecated in commit e9a54265f5 ("hw/rdma: Deprecate the pvrdma device and the rdma subsystem") released in v8.2. Remove: - PVRDMA device - generated vmw_pvrdma/ directory from linux-headers - rdmacm-mux tool from contrib/ Cc: Yuval Shaia Cc: Marcel Apfelbaum Signed-off-by: Philippe Mathieu-Daud=C3=A9 Reviewed-by: Michael S. Tsirkin Reviewed-by: Thomas Huth --- MAINTAINERS | 10 - docs/about/deprecated.rst | 9 - docs/about/removed-features.rst | 4 + docs/pvrdma.txt | 345 ---- docs/system/loongarch/virt.rst | 2 +- meson.build | 36 - qapi/machine.json | 17 - qapi/qapi-schema.json | 1 - qapi/rdma.json | 38 - contrib/rdmacm-mux/rdmacm-mux.h | 61 - hw/rdma/rdma_backend.h | 129 -- hw/rdma/rdma_backend_defs.h | 76 - hw/rdma/rdma_rm.h | 97 -- hw/rdma/rdma_rm_defs.h | 146 -- hw/rdma/rdma_utils.h | 63 - hw/rdma/trace.h | 1 - hw/rdma/vmw/pvrdma.h | 144 -- hw/rdma/vmw/pvrdma_dev_ring.h | 46 - hw/rdma/vmw/pvrdma_qp_ops.h | 28 - hw/rdma/vmw/trace.h | 1 - include/hw/rdma/rdma.h | 37 - include/monitor/hmp.h | 1 - .../infiniband/hw/vmw_pvrdma/pvrdma_dev_api.h | 685 -------- .../infiniband/hw/vmw_pvrdma/pvrdma_verbs.h | 348 ---- .../standard-headers/rdma/vmw_pvrdma-abi.h | 310 ---- contrib/rdmacm-mux/main.c | 831 ---------- hw/core/machine-qmp-cmds.c | 32 - hw/rdma/rdma.c | 30 - hw/rdma/rdma_backend.c | 1401 ----------------- hw/rdma/rdma_rm.c | 812 ---------- hw/rdma/rdma_utils.c | 126 -- hw/rdma/vmw/pvrdma_cmd.c | 815 ---------- hw/rdma/vmw/pvrdma_dev_ring.c | 141 -- hw/rdma/vmw/pvrdma_main.c | 735 --------- hw/rdma/vmw/pvrdma_qp_ops.c | 298 ---- monitor/qmp-cmds.c | 1 - Kconfig.host | 3 - contrib/rdmacm-mux/meson.build | 7 - hmp-commands-info.hx | 13 - hw/Kconfig | 1 - hw/meson.build | 1 - hw/rdma/Kconfig | 3 - hw/rdma/meson.build | 12 - hw/rdma/trace-events | 31 - hw/rdma/vmw/trace-events | 17 - meson_options.txt | 2 - qapi/meson.build | 1 - qemu-options.hx | 3 - .../ci/org.centos/stream/8/x86_64/configure | 1 - scripts/meson-buildoptions.sh | 3 - scripts/update-linux-headers.sh | 27 - 51 files changed, 5 insertions(+), 7977 deletions(-) delete mode 100644 docs/pvrdma.txt delete mode 100644 qapi/rdma.json delete mode 100644 contrib/rdmacm-mux/rdmacm-mux.h delete mode 100644 hw/rdma/rdma_backend.h delete mode 100644 hw/rdma/rdma_backend_defs.h delete mode 100644 hw/rdma/rdma_rm.h delete mode 100644 hw/rdma/rdma_rm_defs.h delete mode 100644 hw/rdma/rdma_utils.h delete mode 100644 hw/rdma/trace.h delete mode 100644 hw/rdma/vmw/pvrdma.h delete mode 100644 hw/rdma/vmw/pvrdma_dev_ring.h delete mode 100644 hw/rdma/vmw/pvrdma_qp_ops.h delete mode 100644 hw/rdma/vmw/trace.h delete mode 100644 include/hw/rdma/rdma.h delete mode 100644 include/standard-headers/drivers/infiniband/hw/vmw_pvrd= ma/pvrdma_dev_api.h delete mode 100644 include/standard-headers/drivers/infiniband/hw/vmw_pvrd= ma/pvrdma_verbs.h delete mode 100644 include/standard-headers/rdma/vmw_pvrdma-abi.h delete mode 100644 contrib/rdmacm-mux/main.c delete mode 100644 hw/rdma/rdma.c delete mode 100644 hw/rdma/rdma_backend.c delete mode 100644 hw/rdma/rdma_rm.c delete mode 100644 hw/rdma/rdma_utils.c delete mode 100644 hw/rdma/vmw/pvrdma_cmd.c delete mode 100644 hw/rdma/vmw/pvrdma_dev_ring.c delete mode 100644 hw/rdma/vmw/pvrdma_main.c delete mode 100644 hw/rdma/vmw/pvrdma_qp_ops.c delete mode 100644 contrib/rdmacm-mux/meson.build delete mode 100644 hw/rdma/Kconfig delete mode 100644 hw/rdma/meson.build delete mode 100644 hw/rdma/trace-events delete mode 100644 hw/rdma/vmw/trace-events diff --git a/MAINTAINERS b/MAINTAINERS index a07af6b9d4..91ab5235b8 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4060,16 +4060,6 @@ F: block/replication.c F: tests/unit/test-replication.c F: docs/block-replication.txt =20 -PVRDMA -M: Yuval Shaia -M: Marcel Apfelbaum -S: Odd Fixes -F: hw/rdma/* -F: hw/rdma/vmw/* -F: docs/pvrdma.txt -F: contrib/rdmacm-mux/* -F: qapi/rdma.json - Semihosting M: Alex Benn=C3=A9e S: Maintained diff --git a/docs/about/deprecated.rst b/docs/about/deprecated.rst index 7b548519b5..29eae69e50 100644 --- a/docs/about/deprecated.rst +++ b/docs/about/deprecated.rst @@ -376,15 +376,6 @@ recommending to switch to their stable counterparts: - "Zve64f" should be replaced with "zve64f" - "Zve64d" should be replaced with "zve64d" =20 -``-device pvrdma`` and the rdma subsystem (since 8.2) -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - -The pvrdma device and the whole rdma subsystem are in a bad shape and -without active maintenance. The QEMU project intends to remove this -device and subsystem from the code base in a future release without -replacement unless somebody steps up and improves the situation. - - Block device options '''''''''''''''''''' =20 diff --git a/docs/about/removed-features.rst b/docs/about/removed-features.= rst index f9cf874f7b..4d5bdc43b4 100644 --- a/docs/about/removed-features.rst +++ b/docs/about/removed-features.rst @@ -909,6 +909,10 @@ contains native support for this feature and thus use = of the option ROM approach was obsolete. The native SeaBIOS support can be activated by using ``-machine graphics=3Doff``. =20 +``pvrdma`` and the RDMA subsystem (removed in 9.1) +'''''''''''''''''''''''''''''''''''''''''''''''''' + +The 'pvrdma' device and the whole RDMA subsystem have been removed. =20 Related binaries ---------------- diff --git a/docs/pvrdma.txt b/docs/pvrdma.txt deleted file mode 100644 index 5c122fe818..0000000000 --- a/docs/pvrdma.txt +++ /dev/null @@ -1,345 +0,0 @@ -Paravirtualized RDMA Device (PVRDMA) -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - - -1. Description -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -PVRDMA is the QEMU implementation of VMware's paravirtualized RDMA device. -It works with its Linux Kernel driver AS IS, no need for any special guest -modifications. - -While it complies with the VMware device, it can also communicate with bare -metal RDMA-enabled machines as peers. - -It does not require an RDMA HCA in the host, it can work with Soft-RoCE (r= xe). - -It does not require the whole guest RAM to be pinned allowing memory -over-commit and, even if not implemented yet, migration support will be -possible with some HW assistance. - -A project presentation accompany this document: -- https://blog.linuxplumbersconf.org/2017/ocw/system/presentations/4730/or= iginal/lpc-2017-pvrdma-marcel-apfelbaum-yuval-shaia.pdf - - - -2. Setup -=3D=3D=3D=3D=3D=3D=3D=3D - - -2.1 Guest setup -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -Fedora 27+ kernels work out of the box, older distributions -require updating the kernel to 4.14 to include the pvrdma driver. - -However the libpvrdma library needed by User Level Software is still -not available as part of the distributions, so the rdma-core library -needs to be compiled and optionally installed. - -Please follow the instructions at: - https://github.com/linux-rdma/rdma-core.git - - -2.2 Host Setup -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -The pvrdma backend is an ibdevice interface that can be exposed -either by a Soft-RoCE(rxe) device on machines with no RDMA device, -or an HCA SRIOV function(VF/PF). -Note that ibdevice interfaces can't be shared between pvrdma devices, -each one requiring a separate instance (rxe or SRIOV VF). - - -2.2.1 Soft-RoCE backend(rxe) -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D -A stable version of rxe is required, Fedora 27+ or a Linux -Kernel 4.14+ is preferred. - -The rdma_rxe module is part of the Linux Kernel but not loaded by default. -Install the User Level library (librxe) following the instructions from: -https://github.com/SoftRoCE/rxe-dev/wiki/rxe-dev:-Home - -Associate an ETH interface with rxe by running: - rxe_cfg add eth0 -An rxe0 ibdevice interface will be created and can be used as pvrdma backe= nd. - - -2.2.2 RDMA device Virtual Function backend -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -Nothing special is required, the pvrdma device can work not only with -Ethernet Links, but also Infinibands Links. -All is needed is an ibdevice with an active port, for Mellanox cards -will be something like mlx5_6 which can be the backend. - - -2.2.3 QEMU setup -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -Configure QEMU with --enable-rdma flag, installing -the required RDMA libraries. - - - -3. Usage -=3D=3D=3D=3D=3D=3D=3D=3D - - -3.1 VM Memory settings -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -Currently the device is working only with memory backed RAM -and it must be mark as "shared": - -m 1G \ - -object memory-backend-ram,id=3Dmb1,size=3D1G,share \ - -numa node,memdev=3Dmb1 \ - - -3.2 MAD Multiplexer -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -MAD Multiplexer is a service that exposes MAD-like interface for VMs in -order to overcome the limitation where only single entity can register with -MAD layer to send and receive RDMA-CM MAD packets. - -To build rdmacm-mux run -# make rdmacm-mux - -Before running the rdmacm-mux make sure that both ib_cm and rdma_cm kernel -modules aren't loaded, otherwise the rdmacm-mux service will fail to start. - -The application accepts 3 command line arguments and exposes a UNIX socket -to pass control and data to it. --d rdma-device-name Name of RDMA device to register with --s unix-socket-path Path to unix socket to listen (default /var/run/rdmac= m-mux) --p rdma-device-port Port number of RDMA device to register with (default = 1) -The final UNIX socket file name is a concatenation of the 3 arguments so -for example for device mlx5_0 on port 2 this /var/run/rdmacm-mux-mlx5_0-2 -will be created. - -pvrdma requires this service. - -Please refer to contrib/rdmacm-mux for more details. - - -3.3 Service exposed by libvirt daemon -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -The control over the RDMA device's GID table is done by updating the -device's Ethernet function addresses. -Usually the first GID entry is determined by the MAC address, the second by -the first IPv6 address and the third by the IPv4 address. Other entries can -be added by adding more IP addresses. The opposite is the same, i.e. -whenever an address is removed, the corresponding GID entry is removed. -The process is done by the network and RDMA stacks. Whenever an address is -added the ib_core driver is notified and calls the device driver add_gid -function which in turn update the device. -To support this in pvrdma device the device hooks into the create_bind and -destroy_bind HW commands triggered by pvrdma driver in guest. - -Whenever changed is made to the pvrdma port's GID table a special QMP -messages is sent to be processed by libvirt to update the address of the -backend Ethernet device. - -pvrdma requires that libvirt service will be up. - - -3.4 PCI devices settings -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -RoCE device exposes two functions - an Ethernet and RDMA. -To support it, pvrdma device is composed of two PCI functions, an Ethernet -device of type vmxnet3 on PCI slot 0 and a PVRDMA device on PCI slot 1. The -Ethernet function can be used for other Ethernet purposes such as IP. - - -3.5 Device parameters -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -- netdev: Specifies the Ethernet device function name on the host for - example enp175s0f0. For Soft-RoCE device (rxe) this would be the Ethernet - device used to create it. -- ibdev: The IB device name on host for example rxe0, mlx5_0 etc. -- mad-chardev: The name of the MAD multiplexer char device. -- ibport: In case of multi-port device (such as Mellanox's HCA) this - specify the port to use. If not set 1 will be used. -- dev-caps-max-mr-size: The maximum size of MR. -- dev-caps-max-qp: Maximum number of QPs. -- dev-caps-max-cq: Maximum number of CQs. -- dev-caps-max-mr: Maximum number of MRs. -- dev-caps-max-pd: Maximum number of PDs. -- dev-caps-max-ah: Maximum number of AHs. - -Notes: -- The first 3 parameters are mandatory settings, the rest have their - defaults. -- The last 8 parameters (the ones that prefixed by dev-caps) defines the t= op - limits but the final values is adjusted by the backend device limitation= s. -- netdev can be extracted from ibdev's sysfs - (/sys/class/infiniband//device/net/) - - -3.6 Example -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -Define bridge device with vmxnet3 network backend: - - - - -
- - -Define pvrdma device: - - - - - - - - - - - - - -4. Implementation details -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D - - -4.1 Overview -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -The device acts like a proxy between the Guest Driver and the host -ibdevice interface. -On configuration path: - - For every hardware resource request (PD/QP/CQ/...) the pvrdma will requ= est - a resource from the backend interface, maintaining a 1-1 mapping - between the guest and host. -On data path: - - Every post_send/receive received from the guest will be converted into - a post_send/receive for the backend. The buffers data will not be touch= ed - or copied resulting in near bare-metal performance for large enough buf= fers. - - Completions from the backend interface will result in completions for - the pvrdma device. - - -4.2 PCI BARs -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -PCI Bars: - BAR 0 - MSI-X - MSI-X vectors: - (0) Command - used when execution of a command is completed. - (1) Async - not in use. - (2) Completion - used when a completion event is placed in - device's CQ ring. - BAR 1 - Registers - -------------------------------------------------------- - | VERSION | DSR | CTL | REQ | ERR | ICR | IMR | MAC | - -------------------------------------------------------- - DSR - Address of driver/device shared memory used - for the command channel, used for passing: - - General info such as driver version - - Address of 'command' and 'response' - - Address of async ring - - Address of device's CQ ring - - Device capabilities - CTL - Device control operations (activate, reset etc) - IMG - Set interrupt mask - REQ - Command execution register - ERR - Operation status - - BAR 2 - UAR - --------------------------------------------------------- - | QP_NUM | SEND/RECV Flag || CQ_NUM | ARM/POLL Flag | - --------------------------------------------------------- - - Offset 0 used for QP operations (send and recv) - - Offset 4 used for CQ operations (arm and poll) - - -4.3 Major flows -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - -4.3.1 Create CQ -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - - Guest driver - - Allocates pages for CQ ring - - Creates page directory (pdir) to hold CQ ring's pages - - Initializes CQ ring - - Initializes 'Create CQ' command object (cqe, pdir etc) - - Copies the command to 'command' address - - Writes 0 into REQ register - - Device - - Reads the request object from the 'command' address - - Allocates CQ object and initialize CQ ring based on pdir - - Creates the backend CQ - - Writes operation status to ERR register - - Posts command-interrupt to guest - - Guest driver - - Reads the HW response code from ERR register - -4.3.2 Create QP -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - - Guest driver - - Allocates pages for send and receive rings - - Creates page directory(pdir) to hold the ring's pages - - Initializes 'Create QP' command object (max_send_wr, - send_cq_handle, recv_cq_handle, pdir etc) - - Copies the object to 'command' address - - Write 0 into REQ register - - Device - - Reads the request object from 'command' address - - Allocates the QP object and initialize - - Send and recv rings based on pdir - - Send and recv ring state - - Creates the backend QP - - Writes the operation status to ERR register - - Posts command-interrupt to guest - - Guest driver - - Reads the HW response code from ERR register - -4.3.3 Post receive -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - - Guest driver - - Initializes a wqe and place it on recv ring - - Write to qpn|qp_recv_bit (31) to QP offset in UAR - - Device - - Extracts qpn from UAR - - Walks through the ring and does the following for each wqe - - Prepares the backend CQE context to be used when - receiving completion from backend (wr_id, op_code, emu_cq_nu= m) - - For each sge prepares backend sge - - Calls backend's post_recv - -4.3.4 Process backend events -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D - - Done by a dedicated thread used to process backend events; - at initialization is attached to the device and creates - the communication channel. - - Thread main loop: - - Polls for completions - - Extracts QEMU _cq_num, wr_id and op_code from context - - Writes CQE to CQ ring - - Writes CQ number to device CQ - - Sends completion-interrupt to guest - - Deallocates context - - Acks the event to backend - - - -5. Limitations -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -- The device obviously is limited by the Guest Linux Driver features imple= mentation - of the VMware device API. -- Memory registration mechanism requires mremap for every page in the buff= er in order - to map it to a contiguous virtual address range. Since this is not the d= ata path - it should not matter much. If the default max mr size is increased, be a= ware that - memory registration can take up to 0.5 seconds for 1GB of memory. -- The device requires target page size to be the same as the host page siz= e, - otherwise it will fail to init. -- QEMU cannot map guest RAM from a file descriptor if a pvrdma device is a= ttached, - so it can't work with huge pages. The limitation will be addressed in th= e future, - however QEMU allocates Guest RAM with MADV_HUGEPAGE so if there are enou= gh huge - pages available, QEMU will use them. QEMU will fail to init if the requi= rements - are not met. - - - -6. Performance -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -By design the pvrdma device exits on each post-send/receive, so for small = buffers -the performance is affected; however for medium buffers it will became clo= se to -bare metal and from 1MB buffers and up it reaches bare metal performance. -(tested with 2 VMs, the pvrdma devices connected to 2 VFs of the same devi= ce) - -All the above assumes no memory registration is done on data path. diff --git a/docs/system/loongarch/virt.rst b/docs/system/loongarch/virt.rst index c37268b404..06d034b8ef 100644 --- a/docs/system/loongarch/virt.rst +++ b/docs/system/loongarch/virt.rst @@ -39,7 +39,7 @@ can be accessed by following steps. =20 .. code-block:: bash =20 - ./configure --disable-rdma --disable-pvrdma --prefix=3D/usr \ + ./configure --disable-rdma --prefix=3D/usr \ --target-list=3D"loongarch64-softmmu" \ --disable-libiscsi --disable-libnfs --disable-libpmem \ --disable-glusterfs --enable-libusb --enable-usb-redir \ diff --git a/meson.build b/meson.build index c9c3217ba4..d6af3cd53a 100644 --- a/meson.build +++ b/meson.build @@ -2829,37 +2829,6 @@ config_host_data.set('CONFIG_ARM_AES_BUILTIN', cc.co= mpiles(''' void foo(uint8x16_t *p) { *p =3D vaesmcq_u8(*p); } ''')) =20 -have_pvrdma =3D get_option('pvrdma') \ - .require(rdma.found(), error_message: 'PVRDMA requires OpenFabrics libra= ries') \ - .require(cc.compiles(gnu_source_prefix + ''' - #include - int main(void) - { - char buf =3D 0; - void *addr =3D &buf; - addr =3D mremap(addr, 0, 1, MREMAP_MAYMOVE | MREMAP_FIXED); - - return 0; - }'''), error_message: 'PVRDMA requires mremap').allowed() - -if have_pvrdma - config_host_data.set('LEGACY_RDMA_REG_MR', not cc.links(''' - #include - int main(void) - { - struct ibv_mr *mr; - struct ibv_pd *pd =3D NULL; - size_t length =3D 10; - uint64_t iova =3D 0; - int access =3D 0; - void *addr =3D NULL; - - mr =3D ibv_reg_mr_iova(pd, addr, length, iova, access); - ibv_dereg_mr(mr); - return 0; - }''')) -endif - if get_option('membarrier').disabled() have_membarrier =3D false elif host_os =3D=3D 'windows' @@ -2993,7 +2962,6 @@ host_kconfig =3D \ (have_vhost_kernel ? ['CONFIG_VHOST_KERNEL=3Dy'] : []) + \ (have_virtfs ? ['CONFIG_VIRTFS=3Dy'] : []) + \ (host_os =3D=3D 'linux' ? ['CONFIG_LINUX=3Dy'] : []) + \ - (have_pvrdma ? ['CONFIG_PVRDMA=3Dy'] : []) + \ (multiprocess_allowed ? ['CONFIG_MULTIPROCESS_ALLOWED=3Dy'] : []) + \ (vfio_user_server_allowed ? ['CONFIG_VFIO_USER_SERVER_ALLOWED=3Dy'] : []= ) + \ (hv_balloon ? ['CONFIG_HV_BALLOON_POSSIBLE=3Dy'] : []) @@ -3357,8 +3325,6 @@ if have_system 'hw/pci', 'hw/pci-host', 'hw/ppc', - 'hw/rdma', - 'hw/rdma/vmw', 'hw/rtc', 'hw/s390x', 'hw/scsi', @@ -4028,7 +3994,6 @@ if have_tools }] endforeach =20 - subdir('contrib/rdmacm-mux') subdir('contrib/elf2dmp') =20 executable('qemu-edid', files('qemu-edid.c', 'hw/display/edid-generate.c= '), @@ -4434,7 +4399,6 @@ summary_info +=3D {'Linux AIO support': libaio} summary_info +=3D {'Linux io_uring support': linux_io_uring} summary_info +=3D {'ATTR/XATTR support': libattr} summary_info +=3D {'RDMA support': rdma} -summary_info +=3D {'PVRDMA support': have_pvrdma} summary_info +=3D {'fdt support': fdt_opt =3D=3D 'disabled' ? false = : fdt_opt} summary_info +=3D {'libcap-ng support': libcap_ng} summary_info +=3D {'bpf support': libbpf} diff --git a/qapi/machine.json b/qapi/machine.json index e8b60641f2..e9f0f0c49a 100644 --- a/qapi/machine.json +++ b/qapi/machine.json @@ -1737,23 +1737,6 @@ 'returns': 'HumanReadableText', 'features': [ 'unstable' ] } =20 -## -# @x-query-rdma: -# -# Query RDMA state -# -# Features: -# -# @unstable: This command is meant for debugging. -# -# Returns: RDMA state -# -# Since: 6.2 -## -{ 'command': 'x-query-rdma', - 'returns': 'HumanReadableText', - 'features': [ 'unstable' ] } - ## # @x-query-roms: # diff --git a/qapi/qapi-schema.json b/qapi/qapi-schema.json index 8304d45625..5e33da7228 100644 --- a/qapi/qapi-schema.json +++ b/qapi/qapi-schema.json @@ -54,7 +54,6 @@ { 'include': 'dump.json' } { 'include': 'net.json' } { 'include': 'ebpf.json' } -{ 'include': 'rdma.json' } { 'include': 'rocker.json' } { 'include': 'tpm.json' } { 'include': 'ui.json' } diff --git a/qapi/rdma.json b/qapi/rdma.json deleted file mode 100644 index 195c001850..0000000000 --- a/qapi/rdma.json +++ /dev/null @@ -1,38 +0,0 @@ -# -*- Mode: Python -*- -# vim: filetype=3Dpython -# - -## -# =3D RDMA device -## - -## -# @RDMA_GID_STATUS_CHANGED: -# -# Emitted when guest driver adds/deletes GID to/from device -# -# @netdev: RoCE Network Device name -# -# @gid-status: Add or delete indication -# -# @subnet-prefix: Subnet Prefix -# -# @interface-id: Interface ID -# -# Since: 4.0 -# -# Example: -# -# <- {"timestamp": {"seconds": 1541579657, "microseconds": 986760}, -# "event": "RDMA_GID_STATUS_CHANGED", -# "data": -# {"netdev": "bridge0", -# "interface-id": 15880512517475447892, -# "gid-status": true, -# "subnet-prefix": 33022}} -## -{ 'event': 'RDMA_GID_STATUS_CHANGED', - 'data': { 'netdev' : 'str', - 'gid-status' : 'bool', - 'subnet-prefix' : 'uint64', - 'interface-id' : 'uint64' } } diff --git a/contrib/rdmacm-mux/rdmacm-mux.h b/contrib/rdmacm-mux/rdmacm-mu= x.h deleted file mode 100644 index 07a4722913..0000000000 --- a/contrib/rdmacm-mux/rdmacm-mux.h +++ /dev/null @@ -1,61 +0,0 @@ -/* - * QEMU paravirtual RDMA - rdmacm-mux declarations - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef RDMACM_MUX_H -#define RDMACM_MUX_H - -#include "linux/if.h" -#include -#include -#include - -typedef enum RdmaCmMuxMsgType { - RDMACM_MUX_MSG_TYPE_REQ =3D 0, - RDMACM_MUX_MSG_TYPE_RESP =3D 1, -} RdmaCmMuxMsgType; - -typedef enum RdmaCmMuxOpCode { - RDMACM_MUX_OP_CODE_REG =3D 0, - RDMACM_MUX_OP_CODE_UNREG =3D 1, - RDMACM_MUX_OP_CODE_MAD =3D 2, -} RdmaCmMuxOpCode; - -typedef enum RdmaCmMuxErrCode { - RDMACM_MUX_ERR_CODE_OK =3D 0, - RDMACM_MUX_ERR_CODE_EINVAL =3D 1, - RDMACM_MUX_ERR_CODE_EEXIST =3D 2, - RDMACM_MUX_ERR_CODE_EACCES =3D 3, - RDMACM_MUX_ERR_CODE_ENOTFOUND =3D 4, -} RdmaCmMuxErrCode; - -typedef struct RdmaCmMuxHdr { - RdmaCmMuxMsgType msg_type; - RdmaCmMuxOpCode op_code; - union ibv_gid sgid; - RdmaCmMuxErrCode err_code; -} RdmaCmUHdr; - -typedef struct RdmaCmUMad { - struct ib_user_mad hdr; - char mad[RDMA_MAX_PRIVATE_DATA]; -} RdmaCmUMad; - -typedef struct RdmaCmMuxMsg { - RdmaCmUHdr hdr; - int umad_len; - RdmaCmUMad umad; -} RdmaCmMuxMsg; - -#endif diff --git a/hw/rdma/rdma_backend.h b/hw/rdma/rdma_backend.h deleted file mode 100644 index 225af481e0..0000000000 --- a/hw/rdma/rdma_backend.h +++ /dev/null @@ -1,129 +0,0 @@ -/* - * RDMA device: Definitions of Backend Device functions - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef RDMA_BACKEND_H -#define RDMA_BACKEND_H - -#include "qapi/error.h" -#include "chardev/char-fe.h" - -#include "rdma_rm_defs.h" -#include "rdma_backend_defs.h" - -/* Vendor Errors */ -#define VENDOR_ERR_FAIL_BACKEND 0x201 -#define VENDOR_ERR_TOO_MANY_SGES 0x202 -#define VENDOR_ERR_NOMEM 0x203 -#define VENDOR_ERR_QP0 0x204 -#define VENDOR_ERR_INV_NUM_SGE 0x205 -#define VENDOR_ERR_MAD_SEND 0x206 -#define VENDOR_ERR_INVLKEY 0x207 -#define VENDOR_ERR_MR_SMALL 0x208 -#define VENDOR_ERR_INV_MAD_BUFF 0x209 -#define VENDOR_ERR_INV_GID_IDX 0x210 - -/* Add definition for QP0 and QP1 as there is no userspace enums for them = */ -enum ibv_special_qp_type { - IBV_QPT_SMI =3D 0, - IBV_QPT_GSI =3D 1, -}; - -static inline uint32_t rdma_backend_qpn(const RdmaBackendQP *qp) -{ - return qp->ibqp ? qp->ibqp->qp_num : 1; -} - -static inline uint32_t rdma_backend_mr_lkey(const RdmaBackendMR *mr) -{ - return mr->ibmr ? mr->ibmr->lkey : 0; -} - -static inline uint32_t rdma_backend_mr_rkey(const RdmaBackendMR *mr) -{ - return mr->ibmr ? mr->ibmr->rkey : 0; -} - -int rdma_backend_init(RdmaBackendDev *backend_dev, PCIDevice *pdev, - RdmaDeviceResources *rdma_dev_res, - const char *backend_device_name, uint8_t port_num, - struct ibv_device_attr *dev_attr, - CharBackend *mad_chr_be); -void rdma_backend_fini(RdmaBackendDev *backend_dev); -int rdma_backend_add_gid(RdmaBackendDev *backend_dev, const char *ifname, - union ibv_gid *gid); -int rdma_backend_del_gid(RdmaBackendDev *backend_dev, const char *ifname, - union ibv_gid *gid); -int rdma_backend_get_gid_index(RdmaBackendDev *backend_dev, - union ibv_gid *gid); -void rdma_backend_start(RdmaBackendDev *backend_dev); -void rdma_backend_stop(RdmaBackendDev *backend_dev); -void rdma_backend_register_comp_handler(void (*handler)(void *ctx, - struct ibv_wc *wc)= ); -void rdma_backend_unregister_comp_handler(void); - -int rdma_backend_query_port(RdmaBackendDev *backend_dev, - struct ibv_port_attr *port_attr); -int rdma_backend_create_pd(RdmaBackendDev *backend_dev, RdmaBackendPD *pd); -void rdma_backend_destroy_pd(RdmaBackendPD *pd); - -int rdma_backend_create_mr(RdmaBackendMR *mr, RdmaBackendPD *pd, void *add= r, - size_t length, uint64_t guest_start, int access= ); -void rdma_backend_destroy_mr(RdmaBackendMR *mr); - -int rdma_backend_create_cq(RdmaBackendDev *backend_dev, RdmaBackendCQ *cq, - int cqe); -void rdma_backend_destroy_cq(RdmaBackendCQ *cq); -void rdma_backend_poll_cq(RdmaDeviceResources *rdma_dev_res, RdmaBackendCQ= *cq); - -int rdma_backend_create_qp(RdmaBackendQP *qp, uint8_t qp_type, - RdmaBackendPD *pd, RdmaBackendCQ *scq, - RdmaBackendCQ *rcq, RdmaBackendSRQ *srq, - uint32_t max_send_wr, uint32_t max_recv_wr, - uint32_t max_send_sge, uint32_t max_recv_sge); -int rdma_backend_qp_state_init(RdmaBackendDev *backend_dev, RdmaBackendQP = *qp, - uint8_t qp_type, uint32_t qkey); -int rdma_backend_qp_state_rtr(RdmaBackendDev *backend_dev, RdmaBackendQP *= qp, - uint8_t qp_type, uint8_t sgid_idx, - union ibv_gid *dgid, uint32_t dqpn, - uint32_t rq_psn, uint32_t qkey, bool use_qke= y); -int rdma_backend_qp_state_rts(RdmaBackendQP *qp, uint8_t qp_type, - uint32_t sq_psn, uint32_t qkey, bool use_qke= y); -int rdma_backend_query_qp(RdmaBackendQP *qp, struct ibv_qp_attr *attr, - int attr_mask, struct ibv_qp_init_attr *init_att= r); -void rdma_backend_destroy_qp(RdmaBackendQP *qp, RdmaDeviceResources *dev_r= es); - -void rdma_backend_post_send(RdmaBackendDev *backend_dev, - RdmaBackendQP *qp, uint8_t qp_type, - struct ibv_sge *sge, uint32_t num_sge, - uint8_t sgid_idx, union ibv_gid *sgid, - union ibv_gid *dgid, uint32_t dqpn, uint32_t d= qkey, - void *ctx); -void rdma_backend_post_recv(RdmaBackendDev *backend_dev, - RdmaBackendQP *qp, uint8_t qp_type, - struct ibv_sge *sge, uint32_t num_sge, void *c= tx); - -int rdma_backend_create_srq(RdmaBackendSRQ *srq, RdmaBackendPD *pd, - uint32_t max_wr, uint32_t max_sge, - uint32_t srq_limit); -int rdma_backend_query_srq(RdmaBackendSRQ *srq, struct ibv_srq_attr *srq_a= ttr); -int rdma_backend_modify_srq(RdmaBackendSRQ *srq, struct ibv_srq_attr *srq_= attr, - int srq_attr_mask); -void rdma_backend_destroy_srq(RdmaBackendSRQ *srq, - RdmaDeviceResources *dev_res); -void rdma_backend_post_srq_recv(RdmaBackendDev *backend_dev, - RdmaBackendSRQ *srq, struct ibv_sge *sge, - uint32_t num_sge, void *ctx); - -#endif diff --git a/hw/rdma/rdma_backend_defs.h b/hw/rdma/rdma_backend_defs.h deleted file mode 100644 index 4e6c0ad695..0000000000 --- a/hw/rdma/rdma_backend_defs.h +++ /dev/null @@ -1,76 +0,0 @@ -/* - * RDMA device: Definitions of Backend Device structures - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef RDMA_BACKEND_DEFS_H -#define RDMA_BACKEND_DEFS_H - -#include "qemu/thread.h" -#include "chardev/char-fe.h" -#include -#include "contrib/rdmacm-mux/rdmacm-mux.h" -#include "rdma_utils.h" - -typedef struct RdmaDeviceResources RdmaDeviceResources; - -typedef struct RdmaBackendThread { - QemuThread thread; - bool run; /* Set by thread manager to let thread know it should exit */ - bool is_running; /* Set by the thread to report its status */ -} RdmaBackendThread; - -typedef struct RdmaCmMux { - CharBackend *chr_be; - int can_receive; -} RdmaCmMux; - -typedef struct RdmaBackendDev { - RdmaBackendThread comp_thread; - PCIDevice *dev; - RdmaDeviceResources *rdma_dev_res; - struct ibv_device *ib_dev; - struct ibv_context *context; - struct ibv_comp_channel *channel; - uint8_t port_num; - RdmaProtectedGQueue recv_mads_list; - RdmaCmMux rdmacm_mux; -} RdmaBackendDev; - -typedef struct RdmaBackendPD { - struct ibv_pd *ibpd; -} RdmaBackendPD; - -typedef struct RdmaBackendMR { - struct ibv_pd *ibpd; - struct ibv_mr *ibmr; -} RdmaBackendMR; - -typedef struct RdmaBackendCQ { - RdmaBackendDev *backend_dev; - struct ibv_cq *ibcq; -} RdmaBackendCQ; - -typedef struct RdmaBackendQP { - struct ibv_pd *ibpd; - struct ibv_qp *ibqp; - uint8_t sgid_idx; - RdmaProtectedGSList cqe_ctx_list; -} RdmaBackendQP; - -typedef struct RdmaBackendSRQ { - struct ibv_srq *ibsrq; - RdmaProtectedGSList cqe_ctx_list; -} RdmaBackendSRQ; - -#endif diff --git a/hw/rdma/rdma_rm.h b/hw/rdma/rdma_rm.h deleted file mode 100644 index d69a917795..0000000000 --- a/hw/rdma/rdma_rm.h +++ /dev/null @@ -1,97 +0,0 @@ -/* - * RDMA device: Definitions of Resource Manager functions - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef RDMA_RM_H -#define RDMA_RM_H - -#include "qapi/error.h" -#include "rdma_backend_defs.h" -#include "rdma_rm_defs.h" - -int rdma_rm_init(RdmaDeviceResources *dev_res, - struct ibv_device_attr *dev_attr); -void rdma_rm_fini(RdmaDeviceResources *dev_res, RdmaBackendDev *backend_de= v, - const char *ifname); - -int rdma_rm_alloc_pd(RdmaDeviceResources *dev_res, RdmaBackendDev *backend= _dev, - uint32_t *pd_handle, uint32_t ctx_handle); -RdmaRmPD *rdma_rm_get_pd(RdmaDeviceResources *dev_res, uint32_t pd_handle); -void rdma_rm_dealloc_pd(RdmaDeviceResources *dev_res, uint32_t pd_handle); - -int rdma_rm_alloc_mr(RdmaDeviceResources *dev_res, uint32_t pd_handle, - uint64_t guest_start, uint64_t guest_length, - void *host_virt, int access_flags, uint32_t *mr_handl= e, - uint32_t *lkey, uint32_t *rkey); -RdmaRmMR *rdma_rm_get_mr(RdmaDeviceResources *dev_res, uint32_t mr_handle); -void rdma_rm_dealloc_mr(RdmaDeviceResources *dev_res, uint32_t mr_handle); - -int rdma_rm_alloc_uc(RdmaDeviceResources *dev_res, uint32_t pfn, - uint32_t *uc_handle); -RdmaRmUC *rdma_rm_get_uc(RdmaDeviceResources *dev_res, uint32_t uc_handle); -void rdma_rm_dealloc_uc(RdmaDeviceResources *dev_res, uint32_t uc_handle); - -int rdma_rm_alloc_cq(RdmaDeviceResources *dev_res, RdmaBackendDev *backend= _dev, - uint32_t cqe, uint32_t *cq_handle, void *opaque); -RdmaRmCQ *rdma_rm_get_cq(RdmaDeviceResources *dev_res, uint32_t cq_handle); -void rdma_rm_req_notify_cq(RdmaDeviceResources *dev_res, uint32_t cq_handl= e, - bool notify); -void rdma_rm_dealloc_cq(RdmaDeviceResources *dev_res, uint32_t cq_handle); - -int rdma_rm_alloc_qp(RdmaDeviceResources *dev_res, uint32_t pd_handle, - uint8_t qp_type, uint32_t max_send_wr, - uint32_t max_send_sge, uint32_t send_cq_handle, - uint32_t max_recv_wr, uint32_t max_recv_sge, - uint32_t recv_cq_handle, void *opaque, uint32_t *qpn, - uint8_t is_srq, uint32_t srq_handle); -RdmaRmQP *rdma_rm_get_qp(RdmaDeviceResources *dev_res, uint32_t qpn); -int rdma_rm_modify_qp(RdmaDeviceResources *dev_res, RdmaBackendDev *backen= d_dev, - uint32_t qp_handle, uint32_t attr_mask, uint8_t sgid= _idx, - union ibv_gid *dgid, uint32_t dqpn, - enum ibv_qp_state qp_state, uint32_t qkey, - uint32_t rq_psn, uint32_t sq_psn); -int rdma_rm_query_qp(RdmaDeviceResources *dev_res, RdmaBackendDev *backend= _dev, - uint32_t qp_handle, struct ibv_qp_attr *attr, - int attr_mask, struct ibv_qp_init_attr *init_attr); -void rdma_rm_dealloc_qp(RdmaDeviceResources *dev_res, uint32_t qp_handle); - -RdmaRmSRQ *rdma_rm_get_srq(RdmaDeviceResources *dev_res, uint32_t srq_hand= le); -int rdma_rm_alloc_srq(RdmaDeviceResources *dev_res, uint32_t pd_handle, - uint32_t max_wr, uint32_t max_sge, uint32_t srq_limi= t, - uint32_t *srq_handle, void *opaque); -int rdma_rm_query_srq(RdmaDeviceResources *dev_res, uint32_t srq_handle, - struct ibv_srq_attr *srq_attr); -int rdma_rm_modify_srq(RdmaDeviceResources *dev_res, uint32_t srq_handle, - struct ibv_srq_attr *srq_attr, int srq_attr_mask); -void rdma_rm_dealloc_srq(RdmaDeviceResources *dev_res, uint32_t srq_handle= ); - -int rdma_rm_alloc_cqe_ctx(RdmaDeviceResources *dev_res, uint32_t *cqe_ctx_= id, - void *ctx); -void *rdma_rm_get_cqe_ctx(RdmaDeviceResources *dev_res, uint32_t cqe_ctx_i= d); -void rdma_rm_dealloc_cqe_ctx(RdmaDeviceResources *dev_res, uint32_t cqe_ct= x_id); - -int rdma_rm_add_gid(RdmaDeviceResources *dev_res, RdmaBackendDev *backend_= dev, - const char *ifname, union ibv_gid *gid, int gid_idx); -int rdma_rm_del_gid(RdmaDeviceResources *dev_res, RdmaBackendDev *backend_= dev, - const char *ifname, int gid_idx); -int rdma_rm_get_backend_gid_index(RdmaDeviceResources *dev_res, - RdmaBackendDev *backend_dev, int sgid_id= x); -static inline union ibv_gid *rdma_rm_get_gid(RdmaDeviceResources *dev_res, - int sgid_idx) -{ - return &dev_res->port.gid_tbl[sgid_idx].gid; -} -void rdma_format_device_counters(RdmaDeviceResources *dev_res, GString *bu= f); - -#endif diff --git a/hw/rdma/rdma_rm_defs.h b/hw/rdma/rdma_rm_defs.h deleted file mode 100644 index 534f2f74d3..0000000000 --- a/hw/rdma/rdma_rm_defs.h +++ /dev/null @@ -1,146 +0,0 @@ -/* - * RDMA device: Definitions of Resource Manager structures - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef RDMA_RM_DEFS_H -#define RDMA_RM_DEFS_H - -#include "rdma_backend_defs.h" - -#define MAX_PORTS 1 /* Do not change - we support only one por= t */ -#define MAX_PORT_GIDS 255 -#define MAX_GIDS MAX_PORT_GIDS -#define MAX_PORT_PKEYS 1 -#define MAX_PKEYS MAX_PORT_PKEYS -#define MAX_UCS 512 -#define MAX_MR_SIZE (1UL << 27) -#define MAX_QP 1024 -#define MAX_SGE 4 -#define MAX_CQ 2048 -#define MAX_MR 1024 -#define MAX_PD 1024 -#define MAX_QP_RD_ATOM 16 -#define MAX_QP_INIT_RD_ATOM 16 -#define MAX_AH 64 -#define MAX_SRQ 512 - -#define MAX_RM_TBL_NAME 16 -#define MAX_CONSEQ_EMPTY_POLL_CQ 4096 /* considered as error above this= */ - -typedef struct RdmaRmResTbl { - char name[MAX_RM_TBL_NAME]; - QemuMutex lock; - unsigned long *bitmap; - size_t tbl_sz; - size_t res_sz; - void *tbl; - uint32_t used; /* number of used entries in the table */ -} RdmaRmResTbl; - -typedef struct RdmaRmPD { - RdmaBackendPD backend_pd; - uint32_t ctx_handle; -} RdmaRmPD; - -typedef enum CQNotificationType { - CNT_CLEAR, - CNT_ARM, - CNT_SET, -} CQNotificationType; - -typedef struct RdmaRmCQ { - RdmaBackendCQ backend_cq; - void *opaque; - CQNotificationType notify; -} RdmaRmCQ; - -/* MR (DMA region) */ -typedef struct RdmaRmMR { - RdmaBackendMR backend_mr; - void *virt; - uint64_t start; - size_t length; - uint32_t pd_handle; - uint32_t lkey; - uint32_t rkey; -} RdmaRmMR; - -typedef struct RdmaRmUC { - uint64_t uc_handle; -} RdmaRmUC; - -typedef struct RdmaRmQP { - RdmaBackendQP backend_qp; - void *opaque; - uint32_t qp_type; - uint32_t qpn; - uint32_t send_cq_handle; - uint32_t recv_cq_handle; - enum ibv_qp_state qp_state; - uint8_t is_srq; -} RdmaRmQP; - -typedef struct RdmaRmSRQ { - RdmaBackendSRQ backend_srq; - uint32_t recv_cq_handle; - void *opaque; -} RdmaRmSRQ; - -typedef struct RdmaRmGid { - union ibv_gid gid; - int backend_gid_index; -} RdmaRmGid; - -typedef struct RdmaRmPort { - RdmaRmGid gid_tbl[MAX_PORT_GIDS]; - enum ibv_port_state state; -} RdmaRmPort; - -typedef struct RdmaRmStats { - uint64_t tx; - uint64_t tx_len; - uint64_t tx_err; - uint64_t rx_bufs; - uint64_t rx_bufs_len; - uint64_t rx_bufs_err; - uint64_t rx_srq; - uint64_t completions; - uint64_t mad_tx; - uint64_t mad_tx_err; - uint64_t mad_rx; - uint64_t mad_rx_err; - uint64_t mad_rx_bufs; - uint64_t mad_rx_bufs_err; - uint64_t poll_cq_from_bk; - uint64_t poll_cq_from_guest; - uint64_t poll_cq_from_guest_empty; - uint64_t poll_cq_ppoll_to; - uint32_t missing_cqe; -} RdmaRmStats; - -struct RdmaDeviceResources { - RdmaRmPort port; - RdmaRmResTbl pd_tbl; - RdmaRmResTbl mr_tbl; - RdmaRmResTbl uc_tbl; - RdmaRmResTbl qp_tbl; - RdmaRmResTbl cq_tbl; - RdmaRmResTbl cqe_ctx_tbl; - RdmaRmResTbl srq_tbl; - GHashTable *qp_hash; /* Keeps mapping between real and emulated */ - QemuMutex lock; - RdmaRmStats stats; -}; - -#endif diff --git a/hw/rdma/rdma_utils.h b/hw/rdma/rdma_utils.h deleted file mode 100644 index 54e4f56edd..0000000000 --- a/hw/rdma/rdma_utils.h +++ /dev/null @@ -1,63 +0,0 @@ -/* - * RDMA device: Debug utilities - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef RDMA_UTILS_H -#define RDMA_UTILS_H - -#include "qemu/error-report.h" -#include "sysemu/dma.h" - -#define rdma_error_report(fmt, ...) \ - error_report("%s: " fmt, "rdma", ## __VA_ARGS__) -#define rdma_warn_report(fmt, ...) \ - warn_report("%s: " fmt, "rdma", ## __VA_ARGS__) -#define rdma_info_report(fmt, ...) \ - info_report("%s: " fmt, "rdma", ## __VA_ARGS__) - -typedef struct RdmaProtectedGQueue { - QemuMutex lock; - GQueue *list; -} RdmaProtectedGQueue; - -typedef struct RdmaProtectedGSList { - QemuMutex lock; - GSList *list; -} RdmaProtectedGSList; - -void *rdma_pci_dma_map(PCIDevice *dev, dma_addr_t addr, dma_addr_t len); -void rdma_pci_dma_unmap(PCIDevice *dev, void *buffer, dma_addr_t len); -void rdma_protected_gqueue_init(RdmaProtectedGQueue *list); -void rdma_protected_gqueue_destroy(RdmaProtectedGQueue *list); -void rdma_protected_gqueue_append_int64(RdmaProtectedGQueue *list, - int64_t value); -int64_t rdma_protected_gqueue_pop_int64(RdmaProtectedGQueue *list); -void rdma_protected_gslist_init(RdmaProtectedGSList *list); -void rdma_protected_gslist_destroy(RdmaProtectedGSList *list); -void rdma_protected_gslist_append_int32(RdmaProtectedGSList *list, - int32_t value); -void rdma_protected_gslist_remove_int32(RdmaProtectedGSList *list, - int32_t value); - -static inline void addrconf_addr_eui48(uint8_t *eui, const char *addr) -{ - memcpy(eui, addr, 3); - eui[3] =3D 0xFF; - eui[4] =3D 0xFE; - memcpy(eui + 5, addr + 3, 3); - eui[0] ^=3D 2; -} - -#endif diff --git a/hw/rdma/trace.h b/hw/rdma/trace.h deleted file mode 100644 index b3fa8ebc51..0000000000 --- a/hw/rdma/trace.h +++ /dev/null @@ -1 +0,0 @@ -#include "trace/trace-hw_rdma.h" diff --git a/hw/rdma/vmw/pvrdma.h b/hw/rdma/vmw/pvrdma.h deleted file mode 100644 index 4cbc10c980..0000000000 --- a/hw/rdma/vmw/pvrdma.h +++ /dev/null @@ -1,144 +0,0 @@ -/* - * QEMU VMWARE paravirtual RDMA device definitions - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef PVRDMA_PVRDMA_H -#define PVRDMA_PVRDMA_H - -#include "qemu/units.h" -#include "qemu/notify.h" -#include "hw/pci/msix.h" -#include "hw/pci/pci_device.h" -#include "chardev/char-fe.h" -#include "hw/net/vmxnet3_defs.h" - -#include "../rdma_backend_defs.h" -#include "../rdma_rm_defs.h" - -#include "standard-headers/drivers/infiniband/hw/vmw_pvrdma/pvrdma_dev_api= .h" -#include "pvrdma_dev_ring.h" -#include "qom/object.h" - -/* BARs */ -#define RDMA_MSIX_BAR_IDX 0 -#define RDMA_REG_BAR_IDX 1 -#define RDMA_UAR_BAR_IDX 2 -#define RDMA_BAR0_MSIX_SIZE (16 * KiB) -#define RDMA_BAR1_REGS_SIZE 64 -#define RDMA_BAR2_UAR_SIZE (0x1000 * MAX_UCS) /* each uc gets page */ - -/* MSIX */ -#define RDMA_MAX_INTRS 3 -#define RDMA_MSIX_TABLE 0x0000 -#define RDMA_MSIX_PBA 0x2000 - -/* Interrupts Vectors */ -#define INTR_VEC_CMD_RING 0 -#define INTR_VEC_CMD_ASYNC_EVENTS 1 -#define INTR_VEC_CMD_COMPLETION_Q 2 - -/* HW attributes */ -#define PVRDMA_HW_NAME "pvrdma" -#define PVRDMA_HW_VERSION 17 -#define PVRDMA_FW_VERSION 14 - -/* Some defaults */ -#define PVRDMA_PKEY 0xFFFF - -typedef struct DSRInfo { - dma_addr_t dma; - struct pvrdma_device_shared_region *dsr; - - union pvrdma_cmd_req *req; - union pvrdma_cmd_resp *rsp; - - PvrdmaRingState *async_ring_state; - PvrdmaRing async; - - PvrdmaRingState *cq_ring_state; - PvrdmaRing cq; -} DSRInfo; - -typedef struct PVRDMADevStats { - uint64_t commands; - uint64_t regs_reads; - uint64_t regs_writes; - uint64_t uar_writes; - uint64_t interrupts; -} PVRDMADevStats; - -struct PVRDMADev { - PCIDevice parent_obj; - MemoryRegion msix; - MemoryRegion regs; - uint32_t regs_data[RDMA_BAR1_REGS_SIZE]; - MemoryRegion uar; - uint32_t uar_data[RDMA_BAR2_UAR_SIZE]; - DSRInfo dsr_info; - int interrupt_mask; - struct ibv_device_attr dev_attr; - uint64_t node_guid; - char *backend_eth_device_name; - char *backend_device_name; - uint8_t backend_port_num; - RdmaBackendDev backend_dev; - RdmaDeviceResources rdma_dev_res; - CharBackend mad_chr; - VMXNET3State *func0; - Notifier shutdown_notifier; - PVRDMADevStats stats; -}; -typedef struct PVRDMADev PVRDMADev; -DECLARE_INSTANCE_CHECKER(PVRDMADev, PVRDMA_DEV, - PVRDMA_HW_NAME) - -static inline int get_reg_val(PVRDMADev *dev, hwaddr addr, uint32_t *val) -{ - int idx =3D addr >> 2; - - if (idx >=3D RDMA_BAR1_REGS_SIZE) { - return -EINVAL; - } - - *val =3D dev->regs_data[idx]; - - return 0; -} - -static inline int set_reg_val(PVRDMADev *dev, hwaddr addr, uint32_t val) -{ - int idx =3D addr >> 2; - - if (idx >=3D RDMA_BAR1_REGS_SIZE) { - return -EINVAL; - } - - dev->regs_data[idx] =3D val; - - return 0; -} - -static inline void post_interrupt(PVRDMADev *dev, unsigned vector) -{ - PCIDevice *pci_dev =3D PCI_DEVICE(dev); - - if (likely(!dev->interrupt_mask)) { - dev->stats.interrupts++; - msix_notify(pci_dev, vector); - } -} - -int pvrdma_exec_cmd(PVRDMADev *dev); - -#endif diff --git a/hw/rdma/vmw/pvrdma_dev_ring.h b/hw/rdma/vmw/pvrdma_dev_ring.h deleted file mode 100644 index d231588ce0..0000000000 --- a/hw/rdma/vmw/pvrdma_dev_ring.h +++ /dev/null @@ -1,46 +0,0 @@ -/* - * QEMU VMWARE paravirtual RDMA ring utilities - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef PVRDMA_DEV_RING_H -#define PVRDMA_DEV_RING_H - - -#define MAX_RING_NAME_SZ 32 - -typedef struct PvrdmaRingState { - int prod_tail; /* producer tail */ - int cons_head; /* consumer head */ -} PvrdmaRingState; - -typedef struct PvrdmaRing { - char name[MAX_RING_NAME_SZ]; - PCIDevice *dev; - uint32_t max_elems; - size_t elem_sz; - PvrdmaRingState *ring_state; /* used only for unmap */ - int npages; - void **pages; -} PvrdmaRing; - -int pvrdma_ring_init(PvrdmaRing *ring, const char *name, PCIDevice *dev, - PvrdmaRingState *ring_state, uint32_t max_elems, - size_t elem_sz, dma_addr_t *tbl, uint32_t npages); -void *pvrdma_ring_next_elem_read(PvrdmaRing *ring); -void pvrdma_ring_read_inc(PvrdmaRing *ring); -void *pvrdma_ring_next_elem_write(PvrdmaRing *ring); -void pvrdma_ring_write_inc(PvrdmaRing *ring); -void pvrdma_ring_free(PvrdmaRing *ring); - -#endif diff --git a/hw/rdma/vmw/pvrdma_qp_ops.h b/hw/rdma/vmw/pvrdma_qp_ops.h deleted file mode 100644 index bf2b15c5ce..0000000000 --- a/hw/rdma/vmw/pvrdma_qp_ops.h +++ /dev/null @@ -1,28 +0,0 @@ -/* - * QEMU VMWARE paravirtual RDMA QP Operations - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef PVRDMA_QP_OPS_H -#define PVRDMA_QP_OPS_H - -#include "pvrdma.h" - -int pvrdma_qp_ops_init(void); -void pvrdma_qp_ops_fini(void); -void pvrdma_qp_send(PVRDMADev *dev, uint32_t qp_handle); -void pvrdma_qp_recv(PVRDMADev *dev, uint32_t qp_handle); -void pvrdma_srq_recv(PVRDMADev *dev, uint32_t srq_handle); -void pvrdma_cq_poll(RdmaDeviceResources *dev_res, uint32_t cq_handle); - -#endif diff --git a/hw/rdma/vmw/trace.h b/hw/rdma/vmw/trace.h deleted file mode 100644 index 3ebc9fb7ad..0000000000 --- a/hw/rdma/vmw/trace.h +++ /dev/null @@ -1 +0,0 @@ -#include "trace/trace-hw_rdma_vmw.h" diff --git a/include/hw/rdma/rdma.h b/include/hw/rdma/rdma.h deleted file mode 100644 index 80b2e531c4..0000000000 --- a/include/hw/rdma/rdma.h +++ /dev/null @@ -1,37 +0,0 @@ -/* - * RDMA device interface - * - * Copyright (C) 2019 Oracle - * Copyright (C) 2019 Red Hat Inc - * - * Authors: - * Yuval Shaia - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#ifndef RDMA_H -#define RDMA_H - -#include "qom/object.h" - -#define INTERFACE_RDMA_PROVIDER "rdma" - -typedef struct RdmaProviderClass RdmaProviderClass; -DECLARE_CLASS_CHECKERS(RdmaProviderClass, RDMA_PROVIDER, - INTERFACE_RDMA_PROVIDER) -#define RDMA_PROVIDER(obj) \ - INTERFACE_CHECK(RdmaProvider, (obj), \ - INTERFACE_RDMA_PROVIDER) - -typedef struct RdmaProvider RdmaProvider; - -struct RdmaProviderClass { - InterfaceClass parent; - - void (*format_statistics)(RdmaProvider *obj, GString *buf); -}; - -#endif diff --git a/include/monitor/hmp.h b/include/monitor/hmp.h index 13f9a2dedb..f4cf8f6717 100644 --- a/include/monitor/hmp.h +++ b/include/monitor/hmp.h @@ -37,7 +37,6 @@ void hmp_info_spice(Monitor *mon, const QDict *qdict); void hmp_info_balloon(Monitor *mon, const QDict *qdict); void hmp_info_irq(Monitor *mon, const QDict *qdict); void hmp_info_pic(Monitor *mon, const QDict *qdict); -void hmp_info_rdma(Monitor *mon, const QDict *qdict); void hmp_info_pci(Monitor *mon, const QDict *qdict); void hmp_info_tpm(Monitor *mon, const QDict *qdict); void hmp_info_iothreads(Monitor *mon, const QDict *qdict); diff --git a/include/standard-headers/drivers/infiniband/hw/vmw_pvrdma/pvrd= ma_dev_api.h b/include/standard-headers/drivers/infiniband/hw/vmw_pvrdma/pv= rdma_dev_api.h deleted file mode 100644 index a5a1c8234e..0000000000 --- a/include/standard-headers/drivers/infiniband/hw/vmw_pvrdma/pvrdma_dev_= api.h +++ /dev/null @@ -1,685 +0,0 @@ -/* - * Copyright (c) 2012-2016 VMware, Inc. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of EITHER the GNU General Public License - * version 2 as published by the Free Software Foundation or the BSD - * 2-Clause License. This program is distributed in the hope that it - * will be useful, but WITHOUT ANY WARRANTY; WITHOUT EVEN THE IMPLIED - * WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. - * See the GNU General Public License version 2 for more details at - * http://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html. - * - * You should have received a copy of the GNU General Public License - * along with this program available in the file COPYING in the main - * directory of this source tree. - * - * The BSD 2-Clause License - * - * Redistribution and use in source and binary forms, with or - * without modification, are permitted provided that the following - * conditions are met: - * - * - Redistributions of source code must retain the above - * copyright notice, this list of conditions and the following - * disclaimer. - * - * - Redistributions in binary form must reproduce the above - * copyright notice, this list of conditions and the following - * disclaimer in the documentation and/or other materials - * provided with the distribution. - * - * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS - * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT - * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS - * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE - * COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, - * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES - * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR - * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) - * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, - * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) - * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED - * OF THE POSSIBILITY OF SUCH DAMAGE. - */ - -#ifndef __PVRDMA_DEV_API_H__ -#define __PVRDMA_DEV_API_H__ - -#include "standard-headers/linux/types.h" - -#include "pvrdma_verbs.h" - -/* - * PVRDMA version macros. Some new features require updates to PVRDMA_VERS= ION. - * These macros allow us to check for different features if necessary. - */ - -#define PVRDMA_ROCEV1_VERSION 17 -#define PVRDMA_ROCEV2_VERSION 18 -#define PVRDMA_PPN64_VERSION 19 -#define PVRDMA_QPHANDLE_VERSION 20 -#define PVRDMA_VERSION PVRDMA_QPHANDLE_VERSION - -#define PVRDMA_BOARD_ID 1 -#define PVRDMA_REV_ID 1 - -/* - * Masks and accessors for page directory, which is a two-level lookup: - * page directory -> page table -> page. Only one directory for now, but we - * could expand that easily. 9 bits for tables, 9 bits for pages, gives one - * gigabyte for memory regions and so forth. - */ - -#define PVRDMA_PDIR_SHIFT 18 -#define PVRDMA_PTABLE_SHIFT 9 -#define PVRDMA_PAGE_DIR_DIR(x) (((x) >> PVRDMA_PDIR_SHIFT) & 0x1) -#define PVRDMA_PAGE_DIR_TABLE(x) (((x) >> PVRDMA_PTABLE_SHIFT) & 0x1ff) -#define PVRDMA_PAGE_DIR_PAGE(x) ((x) & 0x1ff) -#define PVRDMA_PAGE_DIR_MAX_PAGES (1 * 512 * 512) -#define PVRDMA_MAX_FAST_REG_PAGES 128 - -/* - * Max MSI-X vectors. - */ - -#define PVRDMA_MAX_INTERRUPTS 3 - -/* Register offsets within PCI resource on BAR1. */ -#define PVRDMA_REG_VERSION 0x00 /* R: Version of device. */ -#define PVRDMA_REG_DSRLOW 0x04 /* W: Device shared region low PA. */ -#define PVRDMA_REG_DSRHIGH 0x08 /* W: Device shared region high PA. */ -#define PVRDMA_REG_CTL 0x0c /* W: PVRDMA_DEVICE_CTL */ -#define PVRDMA_REG_REQUEST 0x10 /* W: Indicate device request. */ -#define PVRDMA_REG_ERR 0x14 /* R: Device error. */ -#define PVRDMA_REG_ICR 0x18 /* R: Interrupt cause. */ -#define PVRDMA_REG_IMR 0x1c /* R/W: Interrupt mask. */ -#define PVRDMA_REG_MACL 0x20 /* R/W: MAC address low. */ -#define PVRDMA_REG_MACH 0x24 /* R/W: MAC address high. */ - -/* Object flags. */ -#define PVRDMA_CQ_FLAG_ARMED_SOL BIT(0) /* Armed for solicited-only. */ -#define PVRDMA_CQ_FLAG_ARMED BIT(1) /* Armed. */ -#define PVRDMA_MR_FLAG_DMA BIT(0) /* DMA region. */ -#define PVRDMA_MR_FLAG_FRMR BIT(1) /* Fast reg memory region. */ - -/* - * Atomic operation capability (masked versions are extended atomic - * operations. - */ - -#define PVRDMA_ATOMIC_OP_COMP_SWAP BIT(0) /* Compare and swap. */ -#define PVRDMA_ATOMIC_OP_FETCH_ADD BIT(1) /* Fetch and add. */ -#define PVRDMA_ATOMIC_OP_MASK_COMP_SWAP BIT(2) /* Masked compare and swap.= */ -#define PVRDMA_ATOMIC_OP_MASK_FETCH_ADD BIT(3) /* Masked fetch and add. */ - -/* - * Base Memory Management Extension flags to support Fast Reg Memory Regio= ns - * and Fast Reg Work Requests. Each flag represents a verb operation and we - * must support all of them to qualify for the BMME device cap. - */ - -#define PVRDMA_BMME_FLAG_LOCAL_INV BIT(0) /* Local Invalidate. */ -#define PVRDMA_BMME_FLAG_REMOTE_INV BIT(1) /* Remote Invalidate. */ -#define PVRDMA_BMME_FLAG_FAST_REG_WR BIT(2) /* Fast Reg Work Request. */ - -/* - * GID types. The interpretation of the gid_types bit field in the device - * capabilities will depend on the device mode. For now, the device only - * supports RoCE as mode, so only the different GID types for RoCE are - * defined. - */ - -#define PVRDMA_GID_TYPE_FLAG_ROCE_V1 BIT(0) -#define PVRDMA_GID_TYPE_FLAG_ROCE_V2 BIT(1) - -/* - * Version checks. This checks whether each version supports specific - * capabilities from the device. - */ - -#define PVRDMA_IS_VERSION17(_dev) \ - (_dev->dsr_version =3D=3D PVRDMA_ROCEV1_VERSION && \ - _dev->dsr->caps.gid_types =3D=3D PVRDMA_GID_TYPE_FLAG_ROCE_V1) - -#define PVRDMA_IS_VERSION18(_dev) \ - (_dev->dsr_version >=3D PVRDMA_ROCEV2_VERSION && \ - (_dev->dsr->caps.gid_types =3D=3D PVRDMA_GID_TYPE_FLAG_ROCE_V1 || \ - _dev->dsr->caps.gid_types =3D=3D PVRDMA_GID_TYPE_FLAG_ROCE_V2)) \ - -#define PVRDMA_SUPPORTED(_dev) \ - ((_dev->dsr->caps.mode =3D=3D PVRDMA_DEVICE_MODE_ROCE) && \ - (PVRDMA_IS_VERSION17(_dev) || PVRDMA_IS_VERSION18(_dev))) - -/* - * Get capability values based on device version. - */ - -#define PVRDMA_GET_CAP(_dev, _old_val, _val) \ - ((PVRDMA_IS_VERSION18(_dev)) ? _val : _old_val) - -enum pvrdma_pci_resource { - PVRDMA_PCI_RESOURCE_MSIX, /* BAR0: MSI-X, MMIO. */ - PVRDMA_PCI_RESOURCE_REG, /* BAR1: Registers, MMIO. */ - PVRDMA_PCI_RESOURCE_UAR, /* BAR2: UAR pages, MMIO, 64-bit. */ - PVRDMA_PCI_RESOURCE_LAST, /* Last. */ -}; - -enum pvrdma_device_ctl { - PVRDMA_DEVICE_CTL_ACTIVATE, /* Activate device. */ - PVRDMA_DEVICE_CTL_UNQUIESCE, /* Unquiesce device. */ - PVRDMA_DEVICE_CTL_RESET, /* Reset device. */ -}; - -enum pvrdma_intr_vector { - PVRDMA_INTR_VECTOR_RESPONSE, /* Command response. */ - PVRDMA_INTR_VECTOR_ASYNC, /* Async events. */ - PVRDMA_INTR_VECTOR_CQ, /* CQ notification. */ - /* Additional CQ notification vectors. */ -}; - -enum pvrdma_intr_cause { - PVRDMA_INTR_CAUSE_RESPONSE =3D (1 << PVRDMA_INTR_VECTOR_RESPONSE), - PVRDMA_INTR_CAUSE_ASYNC =3D (1 << PVRDMA_INTR_VECTOR_ASYNC), - PVRDMA_INTR_CAUSE_CQ =3D (1 << PVRDMA_INTR_VECTOR_CQ), -}; - -enum pvrdma_gos_bits { - PVRDMA_GOS_BITS_UNK, /* Unknown. */ - PVRDMA_GOS_BITS_32, /* 32-bit. */ - PVRDMA_GOS_BITS_64, /* 64-bit. */ -}; - -enum pvrdma_gos_type { - PVRDMA_GOS_TYPE_UNK, /* Unknown. */ - PVRDMA_GOS_TYPE_LINUX, /* Linux. */ -}; - -enum pvrdma_device_mode { - PVRDMA_DEVICE_MODE_ROCE, /* RoCE. */ - PVRDMA_DEVICE_MODE_IWARP, /* iWarp. */ - PVRDMA_DEVICE_MODE_IB, /* InfiniBand. */ -}; - -struct pvrdma_gos_info { - uint32_t gos_bits:2; /* W: PVRDMA_GOS_BITS_ */ - uint32_t gos_type:4; /* W: PVRDMA_GOS_TYPE_ */ - uint32_t gos_ver:16; /* W: Guest OS version. */ - uint32_t gos_misc:10; /* W: Other. */ - uint32_t pad; /* Pad to 8-byte alignment. */ -}; - -struct pvrdma_device_caps { - uint64_t fw_ver; /* R: Query device. */ - uint64_t node_guid; - uint64_t sys_image_guid; - uint64_t max_mr_size; - uint64_t page_size_cap; - uint64_t atomic_arg_sizes; /* EX verbs. */ - uint32_t ex_comp_mask; /* EX verbs. */ - uint32_t device_cap_flags2; /* EX verbs. */ - uint32_t max_fa_bit_boundary; /* EX verbs. */ - uint32_t log_max_atomic_inline_arg; /* EX verbs. */ - uint32_t vendor_id; - uint32_t vendor_part_id; - uint32_t hw_ver; - uint32_t max_qp; - uint32_t max_qp_wr; - uint32_t device_cap_flags; - uint32_t max_sge; - uint32_t max_sge_rd; - uint32_t max_cq; - uint32_t max_cqe; - uint32_t max_mr; - uint32_t max_pd; - uint32_t max_qp_rd_atom; - uint32_t max_ee_rd_atom; - uint32_t max_res_rd_atom; - uint32_t max_qp_init_rd_atom; - uint32_t max_ee_init_rd_atom; - uint32_t max_ee; - uint32_t max_rdd; - uint32_t max_mw; - uint32_t max_raw_ipv6_qp; - uint32_t max_raw_ethy_qp; - uint32_t max_mcast_grp; - uint32_t max_mcast_qp_attach; - uint32_t max_total_mcast_qp_attach; - uint32_t max_ah; - uint32_t max_fmr; - uint32_t max_map_per_fmr; - uint32_t max_srq; - uint32_t max_srq_wr; - uint32_t max_srq_sge; - uint32_t max_uar; - uint32_t gid_tbl_len; - uint16_t max_pkeys; - uint8_t local_ca_ack_delay; - uint8_t phys_port_cnt; - uint8_t mode; /* PVRDMA_DEVICE_MODE_ */ - uint8_t atomic_ops; /* PVRDMA_ATOMIC_OP_* bits */ - uint8_t bmme_flags; /* FRWR Mem Mgmt Extensions */ - uint8_t gid_types; /* PVRDMA_GID_TYPE_FLAG_ */ - uint32_t max_fast_reg_page_list_len; -}; - -struct pvrdma_ring_page_info { - uint32_t num_pages; /* Num pages incl. header. */ - uint32_t reserved; /* Reserved. */ - uint64_t pdir_dma; /* Page directory PA. */ -}; - -#pragma pack(push, 1) - -struct pvrdma_device_shared_region { - uint32_t driver_version; /* W: Driver version. */ - uint32_t pad; /* Pad to 8-byte align. */ - struct pvrdma_gos_info gos_info; /* W: Guest OS information. */ - uint64_t cmd_slot_dma; /* W: Command slot address. */ - uint64_t resp_slot_dma; /* W: Response slot address. */ - struct pvrdma_ring_page_info async_ring_pages; - /* W: Async ring page info. */ - struct pvrdma_ring_page_info cq_ring_pages; - /* W: CQ ring page info. */ - union { - uint32_t uar_pfn; /* W: UAR pageframe. */ - uint64_t uar_pfn64; /* W: 64-bit UAR page frame. */ - }; - struct pvrdma_device_caps caps; /* R: Device capabilities. */ -}; - -#pragma pack(pop) - -/* Event types. Currently a 1:1 mapping with enum ib_event. */ -enum pvrdma_eqe_type { - PVRDMA_EVENT_CQ_ERR, - PVRDMA_EVENT_QP_FATAL, - PVRDMA_EVENT_QP_REQ_ERR, - PVRDMA_EVENT_QP_ACCESS_ERR, - PVRDMA_EVENT_COMM_EST, - PVRDMA_EVENT_SQ_DRAINED, - PVRDMA_EVENT_PATH_MIG, - PVRDMA_EVENT_PATH_MIG_ERR, - PVRDMA_EVENT_DEVICE_FATAL, - PVRDMA_EVENT_PORT_ACTIVE, - PVRDMA_EVENT_PORT_ERR, - PVRDMA_EVENT_LID_CHANGE, - PVRDMA_EVENT_PKEY_CHANGE, - PVRDMA_EVENT_SM_CHANGE, - PVRDMA_EVENT_SRQ_ERR, - PVRDMA_EVENT_SRQ_LIMIT_REACHED, - PVRDMA_EVENT_QP_LAST_WQE_REACHED, - PVRDMA_EVENT_CLIENT_REREGISTER, - PVRDMA_EVENT_GID_CHANGE, -}; - -/* Event queue element. */ -struct pvrdma_eqe { - uint32_t type; /* Event type. */ - uint32_t info; /* Handle, other. */ -}; - -/* CQ notification queue element. */ -struct pvrdma_cqne { - uint32_t info; /* Handle */ -}; - -enum { - PVRDMA_CMD_FIRST, - PVRDMA_CMD_QUERY_PORT =3D PVRDMA_CMD_FIRST, - PVRDMA_CMD_QUERY_PKEY, - PVRDMA_CMD_CREATE_PD, - PVRDMA_CMD_DESTROY_PD, - PVRDMA_CMD_CREATE_MR, - PVRDMA_CMD_DESTROY_MR, - PVRDMA_CMD_CREATE_CQ, - PVRDMA_CMD_RESIZE_CQ, - PVRDMA_CMD_DESTROY_CQ, - PVRDMA_CMD_CREATE_QP, - PVRDMA_CMD_MODIFY_QP, - PVRDMA_CMD_QUERY_QP, - PVRDMA_CMD_DESTROY_QP, - PVRDMA_CMD_CREATE_UC, - PVRDMA_CMD_DESTROY_UC, - PVRDMA_CMD_CREATE_BIND, - PVRDMA_CMD_DESTROY_BIND, - PVRDMA_CMD_CREATE_SRQ, - PVRDMA_CMD_MODIFY_SRQ, - PVRDMA_CMD_QUERY_SRQ, - PVRDMA_CMD_DESTROY_SRQ, - PVRDMA_CMD_MAX, -}; - -enum { - PVRDMA_CMD_FIRST_RESP =3D (1 << 31), - PVRDMA_CMD_QUERY_PORT_RESP =3D PVRDMA_CMD_FIRST_RESP, - PVRDMA_CMD_QUERY_PKEY_RESP, - PVRDMA_CMD_CREATE_PD_RESP, - PVRDMA_CMD_DESTROY_PD_RESP_NOOP, - PVRDMA_CMD_CREATE_MR_RESP, - PVRDMA_CMD_DESTROY_MR_RESP_NOOP, - PVRDMA_CMD_CREATE_CQ_RESP, - PVRDMA_CMD_RESIZE_CQ_RESP, - PVRDMA_CMD_DESTROY_CQ_RESP_NOOP, - PVRDMA_CMD_CREATE_QP_RESP, - PVRDMA_CMD_MODIFY_QP_RESP, - PVRDMA_CMD_QUERY_QP_RESP, - PVRDMA_CMD_DESTROY_QP_RESP, - PVRDMA_CMD_CREATE_UC_RESP, - PVRDMA_CMD_DESTROY_UC_RESP_NOOP, - PVRDMA_CMD_CREATE_BIND_RESP_NOOP, - PVRDMA_CMD_DESTROY_BIND_RESP_NOOP, - PVRDMA_CMD_CREATE_SRQ_RESP, - PVRDMA_CMD_MODIFY_SRQ_RESP, - PVRDMA_CMD_QUERY_SRQ_RESP, - PVRDMA_CMD_DESTROY_SRQ_RESP, - PVRDMA_CMD_MAX_RESP, -}; - -struct pvrdma_cmd_hdr { - uint64_t response; /* Key for response lookup. */ - uint32_t cmd; /* PVRDMA_CMD_ */ - uint32_t reserved; /* Reserved. */ -}; - -struct pvrdma_cmd_resp_hdr { - uint64_t response; /* From cmd hdr. */ - uint32_t ack; /* PVRDMA_CMD_XXX_RESP */ - uint8_t err; /* Error. */ - uint8_t reserved[3]; /* Reserved. */ -}; - -struct pvrdma_cmd_query_port { - struct pvrdma_cmd_hdr hdr; - uint8_t port_num; - uint8_t reserved[7]; -}; - -struct pvrdma_cmd_query_port_resp { - struct pvrdma_cmd_resp_hdr hdr; - struct pvrdma_port_attr attrs; -}; - -struct pvrdma_cmd_query_pkey { - struct pvrdma_cmd_hdr hdr; - uint8_t port_num; - uint8_t index; - uint8_t reserved[6]; -}; - -struct pvrdma_cmd_query_pkey_resp { - struct pvrdma_cmd_resp_hdr hdr; - uint16_t pkey; - uint8_t reserved[6]; -}; - -struct pvrdma_cmd_create_uc { - struct pvrdma_cmd_hdr hdr; - union { - uint32_t pfn; /* UAR page frame number */ - uint64_t pfn64; /* 64-bit UAR page frame number */ - }; -}; - -struct pvrdma_cmd_create_uc_resp { - struct pvrdma_cmd_resp_hdr hdr; - uint32_t ctx_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_destroy_uc { - struct pvrdma_cmd_hdr hdr; - uint32_t ctx_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_create_pd { - struct pvrdma_cmd_hdr hdr; - uint32_t ctx_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_create_pd_resp { - struct pvrdma_cmd_resp_hdr hdr; - uint32_t pd_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_destroy_pd { - struct pvrdma_cmd_hdr hdr; - uint32_t pd_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_create_mr { - struct pvrdma_cmd_hdr hdr; - uint64_t start; - uint64_t length; - uint64_t pdir_dma; - uint32_t pd_handle; - uint32_t access_flags; - uint32_t flags; - uint32_t nchunks; -}; - -struct pvrdma_cmd_create_mr_resp { - struct pvrdma_cmd_resp_hdr hdr; - uint32_t mr_handle; - uint32_t lkey; - uint32_t rkey; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_destroy_mr { - struct pvrdma_cmd_hdr hdr; - uint32_t mr_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_create_cq { - struct pvrdma_cmd_hdr hdr; - uint64_t pdir_dma; - uint32_t ctx_handle; - uint32_t cqe; - uint32_t nchunks; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_create_cq_resp { - struct pvrdma_cmd_resp_hdr hdr; - uint32_t cq_handle; - uint32_t cqe; -}; - -struct pvrdma_cmd_resize_cq { - struct pvrdma_cmd_hdr hdr; - uint32_t cq_handle; - uint32_t cqe; -}; - -struct pvrdma_cmd_resize_cq_resp { - struct pvrdma_cmd_resp_hdr hdr; - uint32_t cqe; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_destroy_cq { - struct pvrdma_cmd_hdr hdr; - uint32_t cq_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_create_srq { - struct pvrdma_cmd_hdr hdr; - uint64_t pdir_dma; - uint32_t pd_handle; - uint32_t nchunks; - struct pvrdma_srq_attr attrs; - uint8_t srq_type; - uint8_t reserved[7]; -}; - -struct pvrdma_cmd_create_srq_resp { - struct pvrdma_cmd_resp_hdr hdr; - uint32_t srqn; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_modify_srq { - struct pvrdma_cmd_hdr hdr; - uint32_t srq_handle; - uint32_t attr_mask; - struct pvrdma_srq_attr attrs; -}; - -struct pvrdma_cmd_query_srq { - struct pvrdma_cmd_hdr hdr; - uint32_t srq_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_query_srq_resp { - struct pvrdma_cmd_resp_hdr hdr; - struct pvrdma_srq_attr attrs; -}; - -struct pvrdma_cmd_destroy_srq { - struct pvrdma_cmd_hdr hdr; - uint32_t srq_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_create_qp { - struct pvrdma_cmd_hdr hdr; - uint64_t pdir_dma; - uint32_t pd_handle; - uint32_t send_cq_handle; - uint32_t recv_cq_handle; - uint32_t srq_handle; - uint32_t max_send_wr; - uint32_t max_recv_wr; - uint32_t max_send_sge; - uint32_t max_recv_sge; - uint32_t max_inline_data; - uint32_t lkey; - uint32_t access_flags; - uint16_t total_chunks; - uint16_t send_chunks; - uint16_t max_atomic_arg; - uint8_t sq_sig_all; - uint8_t qp_type; - uint8_t is_srq; - uint8_t reserved[3]; -}; - -struct pvrdma_cmd_create_qp_resp { - struct pvrdma_cmd_resp_hdr hdr; - uint32_t qpn; - uint32_t max_send_wr; - uint32_t max_recv_wr; - uint32_t max_send_sge; - uint32_t max_recv_sge; - uint32_t max_inline_data; -}; - -struct pvrdma_cmd_create_qp_resp_v2 { - struct pvrdma_cmd_resp_hdr hdr; - uint32_t qpn; - uint32_t qp_handle; - uint32_t max_send_wr; - uint32_t max_recv_wr; - uint32_t max_send_sge; - uint32_t max_recv_sge; - uint32_t max_inline_data; -}; - -struct pvrdma_cmd_modify_qp { - struct pvrdma_cmd_hdr hdr; - uint32_t qp_handle; - uint32_t attr_mask; - struct pvrdma_qp_attr attrs; -}; - -struct pvrdma_cmd_query_qp { - struct pvrdma_cmd_hdr hdr; - uint32_t qp_handle; - uint32_t attr_mask; -}; - -struct pvrdma_cmd_query_qp_resp { - struct pvrdma_cmd_resp_hdr hdr; - struct pvrdma_qp_attr attrs; -}; - -struct pvrdma_cmd_destroy_qp { - struct pvrdma_cmd_hdr hdr; - uint32_t qp_handle; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_destroy_qp_resp { - struct pvrdma_cmd_resp_hdr hdr; - uint32_t events_reported; - uint8_t reserved[4]; -}; - -struct pvrdma_cmd_create_bind { - struct pvrdma_cmd_hdr hdr; - uint32_t mtu; - uint32_t vlan; - uint32_t index; - uint8_t new_gid[16]; - uint8_t gid_type; - uint8_t reserved[3]; -}; - -struct pvrdma_cmd_destroy_bind { - struct pvrdma_cmd_hdr hdr; - uint32_t index; - uint8_t dest_gid[16]; - uint8_t reserved[4]; -}; - -union pvrdma_cmd_req { - struct pvrdma_cmd_hdr hdr; - struct pvrdma_cmd_query_port query_port; - struct pvrdma_cmd_query_pkey query_pkey; - struct pvrdma_cmd_create_uc create_uc; - struct pvrdma_cmd_destroy_uc destroy_uc; - struct pvrdma_cmd_create_pd create_pd; - struct pvrdma_cmd_destroy_pd destroy_pd; - struct pvrdma_cmd_create_mr create_mr; - struct pvrdma_cmd_destroy_mr destroy_mr; - struct pvrdma_cmd_create_cq create_cq; - struct pvrdma_cmd_resize_cq resize_cq; - struct pvrdma_cmd_destroy_cq destroy_cq; - struct pvrdma_cmd_create_qp create_qp; - struct pvrdma_cmd_modify_qp modify_qp; - struct pvrdma_cmd_query_qp query_qp; - struct pvrdma_cmd_destroy_qp destroy_qp; - struct pvrdma_cmd_create_bind create_bind; - struct pvrdma_cmd_destroy_bind destroy_bind; - struct pvrdma_cmd_create_srq create_srq; - struct pvrdma_cmd_modify_srq modify_srq; - struct pvrdma_cmd_query_srq query_srq; - struct pvrdma_cmd_destroy_srq destroy_srq; -}; - -union pvrdma_cmd_resp { - struct pvrdma_cmd_resp_hdr hdr; - struct pvrdma_cmd_query_port_resp query_port_resp; - struct pvrdma_cmd_query_pkey_resp query_pkey_resp; - struct pvrdma_cmd_create_uc_resp create_uc_resp; - struct pvrdma_cmd_create_pd_resp create_pd_resp; - struct pvrdma_cmd_create_mr_resp create_mr_resp; - struct pvrdma_cmd_create_cq_resp create_cq_resp; - struct pvrdma_cmd_resize_cq_resp resize_cq_resp; - struct pvrdma_cmd_create_qp_resp create_qp_resp; - struct pvrdma_cmd_create_qp_resp_v2 create_qp_resp_v2; - struct pvrdma_cmd_query_qp_resp query_qp_resp; - struct pvrdma_cmd_destroy_qp_resp destroy_qp_resp; - struct pvrdma_cmd_create_srq_resp create_srq_resp; - struct pvrdma_cmd_query_srq_resp query_srq_resp; -}; - -#endif /* __PVRDMA_DEV_API_H__ */ diff --git a/include/standard-headers/drivers/infiniband/hw/vmw_pvrdma/pvrd= ma_verbs.h b/include/standard-headers/drivers/infiniband/hw/vmw_pvrdma/pvrd= ma_verbs.h deleted file mode 100644 index 94d41b202c..0000000000 --- a/include/standard-headers/drivers/infiniband/hw/vmw_pvrdma/pvrdma_verb= s.h +++ /dev/null @@ -1,348 +0,0 @@ -/* - * Copyright (c) 2012-2016 VMware, Inc. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of EITHER the GNU General Public License - * version 2 as published by the Free Software Foundation or the BSD - * 2-Clause License. This program is distributed in the hope that it - * will be useful, but WITHOUT ANY WARRANTY; WITHOUT EVEN THE IMPLIED - * WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. - * See the GNU General Public License version 2 for more details at - * http://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html. - * - * You should have received a copy of the GNU General Public License - * along with this program available in the file COPYING in the main - * directory of this source tree. - * - * The BSD 2-Clause License - * - * Redistribution and use in source and binary forms, with or - * without modification, are permitted provided that the following - * conditions are met: - * - * - Redistributions of source code must retain the above - * copyright notice, this list of conditions and the following - * disclaimer. - * - * - Redistributions in binary form must reproduce the above - * copyright notice, this list of conditions and the following - * disclaimer in the documentation and/or other materials - * provided with the distribution. - * - * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS - * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT - * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS - * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE - * COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, - * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES - * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR - * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) - * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, - * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) - * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED - * OF THE POSSIBILITY OF SUCH DAMAGE. - */ - -#ifndef __PVRDMA_VERBS_H__ -#define __PVRDMA_VERBS_H__ - -#include "standard-headers/linux/types.h" - -union pvrdma_gid { - uint8_t raw[16]; - struct { - uint64_t subnet_prefix; - uint64_t interface_id; - } global; -}; - -enum pvrdma_link_layer { - PVRDMA_LINK_LAYER_UNSPECIFIED, - PVRDMA_LINK_LAYER_INFINIBAND, - PVRDMA_LINK_LAYER_ETHERNET, -}; - -enum pvrdma_mtu { - PVRDMA_MTU_256 =3D 1, - PVRDMA_MTU_512 =3D 2, - PVRDMA_MTU_1024 =3D 3, - PVRDMA_MTU_2048 =3D 4, - PVRDMA_MTU_4096 =3D 5, -}; - -enum pvrdma_port_state { - PVRDMA_PORT_NOP =3D 0, - PVRDMA_PORT_DOWN =3D 1, - PVRDMA_PORT_INIT =3D 2, - PVRDMA_PORT_ARMED =3D 3, - PVRDMA_PORT_ACTIVE =3D 4, - PVRDMA_PORT_ACTIVE_DEFER =3D 5, -}; - -enum pvrdma_port_cap_flags { - PVRDMA_PORT_SM =3D 1 << 1, - PVRDMA_PORT_NOTICE_SUP =3D 1 << 2, - PVRDMA_PORT_TRAP_SUP =3D 1 << 3, - PVRDMA_PORT_OPT_IPD_SUP =3D 1 << 4, - PVRDMA_PORT_AUTO_MIGR_SUP =3D 1 << 5, - PVRDMA_PORT_SL_MAP_SUP =3D 1 << 6, - PVRDMA_PORT_MKEY_NVRAM =3D 1 << 7, - PVRDMA_PORT_PKEY_NVRAM =3D 1 << 8, - PVRDMA_PORT_LED_INFO_SUP =3D 1 << 9, - PVRDMA_PORT_SM_DISABLED =3D 1 << 10, - PVRDMA_PORT_SYS_IMAGE_GUID_SUP =3D 1 << 11, - PVRDMA_PORT_PKEY_SW_EXT_PORT_TRAP_SUP =3D 1 << 12, - PVRDMA_PORT_EXTENDED_SPEEDS_SUP =3D 1 << 14, - PVRDMA_PORT_CM_SUP =3D 1 << 16, - PVRDMA_PORT_SNMP_TUNNEL_SUP =3D 1 << 17, - PVRDMA_PORT_REINIT_SUP =3D 1 << 18, - PVRDMA_PORT_DEVICE_MGMT_SUP =3D 1 << 19, - PVRDMA_PORT_VENDOR_CLASS_SUP =3D 1 << 20, - PVRDMA_PORT_DR_NOTICE_SUP =3D 1 << 21, - PVRDMA_PORT_CAP_MASK_NOTICE_SUP =3D 1 << 22, - PVRDMA_PORT_BOOT_MGMT_SUP =3D 1 << 23, - PVRDMA_PORT_LINK_LATENCY_SUP =3D 1 << 24, - PVRDMA_PORT_CLIENT_REG_SUP =3D 1 << 25, - PVRDMA_PORT_IP_BASED_GIDS =3D 1 << 26, - PVRDMA_PORT_CAP_FLAGS_MAX =3D PVRDMA_PORT_IP_BASED_GIDS, -}; - -enum pvrdma_port_width { - PVRDMA_WIDTH_1X =3D 1, - PVRDMA_WIDTH_4X =3D 2, - PVRDMA_WIDTH_8X =3D 4, - PVRDMA_WIDTH_12X =3D 8, -}; - -enum pvrdma_port_speed { - PVRDMA_SPEED_SDR =3D 1, - PVRDMA_SPEED_DDR =3D 2, - PVRDMA_SPEED_QDR =3D 4, - PVRDMA_SPEED_FDR10 =3D 8, - PVRDMA_SPEED_FDR =3D 16, - PVRDMA_SPEED_EDR =3D 32, -}; - -struct pvrdma_port_attr { - enum pvrdma_port_state state; - enum pvrdma_mtu max_mtu; - enum pvrdma_mtu active_mtu; - uint32_t gid_tbl_len; - uint32_t port_cap_flags; - uint32_t max_msg_sz; - uint32_t bad_pkey_cntr; - uint32_t qkey_viol_cntr; - uint16_t pkey_tbl_len; - uint16_t lid; - uint16_t sm_lid; - uint8_t lmc; - uint8_t max_vl_num; - uint8_t sm_sl; - uint8_t subnet_timeout; - uint8_t init_type_reply; - uint8_t active_width; - uint8_t active_speed; - uint8_t phys_state; - uint8_t reserved[2]; -}; - -struct pvrdma_global_route { - union pvrdma_gid dgid; - uint32_t flow_label; - uint8_t sgid_index; - uint8_t hop_limit; - uint8_t traffic_class; - uint8_t reserved; -}; - -struct pvrdma_grh { - uint32_t version_tclass_flow; - uint16_t paylen; - uint8_t next_hdr; - uint8_t hop_limit; - union pvrdma_gid sgid; - union pvrdma_gid dgid; -}; - -enum pvrdma_ah_flags { - PVRDMA_AH_GRH =3D 1, -}; - -enum pvrdma_rate { - PVRDMA_RATE_PORT_CURRENT =3D 0, - PVRDMA_RATE_2_5_GBPS =3D 2, - PVRDMA_RATE_5_GBPS =3D 5, - PVRDMA_RATE_10_GBPS =3D 3, - PVRDMA_RATE_20_GBPS =3D 6, - PVRDMA_RATE_30_GBPS =3D 4, - PVRDMA_RATE_40_GBPS =3D 7, - PVRDMA_RATE_60_GBPS =3D 8, - PVRDMA_RATE_80_GBPS =3D 9, - PVRDMA_RATE_120_GBPS =3D 10, - PVRDMA_RATE_14_GBPS =3D 11, - PVRDMA_RATE_56_GBPS =3D 12, - PVRDMA_RATE_112_GBPS =3D 13, - PVRDMA_RATE_168_GBPS =3D 14, - PVRDMA_RATE_25_GBPS =3D 15, - PVRDMA_RATE_100_GBPS =3D 16, - PVRDMA_RATE_200_GBPS =3D 17, - PVRDMA_RATE_300_GBPS =3D 18, -}; - -struct pvrdma_ah_attr { - struct pvrdma_global_route grh; - uint16_t dlid; - uint16_t vlan_id; - uint8_t sl; - uint8_t src_path_bits; - uint8_t static_rate; - uint8_t ah_flags; - uint8_t port_num; - uint8_t dmac[6]; - uint8_t reserved; -}; - -enum pvrdma_cq_notify_flags { - PVRDMA_CQ_SOLICITED =3D 1 << 0, - PVRDMA_CQ_NEXT_COMP =3D 1 << 1, - PVRDMA_CQ_SOLICITED_MASK =3D PVRDMA_CQ_SOLICITED | - PVRDMA_CQ_NEXT_COMP, - PVRDMA_CQ_REPORT_MISSED_EVENTS =3D 1 << 2, -}; - -struct pvrdma_qp_cap { - uint32_t max_send_wr; - uint32_t max_recv_wr; - uint32_t max_send_sge; - uint32_t max_recv_sge; - uint32_t max_inline_data; - uint32_t reserved; -}; - -enum pvrdma_sig_type { - PVRDMA_SIGNAL_ALL_WR, - PVRDMA_SIGNAL_REQ_WR, -}; - -enum pvrdma_qp_type { - PVRDMA_QPT_SMI, - PVRDMA_QPT_GSI, - PVRDMA_QPT_RC, - PVRDMA_QPT_UC, - PVRDMA_QPT_UD, - PVRDMA_QPT_RAW_IPV6, - PVRDMA_QPT_RAW_ETHERTYPE, - PVRDMA_QPT_RAW_PACKET =3D 8, - PVRDMA_QPT_XRC_INI =3D 9, - PVRDMA_QPT_XRC_TGT, - PVRDMA_QPT_MAX, -}; - -enum pvrdma_qp_create_flags { - PVRDMA_QP_CREATE_IPOPVRDMA_UD_LSO =3D 1 << 0, - PVRDMA_QP_CREATE_BLOCK_MULTICAST_LOOPBACK =3D 1 << 1, -}; - -enum pvrdma_qp_attr_mask { - PVRDMA_QP_STATE =3D 1 << 0, - PVRDMA_QP_CUR_STATE =3D 1 << 1, - PVRDMA_QP_EN_SQD_ASYNC_NOTIFY =3D 1 << 2, - PVRDMA_QP_ACCESS_FLAGS =3D 1 << 3, - PVRDMA_QP_PKEY_INDEX =3D 1 << 4, - PVRDMA_QP_PORT =3D 1 << 5, - PVRDMA_QP_QKEY =3D 1 << 6, - PVRDMA_QP_AV =3D 1 << 7, - PVRDMA_QP_PATH_MTU =3D 1 << 8, - PVRDMA_QP_TIMEOUT =3D 1 << 9, - PVRDMA_QP_RETRY_CNT =3D 1 << 10, - PVRDMA_QP_RNR_RETRY =3D 1 << 11, - PVRDMA_QP_RQ_PSN =3D 1 << 12, - PVRDMA_QP_MAX_QP_RD_ATOMIC =3D 1 << 13, - PVRDMA_QP_ALT_PATH =3D 1 << 14, - PVRDMA_QP_MIN_RNR_TIMER =3D 1 << 15, - PVRDMA_QP_SQ_PSN =3D 1 << 16, - PVRDMA_QP_MAX_DEST_RD_ATOMIC =3D 1 << 17, - PVRDMA_QP_PATH_MIG_STATE =3D 1 << 18, - PVRDMA_QP_CAP =3D 1 << 19, - PVRDMA_QP_DEST_QPN =3D 1 << 20, - PVRDMA_QP_ATTR_MASK_MAX =3D PVRDMA_QP_DEST_QPN, -}; - -enum pvrdma_qp_state { - PVRDMA_QPS_RESET, - PVRDMA_QPS_INIT, - PVRDMA_QPS_RTR, - PVRDMA_QPS_RTS, - PVRDMA_QPS_SQD, - PVRDMA_QPS_SQE, - PVRDMA_QPS_ERR, -}; - -enum pvrdma_mig_state { - PVRDMA_MIG_MIGRATED, - PVRDMA_MIG_REARM, - PVRDMA_MIG_ARMED, -}; - -enum pvrdma_mw_type { - PVRDMA_MW_TYPE_1 =3D 1, - PVRDMA_MW_TYPE_2 =3D 2, -}; - -struct pvrdma_srq_attr { - uint32_t max_wr; - uint32_t max_sge; - uint32_t srq_limit; - uint32_t reserved; -}; - -struct pvrdma_qp_attr { - enum pvrdma_qp_state qp_state; - enum pvrdma_qp_state cur_qp_state; - enum pvrdma_mtu path_mtu; - enum pvrdma_mig_state path_mig_state; - uint32_t qkey; - uint32_t rq_psn; - uint32_t sq_psn; - uint32_t dest_qp_num; - uint32_t qp_access_flags; - uint16_t pkey_index; - uint16_t alt_pkey_index; - uint8_t en_sqd_async_notify; - uint8_t sq_draining; - uint8_t max_rd_atomic; - uint8_t max_dest_rd_atomic; - uint8_t min_rnr_timer; - uint8_t port_num; - uint8_t timeout; - uint8_t retry_cnt; - uint8_t rnr_retry; - uint8_t alt_port_num; - uint8_t alt_timeout; - uint8_t reserved[5]; - struct pvrdma_qp_cap cap; - struct pvrdma_ah_attr ah_attr; - struct pvrdma_ah_attr alt_ah_attr; -}; - -enum pvrdma_send_flags { - PVRDMA_SEND_FENCE =3D 1 << 0, - PVRDMA_SEND_SIGNALED =3D 1 << 1, - PVRDMA_SEND_SOLICITED =3D 1 << 2, - PVRDMA_SEND_INLINE =3D 1 << 3, - PVRDMA_SEND_IP_CSUM =3D 1 << 4, - PVRDMA_SEND_FLAGS_MAX =3D PVRDMA_SEND_IP_CSUM, -}; - -enum pvrdma_access_flags { - PVRDMA_ACCESS_LOCAL_WRITE =3D 1 << 0, - PVRDMA_ACCESS_REMOTE_WRITE =3D 1 << 1, - PVRDMA_ACCESS_REMOTE_READ =3D 1 << 2, - PVRDMA_ACCESS_REMOTE_ATOMIC =3D 1 << 3, - PVRDMA_ACCESS_MW_BIND =3D 1 << 4, - PVRDMA_ZERO_BASED =3D 1 << 5, - PVRDMA_ACCESS_ON_DEMAND =3D 1 << 6, - PVRDMA_ACCESS_FLAGS_MAX =3D PVRDMA_ACCESS_ON_DEMAND, -}; - -#endif /* __PVRDMA_VERBS_H__ */ diff --git a/include/standard-headers/rdma/vmw_pvrdma-abi.h b/include/stand= ard-headers/rdma/vmw_pvrdma-abi.h deleted file mode 100644 index c30182a7ae..0000000000 --- a/include/standard-headers/rdma/vmw_pvrdma-abi.h +++ /dev/null @@ -1,310 +0,0 @@ -/* SPDX-License-Identifier: ((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Cl= ause) */ -/* - * Copyright (c) 2012-2016 VMware, Inc. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of EITHER the GNU General Public License - * version 2 as published by the Free Software Foundation or the BSD - * 2-Clause License. This program is distributed in the hope that it - * will be useful, but WITHOUT ANY WARRANTY; WITHOUT EVEN THE IMPLIED - * WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. - * See the GNU General Public License version 2 for more details at - * http://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html. - * - * You should have received a copy of the GNU General Public License - * along with this program available in the file COPYING in the main - * directory of this source tree. - * - * The BSD 2-Clause License - * - * Redistribution and use in source and binary forms, with or - * without modification, are permitted provided that the following - * conditions are met: - * - * - Redistributions of source code must retain the above - * copyright notice, this list of conditions and the following - * disclaimer. - * - * - Redistributions in binary form must reproduce the above - * copyright notice, this list of conditions and the following - * disclaimer in the documentation and/or other materials - * provided with the distribution. - * - * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS - * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT - * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS - * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE - * COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, - * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES - * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR - * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) - * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, - * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) - * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED - * OF THE POSSIBILITY OF SUCH DAMAGE. - */ - -#ifndef __VMW_PVRDMA_ABI_H__ -#define __VMW_PVRDMA_ABI_H__ - -#include "standard-headers/linux/types.h" - -#define PVRDMA_UVERBS_ABI_VERSION 3 /* ABI Version. */ -#define PVRDMA_UAR_HANDLE_MASK 0x00FFFFFF /* Bottom 24 bits. */ -#define PVRDMA_UAR_QP_OFFSET 0 /* QP doorbell. */ -#define PVRDMA_UAR_QP_SEND (1 << 30) /* Send bit. */ -#define PVRDMA_UAR_QP_RECV (1 << 31) /* Recv bit. */ -#define PVRDMA_UAR_CQ_OFFSET 4 /* CQ doorbell. */ -#define PVRDMA_UAR_CQ_ARM_SOL (1 << 29) /* Arm solicited bit. */ -#define PVRDMA_UAR_CQ_ARM (1 << 30) /* Arm bit. */ -#define PVRDMA_UAR_CQ_POLL (1 << 31) /* Poll bit. */ -#define PVRDMA_UAR_SRQ_OFFSET 8 /* SRQ doorbell. */ -#define PVRDMA_UAR_SRQ_RECV (1 << 30) /* Recv bit. */ - -enum pvrdma_wr_opcode { - PVRDMA_WR_RDMA_WRITE, - PVRDMA_WR_RDMA_WRITE_WITH_IMM, - PVRDMA_WR_SEND, - PVRDMA_WR_SEND_WITH_IMM, - PVRDMA_WR_RDMA_READ, - PVRDMA_WR_ATOMIC_CMP_AND_SWP, - PVRDMA_WR_ATOMIC_FETCH_AND_ADD, - PVRDMA_WR_LSO, - PVRDMA_WR_SEND_WITH_INV, - PVRDMA_WR_RDMA_READ_WITH_INV, - PVRDMA_WR_LOCAL_INV, - PVRDMA_WR_FAST_REG_MR, - PVRDMA_WR_MASKED_ATOMIC_CMP_AND_SWP, - PVRDMA_WR_MASKED_ATOMIC_FETCH_AND_ADD, - PVRDMA_WR_BIND_MW, - PVRDMA_WR_REG_SIG_MR, - PVRDMA_WR_ERROR, -}; - -enum pvrdma_wc_status { - PVRDMA_WC_SUCCESS, - PVRDMA_WC_LOC_LEN_ERR, - PVRDMA_WC_LOC_QP_OP_ERR, - PVRDMA_WC_LOC_EEC_OP_ERR, - PVRDMA_WC_LOC_PROT_ERR, - PVRDMA_WC_WR_FLUSH_ERR, - PVRDMA_WC_MW_BIND_ERR, - PVRDMA_WC_BAD_RESP_ERR, - PVRDMA_WC_LOC_ACCESS_ERR, - PVRDMA_WC_REM_INV_REQ_ERR, - PVRDMA_WC_REM_ACCESS_ERR, - PVRDMA_WC_REM_OP_ERR, - PVRDMA_WC_RETRY_EXC_ERR, - PVRDMA_WC_RNR_RETRY_EXC_ERR, - PVRDMA_WC_LOC_RDD_VIOL_ERR, - PVRDMA_WC_REM_INV_RD_REQ_ERR, - PVRDMA_WC_REM_ABORT_ERR, - PVRDMA_WC_INV_EECN_ERR, - PVRDMA_WC_INV_EEC_STATE_ERR, - PVRDMA_WC_FATAL_ERR, - PVRDMA_WC_RESP_TIMEOUT_ERR, - PVRDMA_WC_GENERAL_ERR, -}; - -enum pvrdma_wc_opcode { - PVRDMA_WC_SEND, - PVRDMA_WC_RDMA_WRITE, - PVRDMA_WC_RDMA_READ, - PVRDMA_WC_COMP_SWAP, - PVRDMA_WC_FETCH_ADD, - PVRDMA_WC_BIND_MW, - PVRDMA_WC_LSO, - PVRDMA_WC_LOCAL_INV, - PVRDMA_WC_FAST_REG_MR, - PVRDMA_WC_MASKED_COMP_SWAP, - PVRDMA_WC_MASKED_FETCH_ADD, - PVRDMA_WC_RECV =3D 1 << 7, - PVRDMA_WC_RECV_RDMA_WITH_IMM, -}; - -enum pvrdma_wc_flags { - PVRDMA_WC_GRH =3D 1 << 0, - PVRDMA_WC_WITH_IMM =3D 1 << 1, - PVRDMA_WC_WITH_INVALIDATE =3D 1 << 2, - PVRDMA_WC_IP_CSUM_OK =3D 1 << 3, - PVRDMA_WC_WITH_SMAC =3D 1 << 4, - PVRDMA_WC_WITH_VLAN =3D 1 << 5, - PVRDMA_WC_WITH_NETWORK_HDR_TYPE =3D 1 << 6, - PVRDMA_WC_FLAGS_MAX =3D PVRDMA_WC_WITH_NETWORK_HDR_TYPE, -}; - -enum pvrdma_network_type { - PVRDMA_NETWORK_IB, - PVRDMA_NETWORK_ROCE_V1 =3D PVRDMA_NETWORK_IB, - PVRDMA_NETWORK_IPV4, - PVRDMA_NETWORK_IPV6 -}; - -struct pvrdma_alloc_ucontext_resp { - uint32_t qp_tab_size; - uint32_t reserved; -}; - -struct pvrdma_alloc_pd_resp { - uint32_t pdn; - uint32_t reserved; -}; - -struct pvrdma_create_cq { - uint64_t __attribute__((aligned(8))) buf_addr; - uint32_t buf_size; - uint32_t reserved; -}; - -struct pvrdma_create_cq_resp { - uint32_t cqn; - uint32_t reserved; -}; - -struct pvrdma_resize_cq { - uint64_t __attribute__((aligned(8))) buf_addr; - uint32_t buf_size; - uint32_t reserved; -}; - -struct pvrdma_create_srq { - uint64_t __attribute__((aligned(8))) buf_addr; - uint32_t buf_size; - uint32_t reserved; -}; - -struct pvrdma_create_srq_resp { - uint32_t srqn; - uint32_t reserved; -}; - -struct pvrdma_create_qp { - uint64_t __attribute__((aligned(8))) rbuf_addr; - uint64_t __attribute__((aligned(8))) sbuf_addr; - uint32_t rbuf_size; - uint32_t sbuf_size; - uint64_t __attribute__((aligned(8))) qp_addr; -}; - -struct pvrdma_create_qp_resp { - uint32_t qpn; - uint32_t qp_handle; -}; - -/* PVRDMA masked atomic compare and swap */ -struct pvrdma_ex_cmp_swap { - uint64_t __attribute__((aligned(8))) swap_val; - uint64_t __attribute__((aligned(8))) compare_val; - uint64_t __attribute__((aligned(8))) swap_mask; - uint64_t __attribute__((aligned(8))) compare_mask; -}; - -/* PVRDMA masked atomic fetch and add */ -struct pvrdma_ex_fetch_add { - uint64_t __attribute__((aligned(8))) add_val; - uint64_t __attribute__((aligned(8))) field_boundary; -}; - -/* PVRDMA address vector. */ -struct pvrdma_av { - uint32_t port_pd; - uint32_t sl_tclass_flowlabel; - uint8_t dgid[16]; - uint8_t src_path_bits; - uint8_t gid_index; - uint8_t stat_rate; - uint8_t hop_limit; - uint8_t dmac[6]; - uint8_t reserved[6]; -}; - -/* PVRDMA scatter/gather entry */ -struct pvrdma_sge { - uint64_t __attribute__((aligned(8))) addr; - uint32_t length; - uint32_t lkey; -}; - -/* PVRDMA receive queue work request */ -struct pvrdma_rq_wqe_hdr { - uint64_t __attribute__((aligned(8))) wr_id; /* wr id */ - uint32_t num_sge; /* size of s/g array */ - uint32_t total_len; /* reserved */ -}; -/* Use pvrdma_sge (ib_sge) for receive queue s/g array elements. */ - -/* PVRDMA send queue work request */ -struct pvrdma_sq_wqe_hdr { - uint64_t __attribute__((aligned(8))) wr_id; /* wr id */ - uint32_t num_sge; /* size of s/g array */ - uint32_t total_len; /* reserved */ - uint32_t opcode; /* operation type */ - uint32_t send_flags; /* wr flags */ - union { - uint32_t imm_data; - uint32_t invalidate_rkey; - } ex; - uint32_t reserved; - union { - struct { - uint64_t __attribute__((aligned(8))) remote_addr; - uint32_t rkey; - uint8_t reserved[4]; - } rdma; - struct { - uint64_t __attribute__((aligned(8))) remote_addr; - uint64_t __attribute__((aligned(8))) compare_add; - uint64_t __attribute__((aligned(8))) swap; - uint32_t rkey; - uint32_t reserved; - } atomic; - struct { - uint64_t __attribute__((aligned(8))) remote_addr; - uint32_t log_arg_sz; - uint32_t rkey; - union { - struct pvrdma_ex_cmp_swap cmp_swap; - struct pvrdma_ex_fetch_add fetch_add; - } wr_data; - } masked_atomics; - struct { - uint64_t __attribute__((aligned(8))) iova_start; - uint64_t __attribute__((aligned(8))) pl_pdir_dma; - uint32_t page_shift; - uint32_t page_list_len; - uint32_t length; - uint32_t access_flags; - uint32_t rkey; - uint32_t reserved; - } fast_reg; - struct { - uint32_t remote_qpn; - uint32_t remote_qkey; - struct pvrdma_av av; - } ud; - } wr; -}; -/* Use pvrdma_sge (ib_sge) for send queue s/g array elements. */ - -/* Completion queue element. */ -struct pvrdma_cqe { - uint64_t __attribute__((aligned(8))) wr_id; - uint64_t __attribute__((aligned(8))) qp; - uint32_t opcode; - uint32_t status; - uint32_t byte_len; - uint32_t imm_data; - uint32_t src_qp; - uint32_t wc_flags; - uint32_t vendor_err; - uint16_t pkey_index; - uint16_t slid; - uint8_t sl; - uint8_t dlid_path_bits; - uint8_t port_num; - uint8_t smac[6]; - uint8_t network_hdr_type; - uint8_t reserved2[6]; /* Pad to next power of 2 (64). */ -}; - -#endif /* __VMW_PVRDMA_ABI_H__ */ diff --git a/contrib/rdmacm-mux/main.c b/contrib/rdmacm-mux/main.c deleted file mode 100644 index 771ca01e03..0000000000 --- a/contrib/rdmacm-mux/main.c +++ /dev/null @@ -1,831 +0,0 @@ -/* - * QEMU paravirtual RDMA - rdmacm-mux implementation - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include -#include -#include -#include - -#include -#include -#include -#include -#include - -#include "rdmacm-mux.h" - -#define SCALE_US 1000 -#define COMMID_TTL 2 /* How many SCALE_US a context of MAD session is save= d */ -#define SLEEP_SECS 5 /* This is used both in poll() and thread */ -#define SERVER_LISTEN_BACKLOG 10 -#define MAX_CLIENTS 4096 -#define MAD_RMPP_VERSION 0 -#define MAD_METHOD_MASK0 0x8 - -#define IB_USER_MAD_LONGS_PER_METHOD_MASK (128 / (8 * sizeof(long))) - -#define CM_REQ_DGID_POS 80 -#define CM_SIDR_REQ_DGID_POS 44 - -/* The below can be override by command line parameter */ -#define UNIX_SOCKET_PATH "/var/run/rdmacm-mux" -/* Has format %s-%s-%d" -- */ -#define SOCKET_PATH_MAX (PATH_MAX - NAME_MAX - sizeof(int) - 2) -#define RDMA_PORT_NUM 1 - -typedef struct RdmaCmServerArgs { - char unix_socket_path[PATH_MAX]; - char rdma_dev_name[NAME_MAX]; - int rdma_port_num; -} RdmaCMServerArgs; - -typedef struct CommId2FdEntry { - int fd; - int ttl; /* Initialized to 2, decrement each timeout, entry delete whe= n 0 */ - __be64 gid_ifid; -} CommId2FdEntry; - -typedef struct RdmaCmUMadAgent { - int port_id; - int agent_id; - GHashTable *gid2fd; /* Used to find fd of a given gid */ - GHashTable *commid2fd; /* Used to find fd on of a given comm_id */ -} RdmaCmUMadAgent; - -typedef struct RdmaCmServer { - bool run; - RdmaCMServerArgs args; - struct pollfd fds[MAX_CLIENTS]; - int nfds; - RdmaCmUMadAgent umad_agent; - pthread_t umad_recv_thread; - pthread_rwlock_t lock; -} RdmaCMServer; - -static RdmaCMServer server =3D {0}; - -static void usage(const char *progname) -{ - printf("Usage: %s [OPTION]...\n" - "Start a RDMA-CM multiplexer\n" - "\n" - "\t-h Show this help\n" - "\t-d rdma-device-name Name of RDMA device to register with\n" - "\t-s unix-socket-path Path to unix socket to listen on (defa= ult %s)\n" - "\t-p rdma-device-port Port number of RDMA device to register= with (default %d)\n", - progname, UNIX_SOCKET_PATH, RDMA_PORT_NUM); -} - -static void help(const char *progname) -{ - fprintf(stderr, "Try '%s -h' for more information.\n", progname); -} - -static void parse_args(int argc, char *argv[]) -{ - int c; - char unix_socket_path[SOCKET_PATH_MAX]; - - strcpy(server.args.rdma_dev_name, ""); - strcpy(unix_socket_path, UNIX_SOCKET_PATH); - server.args.rdma_port_num =3D RDMA_PORT_NUM; - - while ((c =3D getopt(argc, argv, "hs:d:p:")) !=3D -1) { - switch (c) { - case 'h': - usage(argv[0]); - exit(0); - - case 'd': - strncpy(server.args.rdma_dev_name, optarg, NAME_MAX - 1); - break; - - case 's': - /* This is temporary, final name will build below */ - strncpy(unix_socket_path, optarg, SOCKET_PATH_MAX - 1); - break; - - case 'p': - server.args.rdma_port_num =3D atoi(optarg); - break; - - default: - help(argv[0]); - exit(1); - } - } - - if (!strcmp(server.args.rdma_dev_name, "")) { - fprintf(stderr, "Missing RDMA device name\n"); - help(argv[0]); - exit(1); - } - - /* Build unique unix-socket file name */ - snprintf(server.args.unix_socket_path, PATH_MAX, "%s-%s-%d", - unix_socket_path, server.args.rdma_dev_name, - server.args.rdma_port_num); - - syslog(LOG_INFO, "unix_socket_path=3D%s", server.args.unix_socket_path= ); - syslog(LOG_INFO, "rdma-device-name=3D%s", server.args.rdma_dev_name); - syslog(LOG_INFO, "rdma-device-port=3D%d", server.args.rdma_port_num); -} - -static void hash_tbl_alloc(void) -{ - - server.umad_agent.gid2fd =3D g_hash_table_new_full(g_int64_hash, - g_int64_equal, - g_free, g_free); - server.umad_agent.commid2fd =3D g_hash_table_new_full(g_int_hash, - g_int_equal, - g_free, g_free); -} - -static void hash_tbl_free(void) -{ - if (server.umad_agent.commid2fd) { - g_hash_table_destroy(server.umad_agent.commid2fd); - } - if (server.umad_agent.gid2fd) { - g_hash_table_destroy(server.umad_agent.gid2fd); - } -} - - -static int _hash_tbl_search_fd_by_ifid(__be64 *gid_ifid) -{ - int *fd; - - fd =3D g_hash_table_lookup(server.umad_agent.gid2fd, gid_ifid); - if (!fd) { - /* Let's try IPv4 */ - *gid_ifid |=3D 0x00000000ffff0000; - fd =3D g_hash_table_lookup(server.umad_agent.gid2fd, gid_ifid); - } - - return fd ? *fd : 0; -} - -static int hash_tbl_search_fd_by_ifid(int *fd, __be64 *gid_ifid) -{ - pthread_rwlock_rdlock(&server.lock); - *fd =3D _hash_tbl_search_fd_by_ifid(gid_ifid); - pthread_rwlock_unlock(&server.lock); - - if (!*fd) { - syslog(LOG_WARNING, "Can't find matching for ifid 0x%llx\n", *gid_= ifid); - return -ENOENT; - } - - return 0; -} - -static int hash_tbl_search_fd_by_comm_id(uint32_t comm_id, int *fd, - __be64 *gid_idid) -{ - CommId2FdEntry *fde; - - pthread_rwlock_rdlock(&server.lock); - fde =3D g_hash_table_lookup(server.umad_agent.commid2fd, &comm_id); - pthread_rwlock_unlock(&server.lock); - - if (!fde) { - syslog(LOG_WARNING, "Can't find matching for comm_id 0x%x\n", comm= _id); - return -ENOENT; - } - - *fd =3D fde->fd; - *gid_idid =3D fde->gid_ifid; - - return 0; -} - -static RdmaCmMuxErrCode add_fd_ifid_pair(int fd, __be64 gid_ifid) -{ - int fd1; - - pthread_rwlock_wrlock(&server.lock); - - fd1 =3D _hash_tbl_search_fd_by_ifid(&gid_ifid); - if (fd1) { /* record already exist - an error */ - pthread_rwlock_unlock(&server.lock); - return fd =3D=3D fd1 ? RDMACM_MUX_ERR_CODE_EEXIST : - RDMACM_MUX_ERR_CODE_EACCES; - } - - g_hash_table_insert(server.umad_agent.gid2fd, g_memdup(&gid_ifid, - sizeof(gid_ifid)), g_memdup(&fd, sizeof(fd))); - - pthread_rwlock_unlock(&server.lock); - - syslog(LOG_INFO, "0x%lx registered on socket %d", - be64toh((uint64_t)gid_ifid), fd); - - return RDMACM_MUX_ERR_CODE_OK; -} - -static RdmaCmMuxErrCode delete_fd_ifid_pair(int fd, __be64 gid_ifid) -{ - int fd1; - - pthread_rwlock_wrlock(&server.lock); - - fd1 =3D _hash_tbl_search_fd_by_ifid(&gid_ifid); - if (!fd1) { /* record not exist - an error */ - pthread_rwlock_unlock(&server.lock); - return RDMACM_MUX_ERR_CODE_ENOTFOUND; - } - - g_hash_table_remove(server.umad_agent.gid2fd, g_memdup(&gid_ifid, - sizeof(gid_ifid))); - pthread_rwlock_unlock(&server.lock); - - syslog(LOG_INFO, "0x%lx unregistered on socket %d", - be64toh((uint64_t)gid_ifid), fd); - - return RDMACM_MUX_ERR_CODE_OK; -} - -static void hash_tbl_save_fd_comm_id_pair(int fd, uint32_t comm_id, - uint64_t gid_ifid) -{ - CommId2FdEntry fde =3D {fd, COMMID_TTL, gid_ifid}; - - pthread_rwlock_wrlock(&server.lock); - g_hash_table_insert(server.umad_agent.commid2fd, - g_memdup(&comm_id, sizeof(comm_id)), - g_memdup(&fde, sizeof(fde))); - pthread_rwlock_unlock(&server.lock); -} - -static gboolean remove_old_comm_ids(gpointer key, gpointer value, - gpointer user_data) -{ - CommId2FdEntry *fde =3D (CommId2FdEntry *)value; - - return !fde->ttl--; -} - -static gboolean remove_entry_from_gid2fd(gpointer key, gpointer value, - gpointer user_data) -{ - if (*(int *)value =3D=3D *(int *)user_data) { - syslog(LOG_INFO, "0x%lx unregistered on socket %d", - be64toh(*(uint64_t *)key), *(int *)value); - return true; - } - - return false; -} - -static void hash_tbl_remove_fd_ifid_pair(int fd) -{ - pthread_rwlock_wrlock(&server.lock); - g_hash_table_foreach_remove(server.umad_agent.gid2fd, - remove_entry_from_gid2fd, (gpointer)&fd); - pthread_rwlock_unlock(&server.lock); -} - -static int get_fd(const char *mad, int umad_len, int *fd, __be64 *gid_ifid) -{ - struct umad_hdr *hdr =3D (struct umad_hdr *)mad; - char *data =3D (char *)hdr + sizeof(*hdr); - int32_t comm_id =3D 0; - uint16_t attr_id =3D be16toh(hdr->attr_id); - int rc =3D 0; - - if (umad_len <=3D sizeof(*hdr)) { - rc =3D -EINVAL; - syslog(LOG_DEBUG, "Ignoring MAD packets with header only\n"); - goto out; - } - - switch (attr_id) { - case UMAD_CM_ATTR_REQ: - if (unlikely(umad_len < sizeof(*hdr) + CM_REQ_DGID_POS + - sizeof(*gid_ifid))) { - rc =3D -EINVAL; - syslog(LOG_WARNING, - "Invalid MAD packet size (%d) for attr_id 0x%x\n", umad= _len, - attr_id); - goto out; - } - memcpy(gid_ifid, data + CM_REQ_DGID_POS, sizeof(*gid_ifid)); - rc =3D hash_tbl_search_fd_by_ifid(fd, gid_ifid); - break; - - case UMAD_CM_ATTR_SIDR_REQ: - if (unlikely(umad_len < sizeof(*hdr) + CM_SIDR_REQ_DGID_POS + - sizeof(*gid_ifid))) { - rc =3D -EINVAL; - syslog(LOG_WARNING, - "Invalid MAD packet size (%d) for attr_id 0x%x\n", umad= _len, - attr_id); - goto out; - } - memcpy(gid_ifid, data + CM_SIDR_REQ_DGID_POS, sizeof(*gid_ifid)); - rc =3D hash_tbl_search_fd_by_ifid(fd, gid_ifid); - break; - - case UMAD_CM_ATTR_REP: - /* Fall through */ - case UMAD_CM_ATTR_REJ: - /* Fall through */ - case UMAD_CM_ATTR_DREQ: - /* Fall through */ - case UMAD_CM_ATTR_DREP: - /* Fall through */ - case UMAD_CM_ATTR_RTU: - data +=3D sizeof(comm_id); - /* Fall through */ - case UMAD_CM_ATTR_SIDR_REP: - if (unlikely(umad_len < sizeof(*hdr) + sizeof(comm_id))) { - rc =3D -EINVAL; - syslog(LOG_WARNING, - "Invalid MAD packet size (%d) for attr_id 0x%x\n", umad= _len, - attr_id); - goto out; - } - memcpy(&comm_id, data, sizeof(comm_id)); - if (comm_id) { - rc =3D hash_tbl_search_fd_by_comm_id(comm_id, fd, gid_ifid); - } - break; - - default: - rc =3D -EINVAL; - syslog(LOG_WARNING, "Unsupported attr_id 0x%x\n", attr_id); - } - - syslog(LOG_DEBUG, "mad_to_vm: %d 0x%x 0x%x\n", *fd, attr_id, comm_id); - -out: - return rc; -} - -static void *umad_recv_thread_func(void *args) -{ - int rc; - RdmaCmMuxMsg msg =3D {}; - int fd =3D -2; - - msg.hdr.msg_type =3D RDMACM_MUX_MSG_TYPE_REQ; - msg.hdr.op_code =3D RDMACM_MUX_OP_CODE_MAD; - - while (server.run) { - do { - msg.umad_len =3D sizeof(msg.umad.mad); - rc =3D umad_recv(server.umad_agent.port_id, &msg.umad, &msg.um= ad_len, - SLEEP_SECS * SCALE_US); - if ((rc =3D=3D -EIO) || (rc =3D=3D -EINVAL)) { - syslog(LOG_CRIT, "Fatal error while trying to read MAD"); - } - - if (rc =3D=3D -ETIMEDOUT) { - g_hash_table_foreach_remove(server.umad_agent.commid2fd, - remove_old_comm_ids, NULL); - } - } while (rc && server.run); - - if (server.run) { - rc =3D get_fd(msg.umad.mad, msg.umad_len, &fd, - &msg.hdr.sgid.global.interface_id); - if (rc) { - continue; - } - - send(fd, &msg, sizeof(msg), 0); - } - } - - return NULL; -} - -static int read_and_process(int fd) -{ - int rc; - RdmaCmMuxMsg msg =3D {}; - struct umad_hdr *hdr; - uint32_t *comm_id =3D 0; - uint16_t attr_id; - - rc =3D recv(fd, &msg, sizeof(msg), 0); - syslog(LOG_DEBUG, "Socket %d, recv %d\n", fd, rc); - - if (rc < 0 && errno !=3D EWOULDBLOCK) { - syslog(LOG_ERR, "Fail to read from socket %d\n", fd); - return -EIO; - } - - if (!rc) { - syslog(LOG_ERR, "Fail to read from socket %d\n", fd); - return -EPIPE; - } - - if (msg.hdr.msg_type !=3D RDMACM_MUX_MSG_TYPE_REQ) { - syslog(LOG_WARNING, "Got non-request message (%d) from socket %d\n= ", - msg.hdr.msg_type, fd); - return -EPERM; - } - - switch (msg.hdr.op_code) { - case RDMACM_MUX_OP_CODE_REG: - rc =3D add_fd_ifid_pair(fd, msg.hdr.sgid.global.interface_id); - break; - - case RDMACM_MUX_OP_CODE_UNREG: - rc =3D delete_fd_ifid_pair(fd, msg.hdr.sgid.global.interface_id); - break; - - case RDMACM_MUX_OP_CODE_MAD: - /* If this is REQ or REP then store the pair comm_id,fd to be later - * used for other messages where gid is unknown */ - hdr =3D (struct umad_hdr *)msg.umad.mad; - attr_id =3D be16toh(hdr->attr_id); - if ((attr_id =3D=3D UMAD_CM_ATTR_REQ) || (attr_id =3D=3D UMAD_CM_A= TTR_DREQ) || - (attr_id =3D=3D UMAD_CM_ATTR_SIDR_REQ) || - (attr_id =3D=3D UMAD_CM_ATTR_REP) || (attr_id =3D=3D UMAD_CM_A= TTR_DREP)) { - comm_id =3D (uint32_t *)(msg.umad.mad + sizeof(*hdr)); - hash_tbl_save_fd_comm_id_pair(fd, *comm_id, - msg.hdr.sgid.global.interface_id= ); - } - - syslog(LOG_DEBUG, "vm_to_mad: %d 0x%x 0x%x\n", fd, attr_id, - comm_id ? *comm_id : 0); - rc =3D umad_send(server.umad_agent.port_id, server.umad_agent.agen= t_id, - &msg.umad, msg.umad_len, 1, 0); - if (rc) { - syslog(LOG_ERR, - "Fail to send MAD message (0x%x) from socket %d, err=3D%= d", - attr_id, fd, rc); - } - break; - - default: - syslog(LOG_ERR, "Got invalid op_code (%d) from socket %d", - msg.hdr.msg_type, fd); - rc =3D RDMACM_MUX_ERR_CODE_EINVAL; - } - - msg.hdr.msg_type =3D RDMACM_MUX_MSG_TYPE_RESP; - msg.hdr.err_code =3D rc; - rc =3D send(fd, &msg, sizeof(msg), 0); - - return rc =3D=3D sizeof(msg) ? 0 : -EPIPE; -} - -static int accept_all(void) -{ - int fd, rc =3D 0; - - pthread_rwlock_wrlock(&server.lock); - - do { - if ((server.nfds + 1) > MAX_CLIENTS) { - syslog(LOG_WARNING, "Too many clients (%d)", server.nfds); - rc =3D -EIO; - goto out; - } - - fd =3D accept(server.fds[0].fd, NULL, NULL); - if (fd < 0) { - if (errno !=3D EWOULDBLOCK) { - syslog(LOG_WARNING, "accept() failed"); - rc =3D -EIO; - goto out; - } - break; - } - - syslog(LOG_INFO, "Client connected on socket %d\n", fd); - server.fds[server.nfds].fd =3D fd; - server.fds[server.nfds].events =3D POLLIN; - server.nfds++; - } while (fd !=3D -1); - -out: - pthread_rwlock_unlock(&server.lock); - return rc; -} - -static void compress_fds(void) -{ - int i, j; - int closed =3D 0; - - pthread_rwlock_wrlock(&server.lock); - - for (i =3D 1; i < server.nfds; i++) { - if (!server.fds[i].fd) { - closed++; - for (j =3D i; j < server.nfds - 1; j++) { - server.fds[j] =3D server.fds[j + 1]; - } - } - } - - server.nfds -=3D closed; - - pthread_rwlock_unlock(&server.lock); -} - -static void close_fd(int idx) -{ - close(server.fds[idx].fd); - syslog(LOG_INFO, "Socket %d closed\n", server.fds[idx].fd); - hash_tbl_remove_fd_ifid_pair(server.fds[idx].fd); - server.fds[idx].fd =3D 0; -} - -static void run(void) -{ - int rc, nfds, i; - bool compress =3D false; - - syslog(LOG_INFO, "Service started"); - - while (server.run) { - rc =3D poll(server.fds, server.nfds, SLEEP_SECS * SCALE_US); - if (rc < 0) { - if (errno !=3D EINTR) { - syslog(LOG_WARNING, "poll() failed"); - } - continue; - } - - if (rc =3D=3D 0) { - continue; - } - - nfds =3D server.nfds; - for (i =3D 0; i < nfds; i++) { - syslog(LOG_DEBUG, "pollfd[%d]: revents 0x%x, events 0x%x\n", i, - server.fds[i].revents, server.fds[i].events); - if (server.fds[i].revents =3D=3D 0) { - continue; - } - - if (server.fds[i].revents !=3D POLLIN) { - if (i =3D=3D 0) { - syslog(LOG_NOTICE, "Unexpected poll() event (0x%x)\n", - server.fds[i].revents); - } else { - close_fd(i); - compress =3D true; - } - continue; - } - - if (i =3D=3D 0) { - rc =3D accept_all(); - if (rc) { - continue; - } - } else { - rc =3D read_and_process(server.fds[i].fd); - if (rc) { - close_fd(i); - compress =3D true; - } - } - } - - if (compress) { - compress =3D false; - compress_fds(); - } - } -} - -static void fini_listener(void) -{ - int i; - - if (server.fds[0].fd <=3D 0) { - return; - } - - for (i =3D server.nfds - 1; i >=3D 0; i--) { - if (server.fds[i].fd) { - close(server.fds[i].fd); - } - } - - unlink(server.args.unix_socket_path); -} - -static void fini_umad(void) -{ - if (server.umad_agent.agent_id) { - umad_unregister(server.umad_agent.port_id, server.umad_agent.agent= _id); - } - - if (server.umad_agent.port_id) { - umad_close_port(server.umad_agent.port_id); - } - - hash_tbl_free(); -} - -static void fini(void) -{ - if (server.umad_recv_thread) { - pthread_join(server.umad_recv_thread, NULL); - server.umad_recv_thread =3D 0; - } - fini_umad(); - fini_listener(); - pthread_rwlock_destroy(&server.lock); - - syslog(LOG_INFO, "Service going down"); -} - -static int init_listener(void) -{ - struct sockaddr_un sun; - int rc, on =3D 1; - - server.fds[0].fd =3D socket(AF_UNIX, SOCK_STREAM, 0); - if (server.fds[0].fd < 0) { - syslog(LOG_ALERT, "socket() failed"); - return -EIO; - } - - rc =3D setsockopt(server.fds[0].fd, SOL_SOCKET, SO_REUSEADDR, (char *)= &on, - sizeof(on)); - if (rc < 0) { - syslog(LOG_ALERT, "setsockopt() failed"); - rc =3D -EIO; - goto err; - } - - rc =3D ioctl(server.fds[0].fd, FIONBIO, (char *)&on); - if (rc < 0) { - syslog(LOG_ALERT, "ioctl() failed"); - rc =3D -EIO; - goto err; - } - - if (strlen(server.args.unix_socket_path) >=3D sizeof(sun.sun_path)) { - syslog(LOG_ALERT, - "Invalid unix_socket_path, size must be less than %ld\n", - sizeof(sun.sun_path)); - rc =3D -EINVAL; - goto err; - } - - sun.sun_family =3D AF_UNIX; - rc =3D snprintf(sun.sun_path, sizeof(sun.sun_path), "%s", - server.args.unix_socket_path); - if (rc < 0 || rc >=3D sizeof(sun.sun_path)) { - syslog(LOG_ALERT, "Could not copy unix socket path\n"); - rc =3D -EINVAL; - goto err; - } - - rc =3D bind(server.fds[0].fd, (struct sockaddr *)&sun, sizeof(sun)); - if (rc < 0) { - syslog(LOG_ALERT, "bind() failed"); - rc =3D -EIO; - goto err; - } - - rc =3D listen(server.fds[0].fd, SERVER_LISTEN_BACKLOG); - if (rc < 0) { - syslog(LOG_ALERT, "listen() failed"); - rc =3D -EIO; - goto err; - } - - server.fds[0].events =3D POLLIN; - server.nfds =3D 1; - server.run =3D true; - - return 0; - -err: - close(server.fds[0].fd); - return rc; -} - -static int init_umad(void) -{ - long method_mask[IB_USER_MAD_LONGS_PER_METHOD_MASK]; - - server.umad_agent.port_id =3D umad_open_port(server.args.rdma_dev_name, - server.args.rdma_port_num); - - if (server.umad_agent.port_id < 0) { - syslog(LOG_WARNING, "umad_open_port() failed"); - return -EIO; - } - - memset(&method_mask, 0, sizeof(method_mask)); - method_mask[0] =3D MAD_METHOD_MASK0; - server.umad_agent.agent_id =3D umad_register(server.umad_agent.port_id, - UMAD_CLASS_CM, - UMAD_SA_CLASS_VERSION, - MAD_RMPP_VERSION, method_ma= sk); - if (server.umad_agent.agent_id < 0) { - syslog(LOG_WARNING, "umad_register() failed"); - return -EIO; - } - - hash_tbl_alloc(); - - return 0; -} - -static void signal_handler(int sig, siginfo_t *siginfo, void *context) -{ - static bool warned; - - /* Prevent stop if clients are connected */ - if (server.nfds !=3D 1) { - if (!warned) { - syslog(LOG_WARNING, - "Can't stop while active client exist, resend SIGINT to= overid"); - warned =3D true; - return; - } - } - - if (sig =3D=3D SIGINT) { - server.run =3D false; - fini(); - } - - exit(0); -} - -static int init(void) -{ - int rc; - struct sigaction sig =3D {}; - - rc =3D init_listener(); - if (rc) { - return rc; - } - - rc =3D init_umad(); - if (rc) { - return rc; - } - - pthread_rwlock_init(&server.lock, 0); - - rc =3D pthread_create(&server.umad_recv_thread, NULL, umad_recv_thread= _func, - NULL); - if (rc) { - syslog(LOG_ERR, "Fail to create UMAD receiver thread (%d)\n", rc); - return rc; - } - - sig.sa_sigaction =3D &signal_handler; - sig.sa_flags =3D SA_SIGINFO; - rc =3D sigaction(SIGINT, &sig, NULL); - if (rc < 0) { - syslog(LOG_ERR, "Fail to install SIGINT handler (%d)\n", errno); - return rc; - } - - return 0; -} - -int main(int argc, char *argv[]) -{ - int rc; - - memset(&server, 0, sizeof(server)); - - parse_args(argc, argv); - - rc =3D init(); - if (rc) { - syslog(LOG_ERR, "Fail to initialize server (%d)\n", rc); - rc =3D -EAGAIN; - goto out; - } - - run(); - -out: - fini(); - - return rc; -} diff --git a/hw/core/machine-qmp-cmds.c b/hw/core/machine-qmp-cmds.c index 4b72009cd3..c20829b9ae 100644 --- a/hw/core/machine-qmp-cmds.c +++ b/hw/core/machine-qmp-cmds.c @@ -12,7 +12,6 @@ #include "hw/boards.h" #include "hw/intc/intc.h" #include "hw/mem/memory-device.h" -#include "hw/rdma/rdma.h" #include "qapi/error.h" #include "qapi/qapi-builtin-visit.h" #include "qapi/qapi-commands-machine.h" @@ -291,37 +290,6 @@ MemoryInfo *qmp_query_memory_size_summary(Error **errp) return mem_info; } =20 -static int qmp_x_query_rdma_foreach(Object *obj, void *opaque) -{ - RdmaProvider *rdma; - RdmaProviderClass *k; - GString *buf =3D opaque; - - if (object_dynamic_cast(obj, INTERFACE_RDMA_PROVIDER)) { - rdma =3D RDMA_PROVIDER(obj); - k =3D RDMA_PROVIDER_GET_CLASS(obj); - if (k->format_statistics) { - k->format_statistics(rdma, buf); - } else { - g_string_append_printf(buf, - "RDMA statistics not available for %s.\= n", - object_get_typename(obj)); - } - } - - return 0; -} - -HumanReadableText *qmp_x_query_rdma(Error **errp) -{ - g_autoptr(GString) buf =3D g_string_new(""); - - object_child_foreach_recursive(object_get_root(), - qmp_x_query_rdma_foreach, buf); - - return human_readable_text_from_str(buf); -} - HumanReadableText *qmp_x_query_ramblock(Error **errp) { g_autoptr(GString) buf =3D ram_block_format(); diff --git a/hw/rdma/rdma.c b/hw/rdma/rdma.c deleted file mode 100644 index 7bec0d0d2c..0000000000 --- a/hw/rdma/rdma.c +++ /dev/null @@ -1,30 +0,0 @@ -/* - * RDMA device interface - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include "hw/rdma/rdma.h" -#include "qemu/module.h" - -static const TypeInfo rdma_hmp_info =3D { - .name =3D INTERFACE_RDMA_PROVIDER, - .parent =3D TYPE_INTERFACE, - .class_size =3D sizeof(RdmaProviderClass), -}; - -static void rdma_register_types(void) -{ - type_register_static(&rdma_hmp_info); -} - -type_init(rdma_register_types) diff --git a/hw/rdma/rdma_backend.c b/hw/rdma/rdma_backend.c deleted file mode 100644 index 6dcdfbbbe2..0000000000 --- a/hw/rdma/rdma_backend.c +++ /dev/null @@ -1,1401 +0,0 @@ -/* - * QEMU paravirtual RDMA - Generic RDMA backend - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include "qapi/qapi-events-rdma.h" - -#include - -#include "contrib/rdmacm-mux/rdmacm-mux.h" -#include "trace.h" -#include "rdma_utils.h" -#include "rdma_rm.h" -#include "rdma_backend.h" - -#define THR_NAME_LEN 16 -#define THR_POLL_TO 5000 - -#define MAD_HDR_SIZE sizeof(struct ibv_grh) - -typedef struct BackendCtx { - void *up_ctx; - struct ibv_sge sge; /* Used to save MAD recv buffer */ - RdmaBackendQP *backend_qp; /* To maintain recv buffers */ - RdmaBackendSRQ *backend_srq; -} BackendCtx; - -struct backend_umad { - struct ib_user_mad hdr; - char mad[RDMA_MAX_PRIVATE_DATA]; -}; - -static void (*comp_handler)(void *ctx, struct ibv_wc *wc); - -static void dummy_comp_handler(void *ctx, struct ibv_wc *wc) -{ - rdma_error_report("No completion handler is registered"); -} - -static inline void complete_work(enum ibv_wc_status status, uint32_t vendo= r_err, - void *ctx) -{ - struct ibv_wc wc =3D {}; - - wc.status =3D status; - wc.vendor_err =3D vendor_err; - - comp_handler(ctx, &wc); -} - -static void free_cqe_ctx(gpointer data, gpointer user_data) -{ - BackendCtx *bctx; - RdmaDeviceResources *rdma_dev_res =3D user_data; - unsigned long cqe_ctx_id =3D GPOINTER_TO_INT(data); - - bctx =3D rdma_rm_get_cqe_ctx(rdma_dev_res, cqe_ctx_id); - if (bctx) { - rdma_rm_dealloc_cqe_ctx(rdma_dev_res, cqe_ctx_id); - qatomic_dec(&rdma_dev_res->stats.missing_cqe); - } - g_free(bctx); -} - -static void clean_recv_mads(RdmaBackendDev *backend_dev) -{ - unsigned long cqe_ctx_id; - - do { - cqe_ctx_id =3D rdma_protected_gqueue_pop_int64(&backend_dev-> - recv_mads_list); - if (cqe_ctx_id !=3D -ENOENT) { - qatomic_inc(&backend_dev->rdma_dev_res->stats.missing_cqe); - free_cqe_ctx(GINT_TO_POINTER(cqe_ctx_id), - backend_dev->rdma_dev_res); - } - } while (cqe_ctx_id !=3D -ENOENT); -} - -static int rdma_poll_cq(RdmaDeviceResources *rdma_dev_res, struct ibv_cq *= ibcq) -{ - int i, ne, total_ne =3D 0; - BackendCtx *bctx; - struct ibv_wc wc[2]; - RdmaProtectedGSList *cqe_ctx_list; - - WITH_QEMU_LOCK_GUARD(&rdma_dev_res->lock) { - do { - ne =3D ibv_poll_cq(ibcq, ARRAY_SIZE(wc), wc); - - trace_rdma_poll_cq(ne, ibcq); - - for (i =3D 0; i < ne; i++) { - bctx =3D rdma_rm_get_cqe_ctx(rdma_dev_res, wc[i].wr_id); - if (unlikely(!bctx)) { - rdma_error_report("No matching ctx for req %"PRId64, - wc[i].wr_id); - continue; - } - - comp_handler(bctx->up_ctx, &wc[i]); - - if (bctx->backend_qp) { - cqe_ctx_list =3D &bctx->backend_qp->cqe_ctx_list; - } else { - cqe_ctx_list =3D &bctx->backend_srq->cqe_ctx_list; - } - - rdma_protected_gslist_remove_int32(cqe_ctx_list, wc[i].wr_= id); - rdma_rm_dealloc_cqe_ctx(rdma_dev_res, wc[i].wr_id); - g_free(bctx); - } - total_ne +=3D ne; - } while (ne > 0); - qatomic_sub(&rdma_dev_res->stats.missing_cqe, total_ne); - } - - if (ne < 0) { - rdma_error_report("ibv_poll_cq fail, rc=3D%d, errno=3D%d", ne, err= no); - } - - rdma_dev_res->stats.completions +=3D total_ne; - - return total_ne; -} - -static void *comp_handler_thread(void *arg) -{ - RdmaBackendDev *backend_dev =3D (RdmaBackendDev *)arg; - int rc; - struct ibv_cq *ev_cq; - void *ev_ctx; - int flags; - GPollFD pfds[1]; - - /* Change to non-blocking mode */ - flags =3D fcntl(backend_dev->channel->fd, F_GETFL); - rc =3D fcntl(backend_dev->channel->fd, F_SETFL, flags | O_NONBLOCK); - if (rc < 0) { - rdma_error_report("Failed to change backend channel FD to non-bloc= king"); - return NULL; - } - - pfds[0].fd =3D backend_dev->channel->fd; - pfds[0].events =3D G_IO_IN | G_IO_HUP | G_IO_ERR; - - backend_dev->comp_thread.is_running =3D true; - - while (backend_dev->comp_thread.run) { - do { - rc =3D qemu_poll_ns(pfds, 1, THR_POLL_TO * (int64_t)SCALE_MS); - if (!rc) { - backend_dev->rdma_dev_res->stats.poll_cq_ppoll_to++; - } - } while (!rc && backend_dev->comp_thread.run); - - if (backend_dev->comp_thread.run) { - rc =3D ibv_get_cq_event(backend_dev->channel, &ev_cq, &ev_ctx); - if (unlikely(rc)) { - rdma_error_report("ibv_get_cq_event fail, rc=3D%d, errno= =3D%d", rc, - errno); - continue; - } - - rc =3D ibv_req_notify_cq(ev_cq, 0); - if (unlikely(rc)) { - rdma_error_report("ibv_req_notify_cq fail, rc=3D%d, errno= =3D%d", rc, - errno); - } - - backend_dev->rdma_dev_res->stats.poll_cq_from_bk++; - rdma_poll_cq(backend_dev->rdma_dev_res, ev_cq); - - ibv_ack_cq_events(ev_cq, 1); - } - } - - backend_dev->comp_thread.is_running =3D false; - - qemu_thread_exit(0); - - return NULL; -} - -static inline void disable_rdmacm_mux_async(RdmaBackendDev *backend_dev) -{ - qatomic_set(&backend_dev->rdmacm_mux.can_receive, 0); -} - -static inline void enable_rdmacm_mux_async(RdmaBackendDev *backend_dev) -{ - qatomic_set(&backend_dev->rdmacm_mux.can_receive, sizeof(RdmaCmMuxMsg)= ); -} - -static inline int rdmacm_mux_can_process_async(RdmaBackendDev *backend_dev) -{ - return qatomic_read(&backend_dev->rdmacm_mux.can_receive); -} - -static int rdmacm_mux_check_op_status(CharBackend *mad_chr_be) -{ - RdmaCmMuxMsg msg =3D {}; - int ret; - - ret =3D qemu_chr_fe_read_all(mad_chr_be, (uint8_t *)&msg, sizeof(msg)); - if (ret !=3D sizeof(msg)) { - rdma_error_report("Got invalid message from mux: size %d, expectin= g %d", - ret, (int)sizeof(msg)); - return -EIO; - } - - trace_rdmacm_mux_check_op_status(msg.hdr.msg_type, msg.hdr.op_code, - msg.hdr.err_code); - - if (msg.hdr.msg_type !=3D RDMACM_MUX_MSG_TYPE_RESP) { - rdma_error_report("Got invalid message type %d", msg.hdr.msg_type); - return -EIO; - } - - if (msg.hdr.err_code !=3D RDMACM_MUX_ERR_CODE_OK) { - rdma_error_report("Operation failed in mux, error code %d", - msg.hdr.err_code); - return -EIO; - } - - return 0; -} - -static int rdmacm_mux_send(RdmaBackendDev *backend_dev, RdmaCmMuxMsg *msg) -{ - int rc =3D 0; - - msg->hdr.msg_type =3D RDMACM_MUX_MSG_TYPE_REQ; - trace_rdmacm_mux("send", msg->hdr.msg_type, msg->hdr.op_code); - disable_rdmacm_mux_async(backend_dev); - rc =3D qemu_chr_fe_write(backend_dev->rdmacm_mux.chr_be, - (const uint8_t *)msg, sizeof(*msg)); - if (rc !=3D sizeof(*msg)) { - enable_rdmacm_mux_async(backend_dev); - rdma_error_report("Failed to send request to rdmacm_mux (rc=3D%d)"= , rc); - return -EIO; - } - - rc =3D rdmacm_mux_check_op_status(backend_dev->rdmacm_mux.chr_be); - if (rc) { - rdma_error_report("Failed to execute rdmacm_mux request %d (rc=3D%= d)", - msg->hdr.op_code, rc); - } - - enable_rdmacm_mux_async(backend_dev); - - return 0; -} - -static void stop_backend_thread(RdmaBackendThread *thread) -{ - thread->run =3D false; - while (thread->is_running) { - sleep(THR_POLL_TO / SCALE_US / 2); - } -} - -static void start_comp_thread(RdmaBackendDev *backend_dev) -{ - char thread_name[THR_NAME_LEN] =3D {}; - - stop_backend_thread(&backend_dev->comp_thread); - - snprintf(thread_name, sizeof(thread_name), "rdma_comp_%s", - ibv_get_device_name(backend_dev->ib_dev)); - backend_dev->comp_thread.run =3D true; - qemu_thread_create(&backend_dev->comp_thread.thread, thread_name, - comp_handler_thread, backend_dev, QEMU_THREAD_DETAC= HED); -} - -void rdma_backend_register_comp_handler(void (*handler)(void *ctx, - struct ibv_wc *wc= )) -{ - comp_handler =3D handler; -} - -void rdma_backend_unregister_comp_handler(void) -{ - rdma_backend_register_comp_handler(dummy_comp_handler); -} - -int rdma_backend_query_port(RdmaBackendDev *backend_dev, - struct ibv_port_attr *port_attr) -{ - int rc; - - rc =3D ibv_query_port(backend_dev->context, backend_dev->port_num, por= t_attr); - if (rc) { - rdma_error_report("ibv_query_port fail, rc=3D%d, errno=3D%d", rc, = errno); - return -EIO; - } - - return 0; -} - -void rdma_backend_poll_cq(RdmaDeviceResources *rdma_dev_res, RdmaBackendCQ= *cq) -{ - int polled; - - rdma_dev_res->stats.poll_cq_from_guest++; - polled =3D rdma_poll_cq(rdma_dev_res, cq->ibcq); - if (!polled) { - rdma_dev_res->stats.poll_cq_from_guest_empty++; - } -} - -static GHashTable *ah_hash; - -static struct ibv_ah *create_ah(RdmaBackendDev *backend_dev, struct ibv_pd= *pd, - uint8_t sgid_idx, union ibv_gid *dgid) -{ - GBytes *ah_key =3D g_bytes_new(dgid, sizeof(*dgid)); - struct ibv_ah *ah =3D g_hash_table_lookup(ah_hash, ah_key); - - if (ah) { - trace_rdma_create_ah_cache_hit(be64_to_cpu(dgid->global.subnet_pre= fix), - be64_to_cpu(dgid->global.interface_= id)); - g_bytes_unref(ah_key); - } else { - struct ibv_ah_attr ah_attr =3D { - .is_global =3D 1, - .port_num =3D backend_dev->port_num, - .grh.hop_limit =3D 1, - }; - - ah_attr.grh.dgid =3D *dgid; - ah_attr.grh.sgid_index =3D sgid_idx; - - ah =3D ibv_create_ah(pd, &ah_attr); - if (ah) { - g_hash_table_insert(ah_hash, ah_key, ah); - } else { - g_bytes_unref(ah_key); - rdma_error_report("Failed to create AH for gid <0x%" PRIx64", = 0x%"PRIx64">", - be64_to_cpu(dgid->global.subnet_prefix), - be64_to_cpu(dgid->global.interface_id)); - } - - trace_rdma_create_ah_cache_miss(be64_to_cpu(dgid->global.subnet_pr= efix), - be64_to_cpu(dgid->global.interface= _id)); - } - - return ah; -} - -static void destroy_ah_hash_key(gpointer data) -{ - g_bytes_unref(data); -} - -static void destroy_ah_hast_data(gpointer data) -{ - struct ibv_ah *ah =3D data; - - ibv_destroy_ah(ah); -} - -static void ah_cache_init(void) -{ - ah_hash =3D g_hash_table_new_full(g_bytes_hash, g_bytes_equal, - destroy_ah_hash_key, destroy_ah_hast_d= ata); -} - -#ifdef LEGACY_RDMA_REG_MR -static int build_host_sge_array(RdmaDeviceResources *rdma_dev_res, - struct ibv_sge *sge, uint8_t num_sge, - uint64_t *total_length) -{ - RdmaRmMR *mr; - int idx; - - for (idx =3D 0; idx < num_sge; idx++) { - mr =3D rdma_rm_get_mr(rdma_dev_res, sge[idx].lkey); - if (unlikely(!mr)) { - rdma_error_report("Invalid lkey 0x%x", sge[idx].lkey); - return VENDOR_ERR_INVLKEY | sge[idx].lkey; - } - - sge[idx].addr =3D (uintptr_t)mr->virt + sge[idx].addr - mr->start; - sge[idx].lkey =3D rdma_backend_mr_lkey(&mr->backend_mr); - - *total_length +=3D sge[idx].length; - } - - return 0; -} -#else -static inline int build_host_sge_array(RdmaDeviceResources *rdma_dev_res, - struct ibv_sge *sge, uint8_t num_sg= e, - uint64_t *total_length) -{ - int idx; - - for (idx =3D 0; idx < num_sge; idx++) { - *total_length +=3D sge[idx].length; - } - return 0; -} -#endif - -static void trace_mad_message(const char *title, char *buf, int len) -{ - int i; - char *b =3D g_malloc0(len * 3 + 1); - char b1[4]; - - for (i =3D 0; i < len; i++) { - sprintf(b1, "%.2X ", buf[i] & 0x000000FF); - strcat(b, b1); - } - - trace_rdma_mad_message(title, len, b); - - g_free(b); -} - -static int mad_send(RdmaBackendDev *backend_dev, uint8_t sgid_idx, - union ibv_gid *sgid, struct ibv_sge *sge, uint32_t num= _sge) -{ - RdmaCmMuxMsg msg =3D {}; - char *hdr, *data; - int ret; - - if (num_sge !=3D 2) { - return -EINVAL; - } - - msg.hdr.op_code =3D RDMACM_MUX_OP_CODE_MAD; - memcpy(msg.hdr.sgid.raw, sgid->raw, sizeof(msg.hdr.sgid)); - - msg.umad_len =3D sge[0].length + sge[1].length; - - if (msg.umad_len > sizeof(msg.umad.mad)) { - return -ENOMEM; - } - - msg.umad.hdr.addr.qpn =3D htobe32(1); - msg.umad.hdr.addr.grh_present =3D 1; - msg.umad.hdr.addr.gid_index =3D sgid_idx; - memcpy(msg.umad.hdr.addr.gid, sgid->raw, sizeof(msg.umad.hdr.addr.gid)= ); - msg.umad.hdr.addr.hop_limit =3D 0xFF; - - hdr =3D rdma_pci_dma_map(backend_dev->dev, sge[0].addr, sge[0].length); - if (!hdr) { - return -ENOMEM; - } - data =3D rdma_pci_dma_map(backend_dev->dev, sge[1].addr, sge[1].length= ); - if (!data) { - rdma_pci_dma_unmap(backend_dev->dev, hdr, sge[0].length); - return -ENOMEM; - } - - memcpy(&msg.umad.mad[0], hdr, sge[0].length); - memcpy(&msg.umad.mad[sge[0].length], data, sge[1].length); - - rdma_pci_dma_unmap(backend_dev->dev, data, sge[1].length); - rdma_pci_dma_unmap(backend_dev->dev, hdr, sge[0].length); - - trace_mad_message("send", msg.umad.mad, msg.umad_len); - - ret =3D rdmacm_mux_send(backend_dev, &msg); - if (ret) { - rdma_error_report("Failed to send MAD to rdma_umadmux (%d)", ret); - return -EIO; - } - - return 0; -} - -void rdma_backend_post_send(RdmaBackendDev *backend_dev, - RdmaBackendQP *qp, uint8_t qp_type, - struct ibv_sge *sge, uint32_t num_sge, - uint8_t sgid_idx, union ibv_gid *sgid, - union ibv_gid *dgid, uint32_t dqpn, uint32_t d= qkey, - void *ctx) -{ - BackendCtx *bctx; - uint32_t bctx_id; - int rc; - struct ibv_send_wr wr =3D {}, *bad_wr; - - if (!qp->ibqp) { /* This field is not initialized for QP0 and QP1 */ - if (qp_type =3D=3D IBV_QPT_SMI) { - rdma_error_report("Got QP0 request"); - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_QP0, ctx); - } else if (qp_type =3D=3D IBV_QPT_GSI) { - rc =3D mad_send(backend_dev, sgid_idx, sgid, sge, num_sge); - if (rc) { - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_MAD_SEND, ctx= ); - backend_dev->rdma_dev_res->stats.mad_tx_err++; - } else { - complete_work(IBV_WC_SUCCESS, 0, ctx); - backend_dev->rdma_dev_res->stats.mad_tx++; - } - } - return; - } - - bctx =3D g_malloc0(sizeof(*bctx)); - bctx->up_ctx =3D ctx; - bctx->backend_qp =3D qp; - - rc =3D rdma_rm_alloc_cqe_ctx(backend_dev->rdma_dev_res, &bctx_id, bctx= ); - if (unlikely(rc)) { - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_NOMEM, ctx); - goto err_free_bctx; - } - - rdma_protected_gslist_append_int32(&qp->cqe_ctx_list, bctx_id); - - rc =3D build_host_sge_array(backend_dev->rdma_dev_res, sge, num_sge, - &backend_dev->rdma_dev_res->stats.tx_len); - if (rc) { - complete_work(IBV_WC_GENERAL_ERR, rc, ctx); - goto err_dealloc_cqe_ctx; - } - - if (qp_type =3D=3D IBV_QPT_UD) { - wr.wr.ud.ah =3D create_ah(backend_dev, qp->ibpd, sgid_idx, dgid); - if (!wr.wr.ud.ah) { - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_FAIL_BACKEND, ctx= ); - goto err_dealloc_cqe_ctx; - } - wr.wr.ud.remote_qpn =3D dqpn; - wr.wr.ud.remote_qkey =3D dqkey; - } - - wr.num_sge =3D num_sge; - wr.opcode =3D IBV_WR_SEND; - wr.send_flags =3D IBV_SEND_SIGNALED; - wr.sg_list =3D sge; - wr.wr_id =3D bctx_id; - - rc =3D ibv_post_send(qp->ibqp, &wr, &bad_wr); - if (rc) { - rdma_error_report("ibv_post_send fail, qpn=3D0x%x, rc=3D%d, errno= =3D%d", - qp->ibqp->qp_num, rc, errno); - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_FAIL_BACKEND, ctx); - goto err_dealloc_cqe_ctx; - } - - qatomic_inc(&backend_dev->rdma_dev_res->stats.missing_cqe); - backend_dev->rdma_dev_res->stats.tx++; - - return; - -err_dealloc_cqe_ctx: - backend_dev->rdma_dev_res->stats.tx_err++; - rdma_rm_dealloc_cqe_ctx(backend_dev->rdma_dev_res, bctx_id); - -err_free_bctx: - g_free(bctx); -} - -static unsigned int save_mad_recv_buffer(RdmaBackendDev *backend_dev, - struct ibv_sge *sge, uint32_t num= _sge, - void *ctx) -{ - BackendCtx *bctx; - int rc; - uint32_t bctx_id; - - if (num_sge !=3D 1) { - rdma_error_report("Invalid num_sge (%d), expecting 1", num_sge); - return VENDOR_ERR_INV_NUM_SGE; - } - - if (sge[0].length < RDMA_MAX_PRIVATE_DATA + sizeof(struct ibv_grh)) { - rdma_error_report("Too small buffer for MAD"); - return VENDOR_ERR_INV_MAD_BUFF; - } - - bctx =3D g_malloc0(sizeof(*bctx)); - - rc =3D rdma_rm_alloc_cqe_ctx(backend_dev->rdma_dev_res, &bctx_id, bctx= ); - if (unlikely(rc)) { - g_free(bctx); - return VENDOR_ERR_NOMEM; - } - - bctx->up_ctx =3D ctx; - bctx->sge =3D *sge; - - rdma_protected_gqueue_append_int64(&backend_dev->recv_mads_list, bctx_= id); - - return 0; -} - -void rdma_backend_post_recv(RdmaBackendDev *backend_dev, - RdmaBackendQP *qp, uint8_t qp_type, - struct ibv_sge *sge, uint32_t num_sge, void *c= tx) -{ - BackendCtx *bctx; - uint32_t bctx_id; - int rc; - struct ibv_recv_wr wr =3D {}, *bad_wr; - - if (!qp->ibqp) { /* This field does not get initialized for QP0 and QP= 1 */ - if (qp_type =3D=3D IBV_QPT_SMI) { - rdma_error_report("Got QP0 request"); - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_QP0, ctx); - } - if (qp_type =3D=3D IBV_QPT_GSI) { - rc =3D save_mad_recv_buffer(backend_dev, sge, num_sge, ctx); - if (rc) { - complete_work(IBV_WC_GENERAL_ERR, rc, ctx); - backend_dev->rdma_dev_res->stats.mad_rx_bufs_err++; - } else { - backend_dev->rdma_dev_res->stats.mad_rx_bufs++; - } - } - return; - } - - bctx =3D g_malloc0(sizeof(*bctx)); - bctx->up_ctx =3D ctx; - bctx->backend_qp =3D qp; - - rc =3D rdma_rm_alloc_cqe_ctx(backend_dev->rdma_dev_res, &bctx_id, bctx= ); - if (unlikely(rc)) { - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_NOMEM, ctx); - goto err_free_bctx; - } - - rdma_protected_gslist_append_int32(&qp->cqe_ctx_list, bctx_id); - - rc =3D build_host_sge_array(backend_dev->rdma_dev_res, sge, num_sge, - &backend_dev->rdma_dev_res->stats.rx_bufs_le= n); - if (rc) { - complete_work(IBV_WC_GENERAL_ERR, rc, ctx); - goto err_dealloc_cqe_ctx; - } - - wr.num_sge =3D num_sge; - wr.sg_list =3D sge; - wr.wr_id =3D bctx_id; - rc =3D ibv_post_recv(qp->ibqp, &wr, &bad_wr); - if (rc) { - rdma_error_report("ibv_post_recv fail, qpn=3D0x%x, rc=3D%d, errno= =3D%d", - qp->ibqp->qp_num, rc, errno); - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_FAIL_BACKEND, ctx); - goto err_dealloc_cqe_ctx; - } - - qatomic_inc(&backend_dev->rdma_dev_res->stats.missing_cqe); - backend_dev->rdma_dev_res->stats.rx_bufs++; - - return; - -err_dealloc_cqe_ctx: - backend_dev->rdma_dev_res->stats.rx_bufs_err++; - rdma_rm_dealloc_cqe_ctx(backend_dev->rdma_dev_res, bctx_id); - -err_free_bctx: - g_free(bctx); -} - -void rdma_backend_post_srq_recv(RdmaBackendDev *backend_dev, - RdmaBackendSRQ *srq, struct ibv_sge *sge, - uint32_t num_sge, void *ctx) -{ - BackendCtx *bctx; - uint32_t bctx_id; - int rc; - struct ibv_recv_wr wr =3D {}, *bad_wr; - - bctx =3D g_malloc0(sizeof(*bctx)); - bctx->up_ctx =3D ctx; - bctx->backend_srq =3D srq; - - rc =3D rdma_rm_alloc_cqe_ctx(backend_dev->rdma_dev_res, &bctx_id, bctx= ); - if (unlikely(rc)) { - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_NOMEM, ctx); - goto err_free_bctx; - } - - rdma_protected_gslist_append_int32(&srq->cqe_ctx_list, bctx_id); - - rc =3D build_host_sge_array(backend_dev->rdma_dev_res, sge, num_sge, - &backend_dev->rdma_dev_res->stats.rx_bufs_le= n); - if (rc) { - complete_work(IBV_WC_GENERAL_ERR, rc, ctx); - goto err_dealloc_cqe_ctx; - } - - wr.num_sge =3D num_sge; - wr.sg_list =3D sge; - wr.wr_id =3D bctx_id; - rc =3D ibv_post_srq_recv(srq->ibsrq, &wr, &bad_wr); - if (rc) { - rdma_error_report("ibv_post_srq_recv fail, srqn=3D0x%x, rc=3D%d, e= rrno=3D%d", - srq->ibsrq->handle, rc, errno); - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_FAIL_BACKEND, ctx); - goto err_dealloc_cqe_ctx; - } - - qatomic_inc(&backend_dev->rdma_dev_res->stats.missing_cqe); - backend_dev->rdma_dev_res->stats.rx_bufs++; - backend_dev->rdma_dev_res->stats.rx_srq++; - - return; - -err_dealloc_cqe_ctx: - backend_dev->rdma_dev_res->stats.rx_bufs_err++; - rdma_rm_dealloc_cqe_ctx(backend_dev->rdma_dev_res, bctx_id); - -err_free_bctx: - g_free(bctx); -} - -int rdma_backend_create_pd(RdmaBackendDev *backend_dev, RdmaBackendPD *pd) -{ - pd->ibpd =3D ibv_alloc_pd(backend_dev->context); - - if (!pd->ibpd) { - rdma_error_report("ibv_alloc_pd fail, errno=3D%d", errno); - return -EIO; - } - - return 0; -} - -void rdma_backend_destroy_pd(RdmaBackendPD *pd) -{ - if (pd->ibpd) { - ibv_dealloc_pd(pd->ibpd); - } -} - -int rdma_backend_create_mr(RdmaBackendMR *mr, RdmaBackendPD *pd, void *add= r, - size_t length, uint64_t guest_start, int access) -{ -#ifdef LEGACY_RDMA_REG_MR - mr->ibmr =3D ibv_reg_mr(pd->ibpd, addr, length, access); -#else - mr->ibmr =3D ibv_reg_mr_iova(pd->ibpd, addr, length, guest_start, acce= ss); -#endif - if (!mr->ibmr) { - rdma_error_report("ibv_reg_mr fail, errno=3D%d", errno); - return -EIO; - } - - mr->ibpd =3D pd->ibpd; - - return 0; -} - -void rdma_backend_destroy_mr(RdmaBackendMR *mr) -{ - if (mr->ibmr) { - ibv_dereg_mr(mr->ibmr); - } -} - -int rdma_backend_create_cq(RdmaBackendDev *backend_dev, RdmaBackendCQ *cq, - int cqe) -{ - int rc; - - cq->ibcq =3D ibv_create_cq(backend_dev->context, cqe + 1, NULL, - backend_dev->channel, 0); - if (!cq->ibcq) { - rdma_error_report("ibv_create_cq fail, errno=3D%d", errno); - return -EIO; - } - - rc =3D ibv_req_notify_cq(cq->ibcq, 0); - if (rc) { - rdma_warn_report("ibv_req_notify_cq fail, rc=3D%d, errno=3D%d", rc= , errno); - } - - cq->backend_dev =3D backend_dev; - - return 0; -} - -void rdma_backend_destroy_cq(RdmaBackendCQ *cq) -{ - if (cq->ibcq) { - ibv_destroy_cq(cq->ibcq); - } -} - -int rdma_backend_create_qp(RdmaBackendQP *qp, uint8_t qp_type, - RdmaBackendPD *pd, RdmaBackendCQ *scq, - RdmaBackendCQ *rcq, RdmaBackendSRQ *srq, - uint32_t max_send_wr, uint32_t max_recv_wr, - uint32_t max_send_sge, uint32_t max_recv_sge) -{ - struct ibv_qp_init_attr attr =3D {}; - - qp->ibqp =3D 0; - - switch (qp_type) { - case IBV_QPT_GSI: - return 0; - - case IBV_QPT_RC: - /* fall through */ - case IBV_QPT_UD: - /* do nothing */ - break; - - default: - rdma_error_report("Unsupported QP type %d", qp_type); - return -EIO; - } - - attr.qp_type =3D qp_type; - attr.send_cq =3D scq->ibcq; - attr.recv_cq =3D rcq->ibcq; - attr.cap.max_send_wr =3D max_send_wr; - attr.cap.max_recv_wr =3D max_recv_wr; - attr.cap.max_send_sge =3D max_send_sge; - attr.cap.max_recv_sge =3D max_recv_sge; - if (srq) { - attr.srq =3D srq->ibsrq; - } - - qp->ibqp =3D ibv_create_qp(pd->ibpd, &attr); - if (!qp->ibqp) { - rdma_error_report("ibv_create_qp fail, errno=3D%d", errno); - return -EIO; - } - - rdma_protected_gslist_init(&qp->cqe_ctx_list); - - qp->ibpd =3D pd->ibpd; - - /* TODO: Query QP to get max_inline_data and save it to be used in sen= d */ - - return 0; -} - -int rdma_backend_qp_state_init(RdmaBackendDev *backend_dev, RdmaBackendQP = *qp, - uint8_t qp_type, uint32_t qkey) -{ - struct ibv_qp_attr attr =3D {}; - int rc, attr_mask; - - attr_mask =3D IBV_QP_STATE | IBV_QP_PKEY_INDEX | IBV_QP_PORT; - attr.qp_state =3D IBV_QPS_INIT; - attr.pkey_index =3D 0; - attr.port_num =3D backend_dev->port_num; - - switch (qp_type) { - case IBV_QPT_RC: - attr_mask |=3D IBV_QP_ACCESS_FLAGS; - trace_rdma_backend_rc_qp_state_init(qp->ibqp->qp_num); - break; - - case IBV_QPT_UD: - attr.qkey =3D qkey; - attr_mask |=3D IBV_QP_QKEY; - trace_rdma_backend_ud_qp_state_init(qp->ibqp->qp_num, qkey); - break; - - default: - rdma_error_report("Unsupported QP type %d", qp_type); - return -EIO; - } - - rc =3D ibv_modify_qp(qp->ibqp, &attr, attr_mask); - if (rc) { - rdma_error_report("ibv_modify_qp fail, rc=3D%d, errno=3D%d", rc, e= rrno); - return -EIO; - } - - return 0; -} - -int rdma_backend_qp_state_rtr(RdmaBackendDev *backend_dev, RdmaBackendQP *= qp, - uint8_t qp_type, uint8_t sgid_idx, - union ibv_gid *dgid, uint32_t dqpn, - uint32_t rq_psn, uint32_t qkey, bool use_qke= y) -{ - struct ibv_qp_attr attr =3D {}; - union ibv_gid ibv_gid =3D { - .global.interface_id =3D dgid->global.interface_id, - .global.subnet_prefix =3D dgid->global.subnet_prefix - }; - int rc, attr_mask; - - attr.qp_state =3D IBV_QPS_RTR; - attr_mask =3D IBV_QP_STATE; - - qp->sgid_idx =3D sgid_idx; - - switch (qp_type) { - case IBV_QPT_RC: - attr.path_mtu =3D IBV_MTU_1024; - attr.dest_qp_num =3D dqpn; - attr.max_dest_rd_atomic =3D 1; - attr.min_rnr_timer =3D 12; - attr.ah_attr.port_num =3D backend_dev->port_num; - attr.ah_attr.is_global =3D 1; - attr.ah_attr.grh.hop_limit =3D 1; - attr.ah_attr.grh.dgid =3D ibv_gid; - attr.ah_attr.grh.sgid_index =3D qp->sgid_idx; - attr.rq_psn =3D rq_psn; - - attr_mask |=3D IBV_QP_AV | IBV_QP_PATH_MTU | IBV_QP_DEST_QPN | - IBV_QP_RQ_PSN | IBV_QP_MAX_DEST_RD_ATOMIC | - IBV_QP_MIN_RNR_TIMER; - - trace_rdma_backend_rc_qp_state_rtr(qp->ibqp->qp_num, - be64_to_cpu(ibv_gid.global. - subnet_prefix), - be64_to_cpu(ibv_gid.global. - interface_id), - qp->sgid_idx, dqpn, rq_psn); - break; - - case IBV_QPT_UD: - if (use_qkey) { - attr.qkey =3D qkey; - attr_mask |=3D IBV_QP_QKEY; - } - trace_rdma_backend_ud_qp_state_rtr(qp->ibqp->qp_num, use_qkey ? qk= ey : - 0); - break; - } - - rc =3D ibv_modify_qp(qp->ibqp, &attr, attr_mask); - if (rc) { - rdma_error_report("ibv_modify_qp fail, rc=3D%d, errno=3D%d", rc, e= rrno); - return -EIO; - } - - return 0; -} - -int rdma_backend_qp_state_rts(RdmaBackendQP *qp, uint8_t qp_type, - uint32_t sq_psn, uint32_t qkey, bool use_qke= y) -{ - struct ibv_qp_attr attr =3D {}; - int rc, attr_mask; - - attr.qp_state =3D IBV_QPS_RTS; - attr.sq_psn =3D sq_psn; - attr_mask =3D IBV_QP_STATE | IBV_QP_SQ_PSN; - - switch (qp_type) { - case IBV_QPT_RC: - attr.timeout =3D 14; - attr.retry_cnt =3D 7; - attr.rnr_retry =3D 7; - attr.max_rd_atomic =3D 1; - - attr_mask |=3D IBV_QP_TIMEOUT | IBV_QP_RETRY_CNT | IBV_QP_RNR_RETR= Y | - IBV_QP_MAX_QP_RD_ATOMIC; - trace_rdma_backend_rc_qp_state_rts(qp->ibqp->qp_num, sq_psn); - break; - - case IBV_QPT_UD: - if (use_qkey) { - attr.qkey =3D qkey; - attr_mask |=3D IBV_QP_QKEY; - } - trace_rdma_backend_ud_qp_state_rts(qp->ibqp->qp_num, sq_psn, - use_qkey ? qkey : 0); - break; - } - - rc =3D ibv_modify_qp(qp->ibqp, &attr, attr_mask); - if (rc) { - rdma_error_report("ibv_modify_qp fail, rc=3D%d, errno=3D%d", rc, e= rrno); - return -EIO; - } - - return 0; -} - -int rdma_backend_query_qp(RdmaBackendQP *qp, struct ibv_qp_attr *attr, - int attr_mask, struct ibv_qp_init_attr *init_att= r) -{ - if (!qp->ibqp) { - attr->qp_state =3D IBV_QPS_RTS; - return 0; - } - - return ibv_query_qp(qp->ibqp, attr, attr_mask, init_attr); -} - -void rdma_backend_destroy_qp(RdmaBackendQP *qp, RdmaDeviceResources *dev_r= es) -{ - if (qp->ibqp) { - ibv_destroy_qp(qp->ibqp); - } - g_slist_foreach(qp->cqe_ctx_list.list, free_cqe_ctx, dev_res); - rdma_protected_gslist_destroy(&qp->cqe_ctx_list); -} - -int rdma_backend_create_srq(RdmaBackendSRQ *srq, RdmaBackendPD *pd, - uint32_t max_wr, uint32_t max_sge, - uint32_t srq_limit) -{ - struct ibv_srq_init_attr srq_init_attr =3D {}; - - srq_init_attr.attr.max_wr =3D max_wr; - srq_init_attr.attr.max_sge =3D max_sge; - srq_init_attr.attr.srq_limit =3D srq_limit; - - srq->ibsrq =3D ibv_create_srq(pd->ibpd, &srq_init_attr); - if (!srq->ibsrq) { - rdma_error_report("ibv_create_srq failed, errno=3D%d", errno); - return -EIO; - } - - rdma_protected_gslist_init(&srq->cqe_ctx_list); - - return 0; -} - -int rdma_backend_query_srq(RdmaBackendSRQ *srq, struct ibv_srq_attr *srq_a= ttr) -{ - if (!srq->ibsrq) { - return -EINVAL; - } - - return ibv_query_srq(srq->ibsrq, srq_attr); -} - -int rdma_backend_modify_srq(RdmaBackendSRQ *srq, struct ibv_srq_attr *srq_= attr, - int srq_attr_mask) -{ - if (!srq->ibsrq) { - return -EINVAL; - } - - return ibv_modify_srq(srq->ibsrq, srq_attr, srq_attr_mask); -} - -void rdma_backend_destroy_srq(RdmaBackendSRQ *srq, RdmaDeviceResources *de= v_res) -{ - if (srq->ibsrq) { - ibv_destroy_srq(srq->ibsrq); - } - g_slist_foreach(srq->cqe_ctx_list.list, free_cqe_ctx, dev_res); - rdma_protected_gslist_destroy(&srq->cqe_ctx_list); -} - -#define CHK_ATTR(req, dev, member, fmt) ({ \ - trace_rdma_check_dev_attr(#member, dev.member, req->member); \ - if (req->member > dev.member) { \ - rdma_warn_report("%s =3D "fmt" is higher than host device capabili= ty "fmt, \ - #member, req->member, dev.member); \ - req->member =3D dev.member; \ - } \ -}) - -static int init_device_caps(RdmaBackendDev *backend_dev, - struct ibv_device_attr *dev_attr) -{ - struct ibv_device_attr bk_dev_attr; - int rc; - - rc =3D ibv_query_device(backend_dev->context, &bk_dev_attr); - if (rc) { - rdma_error_report("ibv_query_device fail, rc=3D%d, errno=3D%d", rc= , errno); - return -EIO; - } - - dev_attr->max_sge =3D MAX_SGE; - dev_attr->max_srq_sge =3D MAX_SGE; - - CHK_ATTR(dev_attr, bk_dev_attr, max_mr_size, "%" PRId64); - CHK_ATTR(dev_attr, bk_dev_attr, max_qp, "%d"); - CHK_ATTR(dev_attr, bk_dev_attr, max_sge, "%d"); - CHK_ATTR(dev_attr, bk_dev_attr, max_cq, "%d"); - CHK_ATTR(dev_attr, bk_dev_attr, max_mr, "%d"); - CHK_ATTR(dev_attr, bk_dev_attr, max_pd, "%d"); - CHK_ATTR(dev_attr, bk_dev_attr, max_qp_rd_atom, "%d"); - CHK_ATTR(dev_attr, bk_dev_attr, max_qp_init_rd_atom, "%d"); - CHK_ATTR(dev_attr, bk_dev_attr, max_ah, "%d"); - CHK_ATTR(dev_attr, bk_dev_attr, max_srq, "%d"); - - return 0; -} - -static inline void build_mad_hdr(struct ibv_grh *grh, union ibv_gid *sgid, - union ibv_gid *my_gid, int paylen) -{ - grh->paylen =3D htons(paylen); - grh->sgid =3D *sgid; - grh->dgid =3D *my_gid; -} - -static void process_incoming_mad_req(RdmaBackendDev *backend_dev, - RdmaCmMuxMsg *msg) -{ - unsigned long cqe_ctx_id; - BackendCtx *bctx; - char *mad; - - trace_mad_message("recv", msg->umad.mad, msg->umad_len); - - cqe_ctx_id =3D rdma_protected_gqueue_pop_int64(&backend_dev->recv_mads= _list); - if (cqe_ctx_id =3D=3D -ENOENT) { - rdma_warn_report("No more free MADs buffers, waiting for a while"); - sleep(THR_POLL_TO); - return; - } - - bctx =3D rdma_rm_get_cqe_ctx(backend_dev->rdma_dev_res, cqe_ctx_id); - if (unlikely(!bctx)) { - rdma_error_report("No matching ctx for req %ld", cqe_ctx_id); - backend_dev->rdma_dev_res->stats.mad_rx_err++; - return; - } - - mad =3D rdma_pci_dma_map(backend_dev->dev, bctx->sge.addr, - bctx->sge.length); - if (!mad || bctx->sge.length < msg->umad_len + MAD_HDR_SIZE) { - backend_dev->rdma_dev_res->stats.mad_rx_err++; - complete_work(IBV_WC_GENERAL_ERR, VENDOR_ERR_INV_MAD_BUFF, - bctx->up_ctx); - } else { - struct ibv_wc wc =3D {}; - memset(mad, 0, bctx->sge.length); - build_mad_hdr((struct ibv_grh *)mad, - (union ibv_gid *)&msg->umad.hdr.addr.gid, &msg->hdr.= sgid, - msg->umad_len); - memcpy(&mad[MAD_HDR_SIZE], msg->umad.mad, msg->umad_len); - rdma_pci_dma_unmap(backend_dev->dev, mad, bctx->sge.length); - - wc.byte_len =3D msg->umad_len; - wc.status =3D IBV_WC_SUCCESS; - wc.wc_flags =3D IBV_WC_GRH; - backend_dev->rdma_dev_res->stats.mad_rx++; - comp_handler(bctx->up_ctx, &wc); - } - - g_free(bctx); - rdma_rm_dealloc_cqe_ctx(backend_dev->rdma_dev_res, cqe_ctx_id); -} - -static inline int rdmacm_mux_can_receive(void *opaque) -{ - RdmaBackendDev *backend_dev =3D (RdmaBackendDev *)opaque; - - return rdmacm_mux_can_process_async(backend_dev); -} - -static void rdmacm_mux_read(void *opaque, const uint8_t *buf, int size) -{ - RdmaBackendDev *backend_dev =3D (RdmaBackendDev *)opaque; - RdmaCmMuxMsg *msg =3D (RdmaCmMuxMsg *)buf; - - trace_rdmacm_mux("read", msg->hdr.msg_type, msg->hdr.op_code); - - if (msg->hdr.msg_type !=3D RDMACM_MUX_MSG_TYPE_REQ && - msg->hdr.op_code !=3D RDMACM_MUX_OP_CODE_MAD) { - rdma_error_report("Error: Not a MAD request, skipping"); - return; - } - process_incoming_mad_req(backend_dev, msg); -} - -static int mad_init(RdmaBackendDev *backend_dev, CharBackend *mad_chr_be) -{ - int ret; - - backend_dev->rdmacm_mux.chr_be =3D mad_chr_be; - - ret =3D qemu_chr_fe_backend_connected(backend_dev->rdmacm_mux.chr_be); - if (!ret) { - rdma_error_report("Missing chardev for MAD multiplexer"); - return -EIO; - } - - rdma_protected_gqueue_init(&backend_dev->recv_mads_list); - - enable_rdmacm_mux_async(backend_dev); - - qemu_chr_fe_set_handlers(backend_dev->rdmacm_mux.chr_be, - rdmacm_mux_can_receive, rdmacm_mux_read, NULL, - NULL, backend_dev, NULL, true); - - return 0; -} - -static void mad_stop(RdmaBackendDev *backend_dev) -{ - clean_recv_mads(backend_dev); -} - -static void mad_fini(RdmaBackendDev *backend_dev) -{ - disable_rdmacm_mux_async(backend_dev); - qemu_chr_fe_disconnect(backend_dev->rdmacm_mux.chr_be); - rdma_protected_gqueue_destroy(&backend_dev->recv_mads_list); -} - -int rdma_backend_get_gid_index(RdmaBackendDev *backend_dev, - union ibv_gid *gid) -{ - union ibv_gid sgid; - int ret; - int i =3D 0; - - do { - ret =3D ibv_query_gid(backend_dev->context, backend_dev->port_num,= i, - &sgid); - i++; - } while (!ret && (memcmp(&sgid, gid, sizeof(*gid)))); - - trace_rdma_backend_get_gid_index(be64_to_cpu(gid->global.subnet_prefix= ), - be64_to_cpu(gid->global.interface_id), - i - 1); - - return ret ? ret : i - 1; -} - -int rdma_backend_add_gid(RdmaBackendDev *backend_dev, const char *ifname, - union ibv_gid *gid) -{ - RdmaCmMuxMsg msg =3D {}; - int ret; - - trace_rdma_backend_gid_change("add", be64_to_cpu(gid->global.subnet_pr= efix), - be64_to_cpu(gid->global.interface_id)); - - msg.hdr.op_code =3D RDMACM_MUX_OP_CODE_REG; - memcpy(msg.hdr.sgid.raw, gid->raw, sizeof(msg.hdr.sgid)); - - ret =3D rdmacm_mux_send(backend_dev, &msg); - if (ret) { - rdma_error_report("Failed to register GID to rdma_umadmux (%d)", r= et); - return -EIO; - } - - qapi_event_send_rdma_gid_status_changed(ifname, true, - gid->global.subnet_prefix, - gid->global.interface_id); - - return ret; -} - -int rdma_backend_del_gid(RdmaBackendDev *backend_dev, const char *ifname, - union ibv_gid *gid) -{ - RdmaCmMuxMsg msg =3D {}; - int ret; - - trace_rdma_backend_gid_change("del", be64_to_cpu(gid->global.subnet_pr= efix), - be64_to_cpu(gid->global.interface_id)); - - msg.hdr.op_code =3D RDMACM_MUX_OP_CODE_UNREG; - memcpy(msg.hdr.sgid.raw, gid->raw, sizeof(msg.hdr.sgid)); - - ret =3D rdmacm_mux_send(backend_dev, &msg); - if (ret) { - rdma_error_report("Failed to unregister GID from rdma_umadmux (%d)= ", - ret); - return -EIO; - } - - qapi_event_send_rdma_gid_status_changed(ifname, false, - gid->global.subnet_prefix, - gid->global.interface_id); - - return 0; -} - -int rdma_backend_init(RdmaBackendDev *backend_dev, PCIDevice *pdev, - RdmaDeviceResources *rdma_dev_res, - const char *backend_device_name, uint8_t port_num, - struct ibv_device_attr *dev_attr, CharBackend *mad_c= hr_be) -{ - int i; - int ret =3D 0; - int num_ibv_devices; - struct ibv_device **dev_list; - - memset(backend_dev, 0, sizeof(*backend_dev)); - - backend_dev->dev =3D pdev; - backend_dev->port_num =3D port_num; - backend_dev->rdma_dev_res =3D rdma_dev_res; - - rdma_backend_register_comp_handler(dummy_comp_handler); - - dev_list =3D ibv_get_device_list(&num_ibv_devices); - if (!dev_list) { - rdma_error_report("Failed to get IB devices list"); - return -EIO; - } - - if (num_ibv_devices =3D=3D 0) { - rdma_error_report("No IB devices were found"); - ret =3D -ENXIO; - goto out_free_dev_list; - } - - if (backend_device_name) { - for (i =3D 0; dev_list[i]; ++i) { - if (!strcmp(ibv_get_device_name(dev_list[i]), - backend_device_name)) { - break; - } - } - - backend_dev->ib_dev =3D dev_list[i]; - if (!backend_dev->ib_dev) { - rdma_error_report("Failed to find IB device %s", - backend_device_name); - ret =3D -EIO; - goto out_free_dev_list; - } - } else { - backend_dev->ib_dev =3D *dev_list; - } - - rdma_info_report("uverb device %s", backend_dev->ib_dev->dev_name); - - backend_dev->context =3D ibv_open_device(backend_dev->ib_dev); - if (!backend_dev->context) { - rdma_error_report("Failed to open IB device %s", - ibv_get_device_name(backend_dev->ib_dev)); - ret =3D -EIO; - goto out; - } - - backend_dev->channel =3D ibv_create_comp_channel(backend_dev->context); - if (!backend_dev->channel) { - rdma_error_report("Failed to create IB communication channel"); - ret =3D -EIO; - goto out_close_device; - } - - ret =3D init_device_caps(backend_dev, dev_attr); - if (ret) { - rdma_error_report("Failed to initialize device capabilities"); - ret =3D -EIO; - goto out_destroy_comm_channel; - } - - - ret =3D mad_init(backend_dev, mad_chr_be); - if (ret) { - rdma_error_report("Failed to initialize mad"); - ret =3D -EIO; - goto out_destroy_comm_channel; - } - - backend_dev->comp_thread.run =3D false; - backend_dev->comp_thread.is_running =3D false; - - ah_cache_init(); - - goto out_free_dev_list; - -out_destroy_comm_channel: - ibv_destroy_comp_channel(backend_dev->channel); - -out_close_device: - ibv_close_device(backend_dev->context); - -out_free_dev_list: - ibv_free_device_list(dev_list); - -out: - return ret; -} - - -void rdma_backend_start(RdmaBackendDev *backend_dev) -{ - start_comp_thread(backend_dev); -} - -void rdma_backend_stop(RdmaBackendDev *backend_dev) -{ - mad_stop(backend_dev); - stop_backend_thread(&backend_dev->comp_thread); -} - -void rdma_backend_fini(RdmaBackendDev *backend_dev) -{ - mad_fini(backend_dev); - g_hash_table_destroy(ah_hash); - ibv_destroy_comp_channel(backend_dev->channel); - ibv_close_device(backend_dev->context); -} diff --git a/hw/rdma/rdma_rm.c b/hw/rdma/rdma_rm.c deleted file mode 100644 index 038d564433..0000000000 --- a/hw/rdma/rdma_rm.c +++ /dev/null @@ -1,812 +0,0 @@ -/* - * QEMU paravirtual RDMA - Resource Manager Implementation - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include "qapi/error.h" -#include "cpu.h" -#include "monitor/monitor.h" - -#include "trace.h" -#include "rdma_utils.h" -#include "rdma_backend.h" -#include "rdma_rm.h" - -void rdma_format_device_counters(RdmaDeviceResources *dev_res, GString *bu= f) -{ - g_string_append_printf(buf, "\ttx : %" PRId64 "\n", - dev_res->stats.tx); - g_string_append_printf(buf, "\ttx_len : %" PRId64 "\n", - dev_res->stats.tx_len); - g_string_append_printf(buf, "\ttx_err : %" PRId64 "\n", - dev_res->stats.tx_err); - g_string_append_printf(buf, "\trx_bufs : %" PRId64 "\n", - dev_res->stats.rx_bufs); - g_string_append_printf(buf, "\trx_srq : %" PRId64 "\n", - dev_res->stats.rx_srq); - g_string_append_printf(buf, "\trx_bufs_len : %" PRId64 "\n", - dev_res->stats.rx_bufs_len); - g_string_append_printf(buf, "\trx_bufs_err : %" PRId64 "\n", - dev_res->stats.rx_bufs_err); - g_string_append_printf(buf, "\tcomps : %" PRId64 "\n", - dev_res->stats.completions); - g_string_append_printf(buf, "\tmissing_comps : %" PRId32 "\n", - dev_res->stats.missing_cqe); - g_string_append_printf(buf, "\tpoll_cq (bk) : %" PRId64 "\n", - dev_res->stats.poll_cq_from_bk); - g_string_append_printf(buf, "\tpoll_cq_ppoll_to : %" PRId64 "\n", - dev_res->stats.poll_cq_ppoll_to); - g_string_append_printf(buf, "\tpoll_cq (fe) : %" PRId64 "\n", - dev_res->stats.poll_cq_from_guest); - g_string_append_printf(buf, "\tpoll_cq_empty : %" PRId64 "\n", - dev_res->stats.poll_cq_from_guest_empty); - g_string_append_printf(buf, "\tmad_tx : %" PRId64 "\n", - dev_res->stats.mad_tx); - g_string_append_printf(buf, "\tmad_tx_err : %" PRId64 "\n", - dev_res->stats.mad_tx_err); - g_string_append_printf(buf, "\tmad_rx : %" PRId64 "\n", - dev_res->stats.mad_rx); - g_string_append_printf(buf, "\tmad_rx_err : %" PRId64 "\n", - dev_res->stats.mad_rx_err); - g_string_append_printf(buf, "\tmad_rx_bufs : %" PRId64 "\n", - dev_res->stats.mad_rx_bufs); - g_string_append_printf(buf, "\tmad_rx_bufs_err : %" PRId64 "\n", - dev_res->stats.mad_rx_bufs_err); - g_string_append_printf(buf, "\tPDs : %" PRId32 "\n", - dev_res->pd_tbl.used); - g_string_append_printf(buf, "\tMRs : %" PRId32 "\n", - dev_res->mr_tbl.used); - g_string_append_printf(buf, "\tUCs : %" PRId32 "\n", - dev_res->uc_tbl.used); - g_string_append_printf(buf, "\tQPs : %" PRId32 "\n", - dev_res->qp_tbl.used); - g_string_append_printf(buf, "\tCQs : %" PRId32 "\n", - dev_res->cq_tbl.used); - g_string_append_printf(buf, "\tCEQ_CTXs : %" PRId32 "\n", - dev_res->cqe_ctx_tbl.used); -} - -static inline void res_tbl_init(const char *name, RdmaRmResTbl *tbl, - uint32_t tbl_sz, uint32_t res_sz) -{ - tbl->tbl =3D g_malloc(tbl_sz * res_sz); - - strncpy(tbl->name, name, MAX_RM_TBL_NAME); - tbl->name[MAX_RM_TBL_NAME - 1] =3D 0; - - tbl->bitmap =3D bitmap_new(tbl_sz); - tbl->tbl_sz =3D tbl_sz; - tbl->res_sz =3D res_sz; - tbl->used =3D 0; - qemu_mutex_init(&tbl->lock); -} - -static inline void res_tbl_free(RdmaRmResTbl *tbl) -{ - if (!tbl->bitmap) { - return; - } - qemu_mutex_destroy(&tbl->lock); - g_free(tbl->tbl); - g_free(tbl->bitmap); -} - -static inline void *rdma_res_tbl_get(RdmaRmResTbl *tbl, uint32_t handle) -{ - trace_rdma_res_tbl_get(tbl->name, handle); - - if ((handle < tbl->tbl_sz) && (test_bit(handle, tbl->bitmap))) { - return tbl->tbl + handle * tbl->res_sz; - } else { - rdma_error_report("Table %s, invalid handle %d", tbl->name, handle= ); - return NULL; - } -} - -static inline void *rdma_res_tbl_alloc(RdmaRmResTbl *tbl, uint32_t *handle) -{ - qemu_mutex_lock(&tbl->lock); - - *handle =3D find_first_zero_bit(tbl->bitmap, tbl->tbl_sz); - if (*handle > tbl->tbl_sz) { - rdma_error_report("Table %s, failed to allocate, bitmap is full", - tbl->name); - qemu_mutex_unlock(&tbl->lock); - return NULL; - } - - set_bit(*handle, tbl->bitmap); - - tbl->used++; - - qemu_mutex_unlock(&tbl->lock); - - memset(tbl->tbl + *handle * tbl->res_sz, 0, tbl->res_sz); - - trace_rdma_res_tbl_alloc(tbl->name, *handle); - - return tbl->tbl + *handle * tbl->res_sz; -} - -static inline void rdma_res_tbl_dealloc(RdmaRmResTbl *tbl, uint32_t handle) -{ - trace_rdma_res_tbl_dealloc(tbl->name, handle); - - QEMU_LOCK_GUARD(&tbl->lock); - - if (handle < tbl->tbl_sz) { - clear_bit(handle, tbl->bitmap); - tbl->used--; - } - -} - -int rdma_rm_alloc_pd(RdmaDeviceResources *dev_res, RdmaBackendDev *backend= _dev, - uint32_t *pd_handle, uint32_t ctx_handle) -{ - RdmaRmPD *pd; - int ret =3D -ENOMEM; - - pd =3D rdma_res_tbl_alloc(&dev_res->pd_tbl, pd_handle); - if (!pd) { - goto out; - } - - ret =3D rdma_backend_create_pd(backend_dev, &pd->backend_pd); - if (ret) { - ret =3D -EIO; - goto out_tbl_dealloc; - } - - pd->ctx_handle =3D ctx_handle; - - return 0; - -out_tbl_dealloc: - rdma_res_tbl_dealloc(&dev_res->pd_tbl, *pd_handle); - -out: - return ret; -} - -RdmaRmPD *rdma_rm_get_pd(RdmaDeviceResources *dev_res, uint32_t pd_handle) -{ - return rdma_res_tbl_get(&dev_res->pd_tbl, pd_handle); -} - -void rdma_rm_dealloc_pd(RdmaDeviceResources *dev_res, uint32_t pd_handle) -{ - RdmaRmPD *pd =3D rdma_rm_get_pd(dev_res, pd_handle); - - if (pd) { - rdma_backend_destroy_pd(&pd->backend_pd); - rdma_res_tbl_dealloc(&dev_res->pd_tbl, pd_handle); - } -} - -int rdma_rm_alloc_mr(RdmaDeviceResources *dev_res, uint32_t pd_handle, - uint64_t guest_start, uint64_t guest_length, - void *host_virt, int access_flags, uint32_t *mr_handl= e, - uint32_t *lkey, uint32_t *rkey) -{ - RdmaRmMR *mr; - int ret =3D 0; - RdmaRmPD *pd; - - pd =3D rdma_rm_get_pd(dev_res, pd_handle); - if (!pd) { - return -EINVAL; - } - - mr =3D rdma_res_tbl_alloc(&dev_res->mr_tbl, mr_handle); - if (!mr) { - return -ENOMEM; - } - trace_rdma_rm_alloc_mr(*mr_handle, host_virt, guest_start, guest_lengt= h, - access_flags); - - if (host_virt) { - mr->virt =3D host_virt; - mr->start =3D guest_start; - mr->length =3D guest_length; - mr->virt +=3D (mr->start & (TARGET_PAGE_SIZE - 1)); - - ret =3D rdma_backend_create_mr(&mr->backend_mr, &pd->backend_pd, m= r->virt, - mr->length, guest_start, access_flags= ); - if (ret) { - ret =3D -EIO; - goto out_dealloc_mr; - } -#ifdef LEGACY_RDMA_REG_MR - /* We keep mr_handle in lkey so send and recv get get mr ptr */ - *lkey =3D *mr_handle; -#else - *lkey =3D rdma_backend_mr_lkey(&mr->backend_mr); -#endif - } - - *rkey =3D -1; - - mr->pd_handle =3D pd_handle; - - return 0; - -out_dealloc_mr: - rdma_res_tbl_dealloc(&dev_res->mr_tbl, *mr_handle); - - return ret; -} - -RdmaRmMR *rdma_rm_get_mr(RdmaDeviceResources *dev_res, uint32_t mr_handle) -{ - return rdma_res_tbl_get(&dev_res->mr_tbl, mr_handle); -} - -void rdma_rm_dealloc_mr(RdmaDeviceResources *dev_res, uint32_t mr_handle) -{ - RdmaRmMR *mr =3D rdma_rm_get_mr(dev_res, mr_handle); - - if (mr) { - rdma_backend_destroy_mr(&mr->backend_mr); - trace_rdma_rm_dealloc_mr(mr_handle, mr->start); - if (mr->start) { - mr->virt -=3D (mr->start & (TARGET_PAGE_SIZE - 1)); - munmap(mr->virt, mr->length); - } - rdma_res_tbl_dealloc(&dev_res->mr_tbl, mr_handle); - } -} - -int rdma_rm_alloc_uc(RdmaDeviceResources *dev_res, uint32_t pfn, - uint32_t *uc_handle) -{ - RdmaRmUC *uc; - - /* TODO: Need to make sure pfn is between bar start address and - * bsd+RDMA_BAR2_UAR_SIZE - if (pfn > RDMA_BAR2_UAR_SIZE) { - rdma_error_report("pfn out of range (%d > %d)", pfn, - RDMA_BAR2_UAR_SIZE); - return -ENOMEM; - } - */ - - uc =3D rdma_res_tbl_alloc(&dev_res->uc_tbl, uc_handle); - if (!uc) { - return -ENOMEM; - } - - return 0; -} - -RdmaRmUC *rdma_rm_get_uc(RdmaDeviceResources *dev_res, uint32_t uc_handle) -{ - return rdma_res_tbl_get(&dev_res->uc_tbl, uc_handle); -} - -void rdma_rm_dealloc_uc(RdmaDeviceResources *dev_res, uint32_t uc_handle) -{ - RdmaRmUC *uc =3D rdma_rm_get_uc(dev_res, uc_handle); - - if (uc) { - rdma_res_tbl_dealloc(&dev_res->uc_tbl, uc_handle); - } -} - -RdmaRmCQ *rdma_rm_get_cq(RdmaDeviceResources *dev_res, uint32_t cq_handle) -{ - return rdma_res_tbl_get(&dev_res->cq_tbl, cq_handle); -} - -int rdma_rm_alloc_cq(RdmaDeviceResources *dev_res, RdmaBackendDev *backend= _dev, - uint32_t cqe, uint32_t *cq_handle, void *opaque) -{ - int rc; - RdmaRmCQ *cq; - - cq =3D rdma_res_tbl_alloc(&dev_res->cq_tbl, cq_handle); - if (!cq) { - return -ENOMEM; - } - - cq->opaque =3D opaque; - cq->notify =3D CNT_CLEAR; - - rc =3D rdma_backend_create_cq(backend_dev, &cq->backend_cq, cqe); - if (rc) { - rc =3D -EIO; - goto out_dealloc_cq; - } - - return 0; - -out_dealloc_cq: - rdma_rm_dealloc_cq(dev_res, *cq_handle); - - return rc; -} - -void rdma_rm_req_notify_cq(RdmaDeviceResources *dev_res, uint32_t cq_handl= e, - bool notify) -{ - RdmaRmCQ *cq; - - cq =3D rdma_rm_get_cq(dev_res, cq_handle); - if (!cq) { - return; - } - - if (cq->notify !=3D CNT_SET) { - cq->notify =3D notify ? CNT_ARM : CNT_CLEAR; - } -} - -void rdma_rm_dealloc_cq(RdmaDeviceResources *dev_res, uint32_t cq_handle) -{ - RdmaRmCQ *cq; - - cq =3D rdma_rm_get_cq(dev_res, cq_handle); - if (!cq) { - return; - } - - rdma_backend_destroy_cq(&cq->backend_cq); - - rdma_res_tbl_dealloc(&dev_res->cq_tbl, cq_handle); -} - -RdmaRmQP *rdma_rm_get_qp(RdmaDeviceResources *dev_res, uint32_t qpn) -{ - GBytes *key =3D g_bytes_new(&qpn, sizeof(qpn)); - - RdmaRmQP *qp =3D g_hash_table_lookup(dev_res->qp_hash, key); - - g_bytes_unref(key); - - if (!qp) { - rdma_error_report("Invalid QP handle %d", qpn); - } - - return qp; -} - -int rdma_rm_alloc_qp(RdmaDeviceResources *dev_res, uint32_t pd_handle, - uint8_t qp_type, uint32_t max_send_wr, - uint32_t max_send_sge, uint32_t send_cq_handle, - uint32_t max_recv_wr, uint32_t max_recv_sge, - uint32_t recv_cq_handle, void *opaque, uint32_t *qpn, - uint8_t is_srq, uint32_t srq_handle) -{ - int rc; - RdmaRmQP *qp; - RdmaRmCQ *scq, *rcq; - RdmaRmPD *pd; - RdmaRmSRQ *srq =3D NULL; - uint32_t rm_qpn; - - pd =3D rdma_rm_get_pd(dev_res, pd_handle); - if (!pd) { - return -EINVAL; - } - - scq =3D rdma_rm_get_cq(dev_res, send_cq_handle); - rcq =3D rdma_rm_get_cq(dev_res, recv_cq_handle); - - if (!scq || !rcq) { - rdma_error_report("Invalid send_cqn or recv_cqn (%d, %d)", - send_cq_handle, recv_cq_handle); - return -EINVAL; - } - - if (is_srq) { - srq =3D rdma_rm_get_srq(dev_res, srq_handle); - if (!srq) { - rdma_error_report("Invalid srqn %d", srq_handle); - return -EINVAL; - } - - srq->recv_cq_handle =3D recv_cq_handle; - } - - if (qp_type =3D=3D IBV_QPT_GSI) { - scq->notify =3D CNT_SET; - rcq->notify =3D CNT_SET; - } - - qp =3D rdma_res_tbl_alloc(&dev_res->qp_tbl, &rm_qpn); - if (!qp) { - return -ENOMEM; - } - - qp->qpn =3D rm_qpn; - qp->qp_state =3D IBV_QPS_RESET; - qp->qp_type =3D qp_type; - qp->send_cq_handle =3D send_cq_handle; - qp->recv_cq_handle =3D recv_cq_handle; - qp->opaque =3D opaque; - qp->is_srq =3D is_srq; - - rc =3D rdma_backend_create_qp(&qp->backend_qp, qp_type, &pd->backend_p= d, - &scq->backend_cq, &rcq->backend_cq, - is_srq ? &srq->backend_srq : NULL, - max_send_wr, max_recv_wr, max_send_sge, - max_recv_sge); - - if (rc) { - rc =3D -EIO; - goto out_dealloc_qp; - } - - *qpn =3D rdma_backend_qpn(&qp->backend_qp); - trace_rdma_rm_alloc_qp(rm_qpn, *qpn, qp_type); - g_hash_table_insert(dev_res->qp_hash, g_bytes_new(qpn, sizeof(*qpn)), = qp); - - return 0; - -out_dealloc_qp: - rdma_res_tbl_dealloc(&dev_res->qp_tbl, qp->qpn); - - return rc; -} - -int rdma_rm_modify_qp(RdmaDeviceResources *dev_res, RdmaBackendDev *backen= d_dev, - uint32_t qp_handle, uint32_t attr_mask, uint8_t sgid= _idx, - union ibv_gid *dgid, uint32_t dqpn, - enum ibv_qp_state qp_state, uint32_t qkey, - uint32_t rq_psn, uint32_t sq_psn) -{ - RdmaRmQP *qp; - int ret; - - qp =3D rdma_rm_get_qp(dev_res, qp_handle); - if (!qp) { - return -EINVAL; - } - - if (qp->qp_type =3D=3D IBV_QPT_SMI) { - rdma_error_report("Got QP0 request"); - return -EPERM; - } else if (qp->qp_type =3D=3D IBV_QPT_GSI) { - return 0; - } - - trace_rdma_rm_modify_qp(qp_handle, attr_mask, qp_state, sgid_idx); - - if (attr_mask & IBV_QP_STATE) { - qp->qp_state =3D qp_state; - - if (qp->qp_state =3D=3D IBV_QPS_INIT) { - ret =3D rdma_backend_qp_state_init(backend_dev, &qp->backend_q= p, - qp->qp_type, qkey); - if (ret) { - return -EIO; - } - } - - if (qp->qp_state =3D=3D IBV_QPS_RTR) { - /* Get backend gid index */ - sgid_idx =3D rdma_rm_get_backend_gid_index(dev_res, backend_de= v, - sgid_idx); - if (sgid_idx <=3D 0) { /* TODO check also less than bk.max_sgi= d */ - rdma_error_report("Failed to get bk sgid_idx for sgid_idx = %d", - sgid_idx); - return -EIO; - } - - ret =3D rdma_backend_qp_state_rtr(backend_dev, &qp->backend_qp, - qp->qp_type, sgid_idx, dgid, d= qpn, - rq_psn, qkey, - attr_mask & IBV_QP_QKEY); - if (ret) { - return -EIO; - } - } - - if (qp->qp_state =3D=3D IBV_QPS_RTS) { - ret =3D rdma_backend_qp_state_rts(&qp->backend_qp, qp->qp_type, - sq_psn, qkey, - attr_mask & IBV_QP_QKEY); - if (ret) { - return -EIO; - } - } - } - - return 0; -} - -int rdma_rm_query_qp(RdmaDeviceResources *dev_res, RdmaBackendDev *backend= _dev, - uint32_t qp_handle, struct ibv_qp_attr *attr, - int attr_mask, struct ibv_qp_init_attr *init_attr) -{ - RdmaRmQP *qp; - - qp =3D rdma_rm_get_qp(dev_res, qp_handle); - if (!qp) { - return -EINVAL; - } - - return rdma_backend_query_qp(&qp->backend_qp, attr, attr_mask, init_at= tr); -} - -void rdma_rm_dealloc_qp(RdmaDeviceResources *dev_res, uint32_t qp_handle) -{ - RdmaRmQP *qp; - GBytes *key; - - key =3D g_bytes_new(&qp_handle, sizeof(qp_handle)); - qp =3D g_hash_table_lookup(dev_res->qp_hash, key); - g_hash_table_remove(dev_res->qp_hash, key); - g_bytes_unref(key); - - if (!qp) { - return; - } - - rdma_backend_destroy_qp(&qp->backend_qp, dev_res); - - rdma_res_tbl_dealloc(&dev_res->qp_tbl, qp->qpn); -} - -RdmaRmSRQ *rdma_rm_get_srq(RdmaDeviceResources *dev_res, uint32_t srq_hand= le) -{ - return rdma_res_tbl_get(&dev_res->srq_tbl, srq_handle); -} - -int rdma_rm_alloc_srq(RdmaDeviceResources *dev_res, uint32_t pd_handle, - uint32_t max_wr, uint32_t max_sge, uint32_t srq_limi= t, - uint32_t *srq_handle, void *opaque) -{ - RdmaRmSRQ *srq; - RdmaRmPD *pd; - int rc; - - pd =3D rdma_rm_get_pd(dev_res, pd_handle); - if (!pd) { - return -EINVAL; - } - - srq =3D rdma_res_tbl_alloc(&dev_res->srq_tbl, srq_handle); - if (!srq) { - return -ENOMEM; - } - - rc =3D rdma_backend_create_srq(&srq->backend_srq, &pd->backend_pd, - max_wr, max_sge, srq_limit); - if (rc) { - rc =3D -EIO; - goto out_dealloc_srq; - } - - srq->opaque =3D opaque; - - return 0; - -out_dealloc_srq: - rdma_res_tbl_dealloc(&dev_res->srq_tbl, *srq_handle); - - return rc; -} - -int rdma_rm_query_srq(RdmaDeviceResources *dev_res, uint32_t srq_handle, - struct ibv_srq_attr *srq_attr) -{ - RdmaRmSRQ *srq; - - srq =3D rdma_rm_get_srq(dev_res, srq_handle); - if (!srq) { - return -EINVAL; - } - - return rdma_backend_query_srq(&srq->backend_srq, srq_attr); -} - -int rdma_rm_modify_srq(RdmaDeviceResources *dev_res, uint32_t srq_handle, - struct ibv_srq_attr *srq_attr, int srq_attr_mask) -{ - RdmaRmSRQ *srq; - - srq =3D rdma_rm_get_srq(dev_res, srq_handle); - if (!srq) { - return -EINVAL; - } - - if ((srq_attr_mask & IBV_SRQ_LIMIT) && - (srq_attr->srq_limit =3D=3D 0)) { - return -EINVAL; - } - - if ((srq_attr_mask & IBV_SRQ_MAX_WR) && - (srq_attr->max_wr =3D=3D 0)) { - return -EINVAL; - } - - return rdma_backend_modify_srq(&srq->backend_srq, srq_attr, - srq_attr_mask); -} - -void rdma_rm_dealloc_srq(RdmaDeviceResources *dev_res, uint32_t srq_handle) -{ - RdmaRmSRQ *srq; - - srq =3D rdma_rm_get_srq(dev_res, srq_handle); - if (!srq) { - return; - } - - rdma_backend_destroy_srq(&srq->backend_srq, dev_res); - rdma_res_tbl_dealloc(&dev_res->srq_tbl, srq_handle); -} - -void *rdma_rm_get_cqe_ctx(RdmaDeviceResources *dev_res, uint32_t cqe_ctx_i= d) -{ - void **cqe_ctx; - - cqe_ctx =3D rdma_res_tbl_get(&dev_res->cqe_ctx_tbl, cqe_ctx_id); - if (!cqe_ctx) { - return NULL; - } - - return *cqe_ctx; -} - -int rdma_rm_alloc_cqe_ctx(RdmaDeviceResources *dev_res, uint32_t *cqe_ctx_= id, - void *ctx) -{ - void **cqe_ctx; - - cqe_ctx =3D rdma_res_tbl_alloc(&dev_res->cqe_ctx_tbl, cqe_ctx_id); - if (!cqe_ctx) { - return -ENOMEM; - } - - *cqe_ctx =3D ctx; - - return 0; -} - -void rdma_rm_dealloc_cqe_ctx(RdmaDeviceResources *dev_res, uint32_t cqe_ct= x_id) -{ - rdma_res_tbl_dealloc(&dev_res->cqe_ctx_tbl, cqe_ctx_id); -} - -int rdma_rm_add_gid(RdmaDeviceResources *dev_res, RdmaBackendDev *backend_= dev, - const char *ifname, union ibv_gid *gid, int gid_idx) -{ - int rc; - - rc =3D rdma_backend_add_gid(backend_dev, ifname, gid); - if (rc) { - return -EINVAL; - } - - memcpy(&dev_res->port.gid_tbl[gid_idx].gid, gid, sizeof(*gid)); - - return 0; -} - -int rdma_rm_del_gid(RdmaDeviceResources *dev_res, RdmaBackendDev *backend_= dev, - const char *ifname, int gid_idx) -{ - int rc; - - if (!dev_res->port.gid_tbl[gid_idx].gid.global.interface_id) { - return 0; - } - - rc =3D rdma_backend_del_gid(backend_dev, ifname, - &dev_res->port.gid_tbl[gid_idx].gid); - if (rc) { - return -EINVAL; - } - - memset(dev_res->port.gid_tbl[gid_idx].gid.raw, 0, - sizeof(dev_res->port.gid_tbl[gid_idx].gid)); - dev_res->port.gid_tbl[gid_idx].backend_gid_index =3D -1; - - return 0; -} - -int rdma_rm_get_backend_gid_index(RdmaDeviceResources *dev_res, - RdmaBackendDev *backend_dev, int sgid_id= x) -{ - if (unlikely(sgid_idx < 0 || sgid_idx >=3D MAX_PORT_GIDS)) { - rdma_error_report("Got invalid sgid_idx %d", sgid_idx); - return -EINVAL; - } - - if (unlikely(dev_res->port.gid_tbl[sgid_idx].backend_gid_index =3D=3D = -1)) { - dev_res->port.gid_tbl[sgid_idx].backend_gid_index =3D - rdma_backend_get_gid_index(backend_dev, - &dev_res->port.gid_tbl[sgid_idx].gid); - } - - return dev_res->port.gid_tbl[sgid_idx].backend_gid_index; -} - -static void destroy_qp_hash_key(gpointer data) -{ - g_bytes_unref(data); -} - -static void init_ports(RdmaDeviceResources *dev_res) -{ - int i; - - memset(&dev_res->port, 0, sizeof(dev_res->port)); - - dev_res->port.state =3D IBV_PORT_DOWN; - for (i =3D 0; i < MAX_PORT_GIDS; i++) { - dev_res->port.gid_tbl[i].backend_gid_index =3D -1; - } -} - -static void fini_ports(RdmaDeviceResources *dev_res, - RdmaBackendDev *backend_dev, const char *ifname) -{ - int i; - - dev_res->port.state =3D IBV_PORT_DOWN; - for (i =3D 0; i < MAX_PORT_GIDS; i++) { - rdma_rm_del_gid(dev_res, backend_dev, ifname, i); - } -} - -int rdma_rm_init(RdmaDeviceResources *dev_res, struct ibv_device_attr *dev= _attr) -{ - dev_res->qp_hash =3D g_hash_table_new_full(g_bytes_hash, g_bytes_equal, - destroy_qp_hash_key, NULL); - if (!dev_res->qp_hash) { - return -ENOMEM; - } - - res_tbl_init("PD", &dev_res->pd_tbl, dev_attr->max_pd, sizeof(RdmaRmPD= )); - res_tbl_init("CQ", &dev_res->cq_tbl, dev_attr->max_cq, sizeof(RdmaRmCQ= )); - res_tbl_init("MR", &dev_res->mr_tbl, dev_attr->max_mr, sizeof(RdmaRmMR= )); - res_tbl_init("QP", &dev_res->qp_tbl, dev_attr->max_qp, sizeof(RdmaRmQP= )); - res_tbl_init("CQE_CTX", &dev_res->cqe_ctx_tbl, dev_attr->max_qp * - dev_attr->max_qp_wr, sizeof(void *)); - res_tbl_init("UC", &dev_res->uc_tbl, MAX_UCS, sizeof(RdmaRmUC)); - res_tbl_init("SRQ", &dev_res->srq_tbl, dev_attr->max_srq, - sizeof(RdmaRmSRQ)); - - init_ports(dev_res); - - qemu_mutex_init(&dev_res->lock); - - memset(&dev_res->stats, 0, sizeof(dev_res->stats)); - qatomic_set(&dev_res->stats.missing_cqe, 0); - - return 0; -} - -void rdma_rm_fini(RdmaDeviceResources *dev_res, RdmaBackendDev *backend_de= v, - const char *ifname) -{ - qemu_mutex_destroy(&dev_res->lock); - - fini_ports(dev_res, backend_dev, ifname); - - res_tbl_free(&dev_res->srq_tbl); - res_tbl_free(&dev_res->uc_tbl); - res_tbl_free(&dev_res->cqe_ctx_tbl); - res_tbl_free(&dev_res->qp_tbl); - res_tbl_free(&dev_res->mr_tbl); - res_tbl_free(&dev_res->cq_tbl); - res_tbl_free(&dev_res->pd_tbl); - - if (dev_res->qp_hash) { - g_hash_table_destroy(dev_res->qp_hash); - } -} diff --git a/hw/rdma/rdma_utils.c b/hw/rdma/rdma_utils.c deleted file mode 100644 index c948baf052..0000000000 --- a/hw/rdma/rdma_utils.c +++ /dev/null @@ -1,126 +0,0 @@ -/* - * QEMU paravirtual RDMA - Generic RDMA backend - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include "hw/pci/pci_device.h" -#include "trace.h" -#include "rdma_utils.h" - -void *rdma_pci_dma_map(PCIDevice *dev, dma_addr_t addr, dma_addr_t len) -{ - void *p; - dma_addr_t pci_len =3D len; - - if (!addr) { - rdma_error_report("addr is NULL"); - return NULL; - } - - p =3D pci_dma_map(dev, addr, &pci_len, DMA_DIRECTION_TO_DEVICE); - if (!p) { - rdma_error_report("pci_dma_map fail, addr=3D0x%"PRIx64", len=3D%"P= RId64, - addr, pci_len); - return NULL; - } - - if (pci_len !=3D len) { - rdma_pci_dma_unmap(dev, p, pci_len); - return NULL; - } - - trace_rdma_pci_dma_map(addr, p, pci_len); - - return p; -} - -void rdma_pci_dma_unmap(PCIDevice *dev, void *buffer, dma_addr_t len) -{ - trace_rdma_pci_dma_unmap(buffer); - if (buffer) { - pci_dma_unmap(dev, buffer, len, DMA_DIRECTION_TO_DEVICE, 0); - } -} - -void rdma_protected_gqueue_init(RdmaProtectedGQueue *list) -{ - qemu_mutex_init(&list->lock); - list->list =3D g_queue_new(); -} - -void rdma_protected_gqueue_destroy(RdmaProtectedGQueue *list) -{ - if (list->list) { - g_queue_free_full(list->list, g_free); - qemu_mutex_destroy(&list->lock); - list->list =3D NULL; - } -} - -void rdma_protected_gqueue_append_int64(RdmaProtectedGQueue *list, - int64_t value) -{ - qemu_mutex_lock(&list->lock); - g_queue_push_tail(list->list, g_memdup(&value, sizeof(value))); - qemu_mutex_unlock(&list->lock); -} - -int64_t rdma_protected_gqueue_pop_int64(RdmaProtectedGQueue *list) -{ - int64_t *valp; - int64_t val; - - qemu_mutex_lock(&list->lock); - - valp =3D g_queue_pop_head(list->list); - qemu_mutex_unlock(&list->lock); - - if (!valp) { - return -ENOENT; - } - - val =3D *valp; - g_free(valp); - return val; -} - -void rdma_protected_gslist_init(RdmaProtectedGSList *list) -{ - qemu_mutex_init(&list->lock); -} - -void rdma_protected_gslist_destroy(RdmaProtectedGSList *list) -{ - if (list->list) { - g_slist_free(list->list); - qemu_mutex_destroy(&list->lock); - list->list =3D NULL; - } -} - -void rdma_protected_gslist_append_int32(RdmaProtectedGSList *list, - int32_t value) -{ - qemu_mutex_lock(&list->lock); - list->list =3D g_slist_prepend(list->list, GINT_TO_POINTER(value)); - qemu_mutex_unlock(&list->lock); -} - -void rdma_protected_gslist_remove_int32(RdmaProtectedGSList *list, - int32_t value) -{ - qemu_mutex_lock(&list->lock); - list->list =3D g_slist_remove(list->list, GINT_TO_POINTER(value)); - qemu_mutex_unlock(&list->lock); -} diff --git a/hw/rdma/vmw/pvrdma_cmd.c b/hw/rdma/vmw/pvrdma_cmd.c deleted file mode 100644 index d385d18d9c..0000000000 --- a/hw/rdma/vmw/pvrdma_cmd.c +++ /dev/null @@ -1,815 +0,0 @@ -/* - * QEMU paravirtual RDMA - Command channel - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include "cpu.h" -#include "hw/pci/pci.h" -#include "hw/pci/pci_ids.h" - -#include "../rdma_backend.h" -#include "../rdma_rm.h" -#include "../rdma_utils.h" - -#include "trace.h" -#include "pvrdma.h" -#include "standard-headers/rdma/vmw_pvrdma-abi.h" - -static void *pvrdma_map_to_pdir(PCIDevice *pdev, uint64_t pdir_dma, - uint32_t nchunks, size_t length) -{ - uint64_t *dir, *tbl; - int tbl_idx, dir_idx, addr_idx; - void *host_virt =3D NULL, *curr_page; - - if (!nchunks) { - rdma_error_report("Got nchunks=3D0"); - return NULL; - } - - length =3D ROUND_UP(length, TARGET_PAGE_SIZE); - if (nchunks * TARGET_PAGE_SIZE !=3D length) { - rdma_error_report("Invalid nchunks/length (%u, %lu)", nchunks, - (unsigned long)length); - return NULL; - } - - dir =3D rdma_pci_dma_map(pdev, pdir_dma, TARGET_PAGE_SIZE); - if (!dir) { - rdma_error_report("Failed to map to page directory"); - return NULL; - } - - tbl =3D rdma_pci_dma_map(pdev, dir[0], TARGET_PAGE_SIZE); - if (!tbl) { - rdma_error_report("Failed to map to page table 0"); - goto out_unmap_dir; - } - - curr_page =3D rdma_pci_dma_map(pdev, (dma_addr_t)tbl[0], TARGET_PAGE_S= IZE); - if (!curr_page) { - rdma_error_report("Failed to map the page 0"); - goto out_unmap_tbl; - } - - host_virt =3D mremap(curr_page, 0, length, MREMAP_MAYMOVE); - if (host_virt =3D=3D MAP_FAILED) { - host_virt =3D NULL; - rdma_error_report("Failed to remap memory for host_virt"); - goto out_unmap_tbl; - } - trace_pvrdma_map_to_pdir_host_virt(curr_page, host_virt); - - rdma_pci_dma_unmap(pdev, curr_page, TARGET_PAGE_SIZE); - - dir_idx =3D 0; - tbl_idx =3D 1; - addr_idx =3D 1; - while (addr_idx < nchunks) { - if (tbl_idx =3D=3D TARGET_PAGE_SIZE / sizeof(uint64_t)) { - tbl_idx =3D 0; - dir_idx++; - rdma_pci_dma_unmap(pdev, tbl, TARGET_PAGE_SIZE); - tbl =3D rdma_pci_dma_map(pdev, dir[dir_idx], TARGET_PAGE_SIZE); - if (!tbl) { - rdma_error_report("Failed to map to page table %d", dir_id= x); - goto out_unmap_host_virt; - } - } - - curr_page =3D rdma_pci_dma_map(pdev, (dma_addr_t)tbl[tbl_idx], - TARGET_PAGE_SIZE); - if (!curr_page) { - rdma_error_report("Failed to map to page %d, dir %d", tbl_idx, - dir_idx); - goto out_unmap_host_virt; - } - - mremap(curr_page, 0, TARGET_PAGE_SIZE, MREMAP_MAYMOVE | MREMAP_FIX= ED, - host_virt + TARGET_PAGE_SIZE * addr_idx); - - trace_pvrdma_map_to_pdir_next_page(addr_idx, curr_page, host_virt + - TARGET_PAGE_SIZE * addr_idx); - - rdma_pci_dma_unmap(pdev, curr_page, TARGET_PAGE_SIZE); - - addr_idx++; - - tbl_idx++; - } - - goto out_unmap_tbl; - -out_unmap_host_virt: - munmap(host_virt, length); - host_virt =3D NULL; - -out_unmap_tbl: - rdma_pci_dma_unmap(pdev, tbl, TARGET_PAGE_SIZE); - -out_unmap_dir: - rdma_pci_dma_unmap(pdev, dir, TARGET_PAGE_SIZE); - - return host_virt; -} - -static int query_port(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_query_port *cmd =3D &req->query_port; - struct pvrdma_cmd_query_port_resp *resp =3D &rsp->query_port_resp; - struct ibv_port_attr attrs =3D {}; - - if (cmd->port_num > MAX_PORTS) { - return -EINVAL; - } - - if (rdma_backend_query_port(&dev->backend_dev, &attrs)) { - return -ENOMEM; - } - - memset(resp, 0, sizeof(*resp)); - - /* - * The state, max_mtu and active_mtu fields are enums; the values - * for pvrdma_port_state and pvrdma_mtu match those for - * ibv_port_state and ibv_mtu, so we can cast them safely. - */ - resp->attrs.state =3D dev->func0->device_active ? - (enum pvrdma_port_state)attrs.state : PVRDMA_PORT_DOWN; - resp->attrs.max_mtu =3D (enum pvrdma_mtu)attrs.max_mtu; - resp->attrs.active_mtu =3D (enum pvrdma_mtu)attrs.active_mtu; - resp->attrs.phys_state =3D attrs.phys_state; - resp->attrs.gid_tbl_len =3D MIN(MAX_PORT_GIDS, attrs.gid_tbl_len); - resp->attrs.max_msg_sz =3D 1024; - resp->attrs.pkey_tbl_len =3D MIN(MAX_PORT_PKEYS, attrs.pkey_tbl_len); - resp->attrs.active_width =3D 1; - resp->attrs.active_speed =3D 1; - - return 0; -} - -static int query_pkey(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_query_pkey *cmd =3D &req->query_pkey; - struct pvrdma_cmd_query_pkey_resp *resp =3D &rsp->query_pkey_resp; - - if (cmd->port_num > MAX_PORTS) { - return -EINVAL; - } - - if (cmd->index > MAX_PKEYS) { - return -EINVAL; - } - - memset(resp, 0, sizeof(*resp)); - - resp->pkey =3D PVRDMA_PKEY; - - return 0; -} - -static int create_pd(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_create_pd *cmd =3D &req->create_pd; - struct pvrdma_cmd_create_pd_resp *resp =3D &rsp->create_pd_resp; - - memset(resp, 0, sizeof(*resp)); - return rdma_rm_alloc_pd(&dev->rdma_dev_res, &dev->backend_dev, - &resp->pd_handle, cmd->ctx_handle); -} - -static int destroy_pd(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_destroy_pd *cmd =3D &req->destroy_pd; - - rdma_rm_dealloc_pd(&dev->rdma_dev_res, cmd->pd_handle); - - return 0; -} - -static int create_mr(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_create_mr *cmd =3D &req->create_mr; - struct pvrdma_cmd_create_mr_resp *resp =3D &rsp->create_mr_resp; - PCIDevice *pci_dev =3D PCI_DEVICE(dev); - void *host_virt =3D NULL; - int rc =3D 0; - - memset(resp, 0, sizeof(*resp)); - - if (!(cmd->flags & PVRDMA_MR_FLAG_DMA)) { - host_virt =3D pvrdma_map_to_pdir(pci_dev, cmd->pdir_dma, cmd->nchu= nks, - cmd->length); - if (!host_virt) { - rdma_error_report("Failed to map to pdir"); - return -EINVAL; - } - } - - rc =3D rdma_rm_alloc_mr(&dev->rdma_dev_res, cmd->pd_handle, cmd->start, - cmd->length, host_virt, cmd->access_flags, - &resp->mr_handle, &resp->lkey, &resp->rkey); - if (rc && host_virt) { - munmap(host_virt, cmd->length); - } - - return rc; -} - -static int destroy_mr(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_destroy_mr *cmd =3D &req->destroy_mr; - - rdma_rm_dealloc_mr(&dev->rdma_dev_res, cmd->mr_handle); - - return 0; -} - -static int create_cq_ring(PCIDevice *pci_dev , PvrdmaRing **ring, - uint64_t pdir_dma, uint32_t nchunks, uint32_t cq= e) -{ - uint64_t *dir =3D NULL, *tbl =3D NULL; - PvrdmaRing *r; - int rc =3D -EINVAL; - char ring_name[MAX_RING_NAME_SZ]; - - if (!nchunks || nchunks > PVRDMA_MAX_FAST_REG_PAGES) { - rdma_error_report("Got invalid nchunks: %d", nchunks); - return rc; - } - - dir =3D rdma_pci_dma_map(pci_dev, pdir_dma, TARGET_PAGE_SIZE); - if (!dir) { - rdma_error_report("Failed to map to CQ page directory"); - goto out; - } - - tbl =3D rdma_pci_dma_map(pci_dev, dir[0], TARGET_PAGE_SIZE); - if (!tbl) { - rdma_error_report("Failed to map to CQ page table"); - goto out; - } - - r =3D g_malloc(sizeof(*r)); - *ring =3D r; - - r->ring_state =3D rdma_pci_dma_map(pci_dev, tbl[0], TARGET_PAGE_SIZE); - - if (!r->ring_state) { - rdma_error_report("Failed to map to CQ ring state"); - goto out_free_ring; - } - - sprintf(ring_name, "cq_ring_%" PRIx64, pdir_dma); - rc =3D pvrdma_ring_init(r, ring_name, pci_dev, &r->ring_state[1], - cqe, sizeof(struct pvrdma_cqe), - /* first page is ring state */ - (dma_addr_t *)&tbl[1], nchunks - 1); - if (rc) { - goto out_unmap_ring_state; - } - - goto out; - -out_unmap_ring_state: - /* ring_state was in slot 1, not 0 so need to jump back */ - rdma_pci_dma_unmap(pci_dev, --r->ring_state, TARGET_PAGE_SIZE); - -out_free_ring: - g_free(r); - -out: - rdma_pci_dma_unmap(pci_dev, tbl, TARGET_PAGE_SIZE); - rdma_pci_dma_unmap(pci_dev, dir, TARGET_PAGE_SIZE); - - return rc; -} - -static void destroy_cq_ring(PvrdmaRing *ring) -{ - pvrdma_ring_free(ring); - /* ring_state was in slot 1, not 0 so need to jump back */ - rdma_pci_dma_unmap(ring->dev, --ring->ring_state, TARGET_PAGE_SIZE); - g_free(ring); -} - -static int create_cq(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_create_cq *cmd =3D &req->create_cq; - struct pvrdma_cmd_create_cq_resp *resp =3D &rsp->create_cq_resp; - PvrdmaRing *ring =3D NULL; - int rc; - - memset(resp, 0, sizeof(*resp)); - - resp->cqe =3D cmd->cqe; - - rc =3D create_cq_ring(PCI_DEVICE(dev), &ring, cmd->pdir_dma, cmd->nchu= nks, - cmd->cqe); - if (rc) { - return rc; - } - - rc =3D rdma_rm_alloc_cq(&dev->rdma_dev_res, &dev->backend_dev, cmd->cq= e, - &resp->cq_handle, ring); - if (rc) { - destroy_cq_ring(ring); - } - - resp->cqe =3D cmd->cqe; - - return rc; -} - -static int destroy_cq(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_destroy_cq *cmd =3D &req->destroy_cq; - RdmaRmCQ *cq; - PvrdmaRing *ring; - - cq =3D rdma_rm_get_cq(&dev->rdma_dev_res, cmd->cq_handle); - if (!cq) { - rdma_error_report("Got invalid CQ handle"); - return -EINVAL; - } - - ring =3D (PvrdmaRing *)cq->opaque; - destroy_cq_ring(ring); - - rdma_rm_dealloc_cq(&dev->rdma_dev_res, cmd->cq_handle); - - return 0; -} - -static int create_qp_rings(PCIDevice *pci_dev, uint64_t pdir_dma, - PvrdmaRing **rings, uint32_t scqe, uint32_t sma= x_sge, - uint32_t spages, uint32_t rcqe, uint32_t rmax_s= ge, - uint32_t rpages, uint8_t is_srq) -{ - uint64_t *dir =3D NULL, *tbl =3D NULL; - PvrdmaRing *sr, *rr; - int rc =3D -EINVAL; - char ring_name[MAX_RING_NAME_SZ]; - uint32_t wqe_sz; - - if (!spages || spages > PVRDMA_MAX_FAST_REG_PAGES) { - rdma_error_report("Got invalid send page count for QP ring: %d", - spages); - return rc; - } - - if (!is_srq && (!rpages || rpages > PVRDMA_MAX_FAST_REG_PAGES)) { - rdma_error_report("Got invalid recv page count for QP ring: %d", - rpages); - return rc; - } - - dir =3D rdma_pci_dma_map(pci_dev, pdir_dma, TARGET_PAGE_SIZE); - if (!dir) { - rdma_error_report("Failed to map to QP page directory"); - goto out; - } - - tbl =3D rdma_pci_dma_map(pci_dev, dir[0], TARGET_PAGE_SIZE); - if (!tbl) { - rdma_error_report("Failed to map to QP page table"); - goto out; - } - - if (!is_srq) { - sr =3D g_malloc(2 * sizeof(*rr)); - rr =3D &sr[1]; - } else { - sr =3D g_malloc(sizeof(*sr)); - } - - *rings =3D sr; - - /* Create send ring */ - sr->ring_state =3D rdma_pci_dma_map(pci_dev, tbl[0], TARGET_PAGE_SIZE); - if (!sr->ring_state) { - rdma_error_report("Failed to map to QP ring state"); - goto out_free_sr_mem; - } - - wqe_sz =3D pow2ceil(sizeof(struct pvrdma_sq_wqe_hdr) + - sizeof(struct pvrdma_sge) * smax_sge - 1); - - sprintf(ring_name, "qp_sring_%" PRIx64, pdir_dma); - rc =3D pvrdma_ring_init(sr, ring_name, pci_dev, sr->ring_state, - scqe, wqe_sz, (dma_addr_t *)&tbl[1], spages); - if (rc) { - goto out_unmap_ring_state; - } - - if (!is_srq) { - /* Create recv ring */ - rr->ring_state =3D &sr->ring_state[1]; - wqe_sz =3D pow2ceil(sizeof(struct pvrdma_rq_wqe_hdr) + - sizeof(struct pvrdma_sge) * rmax_sge - 1); - sprintf(ring_name, "qp_rring_%" PRIx64, pdir_dma); - rc =3D pvrdma_ring_init(rr, ring_name, pci_dev, rr->ring_state, - rcqe, wqe_sz, (dma_addr_t *)&tbl[1 + spages], - rpages); - if (rc) { - goto out_free_sr; - } - } - - goto out; - -out_free_sr: - pvrdma_ring_free(sr); - -out_unmap_ring_state: - rdma_pci_dma_unmap(pci_dev, sr->ring_state, TARGET_PAGE_SIZE); - -out_free_sr_mem: - g_free(sr); - -out: - rdma_pci_dma_unmap(pci_dev, tbl, TARGET_PAGE_SIZE); - rdma_pci_dma_unmap(pci_dev, dir, TARGET_PAGE_SIZE); - - return rc; -} - -static void destroy_qp_rings(PvrdmaRing *ring, uint8_t is_srq) -{ - pvrdma_ring_free(&ring[0]); - if (!is_srq) { - pvrdma_ring_free(&ring[1]); - } - - rdma_pci_dma_unmap(ring->dev, ring->ring_state, TARGET_PAGE_SIZE); - g_free(ring); -} - -static int create_qp(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_create_qp *cmd =3D &req->create_qp; - struct pvrdma_cmd_create_qp_resp *resp =3D &rsp->create_qp_resp; - PvrdmaRing *rings =3D NULL; - int rc; - - memset(resp, 0, sizeof(*resp)); - - rc =3D create_qp_rings(PCI_DEVICE(dev), cmd->pdir_dma, &rings, - cmd->max_send_wr, cmd->max_send_sge, cmd->send_ch= unks, - cmd->max_recv_wr, cmd->max_recv_sge, - cmd->total_chunks - cmd->send_chunks - 1, cmd->is= _srq); - if (rc) { - return rc; - } - - rc =3D rdma_rm_alloc_qp(&dev->rdma_dev_res, cmd->pd_handle, cmd->qp_ty= pe, - cmd->max_send_wr, cmd->max_send_sge, - cmd->send_cq_handle, cmd->max_recv_wr, - cmd->max_recv_sge, cmd->recv_cq_handle, rings, - &resp->qpn, cmd->is_srq, cmd->srq_handle); - if (rc) { - destroy_qp_rings(rings, cmd->is_srq); - return rc; - } - - resp->max_send_wr =3D cmd->max_send_wr; - resp->max_recv_wr =3D cmd->max_recv_wr; - resp->max_send_sge =3D cmd->max_send_sge; - resp->max_recv_sge =3D cmd->max_recv_sge; - resp->max_inline_data =3D cmd->max_inline_data; - - return 0; -} - -static int modify_qp(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_modify_qp *cmd =3D &req->modify_qp; - - /* No need to verify sgid_index since it is u8 */ - - return rdma_rm_modify_qp(&dev->rdma_dev_res, &dev->backend_dev, - cmd->qp_handle, cmd->attr_mask, - cmd->attrs.ah_attr.grh.sgid_index, - (union ibv_gid *)&cmd->attrs.ah_attr.grh.dgid, - cmd->attrs.dest_qp_num, - (enum ibv_qp_state)cmd->attrs.qp_state, - cmd->attrs.qkey, cmd->attrs.rq_psn, - cmd->attrs.sq_psn); -} - -static int query_qp(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_query_qp *cmd =3D &req->query_qp; - struct pvrdma_cmd_query_qp_resp *resp =3D &rsp->query_qp_resp; - struct ibv_qp_init_attr init_attr; - - memset(resp, 0, sizeof(*resp)); - - return rdma_rm_query_qp(&dev->rdma_dev_res, &dev->backend_dev, - cmd->qp_handle, - (struct ibv_qp_attr *)&resp->attrs, - cmd->attr_mask, - &init_attr); -} - -static int destroy_qp(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_destroy_qp *cmd =3D &req->destroy_qp; - RdmaRmQP *qp; - PvrdmaRing *ring; - - qp =3D rdma_rm_get_qp(&dev->rdma_dev_res, cmd->qp_handle); - if (!qp) { - return -EINVAL; - } - - ring =3D (PvrdmaRing *)qp->opaque; - destroy_qp_rings(ring, qp->is_srq); - rdma_rm_dealloc_qp(&dev->rdma_dev_res, cmd->qp_handle); - - return 0; -} - -static int create_bind(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_create_bind *cmd =3D &req->create_bind; - union ibv_gid *gid =3D (union ibv_gid *)&cmd->new_gid; - - if (cmd->index >=3D MAX_PORT_GIDS) { - return -EINVAL; - } - - return rdma_rm_add_gid(&dev->rdma_dev_res, &dev->backend_dev, - dev->backend_eth_device_name, gid, cmd->index); -} - -static int destroy_bind(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_destroy_bind *cmd =3D &req->destroy_bind; - - if (cmd->index >=3D MAX_PORT_GIDS) { - return -EINVAL; - } - - return rdma_rm_del_gid(&dev->rdma_dev_res, &dev->backend_dev, - dev->backend_eth_device_name, cmd->index); -} - -static int create_uc(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_create_uc *cmd =3D &req->create_uc; - struct pvrdma_cmd_create_uc_resp *resp =3D &rsp->create_uc_resp; - - memset(resp, 0, sizeof(*resp)); - return rdma_rm_alloc_uc(&dev->rdma_dev_res, cmd->pfn, &resp->ctx_handl= e); -} - -static int destroy_uc(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_destroy_uc *cmd =3D &req->destroy_uc; - - rdma_rm_dealloc_uc(&dev->rdma_dev_res, cmd->ctx_handle); - - return 0; -} - -static int create_srq_ring(PCIDevice *pci_dev, PvrdmaRing **ring, - uint64_t pdir_dma, uint32_t max_wr, - uint32_t max_sge, uint32_t nchunks) -{ - uint64_t *dir =3D NULL, *tbl =3D NULL; - PvrdmaRing *r; - int rc =3D -EINVAL; - char ring_name[MAX_RING_NAME_SZ]; - uint32_t wqe_sz; - - if (!nchunks || nchunks > PVRDMA_MAX_FAST_REG_PAGES) { - rdma_error_report("Got invalid page count for SRQ ring: %d", - nchunks); - return rc; - } - - dir =3D rdma_pci_dma_map(pci_dev, pdir_dma, TARGET_PAGE_SIZE); - if (!dir) { - rdma_error_report("Failed to map to SRQ page directory"); - goto out; - } - - tbl =3D rdma_pci_dma_map(pci_dev, dir[0], TARGET_PAGE_SIZE); - if (!tbl) { - rdma_error_report("Failed to map to SRQ page table"); - goto out; - } - - r =3D g_malloc(sizeof(*r)); - *ring =3D r; - - r->ring_state =3D rdma_pci_dma_map(pci_dev, tbl[0], TARGET_PAGE_SIZE); - if (!r->ring_state) { - rdma_error_report("Failed to map tp SRQ ring state"); - goto out_free_ring_mem; - } - - wqe_sz =3D pow2ceil(sizeof(struct pvrdma_rq_wqe_hdr) + - sizeof(struct pvrdma_sge) * max_sge - 1); - sprintf(ring_name, "srq_ring_%" PRIx64, pdir_dma); - rc =3D pvrdma_ring_init(r, ring_name, pci_dev, &r->ring_state[1], max_= wr, - wqe_sz, (dma_addr_t *)&tbl[1], nchunks - 1); - if (rc) { - goto out_unmap_ring_state; - } - - goto out; - -out_unmap_ring_state: - rdma_pci_dma_unmap(pci_dev, r->ring_state, TARGET_PAGE_SIZE); - -out_free_ring_mem: - g_free(r); - -out: - rdma_pci_dma_unmap(pci_dev, tbl, TARGET_PAGE_SIZE); - rdma_pci_dma_unmap(pci_dev, dir, TARGET_PAGE_SIZE); - - return rc; -} - -static void destroy_srq_ring(PvrdmaRing *ring) -{ - pvrdma_ring_free(ring); - rdma_pci_dma_unmap(ring->dev, ring->ring_state, TARGET_PAGE_SIZE); - g_free(ring); -} - -static int create_srq(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_create_srq *cmd =3D &req->create_srq; - struct pvrdma_cmd_create_srq_resp *resp =3D &rsp->create_srq_resp; - PvrdmaRing *ring =3D NULL; - int rc; - - memset(resp, 0, sizeof(*resp)); - - rc =3D create_srq_ring(PCI_DEVICE(dev), &ring, cmd->pdir_dma, - cmd->attrs.max_wr, cmd->attrs.max_sge, - cmd->nchunks); - if (rc) { - return rc; - } - - rc =3D rdma_rm_alloc_srq(&dev->rdma_dev_res, cmd->pd_handle, - cmd->attrs.max_wr, cmd->attrs.max_sge, - cmd->attrs.srq_limit, &resp->srqn, ring); - if (rc) { - destroy_srq_ring(ring); - return rc; - } - - return 0; -} - -static int query_srq(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_query_srq *cmd =3D &req->query_srq; - struct pvrdma_cmd_query_srq_resp *resp =3D &rsp->query_srq_resp; - - memset(resp, 0, sizeof(*resp)); - - return rdma_rm_query_srq(&dev->rdma_dev_res, cmd->srq_handle, - (struct ibv_srq_attr *)&resp->attrs); -} - -static int modify_srq(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_modify_srq *cmd =3D &req->modify_srq; - - /* Only support SRQ limit */ - if (!(cmd->attr_mask & IBV_SRQ_LIMIT) || - (cmd->attr_mask & IBV_SRQ_MAX_WR)) - return -EINVAL; - - return rdma_rm_modify_srq(&dev->rdma_dev_res, cmd->srq_handle, - (struct ibv_srq_attr *)&cmd->attrs, - cmd->attr_mask); -} - -static int destroy_srq(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp) -{ - struct pvrdma_cmd_destroy_srq *cmd =3D &req->destroy_srq; - RdmaRmSRQ *srq; - PvrdmaRing *ring; - - srq =3D rdma_rm_get_srq(&dev->rdma_dev_res, cmd->srq_handle); - if (!srq) { - return -EINVAL; - } - - ring =3D (PvrdmaRing *)srq->opaque; - destroy_srq_ring(ring); - rdma_rm_dealloc_srq(&dev->rdma_dev_res, cmd->srq_handle); - - return 0; -} - -struct cmd_handler { - uint32_t cmd; - uint32_t ack; - int (*exec)(PVRDMADev *dev, union pvrdma_cmd_req *req, - union pvrdma_cmd_resp *rsp); -}; - -static struct cmd_handler cmd_handlers[] =3D { - {PVRDMA_CMD_QUERY_PORT, PVRDMA_CMD_QUERY_PORT_RESP, query_por= t}, - {PVRDMA_CMD_QUERY_PKEY, PVRDMA_CMD_QUERY_PKEY_RESP, query_pke= y}, - {PVRDMA_CMD_CREATE_PD, PVRDMA_CMD_CREATE_PD_RESP, create_pd= }, - {PVRDMA_CMD_DESTROY_PD, PVRDMA_CMD_DESTROY_PD_RESP_NOOP, destroy_p= d}, - {PVRDMA_CMD_CREATE_MR, PVRDMA_CMD_CREATE_MR_RESP, create_mr= }, - {PVRDMA_CMD_DESTROY_MR, PVRDMA_CMD_DESTROY_MR_RESP_NOOP, destroy_m= r}, - {PVRDMA_CMD_CREATE_CQ, PVRDMA_CMD_CREATE_CQ_RESP, create_cq= }, - {PVRDMA_CMD_RESIZE_CQ, PVRDMA_CMD_RESIZE_CQ_RESP, NULL}, - {PVRDMA_CMD_DESTROY_CQ, PVRDMA_CMD_DESTROY_CQ_RESP_NOOP, destroy_c= q}, - {PVRDMA_CMD_CREATE_QP, PVRDMA_CMD_CREATE_QP_RESP, create_qp= }, - {PVRDMA_CMD_MODIFY_QP, PVRDMA_CMD_MODIFY_QP_RESP, modify_qp= }, - {PVRDMA_CMD_QUERY_QP, PVRDMA_CMD_QUERY_QP_RESP, query_qp}, - {PVRDMA_CMD_DESTROY_QP, PVRDMA_CMD_DESTROY_QP_RESP, destroy_q= p}, - {PVRDMA_CMD_CREATE_UC, PVRDMA_CMD_CREATE_UC_RESP, create_uc= }, - {PVRDMA_CMD_DESTROY_UC, PVRDMA_CMD_DESTROY_UC_RESP_NOOP, destroy_u= c}, - {PVRDMA_CMD_CREATE_BIND, PVRDMA_CMD_CREATE_BIND_RESP_NOOP, create_bi= nd}, - {PVRDMA_CMD_DESTROY_BIND, PVRDMA_CMD_DESTROY_BIND_RESP_NOOP, destroy_b= ind}, - {PVRDMA_CMD_CREATE_SRQ, PVRDMA_CMD_CREATE_SRQ_RESP, create_sr= q}, - {PVRDMA_CMD_QUERY_SRQ, PVRDMA_CMD_QUERY_SRQ_RESP, query_srq= }, - {PVRDMA_CMD_MODIFY_SRQ, PVRDMA_CMD_MODIFY_SRQ_RESP, modify_sr= q}, - {PVRDMA_CMD_DESTROY_SRQ, PVRDMA_CMD_DESTROY_SRQ_RESP, destroy_s= rq}, -}; - -int pvrdma_exec_cmd(PVRDMADev *dev) -{ - int err =3D 0xFFFF; - DSRInfo *dsr_info; - - dsr_info =3D &dev->dsr_info; - - if (!dsr_info->dsr) { - /* Buggy or malicious guest driver */ - rdma_error_report("Exec command without dsr, req or rsp buffer= s"); - goto out; - } - - if (dsr_info->req->hdr.cmd >=3D sizeof(cmd_handlers) / - sizeof(struct cmd_handler)) { - rdma_error_report("Unsupported command"); - goto out; - } - - if (!cmd_handlers[dsr_info->req->hdr.cmd].exec) { - rdma_error_report("Unsupported command (not implemented yet)"); - goto out; - } - - err =3D cmd_handlers[dsr_info->req->hdr.cmd].exec(dev, dsr_info->req, - dsr_info->rsp); - dsr_info->rsp->hdr.response =3D dsr_info->req->hdr.response; - dsr_info->rsp->hdr.ack =3D cmd_handlers[dsr_info->req->hdr.cmd].ack; - dsr_info->rsp->hdr.err =3D err < 0 ? -err : 0; - - trace_pvrdma_exec_cmd(dsr_info->req->hdr.cmd, dsr_info->rsp->hdr.err); - - dev->stats.commands++; - -out: - set_reg_val(dev, PVRDMA_REG_ERR, err); - post_interrupt(dev, INTR_VEC_CMD_RING); - - return (err =3D=3D 0) ? 0 : -EINVAL; -} diff --git a/hw/rdma/vmw/pvrdma_dev_ring.c b/hw/rdma/vmw/pvrdma_dev_ring.c deleted file mode 100644 index 30ce22a5be..0000000000 --- a/hw/rdma/vmw/pvrdma_dev_ring.c +++ /dev/null @@ -1,141 +0,0 @@ -/* - * QEMU paravirtual RDMA - Device rings - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include "hw/pci/pci.h" -#include "cpu.h" -#include "qemu/cutils.h" - -#include "trace.h" - -#include "../rdma_utils.h" -#include "pvrdma_dev_ring.h" - -int pvrdma_ring_init(PvrdmaRing *ring, const char *name, PCIDevice *dev, - PvrdmaRingState *ring_state, uint32_t max_elems, - size_t elem_sz, dma_addr_t *tbl, uint32_t npages) -{ - int i; - int rc =3D 0; - - pstrcpy(ring->name, MAX_RING_NAME_SZ, name); - ring->dev =3D dev; - ring->ring_state =3D ring_state; - ring->max_elems =3D max_elems; - ring->elem_sz =3D elem_sz; - /* TODO: Give a moment to think if we want to redo driver settings - qatomic_set(&ring->ring_state->prod_tail, 0); - qatomic_set(&ring->ring_state->cons_head, 0); - */ - ring->npages =3D npages; - ring->pages =3D g_new0(void *, npages); - - for (i =3D 0; i < npages; i++) { - if (!tbl[i]) { - rdma_error_report("npages=3D%d but tbl[%d] is NULL", npages, i= ); - continue; - } - - ring->pages[i] =3D rdma_pci_dma_map(dev, tbl[i], TARGET_PAGE_SIZE); - if (!ring->pages[i]) { - rc =3D -ENOMEM; - rdma_error_report("Failed to map to page %d in ring %s", i, na= me); - goto out_free; - } - memset(ring->pages[i], 0, TARGET_PAGE_SIZE); - } - - goto out; - -out_free: - while (i--) { - rdma_pci_dma_unmap(dev, ring->pages[i], TARGET_PAGE_SIZE); - } - g_free(ring->pages); - -out: - return rc; -} - -void *pvrdma_ring_next_elem_read(PvrdmaRing *ring) -{ - unsigned int idx, offset; - const uint32_t tail =3D qatomic_read(&ring->ring_state->prod_tail); - const uint32_t head =3D qatomic_read(&ring->ring_state->cons_head); - - if (tail & ~((ring->max_elems << 1) - 1) || - head & ~((ring->max_elems << 1) - 1) || - tail =3D=3D head) { - trace_pvrdma_ring_next_elem_read_no_data(ring->name); - return NULL; - } - - idx =3D head & (ring->max_elems - 1); - offset =3D idx * ring->elem_sz; - return ring->pages[offset / TARGET_PAGE_SIZE] + (offset % TARGET_PAGE_= SIZE); -} - -void pvrdma_ring_read_inc(PvrdmaRing *ring) -{ - uint32_t idx =3D qatomic_read(&ring->ring_state->cons_head); - - idx =3D (idx + 1) & ((ring->max_elems << 1) - 1); - qatomic_set(&ring->ring_state->cons_head, idx); -} - -void *pvrdma_ring_next_elem_write(PvrdmaRing *ring) -{ - unsigned int idx, offset; - const uint32_t tail =3D qatomic_read(&ring->ring_state->prod_tail); - const uint32_t head =3D qatomic_read(&ring->ring_state->cons_head); - - if (tail & ~((ring->max_elems << 1) - 1) || - head & ~((ring->max_elems << 1) - 1) || - tail =3D=3D (head ^ ring->max_elems)) { - rdma_error_report("CQ is full"); - return NULL; - } - - idx =3D tail & (ring->max_elems - 1); - offset =3D idx * ring->elem_sz; - return ring->pages[offset / TARGET_PAGE_SIZE] + (offset % TARGET_PAGE_= SIZE); -} - -void pvrdma_ring_write_inc(PvrdmaRing *ring) -{ - uint32_t idx =3D qatomic_read(&ring->ring_state->prod_tail); - - idx =3D (idx + 1) & ((ring->max_elems << 1) - 1); - qatomic_set(&ring->ring_state->prod_tail, idx); -} - -void pvrdma_ring_free(PvrdmaRing *ring) -{ - if (!ring) { - return; - } - - if (!ring->pages) { - return; - } - - while (ring->npages--) { - rdma_pci_dma_unmap(ring->dev, ring->pages[ring->npages], - TARGET_PAGE_SIZE); - } - - g_free(ring->pages); - ring->pages =3D NULL; -} diff --git a/hw/rdma/vmw/pvrdma_main.c b/hw/rdma/vmw/pvrdma_main.c deleted file mode 100644 index e735ff97eb..0000000000 --- a/hw/rdma/vmw/pvrdma_main.c +++ /dev/null @@ -1,735 +0,0 @@ -/* - * QEMU paravirtual RDMA - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include "qapi/error.h" -#include "qemu/module.h" -#include "hw/pci/pci.h" -#include "hw/pci/pci_ids.h" -#include "hw/pci/msi.h" -#include "hw/pci/msix.h" -#include "hw/qdev-properties.h" -#include "hw/qdev-properties-system.h" -#include "cpu.h" -#include "trace.h" -#include "monitor/monitor.h" -#include "hw/rdma/rdma.h" - -#include "../rdma_rm.h" -#include "../rdma_backend.h" -#include "../rdma_utils.h" - -#include -#include "pvrdma.h" -#include "standard-headers/rdma/vmw_pvrdma-abi.h" -#include "sysemu/runstate.h" -#include "standard-headers/drivers/infiniband/hw/vmw_pvrdma/pvrdma_dev_api= .h" -#include "pvrdma_qp_ops.h" - -static Property pvrdma_dev_properties[] =3D { - DEFINE_PROP_STRING("netdev", PVRDMADev, backend_eth_device_name), - DEFINE_PROP_STRING("ibdev", PVRDMADev, backend_device_name), - DEFINE_PROP_UINT8("ibport", PVRDMADev, backend_port_num, 1), - DEFINE_PROP_UINT64("dev-caps-max-mr-size", PVRDMADev, dev_attr.max_mr_= size, - MAX_MR_SIZE), - DEFINE_PROP_INT32("dev-caps-max-qp", PVRDMADev, dev_attr.max_qp, MAX_Q= P), - DEFINE_PROP_INT32("dev-caps-max-cq", PVRDMADev, dev_attr.max_cq, MAX_C= Q), - DEFINE_PROP_INT32("dev-caps-max-mr", PVRDMADev, dev_attr.max_mr, MAX_M= R), - DEFINE_PROP_INT32("dev-caps-max-pd", PVRDMADev, dev_attr.max_pd, MAX_P= D), - DEFINE_PROP_INT32("dev-caps-qp-rd-atom", PVRDMADev, dev_attr.max_qp_rd= _atom, - MAX_QP_RD_ATOM), - DEFINE_PROP_INT32("dev-caps-max-qp-init-rd-atom", PVRDMADev, - dev_attr.max_qp_init_rd_atom, MAX_QP_INIT_RD_ATOM), - DEFINE_PROP_INT32("dev-caps-max-ah", PVRDMADev, dev_attr.max_ah, MAX_A= H), - DEFINE_PROP_INT32("dev-caps-max-srq", PVRDMADev, dev_attr.max_srq, MAX= _SRQ), - DEFINE_PROP_CHR("mad-chardev", PVRDMADev, mad_chr), - DEFINE_PROP_END_OF_LIST(), -}; - -static void pvrdma_format_statistics(RdmaProvider *obj, GString *buf) -{ - PVRDMADev *dev =3D PVRDMA_DEV(obj); - PCIDevice *pdev =3D PCI_DEVICE(dev); - - g_string_append_printf(buf, "%s, %x.%x\n", - pdev->name, PCI_SLOT(pdev->devfn), - PCI_FUNC(pdev->devfn)); - g_string_append_printf(buf, "\tcommands : %" PRId64 "\n", - dev->stats.commands); - g_string_append_printf(buf, "\tregs_reads : %" PRId64 "\n", - dev->stats.regs_reads); - g_string_append_printf(buf, "\tregs_writes : %" PRId64 "\n", - dev->stats.regs_writes); - g_string_append_printf(buf, "\tuar_writes : %" PRId64 "\n", - dev->stats.uar_writes); - g_string_append_printf(buf, "\tinterrupts : %" PRId64 "\n", - dev->stats.interrupts); - rdma_format_device_counters(&dev->rdma_dev_res, buf); -} - -static void free_dev_ring(PCIDevice *pci_dev, PvrdmaRing *ring, - void *ring_state) -{ - pvrdma_ring_free(ring); - rdma_pci_dma_unmap(pci_dev, ring_state, TARGET_PAGE_SIZE); -} - -static int init_dev_ring(PvrdmaRing *ring, PvrdmaRingState **ring_state, - const char *name, PCIDevice *pci_dev, - dma_addr_t dir_addr, uint32_t num_pages) -{ - uint64_t *dir, *tbl; - int max_pages, rc =3D 0; - - if (!num_pages) { - rdma_error_report("Ring pages count must be strictly positive"); - return -EINVAL; - } - - /* - * Make sure we can satisfy the requested number of pages in a single - * TARGET_PAGE_SIZE sized page table (taking into account that first e= ntry - * is reserved for ring-state) - */ - max_pages =3D TARGET_PAGE_SIZE / sizeof(dma_addr_t) - 1; - if (num_pages > max_pages) { - rdma_error_report("Maximum pages on a single directory must not ex= ceed %d\n", - max_pages); - return -EINVAL; - } - - dir =3D rdma_pci_dma_map(pci_dev, dir_addr, TARGET_PAGE_SIZE); - if (!dir) { - rdma_error_report("Failed to map to page directory (ring %s)", nam= e); - rc =3D -ENOMEM; - goto out; - } - - /* We support only one page table for a ring */ - tbl =3D rdma_pci_dma_map(pci_dev, dir[0], TARGET_PAGE_SIZE); - if (!tbl) { - rdma_error_report("Failed to map to page table (ring %s)", name); - rc =3D -ENOMEM; - goto out_free_dir; - } - - *ring_state =3D rdma_pci_dma_map(pci_dev, tbl[0], TARGET_PAGE_SIZE); - if (!*ring_state) { - rdma_error_report("Failed to map to ring state (ring %s)", name); - rc =3D -ENOMEM; - goto out_free_tbl; - } - /* RX ring is the second */ - (*ring_state)++; - rc =3D pvrdma_ring_init(ring, name, pci_dev, - (PvrdmaRingState *)*ring_state, - (num_pages - 1) * TARGET_PAGE_SIZE / - sizeof(struct pvrdma_cqne), - sizeof(struct pvrdma_cqne), - (dma_addr_t *)&tbl[1], (dma_addr_t)num_pages - 1= ); - if (rc) { - rc =3D -ENOMEM; - goto out_free_ring_state; - } - - goto out_free_tbl; - -out_free_ring_state: - rdma_pci_dma_unmap(pci_dev, *ring_state, TARGET_PAGE_SIZE); - -out_free_tbl: - rdma_pci_dma_unmap(pci_dev, tbl, TARGET_PAGE_SIZE); - -out_free_dir: - rdma_pci_dma_unmap(pci_dev, dir, TARGET_PAGE_SIZE); - -out: - return rc; -} - -static void free_dsr(PVRDMADev *dev) -{ - PCIDevice *pci_dev =3D PCI_DEVICE(dev); - - if (!dev->dsr_info.dsr) { - return; - } - - free_dev_ring(pci_dev, &dev->dsr_info.async, - dev->dsr_info.async_ring_state); - - free_dev_ring(pci_dev, &dev->dsr_info.cq, dev->dsr_info.cq_ring_state); - - rdma_pci_dma_unmap(pci_dev, dev->dsr_info.req, - sizeof(union pvrdma_cmd_req)); - - rdma_pci_dma_unmap(pci_dev, dev->dsr_info.rsp, - sizeof(union pvrdma_cmd_resp)); - - rdma_pci_dma_unmap(pci_dev, dev->dsr_info.dsr, - sizeof(struct pvrdma_device_shared_region)); - - dev->dsr_info.dsr =3D NULL; -} - -static int load_dsr(PVRDMADev *dev) -{ - int rc =3D 0; - PCIDevice *pci_dev =3D PCI_DEVICE(dev); - DSRInfo *dsr_info; - struct pvrdma_device_shared_region *dsr; - - free_dsr(dev); - - /* Map to DSR */ - dev->dsr_info.dsr =3D rdma_pci_dma_map(pci_dev, dev->dsr_info.dma, - sizeof(struct pvrdma_device_shared_region)); - if (!dev->dsr_info.dsr) { - rdma_error_report("Failed to map to DSR"); - rc =3D -ENOMEM; - goto out; - } - - /* Shortcuts */ - dsr_info =3D &dev->dsr_info; - dsr =3D dsr_info->dsr; - - /* Map to command slot */ - dsr_info->req =3D rdma_pci_dma_map(pci_dev, dsr->cmd_slot_dma, - sizeof(union pvrdma_cmd_req)); - if (!dsr_info->req) { - rdma_error_report("Failed to map to command slot address"); - rc =3D -ENOMEM; - goto out_free_dsr; - } - - /* Map to response slot */ - dsr_info->rsp =3D rdma_pci_dma_map(pci_dev, dsr->resp_slot_dma, - sizeof(union pvrdma_cmd_resp)); - if (!dsr_info->rsp) { - rdma_error_report("Failed to map to response slot address"); - rc =3D -ENOMEM; - goto out_free_req; - } - - /* Map to CQ notification ring */ - rc =3D init_dev_ring(&dsr_info->cq, &dsr_info->cq_ring_state, "dev_cq", - pci_dev, dsr->cq_ring_pages.pdir_dma, - dsr->cq_ring_pages.num_pages); - if (rc) { - rc =3D -ENOMEM; - goto out_free_rsp; - } - - /* Map to event notification ring */ - rc =3D init_dev_ring(&dsr_info->async, &dsr_info->async_ring_state, - "dev_async", pci_dev, dsr->async_ring_pages.pdir_dm= a, - dsr->async_ring_pages.num_pages); - if (rc) { - rc =3D -ENOMEM; - goto out_free_rsp; - } - - goto out; - -out_free_rsp: - rdma_pci_dma_unmap(pci_dev, dsr_info->rsp, sizeof(union pvrdma_cmd_res= p)); - -out_free_req: - rdma_pci_dma_unmap(pci_dev, dsr_info->req, sizeof(union pvrdma_cmd_req= )); - -out_free_dsr: - rdma_pci_dma_unmap(pci_dev, dsr_info->dsr, - sizeof(struct pvrdma_device_shared_region)); - dsr_info->dsr =3D NULL; - -out: - return rc; -} - -static void init_dsr_dev_caps(PVRDMADev *dev) -{ - struct pvrdma_device_shared_region *dsr; - - if (!dev->dsr_info.dsr) { - /* Buggy or malicious guest driver */ - rdma_error_report("Can't initialized DSR"); - return; - } - - dsr =3D dev->dsr_info.dsr; - dsr->caps.fw_ver =3D PVRDMA_FW_VERSION; - dsr->caps.mode =3D PVRDMA_DEVICE_MODE_ROCE; - dsr->caps.gid_types |=3D PVRDMA_GID_TYPE_FLAG_ROCE_V1; - dsr->caps.max_uar =3D RDMA_BAR2_UAR_SIZE; - dsr->caps.max_mr_size =3D dev->dev_attr.max_mr_size; - dsr->caps.max_qp =3D dev->dev_attr.max_qp; - dsr->caps.max_qp_wr =3D dev->dev_attr.max_qp_wr; - dsr->caps.max_sge =3D dev->dev_attr.max_sge; - dsr->caps.max_cq =3D dev->dev_attr.max_cq; - dsr->caps.max_cqe =3D dev->dev_attr.max_cqe; - dsr->caps.max_mr =3D dev->dev_attr.max_mr; - dsr->caps.max_pd =3D dev->dev_attr.max_pd; - dsr->caps.max_ah =3D dev->dev_attr.max_ah; - dsr->caps.max_srq =3D dev->dev_attr.max_srq; - dsr->caps.max_srq_wr =3D dev->dev_attr.max_srq_wr; - dsr->caps.max_srq_sge =3D dev->dev_attr.max_srq_sge; - dsr->caps.gid_tbl_len =3D MAX_GIDS; - dsr->caps.sys_image_guid =3D 0; - dsr->caps.node_guid =3D dev->node_guid; - dsr->caps.phys_port_cnt =3D MAX_PORTS; - dsr->caps.max_pkeys =3D MAX_PKEYS; -} - -static void uninit_msix(PCIDevice *pdev, int used_vectors) -{ - PVRDMADev *dev =3D PVRDMA_DEV(pdev); - int i; - - for (i =3D 0; i < used_vectors; i++) { - msix_vector_unuse(pdev, i); - } - - msix_uninit(pdev, &dev->msix, &dev->msix); -} - -static int init_msix(PCIDevice *pdev) -{ - PVRDMADev *dev =3D PVRDMA_DEV(pdev); - int i; - int rc; - - rc =3D msix_init(pdev, RDMA_MAX_INTRS, &dev->msix, RDMA_MSIX_BAR_IDX, - RDMA_MSIX_TABLE, &dev->msix, RDMA_MSIX_BAR_IDX, - RDMA_MSIX_PBA, 0, NULL); - - if (rc < 0) { - rdma_error_report("Failed to initialize MSI-X"); - return rc; - } - - for (i =3D 0; i < RDMA_MAX_INTRS; i++) { - msix_vector_use(PCI_DEVICE(dev), i); - } - - return 0; -} - -static void pvrdma_fini(PCIDevice *pdev) -{ - PVRDMADev *dev =3D PVRDMA_DEV(pdev); - - notifier_remove(&dev->shutdown_notifier); - - pvrdma_qp_ops_fini(); - - rdma_backend_stop(&dev->backend_dev); - - rdma_rm_fini(&dev->rdma_dev_res, &dev->backend_dev, - dev->backend_eth_device_name); - - rdma_backend_fini(&dev->backend_dev); - - free_dsr(dev); - - if (msix_enabled(pdev)) { - uninit_msix(pdev, RDMA_MAX_INTRS); - } - - rdma_info_report("Device %s %x.%x is down", pdev->name, - PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn)); -} - -static void pvrdma_stop(PVRDMADev *dev) -{ - rdma_backend_stop(&dev->backend_dev); -} - -static void pvrdma_start(PVRDMADev *dev) -{ - rdma_backend_start(&dev->backend_dev); -} - -static void activate_device(PVRDMADev *dev) -{ - pvrdma_start(dev); - set_reg_val(dev, PVRDMA_REG_ERR, 0); -} - -static int unquiesce_device(PVRDMADev *dev) -{ - return 0; -} - -static void reset_device(PVRDMADev *dev) -{ - pvrdma_stop(dev); -} - -static uint64_t pvrdma_regs_read(void *opaque, hwaddr addr, unsigned size) -{ - PVRDMADev *dev =3D opaque; - uint32_t val; - - dev->stats.regs_reads++; - - if (get_reg_val(dev, addr, &val)) { - rdma_error_report("Failed to read REG value from address 0x%x", - (uint32_t)addr); - return -EINVAL; - } - - trace_pvrdma_regs_read(addr, val); - - return val; -} - -static void pvrdma_regs_write(void *opaque, hwaddr addr, uint64_t val, - unsigned size) -{ - PVRDMADev *dev =3D opaque; - - dev->stats.regs_writes++; - - if (set_reg_val(dev, addr, val)) { - rdma_error_report("Failed to set REG value, addr=3D0x%"PRIx64 ", v= al=3D0x%"PRIx64, - addr, val); - return; - } - - switch (addr) { - case PVRDMA_REG_DSRLOW: - trace_pvrdma_regs_write(addr, val, "DSRLOW", ""); - dev->dsr_info.dma =3D val; - break; - case PVRDMA_REG_DSRHIGH: - trace_pvrdma_regs_write(addr, val, "DSRHIGH", ""); - dev->dsr_info.dma |=3D val << 32; - load_dsr(dev); - init_dsr_dev_caps(dev); - break; - case PVRDMA_REG_CTL: - switch (val) { - case PVRDMA_DEVICE_CTL_ACTIVATE: - trace_pvrdma_regs_write(addr, val, "CTL", "ACTIVATE"); - activate_device(dev); - break; - case PVRDMA_DEVICE_CTL_UNQUIESCE: - trace_pvrdma_regs_write(addr, val, "CTL", "UNQUIESCE"); - unquiesce_device(dev); - break; - case PVRDMA_DEVICE_CTL_RESET: - trace_pvrdma_regs_write(addr, val, "CTL", "URESET"); - reset_device(dev); - break; - } - break; - case PVRDMA_REG_IMR: - trace_pvrdma_regs_write(addr, val, "INTR_MASK", ""); - dev->interrupt_mask =3D val; - break; - case PVRDMA_REG_REQUEST: - if (val =3D=3D 0) { - trace_pvrdma_regs_write(addr, val, "REQUEST", ""); - pvrdma_exec_cmd(dev); - } - break; - default: - break; - } -} - -static const MemoryRegionOps regs_ops =3D { - .read =3D pvrdma_regs_read, - .write =3D pvrdma_regs_write, - .endianness =3D DEVICE_LITTLE_ENDIAN, - .impl =3D { - .min_access_size =3D sizeof(uint32_t), - .max_access_size =3D sizeof(uint32_t), - }, -}; - -static uint64_t pvrdma_uar_read(void *opaque, hwaddr addr, unsigned size) -{ - return 0xffffffff; -} - -static void pvrdma_uar_write(void *opaque, hwaddr addr, uint64_t val, - unsigned size) -{ - PVRDMADev *dev =3D opaque; - - dev->stats.uar_writes++; - - switch (addr & 0xFFF) { /* Mask with 0xFFF as each UC gets page */ - case PVRDMA_UAR_QP_OFFSET: - if (val & PVRDMA_UAR_QP_SEND) { - trace_pvrdma_uar_write(addr, val, "QP", "SEND", - val & PVRDMA_UAR_HANDLE_MASK, 0); - pvrdma_qp_send(dev, val & PVRDMA_UAR_HANDLE_MASK); - } - if (val & PVRDMA_UAR_QP_RECV) { - trace_pvrdma_uar_write(addr, val, "QP", "RECV", - val & PVRDMA_UAR_HANDLE_MASK, 0); - pvrdma_qp_recv(dev, val & PVRDMA_UAR_HANDLE_MASK); - } - break; - case PVRDMA_UAR_CQ_OFFSET: - if (val & PVRDMA_UAR_CQ_ARM) { - trace_pvrdma_uar_write(addr, val, "CQ", "ARM", - val & PVRDMA_UAR_HANDLE_MASK, - !!(val & PVRDMA_UAR_CQ_ARM_SOL)); - rdma_rm_req_notify_cq(&dev->rdma_dev_res, - val & PVRDMA_UAR_HANDLE_MASK, - !!(val & PVRDMA_UAR_CQ_ARM_SOL)); - } - if (val & PVRDMA_UAR_CQ_ARM_SOL) { - trace_pvrdma_uar_write(addr, val, "CQ", "ARMSOL - not supporte= d", 0, - 0); - } - if (val & PVRDMA_UAR_CQ_POLL) { - trace_pvrdma_uar_write(addr, val, "CQ", "POLL", - val & PVRDMA_UAR_HANDLE_MASK, 0); - pvrdma_cq_poll(&dev->rdma_dev_res, val & PVRDMA_UAR_HANDLE_MAS= K); - } - break; - case PVRDMA_UAR_SRQ_OFFSET: - if (val & PVRDMA_UAR_SRQ_RECV) { - trace_pvrdma_uar_write(addr, val, "QP", "SRQ", - val & PVRDMA_UAR_HANDLE_MASK, 0); - pvrdma_srq_recv(dev, val & PVRDMA_UAR_HANDLE_MASK); - } - break; - default: - rdma_error_report("Unsupported command, addr=3D0x%"PRIx64", val=3D= 0x%"PRIx64, - addr, val); - break; - } -} - -static const MemoryRegionOps uar_ops =3D { - .read =3D pvrdma_uar_read, - .write =3D pvrdma_uar_write, - .endianness =3D DEVICE_LITTLE_ENDIAN, - .impl =3D { - .min_access_size =3D sizeof(uint32_t), - .max_access_size =3D sizeof(uint32_t), - }, -}; - -static void init_pci_config(PCIDevice *pdev) -{ - pdev->config[PCI_INTERRUPT_PIN] =3D 1; -} - -static void init_bars(PCIDevice *pdev) -{ - PVRDMADev *dev =3D PVRDMA_DEV(pdev); - - /* BAR 0 - MSI-X */ - memory_region_init(&dev->msix, OBJECT(dev), "pvrdma-msix", - RDMA_BAR0_MSIX_SIZE); - pci_register_bar(pdev, RDMA_MSIX_BAR_IDX, PCI_BASE_ADDRESS_SPACE_MEMOR= Y, - &dev->msix); - - /* BAR 1 - Registers */ - memset(&dev->regs_data, 0, sizeof(dev->regs_data)); - memory_region_init_io(&dev->regs, OBJECT(dev), ®s_ops, dev, - "pvrdma-regs", sizeof(dev->regs_data)); - pci_register_bar(pdev, RDMA_REG_BAR_IDX, PCI_BASE_ADDRESS_SPACE_MEMORY, - &dev->regs); - - /* BAR 2 - UAR */ - memset(&dev->uar_data, 0, sizeof(dev->uar_data)); - memory_region_init_io(&dev->uar, OBJECT(dev), &uar_ops, dev, "rdma-uar= ", - sizeof(dev->uar_data)); - pci_register_bar(pdev, RDMA_UAR_BAR_IDX, PCI_BASE_ADDRESS_SPACE_MEMORY, - &dev->uar); -} - -static void init_regs(PCIDevice *pdev) -{ - PVRDMADev *dev =3D PVRDMA_DEV(pdev); - - set_reg_val(dev, PVRDMA_REG_VERSION, PVRDMA_HW_VERSION); - set_reg_val(dev, PVRDMA_REG_ERR, 0xFFFF); -} - -static void init_dev_caps(PVRDMADev *dev) -{ - size_t pg_tbl_bytes =3D TARGET_PAGE_SIZE * - (TARGET_PAGE_SIZE / sizeof(uint64_t)); - size_t wr_sz =3D MAX(sizeof(struct pvrdma_sq_wqe_hdr), - sizeof(struct pvrdma_rq_wqe_hdr)); - - dev->dev_attr.max_qp_wr =3D pg_tbl_bytes / - (wr_sz + sizeof(struct pvrdma_sge) * - dev->dev_attr.max_sge) - TARGET_PAGE_SIZE; - /* First page is ring state ^^^^ */ - - dev->dev_attr.max_cqe =3D pg_tbl_bytes / sizeof(struct pvrdma_cqe) - - TARGET_PAGE_SIZE; /* First page is ring state = */ - - dev->dev_attr.max_srq_wr =3D pg_tbl_bytes / - ((sizeof(struct pvrdma_rq_wqe_hdr) + - sizeof(struct pvrdma_sge)) * - dev->dev_attr.max_sge) - TARGET_PAGE_SIZE; -} - -static int pvrdma_check_ram_shared(Object *obj, void *opaque) -{ - bool *shared =3D opaque; - - if (object_dynamic_cast(obj, "memory-backend-ram")) { - *shared =3D object_property_get_bool(obj, "share", NULL); - } - - return 0; -} - -static void pvrdma_shutdown_notifier(Notifier *n, void *opaque) -{ - PVRDMADev *dev =3D container_of(n, PVRDMADev, shutdown_notifier); - PCIDevice *pci_dev =3D PCI_DEVICE(dev); - - pvrdma_fini(pci_dev); -} - -static void pvrdma_realize(PCIDevice *pdev, Error **errp) -{ - int rc =3D 0; - PVRDMADev *dev =3D PVRDMA_DEV(pdev); - Object *memdev_root; - bool ram_shared =3D false; - PCIDevice *func0; - - warn_report_once("pvrdma is deprecated and will be removed in a future= release"); - - rdma_info_report("Initializing device %s %x.%x", pdev->name, - PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn)); - - if (TARGET_PAGE_SIZE !=3D qemu_real_host_page_size()) { - error_setg(errp, "Target page size must be the same as host page s= ize"); - return; - } - - func0 =3D pci_get_function_0(pdev); - /* Break if not vmxnet3 device in slot 0 */ - if (strcmp(object_get_typename(OBJECT(func0)), TYPE_VMXNET3)) { - error_setg(errp, "Device on %x.0 must be %s", PCI_SLOT(pdev->devfn= ), - TYPE_VMXNET3); - return; - } - dev->func0 =3D VMXNET3(func0); - - addrconf_addr_eui48((unsigned char *)&dev->node_guid, - (const char *)&dev->func0->conf.macaddr.a); - - memdev_root =3D object_resolve_path("/objects", NULL); - if (memdev_root) { - object_child_foreach(memdev_root, pvrdma_check_ram_shared, &ram_sh= ared); - } - if (!ram_shared) { - error_setg(errp, "Only shared memory backed ram is supported"); - return; - } - - dev->dsr_info.dsr =3D NULL; - - init_pci_config(pdev); - - init_bars(pdev); - - init_regs(pdev); - - rc =3D init_msix(pdev); - if (rc) { - goto out; - } - - rc =3D rdma_backend_init(&dev->backend_dev, pdev, &dev->rdma_dev_res, - dev->backend_device_name, dev->backend_port_num, - &dev->dev_attr, &dev->mad_chr); - if (rc) { - goto out; - } - - init_dev_caps(dev); - - rc =3D rdma_rm_init(&dev->rdma_dev_res, &dev->dev_attr); - if (rc) { - goto out; - } - - rc =3D pvrdma_qp_ops_init(); - if (rc) { - goto out; - } - - memset(&dev->stats, 0, sizeof(dev->stats)); - - dev->shutdown_notifier.notify =3D pvrdma_shutdown_notifier; - qemu_register_shutdown_notifier(&dev->shutdown_notifier); - -#ifdef LEGACY_RDMA_REG_MR - rdma_info_report("Using legacy reg_mr"); -#else - rdma_info_report("Using iova reg_mr"); -#endif - -out: - if (rc) { - pvrdma_fini(pdev); - error_append_hint(errp, "Device failed to load\n"); - } -} - -static void pvrdma_class_init(ObjectClass *klass, void *data) -{ - DeviceClass *dc =3D DEVICE_CLASS(klass); - PCIDeviceClass *k =3D PCI_DEVICE_CLASS(klass); - RdmaProviderClass *ir =3D RDMA_PROVIDER_CLASS(klass); - - k->realize =3D pvrdma_realize; - k->vendor_id =3D PCI_VENDOR_ID_VMWARE; - k->device_id =3D PCI_DEVICE_ID_VMWARE_PVRDMA; - k->revision =3D 0x00; - k->class_id =3D PCI_CLASS_NETWORK_OTHER; - - dc->desc =3D "RDMA Device"; - device_class_set_props(dc, pvrdma_dev_properties); - set_bit(DEVICE_CATEGORY_NETWORK, dc->categories); - - ir->format_statistics =3D pvrdma_format_statistics; -} - -static const TypeInfo pvrdma_info =3D { - .name =3D PVRDMA_HW_NAME, - .parent =3D TYPE_PCI_DEVICE, - .instance_size =3D sizeof(PVRDMADev), - .class_init =3D pvrdma_class_init, - .interfaces =3D (InterfaceInfo[]) { - { INTERFACE_CONVENTIONAL_PCI_DEVICE }, - { INTERFACE_RDMA_PROVIDER }, - { } - } -}; - -static void register_types(void) -{ - type_register_static(&pvrdma_info); -} - -type_init(register_types) diff --git a/hw/rdma/vmw/pvrdma_qp_ops.c b/hw/rdma/vmw/pvrdma_qp_ops.c deleted file mode 100644 index c30c8344f6..0000000000 --- a/hw/rdma/vmw/pvrdma_qp_ops.c +++ /dev/null @@ -1,298 +0,0 @@ -/* - * QEMU paravirtual RDMA - QP implementation - * - * Copyright (C) 2018 Oracle - * Copyright (C) 2018 Red Hat Inc - * - * Authors: - * Yuval Shaia - * Marcel Apfelbaum - * - * This work is licensed under the terms of the GNU GPL, version 2 or late= r. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" - -#include "../rdma_utils.h" -#include "../rdma_rm.h" -#include "../rdma_backend.h" - -#include "trace.h" - -#include "pvrdma.h" -#include "standard-headers/rdma/vmw_pvrdma-abi.h" -#include "pvrdma_qp_ops.h" - -typedef struct CompHandlerCtx { - PVRDMADev *dev; - uint32_t cq_handle; - struct pvrdma_cqe cqe; -} CompHandlerCtx; - -/* Send Queue WQE */ -typedef struct PvrdmaSqWqe { - struct pvrdma_sq_wqe_hdr hdr; - struct pvrdma_sge sge[]; -} PvrdmaSqWqe; - -/* Recv Queue WQE */ -typedef struct PvrdmaRqWqe { - struct pvrdma_rq_wqe_hdr hdr; - struct pvrdma_sge sge[]; -} PvrdmaRqWqe; - -/* - * 1. Put CQE on send CQ ring - * 2. Put CQ number on dsr completion ring - * 3. Interrupt host - */ -static int pvrdma_post_cqe(PVRDMADev *dev, uint32_t cq_handle, - struct pvrdma_cqe *cqe, struct ibv_wc *wc) -{ - struct pvrdma_cqe *cqe1; - struct pvrdma_cqne *cqne; - PvrdmaRing *ring; - RdmaRmCQ *cq =3D rdma_rm_get_cq(&dev->rdma_dev_res, cq_handle); - - if (unlikely(!cq)) { - return -EINVAL; - } - - ring =3D (PvrdmaRing *)cq->opaque; - - /* Step #1: Put CQE on CQ ring */ - cqe1 =3D pvrdma_ring_next_elem_write(ring); - if (unlikely(!cqe1)) { - return -EINVAL; - } - - memset(cqe1, 0, sizeof(*cqe1)); - cqe1->wr_id =3D cqe->wr_id; - cqe1->qp =3D cqe->qp ? cqe->qp : wc->qp_num; - cqe1->opcode =3D cqe->opcode; - cqe1->status =3D wc->status; - cqe1->byte_len =3D wc->byte_len; - cqe1->src_qp =3D wc->src_qp; - cqe1->wc_flags =3D wc->wc_flags; - cqe1->vendor_err =3D wc->vendor_err; - - trace_pvrdma_post_cqe(cq_handle, cq->notify, cqe1->wr_id, cqe1->qp, - cqe1->opcode, cqe1->status, cqe1->byte_len, - cqe1->src_qp, cqe1->wc_flags, cqe1->vendor_err); - - pvrdma_ring_write_inc(ring); - - /* Step #2: Put CQ number on dsr completion ring */ - cqne =3D pvrdma_ring_next_elem_write(&dev->dsr_info.cq); - if (unlikely(!cqne)) { - return -EINVAL; - } - - cqne->info =3D cq_handle; - pvrdma_ring_write_inc(&dev->dsr_info.cq); - - if (cq->notify !=3D CNT_CLEAR) { - if (cq->notify =3D=3D CNT_ARM) { - cq->notify =3D CNT_CLEAR; - } - post_interrupt(dev, INTR_VEC_CMD_COMPLETION_Q); - } - - return 0; -} - -static void pvrdma_qp_ops_comp_handler(void *ctx, struct ibv_wc *wc) -{ - CompHandlerCtx *comp_ctx =3D (CompHandlerCtx *)ctx; - - pvrdma_post_cqe(comp_ctx->dev, comp_ctx->cq_handle, &comp_ctx->cqe, wc= ); - - g_free(ctx); -} - -static void complete_with_error(uint32_t vendor_err, void *ctx) -{ - struct ibv_wc wc =3D {}; - - wc.status =3D IBV_WC_GENERAL_ERR; - wc.vendor_err =3D vendor_err; - - pvrdma_qp_ops_comp_handler(ctx, &wc); -} - -void pvrdma_qp_ops_fini(void) -{ - rdma_backend_unregister_comp_handler(); -} - -int pvrdma_qp_ops_init(void) -{ - rdma_backend_register_comp_handler(pvrdma_qp_ops_comp_handler); - - return 0; -} - -void pvrdma_qp_send(PVRDMADev *dev, uint32_t qp_handle) -{ - RdmaRmQP *qp; - PvrdmaSqWqe *wqe; - PvrdmaRing *ring; - int sgid_idx; - union ibv_gid *sgid; - - qp =3D rdma_rm_get_qp(&dev->rdma_dev_res, qp_handle); - if (unlikely(!qp)) { - return; - } - - ring =3D (PvrdmaRing *)qp->opaque; - - wqe =3D pvrdma_ring_next_elem_read(ring); - while (wqe) { - CompHandlerCtx *comp_ctx; - - /* Prepare CQE */ - comp_ctx =3D g_new(CompHandlerCtx, 1); - comp_ctx->dev =3D dev; - comp_ctx->cq_handle =3D qp->send_cq_handle; - comp_ctx->cqe.wr_id =3D wqe->hdr.wr_id; - comp_ctx->cqe.qp =3D qp_handle; - comp_ctx->cqe.opcode =3D IBV_WC_SEND; - - sgid =3D rdma_rm_get_gid(&dev->rdma_dev_res, wqe->hdr.wr.ud.av.gid= _index); - if (!sgid) { - rdma_error_report("Failed to get gid for idx %d", - wqe->hdr.wr.ud.av.gid_index); - complete_with_error(VENDOR_ERR_INV_GID_IDX, comp_ctx); - continue; - } - - sgid_idx =3D rdma_rm_get_backend_gid_index(&dev->rdma_dev_res, - &dev->backend_dev, - wqe->hdr.wr.ud.av.gid_ind= ex); - if (sgid_idx <=3D 0) { - rdma_error_report("Failed to get bk sgid_idx for sgid_idx %d", - wqe->hdr.wr.ud.av.gid_index); - complete_with_error(VENDOR_ERR_INV_GID_IDX, comp_ctx); - continue; - } - - if (wqe->hdr.num_sge > dev->dev_attr.max_sge) { - rdma_error_report("Invalid num_sge=3D%d (max %d)", wqe->hdr.nu= m_sge, - dev->dev_attr.max_sge); - complete_with_error(VENDOR_ERR_INV_NUM_SGE, comp_ctx); - continue; - } - - rdma_backend_post_send(&dev->backend_dev, &qp->backend_qp, qp->qp_= type, - (struct ibv_sge *)&wqe->sge[0], wqe->hdr.nu= m_sge, - sgid_idx, sgid, - (union ibv_gid *)wqe->hdr.wr.ud.av.dgid, - wqe->hdr.wr.ud.remote_qpn, - wqe->hdr.wr.ud.remote_qkey, comp_ctx); - - pvrdma_ring_read_inc(ring); - - wqe =3D pvrdma_ring_next_elem_read(ring); - } -} - -void pvrdma_qp_recv(PVRDMADev *dev, uint32_t qp_handle) -{ - RdmaRmQP *qp; - PvrdmaRqWqe *wqe; - PvrdmaRing *ring; - - qp =3D rdma_rm_get_qp(&dev->rdma_dev_res, qp_handle); - if (unlikely(!qp)) { - return; - } - - ring =3D &((PvrdmaRing *)qp->opaque)[1]; - - wqe =3D pvrdma_ring_next_elem_read(ring); - while (wqe) { - CompHandlerCtx *comp_ctx; - - /* Prepare CQE */ - comp_ctx =3D g_new(CompHandlerCtx, 1); - comp_ctx->dev =3D dev; - comp_ctx->cq_handle =3D qp->recv_cq_handle; - comp_ctx->cqe.wr_id =3D wqe->hdr.wr_id; - comp_ctx->cqe.qp =3D qp_handle; - comp_ctx->cqe.opcode =3D IBV_WC_RECV; - - if (wqe->hdr.num_sge > dev->dev_attr.max_sge) { - rdma_error_report("Invalid num_sge=3D%d (max %d)", wqe->hdr.nu= m_sge, - dev->dev_attr.max_sge); - complete_with_error(VENDOR_ERR_INV_NUM_SGE, comp_ctx); - continue; - } - - rdma_backend_post_recv(&dev->backend_dev, &qp->backend_qp, qp->qp_= type, - (struct ibv_sge *)&wqe->sge[0], wqe->hdr.nu= m_sge, - comp_ctx); - - pvrdma_ring_read_inc(ring); - - wqe =3D pvrdma_ring_next_elem_read(ring); - } -} - -void pvrdma_srq_recv(PVRDMADev *dev, uint32_t srq_handle) -{ - RdmaRmSRQ *srq; - PvrdmaRqWqe *wqe; - PvrdmaRing *ring; - - srq =3D rdma_rm_get_srq(&dev->rdma_dev_res, srq_handle); - if (unlikely(!srq)) { - return; - } - - ring =3D (PvrdmaRing *)srq->opaque; - - wqe =3D pvrdma_ring_next_elem_read(ring); - while (wqe) { - CompHandlerCtx *comp_ctx; - - /* Prepare CQE */ - comp_ctx =3D g_new(CompHandlerCtx, 1); - comp_ctx->dev =3D dev; - comp_ctx->cq_handle =3D srq->recv_cq_handle; - comp_ctx->cqe.wr_id =3D wqe->hdr.wr_id; - comp_ctx->cqe.qp =3D 0; - comp_ctx->cqe.opcode =3D IBV_WC_RECV; - - if (wqe->hdr.num_sge > dev->dev_attr.max_sge) { - rdma_error_report("Invalid num_sge=3D%d (max %d)", wqe->hdr.nu= m_sge, - dev->dev_attr.max_sge); - complete_with_error(VENDOR_ERR_INV_NUM_SGE, comp_ctx); - continue; - } - - rdma_backend_post_srq_recv(&dev->backend_dev, &srq->backend_srq, - (struct ibv_sge *)&wqe->sge[0], - wqe->hdr.num_sge, - comp_ctx); - - pvrdma_ring_read_inc(ring); - - wqe =3D pvrdma_ring_next_elem_read(ring); - } - -} - -void pvrdma_cq_poll(RdmaDeviceResources *dev_res, uint32_t cq_handle) -{ - RdmaRmCQ *cq; - - cq =3D rdma_rm_get_cq(dev_res, cq_handle); - if (!cq) { - return; - } - - rdma_backend_poll_cq(dev_res, &cq->backend_cq); -} diff --git a/monitor/qmp-cmds.c b/monitor/qmp-cmds.c index b0f948d337..f84a0dc523 100644 --- a/monitor/qmp-cmds.c +++ b/monitor/qmp-cmds.c @@ -31,7 +31,6 @@ #include "qapi/type-helpers.h" #include "hw/mem/memory-device.h" #include "hw/intc/intc.h" -#include "hw/rdma/rdma.h" =20 NameInfo *qmp_query_name(Error **errp) { diff --git a/Kconfig.host b/Kconfig.host index f496475f8e..f6a2a131e6 100644 --- a/Kconfig.host +++ b/Kconfig.host @@ -35,9 +35,6 @@ config VHOST_KERNEL config VIRTFS bool =20 -config PVRDMA - bool - config MULTIPROCESS_ALLOWED bool imply MULTIPROCESS diff --git a/contrib/rdmacm-mux/meson.build b/contrib/rdmacm-mux/meson.build deleted file mode 100644 index 36c9c89630..0000000000 --- a/contrib/rdmacm-mux/meson.build +++ /dev/null @@ -1,7 +0,0 @@ -if have_pvrdma - # FIXME: broken on big endian architectures - executable('rdmacm-mux', files('main.c'), genh, - dependencies: [glib, libumad], - build_by_default: false, - install: false) -endif diff --git a/hmp-commands-info.hx b/hmp-commands-info.hx index ad1b1306e3..20a9835ea8 100644 --- a/hmp-commands-info.hx +++ b/hmp-commands-info.hx @@ -182,19 +182,6 @@ SRST Show PIC state. ERST =20 - { - .name =3D "rdma", - .args_type =3D "", - .params =3D "", - .help =3D "show RDMA state", - .cmd_info_hrt =3D qmp_x_query_rdma, - }, - -SRST - ``info rdma`` - Show RDMA state. -ERST - { .name =3D "pci", .args_type =3D "", diff --git a/hw/Kconfig b/hw/Kconfig index 2c00936c28..32f876deb0 100644 --- a/hw/Kconfig +++ b/hw/Kconfig @@ -29,7 +29,6 @@ source pci-bridge/Kconfig source pci-host/Kconfig source pcmcia/Kconfig source pci/Kconfig -source rdma/Kconfig source remote/Kconfig source rtc/Kconfig source scsi/Kconfig diff --git a/hw/meson.build b/hw/meson.build index 463d702683..3049a6fab0 100644 --- a/hw/meson.build +++ b/hw/meson.build @@ -28,7 +28,6 @@ subdir('pci') subdir('pci-bridge') subdir('pci-host') subdir('pcmcia') -subdir('rdma') subdir('rtc') subdir('scsi') subdir('sd') diff --git a/hw/rdma/Kconfig b/hw/rdma/Kconfig deleted file mode 100644 index 840320bdc0..0000000000 --- a/hw/rdma/Kconfig +++ /dev/null @@ -1,3 +0,0 @@ -config VMW_PVRDMA - default y if PCI_DEVICES - depends on PVRDMA && MSI_NONBROKEN && VMXNET3_PCI diff --git a/hw/rdma/meson.build b/hw/rdma/meson.build deleted file mode 100644 index 363c9b8c83..0000000000 --- a/hw/rdma/meson.build +++ /dev/null @@ -1,12 +0,0 @@ -system_ss.add(when: 'CONFIG_VMW_PVRDMA', if_true: files( - 'rdma.c', - 'rdma_backend.c', - 'rdma_utils.c', - 'vmw/pvrdma_qp_ops.c', -)) -specific_ss.add(when: 'CONFIG_VMW_PVRDMA', if_true: files( - 'rdma_rm.c', - 'vmw/pvrdma_cmd.c', - 'vmw/pvrdma_dev_ring.c', - 'vmw/pvrdma_main.c', -)) diff --git a/hw/rdma/trace-events b/hw/rdma/trace-events deleted file mode 100644 index c23175120e..0000000000 --- a/hw/rdma/trace-events +++ /dev/null @@ -1,31 +0,0 @@ -# See docs/devel/tracing.rst for syntax documentation. - -# rdma_backend.c -rdma_check_dev_attr(const char *name, int max_bk, int max_fe) "%s: be=3D%d= , fe=3D%d" -rdma_create_ah_cache_hit(uint64_t subnet, uint64_t if_id) "subnet=3D0x%"PR= Ix64",if_id=3D0x%"PRIx64 -rdma_create_ah_cache_miss(uint64_t subnet, uint64_t if_id) "subnet=3D0x%"P= RIx64",if_id=3D0x%"PRIx64 -rdma_poll_cq(int ne, void *ibcq) "Got %d completion(s) from cq %p" -rdmacm_mux(const char *title, int msg_type, int op_code) "%s: msg_type=3D%= d, op_code=3D%d" -rdmacm_mux_check_op_status(int msg_type, int op_code, int err_code) "resp:= msg_type=3D%d, op_code=3D%d, err_code=3D%d" -rdma_mad_message(const char *title, int len, char *data) "mad %s (%d): %s" -rdma_backend_rc_qp_state_init(uint32_t qpn) "RC QP 0x%x switch to INIT" -rdma_backend_ud_qp_state_init(uint32_t qpn, uint32_t qkey) "UD QP 0x%x swi= tch to INIT, qkey=3D0x%x" -rdma_backend_rc_qp_state_rtr(uint32_t qpn, uint64_t subnet, uint64_t ifid,= uint8_t sgid_idx, uint32_t dqpn, uint32_t rq_psn) "RC QP 0x%x switch to RT= R, subnet =3D 0x%"PRIx64", ifid =3D 0x%"PRIx64 ", sgid_idx=3D%d, dqpn=3D0x%= x, rq_psn=3D0x%x" -rdma_backend_ud_qp_state_rtr(uint32_t qpn, uint32_t qkey) "UD QP 0x%x swit= ch to RTR, qkey=3D0x%x" -rdma_backend_rc_qp_state_rts(uint32_t qpn, uint32_t sq_psn) "RC QP 0x%x sw= itch to RTS, sq_psn=3D0x%x, " -rdma_backend_ud_qp_state_rts(uint32_t qpn, uint32_t sq_psn, uint32_t qkey)= "UD QP 0x%x switch to RTS, sq_psn=3D0x%x, qkey=3D0x%x" -rdma_backend_get_gid_index(uint64_t subnet, uint64_t ifid, int gid_idx) "s= ubnet=3D0x%"PRIx64", ifid=3D0x%"PRIx64 ", gid_idx=3D%d" -rdma_backend_gid_change(const char *op, uint64_t subnet, uint64_t ifid) "%= s subnet=3D0x%"PRIx64", ifid=3D0x%"PRIx64 - -# rdma_rm.c -rdma_res_tbl_get(char *name, uint32_t handle) "tbl %s, handle %d" -rdma_res_tbl_alloc(char *name, uint32_t handle) "tbl %s, handle %d" -rdma_res_tbl_dealloc(char *name, uint32_t handle) "tbl %s, handle %d" -rdma_rm_alloc_mr(uint32_t mr_handle, void *host_virt, uint64_t guest_start= , uint64_t guest_length, int access_flags) "mr_handle=3D%d, host_virt=3D%p,= guest_start=3D0x%"PRIx64", length=3D%" PRId64", access_flags=3D0x%x" -rdma_rm_dealloc_mr(uint32_t mr_handle, uint64_t guest_start) "mr_handle=3D= %d, guest_start=3D0x%"PRIx64 -rdma_rm_alloc_qp(uint32_t rm_qpn, uint32_t backend_qpn, uint8_t qp_type) "= rm_qpn=3D%d, backend_qpn=3D0x%x, qp_type=3D%d" -rdma_rm_modify_qp(uint32_t qpn, uint32_t attr_mask, int qp_state, uint8_t = sgid_idx) "qpn=3D0x%x, attr_mask=3D0x%x, qp_state=3D%d, sgid_idx=3D%d" - -# rdma_utils.c -rdma_pci_dma_map(uint64_t addr, void *vaddr, uint64_t len) "0x%"PRIx64" ->= %p (len=3D%" PRIu64")" -rdma_pci_dma_unmap(void *vaddr) "%p" diff --git a/hw/rdma/vmw/trace-events b/hw/rdma/vmw/trace-events deleted file mode 100644 index a6c77e1e10..0000000000 --- a/hw/rdma/vmw/trace-events +++ /dev/null @@ -1,17 +0,0 @@ -# See docs/devel/tracing.rst for syntax documentation. - -# pvrdma_main.c -pvrdma_regs_read(uint64_t addr, uint64_t val) "pvrdma.regs[0x%"PRIx64"]=3D= 0x%"PRIx64 -pvrdma_regs_write(uint64_t addr, uint64_t val, const char *reg_name, const= char *val_name) "pvrdma.regs[0x%"PRIx64"]=3D0x%"PRIx64" (%s %s)" -pvrdma_uar_write(uint64_t addr, uint64_t val, const char *reg_name, const = char *val_name, int val1, int val2) "uar[0x%"PRIx64"]=3D0x%"PRIx64" (cls=3D= %s, op=3D%s, obj=3D%d, val=3D%d)" - -# pvrdma_cmd.c -pvrdma_map_to_pdir_host_virt(void *vfirst, void *vremaped) "mremap %p -> %= p" -pvrdma_map_to_pdir_next_page(int page_idx, void *vnext, void *vremaped) "m= remap [%d] %p -> %p" -pvrdma_exec_cmd(int cmd, int err) "cmd=3D%d, err=3D%d" - -# pvrdma_dev_ring.c -pvrdma_ring_next_elem_read_no_data(char *ring_name) "pvrdma_ring %s is emp= ty" - -# pvrdma_qp_ops.c -pvrdma_post_cqe(uint32_t cq_handle, int notify, uint64_t wr_id, uint64_t q= pn, uint32_t op_code, uint32_t status, uint32_t byte_len, uint32_t src_qp, = uint32_t wc_flags, uint32_t vendor_err) "cq_handle=3D%d, notify=3D%d, wr_id= =3D0x%"PRIx64", qpn=3D0x%"PRIx64", opcode=3D%d, status=3D%d, byte_len=3D%d,= src_qp=3D%d, wc_flags=3D%d, vendor_err=3D%d" diff --git a/meson_options.txt b/meson_options.txt index 0a99a059ec..b5c0bad9e7 100644 --- a/meson_options.txt +++ b/meson_options.txt @@ -198,8 +198,6 @@ option('opengl', type : 'feature', value : 'auto', description: 'OpenGL support') option('rdma', type : 'feature', value : 'auto', description: 'Enable RDMA-based migration') -option('pvrdma', type : 'feature', value : 'auto', - description: 'Enable PVRDMA support') option('gtk', type : 'feature', value : 'auto', description: 'GTK+ user interface') option('sdl', type : 'feature', value : 'auto', diff --git a/qapi/meson.build b/qapi/meson.build index 375d564277..c92af6e063 100644 --- a/qapi/meson.build +++ b/qapi/meson.build @@ -62,7 +62,6 @@ if have_system 'cryptodev', 'qdev', 'pci', - 'rdma', 'rocker', 'tpm', ] diff --git a/qemu-options.hx b/qemu-options.hx index 7fd1713fa8..f7ef9b4e41 100644 --- a/qemu-options.hx +++ b/qemu-options.hx @@ -5113,9 +5113,6 @@ SRST allows a co-operating external process to access the QEMU memory region. =20 - The ``share`` is also required for pvrdma devices due to - limitations in the RDMA API provided by Linux. - Setting share=3Don might affect the ability to configure NUMA bindings for the memory backend under some circumstances, see Documentation/vm/numa\_memory\_policy.txt on the Linux kernel diff --git a/scripts/ci/org.centos/stream/8/x86_64/configure b/scripts/ci/o= rg.centos/stream/8/x86_64/configure index 76781f17f4..868db665f6 100755 --- a/scripts/ci/org.centos/stream/8/x86_64/configure +++ b/scripts/ci/org.centos/stream/8/x86_64/configure @@ -99,7 +99,6 @@ --disable-opengl \ --disable-parallels \ --disable-pie \ ---disable-pvrdma \ --disable-qcow1 \ --disable-qed \ --disable-qom-cast-debug \ diff --git a/scripts/meson-buildoptions.sh b/scripts/meson-buildoptions.sh index 680fa3f581..5ace33f167 100644 --- a/scripts/meson-buildoptions.sh +++ b/scripts/meson-buildoptions.sh @@ -163,7 +163,6 @@ meson_options_help() { printf "%s\n" ' pixman pixman support' printf "%s\n" ' plugins TCG plugins via shared library loading' printf "%s\n" ' png PNG support with libpng' - printf "%s\n" ' pvrdma Enable PVRDMA support' printf "%s\n" ' qcow1 qcow1 image format support' printf "%s\n" ' qed qed image format support' printf "%s\n" ' qga-vss build QGA VSS support (broken with MinG= W)' @@ -428,8 +427,6 @@ _meson_option_parse() { --enable-png) printf "%s" -Dpng=3Denabled ;; --disable-png) printf "%s" -Dpng=3Ddisabled ;; --prefix=3D*) quote_sh "-Dprefix=3D$2" ;; - --enable-pvrdma) printf "%s" -Dpvrdma=3Denabled ;; - --disable-pvrdma) printf "%s" -Dpvrdma=3Ddisabled ;; --enable-qcow1) printf "%s" -Dqcow1=3Denabled ;; --disable-qcow1) printf "%s" -Dqcow1=3Ddisabled ;; --enable-qed) printf "%s" -Dqed=3Denabled ;; diff --git a/scripts/update-linux-headers.sh b/scripts/update-linux-headers= .sh index a0006eec6f..73c292bbac 100755 --- a/scripts/update-linux-headers.sh +++ b/scripts/update-linux-headers.sh @@ -55,7 +55,6 @@ cp_portable() { -e 'linux/if_ether' \ -e 'input-event-codes' \ -e 'sys/' \ - -e 'pvrdma_verbs' \ -e 'drm.h' \ -e 'limits' \ -e 'linux/const' \ @@ -226,32 +225,6 @@ mkdir -p "$output/include/standard-headers/drm" cp_portable "$tmpdir/include/drm/drm_fourcc.h" \ "$output/include/standard-headers/drm" =20 -rm -rf "$output/include/standard-headers/drivers/infiniband/hw/vmw_pvrdma" -mkdir -p "$output/include/standard-headers/drivers/infiniband/hw/vmw_pvrdm= a" - -# Remove the unused functions from pvrdma_verbs.h avoiding the unnecessary -# import of several infiniband/networking/other headers -tmp_pvrdma_verbs=3D"$tmpdir/pvrdma_verbs.h" -# Parse the entire file instead of single lines to match -# function declarations expanding over multiple lines -# and strip the declarations starting with pvrdma prefix. -sed -e '1h;2,$H;$!d;g' -e 's/[^};]*pvrdma[^(| ]*([^)]*);//g' \ - "$linux/drivers/infiniband/hw/vmw_pvrdma/pvrdma_verbs.h" > \ - "$tmp_pvrdma_verbs"; - -for i in "$linux/drivers/infiniband/hw/vmw_pvrdma/pvrdma_dev_api.h" \ - "$tmp_pvrdma_verbs"; do \ - cp_portable "$i" \ - "$output/include/standard-headers/drivers/infiniband/hw/vmw_pvrdm= a/" -done - -rm -rf "$output/include/standard-headers/rdma/" -mkdir -p "$output/include/standard-headers/rdma/" -for i in "$tmpdir/include/rdma/vmw_pvrdma-abi.h"; do - cp_portable "$i" \ - "$output/include/standard-headers/rdma/" -done - cat <$output/include/standard-headers/linux/types.h /* For QEMU all types are already defined via osdep.h, so this * header does not need to do anything. --=20 2.41.0 From nobody Sun Nov 10 21:34:25 2024 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass(p=none dis=none) header.from=linaro.org ARC-Seal: i=1; a=rsa-sha256; t=1711631077; cv=none; d=zohomail.com; s=zohoarc; b=eFmg7rZkjSReaam7OW3JwsrRR3d6X8dzFgIvOG1MMPzGKoyqU7BaFruXFHqVFF6FC/yjhIj8+Skm6rKq4UxLvesOY6cwmKUzVfhPCjTlXXcD5GdKyd9UVSsljPJtQiF4zVUNWn9jKH1hBuEc1/DeU2rBS0exZIF6xE4ia4yf7Zo= ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com; s=zohoarc; t=1711631077; h=Content-Type:Content-Transfer-Encoding:Cc:Cc:Date:Date:From:From:In-Reply-To:List-Subscribe:List-Post:List-Id:List-Archive:List-Help:List-Unsubscribe:MIME-Version:Message-ID:References:Sender:Subject:Subject:To:To:Message-Id:Reply-To; bh=HFWBx5PjuobBYcjdI3h0rBAitT/k2mYN7gZffdXVLXw=; b=FxbWoHNzqrLASkS17zXig/6bY9d7zepAaPlz5WtR5yevnVpzYOjJ6zSzmpPcDd6sTJxvgS0Z5FdJ3GKTAHoO4N4XINUaUK0jyIgmAd77o7POZHO/mEJSPRT39Ib8aCmcbsOzaD0rb4GcRmaHq0pYiZStEOSPckAbFZbzYvcaQms= ARC-Authentication-Results: i=1; mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass header.from= (p=none dis=none) Return-Path: Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 1711631077574418.59923370264926; Thu, 28 Mar 2024 06:04:37 -0700 (PDT) Received: from localhost ([::1] helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1rppQI-0000lY-Mh; Thu, 28 Mar 2024 09:04:03 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1rppPn-0000KC-Id for qemu-devel@nongnu.org; Thu, 28 Mar 2024 09:03:31 -0400 Received: from mail-ej1-x635.google.com ([2a00:1450:4864:20::635]) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1rppPi-00088f-AG for qemu-devel@nongnu.org; Thu, 28 Mar 2024 09:03:31 -0400 Received: by mail-ej1-x635.google.com with SMTP id a640c23a62f3a-a4715d4c2cbso122864966b.1 for ; Thu, 28 Mar 2024 06:03:25 -0700 (PDT) Received: from m1x-phil.lan (pas38-h02-176-184-5-52.dsl.sta.abo.bbox.fr. [176.184.5.52]) by smtp.gmail.com with ESMTPSA id w20-20020a170906481400b00a46aab70226sm728091ejq.180.2024.03.28.06.03.17 (version=TLS1_3 cipher=TLS_CHACHA20_POLY1305_SHA256 bits=256/256); Thu, 28 Mar 2024 06:03:20 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; t=1711631004; x=1712235804; darn=nongnu.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=HFWBx5PjuobBYcjdI3h0rBAitT/k2mYN7gZffdXVLXw=; b=brtmVkoJIyvkknXmWE4zersI27K8Lq2BEQ5RApwUJia1PBnWxhCR726BJ2QRdtrqr3 lkat8vfk3VaLHq/j/HHAAc3w6f8sERFYNE6yo8Oe9GUkXt3k2XEZvALy3vbAmzkbwJq6 roKWsPSemePUWPVPhF378jT3hila1yaGw6gYnOmfcfGMsXm3Hfnc+2w6p7FV77h6vYhu cKco3KAb6+jDE+AI3J65U9XMCst1N0uqOUhycYs4jvsXCjmR2a7yaDwVkbWJOwXB3jOD 4NPMq2Ru4BslFWGoj+qsEqk9Lm9L10/SKDk/svleCoHykljGhycxZpNKu5l686PEgz49 0a5A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1711631004; x=1712235804; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=HFWBx5PjuobBYcjdI3h0rBAitT/k2mYN7gZffdXVLXw=; b=qfm01v74sUCs0M+DP3a652TZ3KKye3s6i4wLtNsBHqoSPidZMuo1/d70WjrgvTwTAK /X/QO0mmkNqAhPiJGS++Q7GoJilz4daAZ6xdxasuwQgDWg9XXEUo3r+9S5dmFgQFdIrd WVO3D08i7MFDzbh/oFsTppeJuxTijz5bnQavg3ritDmqFg5p61hGHWFlpK6S9DNcg9O6 e5xehvMW/DrLyWSpcv8d6shP0rmkg7/WYAaaar00R47jxrReClYIbZbt0Sxq5a3ZYmaM Mii13ijnCX9DGPsnOQ0pOHyfPSwHAplVSYAL6Yi5yjq+6Ri1zKMMh50qATFyRkgE9VBJ OYQw== X-Gm-Message-State: AOJu0YzA45eCyfR8Z3LEy5S5SSqOMu8jisVw4ThT1M7/++YaP/hMqmlX YxuJM7G+av1Uf22oxgQk39nHOR6otr1kN7G3RB6+pmOV9DiUuhXMayAeSdX4YWmEv6gxPuZp5rE o X-Google-Smtp-Source: AGHT+IGcSu8BA3bKL2EDAhVaJrUa0Q7XcP2FvVzOab0gf0MkPowJQk8Fo/yrXhopyeckoQItynaHAg== X-Received: by 2002:a17:906:d553:b0:a4e:14c:227c with SMTP id cr19-20020a170906d55300b00a4e014c227cmr1949830ejc.68.1711631001769; Thu, 28 Mar 2024 06:03:21 -0700 (PDT) From: =?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?= To: qemu-devel@nongnu.org Cc: Yuval Shaia , Kevin Wolf , Prasanna Kumar Kalever , Fabiano Rosas , Cornelia Huck , Michael Roth , Li Zhijian , Prasanna Kumar Kalever , Peter Xu , integration@gluster.org, Paolo Bonzini , qemu-block@nongnu.org, =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= , devel@lists.libvirt.org, Hanna Reitz , "Michael S. Tsirkin" , Thomas Huth , Eric Blake , =?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?= , Song Gao , =?UTF-8?q?Marc-Andr=C3=A9=20Lureau?= , Markus Armbruster , =?UTF-8?q?Alex=20Benn=C3=A9e?= , Wainer dos Santos Moschetta , Beraldo Leal , Peter Maydell Subject: [PATCH-for-9.1 v2 2/3] migration: Remove RDMA protocol handling Date: Thu, 28 Mar 2024 14:02:54 +0100 Message-ID: <20240328130255.52257-3-philmd@linaro.org> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20240328130255.52257-1-philmd@linaro.org> References: <20240328130255.52257-1-philmd@linaro.org> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Received-SPF: pass client-ip=2a00:1450:4864:20::635; envelope-from=philmd@linaro.org; helo=mail-ej1-x635.google.com X-Spam_score_int: -20 X-Spam_score: -2.1 X-Spam_bar: -- X-Spam_report: (-2.1 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, RCVD_IN_DNSWL_NONE=-0.0001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=unavailable autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: qemu-devel-bounces+importer=patchew.org@nongnu.org X-ZohoMail-DKIM: pass (identity @linaro.org) X-ZM-MESSAGEID: 1711631079708100001 The whole RDMA subsystem was deprecated in commit e9a54265f5 ("hw/rdma: Deprecate the pvrdma device and the rdma subsystem") released in v8.2. Remove: - RDMA handling from migration - dependencies on libibumad, libibverbs and librdmacm Keep the RAM_SAVE_FLAG_HOOK definition since it might appears in old migration streams. Cc: Peter Xu Cc: Li Zhijian Acked-by: Fabiano Rosas Signed-off-by: Philippe Mathieu-Daud=C3=A9 Reviewed-by: Michael Galaxy Reviewed-by: Michael S. Tsirkin --- MAINTAINERS | 7 - docs/devel/migration/main.rst | 6 - docs/rdma.txt | 420 -- docs/system/loongarch/virt.rst | 2 +- meson.build | 23 - qapi/migration.json | 31 +- migration/migration-stats.h | 6 +- migration/migration.h | 9 - migration/options.h | 2 - migration/rdma.h | 69 - migration/migration-stats.c | 5 +- migration/migration.c | 31 - migration/options.c | 16 - migration/qemu-file.c | 1 - migration/ram.c | 86 +- migration/rdma.c | 4184 ----------------- migration/savevm.c | 2 +- meson_options.txt | 2 - migration/meson.build | 1 - migration/trace-events | 68 +- qemu-options.hx | 3 - .../org.centos/stream/8/build-environment.yml | 1 - .../ci/org.centos/stream/8/x86_64/configure | 2 - scripts/ci/setup/build-environment.yml | 4 - scripts/coverity-scan/run-coverity-scan | 2 +- scripts/meson-buildoptions.sh | 3 - tests/lcitool/projects/qemu.yml | 3 - tests/migration/guestperf/engine.py | 4 +- 28 files changed, 14 insertions(+), 4979 deletions(-) delete mode 100644 docs/rdma.txt delete mode 100644 migration/rdma.h delete mode 100644 migration/rdma.c diff --git a/MAINTAINERS b/MAINTAINERS index 91ab5235b8..05226cea0a 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3426,13 +3426,6 @@ F: docs/devel/migration.rst F: qapi/migration.json F: tests/migration/ F: util/userfaultfd.c -X: migration/rdma* - -RDMA Migration -R: Li Zhijian -R: Peter Xu -S: Odd Fixes -F: migration/rdma* =20 Migration dirty limit and dirty page rate M: Hyman Huang diff --git a/docs/devel/migration/main.rst b/docs/devel/migration/main.rst index 54385a23e5..70278ce1e3 100644 --- a/docs/devel/migration/main.rst +++ b/docs/devel/migration/main.rst @@ -47,12 +47,6 @@ over any transport. QEMU interference. Note that QEMU does not flush cached file data/metadata at the end of migration. =20 -In addition, support is included for migration using RDMA, which -transports the page data using ``RDMA``, where the hardware takes care of -transporting the pages, and the load on the CPU is much lower. While the -internals of RDMA migration are a bit different, this isn't really visible -outside the RAM migration code. - All these migration protocols use the same infrastructure to save/restore state devices. This infrastructure is shared with the savevm/loadvm functionality. diff --git a/docs/rdma.txt b/docs/rdma.txt deleted file mode 100644 index bd8dd799a9..0000000000 --- a/docs/rdma.txt +++ /dev/null @@ -1,420 +0,0 @@ -(RDMA: Remote Direct Memory Access) -RDMA Live Migration Specification, Version # 1 -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D -Wiki: https://wiki.qemu.org/Features/RDMALiveMigration -Github: git@github.com:hinesmr/qemu.git, 'rdma' branch - -Copyright (C) 2013 Michael R. Hines - -An *exhaustive* paper (2010) shows additional performance details -linked on the QEMU wiki above. - -Contents: -=3D=3D=3D=3D=3D=3D=3D=3D=3D -* Introduction -* Before running -* Running -* Performance -* RDMA Migration Protocol Description -* Versioning and Capabilities -* QEMUFileRDMA Interface -* Migration of VM's ram -* Error handling -* TODO - -Introduction: -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - -RDMA helps make your migration more deterministic under heavy load because -of the significantly lower latency and higher throughput over TCP/IP. This= is -because the RDMA I/O architecture reduces the number of interrupts and -data copies by bypassing the host networking stack. In particular, a TCP-b= ased -migration, under certain types of memory-bound workloads, may take a more -unpredictable amount of time to complete the migration if the amount of -memory tracked during each live migration iteration round cannot keep pace -with the rate of dirty memory produced by the workload. - -RDMA currently comes in two flavors: both Ethernet based (RoCE, or RDMA -over Converged Ethernet) as well as Infiniband-based. This implementation = of -migration using RDMA is capable of using both technologies because of -the use of the OpenFabrics OFED software stack that abstracts out the -programming model irrespective of the underlying hardware. - -Refer to openfabrics.org or your respective RDMA hardware vendor for -an understanding on how to verify that you have the OFED software stack -installed in your environment. You should be able to successfully link -against the "librdmacm" and "libibverbs" libraries and development headers -for a working build of QEMU to run successfully using RDMA Migration. - -BEFORE RUNNING: -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - -Use of RDMA during migration requires pinning and registering memory -with the hardware. This means that memory must be physically resident -before the hardware can transmit that memory to another machine. -If this is not acceptable for your application or product, then the use -of RDMA migration may in fact be harmful to co-located VMs or other -software on the machine if there is not sufficient memory available to -relocate the entire footprint of the virtual machine. If so, then the -use of RDMA is discouraged and it is recommended to use standard TCP migra= tion. - -Experimental: Next, decide if you want dynamic page registration. -For example, if you have an 8GB RAM virtual machine, but only 1GB -is in active use, then enabling this feature will cause all 8GB to -be pinned and resident in memory. This feature mostly affects the -bulk-phase round of the migration and can be enabled for extremely -high-performance RDMA hardware using the following command: - -QEMU Monitor Command: -$ migrate_set_capability rdma-pin-all on # disabled by default - -Performing this action will cause all 8GB to be pinned, so if that's -not what you want, then please ignore this step altogether. - -On the other hand, this will also significantly speed up the bulk round -of the migration, which can greatly reduce the "total" time of your migrat= ion. -Example performance of this using an idle VM in the previous example -can be found in the "Performance" section. - -Note: for very large virtual machines (hundreds of GBs), pinning all -*all* of the memory of your virtual machine in the kernel is very expensive -may extend the initial bulk iteration time by many seconds, -and thus extending the total migration time. However, this will not -affect the determinism or predictability of your migration you will -still gain from the benefits of advanced pinning with RDMA. - -RUNNING: -=3D=3D=3D=3D=3D=3D=3D=3D - -First, set the migration speed to match your hardware's capabilities: - -QEMU Monitor Command: -$ migrate_set_parameter max-bandwidth 40g # or whatever is the MAX of your= RDMA device - -Next, on the destination machine, add the following to the QEMU command li= ne: - -qemu ..... -incoming rdma:host:port - -Finally, perform the actual migration on the source machine: - -QEMU Monitor Command: -$ migrate -d rdma:host:port - -PERFORMANCE -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - -Here is a brief summary of total migration time and downtime using RDMA: -Using a 40gbps infiniband link performing a worst-case stress test, -using an 8GB RAM virtual machine: - -Using the following command: -$ apt-get install stress -$ stress --vm-bytes 7500M --vm 1 --vm-keep - -1. Migration throughput: 26 gigabits/second. -2. Downtime (stop time) varies between 15 and 100 milliseconds. - -EFFECTS of memory registration on bulk phase round: - -For example, in the same 8GB RAM example with all 8GB of memory in -active use and the VM itself is completely idle using the same 40 gbps -infiniband link: - -1. rdma-pin-all disabled total time: approximately 7.5 seconds @ 9.5 Gbps -2. rdma-pin-all enabled total time: approximately 4 seconds @ 26 Gbps - -These numbers would of course scale up to whatever size virtual machine -you have to migrate using RDMA. - -Enabling this feature does *not* have any measurable affect on -migration *downtime*. This is because, without this feature, all of the -memory will have already been registered already in advance during -the bulk round and does not need to be re-registered during the successive -iteration rounds. - -RDMA Protocol Description: -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D - -Migration with RDMA is separated into two parts: - -1. The transmission of the pages using RDMA -2. Everything else (a control channel is introduced) - -"Everything else" is transmitted using a formal -protocol now, consisting of infiniband SEND messages. - -An infiniband SEND message is the standard ibverbs -message used by applications of infiniband hardware. -The only difference between a SEND message and an RDMA -message is that SEND messages cause notifications -to be posted to the completion queue (CQ) on the -infiniband receiver side, whereas RDMA messages (used -for VM's ram) do not (to behave like an actual DMA). - -Messages in infiniband require two things: - -1. registration of the memory that will be transmitted -2. (SEND only) work requests to be posted on both - sides of the network before the actual transmission - can occur. - -RDMA messages are much easier to deal with. Once the memory -on the receiver side is registered and pinned, we're -basically done. All that is required is for the sender -side to start dumping bytes onto the link. - -(Memory is not released from pinning until the migration -completes, given that RDMA migrations are very fast.) - -SEND messages require more coordination because the -receiver must have reserved space (using a receive -work request) on the receive queue (RQ) before QEMUFileRDMA -can start using them to carry all the bytes as -a control transport for migration of device state. - -To begin the migration, the initial connection setup is -as follows (migration-rdma.c): - -1. Receiver and Sender are started (command line or libvirt): -2. Both sides post two RQ work requests -3. Receiver does listen() -4. Sender does connect() -5. Receiver accept() -6. Check versioning and capabilities (described later) - -At this point, we define a control channel on top of SEND messages -which is described by a formal protocol. Each SEND message has a -header portion and a data portion (but together are transmitted -as a single SEND message). - -Header: - * Length (of the data portion, uint32, network byte orde= r) - * Type (what command to perform, uint32, network byte = order) - * Repeat (Number of commands in data portion, same type = only) - -The 'Repeat' field is here to support future multiple page registrations -in a single message without any need to change the protocol itself -so that the protocol is compatible against multiple versions of QEMU. -Version #1 requires that all server implementations of the protocol must -check this field and register all requests found in the array of commands = located -in the data portion and return an equal number of results in the response. -The maximum number of repeats is hard-coded to 4096. This is a conservative -limit based on the maximum size of a SEND message along with empirical -observations on the maximum future benefit of simultaneous page registrati= ons. - -The 'type' field has 12 different command values: - 1. Unused - 2. Error (sent to the source during bad things) - 3. Ready (control-channel is available) - 4. QEMU File (for sending non-live device state) - 5. RAM Blocks request (used right after connection setup) - 6. RAM Blocks result (used right after connection setup) - 7. Compress page (zap zero page and skip registration) - 8. Register request (dynamic chunk registration) - 9. Register result ('rkey' to be used by sender) - 10. Register finished (registration for current iteration fin= ished) - 11. Unregister request (unpin previously registered memory) - 12. Unregister finished (confirmation that unpin completed) - -A single control message, as hinted above, can contain within the data -portion an array of many commands of the same type. If there is more than -one command, then the 'repeat' field will be greater than 1. - -After connection setup, message 5 & 6 are used to exchange ram block -information and optionally pin all the memory if requested by the user. - -After ram block exchange is completed, we have two protocol-level -functions, responsible for communicating control-channel commands -using the above list of values: - -Logically: - -qemu_rdma_exchange_recv(header, expected command type) - -1. We transmit a READY command to let the sender know that - we are *ready* to receive some data bytes on the control channel. -2. Before attempting to receive the expected command, we post another - RQ work request to replace the one we just used up. -3. Block on a CQ event channel and wait for the SEND to arrive. -4. When the send arrives, librdmacm will unblock us. -5. Verify that the command-type and version received matches the one we ex= pected. - -qemu_rdma_exchange_send(header, data, optional response header & data): - -1. Block on the CQ event channel waiting for a READY command - from the receiver to tell us that the receiver - is *ready* for us to transmit some new bytes. -2. Optionally: if we are expecting a response from the command - (that we have not yet transmitted), let's post an RQ - work request to receive that data a few moments later. -3. When the READY arrives, librdmacm will - unblock us and we immediately post a RQ work request - to replace the one we just used up. -4. Now, we can actually post the work request to SEND - the requested command type of the header we were asked for. -5. Optionally, if we are expecting a response (as before), - we block again and wait for that response using the additional - work request we previously posted. (This is used to carry - 'Register result' commands #6 back to the sender which - hold the rkey need to perform RDMA. Note that the virtual address - corresponding to this rkey was already exchanged at the beginning - of the connection (described below). - -All of the remaining command types (not including 'ready') -described above all use the aforementioned two functions to do the hard wo= rk: - -1. After connection setup, RAMBlock information is exchanged using - this protocol before the actual migration begins. This information incl= udes - a description of each RAMBlock on the server side as well as the virtua= l addresses - and lengths of each RAMBlock. This is used by the client to determine t= he - start and stop locations of chunks and how to register them dynamically - before performing the RDMA operations. -2. During runtime, once a 'chunk' becomes full of pages ready to - be sent with RDMA, the registration commands are used to ask the - other side to register the memory for this chunk and respond - with the result (rkey) of the registration. -3. Also, the QEMUFile interfaces also call these functions (described belo= w) - when transmitting non-live state, such as devices or to send - its own protocol information during the migration process. -4. Finally, zero pages are only checked if a page has not yet been registe= red - using chunk registration (or not checked at all and unconditionally - written if chunk registration is disabled. This is accomplished using - the "Compress" command listed above. If the page *has* been registered - then we check the entire chunk for zero. Only if the entire chunk is - zero, then we send a compress command to zap the page on the other side. - -Versioning and Capabilities -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D -Current version of the protocol is version #1. - -The same version applies to both for protocol traffic and capabilities -negotiation. (i.e. There is only one version number that is referred to -by all communication). - -librdmacm provides the user with a 'private data' area to be exchanged -at connection-setup time before any infiniband traffic is generated. - -Header: - * Version (protocol version validated before send/recv occurs), - uint32, network byte order - * Flags (bitwise OR of each capability), - uint32, network byte order - -There is no data portion of this header right now, so there is -no length field. The maximum size of the 'private data' section -is only 192 bytes per the Infiniband specification, so it's not -very useful for data anyway. This structure needs to remain small. - -This private data area is a convenient place to check for protocol -versioning because the user does not need to register memory to -transmit a few bytes of version information. - -This is also a convenient place to negotiate capabilities -(like dynamic page registration). - -If the version is invalid, we throw an error. - -If the version is new, we only negotiate the capabilities that the -requested version is able to perform and ignore the rest. - -Currently there is only one capability in Version #1: dynamic page registr= ation - -Finally: Negotiation happens with the Flags field: If the primary-VM -sets a flag, but the destination does not support this capability, it -will return a zero-bit for that flag and the primary-VM will understand -that as not being an available capability and will thus disable that -capability on the primary-VM side. - -QEMUFileRDMA Interface: -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - -QEMUFileRDMA introduces a couple of new functions: - -1. qemu_rdma_get_buffer() (QEMUFileOps rdma_read_ops) -2. qemu_rdma_put_buffer() (QEMUFileOps rdma_write_ops) - -These two functions are very short and simply use the protocol -describe above to deliver bytes without changing the upper-level -users of QEMUFile that depend on a bytestream abstraction. - -Finally, how do we handoff the actual bytes to get_buffer()? - -Again, because we're trying to "fake" a bytestream abstraction -using an analogy not unlike individual UDP frames, we have -to hold on to the bytes received from control-channel's SEND -messages in memory. - -Each time we receive a complete "QEMU File" control-channel -message, the bytes from SEND are copied into a small local holding area. - -Then, we return the number of bytes requested by get_buffer() -and leave the remaining bytes in the holding area until get_buffer() -comes around for another pass. - -If the buffer is empty, then we follow the same steps -listed above and issue another "QEMU File" protocol command, -asking for a new SEND message to re-fill the buffer. - -Migration of VM's ram: -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - -At the beginning of the migration, (migration-rdma.c), -the sender and the receiver populate the list of RAMBlocks -to be registered with each other into a structure. -Then, using the aforementioned protocol, they exchange a -description of these blocks with each other, to be used later -during the iteration of main memory. This description includes -a list of all the RAMBlocks, their offsets and lengths, virtual -addresses and possibly includes pre-registered RDMA keys in case dynamic -page registration was disabled on the server-side, otherwise not. - -Main memory is not migrated with the aforementioned protocol, -but is instead migrated with normal RDMA Write operations. - -Pages are migrated in "chunks" (hard-coded to 1 Megabyte right now). -Chunk size is not dynamic, but it could be in a future implementation. -There's nothing to indicate that this is useful right now. - -When a chunk is full (or a flush() occurs), the memory backed by -the chunk is registered with librdmacm is pinned in memory on -both sides using the aforementioned protocol. -After pinning, an RDMA Write is generated and transmitted -for the entire chunk. - -Chunks are also transmitted in batches: This means that we -do not request that the hardware signal the completion queue -for the completion of *every* chunk. The current batch size -is about 64 chunks (corresponding to 64 MB of memory). -Only the last chunk in a batch must be signaled. -This helps keep everything as asynchronous as possible -and helps keep the hardware busy performing RDMA operations. - -Error-handling: -=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D - -Infiniband has what is called a "Reliable, Connected" -link (one of 4 choices). This is the mode in which -we use for RDMA migration. - -If a *single* message fails, -the decision is to abort the migration entirely and -cleanup all the RDMA descriptors and unregister all -the memory. - -After cleanup, the Virtual Machine is returned to normal -operation the same way that would happen if the TCP -socket is broken during a non-RDMA based migration. - -TODO: -=3D=3D=3D=3D=3D -1. Currently, 'ulimit -l' mlock() limits as well as cgroups swap limits - are not compatible with infiniband memory pinning and will result in - an aborted migration (but with the source VM left unaffected). -2. Use of the recent /proc//pagemap would likely speed up - the use of KSM and ballooning while using RDMA. -3. Also, some form of balloon-device usage tracking would also - help alleviate some issues. -4. Use LRU to provide more fine-grained direction of UNREGISTER - requests for unpinning memory in an overcommitted environment. -5. Expose UNREGISTER support to the user by way of workload-specific - hints about application behavior. diff --git a/docs/system/loongarch/virt.rst b/docs/system/loongarch/virt.rst index 06d034b8ef..0a8e0766e4 100644 --- a/docs/system/loongarch/virt.rst +++ b/docs/system/loongarch/virt.rst @@ -39,7 +39,7 @@ can be accessed by following steps. =20 .. code-block:: bash =20 - ./configure --disable-rdma --prefix=3D/usr \ + ./configure --prefix=3D/usr \ --target-list=3D"loongarch64-softmmu" \ --disable-libiscsi --disable-libnfs --disable-libpmem \ --disable-glusterfs --enable-libusb --enable-usb-redir \ diff --git a/meson.build b/meson.build index d6af3cd53a..bd65abad13 100644 --- a/meson.build +++ b/meson.build @@ -1854,21 +1854,6 @@ if numa.found() and not cc.links(''' endif endif =20 -rdma =3D not_found -if not get_option('rdma').auto() or have_system - libumad =3D cc.find_library('ibumad', required: get_option('rdma')) - rdma_libs =3D [cc.find_library('rdmacm', has_headers: ['rdma/rdma_cma.h'= ], - required: get_option('rdma')), - cc.find_library('ibverbs', required: get_option('rdma')), - libumad] - rdma =3D declare_dependency(dependencies: rdma_libs) - foreach lib: rdma_libs - if not lib.found() - rdma =3D not_found - endif - endforeach -endif - cacard =3D not_found if not get_option('smartcard').auto() or have_system cacard =3D dependency('libcacard', required: get_option('smartcard'), @@ -2246,7 +2231,6 @@ endif config_host_data.set('CONFIG_OPENGL', opengl.found()) config_host_data.set('CONFIG_PLUGIN', get_option('plugins')) config_host_data.set('CONFIG_RBD', rbd.found()) -config_host_data.set('CONFIG_RDMA', rdma.found()) config_host_data.set('CONFIG_RELOCATABLE', get_option('relocatable')) config_host_data.set('CONFIG_SAFESTACK', get_option('safe_stack')) config_host_data.set('CONFIG_SDL', sdl.found()) @@ -2399,12 +2383,6 @@ if rbd.found() dependencies: rbd, prefix: '#include ')) endif -if rdma.found() - config_host_data.set('HAVE_IBV_ADVISE_MR', - cc.has_function('ibv_advise_mr', - dependencies: rdma, - prefix: '#include ')) -endif =20 have_asan_fiber =3D false if get_option('sanitizers') and \ @@ -4398,7 +4376,6 @@ summary_info +=3D {'Multipath support': mpathpersist} summary_info +=3D {'Linux AIO support': libaio} summary_info +=3D {'Linux io_uring support': linux_io_uring} summary_info +=3D {'ATTR/XATTR support': libattr} -summary_info +=3D {'RDMA support': rdma} summary_info +=3D {'fdt support': fdt_opt =3D=3D 'disabled' ? false = : fdt_opt} summary_info +=3D {'libcap-ng support': libcap_ng} summary_info +=3D {'bpf support': libbpf} diff --git a/qapi/migration.json b/qapi/migration.json index 8c65b90328..9a56d403be 100644 --- a/qapi/migration.json +++ b/qapi/migration.json @@ -221,8 +221,8 @@ # # @setup-time: amount of setup time in milliseconds *before* the # iterations begin but *after* the QMP command is issued. This is -# designed to provide an accounting of any activities (such as -# RDMA pinning) which may be expensive, but do not actually occur +# designed to provide an accounting of any activities which may be +# expensive, but do not actually occur # during the iterative migration rounds themselves. (since 1.6) # # @cpu-throttle-percentage: percentage of time guest cpus are being @@ -430,10 +430,6 @@ # for certain work loads, by sending compressed difference of the # pages # -# @rdma-pin-all: Controls whether or not the entire VM memory -# footprint is mlock()'d on demand or all at once. Refer to -# docs/rdma.txt for usage. Disabled by default. (since 2.0) -# # @zero-blocks: During storage migration encode blocks of zeroes # efficiently. This essentially saves 1MB of zeroes per block on # the wire. Enabling requires source and target VM to support @@ -547,7 +543,7 @@ # Since: 1.2 ## { 'enum': 'MigrationCapability', - 'data': ['xbzrle', 'rdma-pin-all', 'auto-converge', 'zero-blocks', + 'data': ['xbzrle', 'auto-converge', 'zero-blocks', { 'name': 'compress', 'features': [ 'deprecated' ] }, 'events', 'postcopy-ram', { 'name': 'x-colo', 'features': [ 'unstable' ] }, @@ -606,7 +602,6 @@ # -> { "execute": "query-migrate-capabilities" } # <- { "return": [ # {"state": false, "capability": "xbzrle"}, -# {"state": false, "capability": "rdma-pin-all"}, # {"state": false, "capability": "auto-converge"}, # {"state": false, "capability": "zero-blocks"}, # {"state": false, "capability": "compress"}, @@ -1654,14 +1649,12 @@ # # @exec: Direct the migration stream to another process. # -# @rdma: Migrate via RDMA. -# # @file: Direct the migration stream to a file. # # Since: 8.2 ## { 'enum': 'MigrationAddressType', - 'data': [ 'socket', 'exec', 'rdma', 'file' ] } + 'data': [ 'socket', 'exec', 'file' ] } =20 ## # @FileMigrationArgs: @@ -1701,7 +1694,6 @@ 'data': { 'socket': 'SocketAddress', 'exec': 'MigrationExecCommand', - 'rdma': 'InetSocketAddress', 'file': 'FileMigrationArgs' } } =20 ## @@ -1804,14 +1796,6 @@ # -> { "execute": "migrate", # "arguments": { # "channels": [ { "channel-type": "main", -# "addr": { "transport": "rdma", -# "host": "10.12.34.9", -# "port": "1050" } } ] } } -# <- { "return": {} } -# -# -> { "execute": "migrate", -# "arguments": { -# "channels": [ { "channel-type": "main", # "addr": { "transport": "file", # "filename": "/tmp/migfile", # "offset": "0x1000" } } ] } } @@ -1879,13 +1863,6 @@ # "/some/sock" ] } } ] } } # <- { "return": {} } # -# -> { "execute": "migrate-incoming", -# "arguments": { -# "channels": [ { "channel-type": "main", -# "addr": { "transport": "rdma", -# "host": "10.12.34.9", -# "port": "1050" } } ] } } -# <- { "return": {} } ## { 'command': 'migrate-incoming', 'data': {'*uri': 'str', diff --git a/migration/migration-stats.h b/migration/migration-stats.h index 05290ade76..817c53559a 100644 --- a/migration/migration-stats.h +++ b/migration/migration-stats.h @@ -93,10 +93,6 @@ typedef struct { * Maximum amount of data we can send in a cycle. */ Stat64 rate_limit_max; - /* - * Number of bytes sent through RDMA. - */ - Stat64 rdma_bytes; /* * Number of pages transferred that were full of zeros. */ @@ -133,7 +129,7 @@ void migration_rate_set(uint64_t new_rate); * * Returns how many bytes have we transferred since the beginning of * the migration. It accounts for bytes sent through any migration - * channel, multifd, qemu_file, rdma, .... + * channel, multifd, qemu_file, .... */ uint64_t migration_transferred_bytes(void); #endif diff --git a/migration/migration.h b/migration/migration.h index 8045e39c26..d097828580 100644 --- a/migration/migration.h +++ b/migration/migration.h @@ -162,13 +162,6 @@ struct MigrationIncomingState { =20 int state; =20 - /* - * The incoming migration coroutine, non-NULL during qemu_loadvm_state= (). - * Used to wake the migration incoming coroutine from rdma code. How m= uch is - * it safe - it's a question. - */ - Coroutine *loadvm_co; - /* The coroutine we should enter (back) after failover */ Coroutine *colo_incoming_co; QemuSemaphore colo_incoming_sem; @@ -463,8 +456,6 @@ struct MigrationState { * switchover has been received. */ bool switchover_acked; - /* Is this a rdma migration */ - bool rdma_migration; }; =20 void migrate_set_state(int *state, int old_state, int new_state); diff --git a/migration/options.h b/migration/options.h index ab8199e207..c00213973e 100644 --- a/migration/options.h +++ b/migration/options.h @@ -37,7 +37,6 @@ bool migrate_multifd(void); bool migrate_pause_before_switchover(void); bool migrate_postcopy_blocktime(void); bool migrate_postcopy_preempt(void); -bool migrate_rdma_pin_all(void); bool migrate_release_ram(void); bool migrate_return_path(void); bool migrate_validate_uuid(void); @@ -54,7 +53,6 @@ bool migrate_zero_copy_send(void); =20 bool migrate_multifd_flush_after_each_section(void); bool migrate_postcopy(void); -bool migrate_rdma(void); bool migrate_tls(void); =20 /* capabilities helpers */ diff --git a/migration/rdma.h b/migration/rdma.h deleted file mode 100644 index a8d27f33b8..0000000000 --- a/migration/rdma.h +++ /dev/null @@ -1,69 +0,0 @@ -/* - * RDMA protocol and interfaces - * - * Copyright IBM, Corp. 2010-2013 - * Copyright Red Hat, Inc. 2015-2016 - * - * Authors: - * Michael R. Hines - * Jiuxing Liu - * Daniel P. Berrange - * - * This work is licensed under the terms of the GNU GPL, version 2 or - * later. See the COPYING file in the top-level directory. - * - */ - -#include "qemu/sockets.h" - -#ifndef QEMU_MIGRATION_RDMA_H -#define QEMU_MIGRATION_RDMA_H - -#include "exec/memory.h" - -void rdma_start_outgoing_migration(void *opaque, InetSocketAddress *host_p= ort, - Error **errp); - -void rdma_start_incoming_migration(InetSocketAddress *host_port, Error **e= rrp); - -/* - * Constants used by rdma return codes - */ -#define RAM_CONTROL_SETUP 0 -#define RAM_CONTROL_ROUND 1 -#define RAM_CONTROL_FINISH 3 - -/* - * Whenever this is found in the data stream, the flags - * will be passed to rdma functions in the incoming-migration - * side. - */ -#define RAM_SAVE_FLAG_HOOK 0x80 - -#define RAM_SAVE_CONTROL_NOT_SUPP -1000 -#define RAM_SAVE_CONTROL_DELAYED -2000 - -#ifdef CONFIG_RDMA -int rdma_registration_handle(QEMUFile *f); -int rdma_registration_start(QEMUFile *f, uint64_t flags); -int rdma_registration_stop(QEMUFile *f, uint64_t flags); -int rdma_block_notification_handle(QEMUFile *f, const char *name); -int rdma_control_save_page(QEMUFile *f, ram_addr_t block_offset, - ram_addr_t offset, size_t size); -#else -static inline -int rdma_registration_handle(QEMUFile *f) { return 0; } -static inline -int rdma_registration_start(QEMUFile *f, uint64_t flags) { return 0; } -static inline -int rdma_registration_stop(QEMUFile *f, uint64_t flags) { return 0; } -static inline -int rdma_block_notification_handle(QEMUFile *f, const char *name) { return= 0; } -static inline -int rdma_control_save_page(QEMUFile *f, ram_addr_t block_offset, - ram_addr_t offset, size_t size) -{ - return RAM_SAVE_CONTROL_NOT_SUPP; -} -#endif -#endif diff --git a/migration/migration-stats.c b/migration/migration-stats.c index f690b98a03..9bc8d7018f 100644 --- a/migration/migration-stats.c +++ b/migration/migration-stats.c @@ -62,9 +62,8 @@ void migration_rate_reset(void) uint64_t migration_transferred_bytes(void) { uint64_t multifd =3D stat64_get(&mig_stats.multifd_bytes); - uint64_t rdma =3D stat64_get(&mig_stats.rdma_bytes); uint64_t qemu_file =3D stat64_get(&mig_stats.qemu_file_transferred); =20 - trace_migration_transferred_bytes(qemu_file, multifd, rdma); - return qemu_file + multifd + rdma; + trace_migration_transferred_bytes(qemu_file, multifd); + return qemu_file + multifd; } diff --git a/migration/migration.c b/migration/migration.c index 9fe8fd2afd..8e17914c8b 100644 --- a/migration/migration.c +++ b/migration/migration.c @@ -25,7 +25,6 @@ #include "sysemu/runstate.h" #include "sysemu/sysemu.h" #include "sysemu/cpu-throttle.h" -#include "rdma.h" #include "ram.h" #include "ram-compress.h" #include "migration/global_state.h" @@ -545,7 +544,6 @@ bool migrate_uri_parse(const char *uri, MigrationChanne= l **channel, { g_autoptr(MigrationChannel) val =3D g_new0(MigrationChannel, 1); g_autoptr(MigrationAddress) addr =3D g_new0(MigrationAddress, 1); - InetSocketAddress *isock =3D &addr->u.rdma; strList **tail =3D &addr->u.exec.args; =20 if (strstart(uri, "exec:", NULL)) { @@ -558,12 +556,6 @@ bool migrate_uri_parse(const char *uri, MigrationChann= el **channel, QAPI_LIST_APPEND(tail, g_strdup("-c")); #endif QAPI_LIST_APPEND(tail, g_strdup(uri + strlen("exec:"))); - } else if (strstart(uri, "rdma:", NULL)) { - if (inet_parse(isock, uri + strlen("rdma:"), errp)) { - qapi_free_InetSocketAddress(isock); - return false; - } - addr->transport =3D MIGRATION_ADDRESS_TYPE_RDMA; } else if (strstart(uri, "tcp:", NULL) || strstart(uri, "unix:", NULL) || strstart(uri, "vsock:", NULL) || @@ -645,22 +637,6 @@ static void qemu_start_incoming_migration(const char *= uri, bool has_channels, } else if (saddr->type =3D=3D SOCKET_ADDRESS_TYPE_FD) { fd_start_incoming_migration(saddr->u.fd.str, errp); } -#ifdef CONFIG_RDMA - } else if (addr->transport =3D=3D MIGRATION_ADDRESS_TYPE_RDMA) { - if (migrate_compress()) { - error_setg(errp, "RDMA and compression can't be used together"= ); - return; - } - if (migrate_xbzrle()) { - error_setg(errp, "RDMA and XBZRLE can't be used together"); - return; - } - if (migrate_multifd()) { - error_setg(errp, "RDMA and multifd can't be used together"); - return; - } - rdma_start_incoming_migration(&addr->u.rdma, errp); -#endif } else if (addr->transport =3D=3D MIGRATION_ADDRESS_TYPE_EXEC) { exec_start_incoming_migration(addr->u.exec.args, errp); } else if (addr->transport =3D=3D MIGRATION_ADDRESS_TYPE_FILE) { @@ -751,9 +727,7 @@ process_incoming_migration_co(void *opaque) migrate_set_state(&mis->state, MIGRATION_STATUS_SETUP, MIGRATION_STATUS_ACTIVE); =20 - mis->loadvm_co =3D qemu_coroutine_self(); ret =3D qemu_loadvm_state(mis->from_src_file); - mis->loadvm_co =3D NULL; =20 trace_vmstate_downtime_checkpoint("dst-precopy-loadvm-completed"); =20 @@ -1679,7 +1653,6 @@ int migrate_init(MigrationState *s, Error **errp) s->iteration_initial_bytes =3D 0; s->threshold_size =3D 0; s->switchover_acked =3D false; - s->rdma_migration =3D false; /* * set mig_stats memory to zero for a new migration */ @@ -2100,10 +2073,6 @@ void qmp_migrate(const char *uri, bool has_channels, } else if (saddr->type =3D=3D SOCKET_ADDRESS_TYPE_FD) { fd_start_outgoing_migration(s, saddr->u.fd.str, &local_err); } -#ifdef CONFIG_RDMA - } else if (addr->transport =3D=3D MIGRATION_ADDRESS_TYPE_RDMA) { - rdma_start_outgoing_migration(s, &addr->u.rdma, &local_err); -#endif } else if (addr->transport =3D=3D MIGRATION_ADDRESS_TYPE_EXEC) { exec_start_outgoing_migration(s, addr->u.exec.args, &local_err); } else if (addr->transport =3D=3D MIGRATION_ADDRESS_TYPE_FILE) { diff --git a/migration/options.c b/migration/options.c index bfd7753b69..02fc0b9ae8 100644 --- a/migration/options.c +++ b/migration/options.c @@ -185,7 +185,6 @@ Property migration_properties[] =3D { =20 /* Migration capabilities */ DEFINE_PROP_MIG_CAP("x-xbzrle", MIGRATION_CAPABILITY_XBZRLE), - DEFINE_PROP_MIG_CAP("x-rdma-pin-all", MIGRATION_CAPABILITY_RDMA_PIN_AL= L), DEFINE_PROP_MIG_CAP("x-auto-converge", MIGRATION_CAPABILITY_AUTO_CONVE= RGE), DEFINE_PROP_MIG_CAP("x-zero-blocks", MIGRATION_CAPABILITY_ZERO_BLOCKS), DEFINE_PROP_MIG_CAP("x-compress", MIGRATION_CAPABILITY_COMPRESS), @@ -323,13 +322,6 @@ bool migrate_postcopy_ram(void) return s->capabilities[MIGRATION_CAPABILITY_POSTCOPY_RAM]; } =20 -bool migrate_rdma_pin_all(void) -{ - MigrationState *s =3D migrate_get_current(); - - return s->capabilities[MIGRATION_CAPABILITY_RDMA_PIN_ALL]; -} - bool migrate_release_ram(void) { MigrationState *s =3D migrate_get_current(); @@ -393,13 +385,6 @@ bool migrate_postcopy(void) return migrate_postcopy_ram() || migrate_dirty_bitmaps(); } =20 -bool migrate_rdma(void) -{ - MigrationState *s =3D migrate_get_current(); - - return s->rdma_migration; -} - bool migrate_tls(void) { MigrationState *s =3D migrate_get_current(); @@ -458,7 +443,6 @@ INITIALIZE_MIGRATE_CAPS_SET(check_caps_background_snaps= hot, MIGRATION_CAPABILITY_PAUSE_BEFORE_SWITCHOVER, MIGRATION_CAPABILITY_AUTO_CONVERGE, MIGRATION_CAPABILITY_RELEASE_RAM, - MIGRATION_CAPABILITY_RDMA_PIN_ALL, MIGRATION_CAPABILITY_COMPRESS, MIGRATION_CAPABILITY_XBZRLE, MIGRATION_CAPABILITY_X_COLO, diff --git a/migration/qemu-file.c b/migration/qemu-file.c index a10882d47f..ad2efb332e 100644 --- a/migration/qemu-file.c +++ b/migration/qemu-file.c @@ -32,7 +32,6 @@ #include "trace.h" #include "options.h" #include "qapi/error.h" -#include "rdma.h" #include "io/channel-file.h" =20 #define IO_BUF_SIZE 32768 diff --git a/migration/ram.c b/migration/ram.c index 8deb84984f..c81c8a7cff 100644 --- a/migration/ram.c +++ b/migration/ram.c @@ -59,7 +59,6 @@ #include "qemu/iov.h" #include "multifd.h" #include "sysemu/runstate.h" -#include "rdma.h" #include "options.h" #include "sysemu/dirtylimit.h" #include "sysemu/kvm.h" @@ -89,7 +88,7 @@ #define RAM_SAVE_FLAG_EOS 0x10 #define RAM_SAVE_FLAG_CONTINUE 0x20 #define RAM_SAVE_FLAG_XBZRLE 0x40 -/* 0x80 is reserved in rdma.h for RAM_SAVE_FLAG_HOOK */ +#define RAM_SAVE_FLAG_HOOK 0x80 /* was reserved by RDMA */ #define RAM_SAVE_FLAG_COMPRESS_PAGE 0x100 #define RAM_SAVE_FLAG_MULTIFD_FLUSH 0x200 /* We can't use any flag that is bigger than 0x200 */ @@ -1175,32 +1174,6 @@ static int save_zero_page(RAMState *rs, PageSearchSt= atus *pss, return len; } =20 -/* - * @pages: the number of pages written by the control path, - * < 0 - error - * > 0 - number of pages written - * - * Return true if the pages has been saved, otherwise false is returned. - */ -static bool control_save_page(PageSearchStatus *pss, - ram_addr_t offset, int *pages) -{ - int ret; - - ret =3D rdma_control_save_page(pss->pss_channel, pss->block->offset, o= ffset, - TARGET_PAGE_SIZE); - if (ret =3D=3D RAM_SAVE_CONTROL_NOT_SUPP) { - return false; - } - - if (ret =3D=3D RAM_SAVE_CONTROL_DELAYED) { - *pages =3D 1; - return true; - } - *pages =3D ret; - return true; -} - /* * directly send the page to the stream * @@ -2080,11 +2053,6 @@ static bool save_compress_page(RAMState *rs, PageSea= rchStatus *pss, static int ram_save_target_page_legacy(RAMState *rs, PageSearchStatus *pss) { ram_addr_t offset =3D ((ram_addr_t)pss->page) << TARGET_PAGE_BITS; - int res; - - if (control_save_page(pss, offset, &res)) { - return res; - } =20 if (save_compress_page(rs, pss, offset)) { return 1; @@ -3114,18 +3082,6 @@ static int ram_save_setup(QEMUFile *f, void *opaque) } } =20 - ret =3D rdma_registration_start(f, RAM_CONTROL_SETUP); - if (ret < 0) { - qemu_file_set_error(f, ret); - return ret; - } - - ret =3D rdma_registration_stop(f, RAM_CONTROL_SETUP); - if (ret < 0) { - qemu_file_set_error(f, ret); - return ret; - } - migration_ops =3D g_malloc0(sizeof(MigrationOps)); =20 if (migrate_multifd()) { @@ -3221,12 +3177,6 @@ static int ram_save_iterate(QEMUFile *f, void *opaqu= e) /* Read version before ram_list.blocks */ smp_rmb(); =20 - ret =3D rdma_registration_start(f, RAM_CONTROL_ROUND); - if (ret < 0) { - qemu_file_set_error(f, ret); - goto out; - } - t0 =3D qemu_clock_get_ns(QEMU_CLOCK_REALTIME); i =3D 0; while ((ret =3D migration_rate_exceeded(f)) =3D=3D 0 || @@ -3278,15 +3228,6 @@ static int ram_save_iterate(QEMUFile *f, void *opaqu= e) } } =20 - /* - * Must occur before EOS (or any QEMUFile operation) - * because of RDMA protocol. - */ - ret =3D rdma_registration_stop(f, RAM_CONTROL_ROUND); - if (ret < 0) { - qemu_file_set_error(f, ret); - } - out: if (ret >=3D 0 && migration_is_setup_or_active()) { @@ -3332,12 +3273,6 @@ static int ram_save_complete(QEMUFile *f, void *opaq= ue) migration_bitmap_sync_precopy(rs, true); } =20 - ret =3D rdma_registration_start(f, RAM_CONTROL_FINISH); - if (ret < 0) { - qemu_file_set_error(f, ret); - return ret; - } - /* try transferring iterative blocks of memory */ =20 /* flush all remaining blocks regardless of rate limiting */ @@ -3358,12 +3293,6 @@ static int ram_save_complete(QEMUFile *f, void *opaq= ue) qemu_mutex_unlock(&rs->bitmap_mutex); =20 compress_flush_data(); - - ret =3D rdma_registration_stop(f, RAM_CONTROL_FINISH); - if (ret < 0) { - qemu_file_set_error(f, ret); - return ret; - } } =20 ret =3D multifd_send_sync_main(); @@ -3576,8 +3505,7 @@ static inline void *colo_cache_from_block_offset(RAMB= lock *block, /** * ram_handle_zero: handle the zero page case * - * If a page (or a whole RDMA chunk) has been - * determined to be zero, then zap it. + * If a page has been determined to be zero, then zap it. * * @host: host address for the zero page * @ch: what the page is filled from. We only support zero @@ -4161,10 +4089,6 @@ static int parse_ramblock(QEMUFile *f, RAMBlock *blo= ck, ram_addr_t length) return -EINVAL; } } - ret =3D rdma_block_notification_handle(f, block->idstr); - if (ret < 0) { - qemu_file_set_error(f, ret); - } =20 return ret; } @@ -4363,12 +4287,6 @@ static int ram_load_precopy(QEMUFile *f) multifd_recv_sync_main(); } break; - case RAM_SAVE_FLAG_HOOK: - ret =3D rdma_registration_handle(f); - if (ret < 0) { - qemu_file_set_error(f, ret); - } - break; default: error_report("Unknown combination of migration flags: 0x%x", f= lags); ret =3D -EINVAL; diff --git a/migration/rdma.c b/migration/rdma.c deleted file mode 100644 index 855753c671..0000000000 --- a/migration/rdma.c +++ /dev/null @@ -1,4184 +0,0 @@ -/* - * RDMA protocol and interfaces - * - * Copyright IBM, Corp. 2010-2013 - * Copyright Red Hat, Inc. 2015-2016 - * - * Authors: - * Michael R. Hines - * Jiuxing Liu - * Daniel P. Berrange - * - * This work is licensed under the terms of the GNU GPL, version 2 or - * later. See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include "qapi/error.h" -#include "qemu/cutils.h" -#include "exec/target_page.h" -#include "rdma.h" -#include "migration.h" -#include "migration-stats.h" -#include "qemu-file.h" -#include "ram.h" -#include "qemu/error-report.h" -#include "qemu/main-loop.h" -#include "qemu/module.h" -#include "qemu/rcu.h" -#include "qemu/sockets.h" -#include "qemu/bitmap.h" -#include "qemu/coroutine.h" -#include "exec/memory.h" -#include -#include -#include -#include -#include "trace.h" -#include "qom/object.h" -#include "options.h" -#include - -#define RDMA_RESOLVE_TIMEOUT_MS 10000 - -/* Do not merge data if larger than this. */ -#define RDMA_MERGE_MAX (2 * 1024 * 1024) -#define RDMA_SIGNALED_SEND_MAX (RDMA_MERGE_MAX / 4096) - -#define RDMA_REG_CHUNK_SHIFT 20 /* 1 MB */ - -/* - * This is only for non-live state being migrated. - * Instead of RDMA_WRITE messages, we use RDMA_SEND - * messages for that state, which requires a different - * delivery design than main memory. - */ -#define RDMA_SEND_INCREMENT 32768 - -/* - * Maximum size infiniband SEND message - */ -#define RDMA_CONTROL_MAX_BUFFER (512 * 1024) -#define RDMA_CONTROL_MAX_COMMANDS_PER_MESSAGE 4096 - -#define RDMA_CONTROL_VERSION_CURRENT 1 -/* - * Capabilities for negotiation. - */ -#define RDMA_CAPABILITY_PIN_ALL 0x01 - -/* - * Add the other flags above to this list of known capabilities - * as they are introduced. - */ -static uint32_t known_capabilities =3D RDMA_CAPABILITY_PIN_ALL; - -/* - * A work request ID is 64-bits and we split up these bits - * into 3 parts: - * - * bits 0-15 : type of control message, 2^16 - * bits 16-29: ram block index, 2^14 - * bits 30-63: ram block chunk number, 2^34 - * - * The last two bit ranges are only used for RDMA writes, - * in order to track their completion and potentially - * also track unregistration status of the message. - */ -#define RDMA_WRID_TYPE_SHIFT 0UL -#define RDMA_WRID_BLOCK_SHIFT 16UL -#define RDMA_WRID_CHUNK_SHIFT 30UL - -#define RDMA_WRID_TYPE_MASK \ - ((1UL << RDMA_WRID_BLOCK_SHIFT) - 1UL) - -#define RDMA_WRID_BLOCK_MASK \ - (~RDMA_WRID_TYPE_MASK & ((1UL << RDMA_WRID_CHUNK_SHIFT) - 1UL)) - -#define RDMA_WRID_CHUNK_MASK (~RDMA_WRID_BLOCK_MASK & ~RDMA_WRID_TYPE_MASK) - -/* - * RDMA migration protocol: - * 1. RDMA Writes (data messages, i.e. RAM) - * 2. IB Send/Recv (control channel messages) - */ -enum { - RDMA_WRID_NONE =3D 0, - RDMA_WRID_RDMA_WRITE =3D 1, - RDMA_WRID_SEND_CONTROL =3D 2000, - RDMA_WRID_RECV_CONTROL =3D 4000, -}; - -/* - * Work request IDs for IB SEND messages only (not RDMA writes). - * This is used by the migration protocol to transmit - * control messages (such as device state and registration commands) - * - * We could use more WRs, but we have enough for now. - */ -enum { - RDMA_WRID_READY =3D 0, - RDMA_WRID_DATA, - RDMA_WRID_CONTROL, - RDMA_WRID_MAX, -}; - -/* - * SEND/RECV IB Control Messages. - */ -enum { - RDMA_CONTROL_NONE =3D 0, - RDMA_CONTROL_ERROR, - RDMA_CONTROL_READY, /* ready to receive */ - RDMA_CONTROL_QEMU_FILE, /* QEMUFile-transmitted bytes */ - RDMA_CONTROL_RAM_BLOCKS_REQUEST, /* RAMBlock synchronization */ - RDMA_CONTROL_RAM_BLOCKS_RESULT, /* RAMBlock synchronization */ - RDMA_CONTROL_COMPRESS, /* page contains repeat values */ - RDMA_CONTROL_REGISTER_REQUEST, /* dynamic page registration */ - RDMA_CONTROL_REGISTER_RESULT, /* key to use after registration */ - RDMA_CONTROL_REGISTER_FINISHED, /* current iteration finished */ - RDMA_CONTROL_UNREGISTER_REQUEST, /* dynamic UN-registration */ - RDMA_CONTROL_UNREGISTER_FINISHED, /* unpinning finished */ -}; - - -/* - * Memory and MR structures used to represent an IB Send/Recv work request. - * This is *not* used for RDMA writes, only IB Send/Recv. - */ -typedef struct { - uint8_t control[RDMA_CONTROL_MAX_BUFFER]; /* actual buffer to registe= r */ - struct ibv_mr *control_mr; /* registration metadata */ - size_t control_len; /* length of the message */ - uint8_t *control_curr; /* start of unconsumed byte= s */ -} RDMAWorkRequestData; - -/* - * Negotiate RDMA capabilities during connection-setup time. - */ -typedef struct { - uint32_t version; - uint32_t flags; -} RDMACapabilities; - -static void caps_to_network(RDMACapabilities *cap) -{ - cap->version =3D htonl(cap->version); - cap->flags =3D htonl(cap->flags); -} - -static void network_to_caps(RDMACapabilities *cap) -{ - cap->version =3D ntohl(cap->version); - cap->flags =3D ntohl(cap->flags); -} - -/* - * Representation of a RAMBlock from an RDMA perspective. - * This is not transmitted, only local. - * This and subsequent structures cannot be linked lists - * because we're using a single IB message to transmit - * the information. It's small anyway, so a list is overkill. - */ -typedef struct RDMALocalBlock { - char *block_name; - uint8_t *local_host_addr; /* local virtual address */ - uint64_t remote_host_addr; /* remote virtual address */ - uint64_t offset; - uint64_t length; - struct ibv_mr **pmr; /* MRs for chunk-level registration */ - struct ibv_mr *mr; /* MR for non-chunk-level registration= */ - uint32_t *remote_keys; /* rkeys for chunk-level registration = */ - uint32_t remote_rkey; /* rkeys for non-chunk-level registrat= ion */ - int index; /* which block are we */ - unsigned int src_index; /* (Only used on dest) */ - bool is_ram_block; - int nb_chunks; - unsigned long *transit_bitmap; - unsigned long *unregister_bitmap; -} RDMALocalBlock; - -/* - * Also represents a RAMblock, but only on the dest. - * This gets transmitted by the dest during connection-time - * to the source VM and then is used to populate the - * corresponding RDMALocalBlock with - * the information needed to perform the actual RDMA. - */ -typedef struct QEMU_PACKED RDMADestBlock { - uint64_t remote_host_addr; - uint64_t offset; - uint64_t length; - uint32_t remote_rkey; - uint32_t padding; -} RDMADestBlock; - -static const char *control_desc(unsigned int rdma_control) -{ - static const char *strs[] =3D { - [RDMA_CONTROL_NONE] =3D "NONE", - [RDMA_CONTROL_ERROR] =3D "ERROR", - [RDMA_CONTROL_READY] =3D "READY", - [RDMA_CONTROL_QEMU_FILE] =3D "QEMU FILE", - [RDMA_CONTROL_RAM_BLOCKS_REQUEST] =3D "RAM BLOCKS REQUEST", - [RDMA_CONTROL_RAM_BLOCKS_RESULT] =3D "RAM BLOCKS RESULT", - [RDMA_CONTROL_COMPRESS] =3D "COMPRESS", - [RDMA_CONTROL_REGISTER_REQUEST] =3D "REGISTER REQUEST", - [RDMA_CONTROL_REGISTER_RESULT] =3D "REGISTER RESULT", - [RDMA_CONTROL_REGISTER_FINISHED] =3D "REGISTER FINISHED", - [RDMA_CONTROL_UNREGISTER_REQUEST] =3D "UNREGISTER REQUEST", - [RDMA_CONTROL_UNREGISTER_FINISHED] =3D "UNREGISTER FINISHED", - }; - - if (rdma_control > RDMA_CONTROL_UNREGISTER_FINISHED) { - return "??BAD CONTROL VALUE??"; - } - - return strs[rdma_control]; -} - -#if !defined(htonll) -static uint64_t htonll(uint64_t v) -{ - union { uint32_t lv[2]; uint64_t llv; } u; - u.lv[0] =3D htonl(v >> 32); - u.lv[1] =3D htonl(v & 0xFFFFFFFFULL); - return u.llv; -} -#endif - -#if !defined(ntohll) -static uint64_t ntohll(uint64_t v) -{ - union { uint32_t lv[2]; uint64_t llv; } u; - u.llv =3D v; - return ((uint64_t)ntohl(u.lv[0]) << 32) | (uint64_t) ntohl(u.lv[1]); -} -#endif - -static void dest_block_to_network(RDMADestBlock *db) -{ - db->remote_host_addr =3D htonll(db->remote_host_addr); - db->offset =3D htonll(db->offset); - db->length =3D htonll(db->length); - db->remote_rkey =3D htonl(db->remote_rkey); -} - -static void network_to_dest_block(RDMADestBlock *db) -{ - db->remote_host_addr =3D ntohll(db->remote_host_addr); - db->offset =3D ntohll(db->offset); - db->length =3D ntohll(db->length); - db->remote_rkey =3D ntohl(db->remote_rkey); -} - -/* - * Virtual address of the above structures used for transmitting - * the RAMBlock descriptions at connection-time. - * This structure is *not* transmitted. - */ -typedef struct RDMALocalBlocks { - int nb_blocks; - bool init; /* main memory init complete */ - RDMALocalBlock *block; -} RDMALocalBlocks; - -/* - * Main data structure for RDMA state. - * While there is only one copy of this structure being allocated right no= w, - * this is the place where one would start if you wanted to consider - * having more than one RDMA connection open at the same time. - */ -typedef struct RDMAContext { - char *host; - int port; - - RDMAWorkRequestData wr_data[RDMA_WRID_MAX]; - - /* - * This is used by *_exchange_send() to figure out whether or not - * the initial "READY" message has already been received or not. - * This is because other functions may potentially poll() and detect - * the READY message before send() does, in which case we need to - * know if it completed. - */ - int control_ready_expected; - - /* number of outstanding writes */ - int nb_sent; - - /* store info about current buffer so that we can - merge it with future sends */ - uint64_t current_addr; - uint64_t current_length; - /* index of ram block the current buffer belongs to */ - int current_index; - /* index of the chunk in the current ram block */ - int current_chunk; - - bool pin_all; - - /* - * infiniband-specific variables for opening the device - * and maintaining connection state and so forth. - * - * cm_id also has ibv_context, rdma_event_channel, and ibv_qp in - * cm_id->verbs, cm_id->channel, and cm_id->qp. - */ - struct rdma_cm_id *cm_id; /* connection manager ID */ - struct rdma_cm_id *listen_id; - bool connected; - - struct ibv_context *verbs; - struct rdma_event_channel *channel; - struct ibv_qp *qp; /* queue pair */ - struct ibv_comp_channel *recv_comp_channel; /* recv completion channe= l */ - struct ibv_comp_channel *send_comp_channel; /* send completion channe= l */ - struct ibv_pd *pd; /* protection domain */ - struct ibv_cq *recv_cq; /* recvieve completion queue */ - struct ibv_cq *send_cq; /* send completion queue */ - - /* - * If a previous write failed (perhaps because of a failed - * memory registration, then do not attempt any future work - * and remember the error state. - */ - bool errored; - bool error_reported; - bool received_error; - - /* - * Description of ram blocks used throughout the code. - */ - RDMALocalBlocks local_ram_blocks; - RDMADestBlock *dest_blocks; - - /* Index of the next RAMBlock received during block registration */ - unsigned int next_src_index; - - /* - * Migration on *destination* started. - * Then use coroutine yield function. - * Source runs in a thread, so we don't care. - */ - int migration_started_on_destination; - - int total_registrations; - int total_writes; - - int unregister_current, unregister_next; - uint64_t unregistrations[RDMA_SIGNALED_SEND_MAX]; - - GHashTable *blockmap; - - /* the RDMAContext for return path */ - struct RDMAContext *return_path; - bool is_return_path; -} RDMAContext; - -#define TYPE_QIO_CHANNEL_RDMA "qio-channel-rdma" -OBJECT_DECLARE_SIMPLE_TYPE(QIOChannelRDMA, QIO_CHANNEL_RDMA) - - - -struct QIOChannelRDMA { - QIOChannel parent; - RDMAContext *rdmain; - RDMAContext *rdmaout; - QEMUFile *file; - bool blocking; /* XXX we don't actually honour this yet */ -}; - -/* - * Main structure for IB Send/Recv control messages. - * This gets prepended at the beginning of every Send/Recv. - */ -typedef struct QEMU_PACKED { - uint32_t len; /* Total length of data portion */ - uint32_t type; /* which control command to perform */ - uint32_t repeat; /* number of commands in data portion of same type */ - uint32_t padding; -} RDMAControlHeader; - -static void control_to_network(RDMAControlHeader *control) -{ - control->type =3D htonl(control->type); - control->len =3D htonl(control->len); - control->repeat =3D htonl(control->repeat); -} - -static void network_to_control(RDMAControlHeader *control) -{ - control->type =3D ntohl(control->type); - control->len =3D ntohl(control->len); - control->repeat =3D ntohl(control->repeat); -} - -/* - * Register a single Chunk. - * Information sent by the source VM to inform the dest - * to register an single chunk of memory before we can perform - * the actual RDMA operation. - */ -typedef struct QEMU_PACKED { - union QEMU_PACKED { - uint64_t current_addr; /* offset into the ram_addr_t space */ - uint64_t chunk; /* chunk to lookup if unregistering */ - } key; - uint32_t current_index; /* which ramblock the chunk belongs to */ - uint32_t padding; - uint64_t chunks; /* how many sequential chunks to register = */ -} RDMARegister; - -static bool rdma_errored(RDMAContext *rdma) -{ - if (rdma->errored && !rdma->error_reported) { - error_report("RDMA is in an error state waiting migration" - " to abort!"); - rdma->error_reported =3D true; - } - return rdma->errored; -} - -static void register_to_network(RDMAContext *rdma, RDMARegister *reg) -{ - RDMALocalBlock *local_block; - local_block =3D &rdma->local_ram_blocks.block[reg->current_index]; - - if (local_block->is_ram_block) { - /* - * current_addr as passed in is an address in the local ram_addr_t - * space, we need to translate this for the destination - */ - reg->key.current_addr -=3D local_block->offset; - reg->key.current_addr +=3D rdma->dest_blocks[reg->current_index].o= ffset; - } - reg->key.current_addr =3D htonll(reg->key.current_addr); - reg->current_index =3D htonl(reg->current_index); - reg->chunks =3D htonll(reg->chunks); -} - -static void network_to_register(RDMARegister *reg) -{ - reg->key.current_addr =3D ntohll(reg->key.current_addr); - reg->current_index =3D ntohl(reg->current_index); - reg->chunks =3D ntohll(reg->chunks); -} - -typedef struct QEMU_PACKED { - uint32_t value; /* if zero, we will madvise() */ - uint32_t block_idx; /* which ram block index */ - uint64_t offset; /* Address in remote ram_addr_t space */ - uint64_t length; /* length of the chunk */ -} RDMACompress; - -static void compress_to_network(RDMAContext *rdma, RDMACompress *comp) -{ - comp->value =3D htonl(comp->value); - /* - * comp->offset as passed in is an address in the local ram_addr_t - * space, we need to translate this for the destination - */ - comp->offset -=3D rdma->local_ram_blocks.block[comp->block_idx].offset; - comp->offset +=3D rdma->dest_blocks[comp->block_idx].offset; - comp->block_idx =3D htonl(comp->block_idx); - comp->offset =3D htonll(comp->offset); - comp->length =3D htonll(comp->length); -} - -static void network_to_compress(RDMACompress *comp) -{ - comp->value =3D ntohl(comp->value); - comp->block_idx =3D ntohl(comp->block_idx); - comp->offset =3D ntohll(comp->offset); - comp->length =3D ntohll(comp->length); -} - -/* - * The result of the dest's memory registration produces an "rkey" - * which the source VM must reference in order to perform - * the RDMA operation. - */ -typedef struct QEMU_PACKED { - uint32_t rkey; - uint32_t padding; - uint64_t host_addr; -} RDMARegisterResult; - -static void result_to_network(RDMARegisterResult *result) -{ - result->rkey =3D htonl(result->rkey); - result->host_addr =3D htonll(result->host_addr); -}; - -static void network_to_result(RDMARegisterResult *result) -{ - result->rkey =3D ntohl(result->rkey); - result->host_addr =3D ntohll(result->host_addr); -}; - -static int qemu_rdma_exchange_send(RDMAContext *rdma, RDMAControlHeader *h= ead, - uint8_t *data, RDMAControlHeader *resp, - int *resp_idx, - int (*callback)(RDMAContext *rdma, - Error **errp), - Error **errp); - -static inline uint64_t ram_chunk_index(const uint8_t *start, - const uint8_t *host) -{ - return ((uintptr_t) host - (uintptr_t) start) >> RDMA_REG_CHUNK_SHIFT; -} - -static inline uint8_t *ram_chunk_start(const RDMALocalBlock *rdma_ram_bloc= k, - uint64_t i) -{ - return (uint8_t *)(uintptr_t)(rdma_ram_block->local_host_addr + - (i << RDMA_REG_CHUNK_SHIFT)); -} - -static inline uint8_t *ram_chunk_end(const RDMALocalBlock *rdma_ram_block, - uint64_t i) -{ - uint8_t *result =3D ram_chunk_start(rdma_ram_block, i) + - (1UL << RDMA_REG_CHUNK_SHIFT); - - if (result > (rdma_ram_block->local_host_addr + rdma_ram_block->length= )) { - result =3D rdma_ram_block->local_host_addr + rdma_ram_block->lengt= h; - } - - return result; -} - -static void rdma_add_block(RDMAContext *rdma, const char *block_name, - void *host_addr, - ram_addr_t block_offset, uint64_t length) -{ - RDMALocalBlocks *local =3D &rdma->local_ram_blocks; - RDMALocalBlock *block; - RDMALocalBlock *old =3D local->block; - - local->block =3D g_new0(RDMALocalBlock, local->nb_blocks + 1); - - if (local->nb_blocks) { - if (rdma->blockmap) { - for (int x =3D 0; x < local->nb_blocks; x++) { - g_hash_table_remove(rdma->blockmap, - (void *)(uintptr_t)old[x].offset); - g_hash_table_insert(rdma->blockmap, - (void *)(uintptr_t)old[x].offset, - &local->block[x]); - } - } - memcpy(local->block, old, sizeof(RDMALocalBlock) * local->nb_block= s); - g_free(old); - } - - block =3D &local->block[local->nb_blocks]; - - block->block_name =3D g_strdup(block_name); - block->local_host_addr =3D host_addr; - block->offset =3D block_offset; - block->length =3D length; - block->index =3D local->nb_blocks; - block->src_index =3D ~0U; /* Filled in by the receipt of the block lis= t */ - block->nb_chunks =3D ram_chunk_index(host_addr, host_addr + length) + = 1UL; - block->transit_bitmap =3D bitmap_new(block->nb_chunks); - bitmap_clear(block->transit_bitmap, 0, block->nb_chunks); - block->unregister_bitmap =3D bitmap_new(block->nb_chunks); - bitmap_clear(block->unregister_bitmap, 0, block->nb_chunks); - block->remote_keys =3D g_new0(uint32_t, block->nb_chunks); - - block->is_ram_block =3D local->init ? false : true; - - if (rdma->blockmap) { - g_hash_table_insert(rdma->blockmap, (void *)(uintptr_t)block_offse= t, block); - } - - trace_rdma_add_block(block_name, local->nb_blocks, - (uintptr_t) block->local_host_addr, - block->offset, block->length, - (uintptr_t) (block->local_host_addr + block->leng= th), - BITS_TO_LONGS(block->nb_chunks) * - sizeof(unsigned long) * 8, - block->nb_chunks); - - local->nb_blocks++; -} - -/* - * Memory regions need to be registered with the device and queue pairs se= tup - * in advanced before the migration starts. This tells us where the RAM bl= ocks - * are so that we can register them individually. - */ -static int qemu_rdma_init_one_block(RAMBlock *rb, void *opaque) -{ - const char *block_name =3D qemu_ram_get_idstr(rb); - void *host_addr =3D qemu_ram_get_host_addr(rb); - ram_addr_t block_offset =3D qemu_ram_get_offset(rb); - ram_addr_t length =3D qemu_ram_get_used_length(rb); - rdma_add_block(opaque, block_name, host_addr, block_offset, length); - return 0; -} - -/* - * Identify the RAMBlocks and their quantity. They will be references to - * identify chunk boundaries inside each RAMBlock and also be referenced - * during dynamic page registration. - */ -static void qemu_rdma_init_ram_blocks(RDMAContext *rdma) -{ - RDMALocalBlocks *local =3D &rdma->local_ram_blocks; - int ret; - - assert(rdma->blockmap =3D=3D NULL); - memset(local, 0, sizeof *local); - ret =3D foreach_not_ignored_block(qemu_rdma_init_one_block, rdma); - assert(!ret); - trace_qemu_rdma_init_ram_blocks(local->nb_blocks); - rdma->dest_blocks =3D g_new0(RDMADestBlock, - rdma->local_ram_blocks.nb_blocks); - local->init =3D true; -} - -/* - * Note: If used outside of cleanup, the caller must ensure that the desti= nation - * block structures are also updated - */ -static void rdma_delete_block(RDMAContext *rdma, RDMALocalBlock *block) -{ - RDMALocalBlocks *local =3D &rdma->local_ram_blocks; - RDMALocalBlock *old =3D local->block; - - if (rdma->blockmap) { - g_hash_table_remove(rdma->blockmap, (void *)(uintptr_t)block->offs= et); - } - if (block->pmr) { - for (int j =3D 0; j < block->nb_chunks; j++) { - if (!block->pmr[j]) { - continue; - } - ibv_dereg_mr(block->pmr[j]); - rdma->total_registrations--; - } - g_free(block->pmr); - block->pmr =3D NULL; - } - - if (block->mr) { - ibv_dereg_mr(block->mr); - rdma->total_registrations--; - block->mr =3D NULL; - } - - g_free(block->transit_bitmap); - block->transit_bitmap =3D NULL; - - g_free(block->unregister_bitmap); - block->unregister_bitmap =3D NULL; - - g_free(block->remote_keys); - block->remote_keys =3D NULL; - - g_free(block->block_name); - block->block_name =3D NULL; - - if (rdma->blockmap) { - for (int x =3D 0; x < local->nb_blocks; x++) { - g_hash_table_remove(rdma->blockmap, - (void *)(uintptr_t)old[x].offset); - } - } - - if (local->nb_blocks > 1) { - - local->block =3D g_new0(RDMALocalBlock, local->nb_blocks - 1); - - if (block->index) { - memcpy(local->block, old, sizeof(RDMALocalBlock) * block->inde= x); - } - - if (block->index < (local->nb_blocks - 1)) { - memcpy(local->block + block->index, old + (block->index + 1), - sizeof(RDMALocalBlock) * - (local->nb_blocks - (block->index + 1))); - for (int x =3D block->index; x < local->nb_blocks - 1; x++) { - local->block[x].index--; - } - } - } else { - assert(block =3D=3D local->block); - local->block =3D NULL; - } - - trace_rdma_delete_block(block, (uintptr_t)block->local_host_addr, - block->offset, block->length, - (uintptr_t)(block->local_host_addr + block->le= ngth), - BITS_TO_LONGS(block->nb_chunks) * - sizeof(unsigned long) * 8, block->nb_chunks= ); - - g_free(old); - - local->nb_blocks--; - - if (local->nb_blocks && rdma->blockmap) { - for (int x =3D 0; x < local->nb_blocks; x++) { - g_hash_table_insert(rdma->blockmap, - (void *)(uintptr_t)local->block[x].offset, - &local->block[x]); - } - } -} - -/* - * Trace RDMA device open, with device details. - */ -static void qemu_rdma_dump_id(const char *who, struct ibv_context *verbs) -{ - struct ibv_port_attr port; - - if (ibv_query_port(verbs, 1, &port)) { - trace_qemu_rdma_dump_id_failed(who); - return; - } - - trace_qemu_rdma_dump_id(who, - verbs->device->name, - verbs->device->dev_name, - verbs->device->dev_path, - verbs->device->ibdev_path, - port.link_layer, - port.link_layer =3D=3D IBV_LINK_LAYER_INFINIBAND ? "Infini= band" - : port.link_layer =3D=3D IBV_LINK_LAYER_ETHERNET ? "Ethern= et" - : "Unknown"); -} - -/* - * Trace RDMA gid addressing information. - * Useful for understanding the RDMA device hierarchy in the kernel. - */ -static void qemu_rdma_dump_gid(const char *who, struct rdma_cm_id *id) -{ - char sgid[33]; - char dgid[33]; - inet_ntop(AF_INET6, &id->route.addr.addr.ibaddr.sgid, sgid, sizeof sgi= d); - inet_ntop(AF_INET6, &id->route.addr.addr.ibaddr.dgid, dgid, sizeof dgi= d); - trace_qemu_rdma_dump_gid(who, sgid, dgid); -} - -/* - * As of now, IPv6 over RoCE / iWARP is not supported by linux. - * We will try the next addrinfo struct, and fail if there are - * no other valid addresses to bind against. - * - * If user is listening on '[::]', then we will not have a opened a device - * yet and have no way of verifying if the device is RoCE or not. - * - * In this case, the source VM will throw an error for ALL types of - * connections (both IPv4 and IPv6) if the destination machine does not ha= ve - * a regular infiniband network available for use. - * - * The only way to guarantee that an error is thrown for broken kernels is - * for the management software to choose a *specific* interface at bind ti= me - * and validate what time of hardware it is. - * - * Unfortunately, this puts the user in a fix: - * - * If the source VM connects with an IPv4 address without knowing that the - * destination has bound to '[::]' the migration will unconditionally fail - * unless the management software is explicitly listening on the IPv4 - * address while using a RoCE-based device. - * - * If the source VM connects with an IPv6 address, then we're OK because = we can - * throw an error on the source (and similarly on the destination). - * - * But in mixed environments, this will be broken for a while until it is= fixed - * inside linux. - * - * We do provide a *tiny* bit of help in this function: We can list all of= the - * devices in the system and check to see if all the devices are RoCE or - * Infiniband. - * - * If we detect that we have a *pure* RoCE environment, then we can safely - * thrown an error even if the management software has specified '[::]' as= the - * bind address. - * - * However, if there is are multiple hetergeneous devices, then we cannot = make - * this assumption and the user just has to be sure they know what they are - * doing. - * - * Patches are being reviewed on linux-rdma. - */ -static int qemu_rdma_broken_ipv6_kernel(struct ibv_context *verbs, Error *= *errp) -{ - /* This bug only exists in linux, to our knowledge. */ -#ifdef CONFIG_LINUX - struct ibv_port_attr port_attr; - - /* - * Verbs are only NULL if management has bound to '[::]'. - * - * Let's iterate through all the devices and see if there any pure IB - * devices (non-ethernet). - * - * If not, then we can safely proceed with the migration. - * Otherwise, there are no guarantees until the bug is fixed in linux. - */ - if (!verbs) { - int num_devices; - struct ibv_device **dev_list =3D ibv_get_device_list(&num_devices); - bool roce_found =3D false; - bool ib_found =3D false; - - for (int x =3D 0; x < num_devices; x++) { - verbs =3D ibv_open_device(dev_list[x]); - /* - * ibv_open_device() is not documented to set errno. If - * it does, it's somebody else's doc bug. If it doesn't, - * the use of errno below is wrong. - * TODO Find out whether ibv_open_device() sets errno. - */ - if (!verbs) { - if (errno =3D=3D EPERM) { - continue; - } else { - error_setg_errno(errp, errno, - "could not open RDMA device context"); - return -1; - } - } - - if (ibv_query_port(verbs, 1, &port_attr)) { - ibv_close_device(verbs); - error_setg(errp, - "RDMA ERROR: Could not query initial IB port"); - return -1; - } - - if (port_attr.link_layer =3D=3D IBV_LINK_LAYER_INFINIBAND) { - ib_found =3D true; - } else if (port_attr.link_layer =3D=3D IBV_LINK_LAYER_ETHERNET= ) { - roce_found =3D true; - } - - ibv_close_device(verbs); - - } - - if (roce_found) { - if (ib_found) { - warn_report("migrations may fail:" - " IPv6 over RoCE / iWARP in linux" - " is broken. But since you appear to have a" - " mixed RoCE / IB environment, be sure to only" - " migrate over the IB fabric until the kernel " - " fixes the bug."); - } else { - error_setg(errp, "RDMA ERROR: " - "You only have RoCE / iWARP devices in your sys= tems" - " and your management software has specified '[= ::]'" - ", but IPv6 over RoCE / iWARP is not supported = in Linux."); - return -1; - } - } - - return 0; - } - - /* - * If we have a verbs context, that means that some other than '[::]' = was - * used by the management software for binding. In which case we can - * actually warn the user about a potentially broken kernel. - */ - - /* IB ports start with 1, not 0 */ - if (ibv_query_port(verbs, 1, &port_attr)) { - error_setg(errp, "RDMA ERROR: Could not query initial IB port"); - return -1; - } - - if (port_attr.link_layer =3D=3D IBV_LINK_LAYER_ETHERNET) { - error_setg(errp, "RDMA ERROR: " - "Linux kernel's RoCE / iWARP does not support IPv6 " - "(but patches on linux-rdma in progress)"); - return -1; - } - -#endif - - return 0; -} - -/* - * Figure out which RDMA device corresponds to the requested IP hostname - * Also create the initial connection manager identifiers for opening - * the connection. - */ -static int qemu_rdma_resolve_host(RDMAContext *rdma, Error **errp) -{ - Error *err =3D NULL; - int ret; - struct rdma_addrinfo *res; - char port_str[16]; - struct rdma_cm_event *cm_event; - char ip[40] =3D "unknown"; - - if (rdma->host =3D=3D NULL || !strcmp(rdma->host, "")) { - error_setg(errp, "RDMA ERROR: RDMA hostname has not been set"); - return -1; - } - - /* create CM channel */ - rdma->channel =3D rdma_create_event_channel(); - if (!rdma->channel) { - error_setg(errp, "RDMA ERROR: could not create CM channel"); - return -1; - } - - /* create CM id */ - ret =3D rdma_create_id(rdma->channel, &rdma->cm_id, NULL, RDMA_PS_TCP); - if (ret < 0) { - error_setg(errp, "RDMA ERROR: could not create channel id"); - goto err_resolve_create_id; - } - - snprintf(port_str, 16, "%d", rdma->port); - port_str[15] =3D '\0'; - - ret =3D rdma_getaddrinfo(rdma->host, port_str, NULL, &res); - if (ret) { - error_setg(errp, "RDMA ERROR: could not rdma_getaddrinfo address %= s", - rdma->host); - goto err_resolve_get_addr; - } - - /* Try all addresses, saving the first error in @err */ - for (struct rdma_addrinfo *e =3D res; e !=3D NULL; e =3D e->ai_next) { - Error **local_errp =3D err ? NULL : &err; - - inet_ntop(e->ai_family, - &((struct sockaddr_in *) e->ai_dst_addr)->sin_addr, ip, sizeof= ip); - trace_qemu_rdma_resolve_host_trying(rdma->host, ip); - - ret =3D rdma_resolve_addr(rdma->cm_id, NULL, e->ai_dst_addr, - RDMA_RESOLVE_TIMEOUT_MS); - if (ret >=3D 0) { - if (e->ai_family =3D=3D AF_INET6) { - ret =3D qemu_rdma_broken_ipv6_kernel(rdma->cm_id->verbs, - local_errp); - if (ret < 0) { - continue; - } - } - error_free(err); - goto route; - } - } - - rdma_freeaddrinfo(res); - if (err) { - error_propagate(errp, err); - } else { - error_setg(errp, "RDMA ERROR: could not resolve address %s", - rdma->host); - } - goto err_resolve_get_addr; - -route: - rdma_freeaddrinfo(res); - qemu_rdma_dump_gid("source_resolve_addr", rdma->cm_id); - - ret =3D rdma_get_cm_event(rdma->channel, &cm_event); - if (ret < 0) { - error_setg(errp, "RDMA ERROR: could not perform event_addr_resolve= d"); - goto err_resolve_get_addr; - } - - if (cm_event->event !=3D RDMA_CM_EVENT_ADDR_RESOLVED) { - error_setg(errp, - "RDMA ERROR: result not equal to event_addr_resolved %s= ", - rdma_event_str(cm_event->event)); - rdma_ack_cm_event(cm_event); - goto err_resolve_get_addr; - } - rdma_ack_cm_event(cm_event); - - /* resolve route */ - ret =3D rdma_resolve_route(rdma->cm_id, RDMA_RESOLVE_TIMEOUT_MS); - if (ret < 0) { - error_setg(errp, "RDMA ERROR: could not resolve rdma route"); - goto err_resolve_get_addr; - } - - ret =3D rdma_get_cm_event(rdma->channel, &cm_event); - if (ret < 0) { - error_setg(errp, "RDMA ERROR: could not perform event_route_resolv= ed"); - goto err_resolve_get_addr; - } - if (cm_event->event !=3D RDMA_CM_EVENT_ROUTE_RESOLVED) { - error_setg(errp, "RDMA ERROR: " - "result not equal to event_route_resolved: %s", - rdma_event_str(cm_event->event)); - rdma_ack_cm_event(cm_event); - goto err_resolve_get_addr; - } - rdma_ack_cm_event(cm_event); - rdma->verbs =3D rdma->cm_id->verbs; - qemu_rdma_dump_id("source_resolve_host", rdma->cm_id->verbs); - qemu_rdma_dump_gid("source_resolve_host", rdma->cm_id); - return 0; - -err_resolve_get_addr: - rdma_destroy_id(rdma->cm_id); - rdma->cm_id =3D NULL; -err_resolve_create_id: - rdma_destroy_event_channel(rdma->channel); - rdma->channel =3D NULL; - return -1; -} - -/* - * Create protection domain and completion queues - */ -static int qemu_rdma_alloc_pd_cq(RDMAContext *rdma, Error **errp) -{ - /* allocate pd */ - rdma->pd =3D ibv_alloc_pd(rdma->verbs); - if (!rdma->pd) { - error_setg(errp, "failed to allocate protection domain"); - return -1; - } - - /* create receive completion channel */ - rdma->recv_comp_channel =3D ibv_create_comp_channel(rdma->verbs); - if (!rdma->recv_comp_channel) { - error_setg(errp, "failed to allocate receive completion channel"); - goto err_alloc_pd_cq; - } - - /* - * Completion queue can be filled by read work requests. - */ - rdma->recv_cq =3D ibv_create_cq(rdma->verbs, (RDMA_SIGNALED_SEND_MAX *= 3), - NULL, rdma->recv_comp_channel, 0); - if (!rdma->recv_cq) { - error_setg(errp, "failed to allocate receive completion queue"); - goto err_alloc_pd_cq; - } - - /* create send completion channel */ - rdma->send_comp_channel =3D ibv_create_comp_channel(rdma->verbs); - if (!rdma->send_comp_channel) { - error_setg(errp, "failed to allocate send completion channel"); - goto err_alloc_pd_cq; - } - - rdma->send_cq =3D ibv_create_cq(rdma->verbs, (RDMA_SIGNALED_SEND_MAX *= 3), - NULL, rdma->send_comp_channel, 0); - if (!rdma->send_cq) { - error_setg(errp, "failed to allocate send completion queue"); - goto err_alloc_pd_cq; - } - - return 0; - -err_alloc_pd_cq: - if (rdma->pd) { - ibv_dealloc_pd(rdma->pd); - } - if (rdma->recv_comp_channel) { - ibv_destroy_comp_channel(rdma->recv_comp_channel); - } - if (rdma->send_comp_channel) { - ibv_destroy_comp_channel(rdma->send_comp_channel); - } - if (rdma->recv_cq) { - ibv_destroy_cq(rdma->recv_cq); - rdma->recv_cq =3D NULL; - } - rdma->pd =3D NULL; - rdma->recv_comp_channel =3D NULL; - rdma->send_comp_channel =3D NULL; - return -1; - -} - -/* - * Create queue pairs. - */ -static int qemu_rdma_alloc_qp(RDMAContext *rdma) -{ - struct ibv_qp_init_attr attr =3D { 0 }; - - attr.cap.max_send_wr =3D RDMA_SIGNALED_SEND_MAX; - attr.cap.max_recv_wr =3D 3; - attr.cap.max_send_sge =3D 1; - attr.cap.max_recv_sge =3D 1; - attr.send_cq =3D rdma->send_cq; - attr.recv_cq =3D rdma->recv_cq; - attr.qp_type =3D IBV_QPT_RC; - - if (rdma_create_qp(rdma->cm_id, rdma->pd, &attr) < 0) { - return -1; - } - - rdma->qp =3D rdma->cm_id->qp; - return 0; -} - -/* Check whether On-Demand Paging is supported by RDAM device */ -static bool rdma_support_odp(struct ibv_context *dev) -{ - struct ibv_device_attr_ex attr =3D {0}; - - if (ibv_query_device_ex(dev, NULL, &attr)) { - return false; - } - - if (attr.odp_caps.general_caps & IBV_ODP_SUPPORT) { - return true; - } - - return false; -} - -/* - * ibv_advise_mr to avoid RNR NAK error as far as possible. - * The responder mr registering with ODP will sent RNR NAK back to - * the requester in the face of the page fault. - */ -static void qemu_rdma_advise_prefetch_mr(struct ibv_pd *pd, uint64_t addr, - uint32_t len, uint32_t lkey, - const char *name, bool wr) -{ -#ifdef HAVE_IBV_ADVISE_MR - int ret; - int advice =3D wr ? IBV_ADVISE_MR_ADVICE_PREFETCH_WRITE : - IBV_ADVISE_MR_ADVICE_PREFETCH; - struct ibv_sge sg_list =3D {.lkey =3D lkey, .addr =3D addr, .length = =3D len}; - - ret =3D ibv_advise_mr(pd, advice, - IBV_ADVISE_MR_FLAG_FLUSH, &sg_list, 1); - /* ignore the error */ - trace_qemu_rdma_advise_mr(name, len, addr, strerror(ret)); -#endif -} - -static int qemu_rdma_reg_whole_ram_blocks(RDMAContext *rdma, Error **errp) -{ - int i; - RDMALocalBlocks *local =3D &rdma->local_ram_blocks; - - for (i =3D 0; i < local->nb_blocks; i++) { - int access =3D IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_REMOTE_WRITE; - - local->block[i].mr =3D - ibv_reg_mr(rdma->pd, - local->block[i].local_host_addr, - local->block[i].length, access - ); - /* - * ibv_reg_mr() is not documented to set errno. If it does, - * it's somebody else's doc bug. If it doesn't, the use of - * errno below is wrong. - * TODO Find out whether ibv_reg_mr() sets errno. - */ - if (!local->block[i].mr && - errno =3D=3D ENOTSUP && rdma_support_odp(rdma->verbs)) { - access |=3D IBV_ACCESS_ON_DEMAND; - /* register ODP mr */ - local->block[i].mr =3D - ibv_reg_mr(rdma->pd, - local->block[i].local_host_addr, - local->block[i].length, access); - trace_qemu_rdma_register_odp_mr(local->block[i].block_name= ); - - if (local->block[i].mr) { - qemu_rdma_advise_prefetch_mr(rdma->pd, - (uintptr_t)local->block[i].local_host_= addr, - local->block[i].length, - local->block[i].mr->lkey, - local->block[i].block_name, - true); - } - } - - if (!local->block[i].mr) { - error_setg_errno(errp, errno, - "Failed to register local dest ram block!"); - goto err; - } - rdma->total_registrations++; - } - - return 0; - -err: - for (i--; i >=3D 0; i--) { - ibv_dereg_mr(local->block[i].mr); - local->block[i].mr =3D NULL; - rdma->total_registrations--; - } - - return -1; - -} - -/* - * Find the ram block that corresponds to the page requested to be - * transmitted by QEMU. - * - * Once the block is found, also identify which 'chunk' within that - * block that the page belongs to. - */ -static void qemu_rdma_search_ram_block(RDMAContext *rdma, - uintptr_t block_offset, - uint64_t offset, - uint64_t length, - uint64_t *block_index, - uint64_t *chunk_index) -{ - uint64_t current_addr =3D block_offset + offset; - RDMALocalBlock *block =3D g_hash_table_lookup(rdma->blockmap, - (void *) block_offset); - assert(block); - assert(current_addr >=3D block->offset); - assert((current_addr + length) <=3D (block->offset + block->length)); - - *block_index =3D block->index; - *chunk_index =3D ram_chunk_index(block->local_host_addr, - block->local_host_addr + (current_addr - block->offset)); -} - -/* - * Register a chunk with IB. If the chunk was already registered - * previously, then skip. - * - * Also return the keys associated with the registration needed - * to perform the actual RDMA operation. - */ -static int qemu_rdma_register_and_get_keys(RDMAContext *rdma, - RDMALocalBlock *block, uintptr_t host_addr, - uint32_t *lkey, uint32_t *rkey, int chunk, - uint8_t *chunk_start, uint8_t *chunk_end) -{ - if (block->mr) { - if (lkey) { - *lkey =3D block->mr->lkey; - } - if (rkey) { - *rkey =3D block->mr->rkey; - } - return 0; - } - - /* allocate memory to store chunk MRs */ - if (!block->pmr) { - block->pmr =3D g_new0(struct ibv_mr *, block->nb_chunks); - } - - /* - * If 'rkey', then we're the destination, so grant access to the sourc= e. - * - * If 'lkey', then we're the source VM, so grant access only to oursel= ves. - */ - if (!block->pmr[chunk]) { - uint64_t len =3D chunk_end - chunk_start; - int access =3D rkey ? IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_REMOTE_W= RITE : - 0; - - trace_qemu_rdma_register_and_get_keys(len, chunk_start); - - block->pmr[chunk] =3D ibv_reg_mr(rdma->pd, chunk_start, len, acces= s); - /* - * ibv_reg_mr() is not documented to set errno. If it does, - * it's somebody else's doc bug. If it doesn't, the use of - * errno below is wrong. - * TODO Find out whether ibv_reg_mr() sets errno. - */ - if (!block->pmr[chunk] && - errno =3D=3D ENOTSUP && rdma_support_odp(rdma->verbs)) { - access |=3D IBV_ACCESS_ON_DEMAND; - /* register ODP mr */ - block->pmr[chunk] =3D ibv_reg_mr(rdma->pd, chunk_start, len, a= ccess); - trace_qemu_rdma_register_odp_mr(block->block_name); - - if (block->pmr[chunk]) { - qemu_rdma_advise_prefetch_mr(rdma->pd, (uintptr_t)chunk_st= art, - len, block->pmr[chunk]->lkey, - block->block_name, rkey); - - } - } - } - if (!block->pmr[chunk]) { - return -1; - } - rdma->total_registrations++; - - if (lkey) { - *lkey =3D block->pmr[chunk]->lkey; - } - if (rkey) { - *rkey =3D block->pmr[chunk]->rkey; - } - return 0; -} - -/* - * Register (at connection time) the memory used for control - * channel messages. - */ -static int qemu_rdma_reg_control(RDMAContext *rdma, int idx) -{ - rdma->wr_data[idx].control_mr =3D ibv_reg_mr(rdma->pd, - rdma->wr_data[idx].control, RDMA_CONTROL_MAX_BUFFER, - IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_REMOTE_WRITE); - if (rdma->wr_data[idx].control_mr) { - rdma->total_registrations++; - return 0; - } - return -1; -} - -/* - * Perform a non-optimized memory unregistration after every transfer - * for demonstration purposes, only if pin-all is not requested. - * - * Potential optimizations: - * 1. Start a new thread to run this function continuously - - for bit clearing - - and for receipt of unregister messages - * 2. Use an LRU. - * 3. Use workload hints. - */ -static int qemu_rdma_unregister_waiting(RDMAContext *rdma) -{ - Error *err =3D NULL; - - while (rdma->unregistrations[rdma->unregister_current]) { - int ret; - uint64_t wr_id =3D rdma->unregistrations[rdma->unregister_current]; - uint64_t chunk =3D - (wr_id & RDMA_WRID_CHUNK_MASK) >> RDMA_WRID_CHUNK_SHIFT; - uint64_t index =3D - (wr_id & RDMA_WRID_BLOCK_MASK) >> RDMA_WRID_BLOCK_SHIFT; - RDMALocalBlock *block =3D - &(rdma->local_ram_blocks.block[index]); - RDMARegister reg =3D { .current_index =3D index }; - RDMAControlHeader resp =3D { .type =3D RDMA_CONTROL_UNREGISTER_FIN= ISHED, - }; - RDMAControlHeader head =3D { .len =3D sizeof(RDMARegister), - .type =3D RDMA_CONTROL_UNREGISTER_REQUE= ST, - .repeat =3D 1, - }; - - trace_qemu_rdma_unregister_waiting_proc(chunk, - rdma->unregister_current); - - rdma->unregistrations[rdma->unregister_current] =3D 0; - rdma->unregister_current++; - - if (rdma->unregister_current =3D=3D RDMA_SIGNALED_SEND_MAX) { - rdma->unregister_current =3D 0; - } - - - /* - * Unregistration is speculative (because migration is single-thre= aded - * and we cannot break the protocol's inifinband message ordering). - * Thus, if the memory is currently being used for transmission, - * then abort the attempt to unregister and try again - * later the next time a completion is received for this memory. - */ - clear_bit(chunk, block->unregister_bitmap); - - if (test_bit(chunk, block->transit_bitmap)) { - trace_qemu_rdma_unregister_waiting_inflight(chunk); - continue; - } - - trace_qemu_rdma_unregister_waiting_send(chunk); - - ret =3D ibv_dereg_mr(block->pmr[chunk]); - block->pmr[chunk] =3D NULL; - block->remote_keys[chunk] =3D 0; - - if (ret !=3D 0) { - error_report("unregistration chunk failed: %s", - strerror(ret)); - return -1; - } - rdma->total_registrations--; - - reg.key.chunk =3D chunk; - register_to_network(rdma, ®); - ret =3D qemu_rdma_exchange_send(rdma, &head, (uint8_t *) ®, - &resp, NULL, NULL, &err); - if (ret < 0) { - error_report_err(err); - return -1; - } - - trace_qemu_rdma_unregister_waiting_complete(chunk); - } - - return 0; -} - -static uint64_t qemu_rdma_make_wrid(uint64_t wr_id, uint64_t index, - uint64_t chunk) -{ - uint64_t result =3D wr_id & RDMA_WRID_TYPE_MASK; - - result |=3D (index << RDMA_WRID_BLOCK_SHIFT); - result |=3D (chunk << RDMA_WRID_CHUNK_SHIFT); - - return result; -} - -/* - * Consult the connection manager to see a work request - * (of any kind) has completed. - * Return the work request ID that completed. - */ -static int qemu_rdma_poll(RDMAContext *rdma, struct ibv_cq *cq, - uint64_t *wr_id_out, uint32_t *byte_len) -{ - int ret; - struct ibv_wc wc; - uint64_t wr_id; - - ret =3D ibv_poll_cq(cq, 1, &wc); - - if (!ret) { - *wr_id_out =3D RDMA_WRID_NONE; - return 0; - } - - if (ret < 0) { - return -1; - } - - wr_id =3D wc.wr_id & RDMA_WRID_TYPE_MASK; - - if (wc.status !=3D IBV_WC_SUCCESS) { - return -1; - } - - if (rdma->control_ready_expected && - (wr_id >=3D RDMA_WRID_RECV_CONTROL)) { - trace_qemu_rdma_poll_recv(wr_id - RDMA_WRID_RECV_CONTROL, wr_id, - rdma->nb_sent); - rdma->control_ready_expected =3D 0; - } - - if (wr_id =3D=3D RDMA_WRID_RDMA_WRITE) { - uint64_t chunk =3D - (wc.wr_id & RDMA_WRID_CHUNK_MASK) >> RDMA_WRID_CHUNK_SHIFT; - uint64_t index =3D - (wc.wr_id & RDMA_WRID_BLOCK_MASK) >> RDMA_WRID_BLOCK_SHIFT; - RDMALocalBlock *block =3D &(rdma->local_ram_blocks.block[index]); - - trace_qemu_rdma_poll_write(wr_id, rdma->nb_sent, - index, chunk, block->local_host_addr, - (void *)(uintptr_t)block->remote_host_a= ddr); - - clear_bit(chunk, block->transit_bitmap); - - if (rdma->nb_sent > 0) { - rdma->nb_sent--; - } - } else { - trace_qemu_rdma_poll_other(wr_id, rdma->nb_sent); - } - - *wr_id_out =3D wc.wr_id; - if (byte_len) { - *byte_len =3D wc.byte_len; - } - - return 0; -} - -/* Wait for activity on the completion channel. - * Returns 0 on success, none-0 on error. - */ -static int qemu_rdma_wait_comp_channel(RDMAContext *rdma, - struct ibv_comp_channel *comp_chann= el) -{ - struct rdma_cm_event *cm_event; - - /* - * Coroutine doesn't start until migration_fd_process_incoming() - * so don't yield unless we know we're running inside of a coroutine. - */ - if (rdma->migration_started_on_destination && - migration_incoming_get_current()->state =3D=3D MIGRATION_STATUS_AC= TIVE) { - yield_until_fd_readable(comp_channel->fd); - } else { - /* This is the source side, we're in a separate thread - * or destination prior to migration_fd_process_incoming() - * after postcopy, the destination also in a separate thread. - * we can't yield; so we have to poll the fd. - * But we need to be able to handle 'cancel' or an error - * without hanging forever. - */ - while (!rdma->errored && !rdma->received_error) { - GPollFD pfds[2]; - pfds[0].fd =3D comp_channel->fd; - pfds[0].events =3D G_IO_IN | G_IO_HUP | G_IO_ERR; - pfds[0].revents =3D 0; - - pfds[1].fd =3D rdma->channel->fd; - pfds[1].events =3D G_IO_IN | G_IO_HUP | G_IO_ERR; - pfds[1].revents =3D 0; - - /* 0.1s timeout, should be fine for a 'cancel' */ - switch (qemu_poll_ns(pfds, 2, 100 * 1000 * 1000)) { - case 2: - case 1: /* fd active */ - if (pfds[0].revents) { - return 0; - } - - if (pfds[1].revents) { - if (rdma_get_cm_event(rdma->channel, &cm_event) < 0) { - return -1; - } - - if (cm_event->event =3D=3D RDMA_CM_EVENT_DISCONNECTED = || - cm_event->event =3D=3D RDMA_CM_EVENT_DEVICE_REMOVA= L) { - rdma_ack_cm_event(cm_event); - return -1; - } - rdma_ack_cm_event(cm_event); - } - break; - - case 0: /* Timeout, go around again */ - break; - - default: /* Error of some type - - * I don't trust errno from qemu_poll_ns - */ - return -1; - } - - if (migrate_get_current()->state =3D=3D MIGRATION_STATUS_CANCE= LLING) { - /* Bail out and let the cancellation happen */ - return -1; - } - } - } - - if (rdma->received_error) { - return -1; - } - return -rdma->errored; -} - -static struct ibv_comp_channel *to_channel(RDMAContext *rdma, uint64_t wri= d) -{ - return wrid < RDMA_WRID_RECV_CONTROL ? rdma->send_comp_channel : - rdma->recv_comp_channel; -} - -static struct ibv_cq *to_cq(RDMAContext *rdma, uint64_t wrid) -{ - return wrid < RDMA_WRID_RECV_CONTROL ? rdma->send_cq : rdma->recv_cq; -} - -/* - * Block until the next work request has completed. - * - * First poll to see if a work request has already completed, - * otherwise block. - * - * If we encounter completed work requests for IDs other than - * the one we're interested in, then that's generally an error. - * - * The only exception is actual RDMA Write completions. These - * completions only need to be recorded, but do not actually - * need further processing. - */ -static int qemu_rdma_block_for_wrid(RDMAContext *rdma, - uint64_t wrid_requested, - uint32_t *byte_len) -{ - int num_cq_events =3D 0, ret; - struct ibv_cq *cq; - void *cq_ctx; - uint64_t wr_id =3D RDMA_WRID_NONE, wr_id_in; - struct ibv_comp_channel *ch =3D to_channel(rdma, wrid_requested); - struct ibv_cq *poll_cq =3D to_cq(rdma, wrid_requested); - - if (ibv_req_notify_cq(poll_cq, 0)) { - return -1; - } - /* poll cq first */ - while (wr_id !=3D wrid_requested) { - ret =3D qemu_rdma_poll(rdma, poll_cq, &wr_id_in, byte_len); - if (ret < 0) { - return -1; - } - - wr_id =3D wr_id_in & RDMA_WRID_TYPE_MASK; - - if (wr_id =3D=3D RDMA_WRID_NONE) { - break; - } - if (wr_id !=3D wrid_requested) { - trace_qemu_rdma_block_for_wrid_miss(wrid_requested, wr_id); - } - } - - if (wr_id =3D=3D wrid_requested) { - return 0; - } - - while (1) { - ret =3D qemu_rdma_wait_comp_channel(rdma, ch); - if (ret < 0) { - goto err_block_for_wrid; - } - - ret =3D ibv_get_cq_event(ch, &cq, &cq_ctx); - if (ret < 0) { - goto err_block_for_wrid; - } - - num_cq_events++; - - if (ibv_req_notify_cq(cq, 0)) { - goto err_block_for_wrid; - } - - while (wr_id !=3D wrid_requested) { - ret =3D qemu_rdma_poll(rdma, poll_cq, &wr_id_in, byte_len); - if (ret < 0) { - goto err_block_for_wrid; - } - - wr_id =3D wr_id_in & RDMA_WRID_TYPE_MASK; - - if (wr_id =3D=3D RDMA_WRID_NONE) { - break; - } - if (wr_id !=3D wrid_requested) { - trace_qemu_rdma_block_for_wrid_miss(wrid_requested, wr_id); - } - } - - if (wr_id =3D=3D wrid_requested) { - goto success_block_for_wrid; - } - } - -success_block_for_wrid: - if (num_cq_events) { - ibv_ack_cq_events(cq, num_cq_events); - } - return 0; - -err_block_for_wrid: - if (num_cq_events) { - ibv_ack_cq_events(cq, num_cq_events); - } - - rdma->errored =3D true; - return -1; -} - -/* - * Post a SEND message work request for the control channel - * containing some data and block until the post completes. - */ -static int qemu_rdma_post_send_control(RDMAContext *rdma, uint8_t *buf, - RDMAControlHeader *head, - Error **errp) -{ - int ret; - RDMAWorkRequestData *wr =3D &rdma->wr_data[RDMA_WRID_CONTROL]; - struct ibv_send_wr *bad_wr; - struct ibv_sge sge =3D { - .addr =3D (uintptr_t)(wr->control), - .length =3D head->len + sizeof(RDMAControlHeade= r), - .lkey =3D wr->control_mr->lkey, - }; - struct ibv_send_wr send_wr =3D { - .wr_id =3D RDMA_WRID_SEND_CONTROL, - .opcode =3D IBV_WR_SEND, - .send_flags =3D IBV_SEND_SIGNALED, - .sg_list =3D &sge, - .num_sge =3D 1, - }; - - trace_qemu_rdma_post_send_control(control_desc(head->type)); - - /* - * We don't actually need to do a memcpy() in here if we used - * the "sge" properly, but since we're only sending control messages - * (not RAM in a performance-critical path), then its OK for now. - * - * The copy makes the RDMAControlHeader simpler to manipulate - * for the time being. - */ - assert(head->len <=3D RDMA_CONTROL_MAX_BUFFER - sizeof(*head)); - memcpy(wr->control, head, sizeof(RDMAControlHeader)); - control_to_network((void *) wr->control); - - if (buf) { - memcpy(wr->control + sizeof(RDMAControlHeader), buf, head->len); - } - - - ret =3D ibv_post_send(rdma->qp, &send_wr, &bad_wr); - - if (ret > 0) { - error_setg(errp, "Failed to use post IB SEND for control"); - return -1; - } - - ret =3D qemu_rdma_block_for_wrid(rdma, RDMA_WRID_SEND_CONTROL, NULL); - if (ret < 0) { - error_setg(errp, "rdma migration: send polling control error"); - return -1; - } - - return 0; -} - -/* - * Post a RECV work request in anticipation of some future receipt - * of data on the control channel. - */ -static int qemu_rdma_post_recv_control(RDMAContext *rdma, int idx, - Error **errp) -{ - struct ibv_recv_wr *bad_wr; - struct ibv_sge sge =3D { - .addr =3D (uintptr_t)(rdma->wr_data[idx].contr= ol), - .length =3D RDMA_CONTROL_MAX_BUFFER, - .lkey =3D rdma->wr_data[idx].control_mr->lkey, - }; - - struct ibv_recv_wr recv_wr =3D { - .wr_id =3D RDMA_WRID_RECV_CONTROL + id= x, - .sg_list =3D &sge, - .num_sge =3D 1, - }; - - - if (ibv_post_recv(rdma->qp, &recv_wr, &bad_wr)) { - error_setg(errp, "error posting control recv"); - return -1; - } - - return 0; -} - -/* - * Block and wait for a RECV control channel message to arrive. - */ -static int qemu_rdma_exchange_get_response(RDMAContext *rdma, - RDMAControlHeader *head, uint32_t expecting, int idx, - Error **errp) -{ - uint32_t byte_len; - int ret =3D qemu_rdma_block_for_wrid(rdma, RDMA_WRID_RECV_CONTROL + id= x, - &byte_len); - - if (ret < 0) { - error_setg(errp, "rdma migration: recv polling control error!"); - return -1; - } - - network_to_control((void *) rdma->wr_data[idx].control); - memcpy(head, rdma->wr_data[idx].control, sizeof(RDMAControlHeader)); - - trace_qemu_rdma_exchange_get_response_start(control_desc(expecting)); - - if (expecting =3D=3D RDMA_CONTROL_NONE) { - trace_qemu_rdma_exchange_get_response_none(control_desc(head->type= ), - head->type); - } else if (head->type !=3D expecting || head->type =3D=3D RDMA_CONTROL= _ERROR) { - error_setg(errp, "Was expecting a %s (%d) control message" - ", but got: %s (%d), length: %d", - control_desc(expecting), expecting, - control_desc(head->type), head->type, head->len); - if (head->type =3D=3D RDMA_CONTROL_ERROR) { - rdma->received_error =3D true; - } - return -1; - } - if (head->len > RDMA_CONTROL_MAX_BUFFER - sizeof(*head)) { - error_setg(errp, "too long length: %d", head->len); - return -1; - } - if (sizeof(*head) + head->len !=3D byte_len) { - error_setg(errp, "Malformed length: %d byte_len %d", - head->len, byte_len); - return -1; - } - - return 0; -} - -/* - * When a RECV work request has completed, the work request's - * buffer is pointed at the header. - * - * This will advance the pointer to the data portion - * of the control message of the work request's buffer that - * was populated after the work request finished. - */ -static void qemu_rdma_move_header(RDMAContext *rdma, int idx, - RDMAControlHeader *head) -{ - rdma->wr_data[idx].control_len =3D head->len; - rdma->wr_data[idx].control_curr =3D - rdma->wr_data[idx].control + sizeof(RDMAControlHeader); -} - -/* - * This is an 'atomic' high-level operation to deliver a single, unified - * control-channel message. - * - * Additionally, if the user is expecting some kind of reply to this messa= ge, - * they can request a 'resp' response message be filled in by posting an - * additional work request on behalf of the user and waiting for an additi= onal - * completion. - * - * The extra (optional) response is used during registration to us from ha= ving - * to perform an *additional* exchange of message just to provide a respon= se by - * instead piggy-backing on the acknowledgement. - */ -static int qemu_rdma_exchange_send(RDMAContext *rdma, RDMAControlHeader *h= ead, - uint8_t *data, RDMAControlHeader *resp, - int *resp_idx, - int (*callback)(RDMAContext *rdma, - Error **errp), - Error **errp) -{ - int ret; - - /* - * Wait until the dest is ready before attempting to deliver the messa= ge - * by waiting for a READY message. - */ - if (rdma->control_ready_expected) { - RDMAControlHeader resp_ignored; - - ret =3D qemu_rdma_exchange_get_response(rdma, &resp_ignored, - RDMA_CONTROL_READY, - RDMA_WRID_READY, errp); - if (ret < 0) { - return -1; - } - } - - /* - * If the user is expecting a response, post a WR in anticipation of i= t. - */ - if (resp) { - ret =3D qemu_rdma_post_recv_control(rdma, RDMA_WRID_DATA, errp); - if (ret < 0) { - return -1; - } - } - - /* - * Post a WR to replace the one we just consumed for the READY message. - */ - ret =3D qemu_rdma_post_recv_control(rdma, RDMA_WRID_READY, errp); - if (ret < 0) { - return -1; - } - - /* - * Deliver the control message that was requested. - */ - ret =3D qemu_rdma_post_send_control(rdma, data, head, errp); - - if (ret < 0) { - return -1; - } - - /* - * If we're expecting a response, block and wait for it. - */ - if (resp) { - if (callback) { - trace_qemu_rdma_exchange_send_issue_callback(); - ret =3D callback(rdma, errp); - if (ret < 0) { - return -1; - } - } - - trace_qemu_rdma_exchange_send_waiting(control_desc(resp->type)); - ret =3D qemu_rdma_exchange_get_response(rdma, resp, - resp->type, RDMA_WRID_DATA, - errp); - - if (ret < 0) { - return -1; - } - - qemu_rdma_move_header(rdma, RDMA_WRID_DATA, resp); - if (resp_idx) { - *resp_idx =3D RDMA_WRID_DATA; - } - trace_qemu_rdma_exchange_send_received(control_desc(resp->type)); - } - - rdma->control_ready_expected =3D 1; - - return 0; -} - -/* - * This is an 'atomic' high-level operation to receive a single, unified - * control-channel message. - */ -static int qemu_rdma_exchange_recv(RDMAContext *rdma, RDMAControlHeader *h= ead, - uint32_t expecting, Error **errp) -{ - RDMAControlHeader ready =3D { - .len =3D 0, - .type =3D RDMA_CONTROL_READY, - .repeat =3D 1, - }; - int ret; - - /* - * Inform the source that we're ready to receive a message. - */ - ret =3D qemu_rdma_post_send_control(rdma, NULL, &ready, errp); - - if (ret < 0) { - return -1; - } - - /* - * Block and wait for the message. - */ - ret =3D qemu_rdma_exchange_get_response(rdma, head, - expecting, RDMA_WRID_READY, errp= ); - - if (ret < 0) { - return -1; - } - - qemu_rdma_move_header(rdma, RDMA_WRID_READY, head); - - /* - * Post a new RECV work request to replace the one we just consumed. - */ - ret =3D qemu_rdma_post_recv_control(rdma, RDMA_WRID_READY, errp); - if (ret < 0) { - return -1; - } - - return 0; -} - -/* - * Write an actual chunk of memory using RDMA. - * - * If we're using dynamic registration on the dest-side, we have to - * send a registration command first. - */ -static int qemu_rdma_write_one(RDMAContext *rdma, - int current_index, uint64_t current_addr, - uint64_t length, Error **errp) -{ - struct ibv_sge sge; - struct ibv_send_wr send_wr =3D { 0 }; - struct ibv_send_wr *bad_wr; - int reg_result_idx, ret, count =3D 0; - uint64_t chunk, chunks; - uint8_t *chunk_start, *chunk_end; - RDMALocalBlock *block =3D &(rdma->local_ram_blocks.block[current_index= ]); - RDMARegister reg; - RDMARegisterResult *reg_result; - RDMAControlHeader resp =3D { .type =3D RDMA_CONTROL_REGISTER_RESULT }; - RDMAControlHeader head =3D { .len =3D sizeof(RDMARegister), - .type =3D RDMA_CONTROL_REGISTER_REQUEST, - .repeat =3D 1, - }; - -retry: - sge.addr =3D (uintptr_t)(block->local_host_addr + - (current_addr - block->offset)); - sge.length =3D length; - - chunk =3D ram_chunk_index(block->local_host_addr, - (uint8_t *)(uintptr_t)sge.addr); - chunk_start =3D ram_chunk_start(block, chunk); - - if (block->is_ram_block) { - chunks =3D length / (1UL << RDMA_REG_CHUNK_SHIFT); - - if (chunks && ((length % (1UL << RDMA_REG_CHUNK_SHIFT)) =3D=3D 0))= { - chunks--; - } - } else { - chunks =3D block->length / (1UL << RDMA_REG_CHUNK_SHIFT); - - if (chunks && ((block->length % (1UL << RDMA_REG_CHUNK_SHIFT)) =3D= =3D 0)) { - chunks--; - } - } - - trace_qemu_rdma_write_one_top(chunks + 1, - (chunks + 1) * - (1UL << RDMA_REG_CHUNK_SHIFT) / 1024 / 1= 024); - - chunk_end =3D ram_chunk_end(block, chunk + chunks); - - - while (test_bit(chunk, block->transit_bitmap)) { - (void)count; - trace_qemu_rdma_write_one_block(count++, current_index, chunk, - sge.addr, length, rdma->nb_sent, block->nb_chunks); - - ret =3D qemu_rdma_block_for_wrid(rdma, RDMA_WRID_RDMA_WRITE, NULL); - - if (ret < 0) { - error_setg(errp, "Failed to Wait for previous write to complet= e " - "block %d chunk %" PRIu64 - " current %" PRIu64 " len %" PRIu64 " %d", - current_index, chunk, sge.addr, length, rdma->nb_sent); - return -1; - } - } - - if (!rdma->pin_all || !block->is_ram_block) { - if (!block->remote_keys[chunk]) { - /* - * This chunk has not yet been registered, so first check to s= ee - * if the entire chunk is zero. If so, tell the other size to - * memset() + madvise() the entire chunk without RDMA. - */ - - if (buffer_is_zero((void *)(uintptr_t)sge.addr, length)) { - RDMACompress comp =3D { - .offset =3D current_addr, - .value =3D 0, - .block_idx =3D current_index, - .length =3D length, - }; - - head.len =3D sizeof(comp); - head.type =3D RDMA_CONTROL_COMPRESS; - - trace_qemu_rdma_write_one_zero(chunk, sge.length, - current_index, current_addr= ); - - compress_to_network(rdma, &comp); - ret =3D qemu_rdma_exchange_send(rdma, &head, - (uint8_t *) &comp, NULL, NULL, NULL, errp); - - if (ret < 0) { - return -1; - } - - /* - * TODO: Here we are sending something, but we are not - * accounting for anything transferred. The following is = wrong: - * - * stat64_add(&mig_stats.rdma_bytes, sge.length); - * - * because we are using some kind of compression. I - * would think that head.len would be the more similar - * thing to a correct value. - */ - stat64_add(&mig_stats.zero_pages, - sge.length / qemu_target_page_size()); - return 1; - } - - /* - * Otherwise, tell other side to register. - */ - reg.current_index =3D current_index; - if (block->is_ram_block) { - reg.key.current_addr =3D current_addr; - } else { - reg.key.chunk =3D chunk; - } - reg.chunks =3D chunks; - - trace_qemu_rdma_write_one_sendreg(chunk, sge.length, current_i= ndex, - current_addr); - - register_to_network(rdma, ®); - ret =3D qemu_rdma_exchange_send(rdma, &head, (uint8_t *) ®, - &resp, ®_result_idx, NULL, errp); - if (ret < 0) { - return -1; - } - - /* try to overlap this single registration with the one we sen= t. */ - if (qemu_rdma_register_and_get_keys(rdma, block, sge.addr, - &sge.lkey, NULL, chunk, - chunk_start, chunk_end)) { - error_setg(errp, "cannot get lkey"); - return -1; - } - - reg_result =3D (RDMARegisterResult *) - rdma->wr_data[reg_result_idx].control_curr; - - network_to_result(reg_result); - - trace_qemu_rdma_write_one_recvregres(block->remote_keys[chunk], - reg_result->rkey, chunk); - - block->remote_keys[chunk] =3D reg_result->rkey; - block->remote_host_addr =3D reg_result->host_addr; - } else { - /* already registered before */ - if (qemu_rdma_register_and_get_keys(rdma, block, sge.addr, - &sge.lkey, NULL, chunk, - chunk_start, chunk_end)) { - error_setg(errp, "cannot get lkey!"); - return -1; - } - } - - send_wr.wr.rdma.rkey =3D block->remote_keys[chunk]; - } else { - send_wr.wr.rdma.rkey =3D block->remote_rkey; - - if (qemu_rdma_register_and_get_keys(rdma, block, sge.addr, - &sge.lkey, NULL, chun= k, - chunk_start, chunk_en= d)) { - error_setg(errp, "cannot get lkey!"); - return -1; - } - } - - /* - * Encode the ram block index and chunk within this wrid. - * We will use this information at the time of completion - * to figure out which bitmap to check against and then which - * chunk in the bitmap to look for. - */ - send_wr.wr_id =3D qemu_rdma_make_wrid(RDMA_WRID_RDMA_WRITE, - current_index, chunk); - - send_wr.opcode =3D IBV_WR_RDMA_WRITE; - send_wr.send_flags =3D IBV_SEND_SIGNALED; - send_wr.sg_list =3D &sge; - send_wr.num_sge =3D 1; - send_wr.wr.rdma.remote_addr =3D block->remote_host_addr + - (current_addr - block->offset); - - trace_qemu_rdma_write_one_post(chunk, sge.addr, send_wr.wr.rdma.remote= _addr, - sge.length); - - /* - * ibv_post_send() does not return negative error numbers, - * per the specification they are positive - no idea why. - */ - ret =3D ibv_post_send(rdma->qp, &send_wr, &bad_wr); - - if (ret =3D=3D ENOMEM) { - trace_qemu_rdma_write_one_queue_full(); - ret =3D qemu_rdma_block_for_wrid(rdma, RDMA_WRID_RDMA_WRITE, NULL); - if (ret < 0) { - error_setg(errp, "rdma migration: failed to make " - "room in full send queue!"); - return -1; - } - - goto retry; - - } else if (ret > 0) { - error_setg_errno(errp, ret, - "rdma migration: post rdma write failed"); - return -1; - } - - set_bit(chunk, block->transit_bitmap); - stat64_add(&mig_stats.normal_pages, sge.length / qemu_target_page_size= ()); - /* - * We are adding to transferred the amount of data written, but no - * overhead at all. I will assume that RDMA is magicaly and don't - * need to transfer (at least) the addresses where it wants to - * write the pages. Here it looks like it should be something - * like: - * sizeof(send_wr) + sge.length - * but this being RDMA, who knows. - */ - stat64_add(&mig_stats.rdma_bytes, sge.length); - ram_transferred_add(sge.length); - rdma->total_writes++; - - return 0; -} - -/* - * Push out any unwritten RDMA operations. - * - * We support sending out multiple chunks at the same time. - * Not all of them need to get signaled in the completion queue. - */ -static int qemu_rdma_write_flush(RDMAContext *rdma, Error **errp) -{ - int ret; - - if (!rdma->current_length) { - return 0; - } - - ret =3D qemu_rdma_write_one(rdma, rdma->current_index, rdma->current_a= ddr, - rdma->current_length, errp); - - if (ret < 0) { - return -1; - } - - if (ret =3D=3D 0) { - rdma->nb_sent++; - trace_qemu_rdma_write_flush(rdma->nb_sent); - } - - rdma->current_length =3D 0; - rdma->current_addr =3D 0; - - return 0; -} - -static inline bool qemu_rdma_buffer_mergeable(RDMAContext *rdma, - uint64_t offset, uint64_t len) -{ - RDMALocalBlock *block; - uint8_t *host_addr; - uint8_t *chunk_end; - - if (rdma->current_index < 0) { - return false; - } - - if (rdma->current_chunk < 0) { - return false; - } - - block =3D &(rdma->local_ram_blocks.block[rdma->current_index]); - host_addr =3D block->local_host_addr + (offset - block->offset); - chunk_end =3D ram_chunk_end(block, rdma->current_chunk); - - if (rdma->current_length =3D=3D 0) { - return false; - } - - /* - * Only merge into chunk sequentially. - */ - if (offset !=3D (rdma->current_addr + rdma->current_length)) { - return false; - } - - if (offset < block->offset) { - return false; - } - - if ((offset + len) > (block->offset + block->length)) { - return false; - } - - if ((host_addr + len) > chunk_end) { - return false; - } - - return true; -} - -/* - * We're not actually writing here, but doing three things: - * - * 1. Identify the chunk the buffer belongs to. - * 2. If the chunk is full or the buffer doesn't belong to the current - * chunk, then start a new chunk and flush() the old chunk. - * 3. To keep the hardware busy, we also group chunks into batches - * and only require that a batch gets acknowledged in the completion - * queue instead of each individual chunk. - */ -static int qemu_rdma_write(RDMAContext *rdma, - uint64_t block_offset, uint64_t offset, - uint64_t len, Error **errp) -{ - uint64_t current_addr =3D block_offset + offset; - uint64_t index =3D rdma->current_index; - uint64_t chunk =3D rdma->current_chunk; - - /* If we cannot merge it, we flush the current buffer first. */ - if (!qemu_rdma_buffer_mergeable(rdma, current_addr, len)) { - if (qemu_rdma_write_flush(rdma, errp) < 0) { - return -1; - } - rdma->current_length =3D 0; - rdma->current_addr =3D current_addr; - - qemu_rdma_search_ram_block(rdma, block_offset, - offset, len, &index, &chunk); - rdma->current_index =3D index; - rdma->current_chunk =3D chunk; - } - - /* merge it */ - rdma->current_length +=3D len; - - /* flush it if buffer is too large */ - if (rdma->current_length >=3D RDMA_MERGE_MAX) { - return qemu_rdma_write_flush(rdma, errp); - } - - return 0; -} - -static void qemu_rdma_cleanup(RDMAContext *rdma) -{ - Error *err =3D NULL; - - if (rdma->cm_id && rdma->connected) { - if ((rdma->errored || - migrate_get_current()->state =3D=3D MIGRATION_STATUS_CANCELLI= NG) && - !rdma->received_error) { - RDMAControlHeader head =3D { .len =3D 0, - .type =3D RDMA_CONTROL_ERROR, - .repeat =3D 1, - }; - warn_report("Early error. Sending error."); - if (qemu_rdma_post_send_control(rdma, NULL, &head, &err) < 0) { - warn_report_err(err); - } - } - - rdma_disconnect(rdma->cm_id); - trace_qemu_rdma_cleanup_disconnect(); - rdma->connected =3D false; - } - - if (rdma->channel) { - qemu_set_fd_handler(rdma->channel->fd, NULL, NULL, NULL); - } - g_free(rdma->dest_blocks); - rdma->dest_blocks =3D NULL; - - for (int i =3D 0; i < RDMA_WRID_MAX; i++) { - if (rdma->wr_data[i].control_mr) { - rdma->total_registrations--; - ibv_dereg_mr(rdma->wr_data[i].control_mr); - } - rdma->wr_data[i].control_mr =3D NULL; - } - - if (rdma->local_ram_blocks.block) { - while (rdma->local_ram_blocks.nb_blocks) { - rdma_delete_block(rdma, &rdma->local_ram_blocks.block[0]); - } - } - - if (rdma->qp) { - rdma_destroy_qp(rdma->cm_id); - rdma->qp =3D NULL; - } - if (rdma->recv_cq) { - ibv_destroy_cq(rdma->recv_cq); - rdma->recv_cq =3D NULL; - } - if (rdma->send_cq) { - ibv_destroy_cq(rdma->send_cq); - rdma->send_cq =3D NULL; - } - if (rdma->recv_comp_channel) { - ibv_destroy_comp_channel(rdma->recv_comp_channel); - rdma->recv_comp_channel =3D NULL; - } - if (rdma->send_comp_channel) { - ibv_destroy_comp_channel(rdma->send_comp_channel); - rdma->send_comp_channel =3D NULL; - } - if (rdma->pd) { - ibv_dealloc_pd(rdma->pd); - rdma->pd =3D NULL; - } - if (rdma->cm_id) { - rdma_destroy_id(rdma->cm_id); - rdma->cm_id =3D NULL; - } - - /* the destination side, listen_id and channel is shared */ - if (rdma->listen_id) { - if (!rdma->is_return_path) { - rdma_destroy_id(rdma->listen_id); - } - rdma->listen_id =3D NULL; - - if (rdma->channel) { - if (!rdma->is_return_path) { - rdma_destroy_event_channel(rdma->channel); - } - rdma->channel =3D NULL; - } - } - - if (rdma->channel) { - rdma_destroy_event_channel(rdma->channel); - rdma->channel =3D NULL; - } - g_free(rdma->host); - rdma->host =3D NULL; -} - - -static int qemu_rdma_source_init(RDMAContext *rdma, bool pin_all, Error **= errp) -{ - int ret; - - /* - * Will be validated against destination's actual capabilities - * after the connect() completes. - */ - rdma->pin_all =3D pin_all; - - ret =3D qemu_rdma_resolve_host(rdma, errp); - if (ret < 0) { - goto err_rdma_source_init; - } - - ret =3D qemu_rdma_alloc_pd_cq(rdma, errp); - if (ret < 0) { - goto err_rdma_source_init; - } - - ret =3D qemu_rdma_alloc_qp(rdma); - if (ret < 0) { - error_setg(errp, "RDMA ERROR: rdma migration: error allocating qp!= "); - goto err_rdma_source_init; - } - - qemu_rdma_init_ram_blocks(rdma); - - /* Build the hash that maps from offset to RAMBlock */ - rdma->blockmap =3D g_hash_table_new(g_direct_hash, g_direct_equal); - for (int i =3D 0; i < rdma->local_ram_blocks.nb_blocks; i++) { - g_hash_table_insert(rdma->blockmap, - (void *)(uintptr_t)rdma->local_ram_blocks.block[i].offset, - &rdma->local_ram_blocks.block[i]); - } - - for (int i =3D 0; i < RDMA_WRID_MAX; i++) { - ret =3D qemu_rdma_reg_control(rdma, i); - if (ret < 0) { - error_setg(errp, "RDMA ERROR: rdma migration: error " - "registering %d control!", i); - goto err_rdma_source_init; - } - } - - return 0; - -err_rdma_source_init: - qemu_rdma_cleanup(rdma); - return -1; -} - -static int qemu_get_cm_event_timeout(RDMAContext *rdma, - struct rdma_cm_event **cm_event, - long msec, Error **errp) -{ - int ret; - struct pollfd poll_fd =3D { - .fd =3D rdma->channel->fd, - .events =3D POLLIN, - .revents =3D 0 - }; - - do { - ret =3D poll(&poll_fd, 1, msec); - } while (ret < 0 && errno =3D=3D EINTR); - - if (ret =3D=3D 0) { - error_setg(errp, "RDMA ERROR: poll cm event timeout"); - return -1; - } else if (ret < 0) { - error_setg(errp, "RDMA ERROR: failed to poll cm event, errno=3D%i", - errno); - return -1; - } else if (poll_fd.revents & POLLIN) { - if (rdma_get_cm_event(rdma->channel, cm_event) < 0) { - error_setg(errp, "RDMA ERROR: failed to get cm event"); - return -1; - } - return 0; - } else { - error_setg(errp, "RDMA ERROR: no POLLIN event, revent=3D%x", - poll_fd.revents); - return -1; - } -} - -static int qemu_rdma_connect(RDMAContext *rdma, bool return_path, - Error **errp) -{ - RDMACapabilities cap =3D { - .version =3D RDMA_CONTROL_VERSION_CURRENT, - .flags =3D 0, - }; - struct rdma_conn_param conn_param =3D { .initiator_depth =3D 2, - .retry_count =3D 5, - .private_data =3D &cap, - .private_data_len =3D sizeof(cap= ), - }; - struct rdma_cm_event *cm_event; - int ret; - - /* - * Only negotiate the capability with destination if the user - * on the source first requested the capability. - */ - if (rdma->pin_all) { - trace_qemu_rdma_connect_pin_all_requested(); - cap.flags |=3D RDMA_CAPABILITY_PIN_ALL; - } - - caps_to_network(&cap); - - ret =3D qemu_rdma_post_recv_control(rdma, RDMA_WRID_READY, errp); - if (ret < 0) { - goto err_rdma_source_connect; - } - - ret =3D rdma_connect(rdma->cm_id, &conn_param); - if (ret < 0) { - error_setg_errno(errp, errno, - "RDMA ERROR: connecting to destination!"); - goto err_rdma_source_connect; - } - - if (return_path) { - ret =3D qemu_get_cm_event_timeout(rdma, &cm_event, 5000, errp); - } else { - ret =3D rdma_get_cm_event(rdma->channel, &cm_event); - if (ret < 0) { - error_setg_errno(errp, errno, - "RDMA ERROR: failed to get cm event"); - } - } - if (ret < 0) { - goto err_rdma_source_connect; - } - - if (cm_event->event !=3D RDMA_CM_EVENT_ESTABLISHED) { - error_setg(errp, "RDMA ERROR: connecting to destination!"); - rdma_ack_cm_event(cm_event); - goto err_rdma_source_connect; - } - rdma->connected =3D true; - - memcpy(&cap, cm_event->param.conn.private_data, sizeof(cap)); - network_to_caps(&cap); - - /* - * Verify that the *requested* capabilities are supported by the desti= nation - * and disable them otherwise. - */ - if (rdma->pin_all && !(cap.flags & RDMA_CAPABILITY_PIN_ALL)) { - warn_report("RDMA: Server cannot support pinning all memory. " - "Will register memory dynamically."); - rdma->pin_all =3D false; - } - - trace_qemu_rdma_connect_pin_all_outcome(rdma->pin_all); - - rdma_ack_cm_event(cm_event); - - rdma->control_ready_expected =3D 1; - rdma->nb_sent =3D 0; - return 0; - -err_rdma_source_connect: - qemu_rdma_cleanup(rdma); - return -1; -} - -static int qemu_rdma_dest_init(RDMAContext *rdma, Error **errp) -{ - Error *err =3D NULL; - int ret; - struct rdma_cm_id *listen_id; - char ip[40] =3D "unknown"; - struct rdma_addrinfo *res, *e; - char port_str[16]; - int reuse =3D 1; - - for (int i =3D 0; i < RDMA_WRID_MAX; i++) { - rdma->wr_data[i].control_len =3D 0; - rdma->wr_data[i].control_curr =3D NULL; - } - - if (!rdma->host || !rdma->host[0]) { - error_setg(errp, "RDMA ERROR: RDMA host is not set!"); - rdma->errored =3D true; - return -1; - } - /* create CM channel */ - rdma->channel =3D rdma_create_event_channel(); - if (!rdma->channel) { - error_setg(errp, "RDMA ERROR: could not create rdma event channel"= ); - rdma->errored =3D true; - return -1; - } - - /* create CM id */ - ret =3D rdma_create_id(rdma->channel, &listen_id, NULL, RDMA_PS_TCP); - if (ret < 0) { - error_setg(errp, "RDMA ERROR: could not create cm_id!"); - goto err_dest_init_create_listen_id; - } - - snprintf(port_str, 16, "%d", rdma->port); - port_str[15] =3D '\0'; - - ret =3D rdma_getaddrinfo(rdma->host, port_str, NULL, &res); - if (ret) { - error_setg(errp, "RDMA ERROR: could not rdma_getaddrinfo address %= s", - rdma->host); - goto err_dest_init_bind_addr; - } - - ret =3D rdma_set_option(listen_id, RDMA_OPTION_ID, RDMA_OPTION_ID_REUS= EADDR, - &reuse, sizeof reuse); - if (ret < 0) { - error_setg(errp, "RDMA ERROR: Error: could not set REUSEADDR optio= n"); - goto err_dest_init_bind_addr; - } - - /* Try all addresses, saving the first error in @err */ - for (e =3D res; e !=3D NULL; e =3D e->ai_next) { - Error **local_errp =3D err ? NULL : &err; - - inet_ntop(e->ai_family, - &((struct sockaddr_in *) e->ai_dst_addr)->sin_addr, ip, sizeof= ip); - trace_qemu_rdma_dest_init_trying(rdma->host, ip); - ret =3D rdma_bind_addr(listen_id, e->ai_dst_addr); - if (ret < 0) { - continue; - } - if (e->ai_family =3D=3D AF_INET6) { - ret =3D qemu_rdma_broken_ipv6_kernel(listen_id->verbs, - local_errp); - if (ret < 0) { - continue; - } - } - error_free(err); - break; - } - - rdma_freeaddrinfo(res); - if (!e) { - if (err) { - error_propagate(errp, err); - } else { - error_setg(errp, "RDMA ERROR: Error: could not rdma_bind_addr!= "); - } - goto err_dest_init_bind_addr; - } - - rdma->listen_id =3D listen_id; - qemu_rdma_dump_gid("dest_init", listen_id); - return 0; - -err_dest_init_bind_addr: - rdma_destroy_id(listen_id); -err_dest_init_create_listen_id: - rdma_destroy_event_channel(rdma->channel); - rdma->channel =3D NULL; - rdma->errored =3D true; - return -1; - -} - -static void qemu_rdma_return_path_dest_init(RDMAContext *rdma_return_path, - RDMAContext *rdma) -{ - for (int i =3D 0; i < RDMA_WRID_MAX; i++) { - rdma_return_path->wr_data[i].control_len =3D 0; - rdma_return_path->wr_data[i].control_curr =3D NULL; - } - - /*the CM channel and CM id is shared*/ - rdma_return_path->channel =3D rdma->channel; - rdma_return_path->listen_id =3D rdma->listen_id; - - rdma->return_path =3D rdma_return_path; - rdma_return_path->return_path =3D rdma; - rdma_return_path->is_return_path =3D true; -} - -static RDMAContext *qemu_rdma_data_init(InetSocketAddress *saddr, Error **= errp) -{ - RDMAContext *rdma =3D NULL; - - rdma =3D g_new0(RDMAContext, 1); - rdma->current_index =3D -1; - rdma->current_chunk =3D -1; - - rdma->host =3D g_strdup(saddr->host); - rdma->port =3D atoi(saddr->port); - return rdma; -} - -/* - * QEMUFile interface to the control channel. - * SEND messages for control only. - * VM's ram is handled with regular RDMA messages. - */ -static ssize_t qio_channel_rdma_writev(QIOChannel *ioc, - const struct iovec *iov, - size_t niov, - int *fds, - size_t nfds, - int flags, - Error **errp) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(ioc); - RDMAContext *rdma; - int ret; - ssize_t done =3D 0; - size_t len; - - RCU_READ_LOCK_GUARD(); - rdma =3D qatomic_rcu_read(&rioc->rdmaout); - - if (!rdma) { - error_setg(errp, "RDMA control channel output is not set"); - return -1; - } - - if (rdma->errored) { - error_setg(errp, - "RDMA is in an error state waiting migration to abort!"= ); - return -1; - } - - /* - * Push out any writes that - * we're queued up for VM's ram. - */ - ret =3D qemu_rdma_write_flush(rdma, errp); - if (ret < 0) { - rdma->errored =3D true; - return -1; - } - - for (int i =3D 0; i < niov; i++) { - size_t remaining =3D iov[i].iov_len; - uint8_t * data =3D (void *)iov[i].iov_base; - while (remaining) { - RDMAControlHeader head =3D {}; - - len =3D MIN(remaining, RDMA_SEND_INCREMENT); - remaining -=3D len; - - head.len =3D len; - head.type =3D RDMA_CONTROL_QEMU_FILE; - - ret =3D qemu_rdma_exchange_send(rdma, &head, - data, NULL, NULL, NULL, errp); - - if (ret < 0) { - rdma->errored =3D true; - return -1; - } - - data +=3D len; - done +=3D len; - } - } - - return done; -} - -static size_t qemu_rdma_fill(RDMAContext *rdma, uint8_t *buf, - size_t size, int idx) -{ - size_t len =3D 0; - - if (rdma->wr_data[idx].control_len) { - trace_qemu_rdma_fill(rdma->wr_data[idx].control_len, size); - - len =3D MIN(size, rdma->wr_data[idx].control_len); - memcpy(buf, rdma->wr_data[idx].control_curr, len); - rdma->wr_data[idx].control_curr +=3D len; - rdma->wr_data[idx].control_len -=3D len; - } - - return len; -} - -/* - * QEMUFile interface to the control channel. - * RDMA links don't use bytestreams, so we have to - * return bytes to QEMUFile opportunistically. - */ -static ssize_t qio_channel_rdma_readv(QIOChannel *ioc, - const struct iovec *iov, - size_t niov, - int **fds, - size_t *nfds, - int flags, - Error **errp) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(ioc); - RDMAContext *rdma; - RDMAControlHeader head; - int ret; - ssize_t done =3D 0; - size_t len; - - RCU_READ_LOCK_GUARD(); - rdma =3D qatomic_rcu_read(&rioc->rdmain); - - if (!rdma) { - error_setg(errp, "RDMA control channel input is not set"); - return -1; - } - - if (rdma->errored) { - error_setg(errp, - "RDMA is in an error state waiting migration to abort!"= ); - return -1; - } - - for (int i =3D 0; i < niov; i++) { - size_t want =3D iov[i].iov_len; - uint8_t *data =3D (void *)iov[i].iov_base; - - /* - * First, we hold on to the last SEND message we - * were given and dish out the bytes until we run - * out of bytes. - */ - len =3D qemu_rdma_fill(rdma, data, want, 0); - done +=3D len; - want -=3D len; - /* Got what we needed, so go to next iovec */ - if (want =3D=3D 0) { - continue; - } - - /* If we got any data so far, then don't wait - * for more, just return what we have */ - if (done > 0) { - break; - } - - - /* We've got nothing at all, so lets wait for - * more to arrive - */ - ret =3D qemu_rdma_exchange_recv(rdma, &head, RDMA_CONTROL_QEMU_FIL= E, - errp); - - if (ret < 0) { - rdma->errored =3D true; - return -1; - } - - /* - * SEND was received with new bytes, now try again. - */ - len =3D qemu_rdma_fill(rdma, data, want, 0); - done +=3D len; - want -=3D len; - - /* Still didn't get enough, so lets just return */ - if (want) { - if (done =3D=3D 0) { - return QIO_CHANNEL_ERR_BLOCK; - } else { - break; - } - } - } - return done; -} - -/* - * Block until all the outstanding chunks have been delivered by the hardw= are. - */ -static int qemu_rdma_drain_cq(RDMAContext *rdma) -{ - Error *err =3D NULL; - - if (qemu_rdma_write_flush(rdma, &err) < 0) { - error_report_err(err); - return -1; - } - - while (rdma->nb_sent) { - if (qemu_rdma_block_for_wrid(rdma, RDMA_WRID_RDMA_WRITE, NULL) < 0= ) { - error_report("rdma migration: complete polling error!"); - return -1; - } - } - - qemu_rdma_unregister_waiting(rdma); - - return 0; -} - - -static int qio_channel_rdma_set_blocking(QIOChannel *ioc, - bool blocking, - Error **errp) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(ioc); - /* XXX we should make readv/writev actually honour this :-) */ - rioc->blocking =3D blocking; - return 0; -} - - -typedef struct QIOChannelRDMASource QIOChannelRDMASource; -struct QIOChannelRDMASource { - GSource parent; - QIOChannelRDMA *rioc; - GIOCondition condition; -}; - -static gboolean -qio_channel_rdma_source_prepare(GSource *source, - gint *timeout) -{ - QIOChannelRDMASource *rsource =3D (QIOChannelRDMASource *)source; - RDMAContext *rdma; - GIOCondition cond =3D 0; - *timeout =3D -1; - - RCU_READ_LOCK_GUARD(); - if (rsource->condition =3D=3D G_IO_IN) { - rdma =3D qatomic_rcu_read(&rsource->rioc->rdmain); - } else { - rdma =3D qatomic_rcu_read(&rsource->rioc->rdmaout); - } - - if (!rdma) { - error_report("RDMAContext is NULL when prepare Gsource"); - return FALSE; - } - - if (rdma->wr_data[0].control_len) { - cond |=3D G_IO_IN; - } - cond |=3D G_IO_OUT; - - return cond & rsource->condition; -} - -static gboolean -qio_channel_rdma_source_check(GSource *source) -{ - QIOChannelRDMASource *rsource =3D (QIOChannelRDMASource *)source; - RDMAContext *rdma; - GIOCondition cond =3D 0; - - RCU_READ_LOCK_GUARD(); - if (rsource->condition =3D=3D G_IO_IN) { - rdma =3D qatomic_rcu_read(&rsource->rioc->rdmain); - } else { - rdma =3D qatomic_rcu_read(&rsource->rioc->rdmaout); - } - - if (!rdma) { - error_report("RDMAContext is NULL when check Gsource"); - return FALSE; - } - - if (rdma->wr_data[0].control_len) { - cond |=3D G_IO_IN; - } - cond |=3D G_IO_OUT; - - return cond & rsource->condition; -} - -static gboolean -qio_channel_rdma_source_dispatch(GSource *source, - GSourceFunc callback, - gpointer user_data) -{ - QIOChannelFunc func =3D (QIOChannelFunc)callback; - QIOChannelRDMASource *rsource =3D (QIOChannelRDMASource *)source; - RDMAContext *rdma; - GIOCondition cond =3D 0; - - RCU_READ_LOCK_GUARD(); - if (rsource->condition =3D=3D G_IO_IN) { - rdma =3D qatomic_rcu_read(&rsource->rioc->rdmain); - } else { - rdma =3D qatomic_rcu_read(&rsource->rioc->rdmaout); - } - - if (!rdma) { - error_report("RDMAContext is NULL when dispatch Gsource"); - return FALSE; - } - - if (rdma->wr_data[0].control_len) { - cond |=3D G_IO_IN; - } - cond |=3D G_IO_OUT; - - return (*func)(QIO_CHANNEL(rsource->rioc), - (cond & rsource->condition), - user_data); -} - -static void -qio_channel_rdma_source_finalize(GSource *source) -{ - QIOChannelRDMASource *ssource =3D (QIOChannelRDMASource *)source; - - object_unref(OBJECT(ssource->rioc)); -} - -static GSourceFuncs qio_channel_rdma_source_funcs =3D { - qio_channel_rdma_source_prepare, - qio_channel_rdma_source_check, - qio_channel_rdma_source_dispatch, - qio_channel_rdma_source_finalize -}; - -static GSource *qio_channel_rdma_create_watch(QIOChannel *ioc, - GIOCondition condition) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(ioc); - QIOChannelRDMASource *ssource; - GSource *source; - - source =3D g_source_new(&qio_channel_rdma_source_funcs, - sizeof(QIOChannelRDMASource)); - ssource =3D (QIOChannelRDMASource *)source; - - ssource->rioc =3D rioc; - object_ref(OBJECT(rioc)); - - ssource->condition =3D condition; - - return source; -} - -static void qio_channel_rdma_set_aio_fd_handler(QIOChannel *ioc, - AioContext *read_ctx, - IOHandler *io_read, - AioContext *write_ctx, - IOHandler *io_write, - void *opaque) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(ioc); - if (io_read) { - aio_set_fd_handler(read_ctx, rioc->rdmain->recv_comp_channel->fd, - io_read, io_write, NULL, NULL, opaque); - aio_set_fd_handler(read_ctx, rioc->rdmain->send_comp_channel->fd, - io_read, io_write, NULL, NULL, opaque); - } else { - aio_set_fd_handler(write_ctx, rioc->rdmaout->recv_comp_channel->fd, - io_read, io_write, NULL, NULL, opaque); - aio_set_fd_handler(write_ctx, rioc->rdmaout->send_comp_channel->fd, - io_read, io_write, NULL, NULL, opaque); - } -} - -struct rdma_close_rcu { - struct rcu_head rcu; - RDMAContext *rdmain; - RDMAContext *rdmaout; -}; - -/* callback from qio_channel_rdma_close via call_rcu */ -static void qio_channel_rdma_close_rcu(struct rdma_close_rcu *rcu) -{ - if (rcu->rdmain) { - qemu_rdma_cleanup(rcu->rdmain); - } - - if (rcu->rdmaout) { - qemu_rdma_cleanup(rcu->rdmaout); - } - - g_free(rcu->rdmain); - g_free(rcu->rdmaout); - g_free(rcu); -} - -static int qio_channel_rdma_close(QIOChannel *ioc, - Error **errp) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(ioc); - RDMAContext *rdmain, *rdmaout; - struct rdma_close_rcu *rcu =3D g_new(struct rdma_close_rcu, 1); - - trace_qemu_rdma_close(); - - rdmain =3D rioc->rdmain; - if (rdmain) { - qatomic_rcu_set(&rioc->rdmain, NULL); - } - - rdmaout =3D rioc->rdmaout; - if (rdmaout) { - qatomic_rcu_set(&rioc->rdmaout, NULL); - } - - rcu->rdmain =3D rdmain; - rcu->rdmaout =3D rdmaout; - call_rcu(rcu, qio_channel_rdma_close_rcu, rcu); - - return 0; -} - -static int -qio_channel_rdma_shutdown(QIOChannel *ioc, - QIOChannelShutdown how, - Error **errp) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(ioc); - RDMAContext *rdmain, *rdmaout; - - RCU_READ_LOCK_GUARD(); - - rdmain =3D qatomic_rcu_read(&rioc->rdmain); - rdmaout =3D qatomic_rcu_read(&rioc->rdmain); - - switch (how) { - case QIO_CHANNEL_SHUTDOWN_READ: - if (rdmain) { - rdmain->errored =3D true; - } - break; - case QIO_CHANNEL_SHUTDOWN_WRITE: - if (rdmaout) { - rdmaout->errored =3D true; - } - break; - case QIO_CHANNEL_SHUTDOWN_BOTH: - default: - if (rdmain) { - rdmain->errored =3D true; - } - if (rdmaout) { - rdmaout->errored =3D true; - } - break; - } - - return 0; -} - -/* - * Parameters: - * @offset =3D=3D 0 : - * This means that 'block_offset' is a full virtual address that do= es not - * belong to a RAMBlock of the virtual machine and instead - * represents a private malloc'd memory area that the caller wishes= to - * transfer. - * - * @offset !=3D 0 : - * Offset is an offset to be added to block_offset and used - * to also lookup the corresponding RAMBlock. - * - * @size : Number of bytes to transfer - * - * @pages_sent : User-specificed pointer to indicate how many pages were - * sent. Usually, this will not be more than a few bytes = of - * the protocol because most transfers are sent asynchron= ously. - */ -static int qemu_rdma_save_page(QEMUFile *f, ram_addr_t block_offset, - ram_addr_t offset, size_t size) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(qemu_file_get_ioc(f)); - Error *err =3D NULL; - RDMAContext *rdma; - int ret; - - RCU_READ_LOCK_GUARD(); - rdma =3D qatomic_rcu_read(&rioc->rdmaout); - - if (!rdma) { - return -1; - } - - if (rdma_errored(rdma)) { - return -1; - } - - qemu_fflush(f); - - /* - * Add this page to the current 'chunk'. If the chunk - * is full, or the page doesn't belong to the current chunk, - * an actual RDMA write will occur and a new chunk will be formed. - */ - ret =3D qemu_rdma_write(rdma, block_offset, offset, size, &err); - if (ret < 0) { - error_report_err(err); - goto err; - } - - /* - * Drain the Completion Queue if possible, but do not block, - * just poll. - * - * If nothing to poll, the end of the iteration will do this - * again to make sure we don't overflow the request queue. - */ - while (1) { - uint64_t wr_id, wr_id_in; - ret =3D qemu_rdma_poll(rdma, rdma->recv_cq, &wr_id_in, NULL); - - if (ret < 0) { - error_report("rdma migration: polling error"); - goto err; - } - - wr_id =3D wr_id_in & RDMA_WRID_TYPE_MASK; - - if (wr_id =3D=3D RDMA_WRID_NONE) { - break; - } - } - - while (1) { - uint64_t wr_id, wr_id_in; - ret =3D qemu_rdma_poll(rdma, rdma->send_cq, &wr_id_in, NULL); - - if (ret < 0) { - error_report("rdma migration: polling error"); - goto err; - } - - wr_id =3D wr_id_in & RDMA_WRID_TYPE_MASK; - - if (wr_id =3D=3D RDMA_WRID_NONE) { - break; - } - } - - return RAM_SAVE_CONTROL_DELAYED; - -err: - rdma->errored =3D true; - return -1; -} - -int rdma_control_save_page(QEMUFile *f, ram_addr_t block_offset, - ram_addr_t offset, size_t size) -{ - if (!migrate_rdma() || migration_in_postcopy()) { - return RAM_SAVE_CONTROL_NOT_SUPP; - } - - int ret =3D qemu_rdma_save_page(f, block_offset, offset, size); - - if (ret !=3D RAM_SAVE_CONTROL_DELAYED && - ret !=3D RAM_SAVE_CONTROL_NOT_SUPP) { - if (ret < 0) { - qemu_file_set_error(f, ret); - } - } - return ret; -} - -static void rdma_accept_incoming_migration(void *opaque); - -static void rdma_cm_poll_handler(void *opaque) -{ - RDMAContext *rdma =3D opaque; - struct rdma_cm_event *cm_event; - MigrationIncomingState *mis =3D migration_incoming_get_current(); - - if (rdma_get_cm_event(rdma->channel, &cm_event) < 0) { - error_report("get_cm_event failed %d", errno); - return; - } - - if (cm_event->event =3D=3D RDMA_CM_EVENT_DISCONNECTED || - cm_event->event =3D=3D RDMA_CM_EVENT_DEVICE_REMOVAL) { - if (!rdma->errored && - migration_incoming_get_current()->state !=3D - MIGRATION_STATUS_COMPLETED) { - error_report("receive cm event, cm event is %d", cm_event->eve= nt); - rdma->errored =3D true; - if (rdma->return_path) { - rdma->return_path->errored =3D true; - } - } - rdma_ack_cm_event(cm_event); - if (mis->loadvm_co) { - qemu_coroutine_enter(mis->loadvm_co); - } - return; - } - rdma_ack_cm_event(cm_event); -} - -static int qemu_rdma_accept(RDMAContext *rdma) -{ - Error *err =3D NULL; - RDMACapabilities cap; - struct rdma_conn_param conn_param =3D { - .responder_resources =3D 2, - .private_data =3D &cap, - .private_data_len =3D sizeof(c= ap), - }; - RDMAContext *rdma_return_path =3D NULL; - g_autoptr(InetSocketAddress) isock =3D g_new0(InetSocketAddress, 1); - struct rdma_cm_event *cm_event; - struct ibv_context *verbs; - int ret; - - ret =3D rdma_get_cm_event(rdma->channel, &cm_event); - if (ret < 0) { - goto err_rdma_dest_wait; - } - - if (cm_event->event !=3D RDMA_CM_EVENT_CONNECT_REQUEST) { - rdma_ack_cm_event(cm_event); - goto err_rdma_dest_wait; - } - - isock->host =3D g_strdup(rdma->host); - isock->port =3D g_strdup_printf("%d", rdma->port); - - /* - * initialize the RDMAContext for return path for postcopy after first - * connection request reached. - */ - if ((migrate_postcopy() || migrate_return_path()) - && !rdma->is_return_path) { - rdma_return_path =3D qemu_rdma_data_init(isock, NULL); - if (rdma_return_path =3D=3D NULL) { - rdma_ack_cm_event(cm_event); - goto err_rdma_dest_wait; - } - - qemu_rdma_return_path_dest_init(rdma_return_path, rdma); - } - - memcpy(&cap, cm_event->param.conn.private_data, sizeof(cap)); - - network_to_caps(&cap); - - if (cap.version < 1 || cap.version > RDMA_CONTROL_VERSION_CURRENT) { - error_report("Unknown source RDMA version: %d, bailing...", - cap.version); - rdma_ack_cm_event(cm_event); - goto err_rdma_dest_wait; - } - - /* - * Respond with only the capabilities this version of QEMU knows about. - */ - cap.flags &=3D known_capabilities; - - /* - * Enable the ones that we do know about. - * Add other checks here as new ones are introduced. - */ - if (cap.flags & RDMA_CAPABILITY_PIN_ALL) { - rdma->pin_all =3D true; - } - - rdma->cm_id =3D cm_event->id; - verbs =3D cm_event->id->verbs; - - rdma_ack_cm_event(cm_event); - - trace_qemu_rdma_accept_pin_state(rdma->pin_all); - - caps_to_network(&cap); - - trace_qemu_rdma_accept_pin_verbsc(verbs); - - if (!rdma->verbs) { - rdma->verbs =3D verbs; - } else if (rdma->verbs !=3D verbs) { - error_report("ibv context not matching %p, %p!", rdma->verbs, - verbs); - goto err_rdma_dest_wait; - } - - qemu_rdma_dump_id("dest_init", verbs); - - ret =3D qemu_rdma_alloc_pd_cq(rdma, &err); - if (ret < 0) { - error_report_err(err); - goto err_rdma_dest_wait; - } - - ret =3D qemu_rdma_alloc_qp(rdma); - if (ret < 0) { - error_report("rdma migration: error allocating qp!"); - goto err_rdma_dest_wait; - } - - qemu_rdma_init_ram_blocks(rdma); - - for (int i =3D 0; i < RDMA_WRID_MAX; i++) { - ret =3D qemu_rdma_reg_control(rdma, i); - if (ret < 0) { - error_report("rdma: error registering %d control", i); - goto err_rdma_dest_wait; - } - } - - /* Accept the second connection request for return path */ - if ((migrate_postcopy() || migrate_return_path()) - && !rdma->is_return_path) { - qemu_set_fd_handler(rdma->channel->fd, rdma_accept_incoming_migrat= ion, - NULL, - (void *)(intptr_t)rdma->return_path); - } else { - qemu_set_fd_handler(rdma->channel->fd, rdma_cm_poll_handler, - NULL, rdma); - } - - ret =3D rdma_accept(rdma->cm_id, &conn_param); - if (ret < 0) { - error_report("rdma_accept failed"); - goto err_rdma_dest_wait; - } - - ret =3D rdma_get_cm_event(rdma->channel, &cm_event); - if (ret < 0) { - error_report("rdma_accept get_cm_event failed"); - goto err_rdma_dest_wait; - } - - if (cm_event->event !=3D RDMA_CM_EVENT_ESTABLISHED) { - error_report("rdma_accept not event established"); - rdma_ack_cm_event(cm_event); - goto err_rdma_dest_wait; - } - - rdma_ack_cm_event(cm_event); - rdma->connected =3D true; - - ret =3D qemu_rdma_post_recv_control(rdma, RDMA_WRID_READY, &err); - if (ret < 0) { - error_report_err(err); - goto err_rdma_dest_wait; - } - - qemu_rdma_dump_gid("dest_connect", rdma->cm_id); - - return 0; - -err_rdma_dest_wait: - rdma->errored =3D true; - qemu_rdma_cleanup(rdma); - g_free(rdma_return_path); - return -1; -} - -static int dest_ram_sort_func(const void *a, const void *b) -{ - unsigned int a_index =3D ((const RDMALocalBlock *)a)->src_index; - unsigned int b_index =3D ((const RDMALocalBlock *)b)->src_index; - - return (a_index < b_index) ? -1 : (a_index !=3D b_index); -} - -/* - * During each iteration of the migration, we listen for instructions - * by the source VM to perform dynamic page registrations before they - * can perform RDMA operations. - * - * We respond with the 'rkey'. - * - * Keep doing this until the source tells us to stop. - */ -int rdma_registration_handle(QEMUFile *f) -{ - RDMAControlHeader reg_resp =3D { .len =3D sizeof(RDMARegisterResult), - .type =3D RDMA_CONTROL_REGISTER_RESULT, - .repeat =3D 0, - }; - RDMAControlHeader unreg_resp =3D { .len =3D 0, - .type =3D RDMA_CONTROL_UNREGISTER_FINISHED, - .repeat =3D 0, - }; - RDMAControlHeader blocks =3D { .type =3D RDMA_CONTROL_RAM_BLOCKS_RESUL= T, - .repeat =3D 1 }; - QIOChannelRDMA *rioc; - Error *err =3D NULL; - RDMAContext *rdma; - RDMALocalBlocks *local; - RDMAControlHeader head; - RDMARegister *reg, *registers; - RDMACompress *comp; - RDMARegisterResult *reg_result; - static RDMARegisterResult results[RDMA_CONTROL_MAX_COMMANDS_PER_MESSAG= E]; - RDMALocalBlock *block; - void *host_addr; - int ret; - int idx =3D 0; - - if (!migrate_rdma()) { - return 0; - } - - RCU_READ_LOCK_GUARD(); - rioc =3D QIO_CHANNEL_RDMA(qemu_file_get_ioc(f)); - rdma =3D qatomic_rcu_read(&rioc->rdmain); - - if (!rdma) { - return -1; - } - - if (rdma_errored(rdma)) { - return -1; - } - - local =3D &rdma->local_ram_blocks; - do { - trace_rdma_registration_handle_wait(); - - ret =3D qemu_rdma_exchange_recv(rdma, &head, RDMA_CONTROL_NONE, &e= rr); - - if (ret < 0) { - error_report_err(err); - break; - } - - if (head.repeat > RDMA_CONTROL_MAX_COMMANDS_PER_MESSAGE) { - error_report("rdma: Too many requests in this message (%d)." - "Bailing.", head.repeat); - break; - } - - switch (head.type) { - case RDMA_CONTROL_COMPRESS: - comp =3D (RDMACompress *) rdma->wr_data[idx].control_curr; - network_to_compress(comp); - - trace_rdma_registration_handle_compress(comp->length, - comp->block_idx, - comp->offset); - if (comp->block_idx >=3D rdma->local_ram_blocks.nb_blocks) { - error_report("rdma: 'compress' bad block index %u (vs %d)", - (unsigned int)comp->block_idx, - rdma->local_ram_blocks.nb_blocks); - goto err; - } - block =3D &(rdma->local_ram_blocks.block[comp->block_idx]); - - host_addr =3D block->local_host_addr + - (comp->offset - block->offset); - if (comp->value) { - error_report("rdma: Zero page with non-zero (%d) value", - comp->value); - goto err; - } - ram_handle_zero(host_addr, comp->length); - break; - - case RDMA_CONTROL_REGISTER_FINISHED: - trace_rdma_registration_handle_finished(); - return 0; - - case RDMA_CONTROL_RAM_BLOCKS_REQUEST: - trace_rdma_registration_handle_ram_blocks(); - - /* Sort our local RAM Block list so it's the same as the sourc= e, - * we can do this since we've filled in a src_index in the list - * as we received the RAMBlock list earlier. - */ - qsort(rdma->local_ram_blocks.block, - rdma->local_ram_blocks.nb_blocks, - sizeof(RDMALocalBlock), dest_ram_sort_func); - for (int i =3D 0; i < local->nb_blocks; i++) { - local->block[i].index =3D i; - } - - if (rdma->pin_all) { - ret =3D qemu_rdma_reg_whole_ram_blocks(rdma, &err); - if (ret < 0) { - error_report_err(err); - goto err; - } - } - - /* - * Dest uses this to prepare to transmit the RAMBlock descript= ions - * to the source VM after connection setup. - * Both sides use the "remote" structure to communicate and up= date - * their "local" descriptions with what was sent. - */ - for (int i =3D 0; i < local->nb_blocks; i++) { - rdma->dest_blocks[i].remote_host_addr =3D - (uintptr_t)(local->block[i].local_host_addr); - - if (rdma->pin_all) { - rdma->dest_blocks[i].remote_rkey =3D local->block[i].m= r->rkey; - } - - rdma->dest_blocks[i].offset =3D local->block[i].offset; - rdma->dest_blocks[i].length =3D local->block[i].length; - - dest_block_to_network(&rdma->dest_blocks[i]); - trace_rdma_registration_handle_ram_blocks_loop( - local->block[i].block_name, - local->block[i].offset, - local->block[i].length, - local->block[i].local_host_addr, - local->block[i].src_index); - } - - blocks.len =3D rdma->local_ram_blocks.nb_blocks - * sizeof(RDMADestBlock); - - - ret =3D qemu_rdma_post_send_control(rdma, - (uint8_t *) rdma->dest_blocks, &blocks, - &err); - - if (ret < 0) { - error_report_err(err); - goto err; - } - - break; - case RDMA_CONTROL_REGISTER_REQUEST: - trace_rdma_registration_handle_register(head.repeat); - - reg_resp.repeat =3D head.repeat; - registers =3D (RDMARegister *) rdma->wr_data[idx].control_curr; - - for (int count =3D 0; count < head.repeat; count++) { - uint64_t chunk; - uint8_t *chunk_start, *chunk_end; - - reg =3D ®isters[count]; - network_to_register(reg); - - reg_result =3D &results[count]; - - trace_rdma_registration_handle_register_loop(count, - reg->current_index, reg->key.current_addr, reg->c= hunks); - - if (reg->current_index >=3D rdma->local_ram_blocks.nb_bloc= ks) { - error_report("rdma: 'register' bad block index %u (vs = %d)", - (unsigned int)reg->current_index, - rdma->local_ram_blocks.nb_blocks); - goto err; - } - block =3D &(rdma->local_ram_blocks.block[reg->current_inde= x]); - if (block->is_ram_block) { - if (block->offset > reg->key.current_addr) { - error_report("rdma: bad register address for block= %s" - " offset: %" PRIx64 " current_addr: %" PRIx64, - block->block_name, block->offset, - reg->key.current_addr); - goto err; - } - host_addr =3D (block->local_host_addr + - (reg->key.current_addr - block->offset)); - chunk =3D ram_chunk_index(block->local_host_addr, - (uint8_t *) host_addr); - } else { - chunk =3D reg->key.chunk; - host_addr =3D block->local_host_addr + - (reg->key.chunk * (1UL << RDMA_REG_CHUNK_SHIFT)); - /* Check for particularly bad chunk value */ - if (host_addr < (void *)block->local_host_addr) { - error_report("rdma: bad chunk for block %s" - " chunk: %" PRIx64, - block->block_name, reg->key.chunk); - goto err; - } - } - chunk_start =3D ram_chunk_start(block, chunk); - chunk_end =3D ram_chunk_end(block, chunk + reg->chunks); - /* avoid "-Waddress-of-packed-member" warning */ - uint32_t tmp_rkey =3D 0; - if (qemu_rdma_register_and_get_keys(rdma, block, - (uintptr_t)host_addr, NULL, &tmp_rkey, - chunk, chunk_start, chunk_end)) { - error_report("cannot get rkey"); - goto err; - } - reg_result->rkey =3D tmp_rkey; - - reg_result->host_addr =3D (uintptr_t)block->local_host_add= r; - - trace_rdma_registration_handle_register_rkey(reg_result->r= key); - - result_to_network(reg_result); - } - - ret =3D qemu_rdma_post_send_control(rdma, - (uint8_t *) results, ®_resp, &err); - - if (ret < 0) { - error_report_err(err); - goto err; - } - break; - case RDMA_CONTROL_UNREGISTER_REQUEST: - trace_rdma_registration_handle_unregister(head.repeat); - unreg_resp.repeat =3D head.repeat; - registers =3D (RDMARegister *) rdma->wr_data[idx].control_curr; - - for (int count =3D 0; count < head.repeat; count++) { - reg =3D ®isters[count]; - network_to_register(reg); - - trace_rdma_registration_handle_unregister_loop(count, - reg->current_index, reg->key.chunk); - - block =3D &(rdma->local_ram_blocks.block[reg->current_inde= x]); - - ret =3D ibv_dereg_mr(block->pmr[reg->key.chunk]); - block->pmr[reg->key.chunk] =3D NULL; - - if (ret !=3D 0) { - error_report("rdma unregistration chunk failed: %s", - strerror(errno)); - goto err; - } - - rdma->total_registrations--; - - trace_rdma_registration_handle_unregister_success(reg->key= .chunk); - } - - ret =3D qemu_rdma_post_send_control(rdma, NULL, &unreg_resp, &= err); - - if (ret < 0) { - error_report_err(err); - goto err; - } - break; - case RDMA_CONTROL_REGISTER_RESULT: - error_report("Invalid RESULT message at dest."); - goto err; - default: - error_report("Unknown control message %s", control_desc(head.t= ype)); - goto err; - } - } while (1); - -err: - rdma->errored =3D true; - return -1; -} - -/* Destination: - * Called during the initial RAM load section which lists the - * RAMBlocks by name. This lets us know the order of the RAMBlocks on - * the source. We've already built our local RAMBlock list, but not - * yet sent the list to the source. - */ -int rdma_block_notification_handle(QEMUFile *f, const char *name) -{ - int curr; - int found =3D -1; - - if (!migrate_rdma()) { - return 0; - } - - RCU_READ_LOCK_GUARD(); - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(qemu_file_get_ioc(f)); - RDMAContext *rdma =3D qatomic_rcu_read(&rioc->rdmain); - - if (!rdma) { - return -1; - } - - /* Find the matching RAMBlock in our local list */ - for (curr =3D 0; curr < rdma->local_ram_blocks.nb_blocks; curr++) { - if (!strcmp(rdma->local_ram_blocks.block[curr].block_name, name)) { - found =3D curr; - break; - } - } - - if (found =3D=3D -1) { - error_report("RAMBlock '%s' not found on destination", name); - return -1; - } - - rdma->local_ram_blocks.block[curr].src_index =3D rdma->next_src_index; - trace_rdma_block_notification_handle(name, rdma->next_src_index); - rdma->next_src_index++; - - return 0; -} - -int rdma_registration_start(QEMUFile *f, uint64_t flags) -{ - if (!migrate_rdma() || migration_in_postcopy()) { - return 0; - } - - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(qemu_file_get_ioc(f)); - RCU_READ_LOCK_GUARD(); - RDMAContext *rdma =3D qatomic_rcu_read(&rioc->rdmaout); - if (!rdma) { - return -1; - } - - if (rdma_errored(rdma)) { - return -1; - } - - trace_rdma_registration_start(flags); - qemu_put_be64(f, RAM_SAVE_FLAG_HOOK); - return qemu_fflush(f); -} - -/* - * Inform dest that dynamic registrations are done for now. - * First, flush writes, if any. - */ -int rdma_registration_stop(QEMUFile *f, uint64_t flags) -{ - QIOChannelRDMA *rioc; - Error *err =3D NULL; - RDMAContext *rdma; - RDMAControlHeader head =3D { .len =3D 0, .repeat =3D 1 }; - int ret; - - if (!migrate_rdma() || migration_in_postcopy()) { - return 0; - } - - RCU_READ_LOCK_GUARD(); - rioc =3D QIO_CHANNEL_RDMA(qemu_file_get_ioc(f)); - rdma =3D qatomic_rcu_read(&rioc->rdmaout); - if (!rdma) { - return -1; - } - - if (rdma_errored(rdma)) { - return -1; - } - - qemu_fflush(f); - ret =3D qemu_rdma_drain_cq(rdma); - - if (ret < 0) { - goto err; - } - - if (flags =3D=3D RAM_CONTROL_SETUP) { - RDMAControlHeader resp =3D {.type =3D RDMA_CONTROL_RAM_BLOCKS_RESU= LT }; - RDMALocalBlocks *local =3D &rdma->local_ram_blocks; - int reg_result_idx, nb_dest_blocks; - - head.type =3D RDMA_CONTROL_RAM_BLOCKS_REQUEST; - trace_rdma_registration_stop_ram(); - - /* - * Make sure that we parallelize the pinning on both sides. - * For very large guests, doing this serially takes a really - * long time, so we have to 'interleave' the pinning locally - * with the control messages by performing the pinning on this - * side before we receive the control response from the other - * side that the pinning has completed. - */ - ret =3D qemu_rdma_exchange_send(rdma, &head, NULL, &resp, - ®_result_idx, rdma->pin_all ? - qemu_rdma_reg_whole_ram_blocks : NULL, - &err); - if (ret < 0) { - error_report_err(err); - return -1; - } - - nb_dest_blocks =3D resp.len / sizeof(RDMADestBlock); - - /* - * The protocol uses two different sets of rkeys (mutually exclusi= ve): - * 1. One key to represent the virtual address of the entire ram b= lock. - * (dynamic chunk registration disabled - pin everything with o= ne rkey.) - * 2. One to represent individual chunks within a ram block. - * (dynamic chunk registration enabled - pin individual chunks.) - * - * Once the capability is successfully negotiated, the destination= transmits - * the keys to use (or sends them later) including the virtual add= resses - * and then propagates the remote ram block descriptions to his lo= cal copy. - */ - - if (local->nb_blocks !=3D nb_dest_blocks) { - error_report("ram blocks mismatch (Number of blocks %d vs %d)", - local->nb_blocks, nb_dest_blocks); - error_printf("Your QEMU command line parameters are probably " - "not identical on both the source and destination= ."); - rdma->errored =3D true; - return -1; - } - - qemu_rdma_move_header(rdma, reg_result_idx, &resp); - memcpy(rdma->dest_blocks, - rdma->wr_data[reg_result_idx].control_curr, resp.len); - for (int i =3D 0; i < nb_dest_blocks; i++) { - network_to_dest_block(&rdma->dest_blocks[i]); - - /* We require that the blocks are in the same order */ - if (rdma->dest_blocks[i].length !=3D local->block[i].length) { - error_report("Block %s/%d has a different length %" PRIu64 - "vs %" PRIu64, - local->block[i].block_name, i, - local->block[i].length, - rdma->dest_blocks[i].length); - rdma->errored =3D true; - return -1; - } - local->block[i].remote_host_addr =3D - rdma->dest_blocks[i].remote_host_addr; - local->block[i].remote_rkey =3D rdma->dest_blocks[i].remote_rk= ey; - } - } - - trace_rdma_registration_stop(flags); - - head.type =3D RDMA_CONTROL_REGISTER_FINISHED; - ret =3D qemu_rdma_exchange_send(rdma, &head, NULL, NULL, NULL, NULL, &= err); - - if (ret < 0) { - error_report_err(err); - goto err; - } - - return 0; -err: - rdma->errored =3D true; - return -1; -} - -static void qio_channel_rdma_finalize(Object *obj) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(obj); - if (rioc->rdmain) { - qemu_rdma_cleanup(rioc->rdmain); - g_free(rioc->rdmain); - rioc->rdmain =3D NULL; - } - if (rioc->rdmaout) { - qemu_rdma_cleanup(rioc->rdmaout); - g_free(rioc->rdmaout); - rioc->rdmaout =3D NULL; - } -} - -static void qio_channel_rdma_class_init(ObjectClass *klass, - void *class_data G_GNUC_UNUSED) -{ - QIOChannelClass *ioc_klass =3D QIO_CHANNEL_CLASS(klass); - - ioc_klass->io_writev =3D qio_channel_rdma_writev; - ioc_klass->io_readv =3D qio_channel_rdma_readv; - ioc_klass->io_set_blocking =3D qio_channel_rdma_set_blocking; - ioc_klass->io_close =3D qio_channel_rdma_close; - ioc_klass->io_create_watch =3D qio_channel_rdma_create_watch; - ioc_klass->io_set_aio_fd_handler =3D qio_channel_rdma_set_aio_fd_handl= er; - ioc_klass->io_shutdown =3D qio_channel_rdma_shutdown; -} - -static const TypeInfo qio_channel_rdma_info =3D { - .parent =3D TYPE_QIO_CHANNEL, - .name =3D TYPE_QIO_CHANNEL_RDMA, - .instance_size =3D sizeof(QIOChannelRDMA), - .instance_finalize =3D qio_channel_rdma_finalize, - .class_init =3D qio_channel_rdma_class_init, -}; - -static void qio_channel_rdma_register_types(void) -{ - type_register_static(&qio_channel_rdma_info); -} - -type_init(qio_channel_rdma_register_types); - -static QEMUFile *rdma_new_input(RDMAContext *rdma) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(object_new(TYPE_QIO_CHANNEL_= RDMA)); - - rioc->file =3D qemu_file_new_input(QIO_CHANNEL(rioc)); - rioc->rdmain =3D rdma; - rioc->rdmaout =3D rdma->return_path; - - return rioc->file; -} - -static QEMUFile *rdma_new_output(RDMAContext *rdma) -{ - QIOChannelRDMA *rioc =3D QIO_CHANNEL_RDMA(object_new(TYPE_QIO_CHANNEL_= RDMA)); - - rioc->file =3D qemu_file_new_output(QIO_CHANNEL(rioc)); - rioc->rdmaout =3D rdma; - rioc->rdmain =3D rdma->return_path; - - return rioc->file; -} - -static void rdma_accept_incoming_migration(void *opaque) -{ - RDMAContext *rdma =3D opaque; - QEMUFile *f; - - trace_qemu_rdma_accept_incoming_migration(); - if (qemu_rdma_accept(rdma) < 0) { - error_report("RDMA ERROR: Migration initialization failed"); - return; - } - - trace_qemu_rdma_accept_incoming_migration_accepted(); - - if (rdma->is_return_path) { - return; - } - - f =3D rdma_new_input(rdma); - if (f =3D=3D NULL) { - error_report("RDMA ERROR: could not open RDMA for input"); - qemu_rdma_cleanup(rdma); - return; - } - - rdma->migration_started_on_destination =3D 1; - migration_fd_process_incoming(f); -} - -void rdma_start_incoming_migration(InetSocketAddress *host_port, - Error **errp) -{ - MigrationState *s =3D migrate_get_current(); - int ret; - RDMAContext *rdma; - - trace_rdma_start_incoming_migration(); - - /* Avoid ram_block_discard_disable(), cannot change during migration. = */ - if (ram_block_discard_is_required()) { - error_setg(errp, "RDMA: cannot disable RAM discard"); - return; - } - - rdma =3D qemu_rdma_data_init(host_port, errp); - if (rdma =3D=3D NULL) { - goto err; - } - - ret =3D qemu_rdma_dest_init(rdma, errp); - if (ret < 0) { - goto err; - } - - trace_rdma_start_incoming_migration_after_dest_init(); - - ret =3D rdma_listen(rdma->listen_id, 5); - - if (ret < 0) { - error_setg(errp, "RDMA ERROR: listening on socket!"); - goto cleanup_rdma; - } - - trace_rdma_start_incoming_migration_after_rdma_listen(); - s->rdma_migration =3D true; - qemu_set_fd_handler(rdma->channel->fd, rdma_accept_incoming_migration, - NULL, (void *)(intptr_t)rdma); - return; - -cleanup_rdma: - qemu_rdma_cleanup(rdma); -err: - if (rdma) { - g_free(rdma->host); - } - g_free(rdma); -} - -void rdma_start_outgoing_migration(void *opaque, - InetSocketAddress *host_port, Error **errp) -{ - MigrationState *s =3D opaque; - RDMAContext *rdma_return_path =3D NULL; - RDMAContext *rdma; - int ret; - - /* Avoid ram_block_discard_disable(), cannot change during migration. = */ - if (ram_block_discard_is_required()) { - error_setg(errp, "RDMA: cannot disable RAM discard"); - return; - } - - rdma =3D qemu_rdma_data_init(host_port, errp); - if (rdma =3D=3D NULL) { - goto err; - } - - ret =3D qemu_rdma_source_init(rdma, migrate_rdma_pin_all(), errp); - - if (ret < 0) { - goto err; - } - - trace_rdma_start_outgoing_migration_after_rdma_source_init(); - ret =3D qemu_rdma_connect(rdma, false, errp); - - if (ret < 0) { - goto err; - } - - /* RDMA postcopy need a separate queue pair for return path */ - if (migrate_postcopy() || migrate_return_path()) { - rdma_return_path =3D qemu_rdma_data_init(host_port, errp); - - if (rdma_return_path =3D=3D NULL) { - goto return_path_err; - } - - ret =3D qemu_rdma_source_init(rdma_return_path, - migrate_rdma_pin_all(), errp); - - if (ret < 0) { - goto return_path_err; - } - - ret =3D qemu_rdma_connect(rdma_return_path, true, errp); - - if (ret < 0) { - goto return_path_err; - } - - rdma->return_path =3D rdma_return_path; - rdma_return_path->return_path =3D rdma; - rdma_return_path->is_return_path =3D true; - } - - trace_rdma_start_outgoing_migration_after_rdma_connect(); - - s->to_dst_file =3D rdma_new_output(rdma); - s->rdma_migration =3D true; - migrate_fd_connect(s, NULL); - return; -return_path_err: - qemu_rdma_cleanup(rdma); -err: - g_free(rdma); - g_free(rdma_return_path); -} diff --git a/migration/savevm.c b/migration/savevm.c index 388d7af7cd..939d35d69e 100644 --- a/migration/savevm.c +++ b/migration/savevm.c @@ -2970,7 +2970,7 @@ int qemu_loadvm_state(QEMUFile *f) =20 /* We've got to be careful; if we don't read the data and just shut th= e fd * then the sender can error if we close while it's still sending. - * We also mustn't read data that isn't there; some transports (RDMA) + * We also mustn't read data that isn't there; some transports * will stall waiting for that data when the source has already closed. */ if (ret =3D=3D 0 && should_send_vmdesc()) { diff --git a/meson_options.txt b/meson_options.txt index b5c0bad9e7..79b69d4286 100644 --- a/meson_options.txt +++ b/meson_options.txt @@ -196,8 +196,6 @@ option('rbd', type : 'feature', value : 'auto', description: 'Ceph block device driver') option('opengl', type : 'feature', value : 'auto', description: 'OpenGL support') -option('rdma', type : 'feature', value : 'auto', - description: 'Enable RDMA-based migration') option('gtk', type : 'feature', value : 'auto', description: 'GTK+ user interface') option('sdl', type : 'feature', value : 'auto', diff --git a/migration/meson.build b/migration/meson.build index 1eeb915ff6..e2cd92c01f 100644 --- a/migration/meson.build +++ b/migration/meson.build @@ -36,7 +36,6 @@ if get_option('replication').allowed() system_ss.add(files('colo-failover.c', 'colo.c')) endif =20 -system_ss.add(when: rdma, if_true: files('rdma.c')) if get_option('live_block_migration').allowed() system_ss.add(files('block.c')) endif diff --git a/migration/trace-events b/migration/trace-events index f0e1cb80c7..7db3a5194f 100644 --- a/migration/trace-events +++ b/migration/trace-events @@ -193,7 +193,7 @@ process_incoming_migration_co_postcopy_end_main(void) "" postcopy_preempt_enabled(bool value) "%d" =20 # migration-stats -migration_transferred_bytes(uint64_t qemu_file, uint64_t multifd, uint64_t= rdma) "qemu_file %" PRIu64 " multifd %" PRIu64 " RDMA %" PRIu64 +migration_transferred_bytes(uint64_t qemu_file, uint64_t multifd) "qemu_fi= le %" PRIu64 " multifd %" PRIu64 =20 # channel.c migration_set_incoming_channel(void *ioc, const char *ioctype) "ioc=3D%p i= octype=3D%s" @@ -204,72 +204,6 @@ migrate_state_too_big(void) "" migrate_global_state_post_load(const char *state) "loaded state: %s" migrate_global_state_pre_save(const char *state) "saved state: %s" =20 -# rdma.c -qemu_rdma_accept_incoming_migration(void) "" -qemu_rdma_accept_incoming_migration_accepted(void) "" -qemu_rdma_accept_pin_state(bool pin) "%d" -qemu_rdma_accept_pin_verbsc(void *verbs) "Verbs context after listen: %p" -qemu_rdma_block_for_wrid_miss(uint64_t wcomp, uint64_t req) "A Wanted wrid= %" PRIu64 " but got %" PRIu64 -qemu_rdma_cleanup_disconnect(void) "" -qemu_rdma_close(void) "" -qemu_rdma_connect_pin_all_requested(void) "" -qemu_rdma_connect_pin_all_outcome(bool pin) "%d" -qemu_rdma_dest_init_trying(const char *host, const char *ip) "%s =3D> %s" -qemu_rdma_dump_id_failed(const char *who) "%s RDMA Device opened, but can'= t query port information" -qemu_rdma_dump_id(const char *who, const char *name, const char *dev_name,= const char *dev_path, const char *ibdev_path, int transport, const char *t= ransport_name) "%s RDMA Device opened: kernel name %s uverbs device name %s= , infiniband_verbs class device path %s, infiniband class device path %s, t= ransport: (%d) %s" -qemu_rdma_dump_gid(const char *who, const char *src, const char *dst) "%s = Source GID: %s, Dest GID: %s" -qemu_rdma_exchange_get_response_start(const char *desc) "CONTROL: %s recei= ving..." -qemu_rdma_exchange_get_response_none(const char *desc, int type) "Surprise= : got %s (%d)" -qemu_rdma_exchange_send_issue_callback(void) "" -qemu_rdma_exchange_send_waiting(const char *desc) "Waiting for response %s" -qemu_rdma_exchange_send_received(const char *desc) "Response %s received." -qemu_rdma_fill(size_t control_len, size_t size) "RDMA %zd of %zd bytes alr= eady in buffer" -qemu_rdma_init_ram_blocks(int blocks) "Allocated %d local ram block struct= ures" -qemu_rdma_poll_recv(uint64_t comp, int64_t id, int sent) "completion %" PR= Iu64 " received (%" PRId64 ") left %d" -qemu_rdma_poll_write(uint64_t comp, int left, uint64_t block, uint64_t chu= nk, void *local, void *remote) "completions %" PRIu64 " left %d, block %" P= RIu64 ", chunk: %" PRIu64 " %p %p" -qemu_rdma_poll_other(uint64_t comp, int left) "other completion %" PRIu64 = " received left %d" -qemu_rdma_post_send_control(const char *desc) "CONTROL: sending %s.." -qemu_rdma_register_and_get_keys(uint64_t len, void *start) "Registering %"= PRIu64 " bytes @ %p" -qemu_rdma_register_odp_mr(const char *name) "Try to register On-Demand Pag= ing memory region: %s" -qemu_rdma_advise_mr(const char *name, uint32_t len, uint64_t addr, const c= har *res) "Try to advise block %s prefetch at %" PRIu32 "@0x%" PRIx64 ": %s" -qemu_rdma_resolve_host_trying(const char *host, const char *ip) "Trying %s= =3D> %s" -qemu_rdma_signal_unregister_append(uint64_t chunk, int pos) "Appending unr= egister chunk %" PRIu64 " at position %d" -qemu_rdma_signal_unregister_already(uint64_t chunk) "Unregister chunk %" P= RIu64 " already in queue" -qemu_rdma_unregister_waiting_inflight(uint64_t chunk) "Cannot unregister i= nflight chunk: %" PRIu64 -qemu_rdma_unregister_waiting_proc(uint64_t chunk, int pos) "Processing unr= egister for chunk: %" PRIu64 " at position %d" -qemu_rdma_unregister_waiting_send(uint64_t chunk) "Sending unregister for = chunk: %" PRIu64 -qemu_rdma_unregister_waiting_complete(uint64_t chunk) "Unregister for chun= k: %" PRIu64 " complete." -qemu_rdma_write_flush(int sent) "sent total: %d" -qemu_rdma_write_one_block(int count, int block, uint64_t chunk, uint64_t c= urrent, uint64_t len, int nb_sent, int nb_chunks) "(%d) Not clobbering: blo= ck: %d chunk %" PRIu64 " current %" PRIu64 " len %" PRIu64 " %d %d" -qemu_rdma_write_one_post(uint64_t chunk, long addr, long remote, uint32_t = len) "Posting chunk: %" PRIu64 ", addr: 0x%lx remote: 0x%lx, bytes %" PRIu32 -qemu_rdma_write_one_queue_full(void) "" -qemu_rdma_write_one_recvregres(int mykey, int theirkey, uint64_t chunk) "R= eceived registration result: my key: 0x%x their key 0x%x, chunk %" PRIu64 -qemu_rdma_write_one_sendreg(uint64_t chunk, int len, int index, int64_t of= fset) "Sending registration request chunk %" PRIu64 " for %d bytes, index: = %d, offset: %" PRId64 -qemu_rdma_write_one_top(uint64_t chunks, uint64_t size) "Writing %" PRIu64= " chunks, (%" PRIu64 " MB)" -qemu_rdma_write_one_zero(uint64_t chunk, int len, int index, int64_t offse= t) "Entire chunk is zero, sending compress: %" PRIu64 " for %d bytes, index= : %d, offset: %" PRId64 -rdma_add_block(const char *block_name, int block, uint64_t addr, uint64_t = offset, uint64_t len, uint64_t end, uint64_t bits, int chunks) "Added Block= : '%s':%d, addr: %" PRIu64 ", offset: %" PRIu64 " length: %" PRIu64 " end: = %" PRIu64 " bits %" PRIu64 " chunks %d" -rdma_block_notification_handle(const char *name, int index) "%s at %d" -rdma_delete_block(void *block, uint64_t addr, uint64_t offset, uint64_t le= n, uint64_t end, uint64_t bits, int chunks) "Deleted Block: %p, addr: %" PR= Iu64 ", offset: %" PRIu64 " length: %" PRIu64 " end: %" PRIu64 " bits %" PR= Iu64 " chunks %d" -rdma_registration_handle_compress(int64_t length, int index, int64_t offse= t) "Zapping zero chunk: %" PRId64 " bytes, index %d, offset %" PRId64 -rdma_registration_handle_finished(void) "" -rdma_registration_handle_ram_blocks(void) "" -rdma_registration_handle_ram_blocks_loop(const char *name, uint64_t offset= , uint64_t length, void *local_host_addr, unsigned int src_index) "%s: @0x%= " PRIx64 "/%" PRIu64 " host:@%p src_index: %u" -rdma_registration_handle_register(int requests) "%d requests" -rdma_registration_handle_register_loop(int req, int index, uint64_t addr, = uint64_t chunks) "Registration request (%d): index %d, current_addr %" PRIu= 64 " chunks: %" PRIu64 -rdma_registration_handle_register_rkey(int rkey) "0x%x" -rdma_registration_handle_unregister(int requests) "%d requests" -rdma_registration_handle_unregister_loop(int count, int index, uint64_t ch= unk) "Unregistration request (%d): index %d, chunk %" PRIu64 -rdma_registration_handle_unregister_success(uint64_t chunk) "%" PRIu64 -rdma_registration_handle_wait(void) "" -rdma_registration_start(uint64_t flags) "%" PRIu64 -rdma_registration_stop(uint64_t flags) "%" PRIu64 -rdma_registration_stop_ram(void) "" -rdma_start_incoming_migration(void) "" -rdma_start_incoming_migration_after_dest_init(void) "" -rdma_start_incoming_migration_after_rdma_listen(void) "" -rdma_start_outgoing_migration_after_rdma_connect(void) "" -rdma_start_outgoing_migration_after_rdma_source_init(void) "" - # postcopy-ram.c postcopy_discard_send_finish(const char *ramblock, int nwords, int ncmds) = "%s mask words sent=3D%d in %d commands" postcopy_discard_send_range(const char *ramblock, unsigned long start, uns= igned long length) "%s:%lx/%lx" diff --git a/qemu-options.hx b/qemu-options.hx index f7ef9b4e41..4f390c33ef 100644 --- a/qemu-options.hx +++ b/qemu-options.hx @@ -4759,7 +4759,6 @@ ERST =20 DEF("incoming", HAS_ARG, QEMU_OPTION_incoming, \ "-incoming tcp:[host]:port[,to=3Dmaxport][,ipv4=3Don|off][,ipv6=3Don|o= ff]\n" \ - "-incoming rdma:host:port[,ipv4=3Don|off][,ipv6=3Don|off]\n" \ "-incoming unix:socketpath\n" \ " prepare for incoming migration, listen on\n" \ " specified protocol and socket address\n" \ @@ -4773,8 +4772,6 @@ DEF("incoming", HAS_ARG, QEMU_OPTION_incoming, \ QEMU_ARCH_ALL) SRST ``-incoming tcp:[host]:port[,to=3Dmaxport][,ipv4=3Don|off][,ipv6=3Don|off]= `` - \=20 -``-incoming rdma:host:port[,ipv4=3Don|off][,ipv6=3Don|off]`` Prepare for incoming migration, listen on a given tcp port. =20 ``-incoming unix:socketpath`` diff --git a/scripts/ci/org.centos/stream/8/build-environment.yml b/scripts= /ci/org.centos/stream/8/build-environment.yml index 1ead77e2cb..a366bb185b 100644 --- a/scripts/ci/org.centos/stream/8/build-environment.yml +++ b/scripts/ci/org.centos/stream/8/build-environment.yml @@ -68,7 +68,6 @@ - pixman-devel - python38 - python3-sphinx - - rdma-core-devel - redhat-rpm-config - snappy-devel - spice-glib-devel diff --git a/scripts/ci/org.centos/stream/8/x86_64/configure b/scripts/ci/o= rg.centos/stream/8/x86_64/configure index 868db665f6..5dead834fb 100755 --- a/scripts/ci/org.centos/stream/8/x86_64/configure +++ b/scripts/ci/org.centos/stream/8/x86_64/configure @@ -103,7 +103,6 @@ --disable-qed \ --disable-qom-cast-debug \ --disable-rbd \ ---disable-rdma \ --disable-replication \ --disable-rng-none \ --disable-safe-stack \ @@ -175,7 +174,6 @@ --enable-opengl \ --enable-pie \ --enable-rbd \ ---enable-rdma \ --enable-seccomp \ --enable-snappy \ --enable-smartcard \ diff --git a/scripts/ci/setup/build-environment.yml b/scripts/ci/setup/buil= d-environment.yml index f344d1a850..0359b1c023 100644 --- a/scripts/ci/setup/build-environment.yml +++ b/scripts/ci/setup/build-environment.yml @@ -81,8 +81,6 @@ - libglusterfs-dev - libgnutls28-dev - libgtk-3-dev - - libibumad-dev - - libibverbs-dev - libiscsi-dev - libjemalloc-dev - libjpeg-turbo8-dev @@ -99,7 +97,6 @@ - libpng-dev - libpulse-dev - librbd-dev - - librdmacm-dev - libsasl2-dev - libsdl2-dev - libsdl2-image-dev @@ -236,7 +233,6 @@ - pixman-devel - python38 - python3-sphinx - - rdma-core-devel - redhat-rpm-config - snappy-devel - spice-glib-devel diff --git a/scripts/coverity-scan/run-coverity-scan b/scripts/coverity-sca= n/run-coverity-scan index 43cf770f5e..3dd14c3cc4 100755 --- a/scripts/coverity-scan/run-coverity-scan +++ b/scripts/coverity-scan/run-coverity-scan @@ -426,7 +426,7 @@ echo "Configuring..." --enable-libusb --enable-usb-redir \ --enable-libiscsi --enable-libnfs --enable-seccomp \ --enable-tpm --enable-libssh --enable-lzo --enable-snappy --enable-bzi= p2 \ - --enable-numa --enable-rdma --enable-smartcard --enable-virglrenderer \ + --enable-numa --enable-smartcard --enable-virglrenderer \ --enable-mpath --enable-glusterfs \ --enable-virtfs --enable-zstd =20 diff --git a/scripts/meson-buildoptions.sh b/scripts/meson-buildoptions.sh index 5ace33f167..52c34598ba 100644 --- a/scripts/meson-buildoptions.sh +++ b/scripts/meson-buildoptions.sh @@ -167,7 +167,6 @@ meson_options_help() { printf "%s\n" ' qed qed image format support' printf "%s\n" ' qga-vss build QGA VSS support (broken with MinG= W)' printf "%s\n" ' rbd Ceph block device driver' - printf "%s\n" ' rdma Enable RDMA-based migration' printf "%s\n" ' replication replication support' printf "%s\n" ' rutabaga-gfx rutabaga_gfx support' printf "%s\n" ' sdl SDL user interface' @@ -442,8 +441,6 @@ _meson_option_parse() { --disable-qom-cast-debug) printf "%s" -Dqom_cast_debug=3Dfalse ;; --enable-rbd) printf "%s" -Drbd=3Denabled ;; --disable-rbd) printf "%s" -Drbd=3Ddisabled ;; - --enable-rdma) printf "%s" -Drdma=3Denabled ;; - --disable-rdma) printf "%s" -Drdma=3Ddisabled ;; --enable-relocatable) printf "%s" -Drelocatable=3Dtrue ;; --disable-relocatable) printf "%s" -Drelocatable=3Dfalse ;; --enable-replication) printf "%s" -Dreplication=3Denabled ;; diff --git a/tests/lcitool/projects/qemu.yml b/tests/lcitool/projects/qemu.= yml index 149b15de57..511e48a5ec 100644 --- a/tests/lcitool/projects/qemu.yml +++ b/tests/lcitool/projects/qemu.yml @@ -48,8 +48,6 @@ packages: - libfdt - libffi - libgcrypt - - libibumad - - libibverbs - libiscsi - libjemalloc - libjpeg @@ -58,7 +56,6 @@ packages: - libpmem - libpng - librbd - - librdmacm - libseccomp - libselinux - libslirp diff --git a/tests/migration/guestperf/engine.py b/tests/migration/guestper= f/engine.py index 608d7270f6..a704419082 100644 --- a/tests/migration/guestperf/engine.py +++ b/tests/migration/guestperf/engine.py @@ -41,7 +41,7 @@ def __init__(self, binary, dst_host, kernel, initrd, tran= sport=3D"tcp", self._dst_host =3D dst_host # Hostname of target host self._kernel =3D kernel # Path to kernel image self._initrd =3D initrd # Path to stress initrd - self._transport =3D transport # 'unix' or 'tcp' or 'rdma' + self._transport =3D transport # 'unix' or 'tcp' self._sleep =3D sleep self._verbose =3D verbose self._debug =3D debug @@ -427,8 +427,6 @@ def run(self, hardware, scenario, result_dir=3Dos.getcw= d()): =20 if self._transport =3D=3D "tcp": uri =3D "tcp:%s:9000" % self._dst_host - elif self._transport =3D=3D "rdma": - uri =3D "rdma:%s:9000" % self._dst_host elif self._transport =3D=3D "unix": if self._dst_host !=3D "localhost": raise Exception("Running use unix migration transport for = non-local host") --=20 2.41.0 From nobody Sun Nov 10 21:34:25 2024 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass(p=none dis=none) header.from=linaro.org ARC-Seal: i=1; a=rsa-sha256; t=1711631099; cv=none; d=zohomail.com; s=zohoarc; b=lOdUf0L2t8ysiHu5kP0AjsYU7HTM7NMsmtqvQ/IP7bHYIQ/3UKdpXl5RXDcSGPkp+dB/akLlJ4+8FzQBLKjnTANa8g+ga1zdnBNXXFoEvd4G0+rEzApE2AkNAwJdG+RJQzZbwdSAC4jDnaA352jFBhOpEVhZp14SN8v3MVe2iDE= ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com; s=zohoarc; t=1711631099; h=Content-Type:Content-Transfer-Encoding:Cc:Cc:Date:Date:From:From:In-Reply-To:List-Subscribe:List-Post:List-Id:List-Archive:List-Help:List-Unsubscribe:MIME-Version:Message-ID:References:Sender:Subject:Subject:To:To:Message-Id:Reply-To; bh=q2x43f4g/1k9PawbCTBkD3eBnTaZfrxytFTBDY8Dk3g=; b=EsFrehB94uVCp1R1C2iy2EVsOULelBdOAXee5OnEInSXO2HmOb+Y6Yj59Wzi+Ygot2hl+AJH+92re/GXrY9wtdsblMYMPUNBQMN1R1VbeAxKagRtVcqHMpKo1Zui+mRYnIF7aLdsWhz1kHDwnxnVbaN5ZZvza42oj9TyunRfH0o= ARC-Authentication-Results: i=1; mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass header.from= (p=none dis=none) Return-Path: Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 1711631099593554.1648742931851; Thu, 28 Mar 2024 06:04:59 -0700 (PDT) Received: from localhost ([::1] helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1rppQK-0000vG-9g; Thu, 28 Mar 2024 09:04:04 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1rppPp-0000Ld-NC for qemu-devel@nongnu.org; Thu, 28 Mar 2024 09:03:34 -0400 Received: from mail-ej1-x62e.google.com ([2a00:1450:4864:20::62e]) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1rppPn-00089O-Q7 for qemu-devel@nongnu.org; Thu, 28 Mar 2024 09:03:33 -0400 Received: by mail-ej1-x62e.google.com with SMTP id a640c23a62f3a-a4e1742d3a3so113099766b.1 for ; Thu, 28 Mar 2024 06:03:31 -0700 (PDT) Received: from m1x-phil.lan (pas38-h02-176-184-5-52.dsl.sta.abo.bbox.fr. [176.184.5.52]) by smtp.gmail.com with ESMTPSA id gy16-20020a170906f25000b00a45f2dc6795sm724813ejb.137.2024.03.28.06.03.27 (version=TLS1_3 cipher=TLS_CHACHA20_POLY1305_SHA256 bits=256/256); Thu, 28 Mar 2024 06:03:29 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; t=1711631010; x=1712235810; darn=nongnu.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=q2x43f4g/1k9PawbCTBkD3eBnTaZfrxytFTBDY8Dk3g=; b=muj/Opuli31SzElOqB3E6eX5fW/swQMoSoK/5mRc+u8frnmG+83Z7pi/xdwMhC7N77 3AAZmVKIGwTG0TnhZ1QZ0Nw4cSVuEbPRYmf1IqfrUTxQQSeXiME9E6QRpeoEZuslnO2q UjK4Mf3/z+qmgpQ2jF1ZolK1HuiJ1tJEcN6rhD5GgMSkFCbG+eLzrQ+lmZxB5k0sCqye Swp01qKq2qMCvqQjp0camaiRZ4WA+ZsjP7eXOHOr7wyrzRJ4txpyBvjeMBLqx1TPTIr5 rFATVoPfdV6ZpxKglnezjZGoF0C0tKKxSGBey0To5HfBJ5PSW40toof34kS7gQxudvch KDaw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1711631010; x=1712235810; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=q2x43f4g/1k9PawbCTBkD3eBnTaZfrxytFTBDY8Dk3g=; b=SfKkeCU/r49ioC4bBdXKX82FHhyGwsjW/AkY+r1lfGObvOTjZKQw14GoaMTq21hs6S kcRRRZpfIVpBh1Bc4RBnFREsiwaEt1NCJr3g8wubTbgLnwXGFdi3D2cLoK54YmrD2lW4 faNVi3GCadRfw3S2w+XSrLiHwcAgImSfBorqtUyKPO967KoEsHwNi4RKRT6etTg12/1z Sc2BqPG0PD9T9AqeBB94ghpMJQ3pH50jF1v2K5yThP8fcKHYCxz6B7+CCCljFUoEtzgJ h8IJrmiWT9oTMTe1SrxkBCnPuZKHMVAEUejyQENiWGbCmUDqnsx6+X0MI71QYJyIN0dJ r38Q== X-Gm-Message-State: AOJu0YwKINmEq25pvMtN8oHkoaNDobSP5uiTRYp8ZN98PAF7U1yw78P/ JARDDjMdG6t7fqAy272O56JY11/d8HuiSBtY01UPAcu90GHtuZTuMjgoSCLcZJ8XdCTDgxJjMck e X-Google-Smtp-Source: AGHT+IHCYFP2w6Zft+pz6ZBPav1fQP9UvRGriEZ8B33t9g1P3xdXeQrCxLbtekiiNs4Z5kHrqq7YLw== X-Received: by 2002:a17:906:29d1:b0:a47:32b3:18c5 with SMTP id y17-20020a17090629d100b00a4732b318c5mr1695572eje.68.1711631009605; Thu, 28 Mar 2024 06:03:29 -0700 (PDT) From: =?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?= To: qemu-devel@nongnu.org Cc: Yuval Shaia , Kevin Wolf , Prasanna Kumar Kalever , Fabiano Rosas , Cornelia Huck , Michael Roth , Li Zhijian , Prasanna Kumar Kalever , Peter Xu , integration@gluster.org, Paolo Bonzini , qemu-block@nongnu.org, =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= , devel@lists.libvirt.org, Hanna Reitz , "Michael S. Tsirkin" , Thomas Huth , Eric Blake , =?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?= Subject: [PATCH-for-9.1 v2 3/3] block/gluster: Remove RDMA protocol handling Date: Thu, 28 Mar 2024 14:02:55 +0100 Message-ID: <20240328130255.52257-4-philmd@linaro.org> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20240328130255.52257-1-philmd@linaro.org> References: <20240328130255.52257-1-philmd@linaro.org> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Received-SPF: pass client-ip=2a00:1450:4864:20::62e; envelope-from=philmd@linaro.org; helo=mail-ej1-x62e.google.com X-Spam_score_int: -20 X-Spam_score: -2.1 X-Spam_bar: -- X-Spam_report: (-2.1 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, RCVD_IN_DNSWL_NONE=-0.0001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, WEIRD_PORT=0.001 autolearn=unavailable autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: qemu-devel-bounces+importer=patchew.org@nongnu.org X-ZohoMail-DKIM: pass (identity @linaro.org) X-ZM-MESSAGEID: 1711631101582100003 GlusterFS+RDMA has been deprecated 8 years ago in commit 0552ff2465 ("block/gluster: deprecate rdma support"): gluster volfile server fetch happens through unix and/or tcp, it doesn't support volfile fetch over rdma. The rdma code may actually mislead, so to make sure things do not break, for now we fallback to tcp when requested for rdma, with a warning. If you are wondering how this worked all these days, its the gluster libgfapi code which handles anything other than unix transport as socket/tcp, sad but true. Besides, the whole RDMA subsystem was deprecated in commit e9a54265f5 ("hw/rdma: Deprecate the pvrdma device and the rdma subsystem") released in v8.2. Cc: Prasanna Kumar Kalever Signed-off-by: Philippe Mathieu-Daud=C3=A9 Reviewed-by: Michael S. Tsirkin Reviewed-by: Thomas Huth --- docs/system/device-url-syntax.rst.inc | 4 +-- docs/system/qemu-block-drivers.rst.inc | 1 - block/gluster.c | 39 -------------------------- 3 files changed, 2 insertions(+), 42 deletions(-) diff --git a/docs/system/device-url-syntax.rst.inc b/docs/system/device-url= -syntax.rst.inc index 7dbc525fa8..43b5c2596b 100644 --- a/docs/system/device-url-syntax.rst.inc +++ b/docs/system/device-url-syntax.rst.inc @@ -87,8 +87,8 @@ These are specified using a special URL syntax. =20 ``GlusterFS`` GlusterFS is a user space distributed file system. QEMU supports the - use of GlusterFS volumes for hosting VM disk images using TCP, Unix - Domain Sockets and RDMA transport protocols. + use of GlusterFS volumes for hosting VM disk images using TCP and Unix + Domain Sockets transport protocols. =20 Syntax for specifying a VM disk image on GlusterFS volume is =20 diff --git a/docs/system/qemu-block-drivers.rst.inc b/docs/system/qemu-bloc= k-drivers.rst.inc index 105cb9679c..384e95ba76 100644 --- a/docs/system/qemu-block-drivers.rst.inc +++ b/docs/system/qemu-block-drivers.rst.inc @@ -737,7 +737,6 @@ Examples |qemu_system| -drive file=3Dgluster+tcp://[1:2:3:4:5:6:7:8]:24007/testvo= l/dir/a.img |qemu_system| -drive file=3Dgluster+tcp://server.domain.com:24007/testvo= l/dir/a.img |qemu_system| -drive file=3Dgluster+unix:///testvol/dir/a.img?socket=3D/= tmp/glusterd.socket - |qemu_system| -drive file=3Dgluster+rdma://1.2.3.4:24007/testvol/a.img |qemu_system| -drive file=3Dgluster://1.2.3.4/testvol/a.img,file.debug= =3D9,file.logfile=3D/var/log/qemu-gluster.log |qemu_system| 'json:{"driver":"qcow2", "file":{"driver":"gluster", diff --git a/block/gluster.c b/block/gluster.c index cc74af06dc..4253c8db5e 100644 --- a/block/gluster.c +++ b/block/gluster.c @@ -371,9 +371,6 @@ static int qemu_gluster_parse_uri(BlockdevOptionsGluste= r *gconf, } else if (!strcmp(uri->scheme, "gluster+unix")) { gsconf->type =3D SOCKET_ADDRESS_TYPE_UNIX; is_unix =3D true; - } else if (!strcmp(uri->scheme, "gluster+rdma")) { - gsconf->type =3D SOCKET_ADDRESS_TYPE_INET; - warn_report("rdma feature is not supported, falling back to tcp"); } else { ret =3D -EINVAL; goto out; @@ -1638,44 +1635,8 @@ static BlockDriver bdrv_gluster_unix =3D { .strong_runtime_opts =3D gluster_strong_open_opts, }; =20 -/* rdma is deprecated (actually never supported for volfile fetch). - * Let's maintain it for the protocol compatibility, to make sure things - * won't break immediately. For now, gluster+rdma will fall back to gluste= r+tcp - * protocol with a warning. - * TODO: remove gluster+rdma interface support - */ -static BlockDriver bdrv_gluster_rdma =3D { - .format_name =3D "gluster", - .protocol_name =3D "gluster+rdma", - .instance_size =3D sizeof(BDRVGlusterState), - .bdrv_file_open =3D qemu_gluster_open, - .bdrv_reopen_prepare =3D qemu_gluster_reopen_prepare, - .bdrv_reopen_commit =3D qemu_gluster_reopen_commit, - .bdrv_reopen_abort =3D qemu_gluster_reopen_abort, - .bdrv_close =3D qemu_gluster_close, - .bdrv_co_create =3D qemu_gluster_co_create, - .bdrv_co_create_opts =3D qemu_gluster_co_create_opts, - .bdrv_co_getlength =3D qemu_gluster_co_getlength, - .bdrv_co_get_allocated_file_size =3D qemu_gluster_co_get_allocated_fil= e_size, - .bdrv_co_truncate =3D qemu_gluster_co_truncate, - .bdrv_co_readv =3D qemu_gluster_co_readv, - .bdrv_co_writev =3D qemu_gluster_co_writev, - .bdrv_co_flush_to_disk =3D qemu_gluster_co_flush_to_disk, -#ifdef CONFIG_GLUSTERFS_DISCARD - .bdrv_co_pdiscard =3D qemu_gluster_co_pdiscard, -#endif -#ifdef CONFIG_GLUSTERFS_ZEROFILL - .bdrv_co_pwrite_zeroes =3D qemu_gluster_co_pwrite_zeroes, -#endif - .bdrv_co_block_status =3D qemu_gluster_co_block_status, - .bdrv_refresh_limits =3D qemu_gluster_refresh_limits, - .create_opts =3D &qemu_gluster_create_opts, - .strong_runtime_opts =3D gluster_strong_open_opts, -}; - static void bdrv_gluster_init(void) { - bdrv_register(&bdrv_gluster_rdma); bdrv_register(&bdrv_gluster_unix); bdrv_register(&bdrv_gluster_tcp); bdrv_register(&bdrv_gluster); --=20 2.41.0