From nobody Sun Feb 8 05:40:26 2026 Received: from mgamail.intel.com (mgamail.intel.com [192.198.163.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 4FD071427E for ; Fri, 26 Jan 2024 10:59:17 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=192.198.163.11 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1706266768; cv=none; b=AwfsHxOco+CmOCT9pXuLyGjATg+ia3NBNIJxFSFtKKyk1xJHybXxT1dOOIKjtJRVoC9rROfrBI0DDQUSI7ZZNJwTwKSIdtRyzPpo5+REuLVHQxms8zeMy1/9HYugSNyxHyAch2aAXYf1cfWTg+yf66lf3EQQGYNXnr3O+nF3Nss= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1706266768; c=relaxed/simple; bh=zo5LBMZV0B5NEfxZY5cK3SYRyArtsbTV+MJeKCnjfLM=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=Ez85ImVNVw40oYGFZrbgMOpwRkjYrLAPknJRouDGrKwBDwP6jUMa35Bdf4T5JR+SkQbWzrU0++FkqPXgvxTp7679wtAu1j0DcE1j/II4UhgNdm4xnZ3TmT2Kyjpd4ndd8eTYhwnaZJYo6Ds9CZvhm4uqcvODxAKHmyc1o+UwuqU= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.intel.com; spf=none smtp.mailfrom=linux.intel.com; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b=hn9Odf3J; arc=none smtp.client-ip=192.198.163.11 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.intel.com Authentication-Results: smtp.subspace.kernel.org; spf=none smtp.mailfrom=linux.intel.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b="hn9Odf3J" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1706266757; x=1737802757; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=zo5LBMZV0B5NEfxZY5cK3SYRyArtsbTV+MJeKCnjfLM=; b=hn9Odf3JaJL5JyZgOnfiFqE7GolbQX2OhDVHCul1hwDBk/cjZl8iD/Si NicqzVz75HWelJtyvE6BdA298vipgvhCFmZfijaPha+W1qgYfa2Ry7Emp NAvHVT0JuAsyGe/7HJwZ/h6JCjx5fopbsqDgu7WZHhjrS3j9VW0XUxdSm QbFOnH49vyKE8qlSbtdg7pU7cJpc36TgwjGvGBVw/UlgtGmYolfRajetN psl3ZxsLcj4DXKTOzjLdWbRqJb6Zq3oKwQgVSR15XP6OPDDvwr14Ontso oyX+y9Vo8XXCLHzl72Kfq2wsFyp1cyDaYJQJ/MUSKN7ZiWuYbot0y3/30 w==; X-IronPort-AV: E=McAfee;i="6600,9927,10964"; a="9123363" X-IronPort-AV: E=Sophos;i="6.05,216,1701158400"; d="scan'208";a="9123363" Received: from fmsmga004.fm.intel.com ([10.253.24.48]) by fmvoesa105.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 26 Jan 2024 02:59:17 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=McAfee;i="6600,9927,10964"; a="857367842" X-IronPort-AV: E=Sophos;i="6.05,216,1701158400"; d="scan'208";a="857367842" Received: from allen-box.sh.intel.com ([10.239.159.127]) by fmsmga004.fm.intel.com with ESMTP; 26 Jan 2024 02:59:14 -0800 From: Lu Baolu To: Jason Gunthorpe , Kevin Tian , Joerg Roedel , Will Deacon , Robin Murphy Cc: iommu@lists.linux.dev, linux-kernel@vger.kernel.org, Lu Baolu Subject: [PATCH v2 1/2] iommu: Use mutex instead of spinlock for iommu_device_list Date: Fri, 26 Jan 2024 18:53:40 +0800 Message-Id: <20240126105341.78086-2-baolu.lu@linux.intel.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20240126105341.78086-1-baolu.lu@linux.intel.com> References: <20240126105341.78086-1-baolu.lu@linux.intel.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" The iommu_device_lock spinlock was used to protect the iommu device list. However, there is no requirement to access the iommu device list in interrupt context. Therefore, a mutex is sufficient. This also prepares for the next change, which will iterate the iommu device list and call the probe callback within the locking area. Signed-off-by: Lu Baolu Reviewed-by: Jason Gunthorpe Reviewed-by: Kevin Tian --- drivers/iommu/iommu.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 68e648b55767..0af0b5544072 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -146,7 +146,7 @@ struct iommu_group_attribute iommu_group_attr_##_name = =3D \ container_of(_kobj, struct iommu_group, kobj) =20 static LIST_HEAD(iommu_device_list); -static DEFINE_SPINLOCK(iommu_device_lock); +static DEFINE_MUTEX(iommu_device_lock); =20 static const struct bus_type * const iommu_buses[] =3D { &platform_bus_type, @@ -262,9 +262,9 @@ int iommu_device_register(struct iommu_device *iommu, if (hwdev) iommu->fwnode =3D dev_fwnode(hwdev); =20 - spin_lock(&iommu_device_lock); + mutex_lock(&iommu_device_lock); list_add_tail(&iommu->list, &iommu_device_list); - spin_unlock(&iommu_device_lock); + mutex_unlock(&iommu_device_lock); =20 for (int i =3D 0; i < ARRAY_SIZE(iommu_buses) && !err; i++) err =3D bus_iommu_probe(iommu_buses[i]); @@ -279,9 +279,9 @@ void iommu_device_unregister(struct iommu_device *iommu) for (int i =3D 0; i < ARRAY_SIZE(iommu_buses); i++) bus_for_each_dev(iommu_buses[i], NULL, iommu, remove_iommu_group); =20 - spin_lock(&iommu_device_lock); + mutex_lock(&iommu_device_lock); list_del(&iommu->list); - spin_unlock(&iommu_device_lock); + mutex_unlock(&iommu_device_lock); =20 /* Pairs with the alloc in generic_single_device_group() */ iommu_group_put(iommu->singleton_group); @@ -316,9 +316,9 @@ int iommu_device_register_bus(struct iommu_device *iomm= u, if (err) return err; =20 - spin_lock(&iommu_device_lock); + mutex_lock(&iommu_device_lock); list_add_tail(&iommu->list, &iommu_device_list); - spin_unlock(&iommu_device_lock); + mutex_unlock(&iommu_device_lock); =20 err =3D bus_iommu_probe(bus); if (err) { @@ -2033,9 +2033,9 @@ bool iommu_present(const struct bus_type *bus) =20 for (int i =3D 0; i < ARRAY_SIZE(iommu_buses); i++) { if (iommu_buses[i] =3D=3D bus) { - spin_lock(&iommu_device_lock); + mutex_lock(&iommu_device_lock); ret =3D !list_empty(&iommu_device_list); - spin_unlock(&iommu_device_lock); + mutex_unlock(&iommu_device_lock); } } return ret; @@ -2983,13 +2983,13 @@ const struct iommu_ops *iommu_ops_from_fwnode(struc= t fwnode_handle *fwnode) const struct iommu_ops *ops =3D NULL; struct iommu_device *iommu; =20 - spin_lock(&iommu_device_lock); + mutex_lock(&iommu_device_lock); list_for_each_entry(iommu, &iommu_device_list, list) if (iommu->fwnode =3D=3D fwnode) { ops =3D iommu->ops; break; } - spin_unlock(&iommu_device_lock); + mutex_unlock(&iommu_device_lock); return ops; } =20 --=20 2.34.1 From nobody Sun Feb 8 05:40:26 2026 Received: from mgamail.intel.com (mgamail.intel.com [192.198.163.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 90CA51427A for ; Fri, 26 Jan 2024 10:59:19 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=192.198.163.11 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1706266769; cv=none; b=npxYE7XSAMxT19Z80oWaVHcX0E4gOPw1m8wDviyaqLZ9NMv/hvyb39K2l5YbtCizyjFTy2CkZhYq1d84G8nl2+8DxnAoqXuRM1AO2UiguxLvu/ABTKfnjmxLjtdtIX7SPa+U7sqRvir3R0ASDOc1sTeruhGZKhIrI1SFhSsuPJw= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1706266769; c=relaxed/simple; bh=b4leudU2y75I7gVeqH5B2NhuCYSTwTPov4fkwl1ZnBs=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=msKJx9vQY0IoBjmeQd3izeXDLK1Ta/IEY3bcy1WRb7TvWa1rqzo2A3AnSrW0eK9QKWESp+ufZxT31/A0S6QZ7D8VbGhF9GTPlK6yWs5kUedvOvNtYpw8inmmbOsKjInZJXeWIG+rCmtgVcSWNBpNQp36SdhL0CAJl4BT3RjOJ50= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.intel.com; spf=none smtp.mailfrom=linux.intel.com; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b=GrsK9Fmj; arc=none smtp.client-ip=192.198.163.11 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.intel.com Authentication-Results: smtp.subspace.kernel.org; spf=none smtp.mailfrom=linux.intel.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b="GrsK9Fmj" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1706266760; x=1737802760; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=b4leudU2y75I7gVeqH5B2NhuCYSTwTPov4fkwl1ZnBs=; b=GrsK9FmjLx7P+kQXFxngAyvxH57O275Sq3hEglLu+pz+Vck20YHoKZQj Hx5xVUHQh1mASnpN19hwbm/31X0KLumRatain7w3A0//Z2Qoywu62uB0C CticCQsUkULI7rxv7TJnmfAW++EyemuuPLSaUlQS1LeURcZKrBO4DsC2V wRibZYwxTM/zvPcfbW9JefNM/NWJNlbsdTvlYLFL5reZwNux4DZH29tmG JKLGoKziNtLEfVEvmow/9iHkPreaMFEkcfM88IgFDXlh9JL0nuEkFUAW3 7xgtYQW59KpzUFhI69WzNU18oSa1sx1iRMaIRTkD3nlZZuFXZQ7+fyqWy w==; X-IronPort-AV: E=McAfee;i="6600,9927,10964"; a="9123376" X-IronPort-AV: E=Sophos;i="6.05,216,1701158400"; d="scan'208";a="9123376" Received: from fmsmga004.fm.intel.com ([10.253.24.48]) by fmvoesa105.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 26 Jan 2024 02:59:19 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=McAfee;i="6600,9927,10964"; a="857367852" X-IronPort-AV: E=Sophos;i="6.05,216,1701158400"; d="scan'208";a="857367852" Received: from allen-box.sh.intel.com ([10.239.159.127]) by fmsmga004.fm.intel.com with ESMTP; 26 Jan 2024 02:59:16 -0800 From: Lu Baolu To: Jason Gunthorpe , Kevin Tian , Joerg Roedel , Will Deacon , Robin Murphy Cc: iommu@lists.linux.dev, linux-kernel@vger.kernel.org, Lu Baolu Subject: [PATCH v2 2/2] iommu: Probe right iommu_ops for device Date: Fri, 26 Jan 2024 18:53:41 +0800 Message-Id: <20240126105341.78086-3-baolu.lu@linux.intel.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20240126105341.78086-1-baolu.lu@linux.intel.com> References: <20240126105341.78086-1-baolu.lu@linux.intel.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Previously, in the iommu probe device path, __iommu_probe_device() gets the iommu_ops for the device from dev->iommu->fwspec if this field has been initialized before probing. Otherwise, it is assumed that only one of Intel, AMD, s390, PAMU or legacy SMMUv2 can be present, hence it's feasible to lookup the global iommu device list and use the iommu_ops of the first iommu device which has no dev->iommu->fwspec. The assumption mentioned above is no longer correct with the introduction of the IOMMUFD mock device on x86 platforms. There might be multiple instances of iommu drivers, none of which have the dev->iommu->fwspec field populated. Probe the right iommu_ops for device by iterating over all the global drivers and call their probe function to find a match. Fixes: 17de3f5fdd35 ("iommu: Retire bus ops") Cc: Robin Murphy Signed-off-by: Lu Baolu Reviewed-by: Kevin Tian , with a nit: --- drivers/iommu/iommu.c | 76 +++++++++++++++++++++++++++---------------- 1 file changed, 48 insertions(+), 28 deletions(-) diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 0af0b5544072..2cdb01e411fa 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -396,30 +396,69 @@ void dev_iommu_priv_set(struct device *dev, void *pri= v) } EXPORT_SYMBOL_GPL(dev_iommu_priv_set); =20 +static struct iommu_device * +__iommu_do_probe_device(struct device *dev, const struct iommu_ops *ops) +{ + struct iommu_device *iommu_dev; + + if (!try_module_get(ops->owner)) + return ERR_PTR(-EINVAL); + + iommu_dev =3D ops->probe_device(dev); + if (IS_ERR(iommu_dev)) + module_put(ops->owner); + + return iommu_dev; +} + +static struct iommu_device *iommu_probe_device_ops(struct device *dev) +{ + struct iommu_device *iter, *iommu_dev =3D ERR_PTR(-ENODEV); + struct iommu_fwspec *fwspec; + + /* + * For FDT-based systems and ACPI IORT/VIOT, drivers register IOMMU + * instances with non-NULL fwnodes, and client devices should have been + * identified with a fwspec by this point. Otherwise, we will iterate + * over all the global drivers and call their probe function to find a + * match. + */ + fwspec =3D dev_iommu_fwspec_get(dev); + if (fwspec && fwspec->ops) + return __iommu_do_probe_device(dev, fwspec->ops); + + mutex_lock(&iommu_device_lock); + list_for_each_entry(iter, &iommu_device_list, list) { + iommu_dev =3D __iommu_do_probe_device(dev, iter->ops); + if (!IS_ERR(iommu_dev)) + break; + } + mutex_unlock(&iommu_device_lock); + + return iommu_dev; +} + /* * Init the dev->iommu and dev->iommu_group in the struct device and get t= he * driver probed */ -static int iommu_init_device(struct device *dev, const struct iommu_ops *o= ps) +static int iommu_init_device(struct device *dev) { struct iommu_device *iommu_dev; + const struct iommu_ops *ops; struct iommu_group *group; int ret; =20 if (!dev_iommu_get(dev)) return -ENOMEM; =20 - if (!try_module_get(ops->owner)) { - ret =3D -EINVAL; - goto err_free; - } - - iommu_dev =3D ops->probe_device(dev); + iommu_dev =3D iommu_probe_device_ops(dev); if (IS_ERR(iommu_dev)) { ret =3D PTR_ERR(iommu_dev); - goto err_module_put; + goto err_free; } dev->iommu->iommu_dev =3D iommu_dev; + ops =3D dev_iommu_ops(dev); =20 ret =3D iommu_device_link(iommu_dev, dev); if (ret) @@ -444,7 +483,6 @@ static int iommu_init_device(struct device *dev, const = struct iommu_ops *ops) err_release: if (ops->release_device) ops->release_device(dev); -err_module_put: module_put(ops->owner); err_free: dev->iommu->iommu_dev =3D NULL; @@ -499,28 +537,10 @@ DEFINE_MUTEX(iommu_probe_device_lock); =20 static int __iommu_probe_device(struct device *dev, struct list_head *grou= p_list) { - const struct iommu_ops *ops; - struct iommu_fwspec *fwspec; struct iommu_group *group; struct group_device *gdev; int ret; =20 - /* - * For FDT-based systems and ACPI IORT/VIOT, drivers register IOMMU - * instances with non-NULL fwnodes, and client devices should have been - * identified with a fwspec by this point. Otherwise, we can currently - * assume that only one of Intel, AMD, s390, PAMU or legacy SMMUv2 can - * be present, and that any of their registered instances has suitable - * ops for probing, and thus cheekily co-opt the same mechanism. - */ - fwspec =3D dev_iommu_fwspec_get(dev); - if (fwspec && fwspec->ops) - ops =3D fwspec->ops; - else - ops =3D iommu_ops_from_fwnode(NULL); - - if (!ops) - return -ENODEV; /* * Serialise to avoid races between IOMMU drivers registering in * parallel and/or the "replay" calls from ACPI/OF code via client @@ -534,7 +554,7 @@ static int __iommu_probe_device(struct device *dev, str= uct list_head *group_list if (dev->iommu_group) return 0; =20 - ret =3D iommu_init_device(dev, ops); + ret =3D iommu_init_device(dev); if (ret) return ret; =20 --=20 2.34.1