From nobody Mon Nov 25 15:44:07 2024 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (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 11BB018E77D; Sun, 27 Oct 2024 14:22:20 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=10.30.226.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1730038941; cv=none; b=CfK/Q96SOY+YGWRzr1xdsN+UEEfQCi4jhh/l7lqIDKaB55VVscxnddovyXAMXxOKTbT2TGiIjorqIgbG6f3iKuWr3o4E7o6HAVKb2QTOi1FvJOWtHh8fIfdCzyH8NtXU0JHDovcR/4WheyvNXGWkGOMIB83p3mt5gBmO193M/7s= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1730038941; c=relaxed/simple; bh=ufIIJTJY/y6wCplhr9nPwK7JHAPS3UUfFuT2KoVZ//k=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=S7wxdbUEdjLN+jZa60Uo9irPMdoOZqdnKmzmE6OqxPBNdEq3FMSJo7g7S3sOEPZBhoG23Um0eOUGzNw26erNgv+4ULPUtagoOOsuMx1IldoGgnmVepattKYjucqG9qoL25cbsf/QSW02PHhgqjeqRhTzghAVTnFRizGAthknGKE= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=fOZHmxpc; arc=none smtp.client-ip=10.30.226.201 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="fOZHmxpc" Received: by smtp.kernel.org (Postfix) with ESMTPSA id CA8CCC4CEC3; Sun, 27 Oct 2024 14:22:19 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1730038940; bh=ufIIJTJY/y6wCplhr9nPwK7JHAPS3UUfFuT2KoVZ//k=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=fOZHmxpcqkJmbQaKEC3Db5mURkJ2CSvGR9y1sM4c09/cyDn55R5+9tsnoPQFpeVdr utXoa2pz7BSlRuVZpoWyCjic+ZRKWtFbLpgtJcSrJwXkeE6RQN4Dq8z0TNJ2tR4bXZ olsZZFmhb0d+vSDqfiwmtGPh9eemWUC/gsW7Q0yvexTvDwXaWXi1bf7zHmp6DlEpLo ppmd90uVvoo80+kdtFqhayrKQNKFPA6RE5ZgjNkH304e4R4u9aqBcX/cVFm86/+x4W eyQh/jxd2rXoIM9BqLAwdaEprIotSvpUOPSOwvMP7auswWVbeMwV/TI9CNnjIh6/Tq T8YCpqBScneNw== From: Leon Romanovsky To: Jens Axboe , Jason Gunthorpe , Robin Murphy , Joerg Roedel , Will Deacon , Christoph Hellwig , Sagi Grimberg Cc: Leon Romanovsky , Keith Busch , Bjorn Helgaas , Logan Gunthorpe , Yishai Hadas , Shameer Kolothum , Kevin Tian , Alex Williamson , Marek Szyprowski , =?UTF-8?q?J=C3=A9r=C3=B4me=20Glisse?= , Andrew Morton , Jonathan Corbet , linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org, linux-block@vger.kernel.org, linux-rdma@vger.kernel.org, iommu@lists.linux.dev, linux-nvme@lists.infradead.org, linux-pci@vger.kernel.org, kvm@vger.kernel.org, linux-mm@kvack.org Subject: [PATCH 13/18] RDMA/core: Convert UMEM ODP DMA mapping to caching IOVA and page linkage Date: Sun, 27 Oct 2024 16:21:13 +0200 Message-ID: <5b825e07dd9b61d3d543664e3c833ed1aaa12eae.1730037276.git.leon@kernel.org> X-Mailer: git-send-email 2.46.2 In-Reply-To: References: 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" From: Leon Romanovsky Reuse newly added DMA API to cache IOVA and only link/unlink pages in fast path for UMEM ODP flow. Signed-off-by: Leon Romanovsky --- drivers/infiniband/core/umem_odp.c | 101 ++++++--------------------- drivers/infiniband/hw/mlx5/mlx5_ib.h | 11 +-- drivers/infiniband/hw/mlx5/odp.c | 40 +++++++---- drivers/infiniband/hw/mlx5/umr.c | 12 +++- include/rdma/ib_umem_odp.h | 13 +--- 5 files changed, 69 insertions(+), 108 deletions(-) diff --git a/drivers/infiniband/core/umem_odp.c b/drivers/infiniband/core/u= mem_odp.c index 9dba369365af..30cd8f353476 100644 --- a/drivers/infiniband/core/umem_odp.c +++ b/drivers/infiniband/core/umem_odp.c @@ -41,6 +41,7 @@ #include #include #include +#include #include =20 #include @@ -50,6 +51,7 @@ static inline int ib_init_umem_odp(struct ib_umem_odp *umem_odp, const struct mmu_interval_notifier_ops *ops) { + struct ib_device *dev =3D umem_odp->umem.ibdev; int ret; =20 umem_odp->umem.is_odp =3D 1; @@ -59,7 +61,6 @@ static inline int ib_init_umem_odp(struct ib_umem_odp *um= em_odp, size_t page_size =3D 1UL << umem_odp->page_shift; unsigned long start; unsigned long end; - size_t ndmas, npfns; =20 start =3D ALIGN_DOWN(umem_odp->umem.address, page_size); if (check_add_overflow(umem_odp->umem.address, @@ -70,36 +71,23 @@ static inline int ib_init_umem_odp(struct ib_umem_odp *= umem_odp, if (unlikely(end < page_size)) return -EOVERFLOW; =20 - ndmas =3D (end - start) >> umem_odp->page_shift; - if (!ndmas) - return -EINVAL; - - npfns =3D (end - start) >> PAGE_SHIFT; - umem_odp->pfn_list =3D kvcalloc( - npfns, sizeof(*umem_odp->pfn_list), GFP_KERNEL); - if (!umem_odp->pfn_list) - return -ENOMEM; - - umem_odp->dma_list =3D kvcalloc( - ndmas, sizeof(*umem_odp->dma_list), GFP_KERNEL); - if (!umem_odp->dma_list) { - ret =3D -ENOMEM; - goto out_pfn_list; - } + ret =3D hmm_dma_map_alloc(dev->dma_device, &umem_odp->map, + (end - start) >> PAGE_SHIFT, + 1 << umem_odp->page_shift); + if (ret) + return ret; =20 ret =3D mmu_interval_notifier_insert(&umem_odp->notifier, umem_odp->umem.owning_mm, start, end - start, ops); if (ret) - goto out_dma_list; + goto out_free_map; } =20 return 0; =20 -out_dma_list: - kvfree(umem_odp->dma_list); -out_pfn_list: - kvfree(umem_odp->pfn_list); +out_free_map: + hmm_dma_map_free(dev->dma_device, &umem_odp->map); return ret; } =20 @@ -262,6 +250,8 @@ EXPORT_SYMBOL(ib_umem_odp_get); =20 void ib_umem_odp_release(struct ib_umem_odp *umem_odp) { + struct ib_device *dev =3D umem_odp->umem.ibdev; + /* * Ensure that no more pages are mapped in the umem. * @@ -274,48 +264,17 @@ void ib_umem_odp_release(struct ib_umem_odp *umem_odp) ib_umem_end(umem_odp)); mutex_unlock(&umem_odp->umem_mutex); mmu_interval_notifier_remove(&umem_odp->notifier); - kvfree(umem_odp->dma_list); - kvfree(umem_odp->pfn_list); + hmm_dma_map_free(dev->dma_device, &umem_odp->map); } put_pid(umem_odp->tgid); kfree(umem_odp); } EXPORT_SYMBOL(ib_umem_odp_release); =20 -/* - * Map for DMA and insert a single page into the on-demand paging page tab= les. - * - * @umem: the umem to insert the page to. - * @dma_index: index in the umem to add the dma to. - * @page: the page struct to map and add. - * @access_mask: access permissions needed for this page. - * - * The function returns -EFAULT if the DMA mapping operation fails. - * - */ -static int ib_umem_odp_map_dma_single_page( - struct ib_umem_odp *umem_odp, - unsigned int dma_index, - struct page *page) -{ - struct ib_device *dev =3D umem_odp->umem.ibdev; - dma_addr_t *dma_addr =3D &umem_odp->dma_list[dma_index]; - - *dma_addr =3D ib_dma_map_page(dev, page, 0, 1 << umem_odp->page_shift, - DMA_BIDIRECTIONAL); - if (ib_dma_mapping_error(dev, *dma_addr)) { - *dma_addr =3D 0; - return -EFAULT; - } - umem_odp->npages++; - return 0; -} - /** * ib_umem_odp_map_dma_and_lock - DMA map userspace memory in an ODP MR an= d lock it. * * Maps the range passed in the argument to DMA addresses. - * The DMA addresses of the mapped pages is updated in umem_odp->dma_list. * Upon success the ODP MR will be locked to let caller complete its device * page table update. * @@ -372,7 +331,7 @@ int ib_umem_odp_map_dma_and_lock(struct ib_umem_odp *um= em_odp, u64 user_virt, range.default_flags |=3D HMM_PFN_REQ_WRITE; } =20 - range.hmm_pfns =3D &(umem_odp->pfn_list[pfn_start_idx]); + range.hmm_pfns =3D &(umem_odp->map.pfn_list[pfn_start_idx]); timeout =3D jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT); =20 retry: @@ -423,15 +382,6 @@ int ib_umem_odp_map_dma_and_lock(struct ib_umem_odp *u= mem_odp, u64 user_virt, __func__, hmm_order, page_shift); break; } - - ret =3D ib_umem_odp_map_dma_single_page( - umem_odp, dma_index, hmm_pfn_to_page(range.hmm_pfns[pfn_index])); - if (ret < 0) { - ibdev_dbg(umem_odp->umem.ibdev, - "ib_umem_odp_map_dma_single_page failed with error %d\n", ret); - break; - } - range.hmm_pfns[pfn_index] |=3D HMM_PFN_DMA_MAPPED; } /* upon success lock should stay on hold for the callee */ if (!ret) @@ -451,30 +401,23 @@ EXPORT_SYMBOL(ib_umem_odp_map_dma_and_lock); void ib_umem_odp_unmap_dma_pages(struct ib_umem_odp *umem_odp, u64 virt, u64 bound) { - dma_addr_t dma; - int idx; - u64 addr; struct ib_device *dev =3D umem_odp->umem.ibdev; + u64 addr; =20 lockdep_assert_held(&umem_odp->umem_mutex); =20 virt =3D max_t(u64, virt, ib_umem_start(umem_odp)); bound =3D min_t(u64, bound, ib_umem_end(umem_odp)); for (addr =3D virt; addr < bound; addr +=3D BIT(umem_odp->page_shift)) { - unsigned long pfn_idx =3D (addr - ib_umem_start(umem_odp)) >> PAGE_SHIFT; - struct page *page =3D hmm_pfn_to_page(umem_odp->pfn_list[pfn_idx]); - - idx =3D (addr - ib_umem_start(umem_odp)) >> umem_odp->page_shift; - dma =3D umem_odp->dma_list[idx]; + u64 offset =3D addr - ib_umem_start(umem_odp); + size_t idx =3D offset >> umem_odp->page_shift; + unsigned long pfn =3D umem_odp->map.pfn_list[idx]; =20 - if (!(umem_odp->pfn_list[pfn_idx] & HMM_PFN_VALID)) - goto clear; - if (!(umem_odp->pfn_list[pfn_idx] & HMM_PFN_DMA_MAPPED)) + if (!hmm_dma_unmap_pfn(dev->dma_device, &umem_odp->map, idx)) goto clear; =20 - ib_dma_unmap_page(dev, dma, BIT(umem_odp->page_shift), - DMA_BIDIRECTIONAL); - if (umem_odp->pfn_list[pfn_idx] & HMM_PFN_WRITE) { + if (pfn & HMM_PFN_WRITE) { + struct page *page =3D hmm_pfn_to_page(pfn); struct page *head_page =3D compound_head(page); /* * set_page_dirty prefers being called with @@ -489,7 +432,7 @@ void ib_umem_odp_unmap_dma_pages(struct ib_umem_odp *um= em_odp, u64 virt, } umem_odp->npages--; clear: - umem_odp->pfn_list[pfn_idx] &=3D ~HMM_PFN_FLAGS; + umem_odp->map.pfn_list[idx] &=3D ~HMM_PFN_FLAGS; } } EXPORT_SYMBOL(ib_umem_odp_unmap_dma_pages); diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/m= lx5/mlx5_ib.h index 3e4aaa6319db..1bae5595c729 100644 --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h @@ -1444,8 +1444,8 @@ void mlx5_ib_odp_cleanup_one(struct mlx5_ib_dev *ibde= v); int __init mlx5_ib_odp_init(void); void mlx5_ib_odp_cleanup(void); int mlx5_odp_init_mkey_cache(struct mlx5_ib_dev *dev); -void mlx5_odp_populate_xlt(void *xlt, size_t idx, size_t nentries, - struct mlx5_ib_mr *mr, int flags); +int mlx5_odp_populate_xlt(void *xlt, size_t idx, size_t nentries, + struct mlx5_ib_mr *mr, int flags); =20 int mlx5_ib_advise_mr_prefetch(struct ib_pd *pd, enum ib_uverbs_advise_mr_advice advice, @@ -1466,8 +1466,11 @@ static inline int mlx5_odp_init_mkey_cache(struct ml= x5_ib_dev *dev) { return 0; } -static inline void mlx5_odp_populate_xlt(void *xlt, size_t idx, size_t nen= tries, - struct mlx5_ib_mr *mr, int flags) {} +static inline int mlx5_odp_populate_xlt(void *xlt, size_t idx, size_t nent= ries, + struct mlx5_ib_mr *mr, int flags) +{ + return -EOPNOTSUPP; +} =20 static inline int mlx5_ib_advise_mr_prefetch(struct ib_pd *pd, diff --git a/drivers/infiniband/hw/mlx5/odp.c b/drivers/infiniband/hw/mlx5/= odp.c index 78887500ce15..fbb2a5670c32 100644 --- a/drivers/infiniband/hw/mlx5/odp.c +++ b/drivers/infiniband/hw/mlx5/odp.c @@ -35,6 +35,8 @@ #include #include #include +#include +#include =20 #include "mlx5_ib.h" #include "cmd.h" @@ -159,40 +161,50 @@ static void populate_klm(struct mlx5_klm *pklm, size_= t idx, size_t nentries, } } =20 -static void populate_mtt(__be64 *pas, size_t idx, size_t nentries, - struct mlx5_ib_mr *mr, int flags) +static int populate_mtt(__be64 *pas, size_t start, size_t nentries, + struct mlx5_ib_mr *mr, int flags) { struct ib_umem_odp *odp =3D to_ib_umem_odp(mr->umem); bool downgrade =3D flags & MLX5_IB_UPD_XLT_DOWNGRADE; - unsigned long pfn; - dma_addr_t pa; + struct pci_p2pdma_map_state p2pdma_state =3D {}; + struct ib_device *dev =3D odp->umem.ibdev; size_t i; =20 if (flags & MLX5_IB_UPD_XLT_ZAP) - return; + return 0; =20 for (i =3D 0; i < nentries; i++) { - pfn =3D odp->pfn_list[idx + i]; + unsigned long pfn =3D odp->map.pfn_list[start + i]; + dma_addr_t dma_addr; + + pfn =3D odp->map.pfn_list[start + i]; if (!(pfn & HMM_PFN_VALID)) /* ODP initialization */ continue; =20 - pa =3D odp->dma_list[idx + i]; - pa |=3D MLX5_IB_MTT_READ; + dma_addr =3D hmm_dma_map_pfn(dev->dma_device, &odp->map, + start + i, &p2pdma_state); + if (ib_dma_mapping_error(dev, dma_addr)) + return -EFAULT; + + dma_addr |=3D MLX5_IB_MTT_READ; if ((pfn & HMM_PFN_WRITE) && !downgrade) - pa |=3D MLX5_IB_MTT_WRITE; + dma_addr |=3D MLX5_IB_MTT_WRITE; =20 - pas[i] =3D cpu_to_be64(pa); + pas[i] =3D cpu_to_be64(dma_addr); + odp->npages++; } + return 0; } =20 -void mlx5_odp_populate_xlt(void *xlt, size_t idx, size_t nentries, - struct mlx5_ib_mr *mr, int flags) +int mlx5_odp_populate_xlt(void *xlt, size_t idx, size_t nentries, + struct mlx5_ib_mr *mr, int flags) { if (flags & MLX5_IB_UPD_XLT_INDIRECT) { populate_klm(xlt, idx, nentries, mr, flags); + return 0; } else { - populate_mtt(xlt, idx, nentries, mr, flags); + return populate_mtt(xlt, idx, nentries, mr, flags); } } =20 @@ -286,7 +298,7 @@ static bool mlx5_ib_invalidate_range(struct mmu_interva= l_notifier *mni, * estimate the cost of another UMR vs. the cost of bigger * UMR. */ - if (umem_odp->pfn_list[idx] & HMM_PFN_VALID) { + if (umem_odp->map.pfn_list[idx] & HMM_PFN_VALID) { if (!in_block) { blk_start_idx =3D idx; in_block =3D 1; diff --git a/drivers/infiniband/hw/mlx5/umr.c b/drivers/infiniband/hw/mlx5/= umr.c index 887fd6fa3ba9..d7fa94ab23cf 100644 --- a/drivers/infiniband/hw/mlx5/umr.c +++ b/drivers/infiniband/hw/mlx5/umr.c @@ -811,7 +811,17 @@ int mlx5r_umr_update_xlt(struct mlx5_ib_mr *mr, u64 id= x, int npages, size_to_map =3D npages * desc_size; dma_sync_single_for_cpu(ddev, sg.addr, sg.length, DMA_TO_DEVICE); - mlx5_odp_populate_xlt(xlt, idx, npages, mr, flags); + /* + * npages is the maximum number of pages to map, but we + * can't guarantee that all pages are actually mapped. + * + * For example, if page is p2p of type which is not supported + * for mapping, the number of pages mapped will be less than + * requested. + */ + err =3D mlx5_odp_populate_xlt(xlt, idx, npages, mr, flags); + if (err) + return err; dma_sync_single_for_device(ddev, sg.addr, sg.length, DMA_TO_DEVICE); sg.length =3D ALIGN(size_to_map, MLX5_UMR_FLEX_ALIGNMENT); diff --git a/include/rdma/ib_umem_odp.h b/include/rdma/ib_umem_odp.h index a345c26a745d..2a24bf791c10 100644 --- a/include/rdma/ib_umem_odp.h +++ b/include/rdma/ib_umem_odp.h @@ -8,24 +8,17 @@ =20 #include #include -#include +#include =20 struct ib_umem_odp { struct ib_umem umem; struct mmu_interval_notifier notifier; struct pid *tgid; =20 - /* An array of the pfns included in the on-demand paging umem. */ - unsigned long *pfn_list; + struct hmm_dma_map map; =20 /* - * An array with DMA addresses mapped for pfns in pfn_list. - * The lower two bits designate access permissions. - * See ODP_READ_ALLOWED_BIT and ODP_WRITE_ALLOWED_BIT. - */ - dma_addr_t *dma_list; - /* - * The umem_mutex protects the page_list and dma_list fields of an ODP + * The umem_mutex protects the page_list field of an ODP * umem, allowing only a single thread to map/unmap pages. The mutex * also protects access to the mmu notifier counters. */ --=20 2.46.2