Linux-RDMA Archive on lore.kernel.org
 help / color / Atom feed
From: Gal Pressman <galpress@amazon.com>
To: Jason Gunthorpe <jgg@ziepe.ca>, Doug Ledford <dledford@redhat.com>
Cc: <linux-rdma@vger.kernel.org>,
	Alexander Matushevsky <matua@amazon.com>,
	Gal Pressman <galpress@amazon.com>,
	Firas JahJah <firasj@amazon.com>,
	"Yossi Leybovich" <sleybo@amazon.com>
Subject: [PATCH for-next 6/6] RDMA/efa: Do not delay freeing of DMA pages
Date: Tue, 14 Jan 2020 10:57:06 +0200
Message-ID: <20200114085706.82229-7-galpress@amazon.com> (raw)
In-Reply-To: <20200114085706.82229-1-galpress@amazon.com>

When destroying a DMA mmapped object, there is no need to delay the
pages freeing to dealloc_ucontext as the kernel itself will keep
reference count for these pages.

Remove the special handling of DMA pages and call free_pages_exact on
destroy_qp/cq.

Reviewed-by: Firas JahJah <firasj@amazon.com>
Reviewed-by: Yossi Leybovich <sleybo@amazon.com>
Signed-off-by: Gal Pressman <galpress@amazon.com>
---
 drivers/infiniband/hw/efa/efa_verbs.c | 13 ++++---------
 1 file changed, 4 insertions(+), 9 deletions(-)

diff --git a/drivers/infiniband/hw/efa/efa_verbs.c b/drivers/infiniband/hw/efa/efa_verbs.c
index 9a6cff718c49..fd2fc60d569f 100644
--- a/drivers/infiniband/hw/efa/efa_verbs.c
+++ b/drivers/infiniband/hw/efa/efa_verbs.c
@@ -413,6 +413,7 @@ int efa_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata)
 			  &qp->rq_dma_addr);
 		dma_unmap_single(&dev->pdev->dev, qp->rq_dma_addr, qp->rq_size,
 				 DMA_TO_DEVICE);
+		free_pages_exact(qp->rq_cpu_addr, qp->rq_size);
 	}
 
 	efa_qp_user_mmap_entries_remove(qp);
@@ -723,9 +724,7 @@ struct ib_qp *efa_create_qp(struct ib_pd *ibpd,
 	if (qp->rq_size) {
 		dma_unmap_single(&dev->pdev->dev, qp->rq_dma_addr, qp->rq_size,
 				 DMA_TO_DEVICE);
-
-		if (!qp->rq_mmap_entry)
-			free_pages_exact(qp->rq_cpu_addr, qp->rq_size);
+		free_pages_exact(qp->rq_cpu_addr, qp->rq_size);
 	}
 err_free_qp:
 	kfree(qp);
@@ -848,6 +847,7 @@ void efa_destroy_cq(struct ib_cq *ibcq, struct ib_udata *udata)
 	efa_destroy_cq_idx(dev, cq->cq_idx);
 	dma_unmap_single(&dev->pdev->dev, cq->dma_addr, cq->size,
 			 DMA_FROM_DEVICE);
+	free_pages_exact(cq->cpu_addr, cq->size);
 	rdma_user_mmap_entry_remove(cq->mmap_entry);
 }
 
@@ -987,8 +987,7 @@ int efa_create_cq(struct ib_cq *ibcq, const struct ib_cq_init_attr *attr,
 err_free_mapped:
 	dma_unmap_single(&dev->pdev->dev, cq->dma_addr, cq->size,
 			 DMA_FROM_DEVICE);
-	if (!cq->mmap_entry)
-		free_pages_exact(cq->cpu_addr, cq->size);
+	free_pages_exact(cq->cpu_addr, cq->size);
 
 err_out:
 	atomic64_inc(&dev->stats.sw_stats.create_cq_err);
@@ -1549,10 +1548,6 @@ void efa_mmap_free(struct rdma_user_mmap_entry *rdma_entry)
 {
 	struct efa_user_mmap_entry *entry = to_emmap(rdma_entry);
 
-	/* DMA mapping is already gone, now free the pages */
-	if (entry->mmap_flag == EFA_MMAP_DMA_PAGE)
-		free_pages_exact(phys_to_virt(entry->address),
-				 entry->rdma_entry.npages * PAGE_SIZE);
 	kfree(entry);
 }
 
-- 
2.24.1


  parent reply index

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-14  8:57 [PATCH for-next 0/6] EFA updates 2020-01-14 Gal Pressman
2020-01-14  8:57 ` [PATCH for-next 1/6] RDMA/efa: Unified getters/setters for device structs bitmask access Gal Pressman
2020-01-15 19:31   ` Jason Gunthorpe
2020-01-16  7:05     ` Gal Pressman
2020-01-14  8:57 ` [PATCH for-next 2/6] RDMA/efa: Properly document the interrupt mask register Gal Pressman
2020-01-14  8:57 ` [PATCH for-next 3/6] RDMA/efa: Device definitions documentation updates Gal Pressman
2020-01-14  8:57 ` [PATCH for-next 4/6] RDMA/efa: Remove {} brackets from single statement if Gal Pressman
2020-01-14  8:57 ` [PATCH for-next 5/6] RDMA/efa: Remove unused ucontext parameter from efa_qp_user_mmap_entries_remove Gal Pressman
2020-01-14  8:57 ` Gal Pressman [this message]
2020-01-15 19:51   ` [PATCH for-next 6/6] RDMA/efa: Do not delay freeing of DMA pages Jason Gunthorpe
2020-01-16  8:26     ` Gal Pressman
2020-01-15 19:58 ` [PATCH for-next 0/6] EFA updates 2020-01-14 Jason Gunthorpe

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200114085706.82229-7-galpress@amazon.com \
    --to=galpress@amazon.com \
    --cc=dledford@redhat.com \
    --cc=firasj@amazon.com \
    --cc=jgg@ziepe.ca \
    --cc=linux-rdma@vger.kernel.org \
    --cc=matua@amazon.com \
    --cc=sleybo@amazon.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-RDMA Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-rdma/0 linux-rdma/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-rdma linux-rdma/ https://lore.kernel.org/linux-rdma \
		linux-rdma@vger.kernel.org
	public-inbox-index linux-rdma

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-rdma


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git