diff mbox series

udmabuf: fix dma-buf cpu access

Message ID 20191213193359.1603-1-gurchetansingh@chromium.org (mailing list archive)
State New, archived
Headers show
Series udmabuf: fix dma-buf cpu access | expand

Commit Message

Gurchetan Singh Dec. 13, 2019, 7:33 p.m. UTC
I'm just going to put Chia's review comment here since it sums
the issue rather nicely:

"(1) Semantically, a dma-buf is in DMA domain.  CPU access from the
importer must be surrounded by {begin,end}_cpu_access.  This gives the
exporter a chance to move the buffer to the CPU domain temporarily.

(2) When the exporter itself has other means to do CPU access, it is
only reasonable for the exporter to move the buffer to the CPU domain
before access, and to the DMA domain after access.  The exporter can
potentially reuse {begin,end}_cpu_access for that purpose.

Because of (1), udmabuf does need to implement the
{begin,end}_cpu_access hooks.  But "begin" should mean
dma_sync_sg_for_cpu and "end" should mean dma_sync_sg_for_device.

Because of (2), if userspace wants to continuing accessing through the
memfd mapping, it should call udmabuf's {begin,end}_cpu_access to
avoid cache issues."

Reported-by: Chia-I Wu <olvaffe@gmail.com>
Suggested-by: Chia-I Wu <olvaffe@gmail.com>
Fixes: 284562e1f34 ("udmabuf: implement begin_cpu_access/end_cpu_access hooks")
---
 drivers/dma-buf/udmabuf.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

Comments

Gerd Hoffmann Dec. 16, 2019, 11:42 a.m. UTC | #1
On Fri, Dec 13, 2019 at 11:33:59AM -0800, Gurchetan Singh wrote:
> I'm just going to put Chia's review comment here since it sums
> the issue rather nicely:
> 
> "(1) Semantically, a dma-buf is in DMA domain.  CPU access from the
> importer must be surrounded by {begin,end}_cpu_access.  This gives the
> exporter a chance to move the buffer to the CPU domain temporarily.
> 
> (2) When the exporter itself has other means to do CPU access, it is
> only reasonable for the exporter to move the buffer to the CPU domain
> before access, and to the DMA domain after access.  The exporter can
> potentially reuse {begin,end}_cpu_access for that purpose.
> 
> Because of (1), udmabuf does need to implement the
> {begin,end}_cpu_access hooks.  But "begin" should mean
> dma_sync_sg_for_cpu and "end" should mean dma_sync_sg_for_device.
> 
> Because of (2), if userspace wants to continuing accessing through the
> memfd mapping, it should call udmabuf's {begin,end}_cpu_access to
> avoid cache issues."
> 
> Reported-by: Chia-I Wu <olvaffe@gmail.com>
> Suggested-by: Chia-I Wu <olvaffe@gmail.com>
> Fixes: 284562e1f34 ("udmabuf: implement begin_cpu_access/end_cpu_access hooks")

Needs 12 hash digits.

Also your signed-off is missing ...

cheers,
  Gerd
diff mbox series

Patch

diff --git a/drivers/dma-buf/udmabuf.c b/drivers/dma-buf/udmabuf.c
index 61b0a2cff874..acb26c627d27 100644
--- a/drivers/dma-buf/udmabuf.c
+++ b/drivers/dma-buf/udmabuf.c
@@ -122,9 +122,8 @@  static int begin_cpu_udmabuf(struct dma_buf *buf,
 		if (IS_ERR(ubuf->sg))
 			return PTR_ERR(ubuf->sg);
 	} else {
-		dma_sync_sg_for_device(dev, ubuf->sg->sgl,
-				       ubuf->sg->nents,
-				       direction);
+		dma_sync_sg_for_cpu(dev, ubuf->sg->sgl, ubuf->sg->nents,
+				    direction);
 	}
 
 	return 0;
@@ -139,7 +138,7 @@  static int end_cpu_udmabuf(struct dma_buf *buf,
 	if (!ubuf->sg)
 		return -EINVAL;
 
-	dma_sync_sg_for_cpu(dev, ubuf->sg->sgl, ubuf->sg->nents, direction);
+	dma_sync_sg_for_device(dev, ubuf->sg->sgl, ubuf->sg->nents, direction);
 	return 0;
 }