2023-05-09 06:23:03

by Harshit Mogalapalli

[permalink] [raw]
Subject: [PATCH] dmaengine: idxd: Fix passing freed memory in idxd_cdev_open()

Smatch warns:
drivers/dma/idxd/cdev.c:327:
idxd_cdev_open() warn: 'sva' was already freed.

When idxd_wq_set_pasid() fails, the current code unbinds sva and then
goes to 'failed_set_pasid' where iommu_sva_unbind_device is called
again causing the above warning.
[ device_user_pasid_enabled(idxd) is still true when calling
failed_set_pasid ]

Fix this by removing additional unbind when idxd_wq_set_pasid() fails

Fixes: b022f59725f0 ("dmaengine: idxd: add idxd_copy_cr() to copy user completion record during page fault handling")
Signed-off-by: Harshit Mogalapalli <[email protected]>
---
This is purely based on static analysis. Only compile tested.
---
drivers/dma/idxd/cdev.c | 1 -
1 file changed, 1 deletion(-)

diff --git a/drivers/dma/idxd/cdev.c b/drivers/dma/idxd/cdev.c
index ecbf67c2ad2b..d32deb9b4e3d 100644
--- a/drivers/dma/idxd/cdev.c
+++ b/drivers/dma/idxd/cdev.c
@@ -277,7 +277,6 @@ static int idxd_cdev_open(struct inode *inode, struct file *filp)
if (wq_dedicated(wq)) {
rc = idxd_wq_set_pasid(wq, pasid);
if (rc < 0) {
- iommu_sva_unbind_device(sva);
dev_err(dev, "wq set pasid failed: %d\n", rc);
goto failed_set_pasid;
}
--
2.38.1


2023-05-09 15:21:16

by Dave Jiang

[permalink] [raw]
Subject: Re: [PATCH] dmaengine: idxd: Fix passing freed memory in idxd_cdev_open()



On 5/8/23 11:07 PM, Harshit Mogalapalli wrote:
> Smatch warns:
> drivers/dma/idxd/cdev.c:327:
> idxd_cdev_open() warn: 'sva' was already freed.
>
> When idxd_wq_set_pasid() fails, the current code unbinds sva and then
> goes to 'failed_set_pasid' where iommu_sva_unbind_device is called
> again causing the above warning.
> [ device_user_pasid_enabled(idxd) is still true when calling
> failed_set_pasid ]
>
> Fix this by removing additional unbind when idxd_wq_set_pasid() fails
>
> Fixes: b022f59725f0 ("dmaengine: idxd: add idxd_copy_cr() to copy user completion record during page fault handling")
> Signed-off-by: Harshit Mogalapalli <[email protected]>

Acked-by: Dave Jiang <[email protected]>

Thank you!
> ---
> This is purely based on static analysis. Only compile tested.
> ---
> drivers/dma/idxd/cdev.c | 1 -
> 1 file changed, 1 deletion(-)
>
> diff --git a/drivers/dma/idxd/cdev.c b/drivers/dma/idxd/cdev.c
> index ecbf67c2ad2b..d32deb9b4e3d 100644
> --- a/drivers/dma/idxd/cdev.c
> +++ b/drivers/dma/idxd/cdev.c
> @@ -277,7 +277,6 @@ static int idxd_cdev_open(struct inode *inode, struct file *filp)
> if (wq_dedicated(wq)) {
> rc = idxd_wq_set_pasid(wq, pasid);
> if (rc < 0) {
> - iommu_sva_unbind_device(sva);
> dev_err(dev, "wq set pasid failed: %d\n", rc);
> goto failed_set_pasid;
> }

2023-05-10 22:10:39

by Fenghua Yu

[permalink] [raw]
Subject: Re: [PATCH] dmaengine: idxd: Fix passing freed memory in idxd_cdev_open()



On 5/8/23 23:07, Harshit Mogalapalli wrote:
> Smatch warns:
> drivers/dma/idxd/cdev.c:327:
> idxd_cdev_open() warn: 'sva' was already freed.
>
> When idxd_wq_set_pasid() fails, the current code unbinds sva and then
> goes to 'failed_set_pasid' where iommu_sva_unbind_device is called
> again causing the above warning.
> [ device_user_pasid_enabled(idxd) is still true when calling
> failed_set_pasid ]
>
> Fix this by removing additional unbind when idxd_wq_set_pasid() fails
>
> Fixes: b022f59725f0 ("dmaengine: idxd: add idxd_copy_cr() to copy user completion record during page fault handling")
> Signed-off-by: Harshit Mogalapalli <[email protected]>

Acked-by: Fenghua Yu <[email protected]>

Thanks.

-Fenghua

2023-05-16 18:16:36

by Vinod Koul

[permalink] [raw]
Subject: Re: [PATCH] dmaengine: idxd: Fix passing freed memory in idxd_cdev_open()

On 08-05-23, 23:07, Harshit Mogalapalli wrote:
> Smatch warns:
> drivers/dma/idxd/cdev.c:327:
> idxd_cdev_open() warn: 'sva' was already freed.
>
> When idxd_wq_set_pasid() fails, the current code unbinds sva and then
> goes to 'failed_set_pasid' where iommu_sva_unbind_device is called
> again causing the above warning.
> [ device_user_pasid_enabled(idxd) is still true when calling
> failed_set_pasid ]

Applied, thanks

--
~Vinod