uio_pci_sva allocates struct uio_pci_sva_dev with devm_kzalloc() in
probe(), but then calls kfree(udev) both on the probe() error path
(label out_free) and again in remove().

Because devm_kzalloc() allocations are devres-managed and are freed
automatically when the device is detached (including after a failing
probe() and during driver unbind), the explicit kfree() can lead to a
double free.

If probe() fails after devm_kzalloc(), the error path frees udev and
devres cleanup will free it again when the core unwinds the partially
bound device.  On normal driver removal, remove() frees udev and devres
will free it again when the device is detached.

Fix by removing the manual kfree() calls and dropping the now-unused
label.

Fixes: 3397c3cd859a2 ("uio: Add SVA support for PCI devices via uio_pci_generic_sva.c")
Cc: stable@vger.kernel.org
Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
---
v3:
  - Add changelog below the --- line describing changes since v2.

v2:
  - Reflow commit message to keep lines within 75 characters.

 drivers/uio/uio_pci_generic_sva.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/uio/uio_pci_generic_sva.c b/drivers/uio/uio_pci_generic_sva.c
index 4a46acd994a8..152201047334 100644
--- a/drivers/uio/uio_pci_generic_sva.c
+++ b/drivers/uio/uio_pci_generic_sva.c
@@ -129,15 +129,13 @@ static int probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	ret = devm_uio_register_device(&pdev->dev, &udev->info);
 	if (ret) {
 		dev_err(&pdev->dev, "Failed to register uio device\n");
-		goto out_free;
+		goto out_disable;
 	}
 
 	pci_set_drvdata(pdev, udev);
 
 	return 0;
 
-out_free:
-	kfree(udev);
 out_disable:
 	pci_disable_device(pdev);
 
@@ -150,7 +148,6 @@ static void remove(struct pci_dev *pdev)
 
 	pci_release_regions(pdev);
 	pci_disable_device(pdev);
-	kfree(udev);
 }
 
 static ssize_t pasid_show(struct device *dev,
-- 
2.43.0

On Thu, Feb 26, 2026 at 09:16:32AM +0800, Guangshuo Li wrote:
> uio_pci_sva allocates struct uio_pci_sva_dev with devm_kzalloc() in
> probe(), but then calls kfree(udev) both on the probe() error path
> (label out_free) and again in remove().
> 
> Because devm_kzalloc() allocations are devres-managed and are freed
> automatically when the device is detached (including after a failing
> probe() and during driver unbind), the explicit kfree() can lead to a
> double free.
> 
> If probe() fails after devm_kzalloc(), the error path frees udev and
> devres cleanup will free it again when the core unwinds the partially
> bound device.  On normal driver removal, remove() frees udev and devres
> will free it again when the device is detached.
> 
> Fix by removing the manual kfree() calls and dropping the now-unused
> label.
> 
> Fixes: 3397c3cd859a2 ("uio: Add SVA support for PCI devices via uio_pci_generic_sva.c")
> Cc: stable@vger.kernel.org
> Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
> ---
> v3:
>   - Add changelog below the --- line describing changes since v2.
> 
> v2:
>   - Reflow commit message to keep lines within 75 characters.

You forgot my question of "how was this found and tested"?

thanks,

greg k-h

Hi Greg,

Thanks for the reminder.

This was found by a static analysis tool I designed. After a manual
review, I confirmed the issue and sent the fix.

Would you prefer that I include the “how it was found and tested”
information in the commit message?

Thanks,

Guangshuo

Greg Kroah-Hartman <gregkh@linuxfoundation.org> 于2026年2月26日周四 09:38写道：
>
> On Thu, Feb 26, 2026 at 09:16:32AM +0800, Guangshuo Li wrote:
> > uio_pci_sva allocates struct uio_pci_sva_dev with devm_kzalloc() in
> > probe(), but then calls kfree(udev) both on the probe() error path
> > (label out_free) and again in remove().
> >
> > Because devm_kzalloc() allocations are devres-managed and are freed
> > automatically when the device is detached (including after a failing
> > probe() and during driver unbind), the explicit kfree() can lead to a
> > double free.
> >
> > If probe() fails after devm_kzalloc(), the error path frees udev and
> > devres cleanup will free it again when the core unwinds the partially
> > bound device.  On normal driver removal, remove() frees udev and devres
> > will free it again when the device is detached.
> >
> > Fix by removing the manual kfree() calls and dropping the now-unused
> > label.
> >
> > Fixes: 3397c3cd859a2 ("uio: Add SVA support for PCI devices via uio_pci_generic_sva.c")
> > Cc: stable@vger.kernel.org
> > Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
> > ---
> > v3:
> >   - Add changelog below the --- line describing changes since v2.
> >
> > v2:
> >   - Reflow commit message to keep lines within 75 characters.
>
> You forgot my question of "how was this found and tested"?
>
> thanks,
>
> greg k-h

On Thu, Feb 26, 2026 at 06:14:54PM +0800, Guangshuo Li wrote:
> Hi Greg,
> 
> Thanks for the reminder.
> 
> This was found by a static analysis tool I designed. After a manual
> review, I confirmed the issue and sent the fix.
> 
> Would you prefer that I include the “how it was found and tested”
> information in the commit message?

As per our documentation (please go read it again), it is required :)

thanks,

greg k-h

On 2/26/2026 10:32 PM, Greg Kroah-Hartman wrote:
> On Thu, Feb 26, 2026 at 06:14:54PM +0800, Guangshuo Li wrote:
>> Hi Greg,
>>
>> Thanks for the reminder.
>>
>> This was found by a static analysis tool I designed. After a manual
>> review, I confirmed the issue and sent the fix.
>>
>> Would you prefer that I include the “how it was found and tested”
>> information in the commit message?
> 
> As per our documentation (please go read it again), it is required :)
> 
> thanks,
> 

Hi Greg, Guangshuo
I’d be happy to help test this bug and the proposed patch.
  (Well… to be honest, I probably should have fixed it myself earlier — 
I’ve been meaning to, but got caught up with other things lately…)
Thanks for working on it!

Best regards,
Yaxing Guo

> greg k-h