In renesas_usb3_probe, &usb3->role_work is bound with
renesas_usb3_role_work. renesas_usb3_start will be called
to start the work.
If we remove the driver which will call usbhs_remove, there may be
an unfinished work. The possible sequence is as follows:
Fix it by canceling the work before cleanup in the renesas_usb3_remove
CPU0 CPU1
|renesas_usb3_role_work
renesas_usb3_remove |
usb_role_switch_unregister |
device_unregister |
kfree(sw) |
free usb3->role_sw |
| usb_role_switch_set_role
| //use usb3->role_sw
Fixes: 39facfa01c9f ("usb: gadget: udc: renesas_usb3: Add register of usb role switch")
Reviewed-by: Yoshihiro Shimoda <[email protected]>
Signed-off-by: Zheng Wang <[email protected]>
---
v3:
- modify the commit message to make it clearer suggested by Yoshihiro Shimoda
v2:
- fix typo, use clearer commit message and only cancel the UAF-related work suggested by Yoshihiro Shimoda
---
drivers/usb/gadget/udc/renesas_usb3.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/usb/gadget/udc/renesas_usb3.c b/drivers/usb/gadget/udc/renesas_usb3.c
index bee6bceafc4f..a301af66bd91 100644
--- a/drivers/usb/gadget/udc/renesas_usb3.c
+++ b/drivers/usb/gadget/udc/renesas_usb3.c
@@ -2661,6 +2661,7 @@ static int renesas_usb3_remove(struct platform_device *pdev)
debugfs_remove_recursive(usb3->dentry);
device_remove_file(&pdev->dev, &dev_attr_role);
+ cancel_work_sync(&usb3->role_work);
usb_role_switch_unregister(usb3->role_sw);
usb_del_gadget_udc(&usb3->gadget);
--
2.25.1
Hi Zheng,
> From: Zheng Wang, Sent: Tuesday, March 14, 2023 1:51 AM
> Subject: [PATCH v3] usb: gadget: udc: renesas_usb3: Fix use after free bug in renesas_usb3_remove due to race condition
>
> In renesas_usb3_probe, &usb3->role_work is bound with
> renesas_usb3_role_work. renesas_usb3_start will be called
> to start the work.
>
> If we remove the driver which will call usbhs_remove, there may be
> an unfinished work. The possible sequence is as follows:
>
> Fix it by canceling the work before cleanup in the renesas_usb3_remove
>
> CPU0 CPU1
>
> |renesas_usb3_role_work
> renesas_usb3_remove |
> usb_role_switch_unregister |
> device_unregister |
> kfree(sw) |
> free usb3->role_sw |
> | usb_role_switch_set_role
> | //use usb3->role_sw
>
> Fixes: 39facfa01c9f ("usb: gadget: udc: renesas_usb3: Add register of usb role switch")
> Reviewed-by: Yoshihiro Shimoda <[email protected]>
> Signed-off-by: Zheng Wang <[email protected]>
JFYI, but IIUC,
- In general, my Reviewed-by is described here, not top of your Signed-off-by.
- Also, you don't need to submit such this patch. If you would like to submit
such a patch for some reason (ping or so on), incrementing patch version is
better because original v3 [1] doesn't have my Reviewed-by tag.
[1]
https://lore.kernel.org/all/[email protected]/
Best regards,
Yoshihiro Shimoda
> ---
> v3:
> - modify the commit message to make it clearer suggested by Yoshihiro Shimoda
> v2:
> - fix typo, use clearer commit message and only cancel the UAF-related work suggested by Yoshihiro Shimoda
> ---
> drivers/usb/gadget/udc/renesas_usb3.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/usb/gadget/udc/renesas_usb3.c b/drivers/usb/gadget/udc/renesas_usb3.c
> index bee6bceafc4f..a301af66bd91 100644
> --- a/drivers/usb/gadget/udc/renesas_usb3.c
> +++ b/drivers/usb/gadget/udc/renesas_usb3.c
> @@ -2661,6 +2661,7 @@ static int renesas_usb3_remove(struct platform_device *pdev)
> debugfs_remove_recursive(usb3->dentry);
> device_remove_file(&pdev->dev, &dev_attr_role);
>
> + cancel_work_sync(&usb3->role_work);
> usb_role_switch_unregister(usb3->role_sw);
>
> usb_del_gadget_udc(&usb3->gadget);
> --
> 2.25.1
Yoshihiro Shimoda <[email protected]> 于2023年3月14日周二 08:16写道：
>
> Hi Zheng,
>
> > From: Zheng Wang, Sent: Tuesday, March 14, 2023 1:51 AM
> > Subject: [PATCH v3] usb: gadget: udc: renesas_usb3: Fix use after free bug in renesas_usb3_remove due to race condition
> >
> > In renesas_usb3_probe, &usb3->role_work is bound with
> > renesas_usb3_role_work. renesas_usb3_start will be called
> > to start the work.
> >
> > If we remove the driver which will call usbhs_remove, there may be
> > an unfinished work. The possible sequence is as follows:
> >
> > Fix it by canceling the work before cleanup in the renesas_usb3_remove
> >
> > CPU0 CPU1
> >
> > |renesas_usb3_role_work
> > renesas_usb3_remove |
> > usb_role_switch_unregister |
> > device_unregister |
> > kfree(sw) |
> > free usb3->role_sw |
> > | usb_role_switch_set_role
> > | //use usb3->role_sw
> >
> > Fixes: 39facfa01c9f ("usb: gadget: udc: renesas_usb3: Add register of usb role switch")
> > Reviewed-by: Yoshihiro Shimoda <[email protected]>
> > Signed-off-by: Zheng Wang <[email protected]>
>
> JFYI, but IIUC,
> - In general, my Reviewed-by is described here, not top of your Signed-off-by.
> - Also, you don't need to submit such this patch. If you would like to submit
> such a patch for some reason (ping or so on), incrementing patch version is
> better because original v3 [1] doesn't have my Reviewed-by tag.
>
> [1]
> https://lore.kernel.org/all/[email protected]/
>
Dear Yoshihir,
Thanks for your advice. I am not familiar with the submission rules, thanks for
your kind reminder. I'll keep that in mind in the future. Have a good day :)
Best regards,
Zheng
> Best regards,
> Yoshihiro Shimoda
>
> > ---
> > v3:
> > - modify the commit message to make it clearer suggested by Yoshihiro Shimoda
> > v2:
> > - fix typo, use clearer commit message and only cancel the UAF-related work suggested by Yoshihiro Shimoda
> > ---
> > drivers/usb/gadget/udc/renesas_usb3.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/usb/gadget/udc/renesas_usb3.c b/drivers/usb/gadget/udc/renesas_usb3.c
> > index bee6bceafc4f..a301af66bd91 100644
> > --- a/drivers/usb/gadget/udc/renesas_usb3.c
> > +++ b/drivers/usb/gadget/udc/renesas_usb3.c
> > @@ -2661,6 +2661,7 @@ static int renesas_usb3_remove(struct platform_device *pdev)
> > debugfs_remove_recursive(usb3->dentry);
> > device_remove_file(&pdev->dev, &dev_attr_role);
> >
> > + cancel_work_sync(&usb3->role_work);
> > usb_role_switch_unregister(usb3->role_sw);
> >
> > usb_del_gadget_udc(&usb3->gadget);
> > --
> > 2.25.1
>
On Tue, Mar 14, 2023 at 12:51:28AM +0800, Zheng Wang wrote:
> In renesas_usb3_probe, &usb3->role_work is bound with
> renesas_usb3_role_work. renesas_usb3_start will be called
> to start the work.
>
> If we remove the driver which will call usbhs_remove, there may be
> an unfinished work. The possible sequence is as follows:
>
> Fix it by canceling the work before cleanup in the renesas_usb3_remove
>
> CPU0 CPU1
>
> |renesas_usb3_role_work
> renesas_usb3_remove |
> usb_role_switch_unregister |
> device_unregister |
> kfree(sw) |
> free usb3->role_sw |
> | usb_role_switch_set_role
> | //use usb3->role_sw
>
> Fixes: 39facfa01c9f ("usb: gadget: udc: renesas_usb3: Add register of usb role switch")
> Reviewed-by: Yoshihiro Shimoda <[email protected]>
> Signed-off-by: Zheng Wang <[email protected]>
> ---
> v3:
> - modify the commit message to make it clearer suggested by Yoshihiro Shimoda
I see 2 v3 patches in the mailing list, which is obviously not correct.
Please resubmit a v4 properly.
thanks,
greg k-h
Greg KH <[email protected]> 于2023年3月16日周四 18:55写道：
>
> On Tue, Mar 14, 2023 at 12:51:28AM +0800, Zheng Wang wrote:
> > In renesas_usb3_probe, &usb3->role_work is bound with
> > renesas_usb3_role_work. renesas_usb3_start will be called
> > to start the work.
> >
> > If we remove the driver which will call usbhs_remove, there may be
> > an unfinished work. The possible sequence is as follows:
> >
> > Fix it by canceling the work before cleanup in the renesas_usb3_remove
> >
> > CPU0 CPU1
> >
> > |renesas_usb3_role_work
> > renesas_usb3_remove |
> > usb_role_switch_unregister |
> > device_unregister |
> > kfree(sw) |
> > free usb3->role_sw |
> > | usb_role_switch_set_role
> > | //use usb3->role_sw
> >
> > Fixes: 39facfa01c9f ("usb: gadget: udc: renesas_usb3: Add register of usb role switch")
> > Reviewed-by: Yoshihiro Shimoda <[email protected]>
> > Signed-off-by: Zheng Wang <[email protected]>
> > ---
> > v3:
> > - modify the commit message to make it clearer suggested by Yoshihiro Shimoda
>
> I see 2 v3 patches in the mailing list, which is obviously not correct.
>
> Please resubmit a v4 properly.
>
Hi Greg,
Thanks for your kind reminder. I'll resubmit a v4 patch later.
Best regards,
Zheng
> thanks,
>
> greg k-h