2023-03-10 08:42:36

by Zheng Wang

Subject: [PATCH] hwmon: (xgene) Fix use after free bug in xgene_hwmon_remove due to race condition

In xgene_hwmon_probe, &ctx->workq is bound with
xgene_hwmon_evt_work. Then it will be started.

If we remove the driver which will call
xgene_hwmon_remove to make cleanup, there may
be a unfinished work.

The possiblesequence is as follows:

Fix it by finishing the work before cleanup in the
xgene_hwmon_remove

CPU0 CPU1

|xgene_hwmon_evt_work
xgene_hwmon_remove |
kfifo_free(&ctx->async_msg_fifo);|
|
|kfifo_out_spinlocked
|//use &ctx->async_msg_fifo
Fixes: 2ca492e22cb7 ("hwmon: (xgene) Fix crash when alarm occurs before driver probe")
Signed-off-by: Zheng Wang <[email protected]>
---
drivers/hwmon/xgene-hwmon.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/drivers/hwmon/xgene-hwmon.c b/drivers/hwmon/xgene-hwmon.c
index 5cde837bfd09..d1abea49f01b 100644
--- a/drivers/hwmon/xgene-hwmon.c
+++ b/drivers/hwmon/xgene-hwmon.c
@@ -761,6 +761,7 @@ static int xgene_hwmon_remove(struct platform_device *pdev)
{
struct xgene_hwmon_dev *ctx = platform_get_drvdata(pdev);

+ cancel_work_sync(&ctx->workq);
hwmon_device_unregister(ctx->hwmon_dev);
kfifo_free(&ctx->async_msg_fifo);
if (acpi_disabled)
--
2.25.1

2023-03-10 16:36:38

by Guenter Roeck

Subject: Re: [PATCH] hwmon: (xgene) Fix use after free bug in xgene_hwmon_remove due to race condition

On Fri, Mar 10, 2023 at 04:40:07PM +0800, Zheng Wang wrote:
> In xgene_hwmon_probe, &ctx->workq is bound with
> xgene_hwmon_evt_work. Then it will be started.
>
> If we remove the driver which will call
> xgene_hwmon_remove to make cleanup, there may
> be a unfinished work.
>
> The possiblesequence is as follows:
>
> Fix it by finishing the work before cleanup in the
> xgene_hwmon_remove
>
> CPU0 CPU1
>
> |xgene_hwmon_evt_work
> xgene_hwmon_remove |
> kfifo_free(&ctx->async_msg_fifo);|
> |
> |kfifo_out_spinlocked
> |//use &ctx->async_msg_fifo
> Fixes: 2ca492e22cb7 ("hwmon: (xgene) Fix crash when alarm occurs before driver probe")
> Signed-off-by: Zheng Wang <[email protected]>

Applied.

Thanks,
Guenter

> ---
> drivers/hwmon/xgene-hwmon.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/hwmon/xgene-hwmon.c b/drivers/hwmon/xgene-hwmon.c
> index 5cde837bfd09..d1abea49f01b 100644
> --- a/drivers/hwmon/xgene-hwmon.c
> +++ b/drivers/hwmon/xgene-hwmon.c
> @@ -761,6 +761,7 @@ static int xgene_hwmon_remove(struct platform_device *pdev)
> {
> struct xgene_hwmon_dev *ctx = platform_get_drvdata(pdev);
>
> + cancel_work_sync(&ctx->workq);
> hwmon_device_unregister(ctx->hwmon_dev);
> kfifo_free(&ctx->async_msg_fifo);
> if (acpi_disabled)

2023-03-10 16:57:06

by Zheng Hacker

Subject: Re: [PATCH] hwmon: (xgene) Fix use after free bug in xgene_hwmon_remove due to race condition

Guenter Roeck <[email protected]> 于2023年3月11日周六 00:34写道：
>
> On Fri, Mar 10, 2023 at 04:40:07PM +0800, Zheng Wang wrote:
> > In xgene_hwmon_probe, &ctx->workq is bound with
> > xgene_hwmon_evt_work. Then it will be started.
> >
> > If we remove the driver which will call
> > xgene_hwmon_remove to make cleanup, there may
> > be a unfinished work.
> >
> > The possiblesequence is as follows:
> >
> > Fix it by finishing the work before cleanup in the
> > xgene_hwmon_remove
> >
> > CPU0 CPU1
> >
> > |xgene_hwmon_evt_work
> > xgene_hwmon_remove |
> > kfifo_free(&ctx->async_msg_fifo);|
> > |
> > |kfifo_out_spinlocked
> > |//use &ctx->async_msg_fifo
> > Fixes: 2ca492e22cb7 ("hwmon: (xgene) Fix crash when alarm occurs before driver probe")
> > Signed-off-by: Zheng Wang <[email protected]>
>
> Applied.
>
Thanks for your effort.

Best regards,
Zheng
>
> > ---
> > drivers/hwmon/xgene-hwmon.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/hwmon/xgene-hwmon.c b/drivers/hwmon/xgene-hwmon.c
> > index 5cde837bfd09..d1abea49f01b 100644
> > --- a/drivers/hwmon/xgene-hwmon.c
> > +++ b/drivers/hwmon/xgene-hwmon.c
> > @@ -761,6 +761,7 @@ static int xgene_hwmon_remove(struct platform_device *pdev)
> > {
> > struct xgene_hwmon_dev *ctx = platform_get_drvdata(pdev);
> >
> > + cancel_work_sync(&ctx->workq);
> > hwmon_device_unregister(ctx->hwmon_dev);
> > kfifo_free(&ctx->async_msg_fifo);
> > if (acpi_disabled)