When executing SMP task failed, the smp_execute_task_sg()
calls del_timer() to delete the "slow_task->timer". However,
if the timer handler sas_task_internal_timedout() is running,
the del_timer() in smp_execute_task_sg() will not stop it
and the UAF bug will happen. The process is shown below:
(thread 1) | (thread 2)
smp_execute_task_sg() | sas_task_internal_timedout()
... |
del_timer() |
... | ...
sas_free_task(task) |
kfree(task->slow_task) //FREE|
| task->slow_task->... //USE
Fix by calling del_timer_sync() in smp_execute_task_sg(),
which makes sure the timer handler have finished before
the "task->slow_task" is deallocated.
Fixes: 2908d778ab3e ("[SCSI] aic94xx: new driver")
Signed-off-by: Duoming Zhou <[email protected]>
---
drivers/scsi/libsas/sas_expander.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/scsi/libsas/sas_expander.c b/drivers/scsi/libsas/sas_expander.c
index fa2209080cc..5ce25183010 100644
--- a/drivers/scsi/libsas/sas_expander.c
+++ b/drivers/scsi/libsas/sas_expander.c
@@ -67,7 +67,7 @@ static int smp_execute_task_sg(struct domain_device *dev,
res = i->dft->lldd_execute_task(task, GFP_KERNEL);
if (res) {
- del_timer(&task->slow_task->timer);
+ del_timer_sync(&task->slow_task->timer);
pr_notice("executing SMP task failed:%d\n", res);
break;
}
--
2.17.1
On 2022/9/20 22:42, Duoming Zhou wrote:
> When executing SMP task failed, the smp_execute_task_sg()
> calls del_timer() to delete the "slow_task->timer". However,
> if the timer handler sas_task_internal_timedout() is running,
> the del_timer() in smp_execute_task_sg() will not stop it
> and the UAF bug will happen. The process is shown below:
>
> (thread 1) | (thread 2)
> smp_execute_task_sg() | sas_task_internal_timedout()
> ... |
> del_timer() |
> ... | ...
> sas_free_task(task) |
> kfree(task->slow_task) //FREE|
> | task->slow_task->... //USE
>
> Fix by calling del_timer_sync() in smp_execute_task_sg(),
> which makes sure the timer handler have finished before
> the "task->slow_task" is deallocated.
>
> Fixes: 2908d778ab3e ("[SCSI] aic94xx: new driver")
> Signed-off-by: Duoming Zhou <[email protected]>
> ---
> drivers/scsi/libsas/sas_expander.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/scsi/libsas/sas_expander.c b/drivers/scsi/libsas/sas_expander.c
> index fa2209080cc..5ce25183010 100644
> --- a/drivers/scsi/libsas/sas_expander.c
> +++ b/drivers/scsi/libsas/sas_expander.c
> @@ -67,7 +67,7 @@ static int smp_execute_task_sg(struct domain_device *dev,
> res = i->dft->lldd_execute_task(task, GFP_KERNEL);
>
> if (res) {
> - del_timer(&task->slow_task->timer);
> + del_timer_sync(&task->slow_task->timer);
> pr_notice("executing SMP task failed:%d\n", res);
> break;
> }
>
The timer should be triggered 10 seconds later. I am curious why we stay
this long before del_timer().
However this change looks good to me.
Reviewed-by: Jason Yan <[email protected]>
Hello,
On Thu, 22 Sep 2022 22:06:30 +0800 Jason Yan wrote:
> On 2022/9/20 22:42, Duoming Zhou wrote:
> > When executing SMP task failed, the smp_execute_task_sg()
> > calls del_timer() to delete the "slow_task->timer". However,
> > if the timer handler sas_task_internal_timedout() is running,
> > the del_timer() in smp_execute_task_sg() will not stop it
> > and the UAF bug will happen. The process is shown below:
> >
> > (thread 1) | (thread 2)
> > smp_execute_task_sg() | sas_task_internal_timedout()
> > ... |
> > del_timer() |
> > ... | ...
> > sas_free_task(task) |
> > kfree(task->slow_task) //FREE|
> > | task->slow_task->... //USE
> >
> > Fix by calling del_timer_sync() in smp_execute_task_sg(),
> > which makes sure the timer handler have finished before
> > the "task->slow_task" is deallocated.
> >
> > Fixes: 2908d778ab3e ("[SCSI] aic94xx: new driver")
> > Signed-off-by: Duoming Zhou <[email protected]>
> > ---
> > drivers/scsi/libsas/sas_expander.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/scsi/libsas/sas_expander.c b/drivers/scsi/libsas/sas_expander.c
> > index fa2209080cc..5ce25183010 100644
> > --- a/drivers/scsi/libsas/sas_expander.c
> > +++ b/drivers/scsi/libsas/sas_expander.c
> > @@ -67,7 +67,7 @@ static int smp_execute_task_sg(struct domain_device *dev,
> > res = i->dft->lldd_execute_task(task, GFP_KERNEL);
> >
> > if (res) {
> > - del_timer(&task->slow_task->timer);
> > + del_timer_sync(&task->slow_task->timer);
> > pr_notice("executing SMP task failed:%d\n", res);
> > break;
> > }
> >
>
> The timer should be triggered 10 seconds later. I am curious why we stay
> this long before del_timer().
I think the hacker may create a large number of hardware interrupts to
increase the delayed time.
> However this change looks good to me.
> Reviewed-by: Jason Yan <[email protected]>
Thank you!
Best regards,
Duoming Zhou
Duoming,
> When executing SMP task failed, the smp_execute_task_sg() calls
> del_timer() to delete the "slow_task->timer". However, if the timer
> handler sas_task_internal_timedout() is running, the del_timer() in
> smp_execute_task_sg() will not stop it and the UAF bug will happen.
Applied to 6.1/scsi-staging, thanks!
--
Martin K. Petersen Oracle Linux Engineering