Re: [PATCH] mm: kmemleak: Ensure that the task stack is not freed during scanning

[Catalin Marinas]

On Wed, Oct 12, 2016 at 11:54:17AM -0400, CAI Qian wrote:
> ----- Original Message -----
> > From: "Catalin Marinas" <catalin.marinas@xxxxxxx>
> > To: linux-mm@xxxxxxxxx
> > Cc: linux-kernel@xxxxxxxxxxxxxxx, "Andrew Morton" <akpm@xxxxxxxxxxxxxxxxxxxx>, "Andy Lutomirski" <luto@xxxxxxxxxx>,
> > "CAI Qian" <caiqian@xxxxxxxxxx>
> > Sent: Wednesday, October 12, 2016 5:57:03 AM
> > Subject: [PATCH] mm: kmemleak: Ensure that the task stack is not freed during scanning
> > 
> > Commit 68f24b08ee89 ("sched/core: Free the stack early if
> > CONFIG_THREAD_INFO_IN_TASK") may cause the task->stack to be freed
> > during kmemleak_scan() execution, leading to either a NULL pointer
> > fault (if task->stack is NULL) or kmemleak accessing already freed
> > memory. This patch uses the new try_get_task_stack() API to ensure that
> > the task stack is not freed during kmemleak stack scanning.
> > 
> > Fixes: 68f24b08ee89 ("sched/core: Free the stack early if
> > CONFIG_THREAD_INFO_IN_TASK")
> > Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> > Cc: Andy Lutomirski <luto@xxxxxxxxxx>
> > Cc: CAI Qian <caiqian@xxxxxxxxxx>
> > Reported-by: CAI Qian <caiqian@xxxxxxxxxx>
> > Signed-off-by: Catalin Marinas <catalin.marinas@xxxxxxx>
> 
> Tested-by: CAI Qian <caiqian@xxxxxxxxxx>

Thanks.

BTW, I noticed a few false positives reported by kmemleak with the
CONFIG_VMAP_STACK enabled caused by the fact that kmemleak requires two
references (instead of one) to a vmalloc'ed object because of the
vm_struct already containing the address. The cached_stack[] array only
stores the vm_struct rather than the stack address, hence the kmemleak
report. I'll work on a fix/annotation.

-- 
Catalin