On Wed, 14 Sep 2022 14:49:35 +0800 Shuai Xue <xueshuai@xxxxxxxxxxxxxxxxx> wrote:
> The GHES code calls memory_failure_queue() from IRQ context to queue work
> into workqueue and schedule it on the current CPU. Then the work is
> processed in memory_failure_work_func() by kworker and calls
> memory_failure().
>
> When a page is already poisoned, commit a3f5d80ea401 ("mm,hwpoison: send
> SIGBUS with error virutal address") make memory_failure() call
> kill_accessing_process() that:
>
> - holds mmap locking of current->mm
> - does pagetable walk to find the error virtual address
> - and sends SIGBUS to the current process with error info.
>
> However, the mm of kworker is not valid. Therefore, check mm when killing
> accessing process.
Thanks.
When fixing a bug, please always describe the user-visible effects of
tha bug. I'm thinking "null pointer deref crashes the kernel".
> Fixes: a3f5d80ea401 ("mm,hwpoison: send SIGBUS with error virutal address")
> Signed-off-by: Shuai Xue <xueshuai@xxxxxxxxxxxxxxxxx>
I'll add cc:stable.
> --- a/mm/memory-failure.c
> +++ b/mm/memory-failure.c
> @@ -743,6 +743,9 @@ static int kill_accessing_process(struct task_struct *p, unsigned long pfn,
> };
> priv.tk.tsk = p;
>
> + if (!p->mm)
> + return -EFAULT;
> +
> mmap_read_lock(p->mm);
> ret = walk_page_range(p->mm, 0, TASK_SIZE, &hwp_walk_ops,
> (void *)&priv);
> @@ -751,6 +754,7 @@ static int kill_accessing_process(struct task_struct *p, unsigned long pfn,
> else
> ret = 0;
> mmap_read_unlock(p->mm);
> +
> return ret > 0 ? -EHWPOISON : -EFAULT;
> }
This is an unrelated change which doesn't appear to match the style in
memory-failure.c, so I'll drop this hunk.
