2022-09-14 07:57:07

by Shuai Xue

Subject: [PATCH] mm,hwpoison: check mm when killing accessing process

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.

Fixes: a3f5d80ea401 ("mm,hwpoison: send SIGBUS with error virutal address")
Signed-off-by: Shuai Xue <[email protected]>
---
mm/memory-failure.c | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 14439806b5ef..7553917ce820 100644
--- 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;
}

--
2.20.1.12.g72788fdb

2022-09-14 23:48:06

by Andrew Morton

Subject: Re: [PATCH] mm,hwpoison: check mm when killing accessing process

On Wed, 14 Sep 2022 14:49:35 +0800 Shuai Xue <[email protected]> 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 <[email protected]>

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.

2022-09-15 01:52:15

by Miaohe Lin

Subject: Re: [PATCH] mm,hwpoison: check mm when killing accessing process

On 2022/9/14 14:49, Shuai Xue 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.
>
> Fixes: a3f5d80ea401 ("mm,hwpoison: send SIGBUS with error virutal address")
> Signed-off-by: Shuai Xue <[email protected]>

Thanks for fixing.

Reviewed-by: Miaohe Lin <[email protected]>

Thanks,
Miaohe Lin


> ---
> mm/memory-failure.c | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
> index 14439806b5ef..7553917ce820 100644
> --- 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;
> }
>
>

2022-09-15 01:57:39

by Naoya Horiguchi

Subject: Re: [PATCH] mm,hwpoison: check mm when killing accessing process

On Wed, Sep 14, 2022 at 02:49:35PM +0800, Shuai Xue 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.
>
> Fixes: a3f5d80ea401 ("mm,hwpoison: send SIGBUS with error virutal address")
> Signed-off-by: Shuai Xue <[email protected]>

Thank you for the patch, the fix itself makes sense to me.
I agree with this patch with applying comments from Andrew.

Acked-by: Naoya Horiguchi <[email protected]>

2022-09-15 02:15:59

by Shuai Xue

Subject: Re: [PATCH] mm,hwpoison: check mm when killing accessing process

在 2022/9/15 AM6:35, Andrew Morton 写道:
> On Wed, 14 Sep 2022 14:49:35 +0800 Shuai Xue <[email protected]> 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".

Got it. Thank you :)

>
>> Fixes: a3f5d80ea401 ("mm,hwpoison: send SIGBUS with error virutal address")
>> Signed-off-by: Shuai Xue <[email protected]>
>
> I'll add cc:stable.

Thanks.

>
>> --- 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.

I see, thanks.

Cheers,
Shuai

2022-09-15 02:27:11

by Shuai Xue

Subject: Re: [PATCH] mm,hwpoison: check mm when killing accessing process

在 2022/9/15 AM9:48, Naoya Horiguchi 写道:
> On Wed, Sep 14, 2022 at 02:49:35PM +0800, Shuai Xue 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.
>>
>> Fixes: a3f5d80ea401 ("mm,hwpoison: send SIGBUS with error virutal address")
>> Signed-off-by: Shuai Xue <[email protected]>
>
> Thank you for the patch, the fix itself makes sense to me.
> I agree with this patch with applying comments from Andrew.
>
> Acked-by: Naoya Horiguchi <[email protected]>

Thank you for prompt review.

Cheers,
Shuai

2022-09-15 03:07:08

by Shuai Xue

Subject: Re: [PATCH] mm,hwpoison: check mm when killing accessing process

在 2022/9/15 AM9:45, Miaohe Lin 写道:
> On 2022/9/14 14:49, Shuai Xue 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.
>>
>> Fixes: a3f5d80ea401 ("mm,hwpoison: send SIGBUS with error virutal address")
>> Signed-off-by: Shuai Xue <[email protected]>
>
> Thanks for fixing.
>
> Reviewed-by: Miaohe Lin <[email protected]>
>
> Thanks,
> Miaohe Lin

Thank you for your review.

Cheers,
Shuai

>
>
>> ---
>> mm/memory-failure.c | 4 ++++
>> 1 file changed, 4 insertions(+)
>>
>> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
>> index 14439806b5ef..7553917ce820 100644
>> --- 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;
>> }
>>
>>