[v2] filemap: avoid unnecessary major faults in filemap_fault()

Commit Message

Peng Zhang Feb. 6, 2024, 9:26 a.m. UTC

From: ZhangPeng <[e-mail]>

The major fault occurred when using mlockall(MCL_CURRENT | MCL_FUTURE)
in application, which leading to an unexpected performance issue[1].

This caused by temporarily cleared PTE during a read+clear/modify/write
update of the PTE, eg, do_numa_page()/change_pte_range().

For the data segment of the user-mode program, the global variable area
is a private mapping. After the pagecache is loaded, the private anonymous
page is generated after the COW is triggered. Mlockall can lock COW pages
(anonymous pages), but the original file pages cannot be locked and may
be reclaimed. If the global variable (private anon page) is accessed when
vmf->pte is zeroed in numa fault, a file page fault will be triggered.

At this time, the original private file page may have been reclaimed.
If the page cache is not available at this time, a major fault will be
triggered and the file will be read, causing additional overhead.

Fix this by rechecking the PTE without acquiring PTL in filemap_fault()
before triggering a major fault.

Testing file anonymous page read and write page fault performance in ext4
and ramdisk using will-it-scale[2] on a x86 physical machine. The data
is the average change compared with the mainline after the patch is
applied. The test results are within the range of fluctuation, and there
is no obvious difference. The test results are as follows:
	                 processes processes_idle threads threads_idle
ext4    private file write: -1.14%  -0.08%         -1.87%   0.13%
ext4    shared  file write:  0.14%  -0.53%          2.88%  -0.77%
ext4    private file  read:  0.03%  -0.65%         -0.51%  -0.08%
tmpfs   private file write: -0.34%  -0.11%          0.20%   0.15%
tmpfs   shared  file write:  0.96%   0.10%          2.78%  -0.34%
ramdisk private file write: -1.21%  -0.21%         -1.12%   0.11%
ramdisk private file  read:  0.00%  -0.68%         -0.33%  -0.02%

[1] https://lore.kernel.org/linux-mm/[e-mail]/
[2] https://github.com/antonblanchard/will-it-scale/

Suggested-by: "Huang, Ying" <[e-mail]>
Suggested-by: Yin Fengwei <[e-mail]>
Signed-off-by: ZhangPeng <[e-mail]>
Signed-off-by: Kefeng Wang <[e-mail]>
---
v1->v2:
- Add more test results per Huang, Ying
- Add more comments before check PTE per Huang, Ying, David Hildenbrand
  and Yin Fengwei
- Change pte_offset_map_nolock to pte_offset_map as the ptl lock won't
  be used

RFC->v1:
- Add error handling when ptep == NULL per Huang, Ying and Matthew
  Wilcox
- Check the PTE without acquiring PTL in filemap_fault(), suggested by
  Huang, Ying and Yin Fengwei
- Add pmd_none() check before PTE map
- Update commit message and add performance test information

 mm/filemap.c | 34 ++++++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)

Comments

Huang, Ying Feb. 7, 2024, 2:21 a.m. UTC | #1

Peng Zhang <[e-mail]> writes:

> From: ZhangPeng <[e-mail]>
>
> The major fault occurred when using mlockall(MCL_CURRENT | MCL_FUTURE)
> in application, which leading to an unexpected performance issue[1].
>
> This caused by temporarily cleared PTE during a read+clear/modify/write
> update of the PTE, eg, do_numa_page()/change_pte_range().
>
> For the data segment of the user-mode program, the global variable area
> is a private mapping. After the pagecache is loaded, the private anonymous
> page is generated after the COW is triggered. Mlockall can lock COW pages
> (anonymous pages), but the original file pages cannot be locked and may
> be reclaimed. If the global variable (private anon page) is accessed when
> vmf->pte is zeroed in numa fault, a file page fault will be triggered.
>
> At this time, the original private file page may have been reclaimed.
> If the page cache is not available at this time, a major fault will be
> triggered and the file will be read, causing additional overhead.
>
> Fix this by rechecking the PTE without acquiring PTL in filemap_fault()
> before triggering a major fault.
>
> Testing file anonymous page read and write page fault performance in ext4
> and ramdisk using will-it-scale[2] on a x86 physical machine. The data
> is the average change compared with the mainline after the patch is
> applied. The test results are within the range of fluctuation, and there
> is no obvious difference. The test results are as follows:

You still claim that there's no difference in the test results.  If so,
why do you implement the patch?  IMHO, you need to prove your patch can
improve the performance in some cases.

> 	                 processes processes_idle threads threads_idle
> ext4    private file write: -1.14%  -0.08%         -1.87%   0.13%
> ext4    shared  file write:  0.14%  -0.53%          2.88%  -0.77%
> ext4    private file  read:  0.03%  -0.65%         -0.51%  -0.08%
> tmpfs   private file write: -0.34%  -0.11%          0.20%   0.15%
> tmpfs   shared  file write:  0.96%   0.10%          2.78%  -0.34%
> ramdisk private file write: -1.21%  -0.21%         -1.12%   0.11%
> ramdisk private file  read:  0.00%  -0.68%         -0.33%  -0.02%
>
> [1] https://lore.kernel.org/linux-mm/[e-mail]/
> [2] https://github.com/antonblanchard/will-it-scale/
>
> Suggested-by: "Huang, Ying" <[e-mail]>
> Suggested-by: Yin Fengwei <[e-mail]>
> Signed-off-by: ZhangPeng <[e-mail]>
> Signed-off-by: Kefeng Wang <[e-mail]>
> ---
> v1->v2:
> - Add more test results per Huang, Ying
> - Add more comments before check PTE per Huang, Ying, David Hildenbrand
>   and Yin Fengwei
> - Change pte_offset_map_nolock to pte_offset_map as the ptl lock won't
>   be used
>
> RFC->v1:
> - Add error handling when ptep == NULL per Huang, Ying and Matthew
>   Wilcox
> - Check the PTE without acquiring PTL in filemap_fault(), suggested by
>   Huang, Ying and Yin Fengwei
> - Add pmd_none() check before PTE map
> - Update commit message and add performance test information
>
>  mm/filemap.c | 34 ++++++++++++++++++++++++++++++++++
>  1 file changed, 34 insertions(+)
>
> diff --git a/mm/filemap.c b/mm/filemap.c
> index 142864338ca4..a2c1a98bc771 100644
> --- a/mm/filemap.c
> +++ b/mm/filemap.c
> @@ -3238,6 +3238,40 @@ vm_fault_t filemap_fault(struct vm_fault *vmf)
>  			mapping_locked = true;
>  		}
>  	} else {
> +		if (!pmd_none(*vmf->pmd)) {
> +			pte_t *ptep;
> +
> +			ptep = pte_offset_map(vmf->pmd, vmf->address);
> +			if (unlikely(!ptep))
> +				return VM_FAULT_NOPAGE;
> +			/*
> +			 * Recheck PTE as the PTE can be cleared temporarily
> +			 * during a read+clear/modify/write update of the PTE,
> +			 * eg, do_numa_page()/change_pte_range(). This will
> +			 * trigger a major fault, even if we use mlockall,
> +			 * which may affect performance.
> +			 * We don't hold PTL here as acquiring PTL hurts
> +			 * performance. So the check is still racy, but
> +			 * the race window seems small enough.
> +			 *
> +			 * If we lose the race during the check, the page_fault
> +			 * will be triggered. Butthe page table entry lock
> +			 * still make sure the correctness:
> +			 * - If the page cache is not reclaimed, the page_fault
> +			 *   will work like the page fault was served already
> +			 *   and bail out.
> +			 * - If the page cache is reclaimed, the major fault
> +			 *   will be triggered, page cache is filled,
> +			 *   page_fault also work like the page fault was
> +			 *   served already and bail out.
> +			 */

IMHO, this is too long.  It can be shorten to like,

If we lose the race, major fault may be triggered unnecessary.  This
hurts performance but not functionality.

> +			if (unlikely(!pte_none(ptep_get_lockless(ptep))))
> +				ret = VM_FAULT_NOPAGE;
> +			pte_unmap(ptep);
> +			if (unlikely(ret))
> +				return ret;
> +		}
> +
>  		/* No page in the page cache at all */
>  		count_vm_event(PGMAJFAULT);
>  		count_memcg_event_mm(vmf->vma->vm_mm, PGMAJFAULT);

--
Best Regards,
Huang, Ying

Peng Zhang Feb. 26, 2024, 2:55 a.m. UTC | #2

On 2024/2/7 10:21, Huang, Ying wrote:

> Peng Zhang <[e-mail]> writes:
>> From: ZhangPeng <[e-mail]>
>>
>> The major fault occurred when using mlockall(MCL_CURRENT | MCL_FUTURE)
>> in application, which leading to an unexpected performance issue[1].
>>
>> This caused by temporarily cleared PTE during a read+clear/modify/write
>> update of the PTE, eg, do_numa_page()/change_pte_range().
>>
>> For the data segment of the user-mode program, the global variable area
>> is a private mapping. After the pagecache is loaded, the private anonymous
>> page is generated after the COW is triggered. Mlockall can lock COW pages
>> (anonymous pages), but the original file pages cannot be locked and may
>> be reclaimed. If the global variable (private anon page) is accessed when
>> vmf->pte is zeroed in numa fault, a file page fault will be triggered.
>>
>> At this time, the original private file page may have been reclaimed.
>> If the page cache is not available at this time, a major fault will be
>> triggered and the file will be read, causing additional overhead.
>>
>> Fix this by rechecking the PTE without acquiring PTL in filemap_fault()
>> before triggering a major fault.
>>
>> Testing file anonymous page read and write page fault performance in ext4
>> and ramdisk using will-it-scale[2] on a x86 physical machine. The data
>> is the average change compared with the mainline after the patch is
>> applied. The test results are within the range of fluctuation, and there
>> is no obvious difference. The test results are as follows:
> You still claim that there's no difference in the test results.  If so,
> why do you implement the patch?  IMHO, you need to prove your patch can
> improve the performance in some cases.

I'm sorry that maybe I didn't express myself clearly.

The purpose of this patch is to fix the issue that major fault may still be triggered
with mlockall(), thereby improving a little performance. This patch is more of a bugfix
than a performance improvement patch.

This issue affects our traffic analysis service. The inbound traffic is heavy. If a major
fault occurs, the I/O schedule is triggered and the original I/O is suspended. Generally,
the I/O schedule is 0.7 ms. If other applications are operating disks, the system needs
to wait for more than 10 ms. However, the inbound traffic is heavy and the NIC buffer is
small. As a result, packet loss occurs. The traffic analysis service can't tolerate packet
loss.

To prevent packet loss, we use the mlockall() function to prevent I/O. It is unreasonable
that major faults will still be triggered after mlockall() is used.

In our service test environment, the baseline is 7 major faults/12 hours. After applied the
unlock patch, the probability of triggering the major fault is 1 major faults/12 hours. After
applied the lock patch, no major fault will be triggered. So only the locked patch can actually
solve our problem.

The test data provided is intended to prove that the patch does not have a major impact
on the performance of the page fault itself.

>> 	                 processes processes_idle threads threads_idle
>> ext4    private file write: -1.14%  -0.08%         -1.87%   0.13%
>> ext4    shared  file write:  0.14%  -0.53%          2.88%  -0.77%
>> ext4    private file  read:  0.03%  -0.65%         -0.51%  -0.08%
>> tmpfs   private file write: -0.34%  -0.11%          0.20%   0.15%
>> tmpfs   shared  file write:  0.96%   0.10%          2.78%  -0.34%
>> ramdisk private file write: -1.21%  -0.21%         -1.12%   0.11%
>> ramdisk private file  read:  0.00%  -0.68%         -0.33%  -0.02%
>>
>> [1] https://lore.kernel.org/linux-mm/[e-mail]/
>> [2] https://github.com/antonblanchard/will-it-scale/
>>
>> Suggested-by: "Huang, Ying" <[e-mail]>
>> Suggested-by: Yin Fengwei <[e-mail]>
>> Signed-off-by: ZhangPeng <[e-mail]>
>> Signed-off-by: Kefeng Wang <[e-mail]>
>> ---
>> v1->v2:
>> - Add more test results per Huang, Ying
>> - Add more comments before check PTE per Huang, Ying, David Hildenbrand
>>    and Yin Fengwei
>> - Change pte_offset_map_nolock to pte_offset_map as the ptl lock won't
>>    be used
>>
>> RFC->v1:
>> - Add error handling when ptep == NULL per Huang, Ying and Matthew
>>    Wilcox
>> - Check the PTE without acquiring PTL in filemap_fault(), suggested by
>>    Huang, Ying and Yin Fengwei
>> - Add pmd_none() check before PTE map
>> - Update commit message and add performance test information
>>
>>   mm/filemap.c | 34 ++++++++++++++++++++++++++++++++++
>>   1 file changed, 34 insertions(+)
>>
>> diff --git a/mm/filemap.c b/mm/filemap.c
>> index 142864338ca4..a2c1a98bc771 100644
>> --- a/mm/filemap.c
>> +++ b/mm/filemap.c
>> @@ -3238,6 +3238,40 @@ vm_fault_t filemap_fault(struct vm_fault *vmf)
>>   			mapping_locked = true;
>>   		}
>>   	} else {
>> +		if (!pmd_none(*vmf->pmd)) {
>> +			pte_t *ptep;
>> +
>> +			ptep = pte_offset_map(vmf->pmd, vmf->address);
>> +			if (unlikely(!ptep))
>> +				return VM_FAULT_NOPAGE;
>> +			/*
>> +			 * Recheck PTE as the PTE can be cleared temporarily
>> +			 * during a read+clear/modify/write update of the PTE,
>> +			 * eg, do_numa_page()/change_pte_range(). This will
>> +			 * trigger a major fault, even if we use mlockall,
>> +			 * which may affect performance.
>> +			 * We don't hold PTL here as acquiring PTL hurts
>> +			 * performance. So the check is still racy, but
>> +			 * the race window seems small enough.
>> +			 *
>> +			 * If we lose the race during the check, the page_fault
>> +			 * will be triggered. Butthe page table entry lock
>> +			 * still make sure the correctness:
>> +			 * - If the page cache is not reclaimed, the page_fault
>> +			 *   will work like the page fault was served already
>> +			 *   and bail out.
>> +			 * - If the page cache is reclaimed, the major fault
>> +			 *   will be triggered, page cache is filled,
>> +			 *   page_fault also work like the page fault was
>> +			 *   served already and bail out.
>> +			 */
> IMHO, this is too long.  It can be shorten to like,
>
> If we lose the race, major fault may be triggered unnecessary.  This
> hurts performance but not functionality.

OK, I'll fix it in the next version.

>> +			if (unlikely(!pte_none(ptep_get_lockless(ptep))))
>> +				ret = VM_FAULT_NOPAGE;
>> +			pte_unmap(ptep);
>> +			if (unlikely(ret))
>> +				return ret;
>> +		}
>> +
>>   		/* No page in the page cache at all */
>>   		count_vm_event(PGMAJFAULT);
>>   		count_memcg_event_mm(vmf->vma->vm_mm, PGMAJFAULT);
> --
> Best Regards,
> Huang, Ying

Huang, Ying Feb. 26, 2024, 6:04 a.m. UTC | #3

"zhangpeng (AS)" <[e-mail]> writes:

> On 2024/2/7 10:21, Huang, Ying wrote:
>
>> Peng Zhang <[e-mail]> writes:
>>> From: ZhangPeng <[e-mail]>
>>>
>>> The major fault occurred when using mlockall(MCL_CURRENT | MCL_FUTURE)
>>> in application, which leading to an unexpected performance issue[1].
>>>
>>> This caused by temporarily cleared PTE during a read+clear/modify/write
>>> update of the PTE, eg, do_numa_page()/change_pte_range().
>>>
>>> For the data segment of the user-mode program, the global variable area
>>> is a private mapping. After the pagecache is loaded, the private anonymous
>>> page is generated after the COW is triggered. Mlockall can lock COW pages
>>> (anonymous pages), but the original file pages cannot be locked and may
>>> be reclaimed. If the global variable (private anon page) is accessed when
>>> vmf->pte is zeroed in numa fault, a file page fault will be triggered.
>>>
>>> At this time, the original private file page may have been reclaimed.
>>> If the page cache is not available at this time, a major fault will be
>>> triggered and the file will be read, causing additional overhead.
>>>
>>> Fix this by rechecking the PTE without acquiring PTL in filemap_fault()
>>> before triggering a major fault.
>>>
>>> Testing file anonymous page read and write page fault performance in ext4
>>> and ramdisk using will-it-scale[2] on a x86 physical machine. The data
>>> is the average change compared with the mainline after the patch is
>>> applied. The test results are within the range of fluctuation, and there
>>> is no obvious difference. The test results are as follows:
>> You still claim that there's no difference in the test results.  If so,
>> why do you implement the patch?  IMHO, you need to prove your patch can
>> improve the performance in some cases.
>
> I'm sorry that maybe I didn't express myself clearly.
>
> The purpose of this patch is to fix the issue that major fault may still be triggered
> with mlockall(), thereby improving a little performance. This patch is more of a bugfix
> than a performance improvement patch.
>
> This issue affects our traffic analysis service. The inbound traffic is heavy. If a major
> fault occurs, the I/O schedule is triggered and the original I/O is suspended. Generally,
> the I/O schedule is 0.7 ms. If other applications are operating disks, the system needs
> to wait for more than 10 ms. However, the inbound traffic is heavy and the NIC buffer is
> small. As a result, packet loss occurs. The traffic analysis service can't tolerate packet
> loss.
>
> To prevent packet loss, we use the mlockall() function to prevent I/O. It is unreasonable
> that major faults will still be triggered after mlockall() is used.
>
> In our service test environment, the baseline is 7 major faults/12 hours. After applied the
> unlock patch, the probability of triggering the major fault is 1 major faults/12 hours. After
> applied the lock patch, no major fault will be triggered. So only the locked patch can actually
> solve our problem.

This is the data I asked for.

But, you said that this is a feature bug fix instead of performance
improvement.  So, I checked the mlock(2), and found the following words,

"
       mlockall() locks all pages mapped into the address space of the calling
       process.  This includes the pages of the code, data, and stack segment,
       as well as shared libraries, user space kernel data, shared memory, and
       memory-mapped files.  All mapped pages are guaranteed to be resident in
       RAM when the call returns successfully; the  pages  are  guaranteed  to
       stay in RAM until later unlocked.
"

In theory, the locked page are in RAM.  So, IIUC, we don't violate the
ABI.  But, in effect, we does do that.

But, from git history, we have cleared the PTE during modification from
2.6.12-rc2 at least.  I guess that because Linux isn't a hard real time
OS, users don't expect that too.

--
Best Regards,
Huang, Ying

> The test data provided is intended to prove that the patch does not have a major impact
> on the performance of the page fault itself.
>
>>> 	                 processes processes_idle threads threads_idle
>>> ext4    private file write: -1.14%  -0.08%         -1.87%   0.13%
>>> ext4    shared  file write:  0.14%  -0.53%          2.88%  -0.77%
>>> ext4    private file  read:  0.03%  -0.65%         -0.51%  -0.08%
>>> tmpfs   private file write: -0.34%  -0.11%          0.20%   0.15%
>>> tmpfs   shared  file write:  0.96%   0.10%          2.78%  -0.34%
>>> ramdisk private file write: -1.21%  -0.21%         -1.12%   0.11%
>>> ramdisk private file  read:  0.00%  -0.68%         -0.33%  -0.02%
>>>
>>> [1] https://lore.kernel.org/linux-mm/[e-mail]/
>>> [2] https://github.com/antonblanchard/will-it-scale/
>>>
>>> Suggested-by: "Huang, Ying" <[e-mail]>
>>> Suggested-by: Yin Fengwei <[e-mail]>
>>> Signed-off-by: ZhangPeng <[e-mail]>
>>> Signed-off-by: Kefeng Wang <[e-mail]>
>>> ---
>>> v1->v2:
>>> - Add more test results per Huang, Ying
>>> - Add more comments before check PTE per Huang, Ying, David Hildenbrand
>>>    and Yin Fengwei
>>> - Change pte_offset_map_nolock to pte_offset_map as the ptl lock won't
>>>    be used
>>>
>>> RFC->v1:
>>> - Add error handling when ptep == NULL per Huang, Ying and Matthew
>>>    Wilcox
>>> - Check the PTE without acquiring PTL in filemap_fault(), suggested by
>>>    Huang, Ying and Yin Fengwei
>>> - Add pmd_none() check before PTE map
>>> - Update commit message and add performance test information
>>>
>>>   mm/filemap.c | 34 ++++++++++++++++++++++++++++++++++
>>>   1 file changed, 34 insertions(+)
>>>
>>> diff --git a/mm/filemap.c b/mm/filemap.c
>>> index 142864338ca4..a2c1a98bc771 100644
>>> --- a/mm/filemap.c
>>> +++ b/mm/filemap.c
>>> @@ -3238,6 +3238,40 @@ vm_fault_t filemap_fault(struct vm_fault *vmf)
>>>   			mapping_locked = true;
>>>   		}
>>>   	} else {
>>> +		if (!pmd_none(*vmf->pmd)) {
>>> +			pte_t *ptep;
>>> +
>>> +			ptep = pte_offset_map(vmf->pmd, vmf->address);
>>> +			if (unlikely(!ptep))
>>> +				return VM_FAULT_NOPAGE;
>>> +			/*
>>> +			 * Recheck PTE as the PTE can be cleared temporarily
>>> +			 * during a read+clear/modify/write update of the PTE,
>>> +			 * eg, do_numa_page()/change_pte_range(). This will
>>> +			 * trigger a major fault, even if we use mlockall,
>>> +			 * which may affect performance.
>>> +			 * We don't hold PTL here as acquiring PTL hurts
>>> +			 * performance. So the check is still racy, but
>>> +			 * the race window seems small enough.
>>> +			 *
>>> +			 * If we lose the race during the check, the page_fault
>>> +			 * will be triggered. Butthe page table entry lock
>>> +			 * still make sure the correctness:
>>> +			 * - If the page cache is not reclaimed, the page_fault
>>> +			 *   will work like the page fault was served already
>>> +			 *   and bail out.
>>> +			 * - If the page cache is reclaimed, the major fault
>>> +			 *   will be triggered, page cache is filled,
>>> +			 *   page_fault also work like the page fault was
>>> +			 *   served already and bail out.
>>> +			 */
>> IMHO, this is too long.  It can be shorten to like,
>>
>> If we lose the race, major fault may be triggered unnecessary.  This
>> hurts performance but not functionality.
>
> OK, I'll fix it in the next version.
>
>>> +			if (unlikely(!pte_none(ptep_get_lockless(ptep))))
>>> +				ret = VM_FAULT_NOPAGE;
>>> +			pte_unmap(ptep);
>>> +			if (unlikely(ret))
>>> +				return ret;
>>> +		}
>>> +
>>>   		/* No page in the page cache at all */
>>>   		count_vm_event(PGMAJFAULT);
>>>   		count_memcg_event_mm(vmf->vma->vm_mm, PGMAJFAULT);
>> --
>> Best Regards,
>> Huang, Ying

Peng Zhang Feb. 26, 2024, 7:33 a.m. UTC | #4

On 2024/2/26 14:04, Huang, Ying wrote:

> "zhangpeng (AS)" <[e-mail]> writes:
>
>> On 2024/2/7 10:21, Huang, Ying wrote:
>>
>>> Peng Zhang <[e-mail]> writes:
>>>> From: ZhangPeng <[e-mail]>
>>>>
>>>> The major fault occurred when using mlockall(MCL_CURRENT | MCL_FUTURE)
>>>> in application, which leading to an unexpected performance issue[1].
>>>>
>>>> This caused by temporarily cleared PTE during a read+clear/modify/write
>>>> update of the PTE, eg, do_numa_page()/change_pte_range().
>>>>
>>>> For the data segment of the user-mode program, the global variable area
>>>> is a private mapping. After the pagecache is loaded, the private anonymous
>>>> page is generated after the COW is triggered. Mlockall can lock COW pages
>>>> (anonymous pages), but the original file pages cannot be locked and may
>>>> be reclaimed. If the global variable (private anon page) is accessed when
>>>> vmf->pte is zeroed in numa fault, a file page fault will be triggered.
>>>>
>>>> At this time, the original private file page may have been reclaimed.
>>>> If the page cache is not available at this time, a major fault will be
>>>> triggered and the file will be read, causing additional overhead.
>>>>
>>>> Fix this by rechecking the PTE without acquiring PTL in filemap_fault()
>>>> before triggering a major fault.
>>>>
>>>> Testing file anonymous page read and write page fault performance in ext4
>>>> and ramdisk using will-it-scale[2] on a x86 physical machine. The data
>>>> is the average change compared with the mainline after the patch is
>>>> applied. The test results are within the range of fluctuation, and there
>>>> is no obvious difference. The test results are as follows:
>>> You still claim that there's no difference in the test results.  If so,
>>> why do you implement the patch?  IMHO, you need to prove your patch can
>>> improve the performance in some cases.
>> I'm sorry that maybe I didn't express myself clearly.
>>
>> The purpose of this patch is to fix the issue that major fault may still be triggered
>> with mlockall(), thereby improving a little performance. This patch is more of a bugfix
>> than a performance improvement patch.
>>
>> This issue affects our traffic analysis service. The inbound traffic is heavy. If a major
>> fault occurs, the I/O schedule is triggered and the original I/O is suspended. Generally,
>> the I/O schedule is 0.7 ms. If other applications are operating disks, the system needs
>> to wait for more than 10 ms. However, the inbound traffic is heavy and the NIC buffer is
>> small. As a result, packet loss occurs. The traffic analysis service can't tolerate packet
>> loss.
>>
>> To prevent packet loss, we use the mlockall() function to prevent I/O. It is unreasonable
>> that major faults will still be triggered after mlockall() is used.
>>
>> In our service test environment, the baseline is 7 major faults/12 hours. After applied the
>> unlock patch, the probability of triggering the major fault is 1 major faults/12 hours. After
>> applied the lock patch, no major fault will be triggered. So only the locked patch can actually
>> solve our problem.
> This is the data I asked for.
>
> But, you said that this is a feature bug fix instead of performance
> improvement.  So, I checked the mlock(2), and found the following words,
>
> "
>         mlockall() locks all pages mapped into the address space of the calling
>         process.  This includes the pages of the code, data, and stack segment,
>         as well as shared libraries, user space kernel data, shared memory, and
>         memory-mapped files.  All mapped pages are guaranteed to be resident in
>         RAM when the call returns successfully; the  pages  are  guaranteed  to
>         stay in RAM until later unlocked.
> "
>
> In theory, the locked page are in RAM.  So, IIUC, we don't violate the
> ABI.  But, in effect, we does do that.

"mlockall() locks all pages mapped into the address space of the calling process."
For a private mapping, mlockall() can lock COW pages (anonymous pages), but the
original file pages can't be locked. Maybe, we violate the ABI here. This is also
the cause of this issue. The patch fix the impact of this issue: prevent major
faults, reduce IO, and fix the service packet loss issue.

Preventing major faults, and thus reducing IO, could be an important reason to use
mlockall(). Could we fix this with the locked patch? Or is there another way?

> But, from git history, we have cleared the PTE during modification from
> 2.6.12-rc2 at least.  I guess that because Linux isn't a hard real time
> OS, users don't expect that too.
>
> --
> Best Regards,
> Huang, Ying
>
>> The test data provided is intended to prove that the patch does not have a major impact
>> on the performance of the page fault itself.
>>
>>>> 	                 processes processes_idle threads threads_idle
>>>> ext4    private file write: -1.14%  -0.08%         -1.87%   0.13%
>>>> ext4    shared  file write:  0.14%  -0.53%          2.88%  -0.77%
>>>> ext4    private file  read:  0.03%  -0.65%         -0.51%  -0.08%
>>>> tmpfs   private file write: -0.34%  -0.11%          0.20%   0.15%
>>>> tmpfs   shared  file write:  0.96%   0.10%          2.78%  -0.34%
>>>> ramdisk private file write: -1.21%  -0.21%         -1.12%   0.11%
>>>> ramdisk private file  read:  0.00%  -0.68%         -0.33%  -0.02%
>>>>
>>>> [1] https://lore.kernel.org/linux-mm/[e-mail]/
>>>> [2] https://github.com/antonblanchard/will-it-scale/
>>>>
>>>> Suggested-by: "Huang, Ying" <[e-mail]>
>>>> Suggested-by: Yin Fengwei <[e-mail]>
>>>> Signed-off-by: ZhangPeng <[e-mail]>
>>>> Signed-off-by: Kefeng Wang <[e-mail]>
>>>> ---
>>>> v1->v2:
>>>> - Add more test results per Huang, Ying
>>>> - Add more comments before check PTE per Huang, Ying, David Hildenbrand
>>>>     and Yin Fengwei
>>>> - Change pte_offset_map_nolock to pte_offset_map as the ptl lock won't
>>>>     be used
>>>>
>>>> RFC->v1:
>>>> - Add error handling when ptep == NULL per Huang, Ying and Matthew
>>>>     Wilcox
>>>> - Check the PTE without acquiring PTL in filemap_fault(), suggested by
>>>>     Huang, Ying and Yin Fengwei
>>>> - Add pmd_none() check before PTE map
>>>> - Update commit message and add performance test information
>>>>
>>>>    mm/filemap.c | 34 ++++++++++++++++++++++++++++++++++
>>>>    1 file changed, 34 insertions(+)
>>>>
>>>> diff --git a/mm/filemap.c b/mm/filemap.c
>>>> index 142864338ca4..a2c1a98bc771 100644
>>>> --- a/mm/filemap.c
>>>> +++ b/mm/filemap.c
>>>> @@ -3238,6 +3238,40 @@ vm_fault_t filemap_fault(struct vm_fault *vmf)
>>>>    			mapping_locked = true;
>>>>    		}
>>>>    	} else {
>>>> +		if (!pmd_none(*vmf->pmd)) {
>>>> +			pte_t *ptep;
>>>> +
>>>> +			ptep = pte_offset_map(vmf->pmd, vmf->address);
>>>> +			if (unlikely(!ptep))
>>>> +				return VM_FAULT_NOPAGE;
>>>> +			/*
>>>> +			 * Recheck PTE as the PTE can be cleared temporarily
>>>> +			 * during a read+clear/modify/write update of the PTE,
>>>> +			 * eg, do_numa_page()/change_pte_range(). This will
>>>> +			 * trigger a major fault, even if we use mlockall,
>>>> +			 * which may affect performance.
>>>> +			 * We don't hold PTL here as acquiring PTL hurts
>>>> +			 * performance. So the check is still racy, but
>>>> +			 * the race window seems small enough.
>>>> +			 *
>>>> +			 * If we lose the race during the check, the page_fault
>>>> +			 * will be triggered. Butthe page table entry lock
>>>> +			 * still make sure the correctness:
>>>> +			 * - If the page cache is not reclaimed, the page_fault
>>>> +			 *   will work like the page fault was served already
>>>> +			 *   and bail out.
>>>> +			 * - If the page cache is reclaimed, the major fault
>>>> +			 *   will be triggered, page cache is filled,
>>>> +			 *   page_fault also work like the page fault was
>>>> +			 *   served already and bail out.
>>>> +			 */
>>> IMHO, this is too long.  It can be shorten to like,
>>>
>>> If we lose the race, major fault may be triggered unnecessary.  This
>>> hurts performance but not functionality.
>> OK, I'll fix it in the next version.
>>
>>>> +			if (unlikely(!pte_none(ptep_get_lockless(ptep))))
>>>> +				ret = VM_FAULT_NOPAGE;
>>>> +			pte_unmap(ptep);
>>>> +			if (unlikely(ret))
>>>> +				return ret;
>>>> +		}
>>>> +
>>>>    		/* No page in the page cache at all */
>>>>    		count_vm_event(PGMAJFAULT);
>>>>    		count_memcg_event_mm(vmf->vma->vm_mm, PGMAJFAULT);
>>> --
>>> Best Regards,
>>> Huang, Ying

Huang, Ying Feb. 26, 2024, 7:52 a.m. UTC | #5

"zhangpeng (AS)" <[e-mail]> writes:

> On 2024/2/26 14:04, Huang, Ying wrote:
>
>> "zhangpeng (AS)" <[e-mail]> writes:
>>
>>> On 2024/2/7 10:21, Huang, Ying wrote:
>>>
>>>> Peng Zhang <[e-mail]> writes:
>>>>> From: ZhangPeng <[e-mail]>
>>>>>
>>>>> The major fault occurred when using mlockall(MCL_CURRENT | MCL_FUTURE)
>>>>> in application, which leading to an unexpected performance issue[1].
>>>>>
>>>>> This caused by temporarily cleared PTE during a read+clear/modify/write
>>>>> update of the PTE, eg, do_numa_page()/change_pte_range().
>>>>>
>>>>> For the data segment of the user-mode program, the global variable area
>>>>> is a private mapping. After the pagecache is loaded, the private anonymous
>>>>> page is generated after the COW is triggered. Mlockall can lock COW pages
>>>>> (anonymous pages), but the original file pages cannot be locked and may
>>>>> be reclaimed. If the global variable (private anon page) is accessed when
>>>>> vmf->pte is zeroed in numa fault, a file page fault will be triggered.
>>>>>
>>>>> At this time, the original private file page may have been reclaimed.
>>>>> If the page cache is not available at this time, a major fault will be
>>>>> triggered and the file will be read, causing additional overhead.
>>>>>
>>>>> Fix this by rechecking the PTE without acquiring PTL in filemap_fault()
>>>>> before triggering a major fault.
>>>>>
>>>>> Testing file anonymous page read and write page fault performance in ext4
>>>>> and ramdisk using will-it-scale[2] on a x86 physical machine. The data
>>>>> is the average change compared with the mainline after the patch is
>>>>> applied. The test results are within the range of fluctuation, and there
>>>>> is no obvious difference. The test results are as follows:
>>>> You still claim that there's no difference in the test results.  If so,
>>>> why do you implement the patch?  IMHO, you need to prove your patch can
>>>> improve the performance in some cases.
>>> I'm sorry that maybe I didn't express myself clearly.
>>>
>>> The purpose of this patch is to fix the issue that major fault may still be triggered
>>> with mlockall(), thereby improving a little performance. This patch is more of a bugfix
>>> than a performance improvement patch.
>>>
>>> This issue affects our traffic analysis service. The inbound traffic is heavy. If a major
>>> fault occurs, the I/O schedule is triggered and the original I/O is suspended. Generally,
>>> the I/O schedule is 0.7 ms. If other applications are operating disks, the system needs
>>> to wait for more than 10 ms. However, the inbound traffic is heavy and the NIC buffer is
>>> small. As a result, packet loss occurs. The traffic analysis service can't tolerate packet
>>> loss.
>>>
>>> To prevent packet loss, we use the mlockall() function to prevent I/O. It is unreasonable
>>> that major faults will still be triggered after mlockall() is used.
>>>
>>> In our service test environment, the baseline is 7 major faults/12 hours. After applied the
>>> unlock patch, the probability of triggering the major fault is 1 major faults/12 hours. After
>>> applied the lock patch, no major fault will be triggered. So only the locked patch can actually
>>> solve our problem.
>> This is the data I asked for.
>>
>> But, you said that this is a feature bug fix instead of performance
>> improvement.  So, I checked the mlock(2), and found the following words,
>>
>> "
>>         mlockall() locks all pages mapped into the address space of the calling
>>         process.  This includes the pages of the code, data, and stack segment,
>>         as well as shared libraries, user space kernel data, shared memory, and
>>         memory-mapped files.  All mapped pages are guaranteed to be resident in
>>         RAM when the call returns successfully; the  pages  are  guaranteed  to
>>         stay in RAM until later unlocked.
>> "
>>
>> In theory, the locked page are in RAM.  So, IIUC, we don't violate the
>> ABI.  But, in effect, we does do that.
>
> "mlockall() locks all pages mapped into the address space of the calling process."
> For a private mapping, mlockall() can lock COW pages (anonymous pages), but the
> original file pages can't be locked. Maybe, we violate the ABI here.

If so, please make it explicit and loudly.

> This is also
> the cause of this issue. The patch fix the impact of this issue: prevent major
> faults, reduce IO, and fix the service packet loss issue.
>
> Preventing major faults, and thus reducing IO, could be an important reason to use
> mlockall(). Could we fix this with the locked patch? Or is there another way?

Unfortunately, locked patch cause performance regressions for more
common cases.  Is it possible for us to change ptep_modify_prot_start()
to use some magic PTE value instead of 0?  That may be possible.  But,
that isn't enough, you need to change all ptep_get_and_clear() users.

--
Best Regards,
Huang, Ying

>> But, from git history, we have cleared the PTE during modification from
>> 2.6.12-rc2 at least.  I guess that because Linux isn't a hard real time
>> OS, users don't expect that too.
>>
>> --
>> Best Regards,
>> Huang, Ying
>>
>>> The test data provided is intended to prove that the patch does not have a major impact
>>> on the performance of the page fault itself.
>>>
>>>>> 	                 processes processes_idle threads threads_idle
>>>>> ext4    private file write: -1.14%  -0.08%         -1.87%   0.13%
>>>>> ext4    shared  file write:  0.14%  -0.53%          2.88%  -0.77%
>>>>> ext4    private file  read:  0.03%  -0.65%         -0.51%  -0.08%
>>>>> tmpfs   private file write: -0.34%  -0.11%          0.20%   0.15%
>>>>> tmpfs   shared  file write:  0.96%   0.10%          2.78%  -0.34%
>>>>> ramdisk private file write: -1.21%  -0.21%         -1.12%   0.11%
>>>>> ramdisk private file  read:  0.00%  -0.68%         -0.33%  -0.02%
>>>>>
>>>>> [1] https://lore.kernel.org/linux-mm/[e-mail]/
>>>>> [2] https://github.com/antonblanchard/will-it-scale/
>>>>>
>>>>> Suggested-by: "Huang, Ying" <[e-mail]>
>>>>> Suggested-by: Yin Fengwei <[e-mail]>
>>>>> Signed-off-by: ZhangPeng <[e-mail]>
>>>>> Signed-off-by: Kefeng Wang <[e-mail]>
>>>>> ---
>>>>> v1->v2:
>>>>> - Add more test results per Huang, Ying
>>>>> - Add more comments before check PTE per Huang, Ying, David Hildenbrand
>>>>>     and Yin Fengwei
>>>>> - Change pte_offset_map_nolock to pte_offset_map as the ptl lock won't
>>>>>     be used
>>>>>
>>>>> RFC->v1:
>>>>> - Add error handling when ptep == NULL per Huang, Ying and Matthew
>>>>>     Wilcox
>>>>> - Check the PTE without acquiring PTL in filemap_fault(), suggested by
>>>>>     Huang, Ying and Yin Fengwei
>>>>> - Add pmd_none() check before PTE map
>>>>> - Update commit message and add performance test information
>>>>>
>>>>>    mm/filemap.c | 34 ++++++++++++++++++++++++++++++++++
>>>>>    1 file changed, 34 insertions(+)
>>>>>
>>>>> diff --git a/mm/filemap.c b/mm/filemap.c
>>>>> index 142864338ca4..a2c1a98bc771 100644
>>>>> --- a/mm/filemap.c
>>>>> +++ b/mm/filemap.c
>>>>> @@ -3238,6 +3238,40 @@ vm_fault_t filemap_fault(struct vm_fault *vmf)
>>>>>    			mapping_locked = true;
>>>>>    		}
>>>>>    	} else {
>>>>> +		if (!pmd_none(*vmf->pmd)) {
>>>>> +			pte_t *ptep;
>>>>> +
>>>>> +			ptep = pte_offset_map(vmf->pmd, vmf->address);
>>>>> +			if (unlikely(!ptep))
>>>>> +				return VM_FAULT_NOPAGE;
>>>>> +			/*
>>>>> +			 * Recheck PTE as the PTE can be cleared temporarily
>>>>> +			 * during a read+clear/modify/write update of the PTE,
>>>>> +			 * eg, do_numa_page()/change_pte_range(). This will
>>>>> +			 * trigger a major fault, even if we use mlockall,
>>>>> +			 * which may affect performance.
>>>>> +			 * We don't hold PTL here as acquiring PTL hurts
>>>>> +			 * performance. So the check is still racy, but
>>>>> +			 * the race window seems small enough.
>>>>> +			 *
>>>>> +			 * If we lose the race during the check, the page_fault
>>>>> +			 * will be triggered. Butthe page table entry lock
>>>>> +			 * still make sure the correctness:
>>>>> +			 * - If the page cache is not reclaimed, the page_fault
>>>>> +			 *   will work like the page fault was served already
>>>>> +			 *   and bail out.
>>>>> +			 * - If the page cache is reclaimed, the major fault
>>>>> +			 *   will be triggered, page cache is filled,
>>>>> +			 *   page_fault also work like the page fault was
>>>>> +			 *   served already and bail out.
>>>>> +			 */
>>>> IMHO, this is too long.  It can be shorten to like,
>>>>
>>>> If we lose the race, major fault may be triggered unnecessary.  This
>>>> hurts performance but not functionality.
>>> OK, I'll fix it in the next version.
>>>
>>>>> +			if (unlikely(!pte_none(ptep_get_lockless(ptep))))
>>>>> +				ret = VM_FAULT_NOPAGE;
>>>>> +			pte_unmap(ptep);
>>>>> +			if (unlikely(ret))
>>>>> +				return ret;
>>>>> +		}
>>>>> +
>>>>>    		/* No page in the page cache at all */
>>>>>    		count_vm_event(PGMAJFAULT);
>>>>>    		count_memcg_event_mm(vmf->vma->vm_mm, PGMAJFAULT);
>>>> --
>>>> Best Regards,
>>>> Huang, Ying

Peng Zhang Feb. 26, 2024, 8:19 a.m. UTC | #6

On 2024/2/26 15:52, Huang, Ying wrote:

> "zhangpeng (AS)" <[e-mail]> writes:
>
>> On 2024/2/26 14:04, Huang, Ying wrote:
>>
>>> "zhangpeng (AS)" <[e-mail]> writes:
>>>
>>>> On 2024/2/7 10:21, Huang, Ying wrote:
>>>>
>>>>> Peng Zhang <[e-mail]> writes:
>>>>>> From: ZhangPeng <[e-mail]>
>>>>>>
>>>>>> The major fault occurred when using mlockall(MCL_CURRENT | MCL_FUTURE)
>>>>>> in application, which leading to an unexpected performance issue[1].
>>>>>>
>>>>>> This caused by temporarily cleared PTE during a read+clear/modify/write
>>>>>> update of the PTE, eg, do_numa_page()/change_pte_range().
>>>>>>
>>>>>> For the data segment of the user-mode program, the global variable area
>>>>>> is a private mapping. After the pagecache is loaded, the private anonymous
>>>>>> page is generated after the COW is triggered. Mlockall can lock COW pages
>>>>>> (anonymous pages), but the original file pages cannot be locked and may
>>>>>> be reclaimed. If the global variable (private anon page) is accessed when
>>>>>> vmf->pte is zeroed in numa fault, a file page fault will be triggered.
>>>>>>
>>>>>> At this time, the original private file page may have been reclaimed.
>>>>>> If the page cache is not available at this time, a major fault will be
>>>>>> triggered and the file will be read, causing additional overhead.
>>>>>>
>>>>>> Fix this by rechecking the PTE without acquiring PTL in filemap_fault()
>>>>>> before triggering a major fault.
>>>>>>
>>>>>> Testing file anonymous page read and write page fault performance in ext4
>>>>>> and ramdisk using will-it-scale[2] on a x86 physical machine. The data
>>>>>> is the average change compared with the mainline after the patch is
>>>>>> applied. The test results are within the range of fluctuation, and there
>>>>>> is no obvious difference. The test results are as follows:
>>>>> You still claim that there's no difference in the test results.  If so,
>>>>> why do you implement the patch?  IMHO, you need to prove your patch can
>>>>> improve the performance in some cases.
>>>> I'm sorry that maybe I didn't express myself clearly.
>>>>
>>>> The purpose of this patch is to fix the issue that major fault may still be triggered
>>>> with mlockall(), thereby improving a little performance. This patch is more of a bugfix
>>>> than a performance improvement patch.
>>>>
>>>> This issue affects our traffic analysis service. The inbound traffic is heavy. If a major
>>>> fault occurs, the I/O schedule is triggered and the original I/O is suspended. Generally,
>>>> the I/O schedule is 0.7 ms. If other applications are operating disks, the system needs
>>>> to wait for more than 10 ms. However, the inbound traffic is heavy and the NIC buffer is
>>>> small. As a result, packet loss occurs. The traffic analysis service can't tolerate packet
>>>> loss.
>>>>
>>>> To prevent packet loss, we use the mlockall() function to prevent I/O. It is unreasonable
>>>> that major faults will still be triggered after mlockall() is used.
>>>>
>>>> In our service test environment, the baseline is 7 major faults/12 hours. After applied the
>>>> unlock patch, the probability of triggering the major fault is 1 major faults/12 hours. After
>>>> applied the lock patch, no major fault will be triggered. So only the locked patch can actually
>>>> solve our problem.
>>> This is the data I asked for.
>>>
>>> But, you said that this is a feature bug fix instead of performance
>>> improvement.  So, I checked the mlock(2), and found the following words,
>>>
>>> "
>>>          mlockall() locks all pages mapped into the address space of the calling
>>>          process.  This includes the pages of the code, data, and stack segment,
>>>          as well as shared libraries, user space kernel data, shared memory, and
>>>          memory-mapped files.  All mapped pages are guaranteed to be resident in
>>>          RAM when the call returns successfully; the  pages  are  guaranteed  to
>>>          stay in RAM until later unlocked.
>>> "
>>>
>>> In theory, the locked page are in RAM.  So, IIUC, we don't violate the
>>> ABI.  But, in effect, we does do that.
>> "mlockall() locks all pages mapped into the address space of the calling process."
>> For a private mapping, mlockall() can lock COW pages (anonymous pages), but the
>> original file pages can't be locked. Maybe, we violate the ABI here.
> If so, please make it explicit and loudly.
>
>> This is also
>> the cause of this issue. The patch fix the impact of this issue: prevent major
>> faults, reduce IO, and fix the service packet loss issue.
>>
>> Preventing major faults, and thus reducing IO, could be an important reason to use
>> mlockall(). Could we fix this with the locked patch? Or is there another way?
> Unfortunately, locked patch cause performance regressions for more
> common cases.  Is it possible for us to change ptep_modify_prot_start()
> to use some magic PTE value instead of 0?  That may be possible.  But,
> that isn't enough, you need to change all ptep_get_and_clear() users.

Thank you very much for your suggestion!
I'll try it this way.

> --
> Best Regards,
> Huang, Ying
>
>>> But, from git history, we have cleared the PTE during modification from
>>> 2.6.12-rc2 at least.  I guess that because Linux isn't a hard real time
>>> OS, users don't expect that too.
>>>
>>> --
>>> Best Regards,
>>> Huang, Ying
>>>
>>>> The test data provided is intended to prove that the patch does not have a major impact
>>>> on the performance of the page fault itself.
>>>>
>>>>>> 	                 processes processes_idle threads threads_idle
>>>>>> ext4    private file write: -1.14%  -0.08%         -1.87%   0.13%
>>>>>> ext4    shared  file write:  0.14%  -0.53%          2.88%  -0.77%
>>>>>> ext4    private file  read:  0.03%  -0.65%         -0.51%  -0.08%
>>>>>> tmpfs   private file write: -0.34%  -0.11%          0.20%   0.15%
>>>>>> tmpfs   shared  file write:  0.96%   0.10%          2.78%  -0.34%
>>>>>> ramdisk private file write: -1.21%  -0.21%         -1.12%   0.11%
>>>>>> ramdisk private file  read:  0.00%  -0.68%         -0.33%  -0.02%
>>>>>>
>>>>>> [1] https://lore.kernel.org/linux-mm/[e-mail]/
>>>>>> [2] https://github.com/antonblanchard/will-it-scale/
>>>>>>
>>>>>> Suggested-by: "Huang, Ying" <[e-mail]>
>>>>>> Suggested-by: Yin Fengwei <[e-mail]>
>>>>>> Signed-off-by: ZhangPeng <[e-mail]>
>>>>>> Signed-off-by: Kefeng Wang <[e-mail]>
>>>>>> ---
>>>>>> v1->v2:
>>>>>> - Add more test results per Huang, Ying
>>>>>> - Add more comments before check PTE per Huang, Ying, David Hildenbrand
>>>>>>      and Yin Fengwei
>>>>>> - Change pte_offset_map_nolock to pte_offset_map as the ptl lock won't
>>>>>>      be used
>>>>>>
>>>>>> RFC->v1:
>>>>>> - Add error handling when ptep == NULL per Huang, Ying and Matthew
>>>>>>      Wilcox
>>>>>> - Check the PTE without acquiring PTL in filemap_fault(), suggested by
>>>>>>      Huang, Ying and Yin Fengwei
>>>>>> - Add pmd_none() check before PTE map
>>>>>> - Update commit message and add performance test information
>>>>>>
>>>>>>     mm/filemap.c | 34 ++++++++++++++++++++++++++++++++++
>>>>>>     1 file changed, 34 insertions(+)
>>>>>>
>>>>>> diff --git a/mm/filemap.c b/mm/filemap.c
>>>>>> index 142864338ca4..a2c1a98bc771 100644
>>>>>> --- a/mm/filemap.c
>>>>>> +++ b/mm/filemap.c
>>>>>> @@ -3238,6 +3238,40 @@ vm_fault_t filemap_fault(struct vm_fault *vmf)
>>>>>>     			mapping_locked = true;
>>>>>>     		}
>>>>>>     	} else {
>>>>>> +		if (!pmd_none(*vmf->pmd)) {
>>>>>> +			pte_t *ptep;
>>>>>> +
>>>>>> +			ptep = pte_offset_map(vmf->pmd, vmf->address);
>>>>>> +			if (unlikely(!ptep))
>>>>>> +				return VM_FAULT_NOPAGE;
>>>>>> +			/*
>>>>>> +			 * Recheck PTE as the PTE can be cleared temporarily
>>>>>> +			 * during a read+clear/modify/write update of the PTE,
>>>>>> +			 * eg, do_numa_page()/change_pte_range(). This will
>>>>>> +			 * trigger a major fault, even if we use mlockall,
>>>>>> +			 * which may affect performance.
>>>>>> +			 * We don't hold PTL here as acquiring PTL hurts
>>>>>> +			 * performance. So the check is still racy, but
>>>>>> +			 * the race window seems small enough.
>>>>>> +			 *
>>>>>> +			 * If we lose the race during the check, the page_fault
>>>>>> +			 * will be triggered. Butthe page table entry lock
>>>>>> +			 * still make sure the correctness:
>>>>>> +			 * - If the page cache is not reclaimed, the page_fault
>>>>>> +			 *   will work like the page fault was served already
>>>>>> +			 *   and bail out.
>>>>>> +			 * - If the page cache is reclaimed, the major fault
>>>>>> +			 *   will be triggered, page cache is filled,
>>>>>> +			 *   page_fault also work like the page fault was
>>>>>> +			 *   served already and bail out.
>>>>>> +			 */
>>>>> IMHO, this is too long.  It can be shorten to like,
>>>>>
>>>>> If we lose the race, major fault may be triggered unnecessary.  This
>>>>> hurts performance but not functionality.
>>>> OK, I'll fix it in the next version.
>>>>
>>>>>> +			if (unlikely(!pte_none(ptep_get_lockless(ptep))))
>>>>>> +				ret = VM_FAULT_NOPAGE;
>>>>>> +			pte_unmap(ptep);
>>>>>> +			if (unlikely(ret))
>>>>>> +				return ret;
>>>>>> +		}
>>>>>> +
>>>>>>     		/* No page in the page cache at all */
>>>>>>     		count_vm_event(PGMAJFAULT);
>>>>>>     		count_memcg_event_mm(vmf->vma->vm_mm, PGMAJFAULT);
>>>>> --
>>>>> Best Regards,
>>>>> Huang, Ying

David Hildenbrand Feb. 26, 2024, 8:20 a.m. UTC | #7

On 26.02.24 08:52, Huang, Ying wrote:
> "zhangpeng (AS)" <[e-mail]> writes:
> 
>> On 2024/2/26 14:04, Huang, Ying wrote:
>>
>>> "zhangpeng (AS)" <[e-mail]> writes:
>>>
>>>> On 2024/2/7 10:21, Huang, Ying wrote:
>>>>
>>>>> Peng Zhang <[e-mail]> writes:
>>>>>> From: ZhangPeng <[e-mail]>
>>>>>>
>>>>>> The major fault occurred when using mlockall(MCL_CURRENT | MCL_FUTURE)
>>>>>> in application, which leading to an unexpected performance issue[1].
>>>>>>
>>>>>> This caused by temporarily cleared PTE during a read+clear/modify/write
>>>>>> update of the PTE, eg, do_numa_page()/change_pte_range().
>>>>>>
>>>>>> For the data segment of the user-mode program, the global variable area
>>>>>> is a private mapping. After the pagecache is loaded, the private anonymous
>>>>>> page is generated after the COW is triggered. Mlockall can lock COW pages
>>>>>> (anonymous pages), but the original file pages cannot be locked and may
>>>>>> be reclaimed. If the global variable (private anon page) is accessed when
>>>>>> vmf->pte is zeroed in numa fault, a file page fault will be triggered.
>>>>>>
>>>>>> At this time, the original private file page may have been reclaimed.
>>>>>> If the page cache is not available at this time, a major fault will be
>>>>>> triggered and the file will be read, causing additional overhead.
>>>>>>
>>>>>> Fix this by rechecking the PTE without acquiring PTL in filemap_fault()
>>>>>> before triggering a major fault.
>>>>>>
>>>>>> Testing file anonymous page read and write page fault performance in ext4
>>>>>> and ramdisk using will-it-scale[2] on a x86 physical machine. The data
>>>>>> is the average change compared with the mainline after the patch is
>>>>>> applied. The test results are within the range of fluctuation, and there
>>>>>> is no obvious difference. The test results are as follows:
>>>>> You still claim that there's no difference in the test results.  If so,
>>>>> why do you implement the patch?  IMHO, you need to prove your patch can
>>>>> improve the performance in some cases.
>>>> I'm sorry that maybe I didn't express myself clearly.
>>>>
>>>> The purpose of this patch is to fix the issue that major fault may still be triggered
>>>> with mlockall(), thereby improving a little performance. This patch is more of a bugfix
>>>> than a performance improvement patch.
>>>>
>>>> This issue affects our traffic analysis service. The inbound traffic is heavy. If a major
>>>> fault occurs, the I/O schedule is triggered and the original I/O is suspended. Generally,
>>>> the I/O schedule is 0.7 ms. If other applications are operating disks, the system needs
>>>> to wait for more than 10 ms. However, the inbound traffic is heavy and the NIC buffer is
>>>> small. As a result, packet loss occurs. The traffic analysis service can't tolerate packet
>>>> loss.
>>>>
>>>> To prevent packet loss, we use the mlockall() function to prevent I/O. It is unreasonable
>>>> that major faults will still be triggered after mlockall() is used.
>>>>
>>>> In our service test environment, the baseline is 7 major faults/12 hours. After applied the
>>>> unlock patch, the probability of triggering the major fault is 1 major faults/12 hours. After
>>>> applied the lock patch, no major fault will be triggered. So only the locked patch can actually
>>>> solve our problem.
>>> This is the data I asked for.
>>>
>>> But, you said that this is a feature bug fix instead of performance
>>> improvement.  So, I checked the mlock(2), and found the following words,
>>>
>>> "
>>>          mlockall() locks all pages mapped into the address space of the calling
>>>          process.  This includes the pages of the code, data, and stack segment,
>>>          as well as shared libraries, user space kernel data, shared memory, and
>>>          memory-mapped files.  All mapped pages are guaranteed to be resident in
>>>          RAM when the call returns successfully; the  pages  are  guaranteed  to
>>>          stay in RAM until later unlocked.
>>> "
>>>
>>> In theory, the locked page are in RAM.  So, IIUC, we don't violate the
>>> ABI.  But, in effect, we does do that.
>>
>> "mlockall() locks all pages mapped into the address space of the calling process."
>> For a private mapping, mlockall() can lock COW pages (anonymous pages), but the
>> original file pages can't be locked. Maybe, we violate the ABI here.
> 
> If so, please make it explicit and loudly.
> 
>> This is also
>> the cause of this issue. The patch fix the impact of this issue: prevent major
>> faults, reduce IO, and fix the service packet loss issue.
>>
>> Preventing major faults, and thus reducing IO, could be an important reason to use
>> mlockall(). Could we fix this with the locked patch? Or is there another way?
> 
> Unfortunately, locked patch cause performance regressions for more
> common cases.  Is it possible for us to change ptep_modify_prot_start()
> to use some magic PTE value instead of 0?  That may be possible.  But,
> that isn't enough, you need to change all ptep_get_and_clear() users.

Trigger (false) major faults for mlocked memory is suboptimal.

Having such pages temporarily not mapped (e.g., page migration) is 
acceptable (pages are in RAM but are getting moved). We handle that 
using nonswap migration entries.

Let me understand the issue first:

1) MAP_PRIVATE file mapping that is mlocked.

2) We caused COW, so we now have an anonymous page mapped. That anon
    page is mlocked.

3) Change of protection (under PT lock) will temporarily clear the PTE

4) Page fault will trigger and find the PTE still cleared (without PT
    lock)

5) We don't realize that there is a page mapped and, therefore, trigger
    a major fault.

Using the PT lock would fix it properly. Doing it as in this patch can 
only be considered an optimization, not a proper fix.

Using a magic PTE to work around "just use the PT lock like everyone 
else" feels a bit odd. The patch states "We don't hold PTL here as 
acquiring PTL hurts performance" -- do we have any numbers on that?

We could special-case that for MLOCK'ed VMAs with MCL_FUTURE, meaning, 
take the PTL to double-check only in such VMAs.

Peng Zhang Feb. 26, 2024, 8:46 a.m. UTC | #8

On 2024/2/26 16:20, David Hildenbrand wrote:

> On 26.02.24 08:52, Huang, Ying wrote:
>> "zhangpeng (AS)" <[e-mail]> writes:
>>
>>> On 2024/2/26 14:04, Huang, Ying wrote:
>>>
>>>> "zhangpeng (AS)" <[e-mail]> writes:
>>>>
>>>>> On 2024/2/7 10:21, Huang, Ying wrote:
>>>>>
>>>>>> Peng Zhang <[e-mail]> writes:
>>>>>>> From: ZhangPeng <[e-mail]>
>>>>>>>
>>>>>>> The major fault occurred when using mlockall(MCL_CURRENT | 
>>>>>>> MCL_FUTURE)
>>>>>>> in application, which leading to an unexpected performance 
>>>>>>> issue[1].
>>>>>>>
>>>>>>> This caused by temporarily cleared PTE during a 
>>>>>>> read+clear/modify/write
>>>>>>> update of the PTE, eg, do_numa_page()/change_pte_range().
>>>>>>>
>>>>>>> For the data segment of the user-mode program, the global 
>>>>>>> variable area
>>>>>>> is a private mapping. After the pagecache is loaded, the private 
>>>>>>> anonymous
>>>>>>> page is generated after the COW is triggered. Mlockall can lock 
>>>>>>> COW pages
>>>>>>> (anonymous pages), but the original file pages cannot be locked 
>>>>>>> and may
>>>>>>> be reclaimed. If the global variable (private anon page) is 
>>>>>>> accessed when
>>>>>>> vmf->pte is zeroed in numa fault, a file page fault will be 
>>>>>>> triggered.
>>>>>>>
>>>>>>> At this time, the original private file page may have been 
>>>>>>> reclaimed.
>>>>>>> If the page cache is not available at this time, a major fault 
>>>>>>> will be
>>>>>>> triggered and the file will be read, causing additional overhead.
>>>>>>>
>>>>>>> Fix this by rechecking the PTE without acquiring PTL in 
>>>>>>> filemap_fault()
>>>>>>> before triggering a major fault.
>>>>>>>
>>>>>>> Testing file anonymous page read and write page fault 
>>>>>>> performance in ext4
>>>>>>> and ramdisk using will-it-scale[2] on a x86 physical machine. 
>>>>>>> The data
>>>>>>> is the average change compared with the mainline after the patch is
>>>>>>> applied. The test results are within the range of fluctuation, 
>>>>>>> and there
>>>>>>> is no obvious difference. The test results are as follows:
>>>>>> You still claim that there's no difference in the test results.  
>>>>>> If so,
>>>>>> why do you implement the patch?  IMHO, you need to prove your 
>>>>>> patch can
>>>>>> improve the performance in some cases.
>>>>> I'm sorry that maybe I didn't express myself clearly.
>>>>>
>>>>> The purpose of this patch is to fix the issue that major fault may 
>>>>> still be triggered
>>>>> with mlockall(), thereby improving a little performance. This 
>>>>> patch is more of a bugfix
>>>>> than a performance improvement patch.
>>>>>
>>>>> This issue affects our traffic analysis service. The inbound 
>>>>> traffic is heavy. If a major
>>>>> fault occurs, the I/O schedule is triggered and the original I/O 
>>>>> is suspended. Generally,
>>>>> the I/O schedule is 0.7 ms. If other applications are operating 
>>>>> disks, the system needs
>>>>> to wait for more than 10 ms. However, the inbound traffic is heavy 
>>>>> and the NIC buffer is
>>>>> small. As a result, packet loss occurs. The traffic analysis 
>>>>> service can't tolerate packet
>>>>> loss.
>>>>>
>>>>> To prevent packet loss, we use the mlockall() function to prevent 
>>>>> I/O. It is unreasonable
>>>>> that major faults will still be triggered after mlockall() is used.
>>>>>
>>>>> In our service test environment, the baseline is 7 major faults/12 
>>>>> hours. After applied the
>>>>> unlock patch, the probability of triggering the major fault is 1 
>>>>> major faults/12 hours. After
>>>>> applied the lock patch, no major fault will be triggered. So only 
>>>>> the locked patch can actually
>>>>> solve our problem.
>>>> This is the data I asked for.
>>>>
>>>> But, you said that this is a feature bug fix instead of performance
>>>> improvement.  So, I checked the mlock(2), and found the following 
>>>> words,
>>>>
>>>> "
>>>>          mlockall() locks all pages mapped into the address space 
>>>> of the calling
>>>>          process.  This includes the pages of the code, data, and 
>>>> stack segment,
>>>>          as well as shared libraries, user space kernel data, 
>>>> shared memory, and
>>>>          memory-mapped files.  All mapped pages are guaranteed to 
>>>> be resident in
>>>>          RAM when the call returns successfully; the  pages are  
>>>> guaranteed  to
>>>>          stay in RAM until later unlocked.
>>>> "
>>>>
>>>> In theory, the locked page are in RAM.  So, IIUC, we don't violate the
>>>> ABI.  But, in effect, we does do that.
>>>
>>> "mlockall() locks all pages mapped into the address space of the 
>>> calling process."
>>> For a private mapping, mlockall() can lock COW pages (anonymous 
>>> pages), but the
>>> original file pages can't be locked. Maybe, we violate the ABI here.
>>
>> If so, please make it explicit and loudly.
>>
>>> This is also
>>> the cause of this issue. The patch fix the impact of this issue: 
>>> prevent major
>>> faults, reduce IO, and fix the service packet loss issue.
>>>
>>> Preventing major faults, and thus reducing IO, could be an important 
>>> reason to use
>>> mlockall(). Could we fix this with the locked patch? Or is there 
>>> another way?
>>
>> Unfortunately, locked patch cause performance regressions for more
>> common cases.  Is it possible for us to change ptep_modify_prot_start()
>> to use some magic PTE value instead of 0?  That may be possible.  But,
>> that isn't enough, you need to change all ptep_get_and_clear() users.
>
> Trigger (false) major faults for mlocked memory is suboptimal.
>
> Having such pages temporarily not mapped (e.g., page migration) is 
> acceptable (pages are in RAM but are getting moved). We handle that 
> using nonswap migration entries.
>
> Let me understand the issue first:
>
> 1) MAP_PRIVATE file mapping that is mlocked.
>
> 2) We caused COW, so we now have an anonymous page mapped. That anon
>    page is mlocked.
>
> 3) Change of protection (under PT lock) will temporarily clear the PTE
>
> 4) Page fault will trigger and find the PTE still cleared (without PT
>    lock)
>
> 5) We don't realize that there is a page mapped and, therefore, trigger
>    a major fault.
>
> Using the PT lock would fix it properly. Doing it as in this patch can 
> only be considered an optimization, not a proper fix.
>
> Using a magic PTE to work around "just use the PT lock like everyone 
> else" feels a bit odd. The patch states "We don't hold PTL here as 
> acquiring PTL hurts performance" -- do we have any numbers on that?
>
Testing file anonymous page read and write page fault performance in ext4
, tmpfs and ramdisk using will-it-scale[2] on a x86 physical machine. The
data is the average change compared with the mainline after the patch is
applied.

with the locked patch:
                           processes processes_idle threads threads_idle
ext4    private file write: -0.51%    0.08%          -0.03%  -0.04%
ext4    shared  file write:  0.135%  -0.531%          2.883% -0.772%
ramdisk private file write: -0.48%    0.23%          -1.08%   0.27%
ramdisk private file  read:  0.07%   -6.90%          -5.85%  -0.70%
tmpfs   private file write: -0.344%  -0.110%          0.200%  0.145%
tmpfs   shared  file write:  0.958%   0.101%          2.781% -0.337%
tmpfs    private file read: -0.16%    0.00%          -0.12%   0.41%

with the no locked patch:
                           processes processes_idle threads threads_idle
ext4    private file write: -1.14%  -0.08%         -1.87%   0.13%
ext4    private file  read:  0.03%  -0.65%         -0.51%  -0.08%
ramdisk private file write: -1.21%  -0.21%         -1.12%   0.11%
ramdisk private file  read:  0.00%  -0.68%         -0.33%  -0.02%

I could also run other tests if needed.

> We could special-case that for MLOCK'ed VMAs with MCL_FUTURE, meaning, 
> take the PTL to double-check only in such VMAs.
>
Agreed. I think this solution is great. Thanks for your suggestion!

Patch

diff --git a/mm/filemap.c b/mm/filemap.c
index 142864338ca4..a2c1a98bc771 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -3238,6 +3238,40 @@  vm_fault_t filemap_fault(struct vm_fault *vmf)
 			mapping_locked = true;
 		}
 	} else {
+		if (!pmd_none(*vmf->pmd)) {
+			pte_t *ptep;
+
+			ptep = pte_offset_map(vmf->pmd, vmf->address);
+			if (unlikely(!ptep))
+				return VM_FAULT_NOPAGE;
+			/*
+			 * Recheck PTE as the PTE can be cleared temporarily
+			 * during a read+clear/modify/write update of the PTE,
+			 * eg, do_numa_page()/change_pte_range(). This will
+			 * trigger a major fault, even if we use mlockall,
+			 * which may affect performance.
+			 * We don't hold PTL here as acquiring PTL hurts
+			 * performance. So the check is still racy, but
+			 * the race window seems small enough.
+			 *
+			 * If we lose the race during the check, the page_fault
+			 * will be triggered. Butthe page table entry lock
+			 * still make sure the correctness:
+			 * - If the page cache is not reclaimed, the page_fault
+			 *   will work like the page fault was served already
+			 *   and bail out.
+			 * - If the page cache is reclaimed, the major fault
+			 *   will be triggered, page cache is filled,
+			 *   page_fault also work like the page fault was
+			 *   served already and bail out.
+			 */
+			if (unlikely(!pte_none(ptep_get_lockless(ptep))))
+				ret = VM_FAULT_NOPAGE;
+			pte_unmap(ptep);
+			if (unlikely(ret))
+				return ret;
+		}
+
 		/* No page in the page cache at all */
 		count_vm_event(PGMAJFAULT);
 		count_memcg_event_mm(vmf->vma->vm_mm, PGMAJFAULT);