linux-kernel - Re: [PATCH 1/2] mm: Retry migration earlier upon refcount mismatch

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <fd15bf5d-25ec-46d6-aa80-865cf0c925a9@arm.com>
Date: Mon, 12 Aug 2024 12:02:55 +0530
From: Dev Jain <dev.jain@....com>
To: "Huang, Ying" <ying.huang@...el.com>
Cc: akpm@...ux-foundation.org, shuah@...nel.org, david@...hat.com,
 willy@...radead.org, ryan.roberts@....com, anshuman.khandual@....com,
 catalin.marinas@....com, cl@...two.org, vbabka@...e.cz, mhocko@...e.com,
 apopple@...dia.com, osalvador@...e.de, baolin.wang@...ux.alibaba.com,
 dave.hansen@...ux.intel.com, will@...nel.org, baohua@...nel.org,
 ioworker0@...il.com, gshan@...hat.com, mark.rutland@....com,
 kirill.shutemov@...ux.intel.com, hughd@...gle.com, aneesh.kumar@...nel.org,
 yang@...amperecomputing.com, peterx@...hat.com, broonie@...nel.org,
 mgorman@...hsingularity.net, linux-arm-kernel@...ts.infradead.org,
 linux-kernel@...r.kernel.org, linux-mm@...ck.org,
 linux-kselftest@...r.kernel.org
Subject: Re: [PATCH 1/2] mm: Retry migration earlier upon refcount mismatch


On 8/12/24 11:50, Huang, Ying wrote:
> Dev Jain <dev.jain@....com> writes:
>
>> On 8/12/24 11:04, Huang, Ying wrote:
>>> Hi, Dev,
>>>
>>> Dev Jain <dev.jain@....com> writes:
>>>
>>>> As already being done in __migrate_folio(), wherein we backoff if the
>>>> folio refcount is wrong, make this check during the unmapping phase, upon
>>>> the failure of which, the original state of the PTEs will be restored and
>>>> the folio lock will be dropped via migrate_folio_undo_src(), any racing
>>>> thread will make progress and migration will be retried.
>>>>
>>>> Signed-off-by: Dev Jain <dev.jain@....com>
>>>> ---
>>>>    mm/migrate.c | 9 +++++++++
>>>>    1 file changed, 9 insertions(+)
>>>>
>>>> diff --git a/mm/migrate.c b/mm/migrate.c
>>>> index e7296c0fb5d5..477acf996951 100644
>>>> --- a/mm/migrate.c
>>>> +++ b/mm/migrate.c
>>>> @@ -1250,6 +1250,15 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
>>>>    	}
>>>>      	if (!folio_mapped(src)) {
>>>> +		/*
>>>> +		 * Someone may have changed the refcount and maybe sleeping
>>>> +		 * on the folio lock. In case of refcount mismatch, bail out,
>>>> +		 * let the system make progress and retry.
>>>> +		 */
>>>> +		struct address_space *mapping = folio_mapping(src);
>>>> +
>>>> +		if (folio_ref_count(src) != folio_expected_refs(mapping, src))
>>>> +			goto out;
>>>>    		__migrate_folio_record(dst, old_page_state, anon_vma);
>>>>    		return MIGRATEPAGE_UNMAP;
>>>>    	}
>>> Do you have some test results for this?  For example, after applying the
>>> patch, the migration success rate increased XX%, etc.
>> Noting that the migration selftest is operating on a single page,
>> before the patch, the test fails on shared-anon mappings on an
>> average of 10 iterations of move_pages(), and after applying the
>> patch it fails on average of 100 iterations, which makes sense
>> because the unmapping() will get retried 3 + 7 = 10 times.
> Thanks!  What is the test results for
>
> https://lore.kernel.org/all/20240801081657.1386743-1-dev.jain@arm.com/
>
> ?

That solves the problem completely, makes the test pass. Although
that still may not be a good solution since it solves the problem
for this particular case (reader thread faulting and raising refcount).
As David notes, a concurrent read()/write() should also create this
refcount race problem.


>
>>> My understanding for this issue is that the migration success rate can
>>> increase if we undo all changes before retrying.  This is the current
>>> behavior for sync migration, but not for async migration.  If so, we can
>>> use migrate_pages_sync() for async migration too to increase success
>>> rate?  Of course, we need to change the function name and comments.
> --
> Best Regards,
> Huang, Ying
>