| 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
| ||
|
Message-ID: <87bks73ix4.fsf@yhuang6-desk2.ccr.corp.intel.com>
Date: Fri, 26 Aug 2022 09:14:31 +0800
From: "Huang, Ying" <ying.huang@...el.com>
To: Alistair Popple <apopple@...dia.com>
Cc: Peter Xu <peterx@...hat.com>, <linux-mm@...ck.org>,
<akpm@...ux-foundation.org>, Nadav Amit <nadav.amit@...il.com>,
huang ying <huang.ying.caritas@...il.com>,
LKML <linux-kernel@...r.kernel.org>,
"Sierra Guiza, Alejandro (Alex)" <alex.sierra@....com>,
Felix Kuehling <Felix.Kuehling@....com>,
"Jason Gunthorpe" <jgg@...dia.com>,
John Hubbard <jhubbard@...dia.com>,
"David Hildenbrand" <david@...hat.com>,
Ralph Campbell <rcampbell@...dia.com>,
Matthew Wilcox <willy@...radead.org>,
Karol Herbst <kherbst@...hat.com>,
Lyude Paul <lyude@...hat.com>, Ben Skeggs <bskeggs@...hat.com>,
"Logan Gunthorpe" <logang@...tatee.com>, <paulus@...abs.org>,
<linuxppc-dev@...ts.ozlabs.org>, <stable@...r.kernel.org>
Subject: Re: [PATCH v3 2/3] mm/migrate_device.c: Copy pte dirty bit to page
Alistair Popple <apopple@...dia.com> writes:
> Peter Xu <peterx@...hat.com> writes:
>
>> On Fri, Aug 26, 2022 at 08:21:44AM +1000, Alistair Popple wrote:
>>>
>>> Peter Xu <peterx@...hat.com> writes:
>>>
>>> > On Wed, Aug 24, 2022 at 01:03:38PM +1000, Alistair Popple wrote:
>>> >> migrate_vma_setup() has a fast path in migrate_vma_collect_pmd() that
>>> >> installs migration entries directly if it can lock the migrating page.
>>> >> When removing a dirty pte the dirty bit is supposed to be carried over
>>> >> to the underlying page to prevent it being lost.
>>> >>
>>> >> Currently migrate_vma_*() can only be used for private anonymous
>>> >> mappings. That means loss of the dirty bit usually doesn't result in
>>> >> data loss because these pages are typically not file-backed. However
>>> >> pages may be backed by swap storage which can result in data loss if an
>>> >> attempt is made to migrate a dirty page that doesn't yet have the
>>> >> PageDirty flag set.
>>> >>
>>> >> In this case migration will fail due to unexpected references but the
>>> >> dirty pte bit will be lost. If the page is subsequently reclaimed data
>>> >> won't be written back to swap storage as it is considered uptodate,
>>> >> resulting in data loss if the page is subsequently accessed.
>>> >>
>>> >> Prevent this by copying the dirty bit to the page when removing the pte
>>> >> to match what try_to_migrate_one() does.
>>> >>
>>> >> Signed-off-by: Alistair Popple <apopple@...dia.com>
>>> >> Acked-by: Peter Xu <peterx@...hat.com>
>>> >> Reported-by: Huang Ying <ying.huang@...el.com>
>>> >> Fixes: 8c3328f1f36a ("mm/migrate: migrate_vma() unmap page from vma while collecting pages")
>>> >> Cc: stable@...r.kernel.org
>>> >>
>>> >> ---
>>> >>
>>> >> Changes for v3:
>>> >>
>>> >> - Defer TLB flushing
>>> >> - Split a TLB flushing fix into a separate change.
>>> >>
>>> >> Changes for v2:
>>> >>
>>> >> - Fixed up Reported-by tag.
>>> >> - Added Peter's Acked-by.
>>> >> - Atomically read and clear the pte to prevent the dirty bit getting
>>> >> set after reading it.
>>> >> - Added fixes tag
>>> >> ---
>>> >> mm/migrate_device.c | 9 +++++++--
>>> >> 1 file changed, 7 insertions(+), 2 deletions(-)
>>> >>
>>> >> diff --git a/mm/migrate_device.c b/mm/migrate_device.c
>>> >> index 6a5ef9f..51d9afa 100644
>>> >> --- a/mm/migrate_device.c
>>> >> +++ b/mm/migrate_device.c
>>> >> @@ -7,6 +7,7 @@
>>> >> #include <linux/export.h>
>>> >> #include <linux/memremap.h>
>>> >> #include <linux/migrate.h>
>>> >> +#include <linux/mm.h>
>>> >> #include <linux/mm_inline.h>
>>> >> #include <linux/mmu_notifier.h>
>>> >> #include <linux/oom.h>
>>> >> @@ -196,7 +197,7 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
>>> >> anon_exclusive = PageAnon(page) && PageAnonExclusive(page);
>>> >> if (anon_exclusive) {
>>> >> flush_cache_page(vma, addr, pte_pfn(*ptep));
>>> >> - ptep_clear_flush(vma, addr, ptep);
>>> >> + pte = ptep_clear_flush(vma, addr, ptep);
>>> >>
>>> >> if (page_try_share_anon_rmap(page)) {
>>> >> set_pte_at(mm, addr, ptep, pte);
>>> >> @@ -206,11 +207,15 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
>>> >> goto next;
>>> >> }
>>> >> } else {
>>> >> - ptep_get_and_clear(mm, addr, ptep);
>>> >> + pte = ptep_get_and_clear(mm, addr, ptep);
>>> >> }
>>> >
>>> > I remember that in v2 both flush_cache_page() and ptep_get_and_clear() are
>>> > moved above the condition check so they're called unconditionally. Could
>>> > you explain the rational on why it's changed back (since I think v2 was the
>>> > correct approach)?
>>>
>>> Mainly because I agree with your original comments, that it would be
>>> better to keep the batching of TLB flushing if possible. After the
>>> discussion I don't think there is any issues with HW pte dirty bits
>>> here. There are already other cases where HW needs to get that right
>>> anyway (eg. zap_pte_range).
>>
>> Yes tlb batching was kept, thanks for doing that way. Though if only apply
>> patch 1 we'll have both ptep_clear_flush() and batched flush which seems to
>> be redundant.
>>
>>>
>>> > The other question is if we want to split the patch, would it be better to
>>> > move the tlb changes to patch 1, and leave the dirty bit fix in patch 2?
>>>
>>> Isn't that already the case? Patch 1 moves the TLB flush before the PTL
>>> as suggested, patch 2 atomically copies the dirty bit without changing
>>> any TLB flushing.
>>
>> IMHO it's cleaner to have patch 1 fix batch flush, replace
>> ptep_clear_flush() with ptep_get_and_clear() and update pte properly.
>
> Which ptep_clear_flush() are you referring to? This one?
>
> if (anon_exclusive) {
> flush_cache_page(vma, addr, pte_pfn(*ptep));
> ptep_clear_flush(vma, addr, ptep);
>
> My understanding is that we need to do a flush for anon_exclusive.
>
>> No strong opinions on the layout, but I still think we should drop the
>> redundant ptep_clear_flush() above, meanwhile add the flush_cache_page()
>> properly for !exclusive case too.
>
> Good point, we need flush_cache_page() for !exclusive. Will add.
That can be in another patch. For the patch itself, it looks good to
me. Feel free to add,
Reviewed-by: "Huang, Ying" <ying.huang@...el.com>
Best Regards,
Huang, Ying
Powered by blists - more mailing lists