From: Andrew Morton <akpm@linux-foundation.org>
To: akpm@linux-foundation.org, apopple@nvidia.com, hughd@google.com,
kirill.shutemov@linux.intel.com, linux-mm@kvack.org,
mm-commits@vger.kernel.org, peterx@redhat.com,
rcampbell@nvidia.com, shy828301@gmail.com,
stable@vger.kernel.org, torvalds@linux-foundation.org,
wangyugui@e16-tech.com, will@kernel.org, willy@infradead.org,
ziy@nvidia.com
Subject: [patch 09/24] mm/thp: fix page_vma_mapped_walk() if THP mapped by ptes
Date: Thu, 24 Jun 2021 18:39:26 -0700 [thread overview]
Message-ID: <20210625013926.7ZTZ9B0S5%akpm@linux-foundation.org> (raw)
In-Reply-To: <20210624183838.ac3161ca4a43989665ac8b2f@linux-foundation.org>
From: Hugh Dickins <hughd@google.com>
Subject: mm/thp: fix page_vma_mapped_walk() if THP mapped by ptes
Running certain tests with a DEBUG_VM kernel would crash within hours, on
the total_mapcount BUG() in split_huge_page_to_list(), while trying to
free up some memory by punching a hole in a shmem huge page: split's
try_to_unmap() was unable to find all the mappings of the page (which, on
a !DEBUG_VM kernel, would then keep the huge page pinned in memory).
Crash dumps showed two tail pages of a shmem huge page remained mapped by
pte: ptes in a non-huge-aligned vma of a gVisor process, at the end of a
long unmapped range; and no page table had yet been allocated for the head
of the huge page to be mapped into.
Although designed to handle these odd misaligned huge-page-mapped-by-pte
cases, page_vma_mapped_walk() falls short by returning false prematurely
when !pmd_present or !pud_present or !p4d_present or !pgd_present: there
are cases when a huge page may span the boundary, with ptes present in the
next.
Restructure page_vma_mapped_walk() as a loop to continue in these cases,
while keeping its layout much as before. Add a step_forward() helper to
advance pvmw->address across those boundaries: originally I tried to use
mm's standard p?d_addr_end() macros, but hit the same crash 512 times less
often: because of the way redundant levels are folded together, but folded
differently in different configurations, it was just too difficult to use
them correctly; and step_forward() is simpler anyway.
Link: https://lkml.kernel.org/r/fedb8632-1798-de42-f39e-873551d5bc81@google.com
Fixes: ace71a19cec5 ("mm: introduce page_vma_mapped_walk()")
Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Alistair Popple <apopple@nvidia.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Peter Xu <peterx@redhat.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Wang Yugui <wangyugui@e16-tech.com>
Cc: Will Deacon <will@kernel.org>
Cc: Yang Shi <shy828301@gmail.com>
Cc: Zi Yan <ziy@nvidia.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
mm/page_vma_mapped.c | 34 +++++++++++++++++++++++++---------
1 file changed, 25 insertions(+), 9 deletions(-)
--- a/mm/page_vma_mapped.c~mm-thp-fix-page_vma_mapped_walk-if-thp-mapped-by-ptes
+++ a/mm/page_vma_mapped.c
@@ -116,6 +116,13 @@ static bool check_pte(struct page_vma_ma
return pfn_is_match(pvmw->page, pfn);
}
+static void step_forward(struct page_vma_mapped_walk *pvmw, unsigned long size)
+{
+ pvmw->address = (pvmw->address + size) & ~(size - 1);
+ if (!pvmw->address)
+ pvmw->address = ULONG_MAX;
+}
+
/**
* page_vma_mapped_walk - check if @pvmw->page is mapped in @pvmw->vma at
* @pvmw->address
@@ -183,16 +190,22 @@ bool page_vma_mapped_walk(struct page_vm
if (pvmw->pte)
goto next_pte;
restart:
- {
+ do {
pgd = pgd_offset(mm, pvmw->address);
- if (!pgd_present(*pgd))
- return false;
+ if (!pgd_present(*pgd)) {
+ step_forward(pvmw, PGDIR_SIZE);
+ continue;
+ }
p4d = p4d_offset(pgd, pvmw->address);
- if (!p4d_present(*p4d))
- return false;
+ if (!p4d_present(*p4d)) {
+ step_forward(pvmw, P4D_SIZE);
+ continue;
+ }
pud = pud_offset(p4d, pvmw->address);
- if (!pud_present(*pud))
- return false;
+ if (!pud_present(*pud)) {
+ step_forward(pvmw, PUD_SIZE);
+ continue;
+ }
pvmw->pmd = pmd_offset(pud, pvmw->address);
/*
@@ -239,7 +252,8 @@ restart:
spin_unlock(ptl);
}
- return false;
+ step_forward(pvmw, PMD_SIZE);
+ continue;
}
if (!map_pte(pvmw))
goto next_pte;
@@ -269,7 +283,9 @@ next_pte:
spin_lock(pvmw->ptl);
}
goto this_pte;
- }
+ } while (pvmw->address < end);
+
+ return false;
}
/**
_
next prev parent reply other threads:[~2021-06-25 1:39 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-25 1:38 incoming Andrew Morton
2021-06-25 1:39 ` [patch 01/24] mm: page_vma_mapped_walk(): use page for pvmw->page Andrew Morton
2021-06-25 1:39 ` [patch 02/24] mm: page_vma_mapped_walk(): settle PageHuge on entry Andrew Morton
2021-06-25 1:39 ` [patch 03/24] mm: page_vma_mapped_walk(): use pmde for *pvmw->pmd Andrew Morton
2021-06-25 1:39 ` [patch 04/24] mm: page_vma_mapped_walk(): prettify PVMW_MIGRATION block Andrew Morton
2021-06-25 1:39 ` [patch 05/24] mm: page_vma_mapped_walk(): crossing page table boundary Andrew Morton
2021-06-25 1:39 ` [patch 06/24] mm: page_vma_mapped_walk(): add a level of indentation Andrew Morton
2021-06-25 1:39 ` [patch 07/24] mm: page_vma_mapped_walk(): use goto instead of while (1) Andrew Morton
2021-06-25 1:39 ` [patch 08/24] mm: page_vma_mapped_walk(): get vma_address_end() earlier Andrew Morton
2021-06-25 1:39 ` Andrew Morton [this message]
2021-06-25 1:39 ` [patch 10/24] mm/thp: another PVMW_SYNC fix in page_vma_mapped_walk() Andrew Morton
2021-06-25 1:39 ` [patch 11/24] nilfs2: fix memory leak in nilfs_sysfs_delete_device_group Andrew Morton
2021-06-25 1:39 ` [patch 12/24] mm/vmalloc: add vmalloc_no_huge Andrew Morton
2021-06-25 1:39 ` [patch 13/24] KVM: s390: prepare for hugepage vmalloc Andrew Morton
2021-06-25 1:39 ` [patch 14/24] mm/vmalloc: unbreak kasan vmalloc support Andrew Morton
2021-06-25 1:39 ` [patch 15/24] kthread_worker: split code for canceling the delayed work timer Andrew Morton
2021-06-25 1:39 ` [patch 16/24] kthread: prevent deadlock when kthread_mod_delayed_work() races with kthread_cancel_delayed_work_sync() Andrew Morton
2021-06-25 1:39 ` [patch 17/24] mm, futex: fix shared futex pgoff on shmem huge page Andrew Morton
2021-06-25 1:39 ` [patch 18/24] mm/memory-failure: use a mutex to avoid memory_failure() races Andrew Morton
2021-06-25 1:39 ` [patch 19/24] mm,hwpoison: return -EHWPOISON to denote that the page has already been poisoned Andrew Morton
2021-06-25 1:40 ` [patch 20/24] mm/hwpoison: do not lock page again when me_huge_page() successfully recovers Andrew Morton
2021-06-25 1:40 ` [patch 21/24] mm/page_alloc: __alloc_pages_bulk(): do bounds check before accessing array Andrew Morton
2021-06-25 1:40 ` [patch 22/24] mm/page_alloc: do bulk array bounds check after checking populated elements Andrew Morton
2021-06-25 1:40 ` [patch 23/24] MAINTAINERS: fix Marek's identity again Andrew Morton
2021-06-25 1:40 ` [patch 24/24] mailmap: add Marek's other e-mail address and identity without diacritics Andrew Morton
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210625013926.7ZTZ9B0S5%akpm@linux-foundation.org \
--to=akpm@linux-foundation.org \
--cc=apopple@nvidia.com \
--cc=hughd@google.com \
--cc=kirill.shutemov@linux.intel.com \
--cc=linux-mm@kvack.org \
--cc=mm-commits@vger.kernel.org \
--cc=peterx@redhat.com \
--cc=rcampbell@nvidia.com \
--cc=shy828301@gmail.com \
--cc=stable@vger.kernel.org \
--cc=torvalds@linux-foundation.org \
--cc=wangyugui@e16-tech.com \
--cc=will@kernel.org \
--cc=willy@infradead.org \
--cc=ziy@nvidia.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).