[PATCH] mm,numa: remove BUG_ON in __handle_mm_fault

[PATCH] mm,numa: remove BUG_ON in __handle_mm_fault

[Rik van Riel]

Changing PTEs and PMDs to pte_numa & pmd_numa is done with the
mmap_sem held for reading, which means a pmd can be instantiated
and/or turned into a numa one while __handle_mm_fault is examining
the value of orig_pmd.

If that happens, __handle_mm_fault should just return and let
the page fault retry, instead of throwing an oops.

Signed-off-by: Rik van Riel <riel@redhat.com>
Reported-by: Sunil Pandey <sunil.k.pandey@intel.com>
---
 mm/memory.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index d0f0bef..9edccb2 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3900,8 +3900,9 @@ static int __handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
 		}
 	}
 
-	/* THP should already have been handled */
-	BUG_ON(pmd_numa(*pmd));
+	/* The PMD became NUMA while we examined orig_pmd. Return & retry */
+	if (pmd_numa(*pmd))
+		return 0;
 
 	/*
 	 * Use __pte_alloc instead of pte_alloc_map, because we can't

[PATCH -v2] mm,numa: remove BUG_ON in __handle_mm_fault

[Rik van Riel]

Peter pointed out we can do this slightly simpler, since we already
have a test for pmd_trans_huge(*pmd) below...

---8<---

Changing PTEs and PMDs to pte_numa & pmd_numa is done with the
mmap_sem held for reading, which means a pmd can be instantiated
and turned into a numa one while __handle_mm_fault is examining
the value of old_pmd.

If that happens, __handle_mm_fault should just return and let
the page fault retry, instead of throwing an oops. This is
handled by the test for pmd_trans_huge(*pmd) below.

Signed-off-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Reported-by: Sunil Pandey <sunil.k.pandey@intel.com>
Cc: stable@kernel.org
---
 mm/memory.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index d0f0bef..9c2dc65 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3900,9 +3900,6 @@ static int __handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
 		}
 	}
 
-	/* THP should already have been handled */
-	BUG_ON(pmd_numa(*pmd));
-
 	/*
 	 * Use __pte_alloc instead of pte_alloc_map, because we can't
 	 * run pte_offset_map on the pmd, if an huge pmd could

[tip:sched/core] mm/numa: Remove BUG_ON() in __handle_mm_fault()

[tip-bot for Rik van Riel]

Commit-ID:  107437febd495a50e2cd09c81bbaa84d30e57b07
Gitweb:     http://git.kernel.org/tip/107437febd495a50e2cd09c81bbaa84d30e57b07
Author:     Rik van Riel <riel@redhat.com>
AuthorDate: Tue, 29 Apr 2014 15:36:15 -0400
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Wed, 7 May 2014 13:33:48 +0200

mm/numa: Remove BUG_ON() in __handle_mm_fault()

Changing PTEs and PMDs to pte_numa & pmd_numa is done with the
mmap_sem held for reading, which means a pmd can be instantiated
and turned into a numa one while __handle_mm_fault() is examining
the value of old_pmd.

If that happens, __handle_mm_fault() should just return and let
the page fault retry, instead of throwing an oops. This is
handled by the test for pmd_trans_huge(*pmd) below.

Signed-off-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Reported-by: Sunil Pandey <sunil.k.pandey@intel.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: linux-mm@kvack.org
Cc: lwoodman@redhat.com
Cc: dave.hansen@intel.com
Link: http://lkml.kernel.org/r/20140429153615.2d72098e@annuminas.surriel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 mm/memory.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index d0f0bef..9c2dc65 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3900,9 +3900,6 @@ static int __handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
 		}
 	}
 
-	/* THP should already have been handled */
-	BUG_ON(pmd_numa(*pmd));
-
 	/*
 	 * Use __pte_alloc instead of pte_alloc_map, because we can't
 	 * run pte_offset_map on the pmd, if an huge pmd could

Re: [tip:sched/core] mm/numa: Remove BUG_ON() in __handle_mm_fault()

[Patrick McLean]

On Thu, 8 May 2014 03:43:19 -0700

Could this please be included in 3.14 and 3.15 stable as well.

tip-bot for Rik van Riel <tipbot@zytor.com> wrote:

> Commit-ID:  107437febd495a50e2cd09c81bbaa84d30e57b07
> Gitweb:
> http://git.kernel.org/tip/107437febd495a50e2cd09c81bbaa84d30e57b07
> Author:     Rik van Riel <riel@redhat.com> AuthorDate: Tue, 29 Apr
> 2014 15:36:15 -0400 Committer:  Ingo Molnar <mingo@kernel.org>
> CommitDate: Wed, 7 May 2014 13:33:48 +0200
> 
> mm/numa: Remove BUG_ON() in __handle_mm_fault()
> 
> Changing PTEs and PMDs to pte_numa & pmd_numa is done with the
> mmap_sem held for reading, which means a pmd can be instantiated
> and turned into a numa one while __handle_mm_fault() is examining
> the value of old_pmd.
> 
> If that happens, __handle_mm_fault() should just return and let
> the page fault retry, instead of throwing an oops. This is
> handled by the test for pmd_trans_huge(*pmd) below.
> 
> Signed-off-by: Rik van Riel <riel@redhat.com>
> Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
> Reported-by: Sunil Pandey <sunil.k.pandey@intel.com>
> Signed-off-by: Peter Zijlstra <peterz@infradead.org>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Johannes Weiner <hannes@cmpxchg.org>
> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> Cc: Linus Torvalds <torvalds@linux-foundation.org>
> Cc: Mel Gorman <mgorman@suse.de>
> Cc: linux-mm@kvack.org
> Cc: lwoodman@redhat.com
> Cc: dave.hansen@intel.com
> Link:
> http://lkml.kernel.org/r/20140429153615.2d72098e@annuminas.surriel.com
> Signed-off-by: Ingo Molnar <mingo@kernel.org> ---
>  mm/memory.c | 3 ---
>  1 file changed, 3 deletions(-)
> 
> diff --git a/mm/memory.c b/mm/memory.c
> index d0f0bef..9c2dc65 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -3900,9 +3900,6 @@ static int __handle_mm_fault(struct mm_struct
> *mm, struct vm_area_struct *vma, }
>  	}
>  
> -	/* THP should already have been handled */
> -	BUG_ON(pmd_numa(*pmd));
> -
>  	/*
>  	 * Use __pte_alloc instead of pte_alloc_map, because we can't
>  	 * run pte_offset_map on the pmd, if an huge pmd could
> --
> To unsubscribe from this list: send the line "unsubscribe
> linux-kernel" in the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/