* [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
@ 2015-03-03 12:02 ` Leon Yu
0 siblings, 0 replies; 12+ messages in thread
From: Leon Yu @ 2015-03-03 12:02 UTC (permalink / raw)
To: Andrew Morton, Konstantin Khlebnikov, Rik van Riel, Michal Hocko
Cc: Leon Yu, linux-mm, linux-kernel, stable
I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
to 3.19 and had no luck with 4.0-rc1 neither.
So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
prevent endless growth of anon_vma hierarchy"), I found chances are that
unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
error path, its degree will be incorrectly decremented in unlink_anon_vmas()
and eventually underflow when exiting as a result of another call to
unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
for me.
This patch fixes the underflow by dropping dst->anon_vma when allocation
fails. It's safe to do so regardless of original value of dst->anon_vma
because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
Besides, callers don't care dst->anon_vma in such case neither.
Signed-off-by: Leon Yu <chianglungyu@gmail.com>
Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
Cc: stable@vger.kernel.org # v3.19
---
mm/rmap.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/mm/rmap.c b/mm/rmap.c
index 5e3e090..bed3cf2 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -287,6 +287,13 @@ int anon_vma_clone(struct vm_area_struct *dst, struct vm_area_struct *src)
return 0;
enomem_failure:
+ /*
+ * dst->anon_vma is dropped here otherwise its degree can be incorrectly
+ * decremented in unlink_anon_vmas().
+ * We can safely do this because calllers of anon_vma_clone() wouldn't
+ * care dst->anon_vma if anon_vma_clone() failed.
+ */
+ dst->anon_vma = NULL;
unlink_anon_vmas(dst);
return -ENOMEM;
}
--
2.3.1
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
@ 2015-03-03 12:02 ` Leon Yu
0 siblings, 0 replies; 12+ messages in thread
From: Leon Yu @ 2015-03-03 12:02 UTC (permalink / raw)
To: Andrew Morton, Konstantin Khlebnikov, Rik van Riel, Michal Hocko
Cc: Leon Yu, linux-mm, linux-kernel, stable
I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
to 3.19 and had no luck with 4.0-rc1 neither.
So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
prevent endless growth of anon_vma hierarchy"), I found chances are that
unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
error path, its degree will be incorrectly decremented in unlink_anon_vmas()
and eventually underflow when exiting as a result of another call to
unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
for me.
This patch fixes the underflow by dropping dst->anon_vma when allocation
fails. It's safe to do so regardless of original value of dst->anon_vma
because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
Besides, callers don't care dst->anon_vma in such case neither.
Signed-off-by: Leon Yu <chianglungyu@gmail.com>
Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
Cc: stable@vger.kernel.org # v3.19
---
mm/rmap.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/mm/rmap.c b/mm/rmap.c
index 5e3e090..bed3cf2 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -287,6 +287,13 @@ int anon_vma_clone(struct vm_area_struct *dst, struct vm_area_struct *src)
return 0;
enomem_failure:
+ /*
+ * dst->anon_vma is dropped here otherwise its degree can be incorrectly
+ * decremented in unlink_anon_vmas().
+ * We can safely do this because calllers of anon_vma_clone() wouldn't
+ * care dst->anon_vma if anon_vma_clone() failed.
+ */
+ dst->anon_vma = NULL;
unlink_anon_vmas(dst);
return -ENOMEM;
}
--
2.3.1
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
2015-03-03 12:02 ` Leon Yu
@ 2015-03-03 12:36 ` Konstantin Khlebnikov
-1 siblings, 0 replies; 12+ messages in thread
From: Konstantin Khlebnikov @ 2015-03-03 12:36 UTC (permalink / raw)
To: Leon Yu
Cc: Andrew Morton, Rik van Riel, Michal Hocko, linux-mm,
Linux Kernel Mailing List, Stable
On Tue, Mar 3, 2015 at 3:02 PM, Leon Yu <chianglungyu@gmail.com> wrote:
> I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
> to 3.19 and had no luck with 4.0-rc1 neither.
>
> So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
> prevent endless growth of anon_vma hierarchy"), I found chances are that
> unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
> anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
> error path, its degree will be incorrectly decremented in unlink_anon_vmas()
> and eventually underflow when exiting as a result of another call to
> unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
> for me.
>
> This patch fixes the underflow by dropping dst->anon_vma when allocation
> fails. It's safe to do so regardless of original value of dst->anon_vma
> because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
> Besides, callers don't care dst->anon_vma in such case neither.
>
> Signed-off-by: Leon Yu <chianglungyu@gmail.com>
> Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
> Cc: stable@vger.kernel.org # v3.19
Good catch.
Signed-off-by: Konstantin Khlebnikov <koct9i@gmail.com>
That thing already backported into various stable branches so this's
not only for v3.19.
As I see other error paths are fine.
> ---
> mm/rmap.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/mm/rmap.c b/mm/rmap.c
> index 5e3e090..bed3cf2 100644
> --- a/mm/rmap.c
> +++ b/mm/rmap.c
> @@ -287,6 +287,13 @@ int anon_vma_clone(struct vm_area_struct *dst, struct vm_area_struct *src)
> return 0;
>
> enomem_failure:
> + /*
> + * dst->anon_vma is dropped here otherwise its degree can be incorrectly
> + * decremented in unlink_anon_vmas().
> + * We can safely do this because calllers of anon_vma_clone() wouldn't
> + * care dst->anon_vma if anon_vma_clone() failed.
> + */
> + dst->anon_vma = NULL;
> unlink_anon_vmas(dst);
> return -ENOMEM;
> }
> --
> 2.3.1
>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
@ 2015-03-03 12:36 ` Konstantin Khlebnikov
0 siblings, 0 replies; 12+ messages in thread
From: Konstantin Khlebnikov @ 2015-03-03 12:36 UTC (permalink / raw)
To: Leon Yu
Cc: Andrew Morton, Rik van Riel, Michal Hocko, linux-mm,
Linux Kernel Mailing List, Stable
On Tue, Mar 3, 2015 at 3:02 PM, Leon Yu <chianglungyu@gmail.com> wrote:
> I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
> to 3.19 and had no luck with 4.0-rc1 neither.
>
> So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
> prevent endless growth of anon_vma hierarchy"), I found chances are that
> unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
> anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
> error path, its degree will be incorrectly decremented in unlink_anon_vmas()
> and eventually underflow when exiting as a result of another call to
> unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
> for me.
>
> This patch fixes the underflow by dropping dst->anon_vma when allocation
> fails. It's safe to do so regardless of original value of dst->anon_vma
> because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
> Besides, callers don't care dst->anon_vma in such case neither.
>
> Signed-off-by: Leon Yu <chianglungyu@gmail.com>
> Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
> Cc: stable@vger.kernel.org # v3.19
Good catch.
Signed-off-by: Konstantin Khlebnikov <koct9i@gmail.com>
That thing already backported into various stable branches so this's
not only for v3.19.
As I see other error paths are fine.
> ---
> mm/rmap.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/mm/rmap.c b/mm/rmap.c
> index 5e3e090..bed3cf2 100644
> --- a/mm/rmap.c
> +++ b/mm/rmap.c
> @@ -287,6 +287,13 @@ int anon_vma_clone(struct vm_area_struct *dst, struct vm_area_struct *src)
> return 0;
>
> enomem_failure:
> + /*
> + * dst->anon_vma is dropped here otherwise its degree can be incorrectly
> + * decremented in unlink_anon_vmas().
> + * We can safely do this because calllers of anon_vma_clone() wouldn't
> + * care dst->anon_vma if anon_vma_clone() failed.
> + */
> + dst->anon_vma = NULL;
> unlink_anon_vmas(dst);
> return -ENOMEM;
> }
> --
> 2.3.1
>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
2015-03-03 12:02 ` Leon Yu
@ 2015-03-03 13:36 ` Michal Hocko
-1 siblings, 0 replies; 12+ messages in thread
From: Michal Hocko @ 2015-03-03 13:36 UTC (permalink / raw)
To: Leon Yu
Cc: Andrew Morton, Konstantin Khlebnikov, Rik van Riel, linux-mm,
linux-kernel, stable
On Tue 03-03-15 20:02:15, Leon Yu wrote:
> I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
> to 3.19 and had no luck with 4.0-rc1 neither.
>
> So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
> prevent endless growth of anon_vma hierarchy"), I found chances are that
> unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
> anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
> error path, its degree will be incorrectly decremented in unlink_anon_vmas()
> and eventually underflow when exiting as a result of another call to
> unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
> for me.
>
> This patch fixes the underflow by dropping dst->anon_vma when allocation
> fails. It's safe to do so regardless of original value of dst->anon_vma
> because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
> Besides, callers don't care dst->anon_vma in such case neither.
>
> Signed-off-by: Leon Yu <chianglungyu@gmail.com>
> Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
> Cc: stable@vger.kernel.org # v3.19
Reviewed-by: Michal Hocko <mhocko@suse.cz>
I think we can safely remove the following code as well, because it is
anon_vma_clone which is responsible to do all the cleanups.
diff --git a/mm/mmap.c b/mm/mmap.c
index 943c6ad18b1d..06a6076c92e5 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -774,10 +774,8 @@ again: remove_next = 1 + (end > next->vm_end);
importer->anon_vma = exporter->anon_vma;
error = anon_vma_clone(importer, exporter);
- if (error) {
- importer->anon_vma = NULL;
+ if (error)
return error;
- }
}
}
> ---
> mm/rmap.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/mm/rmap.c b/mm/rmap.c
> index 5e3e090..bed3cf2 100644
> --- a/mm/rmap.c
> +++ b/mm/rmap.c
> @@ -287,6 +287,13 @@ int anon_vma_clone(struct vm_area_struct *dst, struct vm_area_struct *src)
> return 0;
>
> enomem_failure:
> + /*
> + * dst->anon_vma is dropped here otherwise its degree can be incorrectly
> + * decremented in unlink_anon_vmas().
> + * We can safely do this because calllers of anon_vma_clone() wouldn't
> + * care dst->anon_vma if anon_vma_clone() failed.
> + */
> + dst->anon_vma = NULL;
> unlink_anon_vmas(dst);
> return -ENOMEM;
> }
> --
> 2.3.1
>
--
Michal Hocko
SUSE Labs
^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
@ 2015-03-03 13:36 ` Michal Hocko
0 siblings, 0 replies; 12+ messages in thread
From: Michal Hocko @ 2015-03-03 13:36 UTC (permalink / raw)
To: Leon Yu
Cc: Andrew Morton, Konstantin Khlebnikov, Rik van Riel, linux-mm,
linux-kernel, stable
On Tue 03-03-15 20:02:15, Leon Yu wrote:
> I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
> to 3.19 and had no luck with 4.0-rc1 neither.
>
> So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
> prevent endless growth of anon_vma hierarchy"), I found chances are that
> unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
> anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
> error path, its degree will be incorrectly decremented in unlink_anon_vmas()
> and eventually underflow when exiting as a result of another call to
> unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
> for me.
>
> This patch fixes the underflow by dropping dst->anon_vma when allocation
> fails. It's safe to do so regardless of original value of dst->anon_vma
> because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
> Besides, callers don't care dst->anon_vma in such case neither.
>
> Signed-off-by: Leon Yu <chianglungyu@gmail.com>
> Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
> Cc: stable@vger.kernel.org # v3.19
Reviewed-by: Michal Hocko <mhocko@suse.cz>
I think we can safely remove the following code as well, because it is
anon_vma_clone which is responsible to do all the cleanups.
diff --git a/mm/mmap.c b/mm/mmap.c
index 943c6ad18b1d..06a6076c92e5 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -774,10 +774,8 @@ again: remove_next = 1 + (end > next->vm_end);
importer->anon_vma = exporter->anon_vma;
error = anon_vma_clone(importer, exporter);
- if (error) {
- importer->anon_vma = NULL;
+ if (error)
return error;
- }
}
}
> ---
> mm/rmap.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/mm/rmap.c b/mm/rmap.c
> index 5e3e090..bed3cf2 100644
> --- a/mm/rmap.c
> +++ b/mm/rmap.c
> @@ -287,6 +287,13 @@ int anon_vma_clone(struct vm_area_struct *dst, struct vm_area_struct *src)
> return 0;
>
> enomem_failure:
> + /*
> + * dst->anon_vma is dropped here otherwise its degree can be incorrectly
> + * decremented in unlink_anon_vmas().
> + * We can safely do this because calllers of anon_vma_clone() wouldn't
> + * care dst->anon_vma if anon_vma_clone() failed.
> + */
> + dst->anon_vma = NULL;
> unlink_anon_vmas(dst);
> return -ENOMEM;
> }
> --
> 2.3.1
>
--
Michal Hocko
SUSE Labs
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
2015-03-03 13:36 ` Michal Hocko
@ 2015-03-03 14:05 ` Leon Yu
-1 siblings, 0 replies; 12+ messages in thread
From: Leon Yu @ 2015-03-03 14:05 UTC (permalink / raw)
To: Michal Hocko
Cc: Andrew Morton, Konstantin Khlebnikov, Rik van Riel, linux-mm,
linux-kernel, stable
On Tue, Mar 3, 2015 at 9:36 PM, Michal Hocko <mhocko@suse.cz> wrote:
> On Tue 03-03-15 20:02:15, Leon Yu wrote:
>> I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
>> to 3.19 and had no luck with 4.0-rc1 neither.
>>
>> So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
>> prevent endless growth of anon_vma hierarchy"), I found chances are that
>> unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
>> anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
>> error path, its degree will be incorrectly decremented in unlink_anon_vmas()
>> and eventually underflow when exiting as a result of another call to
>> unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
>> for me.
>>
>> This patch fixes the underflow by dropping dst->anon_vma when allocation
>> fails. It's safe to do so regardless of original value of dst->anon_vma
>> because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
>> Besides, callers don't care dst->anon_vma in such case neither.
>>
>> Signed-off-by: Leon Yu <chianglungyu@gmail.com>
>> Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
>> Cc: stable@vger.kernel.org # v3.19
>
> Reviewed-by: Michal Hocko <mhocko@suse.cz>
>
> I think we can safely remove the following code as well, because it is
> anon_vma_clone which is responsible to do all the cleanups.
Thanks for the input, I'll send v2 with your cleanup.
- Leon
> diff --git a/mm/mmap.c b/mm/mmap.c
> index 943c6ad18b1d..06a6076c92e5 100644
> --- a/mm/mmap.c
> +++ b/mm/mmap.c
> @@ -774,10 +774,8 @@ again: remove_next = 1 + (end > next->vm_end);
>
> importer->anon_vma = exporter->anon_vma;
> error = anon_vma_clone(importer, exporter);
> - if (error) {
> - importer->anon_vma = NULL;
> + if (error)
> return error;
> - }
> }
> }
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
@ 2015-03-03 14:05 ` Leon Yu
0 siblings, 0 replies; 12+ messages in thread
From: Leon Yu @ 2015-03-03 14:05 UTC (permalink / raw)
To: Michal Hocko
Cc: Andrew Morton, Konstantin Khlebnikov, Rik van Riel, linux-mm,
linux-kernel, stable
On Tue, Mar 3, 2015 at 9:36 PM, Michal Hocko <mhocko@suse.cz> wrote:
> On Tue 03-03-15 20:02:15, Leon Yu wrote:
>> I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
>> to 3.19 and had no luck with 4.0-rc1 neither.
>>
>> So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
>> prevent endless growth of anon_vma hierarchy"), I found chances are that
>> unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
>> anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
>> error path, its degree will be incorrectly decremented in unlink_anon_vmas()
>> and eventually underflow when exiting as a result of another call to
>> unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
>> for me.
>>
>> This patch fixes the underflow by dropping dst->anon_vma when allocation
>> fails. It's safe to do so regardless of original value of dst->anon_vma
>> because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
>> Besides, callers don't care dst->anon_vma in such case neither.
>>
>> Signed-off-by: Leon Yu <chianglungyu@gmail.com>
>> Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
>> Cc: stable@vger.kernel.org # v3.19
>
> Reviewed-by: Michal Hocko <mhocko@suse.cz>
>
> I think we can safely remove the following code as well, because it is
> anon_vma_clone which is responsible to do all the cleanups.
Thanks for the input, I'll send v2 with your cleanup.
- Leon
> diff --git a/mm/mmap.c b/mm/mmap.c
> index 943c6ad18b1d..06a6076c92e5 100644
> --- a/mm/mmap.c
> +++ b/mm/mmap.c
> @@ -774,10 +774,8 @@ again: remove_next = 1 + (end > next->vm_end);
>
> importer->anon_vma = exporter->anon_vma;
> error = anon_vma_clone(importer, exporter);
> - if (error) {
> - importer->anon_vma = NULL;
> + if (error)
> return error;
> - }
> }
> }
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
2015-03-03 12:02 ` Leon Yu
@ 2015-03-03 22:00 ` David Rientjes
-1 siblings, 0 replies; 12+ messages in thread
From: David Rientjes @ 2015-03-03 22:00 UTC (permalink / raw)
To: Leon Yu
Cc: Andrew Morton, Konstantin Khlebnikov, Rik van Riel, Michal Hocko,
linux-mm, linux-kernel, stable
On Tue, 3 Mar 2015, Leon Yu wrote:
> I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
> to 3.19 and had no luck with 4.0-rc1 neither.
>
> So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
> prevent endless growth of anon_vma hierarchy"), I found chances are that
> unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
> anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
> error path, its degree will be incorrectly decremented in unlink_anon_vmas()
> and eventually underflow when exiting as a result of another call to
> unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
> for me.
>
> This patch fixes the underflow by dropping dst->anon_vma when allocation
> fails. It's safe to do so regardless of original value of dst->anon_vma
> because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
> Besides, callers don't care dst->anon_vma in such case neither.
>
> Signed-off-by: Leon Yu <chianglungyu@gmail.com>
> Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
> Cc: stable@vger.kernel.org # v3.19
Acked-by: David Rientjes <rientjes@google.com>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
@ 2015-03-03 22:00 ` David Rientjes
0 siblings, 0 replies; 12+ messages in thread
From: David Rientjes @ 2015-03-03 22:00 UTC (permalink / raw)
To: Leon Yu
Cc: Andrew Morton, Konstantin Khlebnikov, Rik van Riel, Michal Hocko,
linux-mm, linux-kernel, stable
On Tue, 3 Mar 2015, Leon Yu wrote:
> I have constantly stumbled upon "kernel BUG at mm/rmap.c:399!" after upgrading
> to 3.19 and had no luck with 4.0-rc1 neither.
>
> So, after looking into new logic introduced by commit 7a3ef208e662, ("mm:
> prevent endless growth of anon_vma hierarchy"), I found chances are that
> unlink_anon_vmas() is called without incrementing dst->anon_vma->degree in
> anon_vma_clone() due to allocation failure. If dst->anon_vma is not NULL in
> error path, its degree will be incorrectly decremented in unlink_anon_vmas()
> and eventually underflow when exiting as a result of another call to
> unlink_anon_vmas(). That's how "kernel BUG at mm/rmap.c:399!" is triggered
> for me.
>
> This patch fixes the underflow by dropping dst->anon_vma when allocation
> fails. It's safe to do so regardless of original value of dst->anon_vma
> because dst->anon_vma doesn't have valid meaning if anon_vma_clone() fails.
> Besides, callers don't care dst->anon_vma in such case neither.
>
> Signed-off-by: Leon Yu <chianglungyu@gmail.com>
> Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
> Cc: stable@vger.kernel.org # v3.19
Acked-by: David Rientjes <rientjes@google.com>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
2015-03-03 13:36 ` Michal Hocko
@ 2015-03-04 5:54 ` David Rientjes
-1 siblings, 0 replies; 12+ messages in thread
From: David Rientjes @ 2015-03-04 5:54 UTC (permalink / raw)
To: Michal Hocko
Cc: Leon Yu, Andrew Morton, Konstantin Khlebnikov, Rik van Riel,
linux-mm, linux-kernel, stable
On Tue, 3 Mar 2015, Michal Hocko wrote:
> I think we can safely remove the following code as well, because it is
> anon_vma_clone which is responsible to do all the cleanups.
>
> diff --git a/mm/mmap.c b/mm/mmap.c
> index 943c6ad18b1d..06a6076c92e5 100644
> --- a/mm/mmap.c
> +++ b/mm/mmap.c
> @@ -774,10 +774,8 @@ again: remove_next = 1 + (end > next->vm_end);
>
> importer->anon_vma = exporter->anon_vma;
> error = anon_vma_clone(importer, exporter);
> - if (error) {
> - importer->anon_vma = NULL;
> + if (error)
> return error;
> - }
> }
> }
>
When Michal proposes this on top of -mm, feel free to add my
Acked-by: David Rientjes <rientjes@google.com>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention
@ 2015-03-04 5:54 ` David Rientjes
0 siblings, 0 replies; 12+ messages in thread
From: David Rientjes @ 2015-03-04 5:54 UTC (permalink / raw)
To: Michal Hocko
Cc: Leon Yu, Andrew Morton, Konstantin Khlebnikov, Rik van Riel,
linux-mm, linux-kernel, stable
On Tue, 3 Mar 2015, Michal Hocko wrote:
> I think we can safely remove the following code as well, because it is
> anon_vma_clone which is responsible to do all the cleanups.
>
> diff --git a/mm/mmap.c b/mm/mmap.c
> index 943c6ad18b1d..06a6076c92e5 100644
> --- a/mm/mmap.c
> +++ b/mm/mmap.c
> @@ -774,10 +774,8 @@ again: remove_next = 1 + (end > next->vm_end);
>
> importer->anon_vma = exporter->anon_vma;
> error = anon_vma_clone(importer, exporter);
> - if (error) {
> - importer->anon_vma = NULL;
> + if (error)
> return error;
> - }
> }
> }
>
When Michal proposes this on top of -mm, feel free to add my
Acked-by: David Rientjes <rientjes@google.com>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2015-03-04 5:54 UTC | newest]
Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-03-03 12:02 [PATCH] mm: fix anon_vma->degree underflow in anon_vma endless growing prevention Leon Yu
2015-03-03 12:02 ` Leon Yu
2015-03-03 12:36 ` Konstantin Khlebnikov
2015-03-03 12:36 ` Konstantin Khlebnikov
2015-03-03 13:36 ` Michal Hocko
2015-03-03 13:36 ` Michal Hocko
2015-03-03 14:05 ` Leon Yu
2015-03-03 14:05 ` Leon Yu
2015-03-04 5:54 ` David Rientjes
2015-03-04 5:54 ` David Rientjes
2015-03-03 22:00 ` David Rientjes
2015-03-03 22:00 ` David Rientjes
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.