On 10/10/2013 12:52 PM, Eryu Guan wrote:
Gui Hecheng has send a patch for this before:
[PATCH V2] xfstest: an regression test for btrfs balance
https://patchwork.kernel.org/patch/3006261/
Thanks,
Wang
Do subvolume create/delete in loop at back ground and do filesystem
balance at the same time
Do subvolume create/delete in loop at back ground and do filesystem
balance at the same time could trigger
kernel BUG at fs/btrfs/relocation.c:2255!
on 3.12-rc4 kernel.
Signed-off-by: Eryu Guan
---
tests/btrfs/012 | 74 +
tests/btrfs/012.
On 10/10/2013 11:35 AM, Anand Jain wrote:
If 'btrfs_file_extent_item' can contain the ref count it would
solve the current problem quite easily. (problem is that, its
of n * n searches to know data extents with its ref for a given
subvol).
Just considering btrfs_file_extent_item is not eno
If 'btrfs_file_extent_item' can contain the ref count it would
solve the current problem quite easily. (problem is that, its
of n * n searches to know data extents with its ref for a given
subvol).
But what'r the challenge(s) to have ref count in the
btrfs_file_extent_item ? any thoughts
When 'dir' is NULL, after calling extref_get_fields(), add_inode_ref()
can be returning without freeing the 'name' pointer.
Added kfree when necessary.
Signed-off-by: Geyslan G. Bem
---
fs/btrfs/tree-log.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/fs/btrfs/tree-log
Josef,
Thank you. Sending v4.
Geyslan Gregório Bem
hackingbits.com
2013/10/9 Josef Bacik :
> On Wed, Oct 09, 2013 at 08:40:30PM -0300, Geyslan G. Bem wrote:
>> In some cases, add_inode_ref() is returning without freeing
>> the 'name' pointer.
>>
>> Added bail out to explicitly call kfree when n
On Wed, Oct 09, 2013 at 08:40:30PM -0300, Geyslan G. Bem wrote:
> In some cases, add_inode_ref() is returning without freeing
> the 'name' pointer.
>
> Added bail out to explicitly call kfree when necessary.
>
> Signed-off-by: Geyslan G. Bem
> ---
> fs/btrfs/tree-log.c | 10 --
> 1 file
Please,
Analyze [PATCH v3].
Regards,
Geyslan Gregório Bem
hackingbits.com
2013/10/9 Geyslan G. Bem :
> In some cases, add_inode_ref() is returning without freeing
> the 'name' pointer.
>
> Added bail out to explicitly call kfree when necessary.
>
> Signed-off-by: Geyslan G. Bem
> ---
> fs/bt
In some cases, add_inode_ref() is returning without freeing
the 'name' pointer.
Added bail out to explicitly call kfree when necessary.
Signed-off-by: Geyslan G. Bem
---
fs/btrfs/tree-log.c | 10 --
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/fs/btrfs/tree-log.c b/fs/b
Felipe, thank you,
Sending v2.
Geyslan Gregório Bem
hackingbits.com
2013/10/9 Felipe Pena :
> Hi,
>
> On Wed, Oct 9, 2013 at 8:13 PM, Geyslan G. Bem wrote:
>> In some cases, add_inode_ref() is returning without freeing
>> the 'name' pointer.
>>
>> Added bail out to explicitly call kfree when n
In some cases, add_inode_ref() is returning without freeing
the 'name' pointer.
Added bail out to explicitly call kfree when necessary.
Signed-off-by: Geyslan G. Bem
---
fs/btrfs/tree-log.c | 11 ---
1 file changed, 8 insertions(+), 3 deletions(-)
diff --git a/fs/btrfs/tree-log.c b/fs/
Hi,
On Wed, Oct 9, 2013 at 8:13 PM, Geyslan G. Bem wrote:
> In some cases, add_inode_ref() is returning without freeing
> the 'name' pointer.
>
> Added bail out to explicitly call kfree when necessary.
>
> Signed-off-by: Geyslan G. Bem
> ---
> fs/btrfs/tree-log.c | 13 ++---
> 1 file ch
In some cases, add_inode_ref() is returning without freeing
the 'name' pointer.
Added bail out to explicitly call kfree when necessary.
Signed-off-by: Geyslan G. Bem
---
fs/btrfs/tree-log.c | 13 ++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/fs/btrfs/tree-log.c b/
A user reported a regression where we could no longer rename a subvolume into
another subvolume. This is a test case to do just that to make sure we don't
regress on this again. Thanks,
Signed-off-by: Josef Bacik
---
tests/btrfs/016 | 61
On Wed, Oct 09, 2013 at 11:17:43AM -0600, Chris Murphy wrote:
>
> On Oct 9, 2013, at 10:27 AM, Josef Bacik wrote:
>
> > On Tue, Oct 08, 2013 at 03:58:23PM -0600, Chris Murphy wrote:
> >> I don't think this is expected, is it? I can no longer move a subvolume
> >> into another subvolume. I can m
On Oct 9, 2013, at 10:27 AM, Josef Bacik wrote:
> On Tue, Oct 08, 2013 at 03:58:23PM -0600, Chris Murphy wrote:
>> I don't think this is expected, is it? I can no longer move a subvolume into
>> another subvolume. I can move a subvolume into a directory. This happens
>> with 3.11.3, and 3.12 r
On Wed, Oct 09, 2013 at 12:26:42PM -0400, Josef Bacik wrote:
> btrfs_rename was using the root of the old dir instead of the root of the new
> dir when checking for a hash collision, so if you tried to move a file into a
> subvol it would freak out because it would see the file you are trying to mo
On Tue, Oct 08, 2013 at 03:58:23PM -0600, Chris Murphy wrote:
> I don't think this is expected, is it? I can no longer move a subvolume into
> another subvolume. I can move a subvolume into a directory. This happens with
> 3.11.3, and 3.12 rc4. I'm not sure yet when the regression first appeared.
btrfs_rename was using the root of the old dir instead of the root of the new
dir when checking for a hash collision, so if you tried to move a file into a
subvol it would freak out because it would see the file you are trying to move
in its current root. This fixes the bug where this would fail
So both Liu and I made huge messes of find_lock_delalloc_range trying to fix
stuff, me first by fixing extent size, then him by fixing something I broke and
then me again telling him to fix it a different way. So this is obviously a
candidate for some testing. This patch adds a pseudo fs so we ca
In summary:
Looks like minimal damage remains and yet I'm still suffering
"Input/output error" from btrfs and btrfsck appears to have looped...
A diff check suggests the damage to be in one (heavily linked to) tree
of a few MBytes.
Would a scrub clear out the damaged trees?
Worth debugging?
T
On Wed, Oct 09, 2013 at 11:53:23PM +0800, Eryu Guan wrote:
> Hi all,
>
> I hit a BUG_ON while playing with btrfs balance. Kernel is
> 3.12-rc4.
>
> The reproducer is a script doing subvolume create/mount/umount/delete
> in loop at back ground and do balance at the same time.
>
> It's reproducabl
Hi all,
I hit a BUG_ON while playing with btrfs balance. Kernel is
3.12-rc4.
The reproducer is a script doing subvolume create/mount/umount/delete
in loop at back ground and do balance at the same time.
It's reproducable everytime. And seems like this is a regression, I
cannot reproduce it on 3.
Hello David, Anand,
> On Mon, Oct 07, 2013 at 10:47:56AM +0800, Anand Jain wrote:
>> I was also thinking if this should be inside kernel
>> facilitated by a new ioctl? so that we avoid number
>> of search ioctl thats required.
>
> I think so. And for the feature itself, it can be handy in case
On Wed, Oct 09, 2013 at 08:57:12AM -0400, Josef Bacik wrote:
> On Wed, Oct 09, 2013 at 06:07:20PM +0800, Liu Bo wrote:
> > On Tue, Oct 08, 2013 at 01:31:57PM -0400, Josef Bacik wrote:
> > > Liu fixed part of this problem and unfortunately I steered him in
> > > slightly the
> > > wrong direction a
On Wed, Oct 09, 2013 at 06:07:20PM +0800, Liu Bo wrote:
> On Tue, Oct 08, 2013 at 01:31:57PM -0400, Josef Bacik wrote:
> > Liu fixed part of this problem and unfortunately I steered him in slightly
> > the
> > wrong direction and so didn't completely fix the problem. The problem is we
> > limit t
On Tue, Oct 08, 2013 at 01:31:57PM -0400, Josef Bacik wrote:
> Liu fixed part of this problem and unfortunately I steered him in slightly the
> wrong direction and so didn't completely fix the problem. The problem is we
> limit the size of the delalloc range we are looking for to max bytes and the
On 10/09/2013 04:03 PM, Anand Jain wrote:
> Wang,
>
>
>> 1. First to search inline extent_data
>> 2. Secondly search in extent tree to calculate (extent refs=1)
> but these extents will be of entire fs, how do you filter it for a subvol ?
Sorry, this could not be true as i have expected,(we have to
Wang,
> 1. First to search inline extent_data
> 2. Secondly search in extent tree to calculate (extent refs=1)
but these extents will be of entire fs, how do you filter it for a subvol ?
Thanks, Anand
> This will avoid n*n rather than n+n which is better…
--
To unsubscribe from this list: se
Thank a lot for your responses.
I already knew degraded mount but I wanted to solve the problem in a
consistent way and not continuing to mount in degraded mode.
I tried to add another device and "btrfs device delete missing /mnt"
but it failed. I also tried using balance filters and --force opti
30 matches
Mail list logo
