mbox series

[0/6] btrfs-progs: Variant fixes for fuzz-tests

Message ID 20180803055022.9816-1-wqu@suse.com (mailing list archive)
Headers show
Series btrfs-progs: Variant fixes for fuzz-tests | expand

Message

Qu Wenruo Aug. 3, 2018, 5:50 a.m. UTC
This can be fetched from github:
https://github.com/adam900710/btrfs-progs/tree/fixes_for_fuzz_test

The base HEAD is:
commit d7a1b84756157d544a9ddc399ef48c6132eaafcf (david/devel)
Author: Qu Wenruo <wqu@suse.com>
Date:   Thu Jul 5 15:37:31 2018 +0800

    btrfs-progs: check/original: Don't overwrite return value when we failed to repair


Thanks for the already merged fixes for fuzz/003, the remaining part is
pretty small now, +20/-7.

Mostly of the fixes are for fuzz/003, just a small bunch of BUG_ON()
removal. (Patch 1~3 and 5)

There is also a fix for fuzz/003 dead loop. (Patch 4)

Finally we have a fix for fuzz/007, the bug is a segfault triggered by
accessing poisoned list_head, caused by double list freeing. (Patch 6)

Now fuzz-test should finally work without problem.

Qu Wenruo (6):
  btrfs-progs: Exit gracefully if we hit ENOSPC when allocating tree
    block
  btrfs-progs: Exit gracefully when failed to repair root dir item
  btrfs-progs: Don't report dirty leaked eb using BUG_ON
  btrfs-progs: Fix infinite loop when failed to repair bad key order
  btrfs-progs: Exit gracefull when we failed to alloc dev extent
  btrfs-progs: rescue-super: Don't double free fs_devices

 check/main.c    | 10 +++++++++-
 extent-tree.c   |  5 ++++-
 extent_io.c     |  6 +++++-
 super-recover.c |  3 ---
 volumes.c       |  3 ++-
 5 files changed, 20 insertions(+), 7 deletions(-)

Comments

Qu Wenruo Aug. 29, 2018, 5:27 a.m. UTC | #1
Gentle ping.

These fixes are pretty small, I'd like to see them merged before I need
to rebase them again and again.

Thanks,
Qu

On 2018/8/3 下午1:50, Qu Wenruo wrote:
> This can be fetched from github:
> https://github.com/adam900710/btrfs-progs/tree/fixes_for_fuzz_test
> 
> The base HEAD is:
> commit d7a1b84756157d544a9ddc399ef48c6132eaafcf (david/devel)
> Author: Qu Wenruo <wqu@suse.com>
> Date:   Thu Jul 5 15:37:31 2018 +0800
> 
>     btrfs-progs: check/original: Don't overwrite return value when we failed to repair
> 
> 
> Thanks for the already merged fixes for fuzz/003, the remaining part is
> pretty small now, +20/-7.
> 
> Mostly of the fixes are for fuzz/003, just a small bunch of BUG_ON()
> removal. (Patch 1~3 and 5)
> 
> There is also a fix for fuzz/003 dead loop. (Patch 4)
> 
> Finally we have a fix for fuzz/007, the bug is a segfault triggered by
> accessing poisoned list_head, caused by double list freeing. (Patch 6)
> 
> Now fuzz-test should finally work without problem.
> 
> Qu Wenruo (6):
>   btrfs-progs: Exit gracefully if we hit ENOSPC when allocating tree
>     block
>   btrfs-progs: Exit gracefully when failed to repair root dir item
>   btrfs-progs: Don't report dirty leaked eb using BUG_ON
>   btrfs-progs: Fix infinite loop when failed to repair bad key order
>   btrfs-progs: Exit gracefull when we failed to alloc dev extent
>   btrfs-progs: rescue-super: Don't double free fs_devices
> 
>  check/main.c    | 10 +++++++++-
>  extent-tree.c   |  5 ++++-
>  extent_io.c     |  6 +++++-
>  super-recover.c |  3 ---
>  volumes.c       |  3 ++-
>  5 files changed, 20 insertions(+), 7 deletions(-)
>
David Sterba Sept. 11, 2018, 4:38 p.m. UTC | #2
On Wed, Aug 29, 2018 at 01:27:39PM +0800, Qu Wenruo wrote:
> Gentle ping.
> 
> These fixes are pretty small, I'd like to see them merged before I need
> to rebase them again and again.

I've merged them now, thanks. I had to edit all changelogs and remove
the ----- lines, shortened the paths in the test logs and fixed typos.
Please have a look at the committed patches and compare to what you
sent.
Qu Wenruo Sept. 11, 2018, 11:59 p.m. UTC | #3
On 2018/9/12 上午12:38, David Sterba wrote:
> On Wed, Aug 29, 2018 at 01:27:39PM +0800, Qu Wenruo wrote:
>> Gentle ping.
>>
>> These fixes are pretty small, I'd like to see them merged before I need
>> to rebase them again and again.
> 
> I've merged them now, thanks. I had to edit all changelogs and remove
> the ----- lines, shortened the paths in the test logs and fixed typos.

Thanks for the editing.

I'll keep the log format as default for all later patches.

BTW, is there any difference between 2 spaces indent and 1 space indent
for log?

Thanks,
Qu

> Please have a look at the committed patches and compare to what you
> sent.
>
David Sterba Sept. 14, 2018, 2:25 p.m. UTC | #4
On Wed, Sep 12, 2018 at 07:59:34AM +0800, Qu Wenruo wrote:
> 
> 
> On 2018/9/12 上午12:38, David Sterba wrote:
> > On Wed, Aug 29, 2018 at 01:27:39PM +0800, Qu Wenruo wrote:
> >> Gentle ping.
> >>
> >> These fixes are pretty small, I'd like to see them merged before I need
> >> to rebase them again and again.
> > 
> > I've merged them now, thanks. I had to edit all changelogs and remove
> > the ----- lines, shortened the paths in the test logs and fixed typos.
> 
> Thanks for the editing.
> 
> I'll keep the log format as default for all later patches.
> 
> BTW, is there any difference between 2 spaces indent and 1 space indent
> for log?

No, it's for some visual separation so on first glance is obvious what's
the changelog text and what are logs or commands to reproduce. One space
is ok too so the log is not drifting away.