Hi,
This patch works.
From:Dennis Zhou
To: Wang Yugui
Cc: Vlastimil Babka ,
linux...@kvack.org,
linux-btrfs@vger.kernel.org
Date:Thu, 8 Apr 2021 13:48:33 +
Subject: Re: unexpected -ENOMEM from percpu_counter_init()
Ah. Can you try the followi
On 30.03.2021 09:16 Wang Yugui wrote:
H,
On 30.03.21 г. 9:24, Wang Yugui wrote:
Hi, Nikolay Borisov
With a lot of dump_stack()/printk inserted around ENOMEM in btrfs code,
we find out the call stack for ENOMEM.
see the file -btrfs-dump_stack-when-ENOMEM.patch
#cat /usr/hpc-bio/xfstests/
H,
> On 30.03.21 г. 9:24, Wang Yugui wrote:
> > Hi, Nikolay Borisov
> >
> > With a lot of dump_stack()/printk inserted around ENOMEM in btrfs code,
> > we find out the call stack for ENOMEM.
> > see the file -btrfs-dump_stack-when-ENOMEM.patch
> >
> >
> > #cat /usr/hpc-bio/xfstests/results/
On 30.03.21 г. 9:24, Wang Yugui wrote:
> Hi, Nikolay Borisov
>
> With a lot of dump_stack()/printk inserted around ENOMEM in btrfs code,
> we find out the call stack for ENOMEM.
> see the file -btrfs-dump_stack-when-ENOMEM.patch
>
>
> #cat /usr/hpc-bio/xfstests/results//generic/476.dmesg
Hi,
> kmem_cache_zalloc() without __GFP_NOFAIL may fail.
>
> btrfs use kmem_cache_zalloc() with GFP_NOFS mostly,
> and only few place with __GFP_NOFAIL.
>
> xfs use kmem_cache_zalloc() with __GFP_NOFAIL mostly.
>
> It is very difficult to test all case of failure in kmem_cache_zalloc() .
>
>
Hi,
kmem_cache_zalloc() without __GFP_NOFAIL may fail.
btrfs use kmem_cache_zalloc() with GFP_NOFS mostly,
and only few place with __GFP_NOFAIL.
xfs use kmem_cache_zalloc() with __GFP_NOFAIL mostly.
It is very difficult to test all case of failure in kmem_cache_zalloc() .
Should btrfs use kme
Hi,
these callstack have a same root failure.
struct btrfs_path *btrfs_alloc_path(void)
{
return kmem_cache_zalloc(btrfs_path_cachep, GFP_NOFS);
}
fs/btrfs/transaction.c:1679 create_pending_snapshot+0xc1a/0xda0 [btrfs]
(fail)new_root = btrfs_get_new_fs_root(fs_info, objectid, anon_dev
Hi,
SSD/SAS is easy than SSD/NVMe to reproduce this problem.
Yet not able to reproduce this problem on another server.
CPU: Xeon(R) CPU E5-2680 v2(10 core) *2
memory: 192G, no swap
disk: SSD/NVMe with same partition size as SSD/SAS.
And this problem happened in kernel 5.10.26 + btrfs backpor