mbox series

[stable,4.14,0/1] Fix ext4 xfstests failure

Message ID 20230131043815.14989-1-risbhat@amazon.com (mailing list archive)
Headers show
Series Fix ext4 xfstests failure | expand

Message

Rishabh Bhatnagar Jan. 31, 2023, 4:38 a.m. UTC
While running xfstests on 4.14.304 version we see a warning being
generated in one of ext4 tests with the following stack trace.

WARNING: CPU: 4 PID: 15332 at mm/util.c:414
kvmalloc_node+0x67/0x70
ext4_expand_extra_isize_ea+0x2b4/0x870 [ext4]
__ext4_expand_extra_isize+0xcb/0x120 [ext4]
ext4_mark_inode_dirty+0x1a5/0x1d0 [ext4]
ext4_ext_truncate+0x1f/0x90 [ext4]
ext4_truncate+0x363/0x400 [ext4]
ext4_setattr+0x392/0xa00 [ext4]
notify_change+0x300/0x420
? ext4_xattr_security_set+0x20/0x20 [ext4]
do_truncate+0x75/0xc0
? ext4_release_file+0xa0/0xa0 [ext4]
path_openat+0x737/0x16f0
do_filp_open+0x9b/0x110
? __check_object_size+0xb4/0x190
? do_sys_open+0x1bd/0x250
do_sys_open+0x1bd/0x250
do_syscall_64+0x67/0x110
entry_SYSCALL_64_after_hwframe+0x59/0xbe

It seems rebase to 4.14.304 brings a bunch of ext4 changes.
Commit ext4: allocate extended attribute value in vmalloc area
(73c44f61dab180b5f2dee9f15397aba36a75a882) tries to allocate buffer
using kvmalloc with improper flags that generates this warning.
To fix backport an upstream commit mm: kvmalloc does not
fallback to vmalloc for incompatible gfp flags
(170f26afa0481c72af93aa61b7398b5663451651). This removes the WARN_ON and
fallsback to kmalloc if correct flags are not passed.


Michal Hocko (1):
  mm: kvmalloc does not fallback to vmalloc for incompatible gfp flags

 mm/util.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

Comments

Greg Kroah-Hartman Feb. 3, 2023, 8 a.m. UTC | #1
On Tue, Jan 31, 2023 at 04:38:14AM +0000, Rishabh Bhatnagar wrote:
> While running xfstests on 4.14.304 version we see a warning being
> generated in one of ext4 tests with the following stack trace.
> 
> WARNING: CPU: 4 PID: 15332 at mm/util.c:414
> kvmalloc_node+0x67/0x70
> ext4_expand_extra_isize_ea+0x2b4/0x870 [ext4]
> __ext4_expand_extra_isize+0xcb/0x120 [ext4]
> ext4_mark_inode_dirty+0x1a5/0x1d0 [ext4]
> ext4_ext_truncate+0x1f/0x90 [ext4]
> ext4_truncate+0x363/0x400 [ext4]
> ext4_setattr+0x392/0xa00 [ext4]
> notify_change+0x300/0x420
> ? ext4_xattr_security_set+0x20/0x20 [ext4]
> do_truncate+0x75/0xc0
> ? ext4_release_file+0xa0/0xa0 [ext4]
> path_openat+0x737/0x16f0
> do_filp_open+0x9b/0x110
> ? __check_object_size+0xb4/0x190
> ? do_sys_open+0x1bd/0x250
> do_sys_open+0x1bd/0x250
> do_syscall_64+0x67/0x110
> entry_SYSCALL_64_after_hwframe+0x59/0xbe
> 
> It seems rebase to 4.14.304 brings a bunch of ext4 changes.
> Commit ext4: allocate extended attribute value in vmalloc area
> (73c44f61dab180b5f2dee9f15397aba36a75a882) tries to allocate buffer
> using kvmalloc with improper flags that generates this warning.
> To fix backport an upstream commit mm: kvmalloc does not
> fallback to vmalloc for incompatible gfp flags
> (170f26afa0481c72af93aa61b7398b5663451651). This removes the WARN_ON and
> fallsback to kmalloc if correct flags are not passed.
> 
> 
> Michal Hocko (1):
>   mm: kvmalloc does not fallback to vmalloc for incompatible gfp flags
> 
>  mm/util.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> -- 
> 2.38.1
> 

Now queued up, thanks.

greg k-h