Re: [PATCH][CIFS]Call close synchronously during unlink/rename/lease break.

[Steve French <smfrench@gmail.com>]

I added

Cc: stable@vger.kernel.org # 5.13

Sound correct?

On Tue, Aug 10, 2021 at 2:48 PM Rohith Surabattula
<rohiths.msft@gmail.com> wrote:
>
> Hi All,
>
> I have attached an old version of patch by mistake.
> Have attached the correct one.
>
> Sorry for the confusion.
>
> Regards.
> Rohith
>
> On Wed, Aug 11, 2021 at 1:02 AM Steve French <smfrench@gmail.com> wrote:
> >
> > This is in http://smb3-test-rhel-75.southcentralus.cloudapp.azure.com/#/builders/2/builds/766
> >
> > On Tue, Aug 10, 2021 at 2:28 PM Steve French <smfrench@gmail.com> wrote:
> > >
> > > I saw some test failures in
> > >
> > > and looking in /var/log/messages I see e.g.
> > >
> > > Aug 10 14:00:22 fedora29 systemd[1]: Started /usr/bin/bash -c exit 77.
> > > Aug 10 14:00:25 fedora29 kernel: kmemleak: 10 new suspected memory
> > > leaks (see /sys/kernel/debug/kmemleak)
> > > Aug 10 14:00:28 fedora29 kernel: kmemleak: 1 new suspected memory
> > > leaks (see /sys/kernel/debug/kmemleak)
> > > Aug 10 14:00:28 fedora29 kernel: CIFS: Attempting to mount
> > > \\win16.vm.test\Scratch
> > > Aug 10 14:00:28 fedora29 kernel: CIFS: Attempting to mount
> > > \\win16.vm.test\Scratch
> > > Aug 10 14:00:29 fedora29 root[6460]: run xfstest cifs/100
> > > Aug 10 14:00:29 fedora29 journal: run fstests cifs/100 at 2021-08-10 14:00:29
> > > Aug 10 14:00:29 fedora29 systemd[1]: Started /usr/bin/bash -c test -w
> > > /proc/self/oom_score_adj && echo 250 > /proc/self/oom_score_adj; exec
> > > ./tests/cifs/100.
> > > Aug 10 14:00:30 fedora29 kernel: general protection fault, probably
> > > for non-canonical address 0xdead000000000110: 0000 [#4] SMP PTI
> > > Aug 10 14:00:30 fedora29 kernel: CPU: 0 PID: 6624 Comm: rm Tainted: G
> > >     D           5.14.0-rc4 #1
> > > Aug 10 14:00:30 fedora29 kernel: Hardware name: Red Hat KVM, BIOS
> > > 0.5.1 01/01/2011
> > > Aug 10 14:00:30 fedora29 kernel: RIP:
> > > 0010:cifs_close_deferred_file+0x77/0x140 [cifs]
> > > Aug 10 14:00:30 fedora29 kernel: Code: 85 de 00 00 00 4c 89 ef e8 06
> > > 96 2e e8 48 8b 1c 24 4c 39 e3 74 55 49 bd 00 01 00 00 00 00 ad de 48
> > > bd 22 01 00 00 00 00 ad de <48> 8b 7b 10 31 d2 be 01 00 00 00 e8 a9 c5
> > > fe ff 48 89 df e8 61 46
> > > Aug 10 14:00:30 fedora29 kernel: RSP: 0018:ffffb121806dbdd0 EFLAGS: 00010287
> > > Aug 10 14:00:30 fedora29 kernel: RAX: ffff88dc55b8ca01 RBX:
> > > dead000000000100 RCX: 00000000000034d9
> > > Aug 10 14:00:30 fedora29 kernel: RDX: 00000000000034d8 RSI:
> > > 67739b6f43694caa RDI: 0000000000032080
> > >
> > > and
> > >
> > > Aug 10 14:00:30 fedora29 kernel: ---[ end trace b062b12a095f4dd5 ]---
> > > Aug 10 14:00:30 fedora29 kernel: RIP:
> > > 0010:cifs_close_deferred_file+0x77/0x140 [cifs]
> > > Aug 10 14:00:30 fedora29 kernel: Code: 85 de 00 00 00 4c 89 ef e8 06
> > > 96 2e e8 48 8b 1c 24 4c 39 e3 74 55 49 bd 00 01 00 00 00 00 ad de 48
> > > bd 22 01 00 00 00 00 ad de <48> 8b 7b 10 31 d2 be 01 00 00 00 e8 a9 c5
> > > fe ff 48 89 df e8 61 46
> > > Aug 10 14:00:30 fedora29 kernel: RSP: 0018:ffffb12180f0fc80 EFLAGS: 00010283
> > > Aug 10 14:00:30 fedora29 kernel: RAX: ffff88dc5b2ca201 RBX:
> > > dead000000000100 RCX: 0000000000002367
> > > Aug 10 14:00:30 fedora29 kernel: RDX: 0000000000002366 RSI:
> > > 67100f614dfd246a RDI: 0000000000032080
> > > Aug 10 14:00:30 fedora29 kernel: RBP: dead000000000122 R08:
> > > 0000000000000001 R09: 0000000000000001
> > > Aug 10 14:00:30 fedora29 kernel: R10: ffffb12180f0fba0 R11:
> > > 0000000000000001 R12: ffffb12180f0fc80
> > >
> > > On Tue, Aug 10, 2021 at 12:29 PM Rohith Surabattula
> > > <rohiths.msft@gmail.com> wrote:
> > > >
> > > > Hi All,
> > > >
> > > > Have updated the patch to fix the memleak.
> > > >
> > > > Regards,
> > > > Rohith
> > > >
> > > > On Tue, Aug 10, 2021 at 5:18 PM Shyam Prasad N <nspmangalore@gmail.com> wrote:
> > > > >
> > > > > Already gave my review comments along with this one to the other email.
> > > > >
> > > > > Reviewed-by: Shyam Prasad N <sprasad@microsoft.com>
> > > > >
> > > > > On Mon, Aug 9, 2021 at 3:28 PM Rohith Surabattula
> > > > > <rohiths.msft@gmail.com> wrote:
> > > > > >
> > > > > > Hi Steve/All,
> > > > > >
> > > > > > During unlink/rename/lease break, deferred work for close is
> > > > > > scheduled immediately but in asynchronous manner which might
> > > > > > lead to race with actual(unlink/rename) commands.
> > > > > >
> > > > > > This change will schedule close synchronously which will avoid
> > > > > > the race conditions with other commands.
> > > > > >
> > > > > > Regards,
> > > > > > Rohith
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Regards,
> > > > > Shyam
> > >
> > >
> > >
> > > --
> > > Thanks,
> > >
> > > Steve
> >
> >
> >
> > --
> > Thanks,
> >
> > Steve



-- 
Thanks,

Steve