Message-ID: <20090331084546.6b24d66e@tleilax.poochiereds.net>
Date:	Tue, 31 Mar 2009 08:45:46 -0400
From:	Jeff Layton <jlayton@...hat.com>
To:	dhowells@...hat.com
Cc:	linux-kernel@...r.kernel.org, linux-cifs-client@...ts.samba.org
Subject: consistent oops from request_key in 2.6.29

I've started seeing a consistent oops in recent rawhide kernels when I
try to do CIFS mounts with krb5 auth:

BUG: unable to handle kernel NULL pointer dereference at 0000000000000030
IP: [<ffffffff8102bf4c>] __ticket_spin_trylock+0x4/0x21
PGD 0 
Oops: 0000 [#1] SMP 
last sysfs file: /sys/devices/pci0000:00/0000:00:04.0/local_cpus
CPU 1 
Modules linked in: nls_utf8 cifs sunrpc ipv6 dm_multipath uinput 8139too 8139cp i2c_piix4 mii i2c_core pcspkr ata_generic pata_acpi [last unloaded: freq_table]
Pid: 2255, comm: mount.cifs Tainted: G        W  2.6.29-16.fc11.x86_64 #1 HVM domU
RIP: 0010:[<ffffffff8102bf4c>]  [<ffffffff8102bf4c>] __ticket_spin_trylock+0x4/0x21
RSP: 0018:ffff8800370ab9d0  EFLAGS: 00010096
RAX: ffff88003d0d23c0 RBX: 0000000000000030 RCX: 0000000000000000
RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000030
RBP: ffff8800370ab9d0 R08: 0000000000000002 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000030
R13: 0000000000000292 R14: ffff880037088000 R15: ffffffff81597ad0
FS:  00007fa0e7fe56f0(0000) GS:ffff88003ea4b320(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000030 CR3: 00000000370c7000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process mount.cifs (pid: 2255, threadinfo ffff8800370aa000, task ffff88003d0d23c0)
Stack:
 ffff8800370ab9e8 ffffffff811a2af9 0000000000000048 ffff8800370aba28
 ffffffff81399014 ffffffff8119ce86 ffffffff8116470a 0000000000000000
 0000000000000030 0000000000000028 ffffffffa01292a0 ffff8800370aba48
Call Trace:
 [<ffffffff811a2af9>] _raw_spin_trylock+0xd/0x2d
 [<ffffffff81399014>] _spin_lock_irqsave+0x59/0x8b
 [<ffffffff8119ce86>] ? __down_write_trylock+0x16/0x4f
 [<ffffffff8116470a>] ? request_key_and_link+0x220/0x3d5
 [<ffffffff8119ce86>] __down_write_trylock+0x16/0x4f
 [<ffffffff81397a9c>] down_write+0x54/0x7f
 [<ffffffff8116470a>] ? request_key_and_link+0x220/0x3d5
 [<ffffffff8116470a>] request_key_and_link+0x220/0x3d5
 [<ffffffff8119ec8d>] ? vsnprintf+0x2e7/0x4ed
 [<ffffffff81164992>] request_key+0x41/0x72
 [<ffffffffa010f8e5>] cifs_get_spnego_key+0x189/0x1b4 [cifs]
 [<ffffffffa010e4a0>] CIFS_SessSetup+0x442/0xb1e [cifs]
 [<ffffffffa00f91c9>] cifs_setup_session+0x119/0xb61 [cifs]
 [<ffffffff81398cc6>] ? _spin_unlock_irqrestore+0x48/0x58
 [<ffffffff81070acd>] ? trace_hardirqs_on+0xd/0xf
 [<ffffffffa00fd5b6>] cifs_mount+0x17a7/0x1f2b [cifs]
 [<ffffffff810dd173>] ? __kmalloc+0x10b/0x149
 [<ffffffffa00ee8ab>] cifs_get_sb+0x110/0x26a [cifs]
 [<ffffffff810e71a9>] vfs_kern_mount+0xa3/0x13c
 [<ffffffff810e72aa>] do_kern_mount+0x4d/0xe8
 [<ffffffff810fbe0b>] do_mount+0x744/0x790
 [<ffffffff811a2ed2>] ? _raw_spin_lock+0x65/0x111
 [<ffffffff810fbee9>] sys_mount+0x92/0xd5
 [<ffffffff8101133a>] system_call_fastpath+0x16/0x1b
Code: 1b fb ff ff c9 c3 90 55 b8 00 00 01 00 48 89 e5 f0 0f c1 07 0f b7 d0 c1 e8 10 39 c2 74 07 f3 90 0f b7 17 eb f5 c9 c3 55 48 89 e5 <8b> 07 89 c2 c1 c0 10 39 c2 8d 90 00 00 01 00 75 04 f0 0f b1 17 
RIP  [<ffffffff8102bf4c>] __ticket_spin_trylock+0x4/0x21
 RSP <ffff8800370ab9d0>
CR2: 0000000000000030
---[ end trace cee083cffbca031b ]---


...we call request_key from CIFS which calls request_key_and_link
with a NULL dest_keyring. Eventually this calls construct_alloc_key
with the NULL dest_keyring. The following patch seems to have changed
it so that it's no longer safe to call construct_alloc_key this way:

commit 8bbf4976b59fc9fc2861e79cab7beb3f6d647640
Author: David Howells <dhowells@...hat.com>
Date:   Fri Nov 14 10:39:14 2008 +1100

    KEYS: Alter use of key instantiation link-to-keyring argument


...but request_key was never fixed so that it doesn't do this. It's
unclear to me what the correct fix is. Do we need to put the checks for
a NULL dest_keyring back into construct_alloc_key or fix request_key
to call request_key_and_link with a dest_keyring?

Thanks,
-- 
Jeff Layton <jlayton@...hat.com>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists