On 5 Aug 2013, Trond Myklebust told this:
> Does the attached patch fix the problem?
> From 3c50ba80105464a28d456d9a1e0f1d81d4af92a8 Mon Sep 17 00:00:00 2001
> From: Trond Myklebust <trond.mykleb...@netapp.com>
> Date: Mon, 5 Aug 2013 12:06:12 -0400
> Subject: [PATCH] LOCKD: Don't call utsname()->nodename from
> nlmclnt_setlockargs
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
It makes it worse. Much, much worse. From a crash every so often when
I'm doing compilations over NFS, I get an immediate panic on startx,
long long before I even try to replicate the earlier panic:
[ 83.432358] task: ffff88041aaa5ac0 ti: ffff8804199e2000 task.ti:
ffff8804199e2000
[ 83.432428] RIP: 0010:[<ffffffff8124af69>] [<ffffffff8124af69>]
encode_nlm4_lock+0x26/0xbe
[ 83.432512] RSP: 0018:ffff8804199e3a78 EFLAGS: 00010286
[ 83.432564] RAX: 0000000000000000 RBX: ffff88041a577038 RCX: ffffffffffffffff
[ 83.432630] RDX: ffff8804193b3098 RSI: ffff88041a577038 RDI: 000000000000008c
[ 83.432697] RBP: ffff8804199e3aa8 R08: ffff8804193b3098 R09: 0000000000000001
[ 83.432763] R10: ffff88042fa12980 R11: ffff88042fa12980 R12: ffff8804199e3ae8
[ 83.432830] R13: 000000000000008c R14: ffff8804199e3fd8 R15: ffffffff815de80e
[ 83.432898] FS: 00007f594b40c740(0000) GS:ffff88042fa00000(0000)
knlGS:0000000000000000
[ 83.432974] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 83.433028] CR2: 000000000000008c CR3: 000000041ab3d000 CR4: 00000000001407f0
[ 83.433095] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 83.433176] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[ 83.433255] Stack:
[ 83.433276] ffff88041a44fb70 ffff880400000004 ffff8804199e3ae8
ffff88041a577010
[ 83.433360] ffff8804188e0e00 ffff8804199e3fd8 ffff8804199e3ac8
ffffffff8124b0d7
[ 83.433443] ffff8804188e0e00 ffffffff8124b086 ffff8804199e3b38
ffffffff815e6032
[ 83.433616] Call Trace:
[ 83.433646] [<ffffffff8124b0d7>] nlm4_xdr_enc_lockargs+0x51/0x76
[ 83.433707] [<ffffffff8124b086>] ? nlm4_xdr_enc_cancargs+0x56/0x56
[ 83.433769] [<ffffffff815e6032>] rpcauth_wrap_req+0x57/0x62
[ 83.433826] [<ffffffff815de98a>] call_transmit+0x17c/0x1f9
[ 83.433880] [<ffffffff815e4e58>] __rpc_execute+0xe8/0x2ca
[ 83.433935] [<ffffffff815e50f9>] rpc_execute+0x76/0x9d
[ 83.433986] [<ffffffff815debc1>] rpc_run_task+0x78/0x80
[ 83.434039] [<ffffffff815decff>] rpc_call_sync+0x88/0x9e
[ 83.434092] [<ffffffff81244b3c>] nlmclnt_call+0xb5/0x240
[ 83.434146] [<ffffffff812454f0>] nlmclnt_proc+0x226/0x5fb
[ 83.434226] [<ffffffff812209a2>] nfs3_proc_lock+0x21/0x23
[ 83.434280] [<ffffffff81214a5e>] do_setlk+0x65/0xee
[ 83.434329] [<ffffffff81214ca6>] nfs_lock+0x14e/0x162
[ 83.434382] [<ffffffff81199661>] vfs_lock_file+0x29/0x35
[ 83.434435] [<ffffffff8119a51d>] fcntl_setlk+0x139/0x2c5
[ 83.434490] [<ffffffff81169621>] SyS_fcntl+0x2b6/0x47d
[ 83.434543] [<ffffffff81613e92>] system_call_fastpath+0x16/0x1b
[ 83.434600] Code: 5b 41 5c 5d c3 0f 1f 44 00 00 55 31 c0 48 83 c9 ff 48 89
e5 41 56 41 55 41 54 49 89 fc 53 48 89 f3 48 83 ec 10 4c 8b 2e 4c 89 ef <f2> ae
4c 89 e7 48 f7 d1 4c 8d 71 ff 41 8d 76 04 e8 9f 16 3a 00
[ 83.435077] RIP [<ffffffff8124af69>] encode_nlm4_lock+0x26/0xbe
[ 83.435140] RSP <ffff8804199e3a78>
[ 83.435197] CR2: 000000000000008c
That's here:
(gdb) list *(encode_nlm4_lock+0x26)
0xffffffff8124af69 is in encode_nlm4_lock (fs/lockd/clnt4xdr.c:329).
324 * string caller_name<LM_MAXSTRLEN>;
325 */
326 static void encode_caller_name(struct xdr_stream *xdr, const char *name)
327 {
328 /* NB: client-side does not set lock->len */
329 u32 length = strlen(name);
330 __be32 *p;
331
332 p = xdr_reserve_space(xdr, 4 + length);
333 xdr_encode_opaque(p, name, length);
0xffffffff8124af69 <+38>: repnz scas %es:(%rdi),%al
Pretty clearly, "name" can be NULL after this patch...
--
NULL && (void)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
