cifs: ensure that srv_mutex is held when dealing with ssocket pointer
authorJeff Layton <jlayton@redhat.com>
Thu, 5 Sep 2013 12:38:10 +0000 (08:38 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 27 Sep 2013 00:18:04 +0000 (17:18 -0700)
commit9b8ace6745bf4722ff9c48089f5b0723478a849b
treec8048afc91172de31a61adf6b18c061f529e8a58
parentd5cc290b92837eb9844579e297e969cc7a804da0
cifs: ensure that srv_mutex is held when dealing with ssocket pointer

commit 73e216a8a42c0ef3d08071705c946c38fdbe12b0 upstream.

Oleksii reported that he had seen an oops similar to this:

BUG: unable to handle kernel NULL pointer dereference at 0000000000000088
IP: [<ffffffff814dcc13>] sock_sendmsg+0x93/0xd0
PGD 0
Oops: 0000 [#1] PREEMPT SMP
Modules linked in: ipt_MASQUERADE xt_REDIRECT xt_tcpudp iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack ip_tables x_tables carl9170 ath usb_storage f2fs nfnetlink_log nfnetlink md4 cifs dns_resolver hid_generic usbhid hid af_packet uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_core videodev rfcomm btusb bnep bluetooth qmi_wwan qcserial cdc_wdm usb_wwan usbnet usbserial mii snd_hda_codec_hdmi snd_hda_codec_realtek iwldvm mac80211 coretemp intel_powerclamp kvm_intel kvm iwlwifi snd_hda_intel cfg80211 snd_hda_codec xhci_hcd e1000e ehci_pci snd_hwdep sdhci_pci snd_pcm ehci_hcd microcode psmouse sdhci thinkpad_acpi mmc_core i2c_i801 pcspkr usbcore hwmon snd_timer snd_page_alloc snd ptp rfkill pps_core soundcore evdev usb_common vboxnetflt(O) vboxdrv(O)Oops#2 Part8
 loop tun binfmt_misc fuse msr acpi_call(O) ipv6 autofs4
CPU: 0 PID: 21612 Comm: kworker/0:1 Tainted: G        W  O 3.10.1SIGN #28
Hardware name: LENOVO 2306CTO/2306CTO, BIOS G2ET92WW (2.52 ) 02/22/2013
Workqueue: cifsiod cifs_echo_request [cifs]
task: ffff8801e1f416f0 ti: ffff880148744000 task.ti: ffff880148744000
RIP: 0010:[<ffffffff814dcc13>]  [<ffffffff814dcc13>] sock_sendmsg+0x93/0xd0
RSP: 0000:ffff880148745b00  EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff880148745b78 RCX: 0000000000000048
RDX: ffff880148745c90 RSI: ffff880181864a00 RDI: ffff880148745b78
RBP: ffff880148745c48 R08: 0000000000000048 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: ffff880181864a00
R13: ffff880148745c90 R14: 0000000000000048 R15: 0000000000000048
FS:  0000000000000000(0000) GS:ffff88021e200000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000088 CR3: 000000020c42c000 CR4: 00000000001407b0
Oops#2 Part7
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Stack:
 ffff880148745b30 ffffffff810c4af9 0000004848745b30 ffff880181864a00
 ffffffff81ffbc40 0000000000000000 ffff880148745c90 ffffffff810a5aab
 ffff880148745bc0 ffffffff81ffbc40 ffff880148745b60 ffffffff815a9fb8
Call Trace:
 [<ffffffff810c4af9>] ? finish_task_switch+0x49/0xe0
 [<ffffffff810a5aab>] ? lock_timer_base.isra.36+0x2b/0x50
 [<ffffffff815a9fb8>] ? _raw_spin_unlock_irqrestore+0x18/0x40
 [<ffffffff810a673f>] ? try_to_del_timer_sync+0x4f/0x70
 [<ffffffff815aa38f>] ? _raw_spin_unlock_bh+0x1f/0x30
 [<ffffffff814dcc87>] kernel_sendmsg+0x37/0x50
 [<ffffffffa081a0e0>] smb_send_kvec+0xd0/0x1d0 [cifs]
 [<ffffffffa081a263>] smb_send_rqst+0x83/0x1f0 [cifs]
 [<ffffffffa081ab6c>] cifs_call_async+0xec/0x1b0 [cifs]
 [<ffffffffa08245e0>] ? free_rsp_buf+0x40/0x40 [cifs]
Oops#2 Part6
 [<ffffffffa082606e>] SMB2_echo+0x8e/0xb0 [cifs]
 [<ffffffffa0808789>] cifs_echo_request+0x79/0xa0 [cifs]
 [<ffffffff810b45b3>] process_one_work+0x173/0x4a0
 [<ffffffff810b52a1>] worker_thread+0x121/0x3a0
 [<ffffffff810b5180>] ? manage_workers.isra.27+0x2b0/0x2b0
 [<ffffffff810bae00>] kthread+0xc0/0xd0
 [<ffffffff810bad40>] ? kthread_create_on_node+0x120/0x120
 [<ffffffff815b199c>] ret_from_fork+0x7c/0xb0
 [<ffffffff810bad40>] ? kthread_create_on_node+0x120/0x120
Code: 84 24 b8 00 00 00 4c 89 f1 4c 89 ea 4c 89 e6 48 89 df 4c 89 60 18 48 c7 40 28 00 00 00 00 4c 89 68 30 44 89 70 14 49 8b 44 24 28 <ff> 90 88 00 00 00 3d ef fd ff ff 74 10 48 8d 65 e0 5b 41 5c 41
 RIP  [<ffffffff814dcc13>] sock_sendmsg+0x93/0xd0
 RSP <ffff880148745b00>
CR2: 0000000000000088

The client was in the middle of trying to send a frame when the
server->ssocket pointer got zeroed out. In most places, that we access
that pointer, the srv_mutex is held. There's only one spot that I see
that the server->ssocket pointer gets set and the srv_mutex isn't held.
This patch corrects that.

The upstream bug report was here:

    https://bugzilla.kernel.org/show_bug.cgi?id=60557

Reported-by: Oleksii Shevchuk <alxchk@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/cifs/connect.c