Re: [PATCH 2/2 v2] bluetooth: hci_core: fix NULL-pointer dereference at unregister
From: David Herrmann <hidden>
Date: 2012-03-09 14:04:16
Also in:
lkml, netdev, stable
Hi Johan On Fri, Mar 9, 2012 at 1:53 PM, Johan Hovold [off-list ref] wrote:
Make sure hci_dev_open returns immediately if hci_dev_unregister has been called. This fixes a race between hci_dev_open and hci_dev_unregister which can lead to a NULL-pointer dereference. Bug is 100% reproducible using hciattach and a disconnected serial port: 0. # hciattach -n /dev/ttyO1 any noflow 1. hci_dev_open called from hci_power_on grabs req lock 2. hci_init_req executes but device fails to initialise (times out =A0 eventually) 3. hci_dev_open is called from hci_sock_ioctl and sleeps on req lock 4. hci_uart_tty_close calls hci_dev_unregister and sleeps on req lock in =A0 hci_dev_do_close 5. hci_dev_open (1) releases req lock 6. hci_dev_do_close grabs req lock and returns as device is not up 7. hci_dev_unregister sleeps in destroy_workqueue 8. hci_dev_open (3) grabs req lock, calls hci_init_req and eventually sle=
eps
9. hci_dev_unregister finishes, while hci_dev_open is still running... [ =A0 79.627136] INFO: trying to register non-static key. [ =A0 79.632354] the code is fine but needs lockdep annotation. [ =A0 79.638122] turning off the locking correctness validator. [ =A0 79.643920] [<c00188bc>] (unwind_backtrace+0x0/0xf8) from [<c00729c4= ] (__lock_acquire+0x1590/0x1ab0) [ =A0 79.653594] [<c00729c4>] (__lock_acquire+0x1590/0x1ab0) from [<c0073=
3f8>] (lock_acquire+0x9c/0x128)
[ =A0 79.663085] [<c00733f8>] (lock_acquire+0x9c/0x128) from [<c0040a88>]=
(run_timer_softirq+0x150/0x3ac)
[ =A0 79.672668] [<c0040a88>] (run_timer_softirq+0x150/0x3ac) from [<c003=
a3b8>] (__do_softirq+0xd4/0x22c)
[ =A0 79.682281] [<c003a3b8>] (__do_softirq+0xd4/0x22c) from [<c003a924>]=
(irq_exit+0x8c/0x94)
[ =A0 79.690856] [<c003a924>] (irq_exit+0x8c/0x94) from [<c0013a50>] (han=
dle_IRQ+0x34/0x84)
[ =A0 79.699157] [<c0013a50>] (handle_IRQ+0x34/0x84) from [<c0008530>] (o=
map3_intc_handle_irq+0x48/0x4c)
[ =A0 79.708648] [<c0008530>] (omap3_intc_handle_irq+0x48/0x4c) from [<c0=
37499c>] (__irq_usr+0x3c/0x60)
[ =A0 79.718048] Exception stack(0xcf281fb0 to 0xcf281ff8) [ =A0 79.723358] 1fa0: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 =A0 =A0 =A0 =A0 =A0 0001e6a0 be8dab00 0001e698 00036698
[ =A0 79.731933] 1fc0: 0002df98 0002df38 0000001f 00000000 b6f234d0 00000=
000 00000004 00000000
[ =A0 79.740509] 1fe0: 0001e6f8 be8d6aa0 be8dac50 0000aab8 80000010 fffff=
fff
[ =A0 79.747497] Unable to handle kernel NULL pointer dereference at virt=
ual address 00000000
[ =A0 79.756011] pgd =3D cf3b4000 [ =A0 79.758850] [00000000] *pgd=3D8f0c7831, *pte=3D00000000, *ppte=3D000=
00000
[ =A0 79.765502] Internal error: Oops: 80000007 [#1] [ =A0 79.770294] Modules linked in: [ =A0 79.773529] CPU: 0 =A0 =A0Tainted: G =A0 =A0 =A0 =A0W =A0 =A0 (3.3.0=
-rc6-00002-gb5d5c87 #421)
[ =A0 79.781066] PC is at 0x0 [ =A0 79.783721] LR is at run_timer_softirq+0x16c/0x3ac [ =A0 79.788787] pc : [<00000000>] =A0 =A0lr : [<c0040aa4>] =A0 =A0psr: 6=
0000113
[ =A0 79.788787] sp : cf281ee0 =A0ip : 00000000 =A0fp : cf280000 [ =A0 79.800903] r10: 00000004 =A0r9 : 00000100 =A0r8 : b6f234d0 [ =A0 79.806427] r7 : c0519c28 =A0r6 : cf093488 =A0r5 : c0561a00 =A0r4 : =
00000000
[ =A0 79.813323] r3 : 00000000 =A0r2 : c054eee0 =A0r1 : 00000001 =A0r0 : =
00000000
[ =A0 79.820190] Flags: nZCv =A0IRQs on =A0FIQs on =A0Mode SVC_32 =A0ISA =
ARM =A0Segment user
[ =A0 79.827728] Control: 10c5387d =A0Table: 8f3b4019 =A0DAC: 00000015 [ =A0 79.833801] Process gpsd (pid: 1265, stack limit =3D 0xcf2802e8) [ =A0 79.839965] Stack: (0xcf281ee0 to 0xcf282000) [ =A0 79.844573] 1ee0: 00000002 00000000 c0040a24 00000000 00000002 cf281=
f08 00200200 00000000
[ =A0 79.853210] 1f00: 00000000 cf281f18 cf281f08 00000000 00000000 00000=
000 cf281f18 cf281f18
[ =A0 79.861816] 1f20: 00000000 00000001 c056184c 00000000 00000001 b6f23=
4d0 c0561848 00000004
[ =A0 79.870452] 1f40: cf280000 c003a3b8 c051e79c 00000001 00000000 00000=
100 3fa9e7b8 0000000a
[ =A0 79.879089] 1f60: 00000025 cf280000 00000025 00000000 00000000 b6f23=
4d0 00000000 00000004
[ =A0 79.887756] 1f80: 00000000 c003a924 c053ad38 c0013a50 fa200000 cf281=
fb0 ffffffff c0008530
[ =A0 79.896362] 1fa0: 0001e6a0 0000aab8 80000010 c037499c 0001e6a0 be8da=
b00 0001e698 00036698
[ =A0 79.904998] 1fc0: 0002df98 0002df38 0000001f 00000000 b6f234d0 00000=
000 00000004 00000000
[ =A0 79.913665] 1fe0: 0001e6f8 be8d6aa0 be8dac50 0000aab8 80000010 fffff=
fff 00fbf700 04ffff00
[ =A0 79.922302] [<c0040aa4>] (run_timer_softirq+0x16c/0x3ac) from [<c003=
a3b8>] (__do_softirq+0xd4/0x22c)
[ =A0 79.931945] [<c003a3b8>] (__do_softirq+0xd4/0x22c) from [<c003a924>]=
(irq_exit+0x8c/0x94)
[ =A0 79.940582] [<c003a924>] (irq_exit+0x8c/0x94) from [<c0013a50>] (han=
dle_IRQ+0x34/0x84)
[ =A0 79.948913] [<c0013a50>] (handle_IRQ+0x34/0x84) from [<c0008530>] (o=
map3_intc_handle_irq+0x48/0x4c)
[ =A0 79.958404] [<c0008530>] (omap3_intc_handle_irq+0x48/0x4c) from [<c0=
37499c>] (__irq_usr+0x3c/0x60)
[ =A0 79.967773] Exception stack(0xcf281fb0 to 0xcf281ff8) [ =A0 79.973083] 1fa0: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 =A0 =A0 =A0 =A0 =A0 0001e6a0 be8dab00 0001e698 00036698
[ =A0 79.981658] 1fc0: 0002df98 0002df38 0000001f 00000000 b6f234d0 00000=
000 00000004 00000000
[ =A0 79.990234] 1fe0: 0001e6f8 be8d6aa0 be8dac50 0000aab8 80000010 fffff=
fff
quoted hunk ↗ jump to hunk
[ =A0 79.997161] Code: bad PC value [ =A0 80.000396] ---[ end trace 6f6739840475f9ee ]--- [ =A0 80.005279] Kernel panic - not syncing: Fatal exception in interrupt Cc: stable <redacted> Signed-off-by: Johan Hovold <redacted> --- v2: use hdev->dev_flags for internal unregister flag =A0include/net/bluetooth/hci.h | =A0 =A02 ++ =A0net/bluetooth/hci_core.c =A0 =A0| =A0 =A07 +++++++ =A02 files changed, 9 insertions(+), 0 deletions(-)diff --git a/include/net/bluetooth/hci.h b/include/net/bluetooth/hci.h index 00596e8..e8879b9 100644 --- a/include/net/bluetooth/hci.h +++ b/include/net/bluetooth/hci.h@@ -93,6 +93,8 @@ enum {=A0* states from the controller. =A0*/ =A0enum { + =A0 =A0 =A0 HCI_UNREGISTER, + =A0 =A0 =A0 =A0HCI_LE_SCAN, =A0};diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c index d6448f0..22b6781 100644 --- a/net/bluetooth/hci_core.c +++ b/net/bluetooth/hci_core.c@@ -525,6 +525,11 @@ int hci_dev_open(__u16 dev)=A0 =A0 =A0 =A0hci_req_lock(hdev); + =A0 =A0 =A0 if (test_bit(HCI_UNREGISTER, &hdev->dev_flags)) { + =A0 =A0 =A0 =A0 =A0 =A0 =A0 ret =3D -ENODEV; + =A0 =A0 =A0 =A0 =A0 =A0 =A0 goto done; + =A0 =A0 =A0 } +
Isn't it enough to check for HCI_RUNNING here? We obviously have a race here as we take the device with hci_dev_get(), then sleep and then we do not check whether the device is still alive. However, drivers are required to reset HCI_RUNNING before calling hci_unregister_dev() (which is bogus anyway, but its the way we handled it in the past) therefore it should be enough for us to check for HCI_RUNNING. Regards David
quoted hunk ↗ jump to hunk
=A0 =A0 =A0 =A0if (hdev->rfkill && rfkill_blocked(hdev->rfkill)) { =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0ret =3D -ERFKILL; =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto done;@@ -1577,6 +1582,8 @@ void hci_unregister_dev(struct hci_dev *hdev)=A0 =A0 =A0 =A0BT_DBG("%p name %s bus %d", hdev, hdev->name, hdev->bus); + =A0 =A0 =A0 set_bit(HCI_UNREGISTER, &hdev->dev_flags); + =A0 =A0 =A0 =A0write_lock(&hci_dev_list_lock); =A0 =A0 =A0 =A0list_del(&hdev->list); =A0 =A0 =A0 =A0write_unlock(&hci_dev_list_lock); -- 1.7.8.4 -- To unsubscribe from this list: send the line "unsubscribe linux-bluetooth=
" in
the body of a message to majordomo@vger.kernel.org More majordomo info at =A0http://vger.kernel.org/majordomo-info.html