On Tue, Nov 8, 2011 at 4:52 AM, Johannes Berg <johannes@xxxxxxxxxxxxxxxx> wrote:
> After random poking with hwsim:
>
> [ 640.357147] mac80211_hwsim: unregister radios
> [ 640.357151] mac80211_hwsim: closing netlink
> [ 640.871735] BUG: unable to handle kernel paging request at ffff88001a06b5ab
> [ 640.879198] IP: [<ffffffffa030df9a>] reg_device_uevent+0x1a/0x50 [cfg80211]
> [ 640.880507] PGD 1836063 PUD 183a063 PMD 1ffcb067 PTE 1a06b160
> [ 640.880507] Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
> [ 640.880507] CPU 0
> [ 640.880507] Modules linked in: cfg80211(-) [last unloaded: mac80211]
> [ 640.880507]
> [ 640.880507] Pid: 2279, comm: rmmod Tainted: G W 3.1.0-wl+ #663 Bochs Bochs
> [ 640.880507] RIP: 0010:[<ffffffffa030df9a>] [<ffffffffa030df9a>] reg_device_uevent+0x1a/0x50 [cfg80211]
> [ 640.880507] RSP: 0000:ffff88001c5f9d58 EFLAGS: 00010286
> [ 640.880507] RAX: 0000000000000000 RBX: ffff88001d2eda88 RCX: ffff88001c7468fc
> [ 640.880507] RDX: ffff88001a06b5a0 RSI: ffff88001c7467b0 RDI: ffff88001c7467b0
> [ 640.880507] RBP: ffff88001c5f9d58 R08: 000000000000ffff R09: 000000000000ffff
> [ 640.880507] R10: 0000000000000000 R11: 0000000000000001 R12: ffff88001c7467b0
> [ 640.880507] R13: ffff88001d2eda78 R14: ffffffff8164a840 R15: 0000000000000001
> [ 640.880507] FS: 00007f8a91d8a6e0(0000) GS:ffff88001fc00000(0000) knlGS:0000000000000000
> [ 640.880507] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
> [ 640.880507] CR2: ffff88001a06b5ab CR3: 000000001c62e000 CR4: 00000000000006f0
> [ 640.880507] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [ 640.880507] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
> [ 640.880507] Process rmmod (pid: 2279, threadinfo ffff88001c5f8000, task ffff88000023c780)
> [ 640.880507] Stack:
> [ 640.880507] ffff88001c5f9d98 ffffffff812ff7e5 ffffffff8176ab3d ffff88001c7468c2
> [ 640.880507] 000000000000ffff ffff88001d2eda88 ffff88001c7467b0 ffff880000114820
> [ 640.880507] ffff88001c5f9e38 ffffffff81241dc7 ffff88001c5f9db8 ffffffff81040189
> [ 640.880507] Call Trace:
> [ 640.880507] [<ffffffff812ff7e5>] dev_uevent+0xc5/0x170
> [ 640.880507] [<ffffffff81241dc7>] kobject_uevent_env+0x1f7/0x490
> [ 640.880507] [<ffffffff81040189>] ? sub_preempt_count+0x29/0x60
> [ 640.880507] [<ffffffff814cab1a>] ? _raw_spin_unlock_irqrestore+0x4a/0x90
> [ 640.880507] [<ffffffff81305307>] ? devres_release_all+0x27/0x60
> [ 640.880507] [<ffffffff8124206b>] kobject_uevent+0xb/0x10
> [ 640.880507] [<ffffffff812fee27>] device_del+0x157/0x1b0
> [ 640.880507] [<ffffffff8130377d>] platform_device_del+0x1d/0x90
> [ 640.880507] [<ffffffff81303b76>] platform_device_unregister+0x16/0x30
> [ 640.880507] [<ffffffffa030fffd>] regulatory_exit+0x5d/0x180 [cfg80211]
> [ 640.880507] [<ffffffffa032bec3>] cfg80211_exit+0x2b/0x45 [cfg80211]
> [ 640.880507] [<ffffffff8109a84c>] sys_delete_module+0x16c/0x220
> [ 640.880507] [<ffffffff8108a23e>] ? trace_hardirqs_on_caller+0x7e/0x120
> [ 640.880507] [<ffffffff814cba02>] system_call_fastpath+0x16/0x1b
> [ 640.880507] Code: 5b 5d c3 66 0f 1f 44 00 00 bf 10 00 00 00 eb d1 90 55 48 89 e5 66 66 66 66 90 48 8b 15 b8 3b 02 00 31 c0 48 85 d2 48 89 f7 74 06
> [ 640.880507] 7a 0b 00 74 08 5d c3 66 0f 1f 44 00 00 0f be 4a 09 48 c7 c6
> [ 640.880507] RIP [<ffffffffa030df9a>] reg_device_uevent+0x1a/0x50 [cfg80211]
> [ 640.880507] RSP <ffff88001c5f9d58>
> [ 640.880507] CR2: ffff88001a06b5ab
> [ 640.880507] ---[ end trace 147c5099a411e8c0 ]---
> [ 640.880507] BUG: sleeping function called from invalid context at /home/johannes/sys/wireless-testing/kernel/rwsem.c:21
> [ 640.880507] in_atomic(): 0, irqs_disabled(): 1, pid: 2279, name: rmmod
> [ 640.880507] INFO: lockdep is turned off.
> [ 640.880507] irq event stamp: 31652
> [ 640.880507] hardirqs last enabled at (31651): [<ffffffff814bd1b0>] __slab_alloc.isra.58.constprop.63+0x3b8/0x3db
> [ 640.880507] hardirqs last disabled at (31652): [<ffffffff814cb536>] error_sti+0x5/0x6
> [ 640.880507] softirqs last enabled at (31204): [<ffffffff81052ab7>] __do_softirq+0x137/0x3c0
> [ 640.880507] softirqs last disabled at (31185): [<ffffffff814cd07c>] call_softirq+0x1c/0x30
> [ 640.880507] Pid: 2279, comm: rmmod Tainted: G D W 3.1.0-wl+ #663
> [ 640.880507] Call Trace:
> [ 640.880507] [<ffffffff81089740>] ? print_irqtrace_events+0xd0/0xe0
> [ 640.880507] [<ffffffff8103b729>] __might_sleep.part.132+0xb9/0xd0
> [ 640.880507] [<ffffffff8103b791>] __might_sleep+0x51/0x70
> [ 640.880507] [<ffffffff814c9246>] down_read+0x26/0x9c
> [ 640.880507] [<ffffffff810a39cb>] acct_collect+0x4b/0x1b0
> [ 640.880507] [<ffffffff8104f1eb>] do_exit+0x30b/0x470
> [ 640.880507] [<ffffffff81005b19>] oops_end+0x89/0xc0
> [ 640.880507] [<ffffffff814b7864>] no_context+0x146/0x153
> [ 640.880507] [<ffffffff814b7a42>] __bad_area_nosemaphore+0x1d1/0x1f0
> [ 640.880507] [<ffffffff814caaa5>] ? _raw_spin_unlock+0x35/0x60
> [ 640.880507] [<ffffffff814b7a74>] bad_area_nosemaphore+0x13/0x15
> [ 640.880507] [<ffffffff8102a7ed>] do_page_fault+0x40d/0x4e0
> [ 640.880507] [<ffffffff814bd1b0>] ? __slab_alloc.isra.58.constprop.63+0x3b8/0x3db
> [ 640.880507] [<ffffffff8108a23e>] ? trace_hardirqs_on_caller+0x7e/0x120
> [ 640.880507] [<ffffffff81249cde>] ? string.isra.5+0x3e/0xd0
> [ 640.880507] [<ffffffff8124c91d>] ? trace_hardirqs_off_thunk+0x3a/0x3c
> [ 640.880507] [<ffffffff814cb2f5>] page_fault+0x25/0x30
> [ 640.880507] [<ffffffffa030df9a>] ? reg_device_uevent+0x1a/0x50 [cfg80211]
> [ 640.880507] [<ffffffff81303560>] ? platform_uevent+0x30/0x40
> [ 640.880507] [<ffffffff812ff7e5>] dev_uevent+0xc5/0x170
> [ 640.880507] [<ffffffff81241dc7>] kobject_uevent_env+0x1f7/0x490
> [ 640.880507] [<ffffffff81040189>] ? sub_preempt_count+0x29/0x60
> [ 640.880507] [<ffffffff814cab1a>] ? _raw_spin_unlock_irqrestore+0x4a/0x90
> [ 640.880507] [<ffffffff81305307>] ? devres_release_all+0x27/0x60
> [ 640.880507] [<ffffffff8124206b>] kobject_uevent+0xb/0x10
> [ 640.880507] [<ffffffff812fee27>] device_del+0x157/0x1b0
> [ 640.880507] [<ffffffff8130377d>] platform_device_del+0x1d/0x90
> [ 640.880507] [<ffffffff81303b76>] platform_device_unregister+0x16/0x30
> [ 640.880507] [<ffffffffa030fffd>] regulatory_exit+0x5d/0x180 [cfg80211]
> [ 640.880507] [<ffffffffa032bec3>] cfg80211_exit+0x2b/0x45 [cfg80211]
> [ 640.880507] [<ffffffff8109a84c>] sys_delete_module+0x16c/0x220
> [ 640.880507] [<ffffffff8108a23e>] ? trace_hardirqs_on_caller+0x7e/0x120
> [ 640.880507] [<ffffffff814cba02>] system_call_fastpath+0x16/0x1b
>
It seems we cannot tell udev we are committing suicide at __exit() for
whatever reason. Not sure why. I find this odd unless this is
generally true. Can you reproduce easily? Can you try this.
diff --git a/net/wireless/reg.c b/net/wireless/reg.c
index 2520a1b..9f1037e 100644
--- a/net/wireless/reg.c
+++ b/net/wireless/reg.c
@@ -2259,6 +2259,8 @@ void /* __init_or_exit */ regulatory_exit(void)
mutex_lock(&cfg80211_mutex);
mutex_lock(®_mutex);
+ dev_set_uevent_suppress(®_pdev->dev, true);
+
reset_regdomains();
kfree(last_request);
Luis
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
