Messages in this thread | | | From | "Rafael J. Wysocki" <> | Subject | Re: suspend-related lockdep warning | Date | Sun, 2 Dec 2007 21:33:23 +0100 |
| |
On Saturday, 1 of December 2007, Pavel Machek wrote: > Hi! > > > 2.6.24-rc3-mm2 (which will be released if it boots on two more machines and > > if I stay awake) will say this during suspend-to-RAM on the Vaio: > > > > [ 91.876445] Syncing filesystems ... done. > > [ 92.382595] Freezing user space processes ... WARNING: at kernel/lockdep.c:2662 check_flags() > > [ 92.384000] Pid: 1925, comm: dbus-daemon Not tainted 2.6.24-rc3-mm2 #32 > > [ 92.384177] [<c0104a74>] show_trace_log_lvl+0x12/0x25 > > [ 92.384335] [<c01052ff>] show_trace+0xd/0x10 > > [ 92.384469] [<c01059be>] dump_stack+0x55/0x5d > > [ 92.384605] [<c0136a29>] check_flags+0x7f/0x11a > > [ 92.384746] [<c0139cfb>] lock_acquire+0x3a/0x86 > > [ 92.384886] [<c031f4de>] _spin_lock+0x26/0x53 > > [ 92.385023] [<c0141d3b>] refrigerator+0x13/0xc8 > > [ 92.385163] [<c01274c8>] get_signal_to_deliver+0x32/0x3fb > > [ 92.385326] [<c01030f4>] do_notify_resume+0x8c/0x699 > > [ 92.385476] [<c0103c18>] work_notifysig+0x13/0x1b > > [ 92.385620] ======================= > > [ 92.385719] irq event stamp: 309 > > [ 92.385809] hardirqs last enabled at (309): [<c0103c79>] syscall_exit_work+0x11/0x26 > > [ 92.386045] hardirqs last disabled at (308): [<c0103b42>] syscall_exit+0x14/0x25 > > [ 92.386265] softirqs last enabled at (0): [<c011bea5>] copy_process+0x374/0x130e > > [ 92.386491] softirqs last disabled at (0): [<00000000>] 0x0 > > [ 92.392457] (elapsed 0.00 seconds) done. > > [ 92.392581] Freezing remaining freezable tasks ... (elapsed 0.00 seconds) done. > > [ 92.392882] PM: Entering mem sleep > > [ 92.392974] Suspending console(s) > > > > this has been happening for quite some time and might even be happening in > > mainline. > > Is it complaining that we entered refrigerator with irqs disabled?
Or that someone else called task_lock() with irqs disabled at one point ...
Hm, perhaps it's related to kernel preemption. Andrew, I guess the kernel is preemptible?
Greetings, Rafael
| |