kernel NULL pointer dereference in check_spread+0x0/0x26
This is with 2.6.23-mm1 and allmodconfig.
Seems that se in the following is a NULL pointer.
453 static void check_spread(struct cfs_rq *cfs_rq, structsched_entity *se)
454 {
455 #ifdef CONFIG_SCHED_DEBUG
456 s64 d = se->vruntime - cfs_rq->min_vruntime;
457
458 if (d < 0)
459 d = -d;
Cheers,
Jan
--
[ 1.344000] Unable to handle kernel NULL pointer dereference at 0000000000000040 RIP: [ 1.348000] [<ffffffff8022dd7d>] check_spread+0x0/0x26 [ 1.356000] PGD 0 [ 1.360000] Oops: 0000 [1] SMP [ 1.364000] last sysfs file: [ 1.368000] CPU 1 [ 1.368000] Modules linked in: [ 1.372000] Pid: 2, comm: kthreadd Not tainted 2.6.23-mm1-jbl-gab69b1c9 #4 [ 1.380000] RIP: 0010:[<ffffffff8022dd7d>] [<ffffffff8022dd7d>] check_spread+0x0/0x26 [ 1.388000] RSP: 0018:ffff810005753d58 EFLAGS: 00010083 [ 1.392000] RAX: 0000000000001908 RBX: ffff81008006b900 RCX: 0000000000000c31 [ 1.400000] RDX: 0000000003938700 RSI: 0000000000000000 RDI: ffff81008006b900 [ 1.408000] RBP: ffff810005753d90 R08: ffff810005753d40 R09: ffff8100800d0798 [ 1.416000] R10: ffffffffffffffff R11: 0000000000000001 R12: ffff810005c12000 [ 1.424000] R13: ffff810005c12048 R14: 0000000000000000 R15: 0000000000000001 [ 1.428000] FS: 0000000000000000(0000) GS:ffff810005401960(0000) knlGS:0000000000000000 [ 1.440000] CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b [ 1.444000] CR2: 0000000000000040 CR3: 0000000000201000 CR4: 00000000000006e0 [ 1.452000] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 1.460000] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [ 1.468000] Process kthreadd (pid: 2, threadinfo ffff810005752000, task ffff810005750000) [ 1.476000] last branch before last exception/interrupt [ 1.480000] from [<ffffffff802363f2>] task_new_fair+0xd0/0x111 [ 1.488000] to [<ffffffff8022dd7d>] check_spread+0x0/0x26 [ 1.492000] Stack: ffffffff802363f7 ffff8100800d0780 ffff8100800d0780 ffff810005c12000 [ 1.500000] 000000000000000f 0000000000000000 0000000000000000 ffff810005753dc0 [ 1.508000] ffffffff80236cf1 0000000000800711 0000000000000286 ffff810005c12000 [ 1.516000] Call Trace: [ 1.520000] [<ffffffff802363f7>] task_new_fair+0xd5/0x111 [ 1.524000] [<ffffffff80236cf1>] wake_up_new_task+0x84/0xb8 [ 1.532000] [<ffffffff802395be>] do_fork+0x186/0x286 [ 1.536000] [<ffffffff8025e101>] __lock_acquire+0xe4b/0xf40 [ 1.544000] [<ffffffff8020d021>] kernel_thread+0x81/0xde [ 1.548000] [<ffffffff8024fc72>] kthread+0x0/0x76 [ 1.552000] [<ffffffff8020d07e>] child_rip+0x0/0x12 [ 1.560000] [<ffffffff8024fad1>] kthreadd+0xcd/0x143 [ 1.564000] [<ffffffff8020d088>] child_rip+0xa/0x12 [ 1.568000] [<ffffffff8020c79c>] restore_args+0x0/0x30 [ 1.576000] [<ffffffff802a9bea>] __mpol_free+0x3b/0x40 [ 1.580000] [<ffffffff8024fa04>] kthreadd+0x0/0x143 [ 1.584000] [<ffffffff8020d07e>] child_rip+0x0/0x12 [ 1.592000] [ 1.592000] INFO: lockdep is turned off. [ 1.596000] [ 1.596000] Code: 48 8b 46 40 48 2b 47 20 55 48 89 e5 48 99 48 31 d0 48 29 d0 [ 1.604000] RIP [<ffffffff8022dd7d>] check_spread+0x0/0x26 [ 1.612000] RSP <ffff810005753d58> [ 1.616000] CR2: 0000000000000040 [ 1.620000] kthreadd used greatest stack depth: 5512 bytes left [ 36.348000] BUG: spinlock lockup on CPU#2, swapper/0, ffff8100800d0780 [ 36.356000] [ 36.356000] Call Trace: [ 36.360000] <IRQ> [<ffffffff80357e07>] _raw_spin_lock+0x126/0x14e [ 36.368000] [<ffffffff80449e5f>] _spin_lock+0x46/0x53 [ 36.372000] [<ffffffff80235547>] scheduler_tick+0x42/0x1d8 [ 36.376000] [<ffffffff80244684>] update_process_times+0x82/0x92 [ 36.384000] [<ffffffff80257bd6>] tick_periodic+0x6e/0x7a [ 36.388000] [<ffffffff80257c03>] tick_handle_periodic+0x21/0x5e [ 36.396000] [<ffffffff8020b244>] default_idle+0x0/0x64 [ 36.400000] [<ffffffff8021d226>] smp_local_timer_interrupt+0x5a/0x5e [ 36.408000] [<ffffffff8021d83e>] smp_apic_timer_interrupt+0x3a/0x54 [ 36.416000] [<ffffffff8020b244>] default_idle+0x0/0x64 [ 36.420000] [<ffffffff8020ceab>] apic_timer_interrupt+0x6b/0x70 [ 36.428000] <EOI> [<ffffffff8020b28e>] default_idle+0x4a/0x64 [ 36.432000] [<ffffffff8020b28c>] default_idle+0x48/0x64 [ 36.440000] [<ffffffff8020b365>] cpu_idle+0xbd/0xf8 [ 36.444000] [<ffffffff8021cd77>] start_secondary+0x3d6/0x3e7 [ 36.448000] [ 36.452000] INFO: lockdep is turned off.
Quoted 1 line · Jan Blunck
This is with 2.6.23-mm1 and allmodconfig.
This generates a .config with CONFIG_SCHED_DEBUG=y and CONFIG_FAIR_GROUP_SCHED=n (The latter causes parent_entity to return NULL).
Does setting CONFIG_FAIR_GROUP_SCHED=y help?
Regards,
Frederik
Quoted 6 lines · Frederik Deweerdt
Quoted 1 line · Jan Blunck
This is with 2.6.23-mm1 and allmodconfig.
This generates a .config with
CONFIG_SCHED_DEBUG=yandCONFIG_FAIR_GROUP_SCHED=n(The latter causesparent_entityto return NULL).
Does settingCONFIG_FAIR_GROUP_SCHED=yhelp?
No, CONFIG_FAIR_GROUP_SCHED=y was set. Or did you mean CONFIG_FAIR_CGROUP_SCHED?
Quoted 10 lines · Jan Blunck
Quoted 6 lines · Frederik Deweerdt
Quoted 1 line · Jan Blunck
This is with 2.6.23-mm1 and allmodconfig.
This generates a .config with
CONFIG_SCHED_DEBUG=yandCONFIG_FAIR_GROUP_SCHED=n(The latter causesparent_entityto return NULL).
Does settingCONFIG_FAIR_GROUP_SCHED=yhelp?No,
CONFIG_FAIR_GROUP_SCHED=ywas set. Or did you meanCONFIG_FAIR_CGROUP_SCHED?
He no, I misread the code, sorry for the noise.
Quoted 4 lines · Jan Blunck
This is with 2.6.23-mm1 and allmodconfig.
Seems that se in the following is a NULL pointer.
Quoted 8 lines
453 static void check_spread(struct
cfs_rq*cfs_rq, structsched_entity*se)
454 {
455 #ifdefCONFIG_SCHED_DEBUG
456 s64 d =se->vruntime-cfs_rq->min_vruntime;
457
458 if (d < 0)
459 d = -d;
I hit this once yesterday (in sched-devel), and again first boot this morning.
-Mike
Quoted 4 lines · Jan Blunck
This is with 2.6.23-mm1 and allmodconfig.
Seems that se in the following is a NULL pointer.
Quoted 8 lines
453 static void check_spread(struct
cfs_rq*cfs_rq, structsched_entity*se)
454 {
455 #ifdefCONFIG_SCHED_DEBUG
456 s64 d =se->vruntime-cfs_rq->min_vruntime;
457
458 if (d < 0)
459 d = -d;
Have you hit this again recently? You and I seem to be the only ones who have hit it, and here, it either went away, or went into hiding. I've been trying to tickle this (highly improbable) oops with fork/exit bombs while other things are running (scribble scribble), and failing to provoke any adverse reaction.
-Mike