Re: [Xen-devel] intel-pstate driver blows up if cpufreq_register_driver fails.

["Rafael J. Wysocki" <rjw@xxxxxxx>]

Hi,

Thanks for the report!

Dirk, can you please have a look at that?

Rafael


On Tuesday, March 05, 2013 01:25:41 PM Konrad Rzeszutek Wilk wrote:
> This is what we get:
> 
> [   14.429235] calling  intel_pstate_init+0x0/0x14d @ 1^M^M
> [   14.434268] Intel P-state driver initializing.^M^M
> [   14.438775] BUG: unable to handle kernel NULL pointer dereference at 
> 0000000000000030^M^M
> [   14.446649] IP: [<ffffffff8121a254>] sysfs_remove_group+0x14/0x100^M^M
> [   14.452889] PGD 0 ^M^M
> [   14.454968] Oops: 0000 [#1] SMP ^M^M
> [   14.458261] Modules linked in:^M^M
> [   14.461382] CPU 0 ^M^M
> [   14.463204] Pid: 1, comm: swapper/0 Not tainted 
> 3.9.0-rc1upstream-00021-g9e952b4-dirty #8 MSI MS-7680/H61M-P23 (MS-7680)^M^M
> [   14.474382] RIP: e030:[<ffffffff8121a254>]  [<ffffffff8121a254>] 
> sysfs_remove_group+0x14/0x100^M^M
> [   14.483048] RSP: e02b:ffff8801e3489e58  EFLAGS: 00010296^M^M
> [   14.488422] RAX: 00000000ffffffed RBX: ffffffff81a7cd80 RCX: 
> 0000000000000000^M^M
> [   14.495614] RDX: 0000000000000000 RSI: ffffffff81a7cd80 RDI: 
> 0000000000000000^M^M
> [   14.502806] RBP: ffff8801e3489e78 R08: 0000000000000001 R09: 
> 0000000000000008^M^M
> [   14.510001] R10: 0000000000000008 R11: 00000000fffffffa R12: 
> 0000000000000000^M^M
> [   14.517194] R13: ffffffff81afac08 R14: 000000014c7ca386 R15: 
> 0000000000000000^M^M
> [   14.524389] FS:  00007f49debd2700(0000) GS:ffff8801f0e00000(0000) 
> knlGS:0000000000000000^M^M
> [   14.532532] CS:  e033 DS: 0000 ES: 0000 CR0: 0000000080050033^M^M
> [   14.538340] CR2: 0000000000000030 CR3: 0000000001a0c000 CR4: 
> 0000000000042660^M^M
> [   14.545531] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 
> 0000000000000000^M^M
> [   14.552724] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 
> 0000000000000400^M^M
> [   14.559918] Process swapper/0 (pid: 1, threadinfo ffff8801e3488000, task 
> ffff8801e347c800)^M^M
> [   14.568237] Stack:^M^M
> [   14.570316]  00000000ffffffed ffffffff81a9d520 ffffffff81afac08 
> 000000014c7ca386^M^M
> [   14.577771]  ffff8801e3489ea8 ffffffff814e8690 00000000ffffffff 
> 00000000ffffffed^M^M
> [   14.585223]  ffffffff81a9d520 ffffffff81afac08 ffff8801e3489ec8 
> ffffffff81afad42^M^M
> [   14.592677] Call Trace:^M^M
> [   14.595193]  [<ffffffff81afac08>] ? cpufreq_p4_init+0x55/0x55^M^M
> [   14.601000]  [<ffffffff814e8690>] intel_pstate_exit+0x20/0xd0^M^M
> [   14.606805]  [<ffffffff81afac08>] ? cpufreq_p4_init+0x55/0x55^M^M
> [   14.612612]  [<ffffffff81afad42>] intel_pstate_init+0x13a/0x14d^M^M
> [   14.618593]  [<ffffffff81002124>] do_one_initcall+0x124/0x170^M^M
> [   14.624396]  [<ffffffff81ac1a03>] kernel_init_freeable+0x157/0x1e6^M^M
> [   14.630635]  [<ffffffff81ac1a92>] ? kernel_init_freeable+0x1e6/0x1e6^M^M
> [   14.637050]  [<ffffffff8165add0>] ? rest_init+0xa0/0xa0^M^M
> [   14.642337]  [<ffffffff8165add9>] kernel_init+0x9/0xf0^M^M
> [   14.647538]  [<ffffffff8167c9fc>] ret_from_fork+0x7c/0xb0^M^M
> [   14.652998]  [<ffffffff8165add0>] ? rest_init+0xa0/0xa0^M^M
> [   14.658282] Code: 1f 84 00 00 00 00 00 55 48 89 f2 31 f6 48 89 e5 e8 e2 fd 
> ff ff c9 c3 55 48 89 e5 41 56 41 55 41 54 49 89 fc 53 48 8b 16 48 89 f3 <4c> 
> 8b 77 30 48 85 d2 0f 84 87 00 00 00 31 f6 4c 89 f7 e8 d5 e3 ^M^M
> [   14.677174] RIP  [<ffffffff8121a254>] sysfs_remove_group+0x14/0x100^M^M
> [   14.683498]  RSP <ffff8801e3489e58>^M^M
> [   14.687052] CR2: 0000000000000030^M^M
> [   14.690450] ---[ end trace 104ea7a191a991d2 ]---^M^M
> [   14.695120] swapper/0 (1) used greatest stack depth: 3688 bytes left^M^M
> [   14.701532] Kernel panic - not syncing: Attempted to kill init! 
> exitcode=0x00000009^M^M
> 
> 
> The cpufreq is disabled, so we end up with:
> 
> 797         rc = cpufreq_register_driver(&intel_pstate_driver);
> 798         if (rc)
> 799                 goto out;
> 800 
> 801         intel_pstate_debug_expose_params();
> 802         intel_pstate_sysfs_expose_params();
> 803         return rc;
> 804 out:
> 805         intel_pstate_exit();
> 806         return -ENODEV;
> 
> Since line 797 fails, we call intel_pstate_exit() which tries:
> 
> 754         sysfs_remove_group(intel_pstate_kobject,
> 755                                 &intel_pstate_attr_group);
> 
> 
> and that blows up b/c we had never called intel_pstate_debug_expose_params
> (line 801).
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel