Re: [LKP] [cpufreq] BUG: sleeping function called from invalid context at mm/slub.c:1241

[Viresh Kumar]

On Fri, Dec 19, 2014 at 10:17 AM, Huang Ying <ying.huang@xxxxxxxxx> wrote:
> FYI, we noticed the below changes on
>
> https://git.linaro.org/people/vireshk/linux cpufreq/stats/cleanups
> commit 2136fa84639adcc0727e968f99ad222c8146810c ("cpufreq: stats: Fix locking")
>
>
> +----------------------+------------+------------+
> |                      | 0063f8602b | 2136fa8463 |
> +----------------------+------------+------------+
> | boot_successes       | 9          | 9          |
> | boot_failures        | 1          | 1          |
> | BUG:kernel_boot_hang | 1          | 1          |
> +----------------------+------------+------------+
>
>
> [   24.295182] Freeing unused kernel memory: 1972K (ffff880001a13000 - ffff880001c00000)
> [   24.304471] Freeing unused kernel memory: 1780K (ffff880002043000 - ffff880002200000)
> [   24.704332] acpi-cpufreq: overriding BIOS provided _PSD data
> [   24.716589] BUG: sleeping function called from invalid context at mm/slub.c:1241
> [   24.726087] in_atomic(): 1, irqs_disabled(): 0, pid: 2393, name: modprobe
> [   24.726092] CPU: 21 PID: 2393 Comm: modprobe Not tainted 3.18.0-g0a0a28c #1
> [   24.726093] Hardware name: Supermicro H8DGU/H8DGU, BIOS 2.0        09/08/11
> [   24.726097]  00000000000004d9 ffff8810071a3ae8 ffffffff81a02d59 000000000f340f34
> [   24.726099]  ffffffff81effec4 ffff8810071a3af8 ffffffff8110adca ffff8810071a3b28
> [   24.726101]  ffffffff8110ae7d 0000000000000000 00000000000080d0 ffff88040f803b00
> [   24.726102] Call Trace:
> [   24.726111]  [<ffffffff81a02d59>] dump_stack+0x4c/0x65
> [   24.726117]  [<ffffffff8110adca>] ___might_sleep+0x10e/0x110
> [   24.726119]  [<ffffffff8110ae7d>] __might_sleep+0xb1/0xb9
> [   24.726123]  [<ffffffff811ea2af>] kmem_cache_alloc_trace+0x48/0x1e6
> [   24.726128]  [<ffffffff818a25df>] ? __cpufreq_stats_create_table+0x68/0x194
> [   24.726131]  [<ffffffff818a25df>] __cpufreq_stats_create_table+0x68/0x194
> [   24.726138]  [<ffffffff818a27b6>] cpufreq_stat_notifier_policy+0x1b/0x32
> [   24.726141]  [<ffffffff81105a7d>] notifier_call_chain+0x6d/0x95
> [   24.726143]  [<ffffffff81105d20>] __blocking_notifier_call_chain+0x4a/0x63
> [   24.726145]  [<ffffffff81105d4d>] blocking_notifier_call_chain+0x14/0x16
> [   24.726148]  [<ffffffff818a181e>] __cpufreq_add_dev+0x71d/0x8ce
> [   24.726150]  [<ffffffff818a1a3c>] cpufreq_add_dev+0xe/0x10
> [   24.726154]  [<ffffffff8155dd93>] subsys_interface_register+0xb8/0xdf
> [   24.726157]  [<ffffffff818a0792>] cpufreq_register_driver+0x156/0x268
> [   24.726159]  [<ffffffffa009f000>] ? 0xffffffffa009f000
> [   24.726163]  [<ffffffffa009f220>] acpi_cpufreq_init+0x220/0x1000 [acpi_cpufreq]
> [   24.726165]  [<ffffffffa009f000>] ? 0xffffffffa009f000
> [   24.726168]  [<ffffffff8100032a>] do_one_initcall+0xfd/0x18f
> [   24.726172]  [<ffffffff811d4383>] ? __vunmap+0xac/0xb7
> [   24.726176]  [<ffffffff81150cd3>] load_module+0x1a35/0x1ff7
> [   24.726179]  [<ffffffff81205653>] ? kernel_read+0x48/0x5f
> [   24.726183]  [<ffffffff811513ef>] SyS_finit_module+0x85/0x92
> [   24.726187]  [<ffffffff81a0a1e9>] system_call_fastpath+0x12/0x17
> [   25.539283] microcode: CPU0: patch_level=0x06000613
> [   25.544884] microcode: CPU1: patch_level=0x06000613
> [   25.550137] microcode: CPU2: patch_level=0x06000613

Okay, thanks Huang..

The problem is that we are trying to allocate memory from within spinlock's
critical region..

Fixed by replacing spinlock with a mutex. Pushed my branch again.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/