Re: oops on kernel 3.14.17 seems related to EFI

From: Francis Moreau
Date: Mon Oct 06 2014 - 03:44:15 EST


Hello Matt,

On 09/05/2014 02:09 PM, Matt Fleming wrote:
> (Adding linux-efi)
>
> On Fri, 05 Sep, at 08:51:31AM, Francis Moreau wrote:
>> [ +0.000045] RIP: 0010:[<0000000000000000>] [< (null)>] (null)
>> [ +0.000046] RSP: 0018:ffff8800b4001da8 EFLAGS: 00010002
>> [ +0.000032] RAX: 0000000080050033 RBX: ffff880406288000 RCX: ffff880406288000
>> [ +0.000041] RDX: ffff880406288400 RSI: ffff880406288000 RDI: 0000000000000000
>> [ +0.000042] RBP: ffff8800b4001e80 R08: 0000000000000000 R09: ffff8800b4001ec0
>> [ +0.000042] R10: 0000000000000000 R11: 0000000000000246 R12: ffff880406288400
>> [ +0.000042] R13: 0000000000000000 R14: ffff8800b4001ec0 R15: 000000000009b000
>> [ +0.000042] FS: 00007f7720a567c0(0000) GS:ffff88041e2c0000(0000) knlGS:0000000000000000
>> [ +0.000048] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [ +0.000034] CR2: 0000000000000000 CR3: 000000000009b000 CR4: 00000000001427e0
>> [ +0.000042] Stack:
>> [ +0.000014] ffffffff81063281 ffffffff811bd95c 0000000000000246 ffff8800da9e6628
>> [ +0.000052] 8000000300000000 0000000000000000 0000000000000000 ffff8800b4001e50
>> [ +0.000051] 0000000080050033 00007f7722abbc50 00007f7722abbb50 00ffffffffffffff
>> [ +0.000052] Call Trace:
>> [ +0.000022] [<ffffffff81063281>] ? efi_call5+0x71/0xf0
>> [ +0.000035] [<ffffffff811bd95c>] ? getname_flags+0x2c/0x130
>> [ +0.000037] [<ffffffff81062649>] ? virt_efi_get_variable+0x49/0x60
>> [ +0.000051] [<ffffffff813d0181>] efivar_entry_size+0x41/0x80
>> [ +0.000030] [<ffffffff8123aec9>] efivarfs_file_read+0x49/0x100
>> [ +0.000046] [<ffffffff811b2077>] vfs_read+0x97/0x160
>> [ +0.000041] [<ffffffff811b2bc9>] SyS_read+0x59/0xd0
>> [ +0.000041] [<ffffffff81509e29>] system_call_fastpath+0x16/0x1b
>
> This looks like efi.systab->runtime->get_variable is NULL.
>
> Could you send a copy of the dmesg buffer? It might contain some info to
> explain this issue.
>

It happened again and it really does related to loop device: it always
happens after (but not immediately after) I set up/delete loop devices
and maybe after hibernating too.

I attached the dmesg output after the bug triggered again.

Thanks

Attachment: oops.gz
Description: application/gzip