[Linux-cachefs] 'bad page state' error

David Howells dhowells at redhat.com
Thu Sep 5 10:44:20 UTC 2013


Rob Bos <rbos at sfu.ca> wrote:

> Jun 30 11:44:42 lib-hf3 kernel: BUG: Bad page state in process aria2c  pfn:6b5b8
> Jun 30 11:44:42 lib-hf3 kernel: page:ffffea000177c040 flags:0020000000001000 count:0 mapcount:0 mapping:(null) index:0 (Tainted: G           ---------------  T)
> Jun 30 11:44:42 lib-hf3 kernel: Pid: 2121, comm: aria2c Tainted: G           ---------------  T 2.6.32-373.el6.rbos.2.x86_64 #1
> Jun 30 11:44:42 lib-hf3 kernel: Call Trace:
> Jun 30 11:44:42 lib-hf3 kernel: [<ffffffff811291d7>] ? bad_page+0x107/0x160
> Jun 30 11:44:42 lib-hf3 kernel: [<ffffffff8112c959>] ? free_hot_cold_page+0x1c9/0x220
> Jun 30 11:44:42 lib-hf3 kernel: [<ffffffff81175a28>] ? mem_cgroup_cache_charge+0xb8/0xc0
> Jun 30 11:44:42 lib-hf3 kernel: [<ffffffff8112ca6f>] ? free_hot_page+0x2f/0x60
> Jun 30 11:44:42 lib-hf3 kernel: [<ffffffff8112f69e>] ? __put_single_page+0x1e/0x30
> Jun 30 11:44:42 lib-hf3 kernel: [<ffffffff8112f825>] ? put_page+0x25/0x40
> Jun 30 11:44:42 lib-hf3 kernel: [<ffffffffa0240cf8>] ? cifs_readpages+0x2a8/0x6e0 [cifs]

I think I have a rough idea as to the problem that causes this is now.  The
only problem is I don't know exactly which bit of cifs_readpages() is causing
it.

Could you do me a favour and load your cifs.ko into gdb and then do:

	i line *((unsigned long)cifs_readpages + 0x2a8)

That should print the number of the line with the relevant call to put_page().

Thanks,
David




More information about the Linux-cachefs mailing list