Re: [patch] Real-Time Preemption, -VP-2.6.9-rc4-mm1-U3

From: Dominik Karall
Date: Sat Oct 16 2004 - 15:28:43 EST


On Saturday 16 October 2004 17:24, Ingo Molnar wrote:
> * Dominik Karall <dominik.karall@xxxxxxx> wrote:
> > i'm not sure if this bug depends on preemption patch, or if it is a
> > general one in -mm1 tree.
> >
> > kernel BUG at fs/fat/cache.c:150!
> >
> > EIP is at fat_cache_add+0x135/0x151
> >
> > Process mplayer (pid: 9843, threadinfo=ce414000 task=c8128a20)
> >
> > Call Trace:
> > [<c01a11bb>] fat_get_cluster+0x11f/0x1de
> > [<c01a12c1>] fat_bmap_cluster+0x47/0x98
> > [<c01a13f1>] fat_bmap+0xdf/0x101
> > [<c01a36e8>] fat_get_block+0x30/0x198
> > [<c0152866>] alloc_buffer_head+0x33/0x52
> > [<c01501ea>] create_buffers+0x51/0x86
> > [<c015141c>] block_read_full_page+0x1bd/0x2cc
> > [<c01a36b8>] fat_get_block+0x0/0x198
> > [<c0132f0c>] add_to_page_cache+0x58/0x6e
> > [<c013a0e4>] read_pages+0xbe/0x15e
> > [<c013a49c>] do_page_cache_readahead+0x120/0x19b
> > [<c013a606>] page_cache_readahead+0xef/0x1f8
> > [<c01336ab>] do_generic_mapping_read+0x133/0x54b
> > [<c013c911>] lru_cache_add+0xd/0x4f
> > [<c0133d53>] __generic_file_aio_read+0x1a3/0x218
> > [<c0133ac3>] file_read_actor+0x0/0xed
> > [<c0133ec5>] generic_file_read+0x9c/0xba
> > [<c012b84c>] _mutex_lock+0x20/0x36
> > [<c01244c8>] do_sigaction+0x1dc/0x1f7
> > [<c012b468>] autoremove_wake_function+0x0/0x43
> > [<c012b84c>] _mutex_lock+0x20/0x36
> > [<c0167974>] dnotify_parent+0x35/0x95
> > [<c014e0a9>] vfs_read+0xcd/0x126
> > [<c014e36f>] sys_read+0x41/0x6a
> > [<c0103f7b>] syscall_call+0x7/0xb
>
> indeed this does not seem to be related to the preemption patch. How
> hard is it to reproduce this problem? If it's easy then please try with
> vanilla 2.6.9-rc4-mm1 (and if it breaks too, with 2.6.9-rc4 as well), to
> narrow down where the breakage got introduced.
>
> Ingo

sorry, i tried to reproduce this bug, but can't. i even don't know _when_ this
bug occurred, as i just wanted to take a look in the dmesg output after
loading sg module. but it does not depend on sg, as i unloaded it and tried
again to load.

best regards,
dominik

Attachment: pgp00000.pgp
Description: PGP signature