Messages in this thread | | | Date | Tue, 17 Oct 2000 14:24:58 -0400 (EDT) | From | Alexander Viro <> | Subject | Re: [BUG]: Ext2 Corruption in test10pre3 (incl. Oops) |
| |
On Tue, 17 Oct 2000, Udo A. Steinberg wrote:
> Kernel bug at ll_rw_blk.c: 713!
unmapped buffer got to the ll_rw_block()
> Trace; c0184d53 <ll_rw_block+163/1e0> > Trace; c012fa31 <bread+31/70>
What? <thinking> OK, so we got a unmapped bh hashed at some point. Either it was inserted into hash while it was unmapped or it had been hashed and then unmapped. The latter could happen only if block_flushpage() got a bh associated with page _and_ sitting in the hash. That, in turn, means insertion of page-bound bh into hash at some earlier point. IOW, __insert_into_queues() had been called on unmapped or page-bound bh. That returns us to getblk(). OK, unmapped is out of question and page-bound means that we got a page-bound bh on a freelist. Very bad. And I don't believe that it's fs-related. Frankly, I don't see where such thing might happen.
> Trace; c014efde <read_inode_bitmap+3e/90> > Trace; c014f240 <load_inode_bitmap+210/230> > Trace; c014f6af <ext2_new_inode+29f/700> > Trace; c021e87e <unix_write_space+2e/50> Huh? > Trace; c01523af <ext2_create+1f/c0>
The rest of trace is OK, but WTF is net/unix/*.c code is doing here?
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
| |