On Thu, 18 May 2000, Mike Galbraith wrote:
> On Wed, 17 May 2000, Jeff Garzik wrote:
>
> > David Weinehall wrote:
> > > > To Do
> > > > -----
> > > > Loopback fs hangs
> > >
> > > Definitely show-stopper. Almost every distro (if not every) uses Loopback
> > > for their installers, and Loopback is used for a lot of other things too.
> >
> > Are we talking about a loopback filesystem, or the loopback block
> > device? Andries' patch for the loopback block device made it into a
> > recent 2.3.x kernel.
> >
> >
> >
> > So, maybe these two items can be crossed off, after testing?
>
> /dev/loop is busted, and not only wrt lockup. I used to be able to
> beat on it _hard_ if I prevented plugging. Now, it doesn't seem to
> be reading correctly (cpp gripes on simple compile attempt) and disk
> io (iozone/bonnie) blows up the kernel quite spectacularly. The last
> few entries of the longest (kdb) stack trace I've ever seen goes..
> kernel_thread
> bdflush+0x6d
> flush_dirty_buffers+0x84
> ... ~100 lines of processing loop device requests
> do_try_to_free_pages+0x39
> shrink_mmap+0xa0
> try_to_free_buffers+0x149
> sync_page_buffers+0x2f
> ll_rw_block+0x16
> __ll_rw_block+0x146
> generic_make_request+0x5ed
> __get_request_wait+0x279
> schedule+0x2a
> BUG at sched.c:445
> ..and is repeatable.
silence.. <shrug> See busy_buffer_page in buffer.c and ponder the
consequences thereof.. not the only problem (by far), but it's a start.
Later,
-Mike
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue May 23 2000 - 21:00:16 EST