linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* btrfs_search_slot BUG...
@ 2012-03-09  4:25 Daniel J Blueman
  2012-03-09  4:34 ` Liu Bo
  0 siblings, 1 reply; 3+ messages in thread
From: Daniel J Blueman @ 2012-03-09  4:25 UTC (permalink / raw)
  To: Chris Mason; +Cc: Linux BTRFS, Josef Bacik

When testing out 16KB blocks with direct I/O [1] on 3.3-rc6, we
quickly see btrfs_search_slot returning positive numbers, popping an
assertion [2].

Are >4KB block sizes known broken for now?

Thanks,
  Daniel

--- [1]

mkfs.btrfs -m raid1 -d raid1 -l 16k -n 16k /dev/sda /dev/sdb
mount /dev/sda /store && cd /store
fio /usr/share/doc/fio/examples/iometer-file-access-server

--- [2]

kernel BUG at /home/apw/COD/linux/fs/btrfs/extent-tree.c:1481!
-- 
Daniel J Blueman

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: btrfs_search_slot BUG...
  2012-03-09  4:25 btrfs_search_slot BUG Daniel J Blueman
@ 2012-03-09  4:34 ` Liu Bo
  2012-03-09 14:01   ` David Sterba
  0 siblings, 1 reply; 3+ messages in thread
From: Liu Bo @ 2012-03-09  4:34 UTC (permalink / raw)
  To: Daniel J Blueman; +Cc: Chris Mason, Linux BTRFS, Josef Bacik

On 03/09/2012 12:25 PM, Daniel J Blueman wrote:
> When testing out 16KB blocks with direct I/O [1] on 3.3-rc6, we
> quickly see btrfs_search_slot returning positive numbers, popping an
> assertion [2].
> 
> Are >4KB block sizes known broken for now?
> 

Yes, exactly, the kernel side is not ready for >4K block size, but AFAIK it _is_ on going.

thanks,
liubo

> Thanks,
>   Daniel
> 
> --- [1]
> 
> mkfs.btrfs -m raid1 -d raid1 -l 16k -n 16k /dev/sda /dev/sdb
> mount /dev/sda /store && cd /store
> fio /usr/share/doc/fio/examples/iometer-file-access-server
> 
> --- [2]
> 
> kernel BUG at /home/apw/COD/linux/fs/btrfs/extent-tree.c:1481!


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: btrfs_search_slot BUG...
  2012-03-09  4:34 ` Liu Bo
@ 2012-03-09 14:01   ` David Sterba
  0 siblings, 0 replies; 3+ messages in thread
From: David Sterba @ 2012-03-09 14:01 UTC (permalink / raw)
  To: Liu Bo; +Cc: Daniel J Blueman, Chris Mason, Linux BTRFS, Josef Bacik

On Fri, Mar 09, 2012 at 12:34:01PM +0800, Liu Bo wrote:
> On 03/09/2012 12:25 PM, Daniel J Blueman wrote:
> > When testing out 16KB blocks with direct I/O [1] on 3.3-rc6, we
> > quickly see btrfs_search_slot returning positive numbers, popping an
> > assertion [2].
> > 
> > Are >4KB block sizes known broken for now?
> 
> Yes, exactly, the kernel side is not ready for >4K block size, but AFAIK it _is_ on going.

This is connected rather to PAGE_SIZE and PAGE_CACHE_SIZE, which happens
to be 4k on x86_64, but eg. ia64 or powerpc64 are happy with 64k blocks
(leaf and node).


david

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2012-03-09 14:01 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-03-09  4:25 btrfs_search_slot BUG Daniel J Blueman
2012-03-09  4:34 ` Liu Bo
2012-03-09 14:01   ` David Sterba

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).