Zerrout trick allow us to optimize cases where it is more reasonable to explicitly zeroout extent and mark it as initialized instead of splitting to several small ones. But this optimization is not acceptable is extent is beyond i_size Because it is not possible to have initialized blocks after i_size. Fsck treat this as incorrect inode size. BUG# 15742 Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org> --- fs/ext4/extents.c | 49 ++++++++++++++++++++++++++++++++++++++----------- 1 ...

https://bugzilla.kernel.org/show_bug.cgi?id=15742 --- Comment #2 from Dmitry Monakhov <dmonakhov@openvz.org> 2010-04-09 17:24:00 --- Proposed patch: http://patchwork.ozlabs.org/patch/49856/ -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching the assignee of the bug. --

--

Yeah, I'm very much aware of that. What worries me is that locking problems in the jbd2 layer could be very hard to debug, so we need to make sure we have some really good testing as we make any changes. Not taking the j_state_lock spinlock in jbd2_stop_lock() was relatively easy to prove to be safe, but I'm really worried about start_this_handle() the locking around that is going to be subtle, and it's not just the specific fields in the transaction and journal handle. And even with the ...

Nick Piggin's vfs scalability patches takes care of the dcache_lock contention. I'm actually using them with the -rt patch in my testing Yep. This also shows up in my -rt patch testing with Ted's patch. thanks -john --

It seems that if you mount the filesystem with -o sync this problem goes away. --