Re: [PATCH] dax: fix deadlock in __dax_fault

[Dave Chinner <david@fromorbit.com>]

On Tue, Sep 29, 2015 at 08:04:21PM -0600, Ross Zwisler wrote:
> On Tue, Sep 29, 2015 at 12:44:58PM +1000, Dave Chinner wrote:
> <>
> > Already testing a kernel with those reverted. My current DAX patch
> > stack is (bottom is first commit in stack):
> > 
> > f672ae4 xfs: add ->pfn_mkwrite support for DAX
> > 6855c23 xfs: remove DAX complete_unwritten callback
> > e074bdf Revert "dax: fix race between simultaneous faults"
> > 8ba0157 Revert "mm: take i_mmap_lock in unmap_mapping_range() for DAX"
> > a2ce6a5 xfs: DAX does not use IO completion callbacks
> > 246c52a xfs: update size during allocation for DAX
> > 9d10e7b xfs: Don't use unwritten extents for DAX
> > eaef807 xfs: factor out sector mapping.
> > e7f2d50 xfs: introduce per-inode DAX enablement
> 
> Dave, would you be willing to share these patches with me, even if they are
> just RFC?  I'm working through how to add equivalent support in both ext2 and
> ext4, and a conceptual example in XFS would be really helpful.

When I have code that works properly and isn't fundamentally broken,
I'll post it. Hopefully within the next day.

> Eric and Jan, if you guys have any spare cycles to help with the ext2/ext4
> bits, I certainly wouldn't say no.  :)

ext2 already does the block zeroing in allocation, so it only needs
locking help...

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com