On Thu, Jan 28, 2016 at 09:32:11AM -0700, Ross Zwisler wrote:

On Thu, Jan 28, 2016 at 02:16:30PM +0100, Jan Kara wrote:

quoted

On Wed 27-01-16 12:01:48, Ross Zwisler wrote:

quoted

As it is currently written ext4_dax_mkwrite() assumes that the call into
__dax_mkwrite() will not have to do a block allocation so it doesn't create
a journal entry.  For a read that creates a zero page to cover a hole
followed by a write that actually allocates storage this is incorrect.  The
ext4_dax_mkwrite() -> __dax_mkwrite() -> __dax_fault() path calls
get_blocks() to allocate storage.

Fix this by having the ->page_mkwrite fault handler call ext4_dax_fault()
as this function already has all the logic needed to allocate a journal
entry and call __dax_fault().

Also update the ext2 fault handlers in this same way to remove duplicate
code and keep the logic between ext2 and ext4 the same.

Signed-off-by: Ross Zwisler <redacted>

Ah, ok, you are right. The patch looks good but Matthew is reworking the
area more (so ext4_da_mkwrite() is likely to return) so it this worth it?
Or do you expect Matthew's patches to land much later?

Yep, Matthew is in the process of reworking all of the DAX fault handling.

I was thinking that we might want to take this patch for v4.5, since it fixes
a bug that I'm guessing could lead to some sort of corruption (lack of a
journal entry entry for an allocating write), and then Matthew's reworks would
land in v4.6?

Hey Jan,

Looks like this patch didn't ever get merged for v4.5?  Is it still queued for
v4.6?

Thanks,
- Ross