[dm-devel] [PATCH 0/2] patches to improve cluster raid1 performance [V2]

[Dongmao Zhang]

于 2013年10月11日 00:29, Brassow Jonathan 写道:
> Perhaps a better thing to do is to integrate the marks (and possibly
> clears) into the flush - DM_INTEGRATED_FLUSH.  The payload for the
> DM_ULOG_FLUSH flush communication is empty.  It seems to me now that
> it would make more sense to make use of that empty space and fill it
> with mark/clear requests.  What do you think?
>

DM_INTEGRATED_FLUSH is a good idea. It just simplify the confusion of
DM_FLUSH_WITH_MARK thing. My plan is to add (mark and clear) payload to
flush. It is like this:

'X' means we have this kind of request, while '0' means none.

mark request :     X               X               0

clear request:     X               0               X

flush method:    payload_flush   payload_flush   delayed flush.

If there is only clear requests, we can send the clear request first, 
and send normal flush later.  how do you like this strategy?

As for the testing, I only have ocfs2 available.(not GFS2). I will
test on that;

Dongmao Zhang.