Now that we have bio_list in include/linux/bio.h, I wanted to see what
would happen when I replaced rq->{bio,biotail} with rq->bio_list.
Personally, I think the result is more readable, and indicates to all
users that rq->bio is really a list (even if a list with one entry).
Also, it highlights some bio users in downstream drivers, and hopefully
serves to increase the amount of bio-related review in those drivers.
The first patch is a straightforward replacement, with no code or
behavior changes (any such is a bug in the patch...):
- null/not-null tests become bio_list_empty()
- rq->bio becomes rq->bio_list.head
- rq->biotail becomes rq->bio_list.tail
- in a few cases, bio_list_xxx functions are called
as appropriate
The second patch are fixes to what I believe are minor bugs in the
bio-list-aware block core. Even if patch #1 is not accepted, these
fixes are likely needed regardless. Typically the bugs are of the type
"we forgot to update rq->biotail".
But maybe some of those are on purpose. Who knows, give it a
look... it is quite muddled which block core functions want 'bio'
as a list, or just a singleton 'bio'. Maybe I just got confused.
If nothing else, it helps to give this area of block another look, IMO.
Jeff
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
