[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 01/12] nbd: rename read_sync and friends
From: |
Vladimir Sementsov-Ogievskiy |
Subject: |
Re: [Qemu-devel] [PATCH 01/12] nbd: rename read_sync and friends |
Date: |
Fri, 2 Jun 2017 16:54:56 +0300 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.1.1 |
02.06.2017 16:49, Eric Blake wrote:
On 06/02/2017 07:00 AM, Vladimir Sementsov-Ogievskiy wrote:
2. _sync suffix
_sync is related to the fact, that nbd_wr_syncv doesn't return if
s/fact,/fact/
write to socket returns EAGAIN. In first implementation
nbd_wr_syncv
just loops while getting EAGAIN, current implementation yields in
this case.
As mentioned in your followup, you may want to rewrite this to:
_sync was originally used (back in commit 7a5ca864 when it was named
wr_sync) to indicate that we looped rather than returned on EAGAIN. But
now we use qio_channel which yields on our behalf rather than giving us
EAGAIN.
hmm, I like my wording (with adding note "... implementaion
nbd_wr_syncv (was wr_sync in 7a5ca8648b) just ...") more, because:
1. not only nbd_wr_syncv has that suffix, so nbd_wr_syncv should be
mentioned (as we mention wr_sync)
2. I don't say about contrast between old and current, I say that they
are similar.
Finally, are you OK with my wording? If I reroll, can I add your r-b?
What final wording are you proposing (full paragraph, not a snippet)?
2. _sync suffix
_sync is related to the fact that nbd_wr_syncv doesn't return if
write to socket returns EAGAIN. In first implementation nbd_wr_syncv
(was wr_sync in 7a5ca8648b) just loops while getting EAGAIN, current
implementation yields in this case.
Why to get rid of it:
- it is normal for r/w functions to be synchronous, so having
additional suffix for it looks redundant (contrariwise, we have
_aio suffix for async functions)
- _sync suffix in block layer is used when function does flush (so
using it for other thing is confusing a bit)
- keep function names short after adding nbd_ prefix
Reviewed-by: Eric Blake <address@hidden>
At any rate, I already gave R-b for the code, so finessing the commit
message doesn't change that if the code remains unchanged.
--
Best regards,
Vladimir