Question(s) about smbd in respect to preadv2

[Volker Lendecke]

On Sat, Jan 24, 2015 at 03:46:18PM -0500, Milosz Tanski wrote:
> The results were that sync was fastest (no surprise there).
> Pthreadpool + preadv2 was about 8% slower then sync. Plain old
> pthreadpool 26% slower. So not a bad win there. Additionally, it looks
> like the vfs_pread_send and vfs_pread_recv have a bit more overhead
> over plain old vfs_pread in the code path so it's possible to get that
> 8% even closer to the sync case.

Indeed. The change I proposed was really the minimum number
of code lines to be achieved as quickly as possible. If
preadv2 succeeds we don't have to set up the tevent_req
stuff at all, which might add the overhead (basically malloc
cost). But this would happen at a different layer with a
little more code effort (possibly a VFS change).

Volker

-- 
SerNet GmbH, Bahnhofsallee 1b, 37081 Göttingen
phone: +49-551-370000-0, fax: +49-551-370000-9
AG Göttingen, HRB 2816, GF: Dr. Johannes Loxen
http://www.sernet.de, mailto:kontakt at sernet.de