[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: writev is "unfair".
From: |
Sebastian Wilhelmi |
Subject: |
Re: writev is "unfair". |
Date: |
Tue, 28 Nov 2000 10:37:04 +0100 |
Hi Ulrich,
> > I just wanted to know, if such a change would at least have the chance to be
> > included, or if it isn't possible because of e.g. diagnosis semantics, that
> > require to call write inside writev exactly once.
>
> None of this. The caller is responsible. The writev() on all
> architectures guarantees only a given amount of records to written at
> one time. Many OSes use 16, Linux I think has 1024. Use must not
> exceed this limit. If you do we try to recover but it sometimes
> fails.
Ok, but is there a standard out there regarding the maximum number of iovecs?
GLibc seems to be using UIO_MAXIOV, other systems sysconf (_SC_IOV_MAX), other
ones MAXIOV, then we have IOV_MAX... What a mess. Yes, I know, it's not your
fault ;-)
Anyway, I'll shut up soon, but I think, that at least UIO_MAXIOV should be
documented in the writev manpage.Who is responsible for the manpages for
glibc?
Thanks,
Sebastian
--
Sebastian Wilhelmi
mailto:address@hidden
http://goethe.ira.uka.de/~wilhelmi