This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 0/6] Some vfprintf refactoring
- From: Florian Weimer <fweimer at redhat dot com>
- To: "Carlos O'Donell" <carlos at redhat dot com>, libc-alpha at sourceware dot org
- Date: Fri, 06 Mar 2015 15:21:09 +0100
- Subject: Re: [PATCH 0/6] Some vfprintf refactoring
- Authentication-results: sourceware.org; auth=none
- References: <cover dot 1425246936 dot git dot fweimer at redhat dot com> <54F8AFE7 dot 4040603 at redhat dot com> <54F8C25C dot 5010606 at redhat dot com>
On 03/05/2015 09:53 PM, Florian Weimer wrote:
> GCC will inline the new printf_positional function because it is called
> in just one place, so any differences should be just random noise by
> slightly different optimizer choices.
I forgot to mention in my message about the performance impact that this
is not actually true: GCC will not inline printf_positional because of
the way it uses label addresses/computed gotos. But as explained in the
other message, this does not make the code worse (with GCC 4.9 at least):
<https://sourceware.org/ml/libc-alpha/2015-03/msg00266.html>
--
Florian Weimer / Red Hat Product Security