[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150320130313.GB16485@kernel.org>
Date: Fri, 20 Mar 2015 10:03:13 -0300
From: Arnaldo Carvalho de Melo <acme@...nel.org>
To: Vinson Lee <vlee@...pensource.com>
Cc: Jiri Olsa <jolsa@...hat.com>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>,
Paul Mackerras <paulus@...ba.org>,
Ingo Molnar <mingo@...hat.com>,
Adrian Hunter <adrian.hunter@...el.com>,
Masami Hiramatsu <masami.hiramatsu.pt@...achi.com>,
Namhyung Kim <namhyung@...nel.org>,
Yann Droneaud <ydroneaud@...eya.com>,
LKML <linux-kernel@...r.kernel.org>,
linux-perf-users@...r.kernel.org, Vinson Lee <vlee@...tter.com>
Subject: Re: [PATCH v2] perf tools: Work around lack of sched_getcpu in glibc
< 2.6.
Em Thu, Mar 19, 2015 at 02:12:05PM -0700, Vinson Lee escreveu:
> On Thu, Mar 19, 2015 at 6:14 AM, Arnaldo Carvalho de Melo
> <acme@...nel.org> wrote:
> > Em Thu, Mar 19, 2015 at 11:13:04AM +0100, Jiri Olsa escreveu:
> >> On Wed, Mar 18, 2015 at 04:11:52PM -0700, Vinson Lee wrote:
> >> > This patch fixes this build error with glibc < 2.6.
> >
> >> > +#ifdef __GLIBC_PREREQ
> >> > +#if !__GLIBC_PREREQ(2, 6)
> >> > +static int sched_getcpu(void)
> >> > +{
> >> > + errno = ENOSYS;
> >> > + return -1;
> >> > +}
> >> > +#endif
> >> > +#endif
> >
> >> you could mark it as __weak and get rid of those ifdef's
> >> like we do for strlcpy in utilpath.c
> >
> > Agreed.
> I tried this approach but am having trouble getting a patch to compile
> with both older glibc and recent glibc.
> With older glibc I will get a "no previous prototype" compilation
> error. I can fix this error by adding a prototype but this results in
> a "redundant redeclaration" error with newer glibc.
So for strlcpy we have in tools/perf/util/cache.h:
#ifndef __UCLIBC__
/* Matches the libc/libbsd function attribute so we declare this
* unconditionally: */
extern size_t strlcpy(char *dest, const char *src, size_t size);
#endif
I.e. have the ifdef just in the header file.
This way, if we end up needing to use sched_getcpu elsewhere, it will
work and the needed #ifdef will be constrained to the header file.
At some point we should try to move these functions to some
libc-compat.c file and move it to somewhere in tools/lib/.
- Arnaldo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists