On Tue, Jun 04, 2019 at 12:04:47PM -0700, Rosen Penev wrote: > Needed for libraries like musl that do not implement the GNU variant. This patch seems curious. > Signed-off-by: Rosen Penev <rosenp(a)gmail.com&gt; > --- > traceevent/event-parse.c | 5 +++++ > 1 file changed, 5 insertions(+) > > diff --git a/traceevent/event-parse.c b/traceevent/event-parse.c > index 5a717a0..1f7eb80 100644 > --- a/traceevent/event-parse.c > +++ b/traceevent/event-parse.c > @@ -5121,12 +5121,17 @@ int pevent_strerror(struct pevent *pevent __maybe_unused, > const char *msg; > > if (errnum >= 0) { > +#if defined(__GLIBC__) > msg = strerror_r(errnum, buf, buflen); > if (msg != buf) { > size_t len = strlen(msg); > memcpy(buf, msg, min(buflen - 1, len)); > *(buf + min(buflen - 1, len)) = '\0'; > } > +#else > + if (strerror_r(errnum, buf, buflen)) > + snprintf(buf, buflen, "errnum %i", errnum); Was if (a human readable response is returned from strerror_r) clobber it with a numeric version of the error really what you wanted to achieve? I would rather expect something like if (errnum >= 0) { msg = strerror_r(errnum, buf, buflen);

#if defined(__GLIBC__) if (msg != buf) strlcpy(buf, msg, buflen); #endif } where the #ifdef ends up as a pure optimization if strerror_r is the XSI version. /MF > +#endif > return 0; > } > > -- > 2.17.1 > > _______________________________________________ > PowerTop mailing list > PowerTop(a)lists.01.org > https://lists.01.org/mailman/listinfo/powertop