On Tue, May 26, 2020 at 07:38:34PM +0100, Ken Moffat wrote: > On Tue, May 26, 2020 at 01:32:55PM -0400, Don Cross wrote: > > On Tue, May 26, 2020 at 1:17 PM Ken Moffat <zarniwh...@ntlworld.com> wrote: > > > > > I'm trying to discover why two tests in the testsuite for 'check' > > > always fail for me. I've now got a response that the tests raise > > > SIGFPE but do not receive it. The example is > > > > > i.e. it apparently succeeded. > > I then ran strace on the original version - > > write(1, "Before\n", 7) = 7 > rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0 > getpid() = 1567 > gettid() = 1567 > tgkill(1567, 1567, SIGFPE) = 0 > rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 > --- SIGFPE {si_signo=SIGFPE, si_code=SI_TKILL, si_pid=1567, si_uid=1000} --- > write(1, "After\n", 6) = 6 > exit_group(0) = ? > +++ exited with 0 +++ > > Normally I like unexpected successes, but not in this case. >
I've now had a reply on the check list suggesting that I try rebuilding with -O0. Finding a machine, and cpu time, where I can do that (and, I suppose complete enough to boot it if the test still fails in chroot) will take time and get interrupted by known changes due this weekend. Looking back, I think I have been normally using CFLAGS of -O2 -march=native when I first noticed this in 'check'. And now I've moved on to -O3 -march=native. I did do one or two 'generic' builds without -march, but on those I only ran the essential tests. However: 1. My old SandyBridge i3 (now defunct - it was too slow to use, and couldn't support an HDMI monitor so I tried to strip it to reuse the case, only to find the PSU, on/off switch, and case were not standard) seems to have not had problems with SIGFPE, neither in check nor in bash. 2. My current server is an athlon 200ge - installed by hacking a system from a defunct Kaveri enough to get it to build a fresh system - and on that one there were no SIGFPE problems in either bash or check. I tried the test, and there it works (built with CFLAGS of -O2 -march=native etc). Oddly enough, the logs from the Kaveri, also built with -O2 -march=native, did not show any SIGFPE problems. Thanks for the help. ĸen -- Do you not know that, what you belittle by the name tree is but the mere four-dimensional analogue of a whole multidimensional universe which - no, I can see you do not. -- Druellae (a Dryad) -- http://lists.linuxfromscratch.org/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page Do not top post on this list. A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? http://en.wikipedia.org/wiki/Posting_style