On 5/7/19 9:04 PM, Ken Moffat wrote:
On Tue, May 07, 2019 at 05:50:28PM -0400, [email protected] wrote:
I have built a different kernel using pclinuxos config file.
The segfaults only occur when building perl under LFS-8.4. It does not stop
or invalidate the build and the segfault error messages only go to the
screen and to /var/log/sys.log.
Here is some more info, if some one can decipher:
May 7 17:24:23 lfs kernel: perl[5349]: segfault at 7fff0fb57ff8 ip
000000000048ceb9 sp 00007fff0fb58000 error 6 in perl[407000+230000]
May 7 17:24:23 lfs kernel: perl[5351]: segfault at 7ffc09fc7ff8 ip
000000000048bbbf sp 00007ffc09fc7fe0 error 6 in perl[407000+230000]
May 7 17:24:23 lfs kernel: perl[5357]: segfault at 7fff3f7a0ff8 ip
000000000048bba6 sp 00007fff3f7a1000 error 6 in perl[407000+230000]
May 7 17:24:23 lfs kernel: perl[5359]: segfault at 7ffcbf3d2ff8 ip
00000000005a8bc4 sp 00007ffcbf3d3000 error 6 in perl[407000+230000]
May 7 17:24:23 lfs kernel: perl[5365]: segfault at 7fff61303ff8 ip
00000000005a8bc4 sp 00007fff61304000 error 6 in perl[407000+230000]
May 7 17:24:23 lfs kernel: perl[5375]: segfault at 7ffe31a01ff8 ip
000000000048c722 sp 00007ffe31a02000 error 6 in perl[407000+230000]
May 7 17:24:23 lfs kernel: perl[5377]: segfault at 7ffc7703efe8 ip
000000000048bbc9 sp 00007ffc7703eff0 error 6 in perl[407000+230000]
May 7 17:24:23 lfs kernel: perl[5379]: segfault at 7ffd67853ff8 ip
000000000048ceb6 sp 00007ffd67854000 error 6 in perl[407000+230000]
May 7 17:24:23 lfs kernel: perl[5385]: segfault at 7ffcff30dfa8 ip
000000000048dc3d sp 00007ffcff30dfa0 error 6 in perl[407000+230000]
May 7 17:24:24 lfs kernel: perl[5389]: segfault at 7ffe0db99ff8 ip
00007f8f708f1332 sp 00007ffe0db9a000 error 6 in
libc-2.29.so[7f8f70892000+142000]
I don't use gdb, but I think that using it would, at a minimum, mean
you need to know what command was being run. These segfaults all
happened within about 1 second, but we have no indication of when the
various stages of the build happened, so no idea what this might have
been doing.
Nothing odd in the perl logs ? (e.g. different results for tests in
configure between the two kernels, different actions in the build,
probably best identified by opening two term windows side-by-side on
a system where X is working).
I can not run two term windows side by side. I have never been able to
complete LFS because I could never get it to work sufficently to replace
my current desktop system. I have been building LFS off an on since
LFS-5.0 and maintain a repo. on github.
I will also mention that segfaults are not necessarily a bad thing -
see my experience with rustc, where system crashes (possibly, caused
by an -rc kernel) seemed to be associated with segfaults in the
syslog, and it was *much* later before I discovered that they were
actually intentional (a way to force something to die). Similarly
for some of the tests in perl's testsuite.
But since using a distro kernel apparently removed the problem, I
would be inclined to look at both the .config changes (both the
built-in items, and also whatever was loaded when you finished the
perl build) and also at the kernel versions (and theoretically any
non-upstream patches in the distro, if any). But comparing the
kernels sounds a *lot* of work, with not much likelihood of a
result. Ultimately, this might come down to : Your system ... your
problem :-(
Which is why I have never been able to actually use a LFS system.
Problems that I have had with LFS could not be solved or you could say I
have never gotten much help trying to solve these issues on this mailing
list.
And for unexplained segfaults, memtest86 might also be useful - I
bought my main build machine about 3 years ago, up until Easter it
was running fine and then suddenly I started getting segfaults
building large programs (rustc, qt or qtwebengine) - I had 16GB RAM,
everything small still built fine). Running memtest86 showed a
problem within 1 minute, one of the sticks had a problem. Usually,
memtest86 takes a bit longer than that. In that case, it still ran
after pulling one stick, and passed memtest. It's now running ok
with a new stick.
I would not expect it is my system as it builds without the segfaults
under pclinuxos, it just fails while rebuilding its self. Whne building
under pclinuox the system is under heavy usage runing kde5 while
building LFS. I even have vlc and firefox open as well as 6 other
windows and it uses most of the available memory.
I am going to use/build an older perl to see if it segfaults. If it
does I am probably <https://duckduckgo.com/?q=probably&t=ffab> finished
with LFS at this point. No more time for all the work with little to no
result. I have thought about rebuilding LFS-8.2 which worked but at this
point it is not worth the effort.
--
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
