Hi,
seems to me as if f28 has become unusable on non-pae systems.
At least, an attempt to upgrade my PIII from f27 (functional!) to f28 ended up with numerous segfaults, originating from cron, mount.nfs, automount and NetworkManager (and likely more) all similar to this:
...: segfault at bf44bfc4 at b79e35ac sp bf44bfc0 error 6 in libc-2.27.so[b787c000+19000]
I currently don't have any time to investigate any further, but my gut feeling is RHBZ#1482798 might have returned.
Ralf
Interesting. What makes me wonder is how you're getting SIGSEGV, not SIGILL. In my experience, if the binary contains code that your processor doesn't understand, Linux raises SIGILL, not SIGSEGV, as well it should, because they're very different types of problem.
Can you run a test for me? Make a note of the stuff you see crashing, and then use ldd on them to see what dynamic libraries they share. If it's just glibc, well, then it might be a problem with glibc, or maybe, god forbid, either part of or all of the entire Fedora 28 i686 package set is defective and needs to be rebuilt.
This is concerning. I have some old PIII systems I use myself, but I haven't upgraded them to Fedora 28 yet. Hopefully the problem isn't too massive.
-Sub
On 2018-05-08 12:48 am, Ralf Corsepius wrote:
Hi,
seems to me as if f28 has become unusable on non-pae systems.
At least, an attempt to upgrade my PIII from f27 (functional!) to f28 ended up with numerous segfaults, originating from cron, mount.nfs, automount and NetworkManager (and likely more) all similar to this:
...: segfault at bf44bfc4 at b79e35ac sp bf44bfc0 error 6 in libc-2.27.so[b787c000+19000]
I currently don't have any time to investigate any further, but my gut feeling is RHBZ#1482798 might have returned.
Ralf _______________________________________________ X86 mailing list -- x86@lists.fedoraproject.org To unsubscribe send an email to x86-leave@lists.fedoraproject.org
On 05/08/2018 09:48 AM, Ralf Corsepius wrote:
At least, an attempt to upgrade my PIII from f27 (functional!) to f28 ended up with numerous segfaults, originating from cron, mount.nfs, automount and NetworkManager (and likely more) all similar to this:
...: segfault at bf44bfc4 at b79e35ac sp bf44bfc0 error 6 in libc-2.27.so[b787c000+19000]
The fault appears to be at a page boundary. That suggests it's a kernel memory management issue, not something related to userspace instruction selection (e.g., SSE2 instructions, or long NOPs.)
Thanks, Florian
On 2018-05-08 3:53 am, Florian Weimer wrote:
On 05/08/2018 09:48 AM, Ralf Corsepius wrote:
At least, an attempt to upgrade my PIII from f27 (functional!) to f28 ended up with numerous segfaults, originating from cron, mount.nfs, automount and NetworkManager (and likely more) all similar to this:
...: segfault at bf44bfc4 at b79e35ac sp bf44bfc0 error 6 in libc-2.27.so[b787c000+19000]
The fault appears to be at a page boundary. That suggests it's a kernel memory management issue, not something related to userspace instruction selection (e.g., SSE2 instructions, or long NOPs.)
Thanks, Florian _______________________________________________ X86 mailing list -- x86@lists.fedoraproject.org To unsubscribe send an email to x86-leave@lists.fedoraproject.org
Has this issue developed any further? Did it fix itself or is it still a current problem?
On 05/31/2018 06:47 PM, Subsentient wrote:
On 2018-05-08 3:53 am, Florian Weimer wrote:
On 05/08/2018 09:48 AM, Ralf Corsepius wrote:
At least, an attempt to upgrade my PIII from f27 (functional!) to f28 ended up with numerous segfaults, originating from cron, mount.nfs, automount and NetworkManager (and likely more) all similar to this:
...: segfault at bf44bfc4 at b79e35ac sp bf44bfc0 error 6 in libc-2.27.so[b787c000+19000]
The fault appears to be at a page boundary. That suggests it's a kernel memory management issue, not something related to userspace instruction selection (e.g., SSE2 instructions, or long NOPs.)
Has this issue developed any further? Did it fix itself or is it still a current problem?
For me, these issues persist. Despite several further attempts, no further insights nor progress.
Ralf
On 2018-05-31 10:12 am, Ralf Corsepius wrote:
On 05/31/2018 06:47 PM, Subsentient wrote: On 2018-05-08 3:53 am, Florian Weimer wrote:
On 05/08/2018 09:48 AM, Ralf Corsepius wrote: At least, an attempt to upgrade my PIII from f27 (functional!) to f28 ended up with numerous segfaults, originating from cron, mount.nfs, automount and NetworkManager (and likely more) all similar to this:
...: segfault at bf44bfc4 at b79e35ac sp bf44bfc0 error 6 in libc-2.27.so[b787c000+19000] The fault appears to be at a page boundary. That suggests it's a kernel memory management issue, not something related to userspace instruction selection (e.g., SSE2 instructions, or long NOPs.)
Has this issue developed any further? Did it fix itself or is it still a current problem?
For me, these issues persist. Despite several further attempts, no further insights nor progress.
Ralf
_______________________________________________ X86 mailing list -- x86@lists.fedoraproject.org To unsubscribe send an email to x86-leave@lists.fedoraproject.org Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/x86@lists.fedoraproject.org/me...
Have you attempted to use Fedora 28 on a different system of this processor class? And if not, have you attempted the latest kernels?
On 05/31/2018 08:52 PM, Subsentient wrote:
Have you attempted to use Fedora 28 on a different system of this processor class?
Yes, I also tried on an Atom N270, but ... updates there fail for seemingly different causes (SELinux, screwed up uids/gids, yp/nis, dnf bugs, broken packages,...)
And if not, have you attempted the latest kernels?
Also tried this. Things slightly improved on the N270, but not essentially and did not improve at all on the PIII.
The ugliest issue on the PIII is that these crash not generate coredumps at bootup. Seems to me as if something (jourmald, systemd?) is crashing such kind of early, no coredumps are being generated.
Should I find sufficient time (not likely to happen any time soon), I'll likely try a fresh install on the N270.
Ralf
On 2018-05-31 11:12 pm, Ralf Corsepius wrote:
On 05/31/2018 08:52 PM, Subsentient wrote:
Have you attempted to use Fedora 28 on a different system of this processor class?
Yes, I also tried on an Atom N270, but ... updates there fail for seemingly different causes (SELinux, screwed up uids/gids, yp/nis, dnf bugs, broken packages,...)
And if not, have you attempted the latest kernels?
Also tried this. Things slightly improved on the N270, but not essentially and did not improve at all on the PIII.
The ugliest issue on the PIII is that these crash not generate coredumps at bootup. Seems to me as if something (jourmald, systemd?) is crashing such kind of early, no coredumps are being generated.
Should I find sufficient time (not likely to happen any time soon), I'll likely try a fresh install on the N270.
Ralf _______________________________________________ X86 mailing list -- x86@lists.fedoraproject.org To unsubscribe send an email to x86-leave@lists.fedoraproject.org Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/x86@lists.fedoraproject.org/me...
Damn. Interestingly enough I have fc28 running perfectly on an N270 Acer Aspire One ZG5 with SELinux in enforcing. I'll fire up one of my old pentium 3s and see what happens. Might have to wait a day till I have the time to create proper boot media though.
Hi,
On 2018-06-01 06:22 AM, Subsentient wrote:
Damn. Interestingly enough I have fc28 running perfectly on an N270 Acer Aspire One ZG5 with SELinux in enforcing. I'll fire up one of my old pentium 3s and see what happens. Might have to wait a day till I have the time to create proper boot media though.
I have an Asus EeePC with the same CPU upgraded to Fedora 28 recently. One issue I had was that PAE kernels are no longer built, so I had to manually install a regular kernel. The second is: https://bugzilla.redhat.com/show_bug.cgi?id=1498673
Regards, Dominik