Bug 1467833

Reason - Java segfaults when ran in mock chroot, even with no arguments. Only in mock, does not happen on the same host outsides of chroot.

<mock-chroot> sh-4.4# java -version
<mock-chroot> sh-4.4# echo $?
139

(139-128=11 - signal number of SIGSEGV)

<mock-chroot> sh-4.4# rpm -q java-1.8.0-openjdk-headless
java-1.8.0-openjdk-headless-1.8.0.131-7.b12.fc27.ppc64le

The same version works without chroot, same host:

[root@mizdebsk-rhbz1467833 ~]# java -version
openjdk version "1.8.0_131"
OpenJDK Runtime Environment (build 1.8.0_131-b12)
OpenJDK 64-Bit Server VM (build 25.131-b12, mixed mode)
[root@mizdebsk-rhbz1467833 ~]# rpm -q java-1.8.0-openjdk-headless
java-1.8.0-openjdk-headless-1.8.0.131-7.b12.fc27.ppc64le

This is strange as a build of eclipse one day earlier passed fine. Same java version: ppc64le 1:1.8.0.131-7.b12.fc27

See: https://koji.fedoraproject.org/koji/buildinfo?buildID=914908

(In reply to Severin Gehwolf from comment #2)
> This is strange as a build of eclipse one day earlier passed fine. Same java
> version: ppc64le 1:1.8.0.131-7.b12.fc27
> 
> See: https://koji.fedoraproject.org/koji/buildinfo?buildID=914908

Do we use Zero or the JIT on ppc64le?

(In reply to Deepak Bhole from comment #3)
> (In reply to Severin Gehwolf from comment #2)
> > This is strange as a build of eclipse one day earlier passed fine. Same java
> > version: ppc64le 1:1.8.0.131-7.b12.fc27
> > 
> > See: https://koji.fedoraproject.org/koji/buildinfo?buildID=914908
> 
> Do we use Zero or the JIT on ppc64le?

JIT, AFAIK.

Comparing the good vs bad eclipse builds root log I see these differences:

Good:
kernel-headers.ppc64le 4.12.0-0.rc7.git2.1.fc27
glibc.ppc64le 2.25.90-15.fc27

Bad:
kernel-headers-4.12.0-1.fc27.ppc64le
glibc-2.25.90-18.fc27.ppc64le

glibc got some ppc64le changes:
https://koji.fedoraproject.org/koji/buildinfo?buildID=915808

I wonder if downgrading to 2.25.90-15 fixes the issue. Do you still have the chroot available for you to try?

Yes, glibc downgrade from glibc-2.25.90-19.fc27.ppc64le to glibc-2.25.90-15.fc27.ppc64le fixes the issue.

(In reply to Mikolaj Izdebski from comment #6)
> Yes, glibc downgrade from glibc-2.25.90-19.fc27.ppc64le to
> glibc-2.25.90-15.fc27.ppc64le fixes the issue.

Thanks! I'll reassign to glibc as per comment 2 and comment 6.

Additional info:
This started to happen after update from glibc-2.25.90-15.fc27 to glibc-2.25.90-17.fc27.
Reproducible with latest koji build too (glibc-2.25.90-20.fc27).
Breaks other system apps too (including dnf, making upgrade/downgrade impossible once affected glibc version is installed)

(In reply to Mikolaj Izdebski from comment #1)
> Reason - Java segfaults when ran in mock chroot, even with no arguments.
> Only in mock, does not happen on the same host outsides of chroot.

Are the versions of glibc inside and outside the chroot identical?
Does the chroot have mounted /proc and /sys file systems?  Thanks.

Never mind, the chroot setup problems are a red herring.

Backtrace looks like this:

#0  0x0000000000000000 in ?? ()
#1  0x00003fffb6cb2380 in ?? ()
#2  0x00003fffb6cb2838 in ?? ()
#3  0x00003fffb7fba73c in resolve_ifunc (sym_map=<optimized out>, map=<optimized out>, value=70367515977760)
    at ../sysdeps/powerpc/powerpc64/dl-machine.h:674
#4  elf_machine_rela (skip_ifunc=<optimized out>, reloc_addr_arg=0x3fffb6d40098, version=<optimized out>, 
    sym=<optimized out>, reloc=0x3fffb6bf8c48, map=0x20030bb0) at ../sysdeps/powerpc/powerpc64/dl-machine.h:729
#5  elf_dynamic_do_Rela (skip_ifunc=<optimized out>, lazy=<optimized out>, nrelative=<optimized out>, 
    relsize=<optimized out>, reladdr=<optimized out>, map=<optimized out>) at do-rel.h:137
#6  _dl_relocate_object (scope=0x20030f28, reloc_mode=<optimized out>, consider_profiling=<optimized out>)
    at dl-reloc.c:259
#7  0x0000003c00000008 in ?? ()

*** This bug has been marked as a duplicate of bug 1467518 ***