Seems to be LTO runtime related, with %global _lto_cflags %{nil} the rpm is built.

And for the record, build fails early in the build-aux/autogen.sh call when nspawn isolation is used (I guess on all arches), works OK in chroot.

(In reply to Dan Horák from comment #1)
> Seems to be LTO runtime related, with %global _lto_cflags %{nil} the rpm is
> built.

Confirmed and opt-out applied for s390x only, thank you!

Cc'ing Jeff.

(In reply to Dan Horák from comment #2)
> And for the record, build fails early in the build-aux/autogen.sh call when
> nspawn isolation is used (I guess on all arches), works OK in chroot.

It doesn't in koji anymore. glibc-2.32.9000-4.fc34 fixes this.

By the way, could you update the fedora-gpg-keys packages on your lfedora2 s390x test VM? It's unable to build any rawhide packages in mock:
...
Public key for zstd-1.4.5-5.fc33.s390x.rpm is not installed. Failing package is: zstd-1.4.5-5.fc33.s390x
 GPG Keys are configured as: file:///usr/share/distribution-gpg-keys/fedora/RPM-GPG-KEY-fedora-33-primary, file:///usr/share/distribution-gpg-keys/fedora/RPM-GPG-KEY-fedora-33-primary, file:///usr/share/distribution-gpg-keys/fedora/RPM-GPG-KEY-fedora-32-primary
Error: GPG check FAILED

some details from the stuck test, I wonder if there could be a real deadlock hidden somewhere in the app and LTO exposes it

ps output:
 110995 ?        Ss     0:00 /usr/bin/rpmbuild -bb --target s390x --nodeps /builddir/build/SPECS/lbzip2.spec
 131144 ?        S      0:00 /bin/sh -e /var/tmp/rpm-tmp.F1FJ3s
 131145 ?        S      0:00 /usr/bin/make -O -j8 V=1 VERBOSE=1 check
 131146 ?        S      0:00 /bin/sh -c fail=; \ if (target_option=k; case ${target_option-} in ?) ;; *) echo "am__make_running_with_option: internal error: invalid" "target option '${target
 131170 ?        S      0:00 /bin/sh -c fail=; \ if (target_option=k; case ${target_option-} in ?) ;; *) echo "am__make_running_with_option: internal error: invalid" "target option '${target
 131171 ?        S      0:00 /usr/bin/make check
 131205 ?        S      0:00 /usr/bin/make check-TESTS check-local
 131211 ?        S      0:00 /bin/sh -c set +e; bases='manual-compress.log fuzz-collect.log fuzz-divbwt.log manual-expand.log'; bases=`for i in $bases; do echo $i; done | sed 's/\.log$//'`; 
 131220 ?        S      0:00 /usr/bin/make test-suite.log TEST_LOGS=manual-compress.log fuzz-collect.log fuzz-divbwt.log manual-expand.log
 131221 ?        S      0:00 /bin/sh -c p='manual-compress.test'; \ case 'manual-compress.log' in */*) case 'manual-compress' in */*) b='manual-compress';; *) b=`echo 'manual-compress.log' |
 131237 ?        S      0:00 /bin/sh ../build-aux/tap-driver.sh --test-name manual-compress.test --log-file manual-compress.log --trs-file manual-compress.trs --color-tests no --enable-hard-
 131241 ?        S      0:00 /bin/sh ../build-aux/tap-driver.sh --test-name manual-compress.test --log-file manual-compress.log --trs-file manual-compress.trs --color-tests no --enable-hard-
 131242 ?        S      0:00 gawk -v me=tap-driver.sh -v test_script_name=manual-compress.test -v log_file=manual-compress.log -v trs_file=manual-compress.trs -v expect_failure=0 -v merge=0 
 131243 ?        S      0:00 ./driver compress suite/manual-compress
 131811 ?        S      0:00 sh -c cat >&3
 131814 ?        S      0:00 cat
 131933 ?        Sl     0:00 lbzip2

[sharkcz@rock-kvmlp-fedora lbzip2]$ sudo strace -p 131933
[sudo] password for sharkcz: 
strace: Process 131933 attached
rt_sigsuspend([PIPE XFSZ], 8

^Cstrace: Process 131933 detached
 <detached ...>


after attaching gdb to the process

(gdb) thread apply all bt

Thread 10 (LWP 131944):
#0  0x000003ff8298fc88 in pthread_cond_wait@@GLIBC_2.3.2 () from target:/lib64/libpthread.so.0
#1  0x000002aa28d12c70 in worker_thread_proc () at process.c:452
#2  0x000002aa28d123ee in thread_entry (real_entry=<optimized out>) at process.c:82
#3  0x000003ff82989556 in start_thread () from target:/lib64/libpthread.so.0
#4  0x000003ff828a04f6 in thread_start () from target:/lib64/libc.so.6

Thread 9 (LWP 131943):
#0  0x000003ff8298fc88 in pthread_cond_wait@@GLIBC_2.3.2 () from target:/lib64/libpthread.so.0
#1  0x000002aa28d12c70 in worker_thread_proc () at process.c:452
#2  0x000002aa28d123ee in thread_entry (real_entry=<optimized out>) at process.c:82
#3  0x000003ff82989556 in start_thread () from target:/lib64/libpthread.so.0
#4  0x000003ff828a04f6 in thread_start () from target:/lib64/libc.so.6

Thread 8 (LWP 131942):
#0  0x000003ff8298fc88 in pthread_cond_wait@@GLIBC_2.3.2 () from target:/lib64/libpthread.so.0
#1  0x000002aa28d12c70 in worker_thread_proc () at process.c:452
#2  0x000002aa28d123ee in thread_entry (real_entry=<optimized out>) at process.c:82
#3  0x000003ff82989556 in start_thread () from target:/lib64/libpthread.so.0
#4  0x000003ff828a04f6 in thread_start () from target:/lib64/libc.so.6

Thread 7 (LWP 131941):
#0  0x000003ff8298fc88 in pthread_cond_wait@@GLIBC_2.3.2 () from target:/lib64/libpthread.so.0
#1  0x000002aa28d12c70 in worker_thread_proc () at process.c:452
#2  0x000002aa28d123ee in thread_entry (real_entry=<optimized out>) at process.c:82
#3  0x000003ff82989556 in start_thread () from target:/lib64/libpthread.so.0
#4  0x000003ff828a04f6 in thread_start () from target:/lib64/libc.so.6

Thread 6 (LWP 131940):
#0  0x000003ff8298fc88 in pthread_cond_wait@@GLIBC_2.3.2 () from target:/lib64/libpthread.so.0
#1  0x000002aa28d12c70 in worker_thread_proc () at process.c:452
#2  0x000002aa28d123ee in thread_entry (real_entry=<optimized out>) at process.c:82
#3  0x000003ff82989556 in start_thread () from target:/lib64/libpthread.so.0
#4  0x000003ff828a04f6 in thread_start () from target:/lib64/libc.so.6

Thread 5 (LWP 131939):
#0  0x000003ff8298fc88 in pthread_cond_wait@@GLIBC_2.3.2 () from target:/lib64/libpthread.so.0
#1  0x000002aa28d12c70 in worker_thread_proc () at process.c:452
#2  0x000002aa28d123ee in thread_entry (real_entry=<optimized out>) at process.c:82
#3  0x000003ff82989556 in start_thread () from target:/lib64/libpthread.so.0
#4  0x000003ff828a04f6 in thread_start () from target:/lib64/libc.so.6

Thread 4 (LWP 131938):
#0  0x000003ff8298fc88 in pthread_cond_wait@@GLIBC_2.3.2 () from target:/lib64/libpthread.so.0
#1  0x000002aa28d12c70 in worker_thread_proc () at process.c:452
#2  0x000002aa28d123ee in thread_entry (real_entry=<optimized out>) at process.c:82
#3  0x000003ff82989556 in start_thread () from target:/lib64/libpthread.so.0
#4  0x000003ff828a04f6 in thread_start () from target:/lib64/libc.so.6

Thread 3 (LWP 131936):
#0  0x000003ff8298fc88 in pthread_cond_wait@@GLIBC_2.3.2 () from target:/lib64/libpthread.so.0
#1  0x000002aa28d128ba in sink_thread_proc () at process.c:373
#2  0x000002aa28d123ee in thread_entry (real_entry=<optimized out>) at process.c:82
--Type <RET> for more, q to quit, c to continue without paging--
#3  0x000003ff82989556 in start_thread () from target:/lib64/libpthread.so.0
#4  0x000003ff828a04f6 in thread_start () from target:/lib64/libc.so.6

Thread 2 (LWP 131935):
#0  0x000003ff8298fc88 in pthread_cond_wait@@GLIBC_2.3.2 () from target:/lib64/libpthread.so.0
#1  0x000002aa28d152c4 in worker_thread_proc () at process.c:427
#2  primary_thread () at process.c:529
#3  0x000002aa28d123ee in thread_entry (real_entry=<optimized out>) at process.c:82
#4  0x000003ff82989556 in start_thread () from target:/lib64/libpthread.so.0
#5  0x000003ff828a04f6 in thread_start () from target:/lib64/libc.so.6

Thread 1 (LWP 131933):
#0  0x000003ff827cae7a in sigsuspend () from target:/lib64/libc.so.6
#1  0x000002aa28d03458 in halt () at signals.c:233
#2  schedule (proc=0x2aa28d26a28 <compression>) at process.c:611
#3  work () at process.c:654
#4  main (argc=<optimized out>, argv=<optimized out>) at main.c:934

It could well be the case there's a concurrency issue that is pre-existing and possibly made worse by LTO -- particulary if there's shared state across the threads which is not properly marked.  Has anyone tried the thread sanitizer on this code?

This bug appears to have been reported against 'rawhide' during the Fedora 34 development cycle.
Changing version to 34.