Messages in this thread | | | Date | Sun, 9 Jul 2023 17:09:29 +0300 | From | Michal Kubecek <> | Subject | Re: build failure after commit eb0481bbc4ce ("objtool: Fix reloc_hash size") |
| |
On Sat, Jul 08, 2023 at 03:40:50PM +0200, Linux regression tracking (Thorsten Leemhuis) wrote: > On 28.06.23 13:58, Michal Kubecek wrote: > > > > this morning, mainline kernel builds started to fail on my desktop like > > this: > > > > mike@lion:~/work/git/kernel-upstream> mkdir /srv/ram/kobj > > mike@lion:~/work/git/kernel-upstream> make -j24 CC=gcc-13 O=/srv/ram/kobj defconfig >/dev/null > > mike@lion:~/work/git/kernel-upstream> make -j24 CC=gcc-13 O=/srv/ram/kobj all >/dev/null > > net/ipv4/netfilter/iptable_nat.o: warning: objtool: mmap fail reloc > > net/netfilter/xt_mark.o: warning: objtool: mmap fail reloc > > net/netfilter/xt_nat.o: warning: objtool: mmap fail reloc > > net/netfilter/xt_LOG.o: warning: objtool: mmap fail reloc > > net/netfilter/xt_MASQUERADE.o: warning: objtool: mmap fail reloc > > net/netfilter/nf_log_syslog.o: warning: objtool: mmap fail reloc > > net/netfilter/xt_addrtype.o: warning: objtool: mmap fail reloc > > fs/efivarfs/efivarfs.o: warning: objtool: mmap fail reloc > > drivers/thermal/intel/x86_pkg_temp_thermal.o: warning: objtool: mmap fail reloc > > vmlinux.o: warning: objtool: mmap fail reloc > > incomplete ORC unwind tables in file: vmlinux > > Failed to sort kernel tables > > make[2]: *** [/home/mike/work/git/kernel-upstream/scripts/Makefile.vmlinux:35: vmlinux] Error 1 > > make[2]: *** Deleting file 'vmlinux' > > make[1]: *** [/home/mike/work/git/kernel-upstream/Makefile:1256: vmlinux] Error 2 > > make: *** [Makefile:226: __sub-make] Error 2 > > > > I bisected the failure to commit > > > > eb0481bbc4ce ("objtool: Fix reloc_hash size") > > > > I also tried gcc11 and building on a normal filesystem (rather than > > tmpfs) but the result is still the same. Different configurations (e.g. > > allmodconfig) only show more modules with "mmap fail reloc". The machine > > has 64 GB of RAM and both ulimit -m and ulimit -v show "unlimited". > > Anything idea what else I should try? > > Michal, thx for the report. Please correct me if I'm wrong: this as of > now is unfixed, as your patch to fix is afaics wasn't merged yet (at > least I can't see it even in next, but I might be missing something).
Yes, the fix has not been merged yet and current master (commit 1c7873e33645) still shows the issue.
Michal
> Hence adding it to the regression tracking now that the end of the merge > window is near: > > #regzbot ^introduced eb0481bbc4ce > #regzbot title objtool: build failure > #regzbot monitor: > https://lore.kernel.org/all/20230629102051.42E8360467@lion.mk-sys.cz/ > #regzbot fix: objtool: initialize all of struct elf > #regzbot ignore-activity [unhandled content-type:application/pgp-signature] | |