http://bugzilla.opensuse.org/show_bug.cgi?id=905863 --- Comment #11 from Julian Wolf <juwolf@suse.com> --- Thank you for that. The problem in this case appears when extracting the early-microcode in the list_files() function: cat -- testinitrd | cpio --extract --verbose --quiet --list I assume cpio exits after reading the early-microcode (as the rest of the file is a xz archive) which then causes SIGPIPE to be delivered to cat. That's also exactly what I see when checking the exit-code of cat after running above command:
cat -- testinitrd | cpio --extract --verbose --quiet --list [...] echo ${PIPESTATUS[0]} ${PIPESTATUS[1]} 141 0
Exit code 141 (>128) is, afaik, pointing to 128 + Signal-Number, which gives in this case "13 SIGPIPE". Same exit code when running cat and killing it with SIGPIPE manually. So, in my opinion, SIGPIPE is absolutely okay here and the point is more about why the shell (bash?) does report that for you (when running it on a console). -- You are receiving this mail because: You are on the CC list for the bug.