From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.9) Gecko/20020430 Description of problem: if you attach to a running process, and terminate strace with ^C, the running process will remain in a stopped state. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: in window 1: sleep 20 in window 2: strace -p $(pgrep sleep) press ^C wait 20 seconds. the sleep is still running. in window 2: kill -CONT $(pgrep sleep) the sleep terminates immediately. Expected Results: strace should send a SIGCONT or whatever after detaching (I don't know how it works in detail :) Additional info:
*** Bug 66191 has been marked as a duplicate of this bug. ***
*** Bug 71166 has been marked as a duplicate of this bug. ***
Rhis should be better with strace-4.4-8.
*** Bug 73298 has been marked as a duplicate of this bug. ***
Fix confirmed with strace-4.4-8.
strace-4.4-8 fails for me eventually if I keep repeating the test below. This is happening on a RedHat 7.3 box, kernel-smp-2.4.18-18.7.x, dual Intel Pentium 3. Process 1153 is xinetd. # grep ^State /proc/1153/status State: S (sleeping) # ./BUILD/strace-4.4/strace -v -p 1153 select(7, [3 5 6], NULL, NULL, NULL <unfinished ...> # grep ^State /proc/1153/status State: S (sleeping) # ./BUILD/strace-4.4/strace -v -p 1153 select(7, [3 5 6], NULL, NULL, NULL <unfinished ...> # grep ^State /proc/1153/status State: T (stopped) # ./BUILD/strace-4.4/strace -v -p 1153 --- SIGSTOP (Stopped (signal)) --- select(7, [3 5 6], NULL, NULL, NULL <unfinished ...> # grep ^State /proc/1153/status State: S (sleeping) strace-4.4-9.1 worked for me, but only after I applied to detach() the changes from patch21, strace-x86_64.patch. It seems that those changes are not quite architecture specific, but they fix the problem on i386 too.