| Message ID | 1533046208-8757-1-git-send-email-leitao@debian.org (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | powerpc/selftests: Avoid backgroud process/threads | expand |
| Context | Check | Description |
|---|---|---|
| snowpatch_ozlabs/apply_patch | success | next/apply_patch Successfully applied |
| snowpatch_ozlabs/checkpatch | warning | Test checkpatch on branch next |
| snowpatch_ozlabs/build-ppc64le | success | Test build-ppc64le on branch next |
| snowpatch_ozlabs/build-ppc64be | success | Test build-ppc64be on branch next |
| snowpatch_ozlabs/build-ppc64e | success | Test build-ppc64e on branch next |
| snowpatch_ozlabs/build-ppc32 | success | Test build-ppc32 on branch next |
Breno Leitao <leitao@debian.org> writes: > Current tm-unavailable test runs for a long period (>120 seconds), and if it is > interrupted, as pressing CRTL-C (SIGINT), the foreground process (harness) dies > but the child process and threads continue to execute (with PPID = 1 now). > > In this case, you'd think the test is gone, but there are two threads being > executed in background, one of the thread ('pong') consumes 100% of the CPU and > the other one ('ping') dumps output message, from time to time, in the STDOUT, > which is annoying. > > This patch simply gets the child process to be SIGTERMed when the parent dies. Hmm, I think we should fix this in the harness if possible. In run_test() it does: /* Kill anything else in the process group that is still running */ kill(-pid, SIGTERM); But that doesn't work if the harness has been killed with Ctrl-C. I think the harness could have a SIGINT handler that basically does the above and then exits? cheers
diff --git a/tools/testing/selftests/powerpc/tm/tm-unavailable.c b/tools/testing/selftests/powerpc/tm/tm-unavailable.c index 156c8e750259..c42f8b60063c 100644 --- a/tools/testing/selftests/powerpc/tm/tm-unavailable.c +++ b/tools/testing/selftests/powerpc/tm/tm-unavailable.c @@ -23,6 +23,8 @@ #include <stdbool.h> #include <pthread.h> #include <sched.h> +#include <signal.h> +#include <sys/prctl.h> #include "tm.h" @@ -342,6 +344,9 @@ int tm_unavailable_test(void) SKIP_IF(!have_htm()); + /* Send me SIGTERM if PPID is dead (as SIGINTed) */ + prctl(PR_SET_PDEATHSIG, SIGTERM); + /* Set only CPU 0 in the mask. Both threads will be bound to CPU 0. */ CPU_ZERO(&cpuset); CPU_SET(0, &cpuset);
Current tm-unavailable test runs for a long period (>120 seconds), and if it is interrupted, as pressing CRTL-C (SIGINT), the foreground process (harness) dies but the child process and threads continue to execute (with PPID = 1 now). In this case, you'd think the test is gone, but there are two threads being executed in background, one of the thread ('pong') consumes 100% of the CPU and the other one ('ping') dumps output message, from time to time, in the STDOUT, which is annoying. This patch simply gets the child process to be SIGTERMed when the parent dies. Signed-off-by: Breno Leitao <leitao@debian.org> Signed-off-by: Gustavo Romero <gromero@linux.vnet.ibm.com> --- tools/testing/selftests/powerpc/tm/tm-unavailable.c | 5 +++++ 1 file changed, 5 insertions(+)