diff options
| author | Thomas Schwinge <tschwinge@gnu.org> | 2012-11-29 01:33:22 +0100 |
|---|---|---|
| committer | Thomas Schwinge <tschwinge@gnu.org> | 2012-11-29 01:33:22 +0100 |
| commit | 5bd36fdff16871eb7d06fc26cac07e7f2703432b (patch) | |
| tree | b430970a01dfc56b8d41979552999984be5c6dfd /open_issues/fork_deadlock.mdwn | |
| parent | 2603401fa1f899a8ff60ec6a134d5bd511073a9d (diff) | |
| download | web-5bd36fdff16871eb7d06fc26cac07e7f2703432b.tar.gz web-5bd36fdff16871eb7d06fc26cac07e7f2703432b.tar.bz2 web-5bd36fdff16871eb7d06fc26cac07e7f2703432b.zip | |
IRC.
Diffstat (limited to 'open_issues/fork_deadlock.mdwn')
| -rw-r--r-- | open_issues/fork_deadlock.mdwn | 31 |
1 files changed, 31 insertions, 0 deletions
diff --git a/open_issues/fork_deadlock.mdwn b/open_issues/fork_deadlock.mdwn index 6b90aa0a..c1fa9208 100644 --- a/open_issues/fork_deadlock.mdwn +++ b/open_issues/fork_deadlock.mdwn @@ -63,3 +63,34 @@ Another one in `dash`: stopped = 1 i = 6 [...] + + +# IRC, OFTC, #debian-hurd, 2012-11-24 + + <youpi> the lockups are about a SIGCHLD which gets lost + <pinotree> ah, ok + <youpi> which makes bash spin + <pinotree> is that happening more often recently, or it's just something i + just noticed? + <youpi> it's more often recently + <youpi> where "recently" means "some months ago" + <youpi> I didn't notice exactly when + <pinotree> i see + <youpi> it's at most since june, apparently + <youpi> (libtool managed to build without a fuss, while now it's a pain) + <youpi> (libtool building is a good test, it seems to be triggering quite + reliably) + + +## IRC, freenode, #hurd, 2012-11-27 + + <youpi> we also have the shell wait issue + <youpi> it's particularly bad on libtool calls + <youpi> the libtool package (with testsuite) is a good reproducer :) + <youpi> the symptom is shell scripts eating CPU + <youpi> busy-waiting for a SIGCHLD which never gets received + <braunr> that could be what i got + <braunr> + http://www.gnu.org/software/hurd/microkernel/mach/gnumach/memory_management.html + <braunr> last part + <youpi> perhaps watch has the same issue as the shell, yes |