| commit | 4b708e6bd2ba100f804fbe5f47508c92285dcd99 | |
| author | Nigel Cunningham <nigel@suspend2.net> | |
| Tue, 24 Apr 2007 11:01:56 +0000 (24 21:01 +1000) | ||
| committer | Nigel Cunningham <nigel@suspend2.net> | |
| Sun, 29 Apr 2007 11:19:08 +0000 (29 21:19 +1000) | ||
| tree | 00f13fbb4022bcdb1cfbc1aff4e978046a22409c | treesnapshot (tar.gz zip) |
| parent | cd416960d8df6044c3d33806d6c5086c232c8d52 | commitdiff |
Reset cpus allowed on finish anything.
The start/finish_anything modifications uncovered a bug: We're rightly
setting cpus allowed for the thread that starts the suspend/resume to CPU0
when starting, but not resetting it if/when we call finish_anything. On
some computers (why only some?!), this lead to secondary cpus being unused
because the CPU0 mask is apparently duplicated to forked processes. If we
run from process 1, that's bad.
Signed-off-by: Nigel Cunningham < nigel@nigel.suspend2.net>
The start/finish_anything modifications uncovered a bug: We're rightly
setting cpus allowed for the thread that starts the suspend/resume to CPU0
when starting, but not resetting it if/when we call finish_anything. On
some computers (why only some?!), this lead to secondary cpus being unused
because the CPU0 mask is apparently duplicated to forked processes. If we
run from process 1, that's bad.
Signed-off-by: Nigel Cunningham < nigel@nigel.suspend2.net>
| kernel/power/suspend.c | diffblobblamehistory |