diff options
| author |   deraadt <deraadt@openbsd.org> | 2010-09-09 04:13:15 +0000 |
|---|---|---|
| committer | deraadt <deraadt@openbsd.org> | 2010-09-09 04:13:15 +0000 |
| commit | 2422124b87b60708604a53f82574bd716b0a8f86 (patch) | |
| tree | a13af6f6e4fcdd935349e3188af03e46d4947c2d /usr.bin/ssh/ssh-ecdsa.c | |
| parent | when I fixed the ruleset matching on rules that had more than one condition (diff) | |
| download | wireguard-openbsd-2422124b87b60708604a53f82574bd716b0a8f86.tar.xz wireguard-openbsd-2422124b87b60708604a53f82574bd716b0a8f86.zip | |
The resume powerhooks were always in the wrong place. Then when replaced
by the device-tree activate walk it even worse since we were doing so much
more work. apm_suspend() could return into the event handler and thus
tsleep() -- all this while we are supposedly trying to maintain splhigh
because our DVACT_SUSPEND mechanism has temporarily totally wrecked the
machine.
Move the hardware restoration code (DVACT_RESUME + other goo) to the same
function, and recognize that apm_resume() is just a boring event handler
which punts the event to userland.
ok oga
Diffstat (limited to 'usr.bin/ssh/ssh-ecdsa.c')
0 files changed, 0 insertions, 0 deletions