[Pm-utils] [patch] userspace-suspend, revisited
Stefan Seyfried
seife at suse.de
Tue Mar 27 00:55:18 PDT 2007
On Mon, Mar 26, 2007 at 09:03:59PM +0200, Tim Dijkstra wrote:
> On Mon, 26 Mar 2007 20:16:41 +0200
> Stefan Seyfried <seife at suse.de> wrote:
>
> > do_suspend_hybrid()
> > {
> > - return 1
> > + local RET=1
> > + if [ -z "$HIBERNATE_METHOD" ]; then
> > + if [ -x /usr/sbin/s2both -a -c /dev/snapshot ]; then
> > + HIBERNATE_METHOD="userspace"
> > + else
> > + HIBERNATE_METHOD="kernel"
> > + fi
> > + fi
>
> This test doens't seem necessary to me, There is no `kernel' method.
Yes, i thought about this, too. But once somebody works on suspend2 support,
there will be a third method for suspend to disk and a second for suspend to
both.
> Also s2both will return an error when configured wrongly so these
> are in a sense done twice...
>
> > + case $HIBERNATE_METHOD in
> > + userspace)
> > + set -x
> > + /usr/sbin/s2both -f $S2DISK_CONF
> > + RET=$?
> > + set +x
> > + ;;
> > + *)
> > + RET=1
> > + ;;
> > + esac
> > + return $RET
> > }
>
>
> BTW, in pm-action REVERSE is set as "resume" or "thaw" depending on the
> method. With `hybrid' we can get back from both s2ram or s2disk, did you
> think about what to choose for REVERSE from hybrid?
Somebody will come up with a name, i won't go into naming details. You know,
if it was me, we would not have "suspend" and "hibernate" but technically
correct terms... ;-) I really no longer care about the names, i'll just use
what's implemented.
The "set -x" are indeed for more verbose logging, we could also just put a
line with
echo "running /usr/sbin/s2both -f '$S2DISK_CONF'"
in there. The set -x is more a quick hack, since i am right now debugging a
report where apparently some config variables are not set correctly (part of
it due to the non-exporting of the variables at the start of functions, as
fixed by my trivial patch yesterday).
We can argue about logging later IMO :-)
--
Stefan Seyfried
"Any ideas, John?"
"Well, surrounding them's out."
More information about the Pm-utils
mailing list