Ian Collier writes:
> On Sun, 29 Nov 1998 20:56:08 +1100, Richard Gooch said:
> > Could you please extract the CONFIG_APM_RTC_IS_GMT component and
> > include it in a new patch for Linus?
>
> Hmm, that sounds like user-space policy being incorporated into the
> kernel. What if you compile with this option and then decide that the
> RTC should contain local time after all? This should be settable
> via sysctl as was suggested earlier.
I agree that it belongs in a sysctl. That is what I expect is what
will end up happening. Stephan has indicated that's his preferred
solution. However, his current APM patch breaks things for some people
and he feels that it probably won't be ready before 2.3.
In the meantime, the RTC<->GMT calculation is a serious problem for
Dell Inspiron 3200 users: it takes too long to suspend and it looses
several seconds on each suspend/resume cycle. So my patch is a quick
fix for those users until we can get a proper solution. It will avoid
Inspiron 3200 users from having to patch the kernel. We don't want
comments like "Linux 2.2 doesn't work with the Inspiron 3200". People
who have localtime in their RTC will still be stuck, but since there
isn't a safe way to handle this (I considered an offset from GMT
config option, but having to change it twice a year is a pain and will
lead to dumb questions) I don't think supporting that is a good idea.
And in case you're worried that this quick hack will stay in the
kernel forever, let me assure you that once Stephen's new APM code
goes into the kernel, if the hack is still there I will personally
submit a patch to Linus to remove the hack. The hack is only a
stop-gap measure.
Regards,
Richard....
