Resume hpet when back from S3 to S0 Signed-off-by Guanqun Lu <guanqun.lu@intel.com> Signed-off-by Kevin Tian <kevin.tian@intel.com> Thanks, Kevin _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
hpet_setup() is idempotent. We just need to be careful that hpet_broadcast_init() happens after any other caller of hpet_setup(). That happens to be the case currently so I simplified your patch considerably. Changeset 18787. -- Keir On 18/11/08 11:38, "Tian, Kevin" <kevin.tian@intel.com> wrote:> Resume hpet when back from S3 to S0 > > Signed-off-by Guanqun Lu <guanqun.lu@intel.com> > Signed-off-by Kevin Tian <kevin.tian@intel.com> > > Thanks, > Kevin_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
>From: Keir Fraser [mailto:keir.fraser@eu.citrix.com] >Sent: Tuesday, November 18, 2008 11:56 PM > >hpet_setup() is idempotent. We just need to be careful that >hpet_broadcast_init() happens after any other caller of >hpet_setup(). That >happens to be the case currently so I simplified your patch >considerably. >Changeset 18787. >I'm not sure whether it's safe to call hpet_setup twice at boot phase, though it may be ok for S3 resume case. hpet_setup will disable/enable hpet channel, and thus 2nd call may impact system time a bit if hpet is taken as the clocksource (hpet counter is freezed when hpet is disabled) Also regarding to 18786, there's inconsistency between platform timer name and clocksource option, such as "ACPI PM TIMER" and "acpi". Do you intend to use "clocksource=ACPI PM TIMER" as boot option? If you really want to keep such full name, then a speparate short name field can be introduced for comparison. Thanks, Kevin _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 19/11/08 01:25, "Tian, Kevin" <kevin.tian@intel.com> wrote:> I''m not sure whether it''s safe to call hpet_setup twice at boot > phase, though it may be ok for S3 resume case. hpet_setup > will disable/enable hpet channel, and thus 2nd call may impact > system time a bit if hpet is taken as the clocksource (hpet > counter is freezed when hpet is disabled)I''ll have to take a look.> Also regarding to 18786, there''s inconsistency between platform > timer name and clocksource option, such as "ACPI PM TIMER" > and "acpi". Do you intend to use "clocksource=ACPI PM TIMER" > as boot option? If you really want to keep such full name, then > a speparate short name field can be introduced for comparison.Good point. I missed that. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 19/11/08 01:25, "Tian, Kevin" <kevin.tian@intel.com> wrote:> I''m not sure whether it''s safe to call hpet_setup twice at boot > phase, though it may be ok for S3 resume case. hpet_setup > will disable/enable hpet channel, and thus 2nd call may impact > system time a bit if hpet is taken as the clocksource (hpet > counter is freezed when hpet is disabled)Fixed by 18790. I introduced a global S3 counter which subsystems can use to decide whether their hardware needs (re-)initialising. It could be generally useful beyond fixing this HPET issue.> Also regarding to 18786, there''s inconsistency between platform > timer name and clocksource option, such as "ACPI PM TIMER" > and "acpi". Do you intend to use "clocksource=ACPI PM TIMER" > as boot option? If you really want to keep such full name, then > a speparate short name field can be introduced for comparison.Fixed by 18789.> Thanks, > KevinThanks, Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel