[Linux-ia64] possible bug in the efi real time clock

[Linux-ia64] possible bug in the efi real time clock

[aip]

[-- Attachment #1: Type: text/plain, Size: 148 bytes --]


On the BigSur B1 stepping with BIOS 60 machine, it crashes when running
the hwclock.  Could it be a bug in the efi real time clock?

-Andrew 

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: [Linux-ia64] possible bug in the efi real time clock

[Stephane Eranian]

Andrew,

> 
> 
> On the BigSur B1 stepping with BIOS 60 machine, it crashes when running
> the hwclock.  Could it be a bug in the efi real time clock?
> 
Hard to tell from this description. Could you give more details
like what does appear on the screen ? Could you capture the "tombstone"
(regs dump) and send it to me. I don't have a Bigsur B1 so I can't 
really reproduce the problem.

+--------------------------------------------------------------------+
| Ste'phane ERANIAN                       | Email eranian@hpl.hp.com |
| Hewlett-Packard Laboratories            |                          |
| 1501, Page Mill Road MS 1U-15           |                          |
| Palo  Alto, CA 94303-096                |                          |
| USA                                     |                          |
| Tel : (650) 857-7174                    |                          |
| Fax : (650) 857-5548                    |                          |
+--------------------------------------------------------------------+

Re: [Linux-ia64] possible bug in the efi real time clock

[Uros Prestor]

Stephane Eranian wrote:

> > On the BigSur B1 stepping with BIOS 60 machine, it crashes when running
> > the hwclock.  Could it be a bug in the efi real time clock?
> >
> Hard to tell from this description. Could you give more details
> like what does appear on the screen ? Could you capture the "tombstone"
> (regs dump) and send it to me. I don't have a Bigsur B1 so I can't
> really reproduce the problem.

What happens is a complete system lockup.  There is no kernel register dump,
panic or oops.  The screen just goes blank.  You can lock up the box by
simply running /sbin/hwclock.

I think this symptom was around for a long time.  On pretty much all the
distributions we shipped you could observe a brief video glitch during the
boot when the system was setting hardware clock.  Usually, the screen would
go blank for a split second and then get switched on again.  Why this would
happen while running hwclock is beyond me.  I do recall somebody with a Lion
system observing a hard lockup.  The workaround was to disable running
hwclock on boot.  Nobody here had time to poke around what was going on,
though.

Uros

--
Uros Prestor
uros@turbolinux.com

Re: [Linux-ia64] possible bug in the efi real time clock

[Stephane Eranian]

Uros,

> > > On the BigSur B1 stepping with BIOS 60 machine, it crashes when running
> > > the hwclock.  Could it be a bug in the efi real time clock?
> > >

Our 4-way A3 with Build50 is showing similar behavior. However the system 
is up and running fine with test8.

I looked at the efirtc driver one more time. The ioctl() to read the clock
fails because EFI.GetTime() returns an error. This is new to me. I need to
investigate why it is so. The reason why /proc/efirtc does not fails is
because of a missing test on the return values of the 2 EFI calls to read
the time. You get random results back. I will fix this today. 

I am suspecting something is wrong in EFI with Build50 and up.

> I think this symptom was around for a long time.  On pretty much all the
> distributions we shipped you could observe a brief video glitch during the
> boot when the system was setting hardware clock.  Usually, the screen would
> go blank for a split second and then get switched on again.  Why this would
> happen while running hwclock is beyond me.  I do recall somebody with a Lion
> system observing a hard lockup.  The workaround was to disable running
> hwclock on boot.  Nobody here had time to poke around what was going on,
> though.
> 
I saw this thing too recently on an A2 with Build59 but the system works
fine.

--Stephane

Re: [Linux-ia64] possible bug in the efi real time clock

[H . J . Lu]

On Sat, Sep 09, 2000 at 12:51:03PM -0700, Uros Prestor wrote:
> Stephane Eranian wrote:
> 
> > > On the BigSur B1 stepping with BIOS 60 machine, it crashes when running
> > > the hwclock.  Could it be a bug in the efi real time clock?
> > >
> > Hard to tell from this description. Could you give more details
> > like what does appear on the screen ? Could you capture the "tombstone"
> > (regs dump) and send it to me. I don't have a Bigsur B1 so I can't
> > really reproduce the problem.
> 
> What happens is a complete system lockup.  There is no kernel register dump,
> panic or oops.  The screen just goes blank.  You can lock up the box by
> simply running /sbin/hwclock.
> 

In my case, /sbin/hwclock causes the instant reboot on B1 BigSur and
it crashes so bad that I have to unplug the power cord to get my
machine back. B1 stepping is quite different than A2 stepping on
BigSur, even with the same BIOS, Build 59.


H.J.