Skip to contentSkip to conversation

RTC drift on Walnut?

DISCUSSION Torben Mathiasen 2 participants 5 messages Jul 7, 2001 – Jul 10, 2001

lore
Share & tools
lore
#0 lore raw

Hi,

I'm seeing a rather large drift of the RTC on a Walnut system. I haven't got the exact numbers, but its something like 20mins in 24 hours. This is with the montavista 2.4.2 kernel.

Has anyone seen this before?

Thanks in advance,

Torben

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

#1 lore raw
Quoted 7 lines · Torben Mathiasen

Hi,

I'm seeing a rather large drift of the RTC on a Walnut system. I haven't got the exact numbers, but its something like 20mins in 24 hours. This is with the montavista 2.4.2 kernel.

Quoted 6 lines

Has anyone seen this before?

Thanks in advance,

Torben

Torben;

I am not sure. I will have to check, I will let you know tomorrow.

-- armin

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

#2 lore raw
Quoted 10 lines · tn Mon, Jul 09 2001, [email protected] wrote:
Quoted 9 lines · Torben Mathiasen

Hi,

I'm seeing a rather large drift of the RTC on a Walnut system. I haven't got the exact numbers, but its something like 20mins in 24 hours. This is with the montavista 2.4.2 kernel.

Has anyone seen this before?

Quoted 8 lines
Quoted 4 lines

Thanks in advance,

Torben

Torben;

I am not sure. I will have to check, I will let you know tomorrow.

Thanks. I'm pretty sure its kernel related, I bet something's wrong in arch/ppc/kernel/time.c. The kernel just reads the RTC at startup, and maintains the system clok via the timer-interrupts. Coming to think of it, I'm not sure if the hardware RTC drifts, or if its just the kernel 'software' clock. Will check on that.

Maybe we're turning ints off for too long somewhere...

Thanks,
Torben

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

#3 lore raw
Quoted 10 lines · Torben Mathiasen

Thanks. I'm pretty sure its kernel related, I bet something's wrong in arch/ppc/kernel/time.c. The kernel just reads the RTC at startup, and maintains the system clok via the timer-interrupts. Coming to think of it, I'm not sure if the hardware RTC drifts, or if its just the kernel 'software' clock. Will check on that.

Maybe we're turning ints off for too long somewhere...

Thanks,
Torben

Torben;

I am running on what will be the next version of 405 kernel from MontaVista. It is using the new todc_time routines. I saw about a 30 second drift.

-- armin

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

#4 lore raw
Quoted 4 lines · Armin Kuster

I am running on what will be the next version of 405 kernel from MontaVista. It is using the new todc_time routines. I saw about a 30 second drift.

Thanks for the update. I assume you guys are continously merging your code with the official PPC tree, so this bug must also be present there.

I think its weird this issue haven't come up on the list before. There seem to be a lot of people running on the walnut system. It would be interesting to hear from people _not_ having this drift.

Regards,
Torben Mathiasen

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/