[lkml]   [2003]   [Feb]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: High-Res-Timers: Unexpected "lock" during "Calibrating delay loop "
Isabelle, Francois wrote:
> Just wanted to add this: this seems to be reproducible on all 486

I hope to get a chance to look at this today. Thus far, I know that
the PIT version does not keep good time in the face of timers of
resolution > 1/HZ, i.e. the "_HR" clocks.

This has to do with the fact that we have only one interrupt source
and must change it to get the sub jiffies interrupts. In the process
we loose track of where the jiffies interrupt should be. There is an
attempt to use the second PIT timer to cover for this, but assumptions
are made here that may not be right. (This timer is used for memory
refresh and is set up by the BIOS. We can not change it and we assume
that it has a given period. I suppose we could attempt to measure the
period, but thus far we have assumed that these sorts of machines
would not require high res timers and have not tried to run down these

A side issue is that the timers seem to not follow the spec (at least
on my test machine) in the regard that a status bit is supposed to
indicate when the latch has been loaded after a program has been
output. I can easily find cases where a wait for the bit to change is
followed by a read of the latch which returns a stale value. I
suppose some of this has to do with a large difference between the cpu
clock and the timer clock.
>>-----Original Message-----
>>From: Isabelle, Francois []
>>Sent: 4 février, 2003 09:04
>>Subject: High-Res-Timers: Unexpected "lock" during "Calibrating delay
>>loop "
>> we are having an unexpected problem with the HR patch(
>>regardless of
>>the patch you sent which compiled just fine ). The board uses
>>a 486DX cpu,
>>so there is no support for TSC nor ACPI, the only thing we
>>have is the PIT.
>>Without highres, the kernel boots properly, with highres
>>enabled, the kernel
>>passes "time_init()" put it hangs in "calibrate_loop() , ( I
>>though it hung
>>for real, but it get passed the loop after a while) "
>>Seems like the tick is VERY SLOW..
>>The PIT has been tested on this board, and without HR, the
>>kernel boots fine
>>... if you have any hints, they would be welcome.
>>The keyboard detection routine timeouts so the system is
>>quiet unuseable and
>>I can't get the calibration results yet.
>>This SF.NET email is sponsored by:
>>SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
>>to unsubscribe:
>>High-res-timers-discourse mailing list

George Anzinger
Preemption patch:

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:32    [W:0.040 / U:1.320 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site