Re: New Small Cyrix patch

Linux Developer (
Tue, 18 Nov 1997 16:23:59 -0500 (EST)

On Tue, 18 Nov 1997, Andre Derrick Balsa wrote:

> Hi Phillip,
> You seem to be making a small confusion here.
> note the "L" suffix
> The PR200L is always a Rev 4.2 chip (which the Mike Jagdis patch will
> report as step 2 rev 2). Both Rev 2.5 and 2.7 are plain, classic,
> single-voltage 6x86 chips.

Not true. Cyrix, at one point, decided to keep the 6x86 and 6x86L series
at the same step/rev levels, as much as possible. There really isn't much
modification needed.

> If your chip is marked as a rev 2.7 chip or earlier, then it's a single
> voltage 6x86, _not_ the dual voltage 6x86L (which BTW has the cx8 flag
> reported in /proc/cpuinfo).

cx8 is reported in /proc/cpuinfo. Also, another strange discovery. After I
hacked up /usr/src/linux/include/linux/pci.h and
/usr/src/linux/drivers/pci/pci.c, my bogos increased by 10. Up from ~119
to ~130. And all I did was modify it so that it would pass the UMC's
correctly. They were reporting as VIA's.

> 1) Install my patch and cat /proc/cpuinfo, or
> 2) get6x86 -p 0xff. This will report the DIR1 value. Compare to my table
> of 6x86 revisions, DIR 1 column and determine what your _really_ chip
> is, or
> 3) use the previous Mike Jagdis patch. However, from your previous post,
> you either a)patched incorrectly or b)got a buggy version of Mike's
> patch or c) incorrectly read the markings on your 6x86 CPU.

I've worked with these CPUs a long time. Trust me, it is a Step2, Rev7.
I've run every utility in the book, I've patched and repatched, I've used
every patch out there (no wonder lilo.conf is nearly 100k!;), and
everything that reports correctly reports the same; Step 2, Revision 7.

> When you will have (eventually) installed my new small patch, please
> report your findings to the linux-kernel list on this same thread.
> I am very curious to see how the patch does with the unidentifiable
> motherboards you have.

Actually, these motherboards are VERY identifiable. At least ONE is. The
one that reports clocking sans multiplier, to within 20bogos, is an
Amptron PM-8600A, using the VXtwo+ chipset. The other is a slightly newer,
"unlabeled" PM-8600, probably revision B, based on the VXPro chipset. The
only difference between the two is the chipset, and the USB support, or
lack thereof. Otherwise, these boards are exactly identical. I've yet to
be able to get in touch with either Amptron or UMC to check up on much,
but I'm going to try to get in touch today from the office.

-Phillip R. Jaenke [InterNIC Handle: PRJ5] (
MIS Department, PC Importers, Inc. 800.319.9284, x4262
Head of Development, The Improvement Linux Project.
Penguin0: Cyrix Cx6x86 PR200L+, 64M, 1.6G, 1.2G, 4.3G, EE/Pro, 2.0.30
Penguin1: Intel P5-100, 32M, 4G, NE2k, 2.1.60-Improvement
The Diamond FireGL Series - When you're ready for a REAL video card.