[PATCH] 2.0.35: updated Jumbo -7 patch

Andrew Derrick Balsa (andrebalsa@altern.org)
Sat, 25 Jul 1998 02:25:35 +0200


Rev. 7 of the Jumbo patch against 2.0.35 is now available from:
(approx. 38KB file size)

The patch also applies OK against 2.0.36pre1, if you got that one already.

Changes from rev -6:
- I have removed the vt.c beeeep buglet patch, since this one is now included in
2.0.36pre1. Alan's code is - as usual - much cleaner, so please use his patch.
Thanks to Julie Brandon for pointing the problems in the previous fix.

- Slightly improved documentation for UDMA. Please take a look at
/Documentation/udma.txt. Just as a reminder, the UDMA code in this patch
supports a number of chipsets not previously available:
- SiS 5598 and similar SiS controllers.
- VIA Apollo (wanted: MVP3 testers)
- Promise PCI controllers
- More... (RTFM, specially Brion Vibber's UDMA mini-HOWTO, available
from the LDP and mirrors)

- Slightly improved layout of code in triton.c.

- Now includes > 64MB automatic detection, and a small keyboard buglet fix.
This code was contributed by Andre M. Hedrick, who is also the main drive
behind the UDMA code, together with Brion Vibber. Original code by Larry M.

Changes from a stock 2.0.35:
- Fixes Cyrix 6x86 and 6x86L CPU detection.
- Fixes bogomips calibration. (courtesy Richard Johnson).
- Improves Intel CPU detection (partial port of code from 2.1.x).
- Improves AMD CPU detection.
- Has extended capability flags detection for AMD CPUs, code by Rob Dale
(e.g. reports a new "3dnow" flag).
- Inaugurates Centaur C6 detection.
- Improves /arch/i386/kernel/time.c code on CPUs with a Time Stamp Counter (fast
code should work with APM now, also doesn't break with Centaur C6 step 0,
should be more accurate too).
- Reports MHz clock rates on CPUs with a TSC in /proc/cpuinfo.
- Calibrates TSC at boot time (drift and jitter free).
- Adds (U)DMA capabilities on non-Intel chipsets/PCI controller cards.
- Allows control of order of PCI EIDE controller/mainboard EIDE controller.
- Adds > 64MB memory detection.
- Avoids odd (rare) keyboard problems at boot time.

As usual, success/failures reports wanted and welcome. :-)

Andrew D. Balsa

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html