Hi!
While version 2.1 works OK, v2.3 some more cleanups and enhancements
to the driver. These are:
* Added VIA clone chipsets to the comments at the beginning of the file.
* Simpler VIA southbridge detection using a table only now, two version
specific kludges removed.
* Removed 8-bit timing entries (they were #ifdefed out anyway), because
indeed the hardware should take care of that.
* Fixed a misplaced brace in the FIT macro.
* Enhanced the EIDE PIO & MWDMA timing setup to modify the mode timing
based on drive capabilites instead selecting a slower mode based on
the same numbers. This should give better performance on some drives.
* Use the 'recommended timing' value for MWDMA modes to enhance
performance on MWDMA drives. Using 'minimum timing', as done before
results in flowcontrol being triggered too often, which slows down
the transfers considerably.
* Don't try to set PIO_SLOW mode. Drives that need it don't support PIO
mode setting anyway.
* Check for interface presence before trying to tune it. Higher layers
still can request tuning on a nonexistent interface.
Anyone interested, please test this out, if it is as problemless as
version 2.1, I'll send this to Linus for inclusion in the kernel.
-- Vojtech Pavlik SuSE Labs
This archive was generated by hypermail 2b29 : Fri Sep 15 2000 - 21:00:17 EST