ide.2.2.21.05042001-Ole.patch.gz

ide.2.2.21.05042001-Ole.patch.gz

[Krzysztof Oledzki]

Hello,

I have just modified Andre Hedrick's ide.2.2.19.05042001 patch. You can
find it at "http://www.ans.pl/ide/ide.2.2.21.05042001-Ole.patch.gz". It
should patch 2.2.21pre2 kernel without any warnings/errors. I also
bacported new VIA driver from 2.4.x kernel - this one uses correct IO port
as via_base so values displayed in /proc/ide/via file are now properly
calculated. Check the "BM-DMA base" value :)

Hopefully, devices which work in DMA mode, have DMA/UDMA value instead of
PIO in "Transfer Mode" now. It also should works better with some new
chipsets. If I found time I can also try to backport new Intel, Promise
and HPT drivers if no no one else is working on it. I believe I have
required hardware to test it.

I also have one question about IDE support in 2.4.x kernel. Why, with VIA
MVP3 chipset on my FIC 503+ board, with UDMA33/66/100 HDDs using 2.4.x
kernels I have only 13 MB/sec (tested with hdparm -t). With 2.2.x kernels
hdparm shows 17 MB/sec but with IDE support bacported from 2.4.x kernels
(2.2.19/2.2.20+original Andre Hedrick's patch) it is also 13 MB/sec.


Best regards,

                        Krzysztof Oledzki

Re: ide.2.2.21.05042001-Ole.patch.gz

[Vojtech Pavlik]

On Fri, Jan 11, 2002 at 08:58:43PM +0100, Krzysztof Oledzki wrote:
> Hello,
> 
> I have just modified Andre Hedrick's ide.2.2.19.05042001 patch. You can
> find it at "http://www.ans.pl/ide/ide.2.2.21.05042001-Ole.patch.gz". It
> should patch 2.2.21pre2 kernel without any warnings/errors. I also
> bacported new VIA driver from 2.4.x kernel - this one uses correct IO port
> as via_base so values displayed in /proc/ide/via file are now properly
> calculated. Check the "BM-DMA base" value :)
> 
> Hopefully, devices which work in DMA mode, have DMA/UDMA value instead of
> PIO in "Transfer Mode" now. It also should works better with some new
> chipsets. If I found time I can also try to backport new Intel, Promise
> and HPT drivers if no no one else is working on it. I believe I have
> required hardware to test it.
> 
> I also have one question about IDE support in 2.4.x kernel. Why, with VIA
> MVP3 chipset on my FIC 503+ board, with UDMA33/66/100 HDDs using 2.4.x
> kernels I have only 13 MB/sec (tested with hdparm -t). With 2.2.x kernels
> hdparm shows 17 MB/sec but with IDE support bacported from 2.4.x kernels
> (2.2.19/2.2.20+original Andre Hedrick's patch) it is also 13 MB/sec.

If you can send me 'lspci -vvxxx' for both the cases, perhaps I'll be
able to find out why it is so. Is this with all the IDE patches or just
the VIA driver changed?

-- 
Vojtech Pavlik
SuSE Labs

Re: ide.2.2.21.05042001-Ole.patch.gz

[Krzysztof Oledzki]

Hello,

I have just made ide.2.2.21.01152002-Ole patch:

o       backport from ide.2.4.16.12102001.patch: pdc202xx.c - ver 0.30
                - no 48-bit lba - this requires changes in other files:
                   ide-disk.c, ide.c, ide.h, ... Maybe in future...
o       fix missing DEVID_MR_IDE definition in ide-pci.c for VIA_82C576_1
o       add PROMISE_20268R, PROMISE_20269, PROMISE_20275 in ide-pci.c
o       add CONFIG_PDC202XX_FORCE option into Config.in, ide-pci.c

BTW: why the pdc202xx.c file still has "Version 0.30    Mar. 18, 2000"?
It has been changed since that time! :)

After all this changes, pdc202xx.c driver still works for my PDC20265 :)
So, if there is no problem with ide.2.4.16.12102001 my patch should also
works, maybe even for 20268R, 20269 and 20275 chipsets. ;-)

Best regards,

				Krzysztof Oledzki