Hi Jackal1983,

What you are facing here is the infamous UDMA mode bug. This Award BIOS bug was fixed in March 1999 and later BIOS versions.
Unfortunately, the latest ECS P6BX-MS BIOS is Ver 1.1d from 02/08/1999. I assume you are running this BIOS?

This UDMA bug causes the BIOS to report a higher UDMA mode than the chipset can handle, when using an ATA66 or faster drive. Windows then rejects any UMDA mode and reverts to PIO mode 4 to protect your data on the HDD/SSD.

I have a fix for this UDMA mode bug. Give me a few days to patch the 02/08/1999 BIOS with this fix. While I’m at it, I will fix the 32GB HDD limit bug in this BIOS as well. 😀

Cheers, Jan

Chkcpu wrote on Yesterday, 11:40:

Hi Jackal1983, […]
Show full quote

Hi Jackal1983,

What you are facing here is the infamous UDMA mode bug. This Award BIOS bug was fixed in March 1999 and later BIOS versions.
Unfortunately, the latest ECS P6BX-MS BIOS is Ver 1.1d from 02/08/1999. I assume you are running this BIOS?

This UDMA bug causes the BIOS to report a higher UDMA mode than the chipset can handle, when using an ATA66 or faster drive. Windows then rejects any UMDA mode and reverts to PIO mode 4 to protect your data on the HDD/SSD.

I have a fix for this UDMA mode bug. Give me a few days to patch the 02/08/1999 BIOS with this fix. While I’m at it, I will fix the 32GB HDD limit bug in this BIOS as well. 😀

Cheers, Jan

Yep, I'm on v 1.1d. And YES, please send it!!!

By the by, could you add in Coppermine and Tualatin microcode? I'd like to get this board working with C3 Ezra and Nehemiah CPUs and I figure those microcodes might help. Do you have any ideas along those lines?

Jackal1983 wrote on Today, 04:28:

Chkcpu wrote on Yesterday, 11:40:

Hi Jackal1983, […]
Show full quote

Hi Jackal1983,

What you are facing here is the infamous UDMA mode bug. This Award BIOS bug was fixed in March 1999 and later BIOS versions.
Unfortunately, the latest ECS P6BX-MS BIOS is Ver 1.1d from 02/08/1999. I assume you are running this BIOS?

This UDMA bug causes the BIOS to report a higher UDMA mode than the chipset can handle, when using an ATA66 or faster drive. Windows then rejects any UMDA mode and reverts to PIO mode 4 to protect your data on the HDD/SSD.

I have a fix for this UDMA mode bug. Give me a few days to patch the 02/08/1999 BIOS with this fix. While I’m at it, I will fix the 32GB HDD limit bug in this BIOS as well. 😀

Cheers, Jan

Yep, I'm on v 1.1d. And YES, please send it!!!

Here is the patched P6BX-MS Rev 1.1d BIOS.
Both the UMDA Mode bug and 32GB IDE HDD limit bug are fixed.
In the zip, I’ve included a PATCH.TXT file that lists the changes I’ve made.

When ATA/66/100/133 drives are fitted, the BIOS will now limit the reported UDMA Mode to UDMA 2, the maximum supported UDMA Mode on this 440BX chipset. This change is also visible on the BIOS summary screen.

The 32GB limit bugfix consists of one 32GB and two 64GB limit fixes. This patch J.1 BIOS now supports IDE drives up to 128GiB/137GB.
More info about these fixes can be found on my website at http://www.steunebrink.info/k6plus.htm#32GB_bugfix
The “new community fix” described there is the fix I used for your BIOS.

Jackal1983 wrote on Today, 08:00:

By the by, could you add in Coppermine and Tualatin microcode? I'd like to get this board working with C3 Ezra and Nehemiah CPUs and I figure those microcodes might help. Do you have any ideas along those lines?

I’ve looked at the CPU support in the Rev 1.1d BIOS, but I see only Intel CPUs are supported. VIA C3 detection and support code is nowhere to be found. Not surprising for an early 1999 Slot 1 BIOS. 😉

I’m unable to add VIA C3 support. I do mainly Socket 3/5/7 BIOS patches, and my modding repertoire for Slot 1/Socket 370 is very limited.

Do let us know how this P6BX-MS patch J.1 BIOS works.
Cheers, Jan

Chkcpu wrote on Today, 13:46:

Here is the patched P6BX-MS Rev 1.1d BIOS. Both the UMDA Mode bug and 32GB IDE HDD limit bug are fixed. In the zip, I’ve include […]
Show full quote

Jackal1983 wrote on Today, 04:28:

Chkcpu wrote on Yesterday, 11:40:

Hi Jackal1983, […]
Show full quote

Hi Jackal1983,

What you are facing here is the infamous UDMA mode bug. This Award BIOS bug was fixed in March 1999 and later BIOS versions.
Unfortunately, the latest ECS P6BX-MS BIOS is Ver 1.1d from 02/08/1999. I assume you are running this BIOS?

This UDMA bug causes the BIOS to report a higher UDMA mode than the chipset can handle, when using an ATA66 or faster drive. Windows then rejects any UMDA mode and reverts to PIO mode 4 to protect your data on the HDD/SSD.

I have a fix for this UDMA mode bug. Give me a few days to patch the 02/08/1999 BIOS with this fix. While I’m at it, I will fix the 32GB HDD limit bug in this BIOS as well. 😀

Cheers, Jan

Yep, I'm on v 1.1d. And YES, please send it!!!

Here is the patched P6BX-MS Rev 1.1d BIOS.
Both the UMDA Mode bug and 32GB IDE HDD limit bug are fixed.
In the zip, I’ve included a PATCH.TXT file that lists the changes I’ve made.

The attachment BXMS_J1.zip is no longer available

When ATA/66/100/133 drives are fitted, the BIOS will now limit the reported UDMA Mode to UDMA 2, the maximum supported UDMA Mode on this 440BX chipset. This change is also visible on the BIOS summary screen.

The 32GB limit bugfix consists of one 32GB and two 64GB limit fixes. This patch J.1 BIOS now supports IDE drives up to 128GiB/137GB.
More info about these fixes can be found on my website at http://www.steunebrink.info/k6plus.htm#32GB_bugfix
The “new community fix” described there is the fix I used for your BIOS.

Jackal1983 wrote on Today, 08:00:

By the by, could you add in Coppermine and Tualatin microcode? I'd like to get this board working with C3 Ezra and Nehemiah CPUs and I figure those microcodes might help. Do you have any ideas along those lines?

I’ve looked at the CPU support in the Rev 1.1d BIOS, but I see only Intel CPUs are supported. VIA C3 detection and support code is nowhere to be found. Not surprising for an early 1999 Slot 1 BIOS. 😉

I’m unable to add VIA C3 support. I do mainly Socket 3/5/7 BIOS patches, and my modding repertoire for Slot 1/Socket 370 is very limited.

Do let us know how this P6BX-MS patch J.1 BIOS works.
Cheers, Jan

Looks like it works. DMA was already selected when I rebooted after flashing. What's the best way of testing read write speeds? I just did Roadkil's disk speed test and speeds maxed out at 25 mb/s which is faster than PIO-4, but is more along the lines of UDMA-1

Also, it doesn't seem like windows will allow DMA to be enabled for the SD card reader I have on the secondary channel. I'm using a 40 wire cable for it but that should be ok for UDMA 33, correct?

Great to hear the patched BIOS works and Windows allows UDMA now. 😀

My drive read and write speed benchmark of choice is the ATTO Disk Benchmark v2.41 (2007 version). Works fine under Win98.

A 40-wire IDE cable is indeed fine for UDMA 33. Only higher UDMA modes need the 80-wire cable.

DMA on SD card readers can be hit or miss. There are a lot of reports here on Vogons about that, and on SATA2IDE adapters too. Seems to depend a lot on the controller chip used.

Cheers, Jan

Chkcpu wrote on Today, 19:53:

Great to hear the patched BIOS works and Windows allows UDMA now. :) […]
Show full quote

Great to hear the patched BIOS works and Windows allows UDMA now. 😀

My drive read and write speed benchmark of choice is the ATTO Disk Benchmark v2.41 (2007 version). Works fine under Win98.

A 40-wire IDE cable is indeed fine for UDMA 33. Only higher UDMA modes need the 80-wire cable.

DMA on SD card readers can be hit or miss. There are a lot of reports here on Vogons about that, and on SATA2IDE adapters too. Seems to depend a lot on the controller chip used.

Cheers, Jan

So after running ATTO on my HDD with and without DMA enabled there is one helluva difference
DMA Disabled:
Read Max is 5.1 mb/s at 8192 kb
Write Max is 4 mb/s at 256 Kb, with writes at 64 kb and 128 kb being 3.9 mb/s
Write at 8192 kb was 3.8 mb/s

DMA Enabled:
Read Max is 26.2 mb/s at 8192 kb
Write Max is 13.6 mb/s at 64 kb, with writes at 128 kb and 256 kb being 9.89 mb/s and 10.5 mb/s
Writes at 8192 kb was 9 mb/s