[gentoo-amd64] MSI Neo2 Platinum woes follow up

[gentoo-amd64] MSI Neo2 Platinum woes follow up

[Lee Thompson]

Trying to go from BIOS "v1.8" -> "v1.B" in preparation
for a dual core upgrade I ran into kernel hangs
initializing the SATA boot drive (libata).

As Duncan predicted, Knoppix and the 2005.1 LiveCD
booted clean as did a fresh install of 64 bit RH AS4.

I got Gentoo working by downgrading to kernel version
2.6.9-r9.

So, the old BIOS works with all versions of the kernel
going back a year.  The new BIOS doesn't work with
kernel 2.6.11 and 2.6.12.

Anyone familiar with the recent changes to libata or
nv_sata drivers?


		
__________________________________ 
Yahoo! FareChase: Search multiple travel sites in one click.
http://farechase.yahoo.com
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] MSI Neo2 Platinum woes follow up

[Nuitari]

> So, the old BIOS works with all versions of the kernel
> going back a year.  The new BIOS doesn't work with
> kernel 2.6.11 and 2.6.12.
>
> Anyone familiar with the recent changes to libata or
> nv_sata drivers?

No, but I know that there are fixes related to nv_sata in 2.6.14, so it 
might work again.

"sata_nv: Fixed bug introduced by 0.08's MCP51 and MCP55 support"
(the bug seems to have been introduced earlier in 2.6.14 development.

"[PATCH] Add NVIDIA device ID in sata_nv"

There is also a fix atapi_packet_task vs intr race (take 2) which is way 
too technical for me to understand before getting a good night of sleep.


-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] MSI Neo2 Platinum woes follow up

[Lee Thompson]

Okay, Got 3 Neo3 Platinum boxes on BIOS 1.B.

The two 64 bit Gentoo boxes are running clean with
2.6.9-r9 gentoo sources.  One has the dual core
upgrade and its very very nice!  Played with lots of
different kernels including gentoo-sources 2.6.14 and
vanilla-sources 2.6.14 on the single core box and they
all hang like the 2.6.12 and 2.6.13 kernels.

If I can find about 8 hours, I can probably isolate
the kernel code change(s) that causes the freeze on
this MBoard and float a patch into 2.6.14 and see if
some kernel guys can study it.

The third box is a windoze XP and it occasionally
needs the keyboard unplugged and replugged after the
BIOS starts.  This looks like a hardware issue because
it happens with BIOS 1.3, 1.8 and 1.B and occurs far
before the bootloader gets going.

I'm off to Frys because the dual core box has a stuck
chipset fan.  Might as well buy 3 replacements, one
for each board.

--- Nuitari <nuitari@melchior.nuitari.net> wrote:

> > So, the old BIOS works with all versions of the
> kernel
> > going back a year.  The new BIOS doesn't work with
> > kernel 2.6.11 and 2.6.12.
> >
> > Anyone familiar with the recent changes to libata
> or
> > nv_sata drivers?
> 
> No, but I know that there are fixes related to
> nv_sata in 2.6.14, so it 
> might work again.
> 
> "sata_nv: Fixed bug introduced by 0.08's MCP51 and
> MCP55 support"
> (the bug seems to have been introduced earlier in
> 2.6.14 development.
> 
> "[PATCH] Add NVIDIA device ID in sata_nv"
> 
> There is also a fix atapi_packet_task vs intr race
> (take 2) which is way 
> too technical for me to understand before getting a
> good night of sleep.
> 
> 
> -- 
> gentoo-amd64@gentoo.org mailing list
> 
> 



	
		
__________________________________ 
Yahoo! Mail - PC Magazine Editors' Choice 2005 
http://mail.yahoo.com
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] MSI Neo2 Platinum woes follow up

[Matt Randolph]

Lee Thompson wrote:

>Trying to go from BIOS "v1.8" -> "v1.B" in preparation
>for a dual core upgrade I ran into kernel hangs
>initializing the SATA boot drive (libata).
>
>As Duncan predicted, Knoppix and the 2005.1 LiveCD
>booted clean as did a fresh install of 64 bit RH AS4.
>
>I got Gentoo working by downgrading to kernel version
>2.6.9-r9.
>
>So, the old BIOS works with all versions of the kernel
>going back a year.  The new BIOS doesn't work with
>kernel 2.6.11 and 2.6.12.
>
>Anyone familiar with the recent changes to libata or
>nv_sata drivers?
>
>
>		
>__________________________________ 
>Yahoo! FareChase: Search multiple travel sites in one click.
>http://farechase.yahoo.com
>  
>

Same board.  Same BIOS upgrade (1.8 to 1.B).  I got 2.6.13-gentoo-r3 to 
work, but I'm not using SATA so I can't speak to that.  But I couldn't 
get the nvidia-kernel module to load.  I had to upgrade to the latest 
unstable version to get it to work at all.  But then I discovered I 
couldn't even get into the BIOS menu because the keyboard wasn't 
initialized until after it had stopped listening for the DEL keystroke.  
I downgraded back to 1.8 and now it won't POST!

Note that the versions released went from 1.8 to 1.9 to 1.B.  The fact 
that 1.A is missing from the lineup makes me wonder if they were having 
serious problems with it and decided to pull it.  Maybe 1.B still has 
some of those problems.

My ($10) replacement chip is slated to arrive today.  I've had to get by 
with Knoppix on a Windows box for nearly a fortnight.  And this was in 
the middle of my exams! 

I'm half tempted to buy a chip programmer so that this can't happen 
again.  And if they release any more shoddy BIOS versions, I'd be 
tempted to avoid MSI.

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] MSI Neo2 Platinum woes follow up

[Nuitari]

> Same board.  Same BIOS upgrade (1.8 to 1.B).  I got 2.6.13-gentoo-r3 to work, 
> but I'm not using SATA so I can't speak to that.  But I couldn't get the 
> nvidia-kernel module to load.  I had to upgrade to the latest unstable 
> version to get it to work at all.  But then I discovered I couldn't even get 
> into the BIOS menu because the keyboard wasn't initialized until after it had 
> stopped listening for the DEL keystroke.  I downgraded back to 1.8 and now it 
> won't POST!
>
> Note that the versions released went from 1.8 to 1.9 to 1.B.  The fact that 
> 1.A is missing from the lineup makes me wonder if they were having serious 
> problems with it and decided to pull it.  Maybe 1.B still has some of those 
> problems.
>
> My ($10) replacement chip is slated to arrive today.  I've had to get by with 
> Knoppix on a Windows box for nearly a fortnight.  And this was in the middle 
> of my exams! 
> I'm half tempted to buy a chip programmer so that this can't happen again. 
> And if they release any more shoddy BIOS versions, I'd be tempted to avoid 
> MSI.

There was some issue discussed earlier about nvidia and 2.6.13 kernels 
which basically said to use the unstable (and up to date) version of the 
nvidia drivers.

Did you try connecting the keyboard via the PS/2 ports?
I had the same problem with my MSI board until I did it.Enabling Legacy 
USB support also works most of the time. I also found that sometimes the 
USB keyboard won't be initialized between the time the BIOS is done 
posting and until the linux kernel enables it making a mess of LILO but it 
only seems to happen on a boot out of 2.


There are plenty of cases where a BIOS updates skip a version steps.

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] MSI Neo2 Platinum woes follow up

[Matt Randolph]

Nuitari wrote:

>
> There was some issue discussed earlier about nvidia and 2.6.13 kernels 
> which basically said to use the unstable (and up to date) version of 
> the nvidia drivers.

FWIW, I was cheerfully using 2.6.13 with nvidia-kernel-1.0.6629-r6 for 
some time with no ill effects. Starting as of about an hour ago, I'm 
using 1.0.7676-r1 but I may revert to the stable version if I discover 
any issues with my applications.

>
> Did you try connecting the keyboard via the PS/2 ports?
> I had the same problem with my MSI board until I did it.Enabling 
> Legacy USB support also works most of the time. I also found that 
> sometimes the USB keyboard won't be initialized between the time the 
> BIOS is done posting and until the linux kernel enables it making a 
> mess of LILO but it only seems to happen on a boot out of 2.
>
I tried both USB and PS/2 keyboards without luck. Part of the problem is 
that you need to be able to get into the BIOS to change that setting. 
And after a flash, it might have reverted to the default value, which 
may or may not be what you want. That should be a problem that is unique 
to USB keyboards though, so I was surprised that I had a similar problem 
with a PS/2 one.

>
> There are plenty of cases where a BIOS updates skip a version steps.
>
Thank you for the reminder. I notice that they skipped 1.6 too. I know 
that software of any kind sometimes skips a version between releases. I 
expect there can be any of a wide range of reasons for this, but I 
wouldn't be surprised if it is at least occasionally because of problems 
with a release candidate.

BTW, the chip came and I'm back in business. There's no place like 
/home. Do not envy me the two weeks worth of updates I'm about to 
emerge. Fortunately, I have decided that I can still get away with using 
1.8 for now (after all that!).

I shouldn't have been nasty about MSI previously. I have been very happy 
with them by and large. And that's based on owning about half a dozen of 
their motherboards over the last few years. I do hope they resolve any 
issues people are having with 1.B, though.

You may find this even more entertaining than version numbers in hex: 
You know how big a PLCC-32 chip is? That's right, it's about the size of 
your fingernail. The box they sent my chip in is big enough to hold a 
small pizza.

-- 
gentoo-amd64@gentoo.org mailing list