EPIA Linuxbios strangeness

[Jeff Noxon]

I got LB working on my EPIA 533 machine.  It works fine -- BUT,

1) The reset button does nothing.  No kidding.

2) I can't do a software reboot either.  The machine just goes dead if
I try "init 6" or any variation of the reboot command.

3) If I unplug it & plug it back in, the power stays off.

Any ideas?

Thanks!

Jeff

[Jeff Noxon]

The reset button turned out to be a bad connection.

I still can't reboot the system under software control.  Anyone?

Thanks,

Jeff

On Mon, Aug 11, 2003 at 09:42:04AM -0500, Jeff Noxon wrote:
> I can't do a software reboot either.  The machine just goes dead if
>
> I try "init 6" or any variation of the reboot command.
> If I unplug it & plug it back in, the power stays off.

[ron minnich]

On Wed, 13 Aug 2003, Jeff Noxon wrote:
> I still can't reboot the system under software control.  Anyone?

you're going to have to figure out how to make those chips assert pci 
reset. Once you get it, tell me :-)

ron

[Andrew Ip]

Hi,
> you're going to have to figure out how to make those chips assert pci 
> reset. Once you get it, tell me :-)

I don't have this reset problem.

-Andrew

[ron minnich]

On Thu, 14 Aug 2003, Andrew Ip wrote:
> > you're going to have to figure out how to make those chips assert pci 
> > reset. Once you get it, tell me :-)
>
> I don't have this reset problem.

I think overall the EPIA port is shaky. David can't build a working
romimage here. I was able to last decemnber and will try again today.

I think EPIA is a first good candidate for freebios2, dumping assembly and 
going to C.

ron

[Jeff Noxon]

On Thu, Aug 14, 2003 at 02:26:01PM +0800, Andrew Ip wrote:
> Ron Minnich wrote:
> > you're going to have to figure out how to make those chips assert pci 
> > reset. Once you get it, tell me :-)
>
> I don't have this reset problem.

Interesting.  Any suggestions on where I should poke around to help
debug this?

As an aside, I found a Winbond w49f002u chip and decided to back up the
original EPIA BIOS to it.  I had to remove the print statements from the
flash_and_burn w49f002u driver, turn off the chip erase, and program it
repeatedly before it verified correctly!  Apparently that chip is very
timing sensitive!  Aside from that issue, the Winbond chip works fine
on the EPIA.

Thanks,

Jeff

[ron minnich]

On Thu, 14 Aug 2003, Jeff Noxon wrote:
> As an aside, I found a Winbond w49f002u chip and decided to back up the
> original EPIA BIOS to it.  I had to remove the print statements from the
> flash_and_burn w49f002u driver, turn off the chip erase, and program it
> repeatedly before it verified correctly!  Apparently that chip is very
> timing sensitive!  Aside from that issue, the Winbond chip works fine
> on the EPIA.

this seems to be a generic flash programming issue, timing is tricky even 
on parts with clocks.

ron

[SONE Takeshi]

On Mon, Aug 11, 2003 at 09:42:04AM -0500, Jeff Noxon wrote:
> I can't do a software reboot either.  The machine just goes dead if
>
> I try "init 6" or any variation of the reboot command.

Setting bit 1 of ISA bridge register 4F does a system reset.
I have these lines instead of /sbin/reboot in /etc/init.d/reboot
(Debian system).

sleep 1
    setpci -s 0:11.0 4f.b=9

For power-off, I'm using a small program in /etc/init.d/halt.
It is something like the one I posted to this list before
(now with the daemon mode to watch the power-button hit).

However, these things should eventually go in the kernel.
Userspace solutions ignore shutdown callbacks in the kernel,
so it might be dangerous. For me these things work well,
probably helped by "sleep 1" to allow the drive to flush its
writeback cache..
> If I unplug it & plug it back in, the power stays off.

It needs some programming to the supersouth.

[Steve Gehlbach]

SONE Takeshi wrote:
> Setting bit 1 of ISA bridge register 4F does a system reset.
> I have these lines instead of /sbin/reboot in /etc/init.d/reboot
> (Debian system).
> sleep 1
> setpci -s 0:11.0 4f.b=9

From the datasheet:

Via VT8231 fn 0
Offset 4F –PCI Bus and CPU Interface Control ..........RW
----------------------------------------------------------

7-4 Reserved ................always reads 0
3 CPU Reset Source
   0 Use CPURST as CPU Reset..default
   1 Use INIT# as CPU Reset
2 Config Command Reg Rx04 Access (Test Only)
   0 Normal:Bits 0-1=RO,Bit 3=RW...default
   1 Test Mode:Bits 0-1=RW,Bit-3=RO
1 IRDY #Wait States
   0 0 Wait States.............default
   1 1 WaitState
0 Software PCI Reset ...write 1 to generate PCI reset

What about the ACPI I/O space (set same as Award BIOS, I think, 0x4000), 
in particular 0x4002, 0x4003? Anyway this can be used?

-Steve

[SONE Takeshi]

On Sun, Aug 17, 2003 at 01:12:29PM -0700, Steve Gehlbach wrote:
> SONE Takeshi wrote:
> > Setting bit 1 of ISA bridge register 4F does a system reset.

~~~~~
Typo, bit 0.
> > I have these lines instead of /sbin/reboot in /etc/init.d/reboot
> > (Debian system).
> >
> > sleep 1
> >    setpci -s 0:11.0 4f.b=9
>
> From the datasheet:
>
> Via VT8231 fn 0
> Offset 4F ?PCI Bus and CPU Interface Control ..........RW
>
> 7-4 Reserved ................always reads 0
> 3 CPU Reset Source
>   0 Use CPURST as CPU Reset..default
>   1 Use INIT# as CPU Reset
> 2 Config Command Reg Rx04 Access (Test Only)
>   0 Normal:Bits 0-1=RO,Bit 3=RW...default
>   1 Test Mode:Bits 0-1=RW,Bit-3=RO
> 1 IRDY #Wait States
>   0 0 Wait States.............default
>   1 1 WaitState
> 0 Software PCI Reset ...write 1 to generate PCI reset
>
> What about the ACPI I/O space (set same as Award BIOS, I think, 0x4000), 
> in particular 0x4002, 0x4003? Anyway this can be used?

I use it only for power-off.
It does have a reset function, but it didn't always work for me.