Ron,
You need to set the REG in some special position. Because ht_scanbus, may reset some regs according to it's scan result.
I think the best to use VGA is 1. LinuxBIOS init critical regs to make sure VGA can be used in Linux kernel. 2. Linux Kernel does the special work to make use some acceleration function works. 3. Let X Server no changes.
Regards
YH.
-----邮件原件----- 发件人: ron minnich [mailto:rminnich@lanl.gov] 发送时间: 2004年1月8日 14:11 收件人: Stefan Reinauer 抄送: linuxbios@clustermatic.org 主题: Re: K8 northbridge f1 behaviour.
On Thu, 8 Jan 2004, Stefan Reinauer wrote:
Do these registers maybe want an LDTSTOP_L or a reset asserted to become significant?
ouch. maybe.
Are you sure your other operations do access the VGA ports at all and don't operate on the graphics device directly?
These are operations that access 0x3cc and other legacy vga stuff, and they're not getting to vga ...
I got an S2880 up with VGA by using the ATI biosless patch for 2.4.x done by some SGI guy. That is not really flexible, but it helped me to see light for the first time in LinuxBIOS
The right solution. We're doing the wrong solution (running the VGA bios) because VGA people won't let us do the right solution.
ron
_______________________________________________ Linuxbios mailing list Linuxbios@clustermatic.org http://www.clustermatic.org/mailman/listinfo/linuxbios
On Thu, 2004-01-08 at 15:43, YhLu wrote:
Ron,
You need to set the REG in some special position. Because ht_scanbus, may reset some regs according to it's scan result.
I think the best to use VGA is
- LinuxBIOS init critical regs to make sure VGA can be used in Linux
kernel. 2. Linux Kernel does the special work to make use some acceleration function works. 3. Let X Server no changes.
YhLu,
We found that under normal bios, the AMD8151 is not present. Is there any bits to hide it ?
We have tried everything to forward VGAIO behind 8111 and 8151 bridges but fail. Do you know anything about it ?
Ollie
Li-Ta Lo ollie@lanl.gov writes:
On Thu, 2004-01-08 at 15:43, YhLu wrote:
Ron,
You need to set the REG in some special position. Because ht_scanbus, may reset some regs according to it's scan result.
I know the existing code when it was back in the freebios codebase and PCI only handled this correctly, so it should be just a matter of a little bit of debugging to get this to happen automatically if it does not already.
I think the best to use VGA is
- LinuxBIOS init critical regs to make sure VGA can be used in Linux
kernel. 2. Linux Kernel does the special work to make use some acceleration function works. 3. Let X Server no changes.
YhLu,
We found that under normal bios, the AMD8151 is not present. Is there any bits to hide it ?
Hmm. This may because of how things are laid out. I wonder if there are driver issues when you put the 8151 on something besides bus 0.
We have tried everything to forward VGAIO behind 8111 and 8151 bridges but fail. Do you know anything about it ?
That is very weird. Besides the card not working what test are you using to be certain you have failed? Are you failing to read/write an I/O register.
It occurs to me LinuxBIOS might be different enough in pci layout to confuse the option ROM on your card.
Plus I think there are some special bits of AGP initialization that we might not be doing.
Eric
On 8 Jan 2004, Eric W. Biederman wrote:
That is very weird. Besides the card not working what test are you using to be certain you have failed? Are you failing to read/write an I/O register.
on "normal" bios, i/o to 0x3cc produces data. On linuxbios, i/o to 0x3cc always comes back as 0xff.
Here is my understanding of the tyan s2885 layout: cpu0, link2, goes to the 8131, which has a tunnel to 8111, which drives pci.
Under our linuxbios + hack, c0.b is 0x33. c4.l is set for link2. So VGA i/o should in my view go out link2.
the 8111 is the pci connection. register 3e.b in 8111 is set to 0xf, which should pass vga i/o to the pci bus.
the vga card, using Ollie's 'userio' program, does not respond to byte accesses to 0x3cc or 0x3d4. Under normal bios, all these register settings are the same and ... the card responds.
It occurs to me LinuxBIOS might be different enough in pci layout to confuse the option ROM on your card.
no, I don't think that's it.
Plus I think there are some special bits of AGP initialization that we might not be doing.
even on pci it won't work.
ron
On Fri, Jan 09, 2004 at 08:27:55AM -0700, ron minnich wrote:
[..]
the 8111 is the pci connection. register 3e.b in 8111 is set to 0xf, which should pass vga i/o to the pci bus.
the vga card, using Ollie's 'userio' program, does not respond to byte accesses to 0x3cc or 0x3d4. Under normal bios, all these register settings are the same and ... the card responds.
It seems you've solved this now, according to the later success report. Excellent news indeed! :)
I would've suggested a modified POST card or logic analyzer to see if the writes made it to the bus. This is somewhat more of a LAN approach to diagnostics, but might have helped anyway. LAN bridges also tend to just make packets disappear mysteriously.
What was the problem, anyway?
//Peter
On Sat, 10 Jan 2004, Peter Stuge wrote:
I would've suggested a modified POST card or logic analyzer to see if the writes made it to the bus. This is somewhat more of a LAN approach to diagnostics, but might have helped anyway. LAN bridges also tend to just make packets disappear mysteriously.
I finally found my VMETRO and put it in the slot and we were able to work it all out.
What was the problem, anyway?
- incorrect setting of the pci i/o registers in the k8 northbridge (fix in the works) - the stupid vga cards don't respond to I/O in the legacy range until you do a Magic Write to a Magic Place with Magic Data. The emulator helped us figure this out since we can watch I/O
But we're really a lot closer than we ever have been.
ron
-
ebiederman＠lnxi.com -
Li-Ta Lo -
Peter Stuge -
ron minnich -
YhLu