Hi,
Thanks for Takeshi. We now have better ram detection and vgabios support. I also applied hcyun's serial debug patch and epia support for ADLO.
src/northbridge/via/vt8601/vgainit.inc src/arch/i386/lib/vgabios.c src/bioscall/pcibios.c src/mainboard/via/epia/Config util/ADLO/bochs/bios/rombios.c util/ADLO/Makefile
This update should be very limited to epia.
Takeshi, would you mind to verify the tree. Since I still have problem, I might have missed something. However, it still looks much beter than before.
-Andrew
On Thu, May 22, 2003 at 09:05:38PM +0800, Andrew Ip wrote:
Thanks for Takeshi. We now have better ram detection and vgabios support.
Thank you for committing.
I also applied hcyun's serial debug patch and epia support for ADLO.
It doesn't contain any of my work to ADLO. It's just serial patch. EPIA runs just fine anyway.
Takeshi, would you mind to verify the tree. Since I still have problem, I might have missed something. However, it still looks much beter than before.
You missed an important line. Attached vgabios.patch fixes it.
To build ADLO with "make epia" you introduced, pirq table blob is needed, but you didn't add it to the tree.
Also, you need to modify some part of loader.s. My working example is attached adlo-epia.patch. Without it, ADLO screws some part of PCI controller registers. You might also want to change memory size in loader.s.
With above changes, my EPIA runs fine with VGABIOS+memtest86 payload (memtest86 runs on VGA screen), and ADLO which boots to GRUB->Linux. (BOOT_IDE didn't work this morning...)
-- Takeshi
Of course if you commit this ADLO patch for EPIA you will break ADLO on SIS630.
No point to 'fix' one architecture to break another.
On Fri, 23 May 2003, SONE Takeshi wrote:
On Thu, May 22, 2003 at 09:05:38PM +0800, Andrew Ip wrote:
Thanks for Takeshi. We now have better ram detection and vgabios support.
Thank you for committing.
I also applied hcyun's serial debug patch and epia support for ADLO.
It doesn't contain any of my work to ADLO. It's just serial patch. EPIA runs just fine anyway.
Takeshi, would you mind to verify the tree. Since I still have problem, I might have missed something. However, it still looks much beter than before.
You missed an important line. Attached vgabios.patch fixes it.
To build ADLO with "make epia" you introduced, pirq table blob is needed, but you didn't add it to the tree.
Also, you need to modify some part of loader.s. My working example is attached adlo-epia.patch. Without it, ADLO screws some part of PCI controller registers. You might also want to change memory size in loader.s.
With above changes, my EPIA runs fine with VGABIOS+memtest86 payload (memtest86 runs on VGA screen), and ADLO which boots to GRUB->Linux. (BOOT_IDE didn't work this morning...)
-- Takeshi
On Thu, 22 May 2003, Adam Sulmicki wrote:
Of course if you commit this ADLO patch for EPIA you will break ADLO on SIS630.
the, clearly, please don't commit.
ron
On Thu, May 22, 2003 at 07:06:38PM -0600, ron minnich wrote:
On Thu, 22 May 2003, Adam Sulmicki wrote:
Of course if you commit this ADLO patch for EPIA you will break ADLO on SIS630.
the, clearly, please don't commit.
How about applying that patch but disabling it by default by setting DEBUG_SERIAL to 0.
-- Takeshi
Takeshi,
How about applying that patch but disabling it by default by setting DEBUG_SERIAL to 0.
Tried. It only works with DEBUG_SERIAL=1 and DEBUG_ATA=1. BTW, memtest passes. Next, I'll compare the values between LinuxBIOS and regualar.
-Andrew
On Sat, May 24, 2003 at 12:10:37AM +0800, Andrew Ip wrote:
How about applying that patch but disabling it by default
by setting DEBUG_SERIAL to 0.
Tried. It only works with DEBUG_SERIAL=1 and DEBUG_ATA=1. BTW,
Well, I recommended that to commit to cvs, so that that patch doesn't break other boards any more. Not for you to try with EPIA. Sorry for bad English.
Good news: if you add a dummy function, vga_hardware_fixup(), VIDEO_CONSOLE just works. Can we state so at the status page of web? :)
-- Takeshi
It doesn't contain any of my work to ADLO. It's just serial patch. EPIA runs just fine anyway.
If it is not necessary, I probably skip this patch, since it may affect other mb.
You missed an important line. Attached vgabios.patch fixes it.
Fixed
To build ADLO with "make epia" you introduced, pirq table blob is needed, but you didn't add it to the tree.
Done.
Also, you need to modify some part of loader.s. My working example is attached adlo-epia.patch. Without it, ADLO screws some part of PCI controller registers. You might also want to change memory size in loader.s.
I'll try.
-Andrew