I have a Tyan s2895 board and I'm using a PCIe GeForce 6600 Video adapter. With the factory BIOS, lspci shows the interrupt routed to IRQ 5, and it works fine.
With Coreboot, lscpi shows the interrupt routed to IRQ 18 (value from mptable), and it finally loads X after a few tries in low resolution mode.
With Coreboot and ACPI the interrupt gets routed to IRQ 0, even though I put 18 in the _PRT table. X loads more quickly, but there are obvious problems with text highlighting.
In none of the three cases can I find the IRQ listed in /proc/interrupts. I would have expected to see it there.
I don't see anything in dmesg that seems to be related to IRQ 18.
Can anyone suggest the next step?
Thanks, Myles
Hi,
With Coreboot and ACPI the interrupt gets routed to IRQ 0, even though I put 18 in the _PRT table. X loads more quickly, but there are obvious problems with text highlighting.
This is OK, just the PCI line reg is not initialized. Its used only when the routing of PCI is done via 8259. Check http://www.coreboot.org/ACPI_in_coreboot#acpi_fill_madt
There is some pics which explains that.
In none of the three cases can I find the IRQ listed in /proc/interrupts. I would have expected to see it there.
It means that the driver is not using any interrupts Maybe you need to load the "nvidia" proprietary module.
Rudolf
On Tue, Feb 24, 2009 at 12:02 PM, Rudolf Marek r.marek@assembler.cz wrote:
Hi,
With Coreboot and ACPI the interrupt gets routed to IRQ 0, even though I put 18 in the _PRT table. X loads more quickly, but there are obvious problems with text highlighting.
This is OK, just the PCI line reg is not initialized. Its used only when the routing of PCI is done via 8259. Check http://www.coreboot.org/ACPI_in_coreboot#acpi_fill_madt
There is some pics which explains that.
Sorry to make you continually refer to the same place. I thought I understood...
In none of the three cases can I find the IRQ listed in /proc/interrupts. I would have expected to see it there.
It means that the driver is not using any interrupts Maybe you need to load the "nvidia" proprietary module.
I don't really care that it's not using the interrupt. I was just thinking that might be the problem since the factory BIOS uses IRQ 5 and Coreboot uses 18. If interrupts aren't the problem, I wonder what's causing it to fail.
Thanks, Myles
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi,
Sorry to make you continually refer to the same place. I thought I understood...
Well I thought you start to ask question so we can improve the page. Its mostly mine braindump. It must be improved but this is better then nothing.
In none of the three cases can I find the IRQ listed in /proc/interrupts. I would have expected to see it there.
It means that the driver is not using any interrupts Maybe you need to load the "nvidia" proprietary module.
I don't really care that it's not using the interrupt. I was just thinking that might be the problem since the factory BIOS uses IRQ 5 and Coreboot uses 18. If interrupts aren't the problem, I wonder what's causing it to fail.
You see some strange graphical artifacts? Maybe disable 2D/3D accel. I guess perhaps some PCI regions are incorrectly routed.
Rudolf
-
Myles Watson -
Rudolf Marek