After booting and after SeaBIOS picks-up things, the boot process slows when executing the kernel.
Although this box has an add-in PCI SATA card with two SATA drives, I believe I'm seeing the slow down by no matter what boot method I use.
Attaching full log. This one details booting from a SATA hard drive printing specific error codes to stdout.
Hi,
This looks like the IRQ routing for PCI slots is incomplete/wrong. I think you need to fix it in PIR MPTable and in ACPI. Fix whatever tables you are using.
Is this board in SVN already?
Thanks, Rudolf
On Mon, Jan 10, 2011 at 08:26:52AM +0100, Rudolf Marek wrote:
Hi,
This looks like the IRQ routing for PCI slots is incomplete/wrong. I think you need to fix it in PIR MPTable and in ACPI. Fix whatever tables you are using.
I would presume grep the coreboot source for those terms?
Is this board in SVN already?
This is using the Tyan S1846 as a template for my Tyan S1832DL. The S1846 is a single processor, mine is a dual processor.
However, it's probably getting to that point where I need to create a new profile for the S1832DL.
Thanks, Rudolf
Am 10.01.2011 12:18, schrieb Roger:
On Mon, Jan 10, 2011 at 08:26:52AM +0100, Rudolf Marek wrote:
This looks like the IRQ routing for PCI slots is incomplete/wrong. I think you need to fix it in PIR MPTable and in ACPI. Fix whatever tables you are using.
I would presume grep the coreboot source for those terms?
These are managed per board for now. Look in your board's directory for mptable.c, irq_tables.c and dsdt.asl. If these don't exist (or aren't enabled in Kconfig), you're not doing any such configuration yet, which Linux might or might not like, which seems to differ between kernel releases.
However, it's probably getting to that point where I need to create a new profile for the S1832DL.
If you do, please copy the s1846 sources with "svn cp src/mainboard/tyan/s1846 src/mainboard/tyan/s1832dl", and move your changes into the new directory. This ensures that the patches vs. the old board are easier to review (only the changes against the old board), and that your files keep repository history. This should also help a bit with keeping your board up to date, but really, if you manage to get into Linux at all, the code is ready for publication in my opinion. Your benefit of early publication: When we do some sweeping changes across the tree, we try to drag your board along (not always successfully, but the biggest part is done). If you have a local copy, you're on your own.
Patrick