[Please reply-to-all.]
On Sun, 2011-11-27 at 18:44 -0700, Peter Verschoor wrote:
Ben:
Now comes the head-scratching part. When I modify the grub boot
command line to send output to my laptop through the serial port it
boots correctly!
Serial logging will change the timing of the boot sequence.
[ 0.000000] Command line: BOOT_IMAGE=/boot/vmlinuz-3.1.0-1-amd64
root=UUID=155d483d-563c-4c25-b9f2-bf0b61e14041 ro console=ttyS0,9600
console=tty0
and the cciss IRQ gets rerouted to 19.
The output appears on both my laptop and the main console screen. I
can capture the output on the laptop and the kernel captures it
in /var/log.
If I boot it with this command line it will lock up every time.
Command line: BOOT_IMAGE=/boot/vmlinuz-3.1.0-1-amd64
root=UUID=155d483d-563c-4c25-b9f2-bf0b61e14041 ro
The attached file CAPTURE.TXT was created on an old IBM laptop running
Win2k using Hyperterm and sneaker-netted to the Proliant using a USB
stick. That means that at least the front USB port is working.
But not one of the other USB ports, it seems:
[...]
[ 10.834493] cciss 0000:04:03.0: PCI IRQ 51 -> rerouted to legacy IRQ 19
[ 10.834515] cciss 0000:04:03.0: PCI INT A -> GSI 19 (level, low) -> IRQ 19
[ 10.892092] cciss 0000:04:03.0: cciss0:<0x46> at PCI 0000:04:03.0 IRQ 19 using DAC
[...]
[ 13.381713] uhci_hcd 0000:00:1d.1: PCI INT B -> GSI 19 (level, low) -> IRQ 19
[ 13.467200] uhci_hcd 0000:00:1d.1: UHCI Host Controller
[ 13.529832] uhci_hcd 0000:00:1d.1: new USB bus registered, assigned bus number 3
[ 13.618540] uhci_hcd 0000:00:1d.1: request interrupt 19 failed
[...]
Now, because both of these devices are using legacy PCI interrupts, they
should be able to share IRQ 19. But the cciss driver was changed in
Linux 2.6.36 so that it refuses to share an IRQ. Mike, why was that?