On Mon, Oct 05, 2009 at 11:01:24PM +0900, Christophe Prevotaux wrote: > Hi, > > My machine is Running OpenBSD/amd64 4.6 (snapshot 02 Oct 2009) > > > The kernel crashed on AHCI detection. > > ahci0 at pci0 dev 18 function 0 "ATI SB600 SATA" rev 0x00: apic 2 int 22 (irq > 11), AHCI 1.1 > scsibus0 at achci0: 32 targets > sd0 at scsibus0 targ 0 lun 0: <ATA, Hitachi HDS72161, P220> SCSI3 0/direct > fixed > sd0: 152627MB, 512 bytes/sec, 312581808 sec total > panic: kernel diagnostic assertion "ccb->ccb_xa.state == ATA_S_ONCHIP" > failed: "../../../../dev/pci/ahci.c", line 1812 > Stopped at Debugger+0x5: leave > Debugger() at Debugger+0x5 > panic() at panic+0x122 > __assert() at __assert+0x21 > ahci_port_intr() at ahci_port_intr+0x218 > ahci_poll() at ahci_poll+0x4d > ahci_ata_cmd() at ahci_ata_cmd+0x9b > atat_exec() at ata_exec+0x19 > scsi_xs_exec() at scsi_xs_exec+0x38 > scsi_scsi_cmd() at scsi_scsi_cmd+0x98 > scsi_test_unit_ready() at scsi_test_unit_ready+0x43 > end trace frame: 0xffffffff80898a30, count 0 > > > The only way to make the machine boot is > > boot -c > UKC> disable ahci > > then set the fstab to use wd0 (Disable ACPI does not change a thing)
A known problem, fortunately confined to the ATI SB600. Some developers have examined it, including me, and failed to figure out the correct fix/what SB600 is doing wrong. At least one hack is floating around which fixed this issue in an unpleasant way nobody thought correct. Easiest fix is to avoid SB600. I gave mine away so I wouldn't go insane poking at this issue any more. .... Ken
