No NCQ or Speed detected for Intel ICH10 AHCI SATA Controller

Alexander Motin mav at FreeBSD.org
Tue May 14 21:17:24 UTC 2013


On 14.05.2013 16:40, Steven Hartland wrote:
> We've got a number of FUJITSU RX200 S5 boxes here where we're seeing
> I/O performance issues which looks like its down to AHCI not negotiating
> correctly with the controller; specifically the FreeBSD AHCI driver is
> unable to detect both the speed capability and possibly more importantly
> NCQ support.
>
> Our baseline is 8.3-RELEASE but we've also tested 10-Current r250629
>
> I'm currently suspecting either a BIOS or HW issue, but would appreciate
> any ideas people may have, as we do have machines for other manufactures
> with fully working ICH10 AHCI.
>
> Relevant entries from a verbose boot:-
> ahci0: <Intel ICH10 AHCI SATA controller> port
> 0x1c50-0x1c57,0x1c44-0x1c47,0x1c48-0x1c4f,0x1c40-0x1c43,0x18e0-0x18ff
> mem 0xce022000-0xce0227ff irq 16 at device 31.2 on pci0
> ahci0: attempting to allocate 1 MSI vectors (16 supported)
> ahci0: using IRQ 266 for MSI
> ahci0: AHCI v1.20 with 6 ?Gbps ports, Port Multiplier not supported
> ahci0: Caps: 64bit AL ?Gbps PMD 32cmd CCC EM 6ports

AHCI controller capabilities there looks quite pitiful. That is not 
typical for ICH10. On my machine with Supermicro server board and ICH10 
chipset it looks like this:
ahci2: Caps: 64bit NCQ SNTF SS ALP AL CLO 3Gbps PMD SSC PSC 32cmd CCC EM 
eSATA 6ports

Without NCQ capability reported by the hardware driver predictably 
didn't enable it.

I think it may be a BIOS bug. At least I've seen once that AHCI 
capabilities changed after BIOS update.

-- 
Alexander Motin


More information about the freebsd-scsi mailing list