Re: [PATCH] AHCI PCI MSI support, update 1

[Jens Axboe]

On Tue, May 31 2005, Jeff Garzik wrote:
> 
> I just updated the AHCI PCI MSI support patch to support the latest 
> ->host_stop() semantics changes that just made it into upstream.
> 
> This patch, until its merged, may always be found in the 'ahci-msi' 
> branch of
> rsync://rsync.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev.git

Tested on an ich6 based system, with this ahci controller:

centera:~ # lspci -s 0000:00:1f.2 -v
0000:00:1f.2 IDE interface: Intel Corp. I/O Controller Hub SATA cc=raid
(rev 03) (prog-if 8f [Master SecP SecO PriP PriO])
        Subsystem: Tyan Computer: Unknown device 5150
        Flags: bus master, 66Mhz, medium devsel, latency 0, IRQ 169
        I/O ports at e900
        I/O ports at ea00 [size=4]
        I/O ports at eb00 [size=8]
        I/O ports at ec00 [size=4]
        I/O ports at ed00 [size=16]
        Memory at d03c3000 (32-bit, non-prefetchable) [size=1K]
        Capabilities: [70] Power Management version 2

And system boots and works fine, MSI isn't enabled of course.

Also tested on an ich7 based system (945G express chipset), with this
ahci controller:

nelson:~ # lspci -s 0000:00:1f.2 -v
0000:00:1f.2 Class 0106: Intel Corporation I/O Controller Hub SATA
cc=AHCI (rev 01) (prog-if 01)
        Subsystem: Intel Corporation: Unknown device e302
        Flags: bus master, 66Mhz, medium devsel, latency 0, IRQ 201
        I/O ports at 40d8 [size=8]
        I/O ports at 40f4 [size=4]
        I/O ports at 40d0 [size=8]
        I/O ports at 40f0 [size=4]
        I/O ports at 4020 [size=32]
        Memory at 48304000 (32-bit, non-prefetchable) [size=1K]
        Capabilities: [80] Message Signalled Interrupts: 64bit- Queue=0/0 Enable+
        Capabilities: [70] Power Management version 2
        Capabilities: [a8] #12 [0010]

System boots and works fine (typing this message here), MSI is enabled.
I double checked with an added printk :-)

So that's at least one ACK for you. Now running NCQ and MSI enabled
ahci, I'm not sure SATA gets any better!

-- 
Jens Axboe

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/