Re: SATA SB600 works in 2.6.20.4 but not in 2.6.21-rc5 with irqpoll parameter

[Tejun Heo <htejun@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 873 bytes --]

Matí­as Alejandro Torres wrote:
> 
>> Well, making it work without pci=nomsi is the whole point here.  I dunno
>> why this simple thing doesn't work.  Can you post the result of 'lspci
>> -nn' so that we can see the numeric ID?
>>   
> I would really like to know the answer to that, but i'm hundred of miles
> away. Attached to this email is the output of  lspci -nn .
> 
> Maybe you have the wrong id of the pci bridge:
> 
> 00:02.0 PCI bridge [0604]: ATI Technologies Inc RS480 PCI-X Root Port
> [1002:5a34]
> 
> +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x5a3f, quirk_disable_msi);
> /* RS480 PCI bridge */

Yeah, apparently.

00:02.0 PCI bridge [0604]: ATI Technologies Inc RS480 PCI-X Root Port
[1002:5a34]

So, there are two broken PCI bridges.  Argh....  Okay, please try the
attached patch.  It includes both 5a34 and 5a3f.

Thanks for your patience.

-- 
tejun

[-- Attachment #2: patch --]
[-- Type: text/plain, Size: 622 bytes --]

diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
index 3411483..b0e89db 100644
--- a/drivers/pci/quirks.c
+++ b/drivers/pci/quirks.c
@@ -1648,6 +1648,8 @@ static void __devinit quirk_disable_msi(
 	}
 }
 DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_8131_BRIDGE, quirk_disable_msi);
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x5a34, quirk_disable_msi); /* RS480 PCI-X root port */
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x5a3f, quirk_disable_msi); /* RS480 PCI bridge */
 
 /* Go through the list of Hypertransport capabilities and
  * return 1 if a HT MSI capability is found and enabled */