Re: VIA IDE no go with 2.4.5-ac1

[Vojtech Pavlik <vojtech@suse.cz>]

On Sun, May 27, 2001 at 08:28:14PM +0400, Oleg Drokin wrote:

> On Sun, May 27, 2001 at 05:18:20PM +0100, Alan Cox wrote:
> > >   Vanilla 2.4.5 boots ok, but 2.4.5-ac1 finishes kernel initialisation and
> > >   starts to print "hda: lost interrupt", I guess this is related to VIA IDE
> > >   updates in AC kernels. Config for vanilla and AC kernel is the same.
> > >   Here are the kernel logs from 2.4.5 and 2.4.5-ac1 (collected with serial
> > > ACPI: Core Subsystem version [20010208]
> > > ACPI: Subsystem enabled
> > > ACPI: Not using ACPI idle
> > > ACPI: System firmware supports: S0 S1 S4 S5
> > > hda: lost interrupt
> > > hda: lost interrupt
> > Does this still happen if you build without ACPI support. Also does
> > 'noapic' have any impact ?
> I will try this and report.
> I received this patch from Carlos E Gorges <carlos@techlinux.com.br>,
> that allows my box to boot, but DMA is not enabled by default
> (and needs to be explicitly enabled by hdparm -d1 /dev/hda) regardless of
> what is written at boot time.
> 
> --- drivers/ide/via82cxxx.c.orig	Sun May 27 08:10:47 2001
> +++ drivers/ide/via82cxxx.c	Sun May 27 08:11:13 2001
> @@ -105,7 +105,7 @@
>  	{ "vt8233",	PCI_DEVICE_ID_VIA_8233_0,   0x00, 0x2f, VIA_UDMA_100 },
>  	{ "vt8231",	PCI_DEVICE_ID_VIA_8231,     0x00, 0x2f, VIA_UDMA_66 },
>  #endif
> -	{ "vt82c686b",	PCI_DEVICE_ID_VIA_82C686,   0x40, 0x4f, VIA_UDMA_100 | VIA_BAD_PIO },
> +	{ "vt82c686b",	PCI_DEVICE_ID_VIA_82C686,   0x40, 0x4f, VIA_UDMA_100 },
>  	{ "vt82c686a",	PCI_DEVICE_ID_VIA_82C686,   0x10, 0x2f, VIA_UDMA_66 },
>  	{ "vt82c686",	PCI_DEVICE_ID_VIA_82C686,   0x00, 0x0f, VIA_UDMA_33 | VIA_BAD_CLK66 },
>  	{ "vt82c596b",	PCI_DEVICE_ID_VIA_82C596,   0x10, 0x2f, VIA_UDMA_66 },

Not sure what version you're using, but if it's 3.23 (I believe it is)
then this patch does completely nothing, because the VIA_BAD_PIO flag
isn't used.

-- 
Vojtech Pavlik
SuSE Labs