Uhh oh, that's actually interresting and you are right on this point.
One should just add the functionality. If you dare to wait the weekend
it will happen in 2.5 ;-). Or of you care your self, then plase
have a look at the following code in 2.5.6-pre3 in ide-pci:
/*
* Setup DMA transfers on a channel.
*/
static void __init setup_channel_dma(ide_hwif_t *hwif, struct pci_dev *dev,
ide_pci_device_t *d,
int port,
u8 class_rev,
int pciirq, ide_hwif_t **mate,
int autodma, unsigned short *pcicmd)
{
unsigned long dma_base;
if (d->flags & ATA_F_NOADMA)
autodma = 0;
if (autodma)
hwif->autodma = 1;
if (!((d->flags & ATA_F_DMA) || ((dev->class >> 8) ==
PCI_CLASS_STORAGE_IDE && (dev->class & 0x80))))
return;
Would it be sufficient to just prevent the classes you mention to
bail out from initialization?
> It's pity that the linux kernel could agree this point also.
It can be fixed I think.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/