Re: 2.6.2-pa3 framebuffer buglet and broken ALSA harmony driver
On Mon, Feb 16, 2004 at 04:14:58PM +0200, Martin-?ric Racine wrote:
> snd_harmony: Unknown symbol snd_malloc_pci_pages
> snd_harmony: Unknown symbol ccio_get_fake
>
> If I'm not mistaken, the 712 workstation doesn't have any PCI bus. Make me
> wonder why the Harmony driver would insist upon using one, then.
IIRC, the original DMA support code allowed non PCI devices to ask for a
"Fake PCI device" so they could use pci_map_single() and friends.
LASI (ergo Harmony) does live on systems with U2/Uturn in them.
This can be properly implemented in 2.6 by rewriting harmony driver
to use interfaces described in Documentation/DMA-API.txt.
This shouldn't be a big job since the interface is basically
the same as DMA-mapping.txt but uses "struct device" instead
of "struct pci_dev".
hth,
grant
Reply to: