Re: bus_dmamap_sync() for uhci(4)

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Sat, Jun 14, 2008 at 05:00:47AM +0900, Izumi Tsutsui wrote:
> macallan%NetBSD.org@localhost wrote:
> 
> > This works better, devices are found and I don't see those 'generic  
> > HBA error's anymore.
> > Reads seem to work fine, both with uhci and ehci. Writes however  
> > don't seem to work right. Symptoms:
> 
> "offs += len" is needed after "buf += len" in uhci_device_isoc_enter()?

AFAIK it's there, line 2658. I see it it the patch I posted too ...
Or did you mean something else ?

> 
> BTW, patch for ehci.c doesn't seem to have PREREAD/PREWRITE ops
> for xfer->dmabuf.

it's done in ehci_alloc_sqtd_chain() (on dma, we have
usb_dma_t *dma = &xfer->dmabuf in local variable declarations).
There's no usb_syncmem() in ehci_root*, because it's an emulation and
all is filled in by software.

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--