Re: merge for pciide at cardbus?

[KIYOHARA Takashi <kiyohara%kk.iij4u.or.jp@localhost>]

Hi!


I put pciide_cardbus that had done some corrections.

ftp://ftp.netbsd.org/pub/NetBSD/misc/kiyohara/tmp/pciide_cardbus_support-20080912.diff
ftp://ftp.netbsd.org/pub/NetBSD/misc/kiyohara/tmp/satalink_cardbus-20080912.diff


However, I have not tested this enough yet.  Can anyone test?  ;-)


From: Manuel Bouyer <bouyer%antioche.eu.org@localhost>
Date: Sat, 20 Sep 2008 21:27:19 +0200

> On Sun, Sep 21, 2008 at 02:33:20AM +0900, KIYOHARA Takashi wrote:
> > Hi! Bouyer,
> > 
> > 
> > I'd like to my satalink at cardbus patch but I'm still wondering
> > if its design is okay or not:
> > http://mail-index.netbsd.org/current-users/2008/08/18/msg004039.html
> > i.e. if it's okay to put cardbus specific stuff into pciide_softc.
> > 
> > If you have any particular opinion about this, please let me know.
> > Otherwise I'll commit it in a few days.
> 
> If I understand it properly, sc_cmdsize, sc_ctlsize, and sc_dma_size
> are needed if we want to support proper detach on PCI too. If so, 
> pciide_common.c should be fixed to use these softc members instead of
> local on stack variable.
> 
> I think the specific cardbus_chip_map can be avoided, by using a union
> (either for the 2 chip_map pointers, or for the second parameter of
> *chip_map()).
> 
> then you could use:
> struct cardbus_pciide_softc {
>       struct pciide_softc sc_pidesc;
>       cardbus_devfunc_t       sc_ct;          /* Cardbus registers info */
>       cardbus_chipset_tag_t   sc_cc;
>       cardbus_function_tag_t  sc_cf;
>       cardbustag_t            sc_ctag;
> };
> and extending pciide_softc with cardbus infos.


Thanks,
--
kiyohara