On Tuesday 08 August 2006 17:20, Tobias Diedrich wrote:
> Michael Buesch wrote:
> > On Tuesday 08 August 2006 13:16, Tobias Diedrich wrote:
> > > Hi,
> > > 
> > > I'm trying to get the bcm43xx_d80211 driver to work on my wrt54g and
> > > have so far managed to patch the wireless-dev driver into the 2.6.17
> > > buildroot-ng tree, but I get the an oops (data bus error).
> > > I _guess_ the driver is doing an unaligned access?
> > 
> > Yes, the driver is not safe on archs which require get_unaligned(),
> > but is this the case for MIPS? As far as I remember that was not
> > the case, but I am not sure. I always though get_unaligned() was
> > a no-op there, too.
> 
> Hmm, at least for userspace it is handled by
> arch/mips/kernel/unaligned.c
> See also http://pramode-ce.livejournal.com/22594.html
> So I guess get_unaligned is not a no-op.
> 
> Interestingly a simple
> "if (offset & 1) printk(KERN_ERR "read16: unaligned access\n");"
> in bcm43xx_read16() doesn't seem to trigger, so maybe it really is
> a different problem?

We are talking about two different things.
I am talking about unaligned memory access and you are talking
about unaligned MMIO access.
I don't think we do unaligned MMIO access somewhere.

> bcm43xx_sprom_read seems to be the same code in both driver
> versions...

Please insert printks to find out which line is oopsing.

-- 
Greetings Michael.
_______________________________________________
Bcm43xx-dev mailing list
[email protected]
http://bat.berlios.de/mailman/listinfo/bcm43xx-dev

Reply via email to