Subject: Re: Endian bugs in 1.4_ALPHA ep driver
To: Jason Thorpe <thorpej@nas.nasa.gov>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: current-users
Date: 04/13/1999 13:00:54
On  Tue, 13 Apr 1999 12:54:47 -0700,
 Jason Thorpe <thorpej@nas.nasa.gov> writes:


>On Tue, 13 Apr 1999 12:40:26 -0700 
> Jonathan Stone <jonathan@DSG.Stanford.EDU> wrote:
>
> > The changelog in Hayakawa-san's CardBus patches state that, in 1.3I,
> > the elink3 driver (ep) has an endian bug on macppc.
[...]


> > Is the patch below from Tsubai-san a good way to fix this for 1.4?  If
> > not, what's a better way?

>Yes, it is.

> > And, where is the endian-ness of a bus_space_{read,write}_multi_N
> > operation defined? :->.

>"In the spec for the bus."  Basically, on a big-endian machine writing to a
>a natively little-endian bus (like ISA or PCI), non-stream methods must
>ensure the the appropriate byte-swapping occurs.  Post-1.4, the API spec
>for bus_space(9) will be updated to formally include the stream methods,
>and a few other things.

Ok. That's what I guessed; thanks for confirming it.

I'll commit a fix and ask for pullups for this patch and (when it's
confirmed to work) the ep shutdownhook bugs.