On Mon, May 30, 2005 at 11:16:51AM +0100, Phil Blundell wrote:
> On Fri, 2005-05-20 at 16:33 -0700, Erik Hovland wrote:
> > That is what this patch does for the 2.4 kernel. It backports the
> > read_cpuid macro and takes out the CONFIG_PXA_HWUART config item but
> > otherwise does nothing else to Catalin's original work. This means that
> > the patch does indeed have the interrupt skip kludge to deal with the
> > abomination that is drivers/char/serial.c in 2.4. Which means Catalin's
> > caveats for his patch apply for this patch. See Catalin's post here:
> > http://thread.gmane.org/gmane.comp.handhelds.linux.h5400/98
> >
> > Let me know what you think. This patch is meant to be a companion to my
> > previously submitted bluez update patch but the patches don't depend on
> > each other.
>
> This patch scares me a bit,
It scares me too.
> because it seems to contain a pile of changes to the core exception
> handling and irq code. I'm slightly reluctant to apply that kind of
> thing to what's meant to be a stable kernel branch. Are all those
> changes really necessary?
According to Catalin they are. It is my understanding that the changes
are just fixes from -rmk7-pxa2 that were required to remove hard coding
of interrupt numbers. Normally one would not need to do this. But the
serial driver required Catalin to use an HWUART iterrupt of 1 instead of
0 and so the interrupt skip had to be incorporated into all of the core
exception handling irq code to deal with it. It would have been nice if
we could fix drivers/char/serial.c. But RMK already did that in 2.6. So
this was simpler for Catalin to get going.
I am fine with this patch waiting for some more testing before it is
applied. Especially since you just reduced the size of patches I have
been tracking.
Thanks
E
-- Erik Hovland mail: erik_at_hovland.org web: http://hovland.org/ PGP/GPG public key available on requestReceived on Tue May 31 2005 - 00:15:44 EDT
This archive was generated by hypermail 2.2.0 : Mon Jul 25 2005 - 17:20:11 EDT