On Fri, Feb 27, 2009 at 12:13:03PM +0100, Oliver Falk wrote:
Dave Jones wrote:
> On Thu, Feb 26, 2009 at 10:37:27PM +0100, Oliver Falk wrote:
> > * Update Makefile.config to generate *alpha.config and
> *alpha-smp.config
I haven't attached this part, since it's quite clear how it will look
like...
> > * Remove alphaev56 from spec, since nobody will support this subarch
> > * Add a alpha section (%ifarch alpha), defining all the with_* macros
> > - We/I should maybe rework this part...
> > * Fix alpha target, and kernel_image*
See attached specfile diffs (diff -cp).
Notes:
* buildid change will of course not go to the cvs
* The with_* defines, might be better!? Suggestions are welcome!
* Source100 and Source101 for config-alpha* is OK for you?
* Alpha-specific patches that are commented out will be removed before
commit!
* Changelog will be updated to give more information.
> > * Add config-alpha-generic and config-alpha-smp
>
> ok
Attached as config-alpha-generic-$DIST
config-alpha-smp contains only the following two lines:
CONFIG_SMP=y
CONFIG_NR_CPUS=16
I know you're going to tell me that I have to rework the
config-alpha-generic to use the config-generic.
And yes, I'll do that as the next step, but for now, may I check it in
as it is?
> > * And the current patches for F-9:
> > - linux-2.6.28-alpha-exec_range.patch
> > + Add execshield dummy functions for alpha
>
> just fold this into the regular execshield diff
No problem. Since it's a diff against alpha/include/asm/pgalloc.h, I can
easily integrate this hunk into the execshield.patch.
> > - linux-2.6-alpha-pci_get_bus_and_slot.patch
> > + include pci_get_bus_and_slot for alpha
>
> upstreamable?
Yes. Also one of the next steps. :-)
> > - linux-2.6-alpha-eepro100-cleanup.patch
> > + cleanup extraneous "freeing mc frame" message from driver
> > + This is obsolete (with >= 2.6.29) AFAIK, driver dropped!?
> > + Since new driver (e100) shows the same problem, Jay E. will try
> > to fix the new one and/or get in contact with upstream
> > maintainer
>
> yeah, e100 should support everything that eepro100 did now, and
> do a better job at it too.
Well, e100 has problems on alpha, that's why we used to patch eepro100.
However. I guess we can fix this together with upstream. Jay, you
volunteered!? :-)
> > - linux-2.6.28-alpha-pci.h.patch
> > - linux-2.6-alpha-pci.c.patch
> > + Platform support for /proc/bus/pci/X/Y mmap()s
> > + define HAVE_PCI_MMAP
> > + extern int pci_mmap_page_range
>
> also upstreamable?
Yes, of course! If Jay hasn't already done that, we really need to.
> > I can send patches to fedora-kernel-list tomorrow, when I'm back to
> > office. Also a diff to the current F-9 and F-10 spec.
>
> cool. I'll add your acl. It'll take a while to propagate anyway.
Thx a lot Dave for your trust!
I'm not going to commit anything until I have your *GO*...
This all looks fine to me. Sorry to have been so blunt, but I'm fairly
new to Fedora, so I didn't know you were actually working on stuff, and
not just someone asking for random commit access.
I wouldn't worry too much about the linux-2.6- namespace for patches,
I'd prefer if they were just alpha-$patch.patch. davej, thoughts?
regards, Kyle