[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: alpha boot-floppies

Thanks for kind reply.

Adam Di Carlo wrote:
> You've eliminated all the alpha subarchitectures.  That would be nice
> if it were true but I'm afraid it's not.

Right. We have to include all alpha sub-architecture in final stage.
The reason why I limited to 'generic' is just that I was not quite sure
about how various sub-architectures' images are made and
there seem not to be 'generic' in there.
Just adding 'generic' may be correct.

I really like to listen to alpha folks how sub-architecres' images ware made 
automatically in case of slink, especially for generic.

> >@@ -320,7 +321,9 @@
> > 			root.tar.gz
> > else
> > ifeq ($(architecture), alpha)
> >-	$(MAKE) $(alpha_rescue_targets) $(alpha_drivers_targets) root1440.bin
> >+#	$(MAKE) $(alpha_rescue_targets) $(alpha_drivers_targets) root1440.bin
> >+	$(MAKE) $(alpha_rescue_targets) $(alpha_drivers_targets)
> >+#  we can not make root1444.bin for now, because of too large root.bin.
> > else
> > 	echo unsupported architecture: $(architecture); false
> > endif
> We need to solve this problem, not hack-around it in the top level
> makefile.

But we can test root.bin not with FD but with CD-ROM for alpha.
I think we had better to go commenting out 'root1440.bin' for now,
and when root.bin comes under 1.44M, we add root144.bin again.

This way is more efficient than we wait for root.bin to come
under 1.44M, because someone(including me) can start to test
the correctness of root.bin of alpha just now.

> Ditto -- 5400 is too big to actually fit on a floppy.

Yes, not appropriate for floppy now.
But if we could reduce size of root1440.bin, this should be different from
that of i386. Second reason is above.
So leaving 5400 is right, I think.

> This is completely unnecessary.  On line 119 above we make all these
> devices already.


Ok, then I am waiting for someones help to know how 'generic' image was 
made in slink. Any infomations/comment are welcom about this.

At the same time I am going to look into how libc can be reduced,
and will submit revised patch later.

I am happy if acceptable part of my patch is applied.


Tadayoshi Ohkuma / tad@omoikane.co.jp
Omoikane Inc. / http://www.omoikane.co.jp

Reply to: