On Tue, 2007-10-23 at 10:52 -0500, Hollis Blanchard wrote:
> On Tue, 2007-10-23 at 10:45 -0500, Jerone Young wrote:
> > On Tue, 2007-10-23 at 10:21 -0500, Hollis Blanchard wrote:
> > > > diff --git a/user/config-x86-common.mak b/user/config-x86-common.mak
> > > > new file mode 100644
> > > > --- /dev/null
> > > > +++ b/user/config-x86-common.mak
> > > > @@ -0,0 +1,16 @@
> > > > +#This is a make file with common rules for both x86 & x86-64
> > > > +
> > > > +all: kvmctl libkvm.a flatfiles
> > > > +
> > > > +kvmctl: LDFLAGS += -pthread -lrt
> > > > +
> > > > +kvmctl: kvmctl.o main.o
> > > > +
> > > > +libkvm.a: kvmctl.o
> > > > +       $(AR) rcs $@ $^
> > > > +
> > > > +balloon_ctl: balloon_ctl.o
> > > > +
> > > > +flatfiles-common = test/bootstrap test/vmexit.flat test/smp.flat
> > > > +
> > > > +flatfiles: $(flatfiles-common) $(flatfiles)
> > > > diff --git a/user/config-x86_64.mak b/user/config-x86_64.mak
> > > > --- a/user/config-x86_64.mak
> > > > +++ b/user/config-x86_64.mak
> > > > @@ -3,5 +3,9 @@ bits = 64
> > > >  bits = 64
> > > >  ldarch = elf64-x86-64
> > > >  CFLAGS += -m64
> > > > +CFLAGS += -D__x86_64__
> > > > +CFLAGS += -I $(KERNELDIR)/include
> > > > 
> > > >  flatfiles = test/access.flat test/irq.flat test/sieve.flat 
> > > > test/simple.flat test/stringio.flat test/memtest1.flat
> > > > +
> > > > +include config-x86-common.mak
> > > 
> > > I think you're clobbering flatfiles in config-x86-common.mak .
> > 
> > PATCH 2 moves all the rest of the flatfile rules to
> > config-x86-common.mak. So nothing gets clobbered.
> 
> But THIS patch clobbers it.

Oh, It does not clobber flatfiles. It uses it. So in
config-x86-common.mak there is a rule

flatfiles: $(flatfiles-common) $(flatfiles)

Acutally I thought the same too. But this is actually in the original
makefile. It doesn't clobber it apparently. But for saftey I'll change
the name of this rule as it shouldn't be the same name as variable.



> 
> > > Philosophically, I don't think that targets like "all" should be moved
> > > into arch-specific makefiles. Also, the kvmctl and libkvm.a targets
> > > should both be in the common Makefile.
> > 
> > So "all" is going to be different for everyone. Since all the test are
> > not common, and all will not use flat files.
> > 
> > Also kvmctl ad libkvm.a will be diffrenent for everyone. But this could
> > also be made into a variable. With new split code (soon to be coming to
> > a kvm-devel list near you). Here is an example of how it will look
> > later. The names may change but you get the idea.
> > 
> > Examples x86:
> >    kvmctl: kvmctl-x86.o kvmctl.o main-x86.o main.o
> > 
> >    libkvm.a: kvmctl.o kvmctl-x86.o
> > 
> >    This split may also grow more .. as maybe the apic stuff will get
> > moved to it's own file.
> > 
> > 
> > Example Powerpc:
> >    kvmctl: kvmctl-powerpc.o kvmctl.o main.o main-x86.o
> > 
> > 
> > Now one way could be is to create a variable in the arch specific
> > makefiles for kvmctl & libkvm . So that for x86  you woud have
> > 
> > KVMCTL= kvmctl-x86.o kvmctl.o main-x86.o main.o
> > 
> > Then in the main Makefile.
> > 
> > kvmctl: $(KVMCTL)
> 
> Exactly: that way we can ensure that every arch builds a common set of
> targets.
> 
> > I personally think having the declaration in each arch makes things less
> > confusing.
> 
> I can't tell if you're agreeing or if you forgot a negative. I agree
> that each arch should define e.g. "kvmctl_objs".

Well I guess I'm doing both :-). Either way will work. Having a variable
"kvmctl_objs" may be a better way. I'll redo and send again.

> 


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
kvm-devel mailing list
kvm-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/kvm-devel

Reply via email to