Re: CVS commit: src/sys/rump

[Antti Kantee <pooka%netbsd.org@localhost>]

On Sat Dec 13 2008 at 13:05:23 +0000, David Laight wrote:
> > Add a few MD ifdefs into the rump pmap stubs to enable us to get
> > rid of the local pmap.h and unify the pmap ABI with the kernel.
> > (XXX: compiles but doesn't work on vax)
> 
> This doesn't seem to build on amd64.
> 
> In file included from ./i386/pmap.h:318,
>      from ./machine/pmap.h:348,
>      from src/lib/librump/../../sys/rump/../uvm/uvm_pmap.h:92,
>      from src/lib/librump/../../sys/rump/../uvm/uvm_extern.h:474,
>      from src/lib/librump/../../sys/rump/../sys/vnode.h:73,
>      from src/lib/librump/../../sys/rump/librump/rumpkern/emul.c:35:
> 
> ./x86/pmap.h: In function 'vtopte':
> ./x86/pmap.h:320: warning: integer constant is too large for 'long' type
> ./x86/pmap.h:320: warning: comparison is always true due to limited range of 
> data type
> ./x86/pmap.h: In function 'kvtopte':
> ./x86/pmap.h:330: warning: integer constant is too large for 'long' type
> ./x86/pmap.h:330: warning: comparison is always false due to limited range of 
> data type
> 
> nbmake: stopped in /bsd/current/src/compat/lib/librump
> 
> The problems seem to be with:
>         KASSERT(va < VM_MIN_KERNEL_ADDRESS);

Nobody expects the libcompat assertion!

Can you cvs up sys/arch/amd64/include/vmparam.h and try again (it will
take a while for my build to get that far)?