On Wed, 12 Feb 1997 19:02:00 +0100 (MET) 
 Matthias Drochner <drochner@zelz28.zel.kfa-juelich.de> wrote:

 > 
 > >Description:
 > 	In sys/dev/isa/if_le_isa.c, function depca_isa_probe(), the second
 > argument to bus_space_unmap() is the physical memory address, not the
 > handle as
 > returned by bus_space_map() before. For the i386, this causes a panic
 > in extent_free()
 > because the pmap_extract() in bus_space_unmap() returned a "0".

"oops!"  Fixed.

 > Possible / future bugs:
 > To check if this bug or similar ones occur in other drivers too, I made
 > bus_space_handle_t
 > more abstract and compiled a "GENERIC" kernel. The following compiler
 > warnings appeared:
 > -sys/dev/isa/sb_isa.c: similar bug in sb_isa_match() (but without
 > consequences up to now
 >  because handles and addresses are identical for ISA on i386)

Fixed that one, too.

 > -sys/i386/isa/if_fea.c: in pdq_eisa_devinit(), arithmetics with handles

I've fixed this, too... that one just sort of slipped by me...

 > -sys/i386/isa/isapnp_machdep.c: in isapnp_map() and isapnp_map_readport(),
 > use of
 >  integer constants as handles

This is OK ... this is machine-dependent code, which is allowed to know
about the guts.

Thanks for pointing these out!

Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939