Re: CVS commit: src/sys/arch/alpha/alpha

[christos%astron.com@localhost (Christos Zoulas)]

In article <1125541666.969130.709.nullmailer%yamt.dyndns.org@localhost>,
YAMAMOTO Takashi  <yamt%mwd.biglobe.ne.jp@localhost> wrote:
>> > he%NetBSD.org@localhost said:
>> > > Modified Files:
>> > >  src/sys/arch/alpha/alpha: trap.c
>> > > Log Message: Now that uvm_extern.h no longer includes uvm_fault.h, we
>> > > need to include the latter explicitly. 
>> >
>> > I have the impression that the VM_FAULT_INVALID and
>> > VM_FAULT_PROTECT flags are not used anywhere and
>> > can be removed completely.
>> 
>> Well, a grep and test-compile says otherwise:
>
>i think what he meant is that there is no point to distinguish them.
>
>as it seems that there are some code to rely on uvm-internal decls
>(eg. uvm_fault_wire), i put back uvm_fault.h for now.
>
>YAMAMOTO Takashi
>
Aren't those needed for:

/* SIGSEGV */
#define SEGV_MAPERR     1       /* Address not mapped to object         */
#define SEGV_ACCERR     2       /* Invalid permissions for mapped object*/

/* SIGBUS */
#define BUS_ADRALN      1       /* Invalid address alignment            */
#define BUS_ADRERR      2       /* Non-existent physical address        */
#define BUS_OBJERR      3       /* Object specific hardware error       */

christos