Re: CVS commit: src/sys/arch/i386/include

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

In article <200309101712.TAA0000170528%zel459.zel.kfa-juelich.de@localhost>,
Matthias Drochner <M.Drochner%fz-juelich.de@localhost> wrote:
>
>christos%NetBSD.org@localhost said:
>> XXX1: We need to fix the vm86 api. 
>
>What's wrong with it?
>As you see in regress/.../vm86, one can live with it.
>One thing which is strange is that we have a si_trap
>which noone else has -- I'd expect that value in si_code.
No, si_code is reserved for MI stuff so we cannot use this. Read the
new manpage for siginfo...

The vm.substr.vmsc.sc_* stuff refers to the old struct sigcontext.
It should be changed to use __gregset_t, and vmsc should do the same.
i386_vm86 needs to be versioned if that happens, so I did not do it (yet).
This is what I meant by saying we need to fix the vm86 api.

christos