[Qemu-devel] Re: [PATCH 04/15] Improve vm_stop reason declarations

[Marcelo Tosatti]

On Wed, Feb 09, 2011 at 09:07:01AM +0100, Jan Kiszka wrote:
> On 2011-02-08 19:59, Marcelo Tosatti wrote:
> > On Mon, Feb 07, 2011 at 12:19:15PM +0100, Jan Kiszka wrote:
> >> index d6556c9..3397566 100644
> >> --- a/gdbstub.c
> >> +++ b/gdbstub.c
> >> @@ -2194,14 +2194,14 @@ static void gdb_vm_state_change(void *opaque, int 
> >> running, int reason)
> >>      const char *type;
> >>      int ret;
> >>  
> >> -    if (running || (reason != EXCP_DEBUG && reason != EXCP_INTERRUPT) ||
> >> -        s->state == RS_INACTIVE || s->state == RS_SYSCALL)
> >> +    if (running || (reason != VMSTOP_DEBUG && reason != VMSTOP_INTERRUPT) 
> >> ||
> >> +        s->state == RS_INACTIVE || s->state == RS_SYSCALL) {
> >>          return;
> > 
> > What about VMSTOP_USER ?
> > 
> > VMSTOP_INTERRUPT -> "the VM is stopped by an interrupt".
> > 
> > VMSTOP_USER -> "the VM is stopped by the user".
> 
> Makes a lot of sense, will change.
> 
> > 
> >> diff --git a/migration.c b/migration.c
> >> index 3612572..20ea113 100644
> >> --- a/migration.c
> >> +++ b/migration.c
> >> @@ -378,7 +378,7 @@ void migrate_fd_put_ready(void *opaque)
> >>          int old_vm_running = vm_running;
> >>  
> >>          DPRINTF("done iterating\n");
> >> -        vm_stop(0);
> >> +        vm_stop(VMSTOP_RWSTATE);
> > 
> > VMSTOP_RWSTATE is cryptic. What about VMSTOP_SAVEVM, MIGRATE, etc.
> 
> Both alternatives are not completely accurate as we also stop of vmload.
> What about VMSTOP_SYNC_STATE or UPDATE_STATE?

IMO it would be better to state the specific reason why the vm stopped,
not some generic action. VMSTOP_SAVEVM, VMSTOP_VMLOAD and
VMSTOP_MIGRATE.