[LinuxBIOS] Fix to vga emulator (from MSI)

[ron minnich]

This change fixes a long-standing bug, whereby we do not set ret for an
un-inited vector, which we should have done.
Signed-off-by: Ronald G. Minnich

---

Index: devices/emulator/biosemu.c
===================================================================
--- devices/emulator/biosemu.c  (revision 2470)
+++ devices/emulator/biosemu.c  (working copy)
@@ -122,6 +122,7 @@
        case 0x6D:
                if (getIntVect(num) == 0x0000) {
                        printk_debug("un-inited int vector\n");
+                       ret = 1;
                }
                if (getIntVect(num) == 0xFF065) {
                        //ret = int42_handler();

[Stefan Reinauer]

* ron minnich  [061027 17:13]:
> This change fixes a long-standing bug, whereby we do not set ret for an
> un-inited vector, which we should have done.
> Signed-off-by: Ronald G. Minnich
> Index: devices/emulator/biosemu.c
>
> --- devices/emulator/biosemu.c  (revision 2470)
> +++ devices/emulator/biosemu.c  (working copy)
> @@ -122,6 +122,7 @@
>         case 0x6D:
>                 if (getIntVect(num) == 0x0000) {

right, this is not AH/AL as I assumed earlier. what is it supposed to
be?
>                     printk_debug("un-inited int vector\n");
>                   ret = 1;
>           }
>           if (getIntVect(num) == 0xFF065) {
>                   //ret = int42_handler();

Acked-by: Stefan Reinauer

[Peter Stuge]

On Fri, Oct 27, 2006 at 07:37:15PM +0200, Stefan Reinauer wrote:
> > @@ -122,6 +122,7 @@
> >         case 0x6D:
> >                 if (getIntVect(num) == 0x0000) {
>
> right, this is not AH/AL as I assumed earlier. what is it supposed
> to be?

getIntVect() returns the real mode address of a vector AFAICT.
> >             if (getIntVect(num) == 0xFF065) {

//Peter

[Stefan Reinauer]

* Peter Stuge  [061027 20:36]:
> On Fri, Oct 27, 2006 at 07:37:15PM +0200, Stefan Reinauer wrote:
> > > @@ -122,6 +122,7 @@
> > >         case 0x6D:
> > >                 if (getIntVect(num) == 0x0000) {
> >
> > right, this is not AH/AL as I assumed earlier. what is it supposed
> > to be?
>
> getIntVect() returns the real mode address of a vector AFAICT.

so in our case this is a fantasy address, right?

[Uwe Hermann]

On Fri, Oct 27, 2006 at 09:13:04AM -0600, ron minnich wrote:
> This change fixes a long-standing bug, whereby we do not set ret for an
> un-inited vector, which we should have done.
> Signed-off-by: Ronald G. Minnich
> Index: devices/emulator/biosemu.c
>
> --- devices/emulator/biosemu.c  (revision 2470)
> +++ devices/emulator/biosemu.c  (working copy)
> @@ -122,6 +122,7 @@
>        case 0x6D:
>                if (getIntVect(num) == 0x0000) {
>                        printk_debug("un-inited int vector\n");
>                   ret = 1;
>          }
>          if (getIntVect(num) == 0xFF065) {
>                  //ret = int42_handler();

I think this is fixed and obsolete now since the MSI patch got
committed, right?

Uwe.

[Stefan Reinauer]

* Uwe Hermann  [061102 22:36]:
> On Fri, Oct 27, 2006 at 09:13:04AM -0600, ron minnich wrote:
> > This change fixes a long-standing bug, whereby we do not set ret for an
> > un-inited vector, which we should have done.
> > Signed-off-by: Ronald G. Minnich
>
> I think this is fixed and obsolete now since the MSI patch got
> committed, right?

This got committed ad 2479.

http://snapshots.linuxbios.org/stats/commit-LinuxBIOSv2-2479.log