Re: [Qemu-devel] [RFC PATCH] Distinguish between reset types

[David Woodhouse]

On Wed, 2013-02-20 at 12:12 -0600, Anthony Liguori wrote:
> The above implements a CPU-only soft reset that should fix the problem
> you're having with PAM resetting unconditionally.  If it works, I'll
> fixup the other PC callers of reset too.

To be clear: that wasn't my problem :)

If people want to run OS/2, VDISK or whatever else needs to get back
from protected mode to real mode with a CPU reset but *without* a full
system reset, it's going to be *their* problem. And I suspect they have
other problems already, so I don't feel too guilty about the final patch
in my tree at git.infradead.org/users/dwmw2/qemu.git doing the PAM reset
unconditionally. That works just fine for me.


     if (val & 4) {
-        qemu_system_reset_request();
+        if (val & 2) {
+            piix_soft_reset_request();
+        } else {
+            qemu_system_reset_request();
+        }
         return;

Those are the wrong way round, I think. If bit 1 is set, that's a hard
reset and should use qemu_system_reset_request(). If it's *not* set,
then it's supposed to be a soft reset (like a port 92 or keyboard
controller reset).

-- 
dwmw2

smime.p7s
Description: S/MIME cryptographic signature