[Qemu-devel] Running Plan9

[Qemu-devel] Running Plan9

[Brian Campbell]

Being a little curious about Bell Lab's Plan 9 I tried to get the
installation/demo CD running under qemu.  The major stumbling block I
encountered was that the IDE emulation wasn't properly detected.  A
little bit of debugging revealed that Plan 9 was relying on an
an IDE feature that hasn't been implemented in qemu.

Adding some very shoddy (and almost certainly wrong) support for
WIN_DIAGNOSE to ide_ioport_write in hw/ide.c was enough to get it to boot
and run the live version.  Is anyone likely to provide a proper
implementation?  I'd do it myself, but the (draft) ATA specs I had a look
through appeared to have some oddities I'd rather not have to deal with,
given that I've never done any ATA work before.

-- 
Brian Campbell

Re: [Qemu-devel] Running Plan9

[Lars Munch]

On Wed, Dec 01, 2004 at 10:32:53PM +0000, Brian Campbell wrote:
> Being a little curious about Bell Lab's Plan 9 I tried to get the
> installation/demo CD running under qemu.  The major stumbling block I
> encountered was that the IDE emulation wasn't properly detected.  A
> little bit of debugging revealed that Plan 9 was relying on an
> an IDE feature that hasn't been implemented in qemu.
> 
> Adding some very shoddy (and almost certainly wrong) support for
> WIN_DIAGNOSE to ide_ioport_write in hw/ide.c was enough to get it to boot
> and run the live version.  Is anyone likely to provide a proper
> implementation?  I'd do it myself, but the (draft) ATA specs I had a look
> through appeared to have some oddities I'd rather not have to deal with,
> given that I've never done any ATA work before.

I had the same problem when I tried to get the ide driver going in 
RTEMS, since it also depends on WIN_DIAGNOSE. I also added quick version
of the WIN_DIAGNOSE command (just setting error register to 01h) and
RTEMS found the disc.

I just looked at the ATA specs. What oddities are you referring to?

Regards
Lars Munch

Re: [Qemu-devel] Running Plan9

[Fabrice Bellard]

It should work now!

Fabrice.

Brian Campbell wrote:
> Being a little curious about Bell Lab's Plan 9 I tried to get the
> installation/demo CD running under qemu.  The major stumbling block I
> encountered was that the IDE emulation wasn't properly detected.  A
> little bit of debugging revealed that Plan 9 was relying on an
> an IDE feature that hasn't been implemented in qemu.
> 
> Adding some very shoddy (and almost certainly wrong) support for
> WIN_DIAGNOSE to ide_ioport_write in hw/ide.c was enough to get it to boot
> and run the live version.  Is anyone likely to provide a proper
> implementation?  I'd do it myself, but the (draft) ATA specs I had a look
> through appeared to have some oddities I'd rather not have to deal with,
> given that I've never done any ATA work before.
> 

Re: [Qemu-devel] Running Plan9

[Brian Campbell]

On Thu, Dec 02, 2004 at 10:51:26AM +0100, Lars Munch wrote:
> On Wed, Dec 01, 2004 at 10:32:53PM +0000, Brian Campbell wrote:
> > 
> > Adding some very shoddy (and almost certainly wrong) support for
> > WIN_DIAGNOSE to ide_ioport_write in hw/ide.c was enough to get it to boot
> > and run the live version.  Is anyone likely to provide a proper
> > implementation?  I'd do it myself, but the (draft) ATA specs I had a look
> > through appeared to have some oddities I'd rather not have to deal with,
> > given that I've never done any ATA work before.
> 
> I had the same problem when I tried to get the ide driver going in 
> RTEMS, since it also depends on WIN_DIAGNOSE. I also added quick version
> of the WIN_DIAGNOSE command (just setting error register to 01h) and
> RTEMS found the disc.
> 
> I just looked at the ATA specs. What oddities are you referring to?

The draft I looked at (d1410r3b ATA/ATAPI-6 revision 3b) has a section
on the "EXECUTE DEVICE DIAGNOSTIC command protocol", with various state
diagrams and timing information.  The problem is that I've no idea if
this stuff would even be visible from the software's point of view, and
if it is, whether it's worth emulating it.

Cheers,

Brian

Re: [Qemu-devel] Running Plan9

[Brian Campbell]

On Thu, Dec 02, 2004 at 09:18:27PM +0100, Fabrice Bellard wrote:
> It should work now!

Works fine.  Thanks.

Re: [Qemu-devel] Running Plan9

[Lars Munch]

Thanks Fabrice, but I still have some problems with this command.

Why is READY not set? According to the specs:

"g) Device 0 sets the DRDY bit when ready to accept any command."

aren't we ready to accept any command at this point?

Another note, should't there be a check for device 1 present?

Thanks
Lars Munch

On Thu, Dec 02, 2004 at 09:18:27PM +0100, Fabrice Bellard wrote:
> It should work now!
> 
> Fabrice.
> 
> Brian Campbell wrote:
> >Being a little curious about Bell Lab's Plan 9 I tried to get the
> >installation/demo CD running under qemu.  The major stumbling block I
> >encountered was that the IDE emulation wasn't properly detected.  A
> >little bit of debugging revealed that Plan 9 was relying on an
> >an IDE feature that hasn't been implemented in qemu.
> >
> >Adding some very shoddy (and almost certainly wrong) support for
> >WIN_DIAGNOSE to ide_ioport_write in hw/ide.c was enough to get it to boot
> >and run the live version.  Is anyone likely to provide a proper
> >implementation?  I'd do it myself, but the (draft) ATA specs I had a look
> >through appeared to have some oddities I'd rather not have to deal with,
> >given that I've never done any ATA work before.
> >
> 
> 
> 
> 
> _______________________________________________
> Qemu-devel mailing list
> Qemu-devel@nongnu.org
> http://lists.nongnu.org/mailman/listinfo/qemu-devel
>