[openfirmware] Troubleshooting VGA under qemu? [patch]

[James Cameron]

Summary: probe in .bth stops build?

On Thu, Oct 01, 2015 at 10:36:24PM +0000, Marcin Cieslak wrote:
> On Fri, 2 Oct 2015, James Cameron wrote:
> 
> > Thanks.
> > 
> > Some quick thoughts;
> > 
> > Suggest borrowing my-w@ and my-w! from other drivers instead
> 
> done

Looks fine.

> > Several other drivers also impose a symmetry and undo in
> > unmap-regs
> 
> done

Looks fine.

> > Is needed?
> 
> not sure, it works. Only right now qemu reports 4x more PCI
> remappings done
> 
> > > +: set-reg-property [...]
> > 
> > Most other drivers set this property at load time outside the
> > colon definitions; is there a reason for doing it after map-regs?
> 
> That's a whole point here - /fb and /fifo are taken from the mapped
> registers.
> 
> Some other video drivers have a more sophisticated machinery to
> construct that property - encode-reg-property, >reg-props

Thanks.

> Unfortunately I cannot test it on a real "hardware", i.e. under
> VMware :)

Nor I.

> [...]
> Index: dev/video/vmsvga.bth
> ===================================================================
> --- dev/video/vmsvga.bth	(revision 3784)
> +++ dev/video/vmsvga.bth	(working copy)
> @@ -12,6 +12,7 @@
>  " display" device-name
>  
>  fload ${BP}/dev/video/controlr/vmsvga.fth	\ VMware-specific words
> +probe
>  
>  end0

Is this the right place for probe?  The build now pauses in the
builder;

host:~/openfirmware$ (cd cpu/x86/pc/emu/build; make clean)
rm -f *.dic *.log headers *~ *.elf *.di *.img builton.fth build *.c32 *.rom
host:~/openfirmware$ (cd cpu/x86/pc/emu/build; make)
./build emuofw.rom
--- Rebuilding rmreset.img
--- Cmd: ${HOSTDIR}/forth ${HOSTDIR}/../build/builder.dic ../rmreset.bth
--- Rebuilding pmreset.di
--- Cmd: ${HOSTDIR}/forth ${HOSTDIR}/../build/builder.dic ../pmreset.bth
--- Rebuilding paging.di
--- Cmd: ${HOSTDIR}/forth ${HOSTDIR}/../build/builder.dic ../../paging.bth
--- Rebuilding fw.img
--- Cmd: ${HOSTDIR}/x86forth ${BP}/cpu/x86/build/basefw.dic ../fw.bth
Page Fault
ok ftrace
@                 Called from fdavail?                  at  5575ac58
        0
fdavail?          Called from fclose                    at  5575b09c
  868278c
execute           Called from interpret-do-defined      at  5575804c
do-defined        Called from $compile                  at  55757ff0
$compile          Called from (interpret                at  557582b4
interpret         Called from interpret-lines           at  5575ce08
   Catch frame - SP: 55869058   my-self: 0   handler: 55868880 
catch             Called from include-file              at  5575cf3c
       47
  8685880
       44
  86826f8
        5
        0
include-file      Called from (included)                at  5575d064
   Catch frame - SP: 55869070   my-self: 0   handler: 55868878 
catch             Called from included                  at  5575d08c
included          Called from process-argument          at  5575d5d0
   Catch frame - SP: 558166b8   my-self: 0   handler: 55868878 
catch             Called from process-command-line      at  5575d630
process-command-line  Called from cold                  at  5575b6b4
ok 

-- 
James Cameron
http://quozl.linux.org.au/