October 2015 - openfirmware - mail.coreboot.org

Fix for /pci/isa@1/ide-controller@i1f0/ide@0/<Unnamed>@0
by Marcin Cieslak Oct. 22, 2015

Oct. 22, 2015

The following patch improves ISA IDE discovery. On a modern system those devices will probably be the same as under pci-ide tree of devices. * Discover up to 4 IDE devices * Fix "Unimplemented package interface procedure" message during "Probe IDE" Before patch: show-devs /pci/isa@1/ide-controller@i1f0/ide@1 /pci/isa@1/ide-controller@i1f0/ide@0 /pci/isa@1/ide-controller@i1f0/ide@0/<Unnamed>@0 ok select /pci/isa@1/ide-controller@i1f0/ide@0/<Unnamed>@0 Invalid ihandle After patch: /pci/isa@1/ide-controller@i1f0/ide@1 /pci/isa@1/ide-controller@i1f0/ide@0 /pci/isa@1/ide-controller@i1f0/ide@1/cdrom@2 /pci/isa@1/ide-controller@i1f0/ide@0/disk@1 /pci/isa@1/ide-controller@i1f0/ide@0/disk@0 Open issues: ok select /pci/isa@1/ide-controller@i1f0/ide@1/cdrom@2 IDE data error: 0 Can't open deblocker package ok It does not matter whether Qemu's CDROM device is attached or not. (For /pci/pci-ide@1,1/ide@1/cdrom@0 works when attached). What is file /dev/ide/node.fc binary? Marcin Index: dev/ide/idenode.fth =================================================================== --- dev/ide/idenode.fth (revision 3784) +++ dev/ide/idenode.fth (working copy) @@ -22,6 +22,7 @@ : any-blocks? " any-blocks?" $call-parent ; : cdrom? " cdrom?" $call-parent ; +: node-fcode " node-fcode" $call-parent ; : set-address ( unit -- dummy unit ) @@ -65,7 +66,7 @@ finish-device ( ) ; -: probe ( -- ) 2 0 do i probe-child loop ; +: probe ( -- ) 3 0 do i probe-child loop ; headers \ LICENSE_BEGIN

2 3

[commit] r3786 - dev/ide
by repository service Oct. 22, 2015

Oct. 22, 2015

Author: quozl Date: Thu Oct 22 03:13:19 2015 New Revision: 3786 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/3786 Log: IDE: fix probe message, from Marcin Cieslak Fix "Unimplemented package interface procedure" message during "Probe IDE" Modified: dev/ide/idenode.fth Modified: dev/ide/idenode.fth ============================================================================== --- dev/ide/idenode.fth Thu Oct 22 03:08:55 2015 (r3785) +++ dev/ide/idenode.fth Thu Oct 22 03:13:19 2015 (r3786) @@ -22,6 +22,7 @@ : any-blocks? " any-blocks?" $call-parent ; : cdrom? " cdrom?" $call-parent ; +: node-fcode " node-fcode" $call-parent ; : set-address ( unit -- dummy unit )

1 0

Troubleshooting VGA under qemu?
by Marcin Cieslak Oct. 22, 2015

Oct. 22, 2015

Hello, I have tried running ofw without BIOS (as per http://www.openfirmware.info/Building_OFW_for_QEMU) and it seems that VGA support does not work. I am using config-virtualbox.fth as advised. I have uncommented "console-serial" in config.fth to be able to interact with the console and run: qemu-system-i386 -m 32 -bios emuofw.rom -hda fat:. -serial stdio (this qemu 2.4.0 on FreeBSD/amd64) "show-devs" reports (this is with video.fc included) ok show-devs /cpus /dropin-fs /flash@fff80000 /pci /dropin-fs /mmu /memory@0 /aliases /options /openprom /chosen /packages /cpus/cpu@0 /pci/ethernet@3 /pci/display@2 /pci/pci8086,7113@1,3 /pci/pci-ide@1,1 /pci/host@0 /pci/isa@1 /pci/pci-ide@1,1/ide@1 /pci/pci-ide@1,1/ide@0 /pci/pci-ide@1,1/ide@1/cdrom@0 /pci/pci-ide@1,1/ide@0/disk@0 /pci/isa@1/ide-controller@i1f0 /pci/isa@1/sound@i220 /pci/isa@1/rtc@i70 /pci/isa@1/8042@i60 /pci/isa@1/fdc@i3f0 /pci/isa@1/parallel@i378 /pci/isa@1/serial@i2f8 /pci/isa@1/serial@i3f8 /pci/isa@1/timer@i40 /pci/isa@1/interrupt-controller@i20 /pci/isa@1/dma-controller@i00 /pci/isa@1/ide-controller@i1f0/ide@1 /pci/isa@1/ide-controller@i1f0/ide@0 /pci/isa@1/ide-controller@i1f0/ide@0/<Unnamed>@0 /pci/isa@1/8042@i60/mouse@aux /pci/isa@1/8042@i60/keyboard@kbd /pci/isa@1/fdc@i3f0/disk /packages/http /packages/tcp ... I have tried uncommenting in: ${BP}/dev/video/build/video.fc in cpu/x86/pc/emu/emuofw.bth as well as uncommenting create use-vga in cpu/x86/pc/emu/config.fth, none of which improves the situation. When I switch consoles in qemu the VGA screen says "Guest has not initialized the display (yet)" Any hints to troubleshoot this? Additional questions: 1) Would that be possible to have a dual-console mode (similar to what FreeBSD early bootloader has) so that both vga and serial console are active? ofw prints messages on the serial console even if disabled, only input is not working. 2) Initially I thought ofw is completely dead and I have run it under gdb, only to figure out I am running in the Forth inner interpreter: 0xff9fc000: lods %ds:(%esi),%eax => 0xff9fc001: jmp *(%eax) Is there any way to have "symbol table" to map word addresses to the word names? (I figured out I can tick the word to get its address). Thank you, Marcin

2 20

[commit] r3785 - in dev: . video video/controlr
by repository service Oct. 22, 2015

Oct. 22, 2015

Author: quozl Date: Thu Oct 22 03:08:55 2015 New Revision: 3785 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/3785 Log: Emulator ROM build - fix VMWare VGA adapter, from Marcin Cieslak Send command to the PCI command port to inform the device that I/O port has been enabled. Issue a warning during PCI resouce allocation if "reg" property has not been provided (with diag-switch? true). Generate "reg" property for the PCI resource allocator. Known to work with qemu-2.4.0. Does not rewrite all changed rectangles on qemu-2.0.0 Modified: dev/pcibus.fth dev/video/controlr/vmsvga.fth dev/video/vmsvga.bth Modified: dev/pcibus.fth ============================================================================== --- dev/pcibus.fth Thu Oct 1 09:17:25 2015 (r3784) +++ dev/pcibus.fth Thu Oct 22 03:08:55 2015 (r3785) @@ -1183,7 +1183,8 @@ 4 ( increment ) then ( increment ) +loop - + else + " reg property missing!" diag-type-cr then r> finish-aa-property ; Modified: dev/video/controlr/vmsvga.fth ============================================================================== --- dev/video/controlr/vmsvga.fth Thu Oct 1 09:17:25 2015 (r3784) +++ dev/video/controlr/vmsvga.fth Thu Oct 22 03:08:55 2015 (r3785) @@ -39,6 +39,9 @@ : reg@ ( index -- value ) regs rl! regs 1+ rl@ ; : reg! ( value index -- ) regs rl! regs 1+ rl! ; +: my-w@ ( offset -- w ) my-space + " config-w@" $call-parent ; +: my-w! ( w offset -- ) my-space + " config-w!" $call-parent ; + \ Here are the register numbers. Most of these registers are \ accessed only once, so I don't define access words for many of them. \ (W) means that it is meaningful for you to write to this register @@ -85,6 +88,7 @@ h# 200.0000 instance value /mem : map-regs ( -- ) 0 0 my-space h# 0100.0010 + h# 10 " map-in" $call-parent to regs + 4 my-w@ 1 or 4 my-w! ; : map-mem ( -- ) my-space h# 14 + " config-l@" $call-parent if @@ -94,12 +98,16 @@ 0 0 my-space h# 0200.0018 + /fb " map-in" $call-parent to frame-buffer-adr /fb 0 my-space h# 0200.0018 + /fifo " map-in" $call-parent to fifo then - 3 my-space h# 04 + " config-w!" $call-parent + 4 my-w@ 2 or 4 my-w! +; +: unmap-regs ( -- ) + regs h# 10 " map-out" $call-parent + 4 my-w@ 1 invert and 4 my-w! ; -: unmap-regs ( -- ) regs h# 10 " map-out" $call-parent ; : unmap-mem ( -- ) fifo /fifo " map-out" $call-parent frame-buffer-adr /fb " map-out" $call-parent + 4 my-w@ 2 invert and 4 my-w! ; \ Min and Max are the static limits of the FIFO area. @@ -116,6 +124,23 @@ unmap-regs abort \ We don't support version 0 ; + +: int+ ( adr len n -- adr' len' ) encode-int encode+ ; + +: set-reg-property + my-address my-space encode-phys 0 int+ h# 0 int+ + my-address my-space h# 0100.0010 + encode-phys encode+ 0 int+ h# 10 int+ + my-address my-space h# 0200.0014 + encode-phys encode+ 0 int+ /fb int+ + my-address my-space h# 0200.0018 + encode-phys encode+ 0 int+ /fifo int+ + " reg" property +; + +: probe ( -- ) + map-regs + set-reg-property + unmap-regs +; + : init-fb ( -- ) depth 7 reg! 7 reg@ depth <> if 7 reg@ to depth then Modified: dev/video/vmsvga.bth ============================================================================== --- dev/video/vmsvga.bth Thu Oct 1 09:17:25 2015 (r3784) +++ dev/video/vmsvga.bth Thu Oct 22 03:08:55 2015 (r3785) @@ -12,6 +12,7 @@ " display" device-name fload ${BP}/dev/video/controlr/vmsvga.fth \ VMware-specific words +probe end0

1 0

gmane.comp.boot-loaders.openfirmware
by Marcin Cieslak Oct. 5, 2015

Oct. 5, 2015

We are now on gmane: http://news.gmane.org/gmane.comp.boot-loaders.openfirmware Also the archive has been imported. Marcin

1 0

[commit] r3784 - cpu/x86/pc/emu
by repository service Oct. 1, 2015

Oct. 1, 2015

Author: quozl Date: Thu Oct 1 07:17:25 2015 New Revision: 3784 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/3784 Log: Emulator ROM build - added Bochs driver to the dropin set, from Marcin Cieslak Modified: cpu/x86/pc/emu/emuofw.bth Modified: cpu/x86/pc/emu/emuofw.bth ============================================================================== --- cpu/x86/pc/emu/emuofw.bth Thu Oct 1 06:36:31 2015 (r3783) +++ cpu/x86/pc/emu/emuofw.bth Thu Oct 1 07:17:25 2015 (r3784) @@ -25,6 +25,7 @@ in: ${BP}/dev/video/build/vmsvga.fc in: ${BP}/dev/ne2000/build/ne2kpci.fc in: ${BP}/dev/amd79970/build/vmlance.fc +in: ${BP}/dev/video/build/bga.fc \ in: ${BP}/dev/video/build/video.fc in: ${BP}/cpu/x86/pc/biosload/linux.fth @@ -58,6 +59,7 @@ \ Just use a driver targeted at the specific chipset that QEMU emulates " ${BP}/dev/video/build/cirrus.fc" " pci1013,b8" $add-deflated-dropin " ${BP}/dev/video/build/vmsvga.fc" " pci15ad,405" $add-deflated-dropin + " ${BP}/dev/video/build/bga.fc" " pci1234,1111" $add-deflated-dropin " ${BP}/dev/ne2000/build/ne2kpci.fc" " pci10ec,8029" $add-deflated-dropin " ${BP}/dev/amd79970/build/vmlance.fc" " pci1022,2000" $add-deflated-dropin

1 0

Copy paste bug in der! (PowerPC)
by Marcin Cieslak Oct. 1, 2015

Oct. 1, 2015

There is a copy+paste bug: Index: cpu/ppc/spr.fth =================================================================== --- cpu/ppc/spr.fth (wersja 3782) +++ cpu/ppc/spr.fth (kopia robocza) @@ -426,7 +426,7 @@ mfspr tos,der c; code der! ( n -- ) - mtspr immr,tos + mtspr der,tos lwz tos,0(sp) addi sp,sp,1cell c;

2 1

[commit] r3783 - cpu/ppc
by repository service Oct. 1, 2015

Oct. 1, 2015

Author: quozl Date: Thu Oct 1 06:36:31 2015 New Revision: 3783 URL: http://tracker.coreboot.org/trac/openfirmware/changeset/3783 Log: PPC - der! writes to wrong register, der! is unused elsewhere in Open Firmware, reported by Marcin Cieslak, reviewed by James Cameron. Modified: cpu/ppc/spr.fth Modified: cpu/ppc/spr.fth ============================================================================== --- cpu/ppc/spr.fth Wed Sep 30 07:07:04 2015 (r3782) +++ cpu/ppc/spr.fth Thu Oct 1 06:36:31 2015 (r3783) @@ -426,7 +426,7 @@ mfspr tos,der c; code der! ( n -- ) - mtspr immr,tos + mtspr der,tos lwz tos,0(sp) addi sp,sp,1cell c;

1 0