See patch. See also http://www.coreboot.org/OpenFirmware for the beginnings of a HOWTO page for manual builds (needs lots of work). Uwe. -- http://www.hermann-uwe.de | http://www.holsham-traders.de http://www.crazy-hacks.org | http://www.unmaintained-free-software.org
Uwe Hermann wrote:
See patch.
See also http://www.coreboot.org/OpenFirmware for the beginnings of a HOWTO page for manual builds (needs lots of work).
Uwe.
Initial experimental support for OpenFirmware as payload.
As the OFW image is quite big, increase QEMU's BIOS size to 512 KB.
Tested in QEMU only so far. There are some issues with screen display, you should test it with '-serial stdio' in QEMU.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de> Acked-by: Stefan Reinauer <stepan@coresystems.de>
-- coresystems GmbH • Brahmsstr. 16 • D-79104 Freiburg i. Br. Tel.: +49 761 7668825 • Fax: +49 761 7664613 Email: info@coresystems.de • http://www.coresystems.de/ Registergericht: Amtsgericht Freiburg • HRB 7656 Geschäftsführer: Stefan Reinauer • Ust-IdNr.: DE245674866
On Thu, Apr 24, 2008 at 12:45:27PM +0200, Stefan Reinauer wrote:
Acked-by: Stefan Reinauer <stepan@coresystems.de>
Thanks, r157. Uwe. -- http://www.hermann-uwe.de | http://www.holsham-traders.de http://www.crazy-hacks.org | http://www.unmaintained-free-software.org
Index: packages/coreboot-v2/patches/buildrom_qemu_size.patch =================================================================== --- packages/coreboot-v2/patches/buildrom_qemu_size.patch (Revision 0) +++ packages/coreboot-v2/patches/buildrom_qemu_size.patch (Revision 0) @@ -0,0 +1,14 @@ +Index: targets/emulation/qemu-x86/Config.lb +=================================================================== +--- a/argets/emulation/qemu-x86/Config.lb (Revision 3244) ++++ a/targets/emulation/qemu-x86/Config.lb (Arbeitskopie) +@@ -3,7 +3,8 @@ + target qemu-x86 + mainboard emulation/qemu-x86 + +-option ROM_SIZE=256*1024 ++# option ROM_SIZE=256*1024 ++option ROM_SIZE=512*1024 +
I prefer using a different Config file to patching the Config file. I think it makes it a cleaner design. It also makes it so that the ROM size doesn't increase unless you use a larger payload. I know that's less important for QEMU, but... See the Tyan boards for examples. They have Config.lb and Config-lab.lb for larger payloads. Myles
buildrom can be fragile on the clean. I ^C in the middle of it wget'ing OFW (long story) and now: [rminnich@xcpu buildrom-devel]$ make clean maCleaning coreboot v3... ke Cleaning Openfirmware... make: *** [ofw-clean] Error 2 [rminnich@xcpu buildrom-devel]$ make Fetching OpenFirmware... SVN Checkout rev 720 Fetching svn://openbios.org/openfirmware... ^C^C^C^C^Z [1]+ Stopped make [rminnich@xcpu buildrom-devel]$ kill -9 %1 [1]+ Killed make [rminnich@xcpu buildrom-devel]$ [rminnich@xcpu buildrom-devel]$ [rminnich@xcpu buildrom-devel]$ make menuconfig # # using defaults found in .config # .config:93:warning: symbol value '' invalid for LAB_PAUSE *** End of buildrom configuration. *** Execute 'make' to build buildrom or try 'make help'. [rminnich@xcpu buildrom-devel]$ make clean all Cleaning coreboot v3... Cleaning Openfirmware... make: *** [ofw-clean] Error 2 [rminnich@xcpu buildrom-devel]$ mak eclean bash: mak: command not found [rminnich@xcpu buildrom-devel]$ make clean Cleaning coreboot v3... Cleaning Openfirmware... make: *** [ofw-clean] Error 2 [rminnich@xcpu buildrom-devel]$ [rminnich@xcpu buildrom-devel]$ make Patching OpenFirmware... Applying patch patches/ofw_coreboot_qemu.patch The text leading up to this was: -------------------------- |Index: cpu/x86/pc/biosload/config.fth |=================================================================== |--- a/cpu/x86/pc/biosload/config.fth (Revision 720) |+++ b/cpu/x86/pc/biosload/config.fth (Arbeitskopie) -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored Patch patches/ofw_coreboot_qemu.patch does not apply (enforce with -f) make: *** [/home/rminnich/src/bios/buildrom/buildrom-devel/work/ofw/stamps/.patched] Error 1 [rminnich@xcpu buildrom-devel]$ make distclean [rminnich@xcpu buildrom-devel]$ make clean Cleaning coreboot v3... make: *** [coreboot-v3-clean] Error 2 it's just all kindsa confused and I can't get it undone. Maybe a 'fuxored' target that would do what needs to be done, without regard to the state? distclean seems not to help. ron
Try rm -rf work/ofw rm -rf sources/ofw rm sources/openfi* Myles
On 24/04/08 10:47 -0700, ron minnich wrote:
buildrom can be fragile on the clean. I ^C in the middle of it wget'ing OFW (long story) and now: [rminnich@xcpu buildrom-devel]$ make clean maCleaning coreboot v3... ke Cleaning Openfirmware... make: *** [ofw-clean] Error 2 [rminnich@xcpu buildrom-devel]$ make Fetching OpenFirmware... SVN Checkout rev 720 Fetching svn://openbios.org/openfirmware... ^C^C^C^C^Z [1]+ Stopped make [rminnich@xcpu buildrom-devel]$ kill -9 %1 [1]+ Killed make [rminnich@xcpu buildrom-devel]$ [rminnich@xcpu buildrom-devel]$ [rminnich@xcpu buildrom-devel]$ make menuconfig # # using defaults found in .config # .config:93:warning: symbol value '' invalid for LAB_PAUSE
*** End of buildrom configuration. *** Execute 'make' to build buildrom or try 'make help'.
[rminnich@xcpu buildrom-devel]$ make clean all Cleaning coreboot v3... Cleaning Openfirmware... make: *** [ofw-clean] Error 2 [rminnich@xcpu buildrom-devel]$ mak eclean bash: mak: command not found [rminnich@xcpu buildrom-devel]$ make clean Cleaning coreboot v3... Cleaning Openfirmware... make: *** [ofw-clean] Error 2 [rminnich@xcpu buildrom-devel]$
[rminnich@xcpu buildrom-devel]$ make Patching OpenFirmware... Applying patch patches/ofw_coreboot_qemu.patch The text leading up to this was: -------------------------- |Index: cpu/x86/pc/biosload/config.fth |=================================================================== |--- a/cpu/x86/pc/biosload/config.fth (Revision 720) |+++ b/cpu/x86/pc/biosload/config.fth (Arbeitskopie) -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored Patch patches/ofw_coreboot_qemu.patch does not apply (enforce with -f) make: *** [/home/rminnich/src/bios/buildrom/buildrom-devel/work/ofw/stamps/.patched] Error 1 [rminnich@xcpu buildrom-devel]$ make distclean [rminnich@xcpu buildrom-devel]$ make clean Cleaning coreboot v3... make: *** [coreboot-v3-clean] Error 2
it's just all kindsa confused and I can't get it undone. Maybe a 'fuxored' target that would do what needs to be done, without regard to the state?
distclean seems not to help.
Distclean actually should have what you needed. That last error message is an issue, but easily fixed. DIstclean should always recover you - try make distclean; make - it should do the right thing. Jordan -- Jordan Crouse Systems Software Development Engineer Advanced Micro Devices, Inc.
On Thu, Apr 24, 2008 at 11:00 AM, Jordan Crouse <jordan.crouse@amd.com> wrote:
Distclean actually should have what you needed. That last error message is an issue, but easily fixed. DIstclean should always recover you - try make distclean; make - it should do the right thing.
no joy. I did an rm -rf and an svn co, this should work ron
On 24/04/08 10:55 -0700, ron minnich wrote:
On Thu, Apr 24, 2008 at 11:00 AM, Jordan Crouse <jordan.crouse@amd.com> wrote:
Distclean actually should have what you needed. That last error message is an issue, but easily fixed. DIstclean should always recover you - try make distclean; make - it should do the right thing.
no joy.
I did an rm -rf and an svn co, this should work
What exactly is breaking?
ron
-- Jordan Crouse Systems Software Development Engineer Advanced Micro Devices, Inc.
On Thu, Apr 24, 2008 at 11:05 AM, Jordan Crouse <jordan.crouse@amd.com> wrote:
What exactly is breaking?
my guess: I interrupted the wget or untar of ofw. The patch was failing (file did not exist) make clean failed: things were not there make distclean failed: it would do the distclean, but then a make would fail in coreboot. One other idea: make run-qemu given that you had built with qemu it would run it for you. Or is that there and I don't know it :-) ron
On 24/04/08 11:01 -0700, ron minnich wrote:
On Thu, Apr 24, 2008 at 11:05 AM, Jordan Crouse <jordan.crouse@amd.com> wrote:
What exactly is breaking?
my guess: I interrupted the wget or untar of ofw. The patch was failing (file did not exist)
If you broke the wget, thats bad. rm -rf the source and start over. I don't know any way to cleanly handle this - unless we start recording MD5SUMS somehow.
make clean failed: things were not there This is a known issue. Don't worry about it. I'll fix it.
make distclean failed: it would do the distclean, but then a make would fail in coreboot.
Logs! We need logs!
One other idea: make run-qemu
Hmm - okay, I could buy into this. wget the VGA BIOS and do the right thing. Jordan -- Jordan Crouse Systems Software Development Engineer Advanced Micro Devices, Inc.
participants (5)
-
Jordan Crouse -
Myles Watson -
ron minnich -
Stefan Reinauer -
Uwe Hermann