Hello,
I want to build the openbios image for use in coreboot, but it fails. I am running a 32bit debian 8.10 vm.
$ make /usr/bin/xsltproc Building OpenBIOS for x86 Building... error: AR libfs.a CC target/libgcc/ashldi3.o CC target/libgcc/ashrdi3.o CC target/libgcc/__lshrdi3.o CC target/libgcc/__divdi3.o CC target/libgcc/__udivdi3.o CC target/libgcc/__udivmoddi4.o CC target/libgcc/__umoddi3.o AR libgcc.a LINK openbios.multiboot libdrivers.a(pci.o): In function `ob_pci_unmap': /tmp/openbios/drivers/pci.c:405: undefined reference to `ofmem_unmap' rules.mak:235: recipe for target 'openbios.multiboot' failed make[1]: *** [openbios.multiboot] Error 1 make[1]: Leaving directory '/tmp/openbios/obj-x86' Makefile:19: recipe for target 'build' failed make: *** [build] Error 1
if i select amd64 on a 64bit machine it works without a problem.
What is the problem here?
Thanks Jens
On 03/01/18 00:25, mr.jb@gmx.de wrote:
Hello,
I want to build the openbios image for use in coreboot, but it fails. I am running a 32bit debian 8.10 vm.
$ make /usr/bin/xsltproc Building OpenBIOS for x86 Building... error: AR libfs.a CC target/libgcc/ashldi3.o CC target/libgcc/ashrdi3.o CC target/libgcc/__lshrdi3.o CC target/libgcc/__divdi3.o CC target/libgcc/__udivdi3.o CC target/libgcc/__udivmoddi4.o CC target/libgcc/__umoddi3.o AR libgcc.a LINK openbios.multiboot libdrivers.a(pci.o): In function `ob_pci_unmap': /tmp/openbios/drivers/pci.c:405: undefined reference to `ofmem_unmap' rules.mak:235: recipe for target 'openbios.multiboot' failed make[1]: *** [openbios.multiboot] Error 1 make[1]: Leaving directory '/tmp/openbios/obj-x86' Makefile:19: recipe for target 'build' failed make: *** [build] Error 1
if i select amd64 on a 64bit machine it works without a problem.
What is the problem here?
Oops, that was me not handling the non-OFMEM PCI mapping case in the last set of updates.
The attached patch squash is enough to compile x86 for me, can you confirm that it actually boots on your setup?
ATB,
Mark.
Hi Mark,
thanks for the quick reply. WIth this patch I can build the x86 version.
I should have said that I attempt to flash coreboot the first time and have not the equipment yet to do the flashing (but its ordered). I hope that the stuff is coming in a few days so I can report if its working on real hardware.
Best Regards Jens
On 01/03/2018 10:25 AM, Mark Cave-Ayland wrote:
On 03/01/18 00:25, mr.jb@gmx.de wrote:
Hello,
I want to build the openbios image for use in coreboot, but it fails. I am running a 32bit debian 8.10 vm.
$ make /usr/bin/xsltproc Building OpenBIOS for x86 Building... error: AR libfs.a CC target/libgcc/ashldi3.o CC target/libgcc/ashrdi3.o CC target/libgcc/__lshrdi3.o CC target/libgcc/__divdi3.o CC target/libgcc/__udivdi3.o CC target/libgcc/__udivmoddi4.o CC target/libgcc/__umoddi3.o AR libgcc.a LINK openbios.multiboot libdrivers.a(pci.o): In function `ob_pci_unmap': /tmp/openbios/drivers/pci.c:405: undefined reference to `ofmem_unmap' rules.mak:235: recipe for target 'openbios.multiboot' failed make[1]: *** [openbios.multiboot] Error 1 make[1]: Leaving directory '/tmp/openbios/obj-x86' Makefile:19: recipe for target 'build' failed make: *** [build] Error 1
if i select amd64 on a 64bit machine it works without a problem.
What is the problem here?
Oops, that was me not handling the non-OFMEM PCI mapping case in the last set of updates.
The attached patch squash is enough to compile x86 for me, can you confirm that it actually boots on your setup?
ATB,
Mark.
On 04/01/18 00:29, Jens Bauernfeind wrote:
Hi Mark,
thanks for the quick reply. WIth this patch I can build the x86 version.
I should have said that I attempt to flash coreboot the first time and have not the equipment yet to do the flashing (but its ordered). I hope that the stuff is coming in a few days so I can report if its working on real hardware.
Best Regards Jens
Hi Jens,
That sounds really interesting! Sadly there isn't much x86 development happening on OpenBIOS at the moment since the current focus for x86 is on the UEFI BIOSs, however I generally try and make sure that it builds.
I was able to boot coreboot/OpenBIOS in QEMU as per the instructions at https://www.coreboot.org/QEMU#coreboot_v3_.2B_OpenBIOS, although that was a few years ago and it's not something I regularly do.
However I would expect that the basic functionality of booting into OpenBIOS with serial console/VGA driver and running an ELF executable in real mode to be working fine - if it's not then please let us know :)
ATB,
Mark.
Hi Mark,
I installed coreboot succesfully on my t430. I build openbios in the x86 version and added the openbios-bultin.elf as addtional payload to the image (with seabios as primary, so i could easily test openbios ) Unfortunately when I start it, it doesn't recognize the keyboard and several messages "cannot manage <device name>...." I know that the laptop keyboard is attached via PS/2, and in the drivers/pci.c is a PS2 Keyboard block. Is there anything I can debug?
Best Regards, Jens
Gesendet: Freitag, 05. Januar 2018 um 09:29 Uhr Von: "Mark Cave-Ayland" mark.cave-ayland@ilande.co.uk An: "Jens Bauernfeind" mr.jb@gmx.de, openbios@openbios.org Betreff: Re: [OpenBIOS] failed to build
On 04/01/18 00:29, Jens Bauernfeind wrote:
Hi Mark,
thanks for the quick reply. WIth this patch I can build the x86 version.
I should have said that I attempt to flash coreboot the first time and have not the equipment yet to do the flashing (but its ordered). I hope that the stuff is coming in a few days so I can report if its working on real hardware.
Best Regards Jens
Hi Jens,
That sounds really interesting! Sadly there isn't much x86 development happening on OpenBIOS at the moment since the current focus for x86 is on the UEFI BIOSs, however I generally try and make sure that it builds.
I was able to boot coreboot/OpenBIOS in QEMU as per the instructions at https://www.coreboot.org/QEMU#coreboot_v3_.2B_OpenBIOS, although that was a few years ago and it's not something I regularly do.
However I would expect that the basic functionality of booting into OpenBIOS with serial console/VGA driver and running an ELF executable in real mode to be working fine - if it's not then please let us know :)
ATB,
Mark.