OpenBIOS December 2009

openbios@openbios.org

19 participants
44 discussions

r645 - trunk/openbios-devel/config/scripts
by svn＠openbios.org 18 Dec '09

18 Dec '09

Author: laurent Date: 2009-12-18 21:41:51 +0100 (Fri, 18 Dec 2009) New Revision: 645 Modified: trunk/openbios-devel/config/scripts/switch-arch Log: I usually compile openbios on PowerPC for PowerPC. Apparently I'm the only onedoing that, as that (really easy) case got broken by the cross compilation overhaul. Let's allow host = target gcc configurations again. Signed-off-by: Alexander Graf <agraf(a)suse.de> Modified: trunk/openbios-devel/config/scripts/switch-arch =================================================================== --- trunk/openbios-devel/config/scripts/switch-arch 2009-12-10 00:23:02 UTC (rev 644) +++ trunk/openbios-devel/config/scripts/switch-arch 2009-12-18 20:41:51 UTC (rev 645) @@ -93,6 +93,9 @@ return fi done + if [ "$ARCH" = "$HOSTARCH" ]; then + return + fi echo "ERROR: no ${1} cross-compiler found !" 1>&2 exit 1 }

1 0

Re: [OpenBIOS] r640 - trunk/openbios-devel/arch/sparc64
by Andreas Färber 15 Dec '09

15 Dec '09

Hi, Just updated to this revision from r636 and did a `make`: Building OpenBIOS for ppc sparc32 sparc64 Building...error: GEN openbios.dict undefined word. make[1]: *** [openbios.dict] Error 1 error: GEN openbios.dict undefined word. make[1]: *** [openbios.dict] Error 1 error: GEN openbios.dict undefined word. make[1]: *** [openbios.dict] Error 1 make: *** [build] Error 1 After `make clean` it worked fine. Could there be something wrong with the dependency modelling? In obj-sparc32/rules.mak I found this line: bootstrap-DICTIONARY:=$(bootstrap-DICTIONARY) $(SRCDIR)/forth/ bootstrap/start.fs But I believe bootstrap.dict should also depend on $(SRCDIR)/forth/ bootstrap/bootstrap.fs, which was modified in r638. After a switch- arch, it is then rebuilt before openbios.dict. See attached patch. However, there are some more .fs files in the forth/bootstrap folder that are not referenced in build.xml. Should all such files be added to the build.xml file in their folder, or are they / should they be referenced from somewhere else? Andreas

2 2

Re: [OpenBIOS] Starting QEMU with OpenBIOS for Sparc32
by gabriel omar batistuta 11 Dec '09

11 Dec '09

Thanks for your reply the problem is as follows: Boot hangs after the following console output Loading a.out image... Loaded 7483136 bytes entry point is 0x4000 Jumping to entry point... Host operating system:Windows XP SP2. Version Qemu: Qemu manager for windows V6.0 (qemu 0.9.1). Guest operating system : Solaris 2.4.

1 0

Re: [OpenBIOS] [PATCH] pci: enable io/memory unconditionally
by Laurent Vivier 10 Dec '09

10 Dec '09

>On Thu, Dec 10, 2009 at 01:09:34PM +0100, Laurent Vivier wrote: >> >On Thu, Dec 10, 2009 at 01:20:23AM +0100, Laurent Vivier wrote: >> >> Le mercredi 09 décembre 2009 à 17:53 +0200, Michael S. Tsirkin a écrit : >> >> > VGA adapters need to claim memory and i/o >> >> > transactions even if they do not have any >> >> > i/o or memory bars. E.g. PCI spec, page 297, >> >> > gives an example of such a device: >> >> > >> >> > Programming interface 0000 0000b >> >> > VGA-compatible controller. Memory >> >> > addresses 0A 0000h through 0B >> >> > FFFFh. I/O addresses 3B0h to 3BBh >> >> > and 3C0h to 3DFh and all aliases of >> >> > these addresses. >> >> > >> >> > While bios could check for these devices and special-case them, it is >> >> > easier to fix this by enabling i/o and memory space unconditionally: >> >> > devices that do not support it will just ignore this setting. >> >> > >> >> > Signed-off-by: Michael S. Tsirkin <mst(a)redhat.com> >> >> >> >> Applied, but could you give me an use case where previous behavior >> >> breaks something ? >> >> >> >> Regards, >> >> Laurent >> > >> >qemu uses openbios and implements cirrus vga emulation, >> >which has this device class and no IO bars. >> >> qemu-system-ppc doesn't support cirrus vga emulation (only vga emulation), >moreover openbios is not able to manage cirrus vga card. > >Maybe it will support it in the future :). Are you working on this ? >> So why this modification is needed ? >> (even if it is correct according to the specs, it's why I've applied it) > >It's not fixing any breakage I have observed, if that's what you mean. >I found the issue in bochs bios and seabios with >cirrus vga emulation, so I thought I'd fix up openbios as well. OK, it's fine for me. Thank you. Laurent -- --------------------- Laurent(a)vivier.eu --------------------- "Tout ce qui est impossible reste à accomplir" Jules Verne "Things are only impossible until they're not" Jean-Luc Picard

2 1

Re: [OpenBIOS] [PATCH] pci: enable io/memory unconditionally
by Laurent Vivier 10 Dec '09

10 Dec '09

>On Thu, Dec 10, 2009 at 01:20:23AM +0100, Laurent Vivier wrote: >> Le mercredi 09 décembre 2009 à 17:53 +0200, Michael S. Tsirkin a écrit : >> > VGA adapters need to claim memory and i/o >> > transactions even if they do not have any >> > i/o or memory bars. E.g. PCI spec, page 297, >> > gives an example of such a device: >> > >> > Programming interface 0000 0000b >> > VGA-compatible controller. Memory >> > addresses 0A 0000h through 0B >> > FFFFh. I/O addresses 3B0h to 3BBh >> > and 3C0h to 3DFh and all aliases of >> > these addresses. >> > >> > While bios could check for these devices and special-case them, it is >> > easier to fix this by enabling i/o and memory space unconditionally: >> > devices that do not support it will just ignore this setting. >> > >> > Signed-off-by: Michael S. Tsirkin <mst(a)redhat.com> >> >> Applied, but could you give me an use case where previous behavior >> breaks something ? >> >> Regards, >> Laurent > >qemu uses openbios and implements cirrus vga emulation, >which has this device class and no IO bars. qemu-system-ppc doesn't support cirrus vga emulation (only vga emulation), moreover openbios is not able to manage cirrus vga card. So why this modification is needed ? (even if it is correct according to the specs, it's why I've applied it) Regards, Laurent -- --------------------- Laurent(a)vivier.eu --------------------- "Tout ce qui est impossible reste à accomplir" Jules Verne "Things are only impossible until they're not" Jean-Luc Picard

2 1

[PATCH] pci: enable io/memory unconditionally
by Michael S. Tsirkin 10 Dec '09

10 Dec '09

VGA adapters need to claim memory and i/o transactions even if they do not have any i/o or memory bars. E.g. PCI spec, page 297, gives an example of such a device: Programming interface 0000 0000b VGA-compatible controller. Memory addresses 0A 0000h through 0B FFFFh. I/O addresses 3B0h to 3BBh and 3C0h to 3DFh and all aliases of these addresses. While bios could check for these devices and special-case them, it is easier to fix this by enabling i/o and memory space unconditionally: devices that do not support it will just ignore this setting. Signed-off-by: Michael S. Tsirkin <mst(a)redhat.com> --- Pls Cc me on responses, I'm not on the list. Thanks! Index: drivers/pci.c =================================================================== --- drivers/pci.c (revision 641) +++ drivers/pci.c (working copy) @@ -732,10 +732,6 @@ base = *io_base; min_align = 1 << 7; amask = 0x00000001; - pci_config_write16(addr, PCI_COMMAND, - pci_config_read16(addr, - PCI_COMMAND) | - PCI_COMMAND_IO); } else { /* Memory Space */ base = *mem_base; @@ -744,10 +740,6 @@ if (reg == 6) { smask |= 1; /* ROM */ } - pci_config_write16(addr, PCI_COMMAND, - pci_config_read16(addr, - PCI_COMMAND) | - PCI_COMMAND_MEMORY); } *p_omask = smask & amask; smask &= ~amask; @@ -788,6 +780,7 @@ { uint32_t omask; + uint16_t cmd; int reg; pci_addr config_addr; @@ -807,6 +800,9 @@ ob_pci_configure_bar(addr, config, reg, config_addr, &omask, mem_base, io_base); } + cmd = pci_config_read16(addr, PCI_COMMAND); + cmd |= PCI_COMMAND_IO | PCI_COMMAND_MEMORY; + pci_config_write16(addr, PCI_COMMAND, cmd); } static void ob_scan_pci_bus(int bus, unsigned long *mem_base,

2 2

r644 - trunk/openbios-devel/config/scripts
by svn＠openbios.org 10 Dec '09

10 Dec '09

Author: laurent Date: 2009-12-10 01:23:02 +0100 (Thu, 10 Dec 2009) New Revision: 644 Modified: trunk/openbios-devel/config/scripts/switch-arch Log: Autodetect cross-compiler name. Signed-off-by: Laurent Vivier <Laurent(a)vivier.eu> Modified: trunk/openbios-devel/config/scripts/switch-arch =================================================================== --- trunk/openbios-devel/config/scripts/switch-arch 2009-12-10 00:17:17 UTC (rev 643) +++ trunk/openbios-devel/config/scripts/switch-arch 2009-12-10 00:23:02 UTC (rev 644) @@ -84,6 +84,19 @@ -e "s/Power Macintosh/ppc/"` } +select_prefix() +{ + for TARGET in ${1}-linux-gnu- ${1}-elf- ${1}-eabi- + do + if type -p ${TARGET}gcc > /dev/null + then + return + fi + done + echo "ERROR: no ${1} cross-compiler found !" 1>&2 + exit 1 +} + if ! test -f utils/dist/debian/rules; then echo "switch-arch must be run from the top-level source directory" >&2 exit 1 @@ -110,31 +123,31 @@ ARCH=`echo $RULES_ARCH | sed s/cross-//g` case $ARCH in amd64) - TARGET="x86_64-linux-gnu-" + select_prefix x86_64 CFLAGS="-fno-builtin" AS_FLAGS= ;; ppc) - TARGET="powerpc-linux-gnu-" + select_prefix powerpc CFLAGS="-msoft-float -fno-builtin-bcopy -fno-builtin-log2" AS_FLAGS= ;; sparc32) - TARGET="sparc-elf-" + select_prefix sparc CFLAGS="-Wa,-xarch=v8 -Wa,-32 -m32 -mcpu=supersparc -fno-builtin" AS_FLAGS="-Wa,-xarch=v8 -Wa,-32" ;; sparc64) - TARGET="sparc64-elf-" + select_prefix sparc64 CFLAGS="-Wa,-xarch=v9b -Wa,-64 -m64 -mcpu=ultrasparc -mcmodel=medany -fno-builtin" AS_FLAGS="-Wa,-xarch=v9b -Wa,-64" ;; x86) - TARGET="i486-linux-gnu-" + select_prefix i486 CFLAGS="-fno-builtin -m32" AS_FLAGS="-Wa,-32" ;;

1 0

r643 - trunk/openbios-devel/drivers
by svn＠openbios.org 10 Dec '09

10 Dec '09

Author: laurent Date: 2009-12-10 01:17:17 +0100 (Thu, 10 Dec 2009) New Revision: 643 Modified: trunk/openbios-devel/drivers/pci.c Log: VGA adapters need to claim memory and i/o transactions even if they do not have any i/o or memory bars. E.g. PCI spec, page 297, gives an example of such a device: Programming interface 0000 0000b VGA-compatible controller. Memory addresses 0A 0000h through 0B FFFFh. I/O addresses 3B0h to 3BBh and 3C0h to 3DFh and all aliases of these addresses. While bios could check for these devices and special-case them, it is easier to fix this by enabling i/o and memory space unconditionally: devices that do not support it will just ignore this setting. Signed-off-by: Michael S. Tsirkin <mst(a)redhat.com> Modified: trunk/openbios-devel/drivers/pci.c =================================================================== --- trunk/openbios-devel/drivers/pci.c 2009-12-09 18:44:01 UTC (rev 642) +++ trunk/openbios-devel/drivers/pci.c 2009-12-10 00:17:17 UTC (rev 643) @@ -732,10 +732,6 @@ base = *io_base; min_align = 1 << 7; amask = 0x00000001; - pci_config_write16(addr, PCI_COMMAND, - pci_config_read16(addr, - PCI_COMMAND) | - PCI_COMMAND_IO); } else { /* Memory Space */ base = *mem_base; @@ -744,10 +740,6 @@ if (reg == 6) { smask |= 1; /* ROM */ } - pci_config_write16(addr, PCI_COMMAND, - pci_config_read16(addr, - PCI_COMMAND) | - PCI_COMMAND_MEMORY); } *p_omask = smask & amask; smask &= ~amask; @@ -788,6 +780,7 @@ { uint32_t omask; + uint16_t cmd; int reg; pci_addr config_addr; @@ -807,6 +800,9 @@ ob_pci_configure_bar(addr, config, reg, config_addr, &omask, mem_base, io_base); } + cmd = pci_config_read16(addr, PCI_COMMAND); + cmd |= PCI_COMMAND_IO | PCI_COMMAND_MEMORY; + pci_config_write16(addr, PCI_COMMAND, cmd); } static void ob_scan_pci_bus(int bus, unsigned long *mem_base,

1 0

Starting QEMU with OpenBIOS for Sparc32
by gabriel omar batistuta 09 Dec '09

09 Dec '09

Can someone with the following question? I'm trying to use Qemu to load a solaris 2.4 on an x86 machine running a Windows XP. I know how to configure both OpenBoot Qemu as to emulate the sistems sparc32. >From already thank you very much.

2 1

Milax and the alarm word
by Mark Cave-Ayland 09 Dec '09

09 Dec '09

Hi all, With the previous commit, I seem to be getting to the point where files are now being read from the ISO into memory. However, on Milax things are still failing because the Fcode tries to set up a callback function repeated every second using the "alarm" word. Does SPARC64 have a way of counting elapsed time (some kind of clock tick register) so that it would be possible to measure this in the main loop? Or alternatively set up some kind of timer interrupt after X ms? Incidentally I notice that the latest Solaris code doesn't seem to use alarm anymore; Nick - does your Nevada boot get any further with current OpenBIOS SVN? ATB, Mark. -- Mark Cave-Ayland - Senior Technical Architect PostgreSQL - PostGIS Sirius Corporation plc - control through freedom http://www.siriusit.co.uk t: +44 870 608 0063 Sirius Labs: http://www.siriusit.co.uk/labs

4 12