Don't trace the buffer length in place of the property size.
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@889 f158a5a8-5612-0410-a976-696ce0be7e32
A printk statement was commented out and there were format string mismatches for ppc target.
Use OFMEM_TRACE to more easily enable/disable it and use FMT_ucellx as in other trace output.
Note that this changes the alignment from decimal to hexidecimal notation.
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@881 f158a5a8-5612-0410-a976-696ce0be7e32
Each console write caused temporary buffer allocation.
Avoid allocations by changing console_draw_str() to use Forth
string parameters, which are usually readily available.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@872 f158a5a8-5612-0410-a976-696ce0be7e32
The maximum allowed value for x axis is one less than cons.w.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@870 f158a5a8-5612-0410-a976-696ce0be7e32
is emitted by dump_service(). After successful execution the return value
and a linebreak are emitted by dump_return(). If however the method fails,
nothing is emitted, so that subsequent method calls end up in place of the
return value. Observed for getproplen with AIX.
Emit an error indication and a linebreak when DEBUG_CIF is defined.
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@868 f158a5a8-5612-0410-a976-696ce0be7e32
OFmem is not available on x86.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@860 f158a5a8-5612-0410-a976-696ce0be7e32
Can't enable Sparc32 yet, there is no ofmem_claim().
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@858 f158a5a8-5612-0410-a976-696ce0be7e32
If a loader succeeds, don't try the others.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@847 f158a5a8-5612-0410-a976-696ce0be7e32
open_ih() returns -1 on error and 0 is a valid return value.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@839 f158a5a8-5612-0410-a976-696ce0be7e32
Replace uchar, uint, ulong, u_char, u_int, u_long, u_int* with
their standard equivalents.
Fixes warnings like these on OpenBSD:
CC target/arch/unix/unix.o
In file included from ../arch/unix/unix.c:29:
../include/config.h:26: warning: redefinition of `ulong'
/usr/include/sys/types.h:56: warning: `ulong' previously declared here
../include/config.h:26: warning: redundant redeclaration of `ulong' in same scope
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@830 f158a5a8-5612-0410-a976-696ce0be7e32
Fix build failure:
LINK openbios-unix
libopenbios.a(load.o)(.text+0xe0): In function `load':
../libopenbios/load.c:58: undefined reference to `elf_boot_notes'
libopenbios.a(load.o)(.text+0xe8):../libopenbios/load.c:58: undefined reference to `elf_boot_notes'
Move elf_boot_notes variable to load.c to make it also available on
Unix build. Also fix a spurious sys_info variable definition.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@829 f158a5a8-5612-0410-a976-696ce0be7e32
IEEE-1275 spec compliance.
This patch implements the following:
1) Fix bootpath/bootargs handling so that default values are read from NVRAM, and allow multiple space-separated values to be
specified.
2) With correct bootargs handling in place, move the ELF loader over to the new libopenbios unified loaders.
3) Remove all the loader code from all architecture directories sine we don't need it anymore.
4) Simplify the boot word so it invokes platform-specific code where required, then calls load and go as per the specification.
Tested on all my available images for SPARC32, SPARC64 and PPC, and compile-tested on x86.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@828 f158a5a8-5612-0410-a976-696ce0be7e32
Move sys_info to load.c to make it also available on Unix build.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@826 f158a5a8-5612-0410-a976-696ce0be7e32
Also add a little bootscript debugging for those people that may need it.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@823 f158a5a8-5612-0410-a976-696ce0be7e32
hierarchy is correct, we can simply pass an ihandle into the loader and it will work correctly regardless of whether it is being
invoked on an entire disk, partition or individual file.
In order to test the new code, start by switching the Fcode loader over to the new infrastructure for testing on SPARC64. Note
this patch also contains a correction to load-base on SPARC which was being set to a value other than 0x4000 which is the
documented default.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@798 f158a5a8-5612-0410-a976-696ce0be7e32
- refactor scan procedure to start with PCI host controller
- initiate scan of subordinate PCI bus from PCI host and PCI-to-PCI bridges
- find out PCI subordinate bus numbers and write them to bride devices
- automated assignment of "reg", "ranges", and "bus-range" properties
Signed-off-by: Igor V. Kovalenko <igor.v.kovalenko@gmail.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@786 f158a5a8-5612-0410-a976-696ce0be7e32
different translation entries described within the OF platform bindings. With thanks to Andreas Färber and Igor Kovalenko.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@772 f158a5a8-5612-0410-a976-696ce0be7e32
to finish the code and test retention after a restart since the OpenBIOS words reset and reset-all don't seem to work at the
moment. However, it enables OpenSolaris boot to get further in the meantime.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@766 f158a5a8-5612-0410-a976-696ce0be7e32
not allocated enough memory in qemu without having to rebuild with CONFIG_DEBUG_OFMEM enabled.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@763 f158a5a8-5612-0410-a976-696ce0be7e32
Fix clang warnings:
../kernel/dict.c:289:2: warning: Value stored to 'len' is never read
len -= sizeof(dictionary_header_t);
../packages/cmdline.c:181:8: warning: Value stored to 'buf' during its
initialization is never read
char *buf = ci->buf;
../libopenbios/elf_info.c:126:2: warning: Value stored to 'name' is never read
name = addr;
../libopenbios/elf_load.c:277:5: warning: Value stored to 'addr' is never read
addr += pad;
../drivers/ide.c:209:3: warning: Value stored to 'err' is never read
err = ob_ide_pio_readb(drive, IDEREG_ERROR);
../drivers/ide.c:219:17: warning: Value stored to 'old_cdb' during its initialization is never read
unsigned char old_cdb = cmd->cdb[0];
../drivers/ide.c:222:4: warning: Value stored to 'old_cdb' is never read
old_cdb = cmd->old_cdb;
../drivers/pci.c:103:2: warning: Value stored to 'ss' is never read
ss = 0;
../drivers/pci.c:100:2: warning: Value stored to 'dev' is never read
dev = 0;
../drivers/pci.c:212:2: warning: Value stored to 'bus' is never read
bus = (hi >> 16) & 0xFF;
../drivers/pci.c:953:4: warning: Value stored to 'rev' is never read
rev = pci_config_read8(addr, PCI_REVISION_ID);
../packages/video.c:331:2: warning: Value stored to 's' is never read
s = video.fb.mphys - s;
../packages/video.c:330:2: warning: Value stored to 'size' is never read
size = ((video.fb.h * video.fb.rb + s) + 0xfff) & ~0xfff;
../fs/hfsplus/btree.c:229:5: warning: Value stored to 'p' is never read
p = btree_readhead(&bt->head, p);
../fs/hfsplus/volume.c:149:2: warning: Value stored to 'p' is never read
p = volume_readfork(p, &vh->start_file );
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@751 f158a5a8-5612-0410-a976-696ce0be7e32
issue whereby Forth starts doing strange things just after attemping a claim on physical memory resource.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@739 f158a5a8-5612-0410-a976-696ce0be7e32
returned from the CIF, but with nret set to zero then the size argument is never pushed onto the stack and so the size would be
set to whatever junk was left in the argument block. Change the code to use the size expected by the client instead.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@736 f158a5a8-5612-0410-a976-696ce0be7e32
i) Ensure that the CIF caller's idea of pb->nret is always respected, but if there is a mismatch then log a warning if DEBUG_CIF
is enabled, then correct the stack and continue.
ii) Correct the code path for call-method and interpret to ensure they ignore the extra status variable pushed onto the stack by
client-call-iface.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@733 f158a5a8-5612-0410-a976-696ce0be7e32
i) client-call-iface did not set a return value of 0 on success; hence the top stack argument was dropped as part of the status
check causing an off-by-one error.
ii) instead of setting pb->nret to be the number of arguments returned, the (random) value being passed in was being used to
control the number of arguments being returned.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@732 f158a5a8-5612-0410-a976-696ce0be7e32
accordingly. Hence since the loaders are currently written in C, it makes more sense to drop down into C and re-use the same
code rather than have a second implementation in Forth.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@726 f158a5a8-5612-0410-a976-696ce0be7e32
architectures are updated to use it. This is required to allow binding of C functions into Forth by libopenbios (i.e. cross
architecture) rather than having to update everything in every arch/*/ initialisation file.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@725 f158a5a8-5612-0410-a976-696ce0be7e32
is being set correctly in all of the loaders, and secondly we change the bootinfo loader so that the bootscript is executed as
part of init-program and not go. This seems to reflect the idea of execute-buffer in OpenBOOT that any Forth code is executed
directly during init-program, rather than being deferred to go.
Note: the latter change is fairly simple, but I have been unable to test it myself other than verifying it compiles.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@724 f158a5a8-5612-0410-a976-696ce0be7e32
involves taking the existing parts of the XCOFF loader and putting them into a new loader file xcoff_load.c. Also fix a
dependency order change these changes introduce when building openbios-unix.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@720 f158a5a8-5612-0410-a976-696ce0be7e32
included and built as part of the SPARC64 and SPARC32 builds, it was never actually invoked in the boot sequence. Hence this
patch maintains the existing behaviour in that only X86 builds included the Forth loader.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@712 f158a5a8-5612-0410-a976-696ce0be7e32
that each arch no longer needs its own ELF loader implementation.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@710 f158a5a8-5612-0410-a976-696ce0be7e32
from lots of different C files into a single header.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@696 f158a5a8-5612-0410-a976-696ce0be7e32
configuration header file to include/arch/common. Following through the code showed that elfnote.c and linuxbios.c were being
used by sys_info and so these are also renamed to elf_info.c and linuxbios_info.c for clarity.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@693 f158a5a8-5612-0410-a976-696ce0be7e32
the source files into a structure whereby it is easier to find the relationship between a forth package word and its C
implementation, while also creating a libopenbios library into which more common code from the arch/ tree can eventually be
moved.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@siriusit.co.uk>
git-svn-id: svn://coreboot.org/openbios/trunk/openbios-devel@685 f158a5a8-5612-0410-a976-696ce0be7e32
