coreboot.git
14 years agoAdd initial support for a CBFS module for coreinfo.
Uwe Hermann [Tue, 7 Jul 2009 15:10:13 +0000 (15:10 +0000)]
Add initial support for a CBFS module for coreinfo.

Currently it prints a list of components in CBFS and their size/type.
There's a bunch of additional output that could be printed, but that's
for another patch.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4404 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd the IORESOURCE_BRIDGE flag to the fam10 resources for the benefit of the resource...
Myles Watson [Tue, 7 Jul 2009 13:26:35 +0000 (13:26 +0000)]
Add the IORESOURCE_BRIDGE flag to the fam10 resources for the benefit of the resource allocator.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Zheng Bao <zheng.bao@amd.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4403 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoEnable onboard-VGA on the Mitac 6513WU board.
Michael Gold [Mon, 6 Jul 2009 16:05:54 +0000 (16:05 +0000)]
Enable onboard-VGA on the Mitac 6513WU board.

Signed-off-by: Michael Gold <mgold@ncf.ca>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4402 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd support for the Mitac 6513WU mainboard, a Compaq OEM board using the
Michael Gold [Sun, 5 Jul 2009 19:29:39 +0000 (19:29 +0000)]
Add support for the Mitac 6513WU mainboard, a Compaq OEM board using the
i810 chipset. Not all hardware has been tested, but my test PC boots Linux
(via FILO) without any problems.

Also: Add support for the SMSC LPC47U33X to the generic 'smscsuperio' driver.

Signed-off-by: Michael Gold <mgold@ncf.ca>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4401 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix build for i810 boards that don't enable onboard VGA, yet.
Uwe Hermann [Sun, 5 Jul 2009 16:23:43 +0000 (16:23 +0000)]
Fix build for i810 boards that don't enable onboard VGA, yet.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4400 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoEnable onboard VGA on the MS-6178 (i810 chipset) board (trivial).
Uwe Hermann [Sun, 5 Jul 2009 16:01:57 +0000 (16:01 +0000)]
Enable onboard VGA on the MS-6178 (i810 chipset) board (trivial).

Tested on hardware with the patch from r4398 and works fine as soon
as Linux boots (no VGA in FILO for some reason, will investigate).

In order to make the 'i810.vga' VGA blob from the vendor BIOS work
you have to make the check for PCI device ID mismatches non-fatal
(for now) in the src/devices/pci_rom.c file like this:

Index: src/devices/pci_rom.c
===================================================================
--- src/devices/pci_rom.c       (Revision 4393)
+++ src/devices/pci_rom.c       (Arbeitskopie)
@@ -87,7 +87,7 @@
        if (dev->vendor != rom_data->vendor || dev->device != rom_data->device) {
                printk_err("Device or Vendor ID mismatch Vendor %04x, Device %04x\n",
                           rom_data->vendor, rom_data->device);
-               return NULL;
+               // return NULL;
        }

        printk_spew("PCI ROM Image,  Class Code %04x%02x, Code Type %02x\n",

The reason is that the VGA blob thinks the proper VGA device ID is 0x7123
whereas it really is 0x7121 on hardware. There are multiple ways to work
around this (there have been many discussions in the past), we'll see which
method will be used in future...

Note: This has been tested against r4393 only for now to make sure there
are no problems because of the recent resource allocator changes, see
http://www.coreboot.org/pipermail/coreboot/2009-July/050486.html.
Tests with trunk will follow.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4399 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoVarious Intel 82810/82810E changes which allow onboard VGA to work.
Elia Yehuda [Sun, 5 Jul 2009 15:50:30 +0000 (15:50 +0000)]
Various Intel 82810/82810E changes which allow onboard VGA to work.
At the same time also make the 82810 code handle 82810E.

 - Set SMRAM register according to CONFIG_VIDEO_MB value:
    - 512 means 512 KB
    - 1 means 1 MB
    - Every other value for CONFIG_VIDEO_MB (e.g. 0) disables VGA.
   This is not very clean, changing CONFIG_VIDEO_MB to CONFIG_VIDEO_KB
   in a future patch may be nicer.

 - Set MISSC2 register bits as required per datasheet to make VGA work.
   The code handles both 82810 and 82810E.

 - northbridge.c: Add __pci_driver entry for the Intel 82810E.

Also:

 - Rename PAM register #define to PAMR as per datasheet.

 - Drop unused/commented code for now.

 - Don't explicitly set GMCHCFG for now, the default works ok. We'll
   have to figure out the proper/ideal settings later.

The code is based on a patch from Elia Yehuda <z4ziggy@gmail.com> but
has been modified quite a bit for correctness and minimalism.

Tested on hardware with a slightly modified MS-6178 target,
patches to enable onboard-VGA for MS-6178 will follow.

Signed-off-by: Elia Yehuda <z4ziggy@gmail.com>
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4398 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix many things for via/epia-m700 to build.
Myles Watson [Thu, 2 Jul 2009 21:19:33 +0000 (21:19 +0000)]
Fix many things for via/epia-m700 to build.

Unfortunately it still doesn't.  I think it's close, though.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4397 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoChangeLog:
Harald Gutmann [Thu, 2 Jul 2009 19:06:01 +0000 (19:06 +0000)]
ChangeLog:

Turn on Parallel Port and Floppy in Config.lb

Signed-off-by: Harald Gutmann <harald.gutmann@gmx.net>
Acked-by: Andreas B. Mundt <andi.mundt@web.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4396 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoUpdate the k8 code for the v3 resource allocator.
Myles Watson [Thu, 2 Jul 2009 19:02:33 +0000 (19:02 +0000)]
Update the k8 code for the v3 resource allocator.
The major change is that the K8 registers don't get touched until the end of
resource allocation.

Fam10 code could be updated the same way.

Move VGA code before resource allocation but after device enumeration.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4395 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMove the v3 resource allocator to v2.
Myles Watson [Thu, 2 Jul 2009 18:56:24 +0000 (18:56 +0000)]
Move the v3 resource allocator to v2.

Major changes:
1. Separate resource allocation into:
A. Read Resources
B. Avoid fixed resources (constrain limits)
C. Allocate resources
D. Set resources

Usage notes:
Resources which have IORESOURCE_FIXED set in the flags constrain the placement
of other resources.  All fixed resources will end up outside (above or below)
the allocated resources.

Domains usually start with base = 0 and limit = 2^address_bits - 1.

I've added an IOAPIC to all platforms so that the old limit of 0xfec00000 is
still there for resources.  Some platforms may want to change that, but I didn't
want to break anyone's board.

Resources are allocated in a single block for memory and another for I/O.
Currently the resource allocator doesn't support holes.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
Acked-by: Patrick Georgi <patrick.georgi@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4394 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoConvert Supermicro H8DMR to CBFS. Also clean up some whitespace in
Ward Vandewege [Thu, 2 Jul 2009 18:27:02 +0000 (18:27 +0000)]
Convert Supermicro H8DMR to CBFS. Also clean up some whitespace in
targets/supermicro/h8dmr/Config.lb and Config-abuild.lb.

Importantly, this also sets

  default CONFIG_AP_CODE_IN_CAR=0

in

  src/mainboard/supermicro/h8dmr/Options.lb

which is required to make this box boot since the changes that went in in
r4315.

At Myles' suggestion, this patch also sets

  default CONFIG_USE_FAILOVER_IMAGE=0
  default CONFIG_USE_FALLBACK_IMAGE=0
  default CONFIG_XIP_ROM_SIZE=CONFIG_FAILOVER_SIZE

in src/mainboard/supermicro/h8dmr/Options.lb to simplify
targets/supermicro/h8dmr/Config.lb a bit further.

Build tested with abuild, boot tested on physical hardware.

Signed-off-by: Ward Vandewege <ward@gnu.org>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4393 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd support for the Intel Eagle Heights development board.
Thomas Jourdan [Wed, 1 Jul 2009 17:01:17 +0000 (17:01 +0000)]
Add support for the Intel Eagle Heights development board.

Signed-off-by: Thomas Jourdan <thomas.jourdan@gmail.com>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4392 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix typo and only output post code if the work was done.
Myles Watson [Wed, 1 Jul 2009 16:34:03 +0000 (16:34 +0000)]
Fix typo and only output post code if the work was done.

Thanks to Thomas Jourdan <thomas.jourdan@gmail.com> for reporting it.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4391 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix abuild for via/epia-n. Trivial.
Myles Watson [Wed, 1 Jul 2009 15:08:19 +0000 (15:08 +0000)]
Fix abuild for via/epia-n.  Trivial.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4390 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoI missed three files.
Jon Harrison [Wed, 1 Jul 2009 13:19:25 +0000 (13:19 +0000)]
I missed three files.
Signed-off-by: Jon Harrison <bothlyn@blueyonder.co.uk>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4389 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agolet abuild autodetect the coreboot path a bit better. So in the top level of
Stefan Reinauer [Wed, 1 Jul 2009 12:26:11 +0000 (12:26 +0000)]
let abuild autodetect the coreboot path a bit better. So in the top level of
coreboot you can now do:
  $ util/abuild/abuild -t foo/bar
instead of
  $ util/abuild/abuild -t foo/bar $PWD

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4388 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agothe file was not really different, so use the default file (trivial, since it
Stefan Reinauer [Wed, 1 Jul 2009 12:22:26 +0000 (12:22 +0000)]
the file was not really different, so use the default file (trivial, since it
didn't build before, and it still doesn't)
Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4387 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoRon,
Jon Harrison [Wed, 1 Jul 2009 10:57:25 +0000 (10:57 +0000)]
Ron,

Attached is the third revision of the CN400/EPIA-N(L) patch for CB V2.

Patch should work against r4381 (or later ?)

This version now boots all of the way through to attempting to launch a
payload (I'm trying FILO right now), where it falls over with exception
6 (invalid opcode)

The coreboot_table issue seems to have been automagically resolved by
the latest core files.

It may still be that the reason for the payload not starting is down to
some issue with the tables initialising, I'll look closer at that.

Signed-off-by: Jon Harrison <bothlyn@blueyonder.co.uk>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4386 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd AMD family 10 AM2r2 support.
Zheng Bao [Wed, 1 Jul 2009 07:01:32 +0000 (07:01 +0000)]
Add AMD family 10 AM2r2 support.

Coreboot used to take SYSTEM_TYPE as a lable to tell what the socket is.

This patch replaces (some of, not all) CONFIG_SYSTEM_TYPE with CONFIG_SOCKET_TYPE.
It also fix some compiling error in src/northbridge/amd/amdmct/mct/mctardk4.c

Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4385 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agothe tool chain settings should not be in renamed (as they will never live in
Stefan Reinauer [Tue, 30 Jun 2009 17:13:58 +0000 (17:13 +0000)]
the tool chain settings should not be in renamed (as they will never live in
Kconfig)
Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4384 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agowhoops. missed those two..
Stefan Reinauer [Tue, 30 Jun 2009 15:54:53 +0000 (15:54 +0000)]
whoops. missed those two..

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4383 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agocommit svn:externals for last commit (what a pain) (1/2)
Stefan Reinauer [Tue, 30 Jun 2009 15:23:20 +0000 (15:23 +0000)]
commit svn:externals for last commit (what a pain) (1/2)
Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4382 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoThis patch unifies the use of config options in v2 to all start with CONFIG_
Stefan Reinauer [Tue, 30 Jun 2009 15:17:49 +0000 (15:17 +0000)]
This patch unifies the use of config options in v2 to all start with CONFIG_

It's basically done with the following script and some manual fixup:

VARS=`grep ^define src/config/Options.lb | cut -f2 -d\ | grep -v ^CONFIG | grep -v ^COREBOOT |grep -v ^CC`
for VAR in $VARS; do
find . -name .svn -prune -o -type f -exec perl -pi -e "s/(^|[^0-9a-zA-Z_]+)$VAR($|[^0-9a-zA-Z_]+)/\1CONFIG_$VAR\2/g" {} \;
done

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4381 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoadd new supported chipset
Warren Turkal [Tue, 30 Jun 2009 14:11:42 +0000 (14:11 +0000)]
add new supported chipset
Add identification for X58 and ICH10R.

Signed-off-by: Warren Turkal <wt@penguintechs.org>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4380 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoThis copy of flashrom is old and unmaintained. Delete it to avoid
Patrick Georgi [Sun, 28 Jun 2009 22:47:06 +0000 (22:47 +0000)]
This copy of flashrom is old and unmaintained. Delete it to avoid
confusion, after we had a grace period to get people to move over to the
new repository.

History stays available for the archaeologists and for ancient
checkouts (potentially with local changes). Use "svn diff" to get local
changes, and apply it to a newer tree (if you manage to apply it, that
is).

The new repository is available at:
svn://coreboot.org/flashrom/trunk

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4379 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoflashrom is a separate project now. Thus we don't enforce it on our users.
Stefan Reinauer [Sun, 28 Jun 2009 22:40:48 +0000 (22:40 +0000)]
flashrom is a separate project now. Thus we don't enforce it on our users.

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
Acked-by: Patrick Georgi <patrick.georgi@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4378 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoImpact: fix vmlinux from 2.6.30
Yinghai Lu [Sun, 28 Jun 2009 15:04:06 +0000 (15:04 +0000)]
Impact: fix vmlinux from 2.6.30

from 2.6.30 (?)
the new vmlinux with per_cpu and brk support will have more sections.

Elf file type is EXEC (Executable file)
Entry point 0x200000
There are 7 program headers, starting at offset 64

Program Headers:
  Type           Offset             VirtAddr           PhysAddr
                 FileSiz            MemSiz              Flags  Align
  LOAD           0x0000000000200000 0xffffffff80200000 0x0000000000200000
                 0x0000000000e46000 0x0000000000e46000  R E    200000
  LOAD           0x0000000001046000 0xffffffff81046000 0x0000000001046000
                 0x00000000001406e0 0x00000000001406e0  RWE    200000
  LOAD           0x0000000001200000 0xffffffffff600000 0x0000000001187000
                 0x0000000000000888 0x0000000000000888  RWE    200000
  LOAD           0x0000000001388000 0xffffffff81188000 0x0000000001188000
                 0x000000000008a086 0x000000000008a086  RWE    200000
  LOAD           0x0000000001600000 0x0000000000000000 0x0000000001213000
                 0x0000000000015e20 0x0000000000015e20  RWE    200000
  LOAD           0x0000000001629000 0xffffffff81229000 0x0000000001229000
                 0x0000000000000000 0x0000000000208000  RWE    200000
  NOTE           0x0000000000b3c7e8 0xffffffff80b3c7e8 0x0000000000b3c7e8
                 0x0000000000000024 0x0000000000000024         4

 Section to Segment mapping:
  Segment Sections...
   00     .text .notes __ex_table .rodata __bug_table .pci_fixup .builtin_fw __ksymtab __ksymtab_gpl __ksymtab_strings __init_rodata __param
   01     .data .init.rodata .data.cacheline_aligned .data.read_mostly
   02     .vsyscall_0 .vsyscall_fn .vsyscall_gtod_data .vsyscall_1 .vsyscall_2 .vgetcpu_mode .jiffies
   03     .data.init_task .smp_locks .init.text .init.data .init.setup .initcall.init .con_initcall.init .x86_cpu_dev.init .altinstructions .altinstr_replacement .exit.text .init.ramfs
   04     .data.percpu
   05     .bss .brk
   06     .notes

So need to increase NR_SECTIONS.

also fix one typo about phys address mask.

Peter says: A similar fix was also implemented by Maciej Pijanka, so let's
commit this now, eh.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4377 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd a target/ directory for the VIA EPIA-M700 board, so we can build it.
Uwe Hermann [Sat, 27 Jun 2009 16:06:51 +0000 (16:06 +0000)]
Add a target/ directory for the VIA EPIA-M700 board, so we can build it.

Note that this board is nowhere near usable, further patches will follow
and hopefully get this into buildable and usable shape.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4376 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd dump support for SMSC LPC47N252.
François-Regis Vuillemin [Sat, 27 Jun 2009 02:23:26 +0000 (02:23 +0000)]
Add dump support for SMSC LPC47N252.

Signed-off-by: François-Regis Vuillemin <coreboot@miradou.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4375 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoDon't dump 0x07 registers, they're useless as they change every time
Uwe Hermann [Fri, 26 Jun 2009 15:53:07 +0000 (15:53 +0000)]
Don't dump 0x07 registers, they're useless as they change every time
some software wants to do an LDN access (trivial).

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4374 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoThis adds register definitions for all logical devices on the SMSC
Michael Gold [Fri, 26 Jun 2009 15:25:04 +0000 (15:25 +0000)]
This adds register definitions for all logical devices on the SMSC
LPC47U33x, allowing 'superiotool -d' to work.

Also, some consistency string fixes.

Signed-off-by: Michael Gold <mgold@ncf.ca>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4373 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd dump information for F71862FG and F71863FG.
Arjan Koers [Fri, 26 Jun 2009 15:16:21 +0000 (15:16 +0000)]
Add dump information for F71862FG and F71863FG.

Signed-off-by: Arjan Koers <0h3q2rmn2bdb@list.nospam.xutrox.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4372 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoRemove the object files for cbfs from target directories and add a
Myles Watson [Thu, 25 Jun 2009 21:44:38 +0000 (21:44 +0000)]
Remove the object files for cbfs from target directories and add a
cbfstool-clean target to the Makefile.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4371 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoThe problem is that the check to see if we're at the end is never reached. I
Myles Watson [Thu, 25 Jun 2009 15:54:57 +0000 (15:54 +0000)]
The problem is that the check to see if we're at the end is never reached.  I
didn't look into it enough to know why fssize is 32 bytes larger than the
offset.  There may be another bug here.  Maybe something with the CBFS header
not being included or excluded from the calculation?

Anyway, this patch fixes it for all cases size > 32.

I also changed the error message so that it doesn't look like the ROM is full
just because it can't find room for a file.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4370 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd support for the Soyo SY-6BA+ III board.
Uwe Hermann [Wed, 24 Jun 2009 00:35:07 +0000 (00:35 +0000)]
Add support for the Soyo SY-6BA+ III board.

Tested on hardware by Andrew Morgan <ziltro@ziltro.com>, boots
Linux fine. Detailed status at:

  http://www.coreboot.org/Soyo_SY-6BA_Plus_III

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Andrew Morgan <ziltro@ziltro.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4369 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoThis patch adds a proper namestring generation to our ACPIgen generator.
Rudolf Marek [Sun, 21 Jun 2009 20:26:13 +0000 (20:26 +0000)]
This patch adds a proper namestring generation to our ACPIgen generator.
Its used for Name and Scope and Processor now. As bonus, it allows to
create a multi name paths too. Like Scope(\ALL.YOUR.BASE).

Signed-off-by: Rudolf Marek <r.marek@assembler.cz>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4368 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoUndo my ugly commit that added uses clauses in lots of places instead of one.
Myles Watson [Fri, 19 Jun 2009 21:18:14 +0000 (21:18 +0000)]
Undo my ugly commit that added uses clauses in lots of places instead of one.
Fix configuration of all boards. (Abuild tested)
Hopefully fix compilation of PPC boards (they've never compiled for me.)

Apologize profusely.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4367 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agowork around initobject breakage in pc80/Config.lb
Stefan Reinauer [Fri, 19 Jun 2009 19:50:58 +0000 (19:50 +0000)]
work around initobject breakage in pc80/Config.lb

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4366 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoConvert the MSI MS-6178 board to CBFS.
Uwe Hermann [Fri, 19 Jun 2009 15:41:49 +0000 (15:41 +0000)]
Convert the MSI MS-6178 board to CBFS.
Also, enable HIGH_TABLES support for this board.

The HIGH_TABLES failed with:

  No matching ram area found for range:
    [0x00000000000f0000, 0x0000000000100000)
  Ram areas
    [0x0000000000000000, 0x0000000000001000) Reserved
    [0x0000000000001000, 0x00000000000a0000) RAM
    [0x0000000000100000, 0x000000000fff0000) RAM
    [0x000000000fff0000, 0x0000000010000000) Reserved
  SELFBOOT RETURNED!
  Boot failed.

The fix was to change northbridge.c as follows:

  - ram_resource(dev, idx++, 1024, tolmk - 1024);
  + ram_resource(dev, idx++, 768, tolmk - 768);

This is build-tested and tested on hardware by me. It boots fine,
for instace with SeaBIOS and the standard GRUB1 from my disk.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4365 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoChangeLog:
Harald Gutmann [Fri, 19 Jun 2009 12:03:40 +0000 (12:03 +0000)]
ChangeLog:
Add initial ACPI support for M57SLI.
Activates/Enables:
* native Coreboot ACPI for M57SLI
* Soft-Power-Off
* PowerNow!
* High Precision Event Timer
* Windows booting with ACPI support

Signed-off-by: Harald Gutmann <harald.gutmann@gmx.net>
Acked-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4364 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMake sure the address variable is initialized to zero - it is only set when a
Ward Vandewege [Thu, 18 Jun 2009 16:38:35 +0000 (16:38 +0000)]
Make sure the address variable is initialized to zero - it is only set when a
[base address] parameter is supplied on the command line... This patch fixes
random segfaults when using 'cbfstool add'.

Signed-off-by: Ward Vandewege <ward@gnu.org>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4363 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoChange Log:
Harald Gutmann [Thu, 18 Jun 2009 10:05:41 +0000 (10:05 +0000)]
Change Log:
Fix interrupt handling in mptable.c on M57SLI.

Signed-off-by: Harald Gutmann <harald.gutmann@gmx.net>
Acked-by: Luc Verhaegen <libv@skynet.be>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4362 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoChange Log:
Harald Gutmann [Thu, 18 Jun 2009 09:46:11 +0000 (09:46 +0000)]
Change Log:
Activate HIGH_TABES on M57SLI

Signed-off-by: Harald Gutmann <harald.gutmann@gmx.net>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4361 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix configuration of boards that didn't have uses CONFIG_USE_INIT. Trivial.
Myles Watson [Wed, 17 Jun 2009 16:38:43 +0000 (16:38 +0000)]
Fix configuration of boards that didn't have uses CONFIG_USE_INIT. Trivial.

Abuild tested with -C.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4360 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoPatch AMD Fam10 C2 for errata 327, 344, 346, 354, 351.
Marc Jones [Wed, 17 Jun 2009 15:33:57 +0000 (15:33 +0000)]
Patch AMD Fam10 C2 for errata 327, 344, 346, 354, 351.
Removed c2 HT Phy 520a/530a reserved bit.

Signed-off-by: Marc Jones <marcj303@gmail.com>
Acked-by: Ward Vandewege <ward@gnu.org>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4359 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMaximilian Thuermer found a bug where the HT link capability code was always
Marc Jones [Tue, 16 Jun 2009 23:02:39 +0000 (23:02 +0000)]
Maximilian Thuermer found a bug where the HT link capability code was always
updating the passed value to the next link offset even when it was on the
requested link (cap_count).

Maximilian also found a bug where the linktype was still getting attributes
even when it wasn't initialized.

This should fix the HT problems for Fam10 C2. There are still issues with the
microcode which need to be resolved.

Signed-off-by: Marc Jones <marcj303@gmail.com>
Acked-by: Ward Vandewege <ward@gnu.org>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4358 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoThese changes implement car in qemu. The implementation is in several
Ronald G. minnich [Tue, 16 Jun 2009 15:02:52 +0000 (15:02 +0000)]
These changes implement car in qemu.  The implementation is in several
ways superior to v3, while lacking its completeness. But, one nice
thing: no more included .S or .c files. It's all separate compilation.
That should allow our Makefiles to work much better.

Note that the current non-CAR implementation is the default and
continues to work (tested FILO boot to Linux on both CAR and non-CAR).

Index: src/mainboard/emulation/qemu-x86/Config.lb
Change this to be sensitive to USE_DCACHE_RAM. All settings etc. that
depend on this variable are grouped in one if, and the other parts
(romcc etc.) are in the else. This change is a model of how we should be
able to do other motherboards.

Index: src/mainboard/emulation/qemu-x86/Options.lb
add needed options.

Index: src/mainboard/emulation/qemu-x86/failover.c
remove code inclusion from this not-yet-used file.

Index: src/mainboard/emulation/qemu-x86/rom.c
This is the entry point for the rom-based code. Called stage1.c in v3.

Index: src/lib/Config.lb
change initobject to a .o from a .c; this fixed a build problem.

Index: src/pc80/serial.c
make uart_init non-static.

Index: src/pc80/Config.lb
add initobject

Index: src/arch/i386/init/entry.S
Entry point. Unify a bunch of files that were fiddly lttle includes. From v3.

Index: src/arch/i386/init/ldscript.ld
new file. The goal is to hang all init changes for CAR here, to minimize other changes to any
other ldscript. Besides, putting this in init makes sense; entry and car are manage init.

Index: src/arch/i386/init/car.S
generic i386 car code from v3.

Index: src/arch/i386/init/ldscript_fallback_cbfs.lb
Fix what looks like a bug: this was not including the init.text section.

Index: targets/emulation/qemu-x86/Config.lb
push up the console loglevel. qemu is for debugging so we might as well
get all the debugging we can.

Index: targets/emulation/qemu-x86/Config-car.lb
For CAR bullds.

Signed-off-by: Ronald G. minnich <rminnich@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4357 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix typo in Winbond W83977TF register listing.
Ioannis Barkas [Fri, 12 Jun 2009 14:20:27 +0000 (14:20 +0000)]
Fix typo in Winbond W83977TF register listing.

Signed-off-by: Ioannis Barkas <tripl3fault@yahoo.com>
Signed-off-by: Nikos Barkas <levelwol@gmail.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4356 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix s2895 failover booting.
Myles Watson [Thu, 11 Jun 2009 18:27:41 +0000 (18:27 +0000)]
Fix s2895 failover booting.

Abuild tested and boot tested on s2895 and serengeti_cheetah.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4355 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agothis port is horribly broken and should not have been checked in. This patch
Stefan Reinauer [Tue, 9 Jun 2009 15:22:47 +0000 (15:22 +0000)]
this port is horribly broken and should not have been checked in. This patch
gets us through config, but it fails during build because the original patch
duplicated some files for VIA systems.

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4354 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix a little white space issue. Also, don't copy the rom image
Ronald G. Minnich [Tue, 9 Jun 2009 14:44:37 +0000 (14:44 +0000)]
Fix a little white space issue. Also, don't copy the rom image
 if it is already in its correct location.

Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4353 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd (commented) line for VGA blob adding (CBFS version) to simplify things for users...
Uwe Hermann [Mon, 8 Jun 2009 13:05:47 +0000 (13:05 +0000)]
Add (commented) line for VGA blob adding (CBFS version) to simplify things for users a bit.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4352 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoThis is transition code for cbfs to implement
Ronald G. Minnich [Mon, 8 Jun 2009 03:33:57 +0000 (03:33 +0000)]
This is transition code for cbfs to implement
cbfs files at fixed addresses.

I call this transitional as the approach I am taking is to add
capability to cbfstool but not change code in a way that will break
existing usages. Later, once we're sure nothing has broken, we can start to
smooth the edges.

Right now, fixed address file are only supported via the add command.

There is one additional command syntax, so, example:
cbfstool add rom romstrap optionrom 0xffffd000

Will add the file to that fix location for a romstrap.

The assumption is that the ROM is based at the end of a 32-bit address
space. As you can see from the code, that assumption can easily be
over-ridden, if we ever need to, with a command option.

Here is one example output result.

rminnich@xcpu2:~/src/bios/coreboot-v2/util/cbfstool$ ./cbfstool x.cbf print
x.cbf: 1024 kB, bootblocksize 32768, romsize 1048576, offset 0x0
Alignment: 16 bytes

Name                           Offset     Type         Size
h                              0x0        optionrom   251
                             0x130      free         917120
h3                             0xdffe0    optionrom    251
                             0xe0110    free         97960

The way this is implemented is pretty simple. I introduce a new
operator, split, that splits an unallocated area into two unallocated
areas. Then, allocation merely becomes a matter of 0, 1, or 2 splits:
0 split -- the free area is the exact fit
1 splits -- need to split some off the front or back
2 splits -- need to split off BOTH the front and back

I think you'll be able to see what I've done. I call this transitional
because, in the end state, we only need one allocate function; for now
I've left two in, to make sure I don't break compatibilty.

Why I like this better than ldscript approach: I like having the
ROMSTRAP located by cbfs, not linker scripts. For one thing, it makes
romstrap visible as a first class object. I think I would have latched
onto a problem I was having much more quickly had I remembered the
ROMSTRAP. It gets lost in the linker scripts.

At this point, we should be able to start removing special ROMSTRAP location
code from linker scripts.

Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Acked-by: Patrick Georgi <patrick.georgi@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4351 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoA bunch of additional EPIA-M700 cleanups and also some non-cosmetic changes:
Uwe Hermann [Sun, 7 Jun 2009 14:38:32 +0000 (14:38 +0000)]
A bunch of additional EPIA-M700 cleanups and also some non-cosmetic changes:

 - Make get_dsdt script executable.

 - Rename DrivingClkPhaseData.c to driving_clk_phase_data.c.

 - Set proper IRQ_SLOT_COUNT value in the hope that the '14' from irq_table.c
   is correct.

 - Fix broken or incorrect #include names to increase likelyhood of a
   successful compile.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4350 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFirst bunch of coding style and consistency cleanups for the
Uwe Hermann [Sun, 7 Jun 2009 13:46:50 +0000 (13:46 +0000)]
First bunch of coding style and consistency cleanups for the
EPIA-M700 target.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4349 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoWhen I started refactoring mainboard Config.lb, I added two different
Carl-Daniel Hailfinger [Sat, 6 Jun 2009 16:50:38 +0000 (16:50 +0000)]
When I started refactoring mainboard Config.lb, I added two different
files for targets without failover:
src/config/nofailovercalculation.lb (64 kB XIP)
src/config/nofailovercalculation128.lb (128 kB XIP)
Targets with other XIP sizes were ignored.

This patch moves XIP size back into mainboard code.

Benefits from this patch:
- src/config/nofailovercalculation128.lb is no longer needed
- Targets with XIP sizes besides 64k and 128k benefit from refactoring
- Conceptually, this makes the include files pure calculation files
without settings.

Abuild tested.

Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Patrick Georgi <patrick.georgi@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4348 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMake failover larger and decrease fallback's size so the total stays the
Patrick Georgi [Sat, 6 Jun 2009 12:19:59 +0000 (12:19 +0000)]
Make failover larger and decrease fallback's size so the total stays the
same. The errata need some extra room in failover.

Trivial and abuild tested

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Patrick Georgi <patrick.georgi@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4347 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix for Erratum 350 for AMD Fam10h CPUs.
Marco Schmidt [Sat, 6 Jun 2009 11:33:58 +0000 (11:33 +0000)]
Fix for Erratum 350 for AMD Fam10h CPUs.

Compared to posted patch, there are whitespace fixes
(request by Uwe), and a guard to run the erratum only
on AMD_RB_C2 (request by Marc).

Signed-off-by: Marco Schmidt <mashpb@gmail.com>
Acked-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4346 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix for Erratum 343 for AMD Fam10h CPUs.
Marco Schmidt [Sat, 6 Jun 2009 11:21:52 +0000 (11:21 +0000)]
Fix for Erratum 343 for AMD Fam10h CPUs.

Signed-off-by: Marco Schmidt <mashpb@gmail.com>
Acked-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4345 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoChange the CBFS build process to use coreboot.rom
Patrick Georgi [Sat, 6 Jun 2009 07:19:53 +0000 (07:19 +0000)]
Change the CBFS build process to use coreboot.rom
instead of coreboot.strip. That fixes the normal
image because the calculations for its offset in
the ROM match reality again.

This requires changes in CBFS configurations to
minimize the bootblock size. These are also done
for CBFS boards.

Other than this a couple of minor fixes are in this
patch:
- make asus/m2v-mx_se build with abuild with a
  crosscompiler
- move CONFIG_CBFS for hp/dl145_g3 to Options.lb
  as it's done everywhere else
- change the default config of abuild to not
  provide ROM_IMAGE_SIZE values for the images
  in a CBFS configuration
- change abuild's crosscompile autodetection to
  not try to use "i386-elf-i386-elf-gcc" (which
  is bogus)

Except for the latter two abuild changes (both
in util/abuild/abuild), they're available as
patch set on the mailing list in a mail from
2009-06-05 titled
[PATCH]es to get normal image to work again with CBFS

The changes in util/abuild/abuild are trivial and
abuild tested.

As discussed on the list,
targets/hp/dl145_g3/Config-abuild.lb is
deleted, now that Config.lb works again.

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4344 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix non-revF K8 ram init compilation which was broken in r4341.
Carl-Daniel Hailfinger [Fri, 5 Jun 2009 23:43:11 +0000 (23:43 +0000)]
Fix non-revF K8 ram init compilation which was broken in r4341.
Change all printk_raminit to printk_spew.

Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4343 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoInitial untested board code for the VIA EPIA-M700 Mini-ITX board.
Uwe Hermann [Fri, 5 Jun 2009 23:02:43 +0000 (23:02 +0000)]
Initial untested board code for the VIA EPIA-M700 Mini-ITX board.

The patch has been submitted by bari <bari@onelabs.com> and written
by OLPC.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4342 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoK8 RAM init debug messages are pretty short and sometimes cryptic. Make
Carl-Daniel Hailfinger [Fri, 5 Jun 2009 20:37:35 +0000 (20:37 +0000)]
K8 RAM init debug messages are pretty short and sometimes cryptic. Make
them a bit more verbose and hopefully more understandable.

Old messages for my machine with 5 GB:
RAM: 0x00400000 kB
Ram3
[...]
Initializing memory:  done
RAM: 0x00500000 kB

New messages:
RAM end at 0x00400000 kB
Adjusting lower RAM end
Lower RAM end at 0x003f0000 kB
Ram3
[...]
Initializing memory:  done
Handling memory hole at 0x00300000 (default)
RAM end at 0x00500000 kB
Handling memory mapped above 4 GB
Upper RAM end at 0x00500000 kB
Correcting memory amount mapped below 4 GB
Adjusting lower RAM end
Lower RAM end at 0x00300000 kB

Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4341 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agodie() does never return. Annotate it as such.
Carl-Daniel Hailfinger [Fri, 5 Jun 2009 11:41:51 +0000 (11:41 +0000)]
die() does never return. Annotate it as such.
Any endless loop after die() can be eliminated.
Dereferencing a NULL pointer is bad. die() instead.
Replace endless loops with die().

Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4340 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAfter I modify the pci_ext_read_config32 and pci_ext_read_config32, the step 6a
Zheng Bao [Fri, 5 Jun 2009 10:00:07 +0000 (10:00 +0000)]
After I modify the pci_ext_read_config32 and pci_ext_read_config32, the step 6a
starts to play its role. Then the system hangs at HDA init. I dont know what the
VC1 is. The RPR says "Optional Features (only needed if CMOS option is enabled)"
in 5.10.2. Before I know what it is, I think it is better to skip it.

Tested on dbm690t.

Add comment from Rudolf,
"
VC is virtual channel. Its used for isochronous transfer of data to sound card.
The virtual channel guarantee "on time" delivery. In other words it sets up a
channel for data to sound card, which means that that arrivs in time and there will
be no interuptions in audio stream.

http://www.microsoft.com/whdc/connect/pci/wlp_interrupt.mspx
"

Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4339 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd a hopefully more correct and flexible set_dram_buffer_strength()
Elia Yehuda [Fri, 5 Jun 2009 00:22:25 +0000 (00:22 +0000)]
Add a hopefully more correct and flexible set_dram_buffer_strength()
function based on test results with many different DIMMs.

Tested by Uwe Hermann <uwe@hermann-uwe.de> on hardware.

Might need a small increase of ROM_IMAGE_SIZE for some boards, we'll see.

Signed-off-by: Elia Yehuda <z4ziggy@gmail.com>
Acked-by: Corey Osgood <corey.osgood@gmail.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4338 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoThe point of the patch is to make it easier to understand the raminit
Myles Watson [Thu, 4 Jun 2009 20:18:42 +0000 (20:18 +0000)]
The point of the patch is to make it easier to understand the raminit
code, specifically the difference between pre_f and f code.

The only functional changes are in printk statements.  The rest is white space.

1. Remove some #if 0 and #if 1 blocks
2. Remove #if USE_DCACHE_RAM blocks.  All K8 boards use CAR.
2. Correct typos (canidate -> candidate)
3. Try to minimize the differences between amdk8_f.h and amdk8_pre_f.h
4. Try to minimize the differences between raminit.c and raminit_f.c
5. Make boards that have rev_f processors include the correct raminit code

There is much more that could be done, but it's a start.

Abuild tested and boot tested on s2892 and serengeti_cheetah.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4337 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoThis patch is about some noticable bugs which was made by no reason.
Zheng Bao [Thu, 4 Jun 2009 01:57:03 +0000 (01:57 +0000)]
This patch is about some noticable bugs which was made by no reason.
1. In rs690_cmn.c, mask the lower 4 bits of the BAR3. No doubt, right?
2. In rs690_pcie.c,
  (1) Obviously, the mask should be 0xF, and bit 19 should be set to 1 (in comment).
      In rpr 5.10.2, step 2, step 2.1 & step 2.6
  (2) The dynamic buffer allocation is enabled by setting bit 11 of PCIEIND: 0x20,
      instead of PCIEIND_P: 0x20.
      In rpr 5.10.2, step 5. Dynamic Slave CPL Buffer Allocation

Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4336 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoRevert "CMOS: Add set_option and rework get_option."
Luc Verhaegen [Wed, 3 Jun 2009 14:19:33 +0000 (14:19 +0000)]
Revert "CMOS: Add set_option and rework get_option."

This reverts commit eb7bb49eb5b48c39baf7a256b7c74e23e3da5660.

Stepan pointed out that "s" means string, which makes the following statement
in this commit message invalid: "Since we either have reserved space (which
we shouldn't do anything with in these two functions), an enum or a
hexadecimal value, unsigned int seemed like the way to go."

Signed-off-by: Luc Verhaegen <libv@skynet.be>
Acked-by: Luc Verhaegen <libv@skynet.be>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4335 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoRevert "kontron 986lcd_m: cmos.layout: mark boot_devices as reserved."
Luc Verhaegen [Wed, 3 Jun 2009 14:19:20 +0000 (14:19 +0000)]
Revert "kontron 986lcd_m: cmos.layout: mark boot_devices as reserved."

This reverts commit c03527377db5951f0d3228e2a93b4c57dd81b8ec.

Stepan pointed out that 's' means string, and that therefor strings do exist.
Marking this as reserved breaks some payloads.

Signed-off-by: Luc Verhaegen <libv@skynet.be>
Acked-by: Luc Verhaegen <libv@skynet.be>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4334 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agokontron 986lcd_m: cmos.layout: mark boot_devices as reserved.
Luc Verhaegen [Wed, 3 Jun 2009 11:53:54 +0000 (11:53 +0000)]
kontron 986lcd_m: cmos.layout: mark boot_devices as reserved.

The kontron 986lcd_m cmos.layout had a 512bit area claimed for "boot_devices".
The changes to the cmos code no longer allow usage of values larger than
32bits. Since this option was completely unused, mark it as reserved.

Fixes build after the get_option change (r4332)..

Signed-off-by: Luc Verhaegen <libv@skynet.be>
Acked-by: Luc Verhaegen <libv@skynet.be>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4333 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoCMOS: Add set_option and rework get_option.
Luc Verhaegen [Wed, 3 Jun 2009 10:47:19 +0000 (10:47 +0000)]
CMOS: Add set_option and rework get_option.

To ease some of my debugging pain on the unichrome, i decided i needed to
move FB size selection into cmos, so i could test a size and then reset it
to the default after loading this value so that the next reboot uses the
(working) default again. This meant implementing set_option in parallel to
get_option.

get_option was then found to have inversed argument ordering (like outb) and
passing char * and then depending on the cmos layout length, which made me
feel quite uncomfortable. Since we either have reserved space (which we
shouldn't do anything with in these two functions), an enum or a
hexadecimal value, unsigned int seemed like the way to go. So all users of
get_option now have their arguments inversed and switched from using ints
to unsigned ints now.

The way get_cmos_value was implemented forced us to not overlap byte and to
have multibyte values be byte aligned. This logic is now adapted to do a
full uint32_t read (when needed) at any offset and any length up to 32, and
the shifting all happens inside an uint32_t as well. set_cmos_value was
implemented similarly. Both routines have been extensively tested in a
quick separate little program as it is not easy to get this stuff right.

build_opt_tbl.c was altered to function correctly within these new
parameters. The enum value retrieval has been changed strol(..., NULL, 10)
to stroul(..., NULL, 0), so that we not only are able to use unsigned ints
now but so that we also interprete hex values correctly. The 32bit limit
gets imposed on all entries not marked reserved, an unused "user_data" field
that appeared in a lot of cmos.layouts has been changed to reserved as well.

Signed-off-by: Luc Verhaegen <libv@skynet.be>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4332 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoModify it based on the RPR 5.7.7. Switching GGSP Configuration By Register Programming.
Zheng Bao [Wed, 3 Jun 2009 03:15:05 +0000 (03:15 +0000)]
Modify it based on the RPR 5.7.7. Switching GGSP Configuration By Register Programming.
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4331 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMore compact format for wiki output at
Uwe Hermann [Tue, 2 Jun 2009 23:49:00 +0000 (23:49 +0000)]
More compact format for wiki output at
http://www.coreboot.org/Coreboot_Options (trivial).

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4330 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agocbfstool reacts to a too large bootblock file by stopping
Patrick Georgi [Mon, 1 Jun 2009 20:02:21 +0000 (20:02 +0000)]
cbfstool reacts to a too large bootblock file by stopping
with an error code now.

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4329 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoCosmetics and consistency fixes in src/superio/serverengines/pilot/ (trivial).
Uwe Hermann [Mon, 1 Jun 2009 01:38:29 +0000 (01:38 +0000)]
Cosmetics and consistency fixes in src/superio/serverengines/pilot/ (trivial).

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4328 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFollowing patch moves all vt8237 fadt.c from mainboard/* file to chipset
Rudolf Marek [Sun, 31 May 2009 17:00:25 +0000 (17:00 +0000)]
Following patch moves all vt8237 fadt.c from mainboard/* file to chipset
directory just with one common file.

Changes to FADT: move to rev4, fix the generic register descriptors, detect additional VT8237S features.
Change the compiler to CORE , its revision to 42.

Signed-off-by: Rudolf Marek <r.marek@assembler.cz>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4327 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMany Kudos go to Segher Boessenkool and Patrick Georgi for figuring this one
Stefan Reinauer [Sat, 30 May 2009 15:12:33 +0000 (15:12 +0000)]
Many Kudos go to Segher Boessenkool and Patrick Georgi for figuring this one
out. Fix the libgcc dependency on abort() due to nested functions.

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4326 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agorename the option CONFIG_PCI_OPTION_ROM_RUN_VM86 to CONFIG_PCI_OPTION_ROM_RUN_REALMODE.
Joseph Smith [Fri, 29 May 2009 18:41:09 +0000 (18:41 +0000)]
rename the option CONFIG_PCI_OPTION_ROM_RUN_VM86 to CONFIG_PCI_OPTION_ROM_RUN_REALMODE.
Signed-off-by: Joseph Smith <joe@settoplinux.org>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4325 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoenable/disable IDE 0/1 (Primary/Secondary) interfaces on the i82801xx southbridge.
Joseph Smith [Fri, 29 May 2009 13:45:22 +0000 (13:45 +0000)]
enable/disable IDE 0/1 (Primary/Secondary) interfaces on the i82801xx southbridge.
Signed-off-by: Joseph Smith <joe@settoplinux.org>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4324 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agodrop most of the crappy vm86 code and replace it with a rewritten
Stefan Reinauer [Fri, 29 May 2009 13:08:27 +0000 (13:08 +0000)]
drop most of the crappy vm86 code and replace it with a rewritten
version that has all assembler in a .S file and all C code in a .c
file. Also, remove requirement to move around between GDTs.

This version includes the suggestions from Peter to clean up CR0 manipulation
and to guard critical code paths by cli/sti. Tested and working on my hardware.

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4323 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFix build with CONFIG_*_ROM_RUN.
Luc Verhaegen [Fri, 29 May 2009 03:44:47 +0000 (03:44 +0000)]
Fix build with CONFIG_*_ROM_RUN.

Last commit broke it due to leftover "void" from prototype.

Signed-off-by: Luc Verhaegen <libv@skynet.be>
Acked-by: Luc Verhaegen <libv@skynet.be>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4322 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoImplement native VGA Support.
Luc Verhaegen [Fri, 29 May 2009 03:04:16 +0000 (03:04 +0000)]
Implement native VGA Support.

This code brings a rather complete set of VGA IO routines for whoever wants it.
These consist of the by now familiar read/write/mask sets. Due to the crazy
nature of VGA, an ancient standard with bits all over the place, it makes no
sense to define individual registers. You need a vga register spec at hand if
you want to do anything anyway. These IO routines are always exposed.

It also provides code to natively set up a 640x400 VGA textmode with an 8x16
font. The native VGA mode code is behind the OPTION_VGA option, as the font
really adds to the size of the compiled/compressed rom. The font is the one
also present in the linux kernel, but this file is unlicensed. Another copy of
this is also present in coreboot in the deprecated console/btext code.

The vga console code has been cleaned up, but it still has some TODO's left
open, but that's for when i finally have found the remaining issue with the
epia-m. Right now, it is important to get parts of my work out already and to
make the remainder managable again.

Signed-off-by: Luc Verhaegen <libv@skynet.be>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4321 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoAdd lzma.o for CBFS, regardless if CONFIG_COMPRESSED_PAYLOAD_LZMA is
Patrick Georgi [Thu, 28 May 2009 22:18:25 +0000 (22:18 +0000)]
Add lzma.o for CBFS, regardless if CONFIG_COMPRESSED_PAYLOAD_LZMA is
enabled or not.
CONFIG_COMPRESSED_PAYLOAD_LZMA is set only if the lzma utility is found
on the system - at least when using abuild. CBFS doesn't use this tool
for compression.

The result was a failed build if lzma (the tool) wasn't found, or
failed runtime (if src/lib/cbfs.c disables lzma decompression based on
CONFIG_COMPRESSED_PAYLOAD_LZMA)

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4320 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMake memmove copy (dev->resources -i) resource structs instead of
Myles Watson [Thu, 28 May 2009 21:57:11 +0000 (21:57 +0000)]
Make memmove copy (dev->resources -i) resource structs instead of
(dev->resources-i) bytes in compact_resources.

Signed-off-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4319 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoFirst batch of indent-aided code cleanups, more will follow.
Uwe Hermann [Wed, 27 May 2009 18:55:19 +0000 (18:55 +0000)]
First batch of indent-aided code cleanups, more will follow.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4318 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMake directory hierarchy flat to match the same layout we use
Uwe Hermann [Wed, 27 May 2009 17:06:54 +0000 (17:06 +0000)]
Make directory hierarchy flat to match the same layout we use
for other chipsets, as suggested on IRC.

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4317 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoTrivial, but brown paper bag worthy:
Patrick Georgi [Wed, 27 May 2009 14:57:53 +0000 (14:57 +0000)]
Trivial, but brown paper bag worthy:
#ifdef CONFIG_foo
is a bad idea with our build system

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Patrick Georgi <patrick.georgi@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4316 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMove coreboot_ram and coreboot_apc to CBFS. This allows to
Patrick Georgi [Wed, 27 May 2009 14:19:31 +0000 (14:19 +0000)]
Move coreboot_ram and coreboot_apc to CBFS. This allows to
reduce the size of the bootblock (done for kontron/986lcd-m)

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4315 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoChange all vx800 file names from CamelCase to camel_case to match
Uwe Hermann [Wed, 27 May 2009 13:46:37 +0000 (13:46 +0000)]
Change all vx800 file names from CamelCase to camel_case to match
our coding guidelines (trivial).

Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4314 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoHere's the VIA vx800 patch from OLPC.
Bari Ari [Wed, 27 May 2009 13:12:42 +0000 (13:12 +0000)]
Here's the VIA vx800 patch from OLPC.

It's untested, but a good starting point for everyone.

Signed-off-by: Bari Ari <bari@onelabs.com>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4313 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoutil/vgabios: build/warning fixes.
Luc Verhaegen [Wed, 27 May 2009 11:39:16 +0000 (11:39 +0000)]
util/vgabios: build/warning fixes.

Signed-off-by: Luc Verhaegen <libv@skynet.be>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4312 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoClean up acpi table writing code, and don't rely
Patrick Georgi [Tue, 26 May 2009 19:39:14 +0000 (19:39 +0000)]
Clean up acpi table writing code, and don't rely
on a given alignment for the RSDP and RSDT - look
it up instead.

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4311 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoTell lpgcc about the target architecture directory. This slipped through since
Stefan Reinauer [Tue, 26 May 2009 18:01:53 +0000 (18:01 +0000)]
Tell lpgcc about the target architecture directory. This slipped through since
FILO does not use lpgcc (yet)

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4310 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMake printk_* behaviour more consistent. Without it, side
Patrick Georgi [Tue, 26 May 2009 14:49:59 +0000 (14:49 +0000)]
Make printk_* behaviour more consistent. Without it, side
effects in the arguments (eg. a pci config read, or variable increment)
"vanish" with the message, and the behaviour changes.

Some of these effects might be unwanted, but at least they are consistent now.
To reduce the memory footprint slightly, the formatted strings are discarded.

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Myles Watson <mylesgw@gmail.com>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4309 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoencapsule mbi initialization in write_multiboot_table, where it belongs. (very
Stefan Reinauer [Tue, 26 May 2009 14:37:17 +0000 (14:37 +0000)]
encapsule mbi initialization in write_multiboot_table, where it belongs. (very
simple and trivial)

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4308 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMake vsprintf reentrant. More importantly, eliminate global variable.
Patrick Georgi [Tue, 26 May 2009 14:31:37 +0000 (14:31 +0000)]
Make vsprintf reentrant. More importantly, eliminate global variable.

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4307 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoMajor cleanup of i386 tables.c:
Stefan Reinauer [Tue, 26 May 2009 14:07:44 +0000 (14:07 +0000)]
Major cleanup of i386 tables.c:

* fix copyright messages
* remove all HAVE_HIGH_TABLES and HAVE_LOW_TABLES preprocessor hackery
  and instead use high_tables_base to find out if high tables should be used.
  The code path with high tables disabled and high tables not available for
  another reason should be the same.
* put MP-table into Fseg instead of 0x10. This allows us to drop an huge and ugly
  portion of code. And it will make some ugly Linux warnings go away.
* use ALIGN macro instead of hand crafted aligning.
* renumber post codes in this piece of code (don't jump ahead and
  back anymore)

Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4306 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

14 years agoVarious fixes to the tree to get coreboot-v2 to build on Solaris
Patrick Georgi [Tue, 26 May 2009 14:03:51 +0000 (14:03 +0000)]
Various fixes to the tree to get coreboot-v2 to build on Solaris

- Replace $(PWD) with $(CURDIR) in Makefiles. I don't know why
  the Solaris version behaves differently, but CURDIR is a safe
  choice on gnu make (and we require gnu make already)
- Use tail -1 instead of tail -n1 in a file that already relies on
  tail -1 support in another place
- Use tail -1 as alternative to tail -n1 in another place
- Use #define for ulong_t in romcc, as that name is used on Solaris
- Avoid fprinting a null pointer. The standard doesn't mandate that
  this is a special case, and Solaris doesn't implement it that way.

Signed-off-by: Patrick Georgi <patrick.georgi@coresystems.de>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4305 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1