Use a simple one-liner for each of the functions. You can surely optimize
the code some more, but I chose not to do that in order to keep it readable.
When compiling with -Os the size differences are minimal.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Jordan Crouse <jordan.crouse@amd.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3236 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
of functions while we're at it.
hex2bin() now also supports upper-case input characters.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Jordan Crouse <jordan.crouse@amd.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3235 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
- Use _FOO_H include guard format everywhere.
- Add missing speaker.c prototypes to libpayload.h.
- Consistently use short form u8/u16/u32 instead of uint8_t et. al.
- kcofig: Use 'depends on' instead of 'depends', which seems deprecated.
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@3234 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Add a Geode video driver in lieu of VGA on Geode LX devices
Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3233 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
recomendations.
Changes include the following:
fix > 4GB dqs tests
fix channel interleaving
ecc memory scrub updates
MC tristating updates
debug print changes
fix memory hoisting across nodes -
The DRAM Hole Address Register is set via devx in each node, but the Node
number <-> DRAM Base mapping and the Node number <-> DstNode mapping is
set in Node 0. The memmap is setup on node0 and copied to the other nodes
later. so dev, not devx. The bug was the hole was always being set on the
first node.
Signed-off-by: Marc Jones (marc.jones@amd.com)
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3232 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
The Geode video driver will require a number of support functions,
including udelay(), PCI bus walking and MSRs. This adds those functions
in preparation for the actual code.
Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3231 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Add a framework for multiple video console drivers. This is to prepare
for the Geode driver.
Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3230 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Slight tweak to to the install target to avoid copying .svn
files. Trivial self ack.
Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Acked-by: Jordan Crouse <jordan.crouse@amd.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3229 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
easily for other payloads. Also, add a 'distclean' target (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@3227 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
libpayload uses a ton of flags and other scary gcc and ld options. These
wrappers hide most of that from the user, so that using libpayload is as
easy as lpgcc -o hello hello.c
Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3226 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
- Properly set LINES and COLS, needed for a real curses application.
- Implement notimeout() and wtimeout(), which are trivial.
- Implement a dummy flushinp() for now as it's needed by a curses
application I'm porting, will be replaced by something useful later.
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@3224 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Implement a timeout option for getchar() to return after so many
milliseconds. Also implement the same thing for curses using
the halfdelay() function.
Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3223 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Signed-off-by: Ed Swierk <eswierk@arastra.com>
Acked-by: Joseph Smith <joe@smittys.pointclark.net>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3221 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
patch.
Signed-off-by: Marc Jones (marc.jones@amd.com)
Acked-by: Marc Jones (marc.jones@amd.com)
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3220 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
This field is rarely used (and not used in the LX tables).
There is not a good reason to mask off non-exclusive IRQs.
Signed-off-by: Marc Jones(marc.jones@amd.com)
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3219 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
It also keeps the boot processes from rebooting through out the coreboot process.
Signed-off-by: Joseph Smith <joe@smittys.pointclark.net>
Acked-by: Corey Osgood <corey.osgood@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3218 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Following on the previous code to streamline the libpayload init code,
this removes the now unneeded stack structures.
Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3217 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Until now, the build would break, and even if it didn't the ELF would
triple-fault in QEMU.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3216 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
interpret whitespace as macro argument delimiter. Since the code is
preprocessed by gcc and the tokenizer may insert whitespace, that can
fail. http://sourceware.org/bugzilla/show_bug.cgi?id=669
The same change was committed in r3044 to the AMD CAR code.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Jordan Crouse <jordan.crouse@amd.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3215 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
coreinfo features / modules will consume (trivial).
The measurements were done with libpayload r3213 (but compiled with -Os),
and coreinfo r3211 (also compiled with -Os).
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@3214 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
They're generally useful for lots of stuff, but especially for converting
to/from the compact 160 bit (20 byte) representation of SHA-1 hashes to
the "hex" representation (same as 'sha1sum' output), which is 40 bytes long.
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@3213 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
by not deleting build/config.h during 'make clean' (only in 'make distclean').
Also, change the default behaviour of 'make' from asking the user to
run 'make config' (or similar) to actually _run_ 'make config' without
asking questions. It's always possible to explicitly invoke
'make menuconfig' or 'make xconfig' and so on, of course.
Finally, make _all_ targets (allyesconfig, randconfig, and so on)
generate a build/config.h file, as we always #include it.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Jordan Crouse <jordan.crouse@amd.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3211 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
This is also taken from the HelenOS project.
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@3210 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
"off" in Config.lb should cause the PCI device not to respond to
configuration requests.
Replace the existing code that I naively copied from esb6300 with
something that actually works on the 3100.
Signed-off-by: Ed Swierk <eswierk@arastra.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3209 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Signed-off-by: Joseph Smith <joe@smittys.pointclark.net>
Acked-by: Corey Osgood <corey.osgood@gmail.com>
Acked-by: Ed Swierk <eswierk@arastra.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3208 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
There are no boards that use the i82801DB (ICH4). The code does NOT work.
Signed-off-by: Joseph Smith <joe@smittys.pointclark.net>
Acked-by: Corey Osgood <corey.osgood@gmail.com>
Acked-by: Ed Swierk <eswierk@arastra.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3206 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
pci_locate_device() to locate the SMBus controller and LPC bridge
devices on the PCI bus. Since these devices are always located at a
fixed PCI bus:device:function, the code can be simplified by
hardcoding the devices.
Signed-off-by: Ed Swierk <eswierk@arastra.com>
Acked-by: Corey Osgood <corey.osgood@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3205 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Let the linker figure out the correct address and just CALL the
start_main entry point.
Signed-off-by: Klaus Schnass <dev@stuffit.at>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3204 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
displaying the current date/time in the lower-right corner (optional).
Also, only build/use coreinfo modules which were selected in kconfig. This
makes coreinfo truly modular, and you can save quite a bit of ROM space
by disabling unwanted parts of coreinfo.
Finally, simplify the Makefile a bit by getting rid of MODULES (and only
using OBJECTS).
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Jordan Crouse <jordan.crouse@amd.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3203 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
upper 128 bytes of NVRAM (if enabled).
For most chipsets this means using I/O ports 0x72/0x73, but at least
on some VIA chipsets (I tested the VIA VT8237R on actual hardware)
these ports won't work and you have to use 0x74/0x75. Thus, make this
a Kconfig option for now.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Jordan Crouse <jordan.crouse@amd.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3202 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
the system automatically unless software resets the timer
periodically. The extra reboot extends boot time by several seconds.
The attached patch adds a function to the Intel 3100 southbridge code
that halts the TCO timer, thus preventing this extra reboot, and calls
the function early in the boot process on the Mt. Arvon board.
It also fixes a bug in the LPC device initialization -- the ACPI BAR
enable flag is bit 7, not bit 4.
Signed-off-by: Ed Swierk <eswierk@arastra.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3198 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
You can see this by executing commands like this:
grep -r pci_assign_irqs coreboot/src/*
This basically AMD/LX based boards: pcengines/alix1c,
digitallogic/msm800sev, artecgroup/dbe61, amd/norwich, amd/db800.
Also for AMD/GX1 based boards need a patch
[http://www.pengutronix.de/software/ptxdist/temporary-src/references/geode-5530.patch]
for the right IRQ setup.
AMD/GX1 based boards is: advantech/pcm-5820, asi/mb_5blmp, axus/tc320,
bcom/winnet100, eaglelion/5bcm, iei/nova4899r, iei/juki-511p.
I have two ideas.
1. Delete duplicate code from AMD/LX based boards.
2. Add IRQ routing for AMD/GX1 boards in coreboot.
The pirq.patch for IRQ routing logically consist from of two parts:
First part of pirq.patch independent from type chipsets and assign IRQ for
ever PCI device. It part based on AMD/LX write_pirq_routing_table() function.
Second part of pirq.patch depends of type chipset and set PIRQx lines
in interrupt router. This part supports only CS5530/5536 interrupt routers.
IRQ routing functionality is included through PIRQ_ROUTE in Config.lb.
Tested on iei/juki-511p(cs5530a), iei/pcisa-lx(cs5536) and also on
TeleVideo TC7020, see
http://www.coreboot.org/pipermail/coreboot/2007-December/027973.html.
Signed-off-by: Nikolay Petukhov <nikolay.petukhov@gmail.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3196 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Somewhere characters are getting sign-extended, meaning that the
attributes of the drawing chars (>= 128) are wrong. Cast the value
before sending it to VGA.
Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3195 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
This decreases the size of the superiotool binary from ca. 1.1 MB to 600 KB.
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@3194 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Signed-off-by: Joseph Smith <joe@smittys.pointclark.net>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3193 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
bcd2dec()/dec2bcd() functions we'll need for (among other things)
converting some date/time parameters in CMOS.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Jordan Crouse <jordan.crouse@amd.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@3192 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
- Show PCI IDs as 4-digit numbers always.
- Cosmetic changes to make UI look more consistent.
- Drop MODULE_COUNT #define and use ARRAY_SIZE() where needed.
- Small fix to improve build system (create build/ when not there).
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@3191 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
strange reason (broken curses/VGA display, maybe others).
It'll be re-enabled when we fixed that.
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@3190 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
supported features configurable later (currently unused). Store all
build files and results (coreinfo.elf) in build/ now.
I'm self-acking this as (though non-"trivial") it doesn't affect coreinfo
in its functionality, this is more or less a "cosmetic" change to the
build interface ("make" -> "make menuconfig && make").
This is a kconfig checkout from the Linux kernel (where kconfig is being
actively maintained) from 03/2008. The hash identifying the last commit
to kconfig is 587c90616a5b44e6ccfac38e64d4fecee51d588c.
The amount of changes to kconfig itself has been kept as small as possible
to keep the diff small and to ease updating/porting to newer kconfig versions.
The following changes were performed on the upstream Linux kconfig:
- s/kernel/coreinfo/, and s/Linux/coreinfo/ in various strings.
- Consistently use the env. variable KERNELVERSION in all kconfig
interfaces -- e.g. config/menuconfig/gconfig/xconfig -- as version number.
- Hardcode our paths/filenames in some places (could be improved upstream).
- Always write .config and build/config.h, no matter which kconfig
interface is used (config/menuconfig/gconfig/xconfig). We want to
include build/config.h in our code.
- Adapt the kconfig Makefile for our purposes (build/ directory, rules, etc).
In addition, a few items in the coreinfo Makefile are needed for this to work.
This kconfig setup is successfully tested with all targets from 'make help':
config - Update current config utilising a line-oriented program
menuconfig - Update current config utilising a menu based program
xconfig - Update current config utilising a QT based front-end
gconfig - Update current config utilising a GTK based front-end
oldconfig - Update current config utilising a provided .config as base
silentoldconfig - Same as oldconfig, but quietly
randconfig - New config with random answer to all options
defconfig - New config with default answer to all options
allmodconfig - New config selecting modules when possible
allyesconfig - New config where all options are accepted with yes
allnoconfig - New config where all options are answered with no
For 'make defconfig' to work you have to do (which we don't need in coreinfo):
$ mkdir configs; touch configs/defconfig
You can also use 'make foo_defconfig' in which case kconfig will use a
file called 'configs/foo_defconfig' as basis.
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@3188 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Also, disable header() for now, as it's not being used.
Here are some stats on size differences:
- ls
23820 coreinfo.old.elf
23564 coreinfo.new.elf
- size *elf
text data bss dec hex filename
15199 2468 181904 199571 30b93 coreinfo.old.elf
14934 2468 181912 199314 30a92 coreinfo.new.elf
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@3187 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Uwe Hermann