Void pointer arithmetics are forbidden in standard C but GCC has
an extension that allows it.
Change-Id: I43029b2ab2f7709b8e1ba85eb05c31341b8ac16f
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/18293
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
Reviewed-by: Nico Huber <nico.h@gmx.de>
This also selects RELOCATABLE_RAMSTAGE and
CACHE_RELOCATABLE_RAMSTAGE_OUTSIDE_CBMEM by default on Haswell.
Change-Id: I50b9ee8bbfb3611fccfd1cfde58c6c9f46b189ca
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/18232
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
The selection of the SSC reference frequency for LVDS was based on a
completely unrelated clock.
The `ssc_freq` flag should be set when the SSC reference runs at a
different frequency than the general display reference clock (DREF).
For most platforms, there is no choice, i.e. for i945 and gm45 the SSC
reference always differs from the display reference clock (i945: 66Mhz
SSC vs. 48MHz DREF; gm45: 100MHz SSC vs. 96Mhz DREF), for Nehalem and
newer, it's the same frequency for SSC/non-SSC (120MHz). The only,
currently supported platform with a choice seems to be Pineview, where
the alternative is 100MHz vs. the default 96MHz.
Change-Id: I7791754bd366c9fe6832c32eccef4657ba5f309b
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: https://review.coreboot.org/18186
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The results were obtained by comparing the MCHBAR registers of vendor bios
with coreboot at the same dram timings.
This fixes 2 issues:
* 1333MHz fsb CPUs were limited to 667MHz ddr2 speeds, because with
800MHz raminit failed;
* 1067MHz fsb CPUs did not boot when second dimm slot was populated.
TESTED on ga-g41m-es2l on 800, 1067 and 1333MHz CPUs with
DDR2 667 and 800MHz dimms.
Change-Id: I70f554f97b44947c2c78713b4d73a47c06d7ba60
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/18022
Tested-by: build bot (Jenkins)
Reviewed-by: Nico Huber <nico.h@gmx.de>
The code to set the igd frequencies is written with the mobile version
of the 945 chipset in mind and seems to cause cause strange igd
related problems on the desktop versions.
Some possible problems are:
* on 800MHz fsb CPUs the igd sometimes has artifacts on the screen;
* on 800MHz fsb CPU memtest results vary a lot;
* since a commit 45e11aa0a5 "Add/Combine Broadwell Chromebooks using
variant board scheme" that does not affect this northbridge, the
display shows garbage as soon as Linux (4.8) modesets the display.
A fix is to hardcode the core display and render clocks to their
maximum, potentially also improving graphical performance.
Vendor bios on all boards in coreboot with this northbridge have the
same value in this PCI config address.
TESTED on P5GC-MX (display works fine again in Linux) and
user reports of it making GA-945GCM-S2L run more stable.
Change-Id: I8b046edbc952631d9b79023e3d385160ff682c24
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/17981
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Nothing from that header is used or even declared since
CONFIG_HYPERTRANSPORT_PLUGIN_SUPPORT is not selected on Intel
hardware.
Change-Id: I9101eb6ffa6664a2ab45bc0b247279c916266537
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/18044
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Idwer Vollering <vidwer@gmail.com>
This is more consistent with newer Intel targets.
This a static struct so it is initialized to 0 by default.
To make it more readable:
* only setting to GPIO mode is made explicit;
* only pins in GPIO mode are either set to input or output since this
is ignored in native mode;
* only output pins are set high or low, since this is read-only on
input;
* blink is only operational on output pins, non-blink is not set
explicitly;
* invert is only operational on input pins, non-invert is not set
explicitly.
Change-Id: I05f9c52dee78b7120b225982c040e3dcc8ee3e4e
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/17639
Tested-by: build bot (Jenkins)
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Values based on vendor bios and suggested by Arthur Heymans for FSB1067.
FSB1067:
The ratio 1067/800 is proportional to the ratio of EPBAR32(0x2c) bits:
0x1a / 0x14 ~ 1067/800
EPVC1IST:
The ratio is also proportional to FSB ratios: 0x9c / 0xf0 ~ 533/800.
Change-Id: Ib90e8ea1b82f2fcc3b5c199cace32a7f0aff4b5c
Signed-off-by: Elyes HAOUAS <ehaouas@noos.fr>
Reviewed-on: https://review.coreboot.org/17198
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Having same memory region set as both WRPROT and WRBACK
using MTRRs is undefined behaviour. This could happen if
we allow DCACHE_RAM_BASE to be located within CBFS in SPI
flash memory and XIP romstage is at the same location.
As SPI master by default decodes all of top 16MiB below
4GiB, initial cache-as-ram line fills may have actually
read from SPI flash even in the case DCACHE_RAM_BASE was
below the nominal 4GiB - ROM_SIZE.
There are no reasons to have this as board-specific setting.
Change-Id: I2cce80731ede2e7f78197d9b0c77c7e9957a81b5
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17806
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Currently only there is only one eaglelake board in coreboot
(ga-g41m-es2l) featuring a G41 variant northbridge.
Adding boards with a different variant (Q43, Q45, G43, G45, B43) will
require this change for graphic initialisation.
Change-Id: Ida32c563a99576b66685dfdadf9a534fd6e197dc
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/17900
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Add custom files for Sandybridge and IvyBridge functions.
Move only the minimal required functions into separate files.
Both files' functions are going to call raminit_common functions.
No functionality is changed.
Sandybridge code path tested on Lenovo T420.
Change-Id: I1b1dfbd0857b59d3ae4392b73c033ee7a5aed243
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/17605
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
This is a follow-on patch to commit 10141c30 -
(nb/intel/gm45: Use LAPIC udelay instead of custom version)
which removed the custom udelay from everywhere except SMM.
This patch removes it from SMM as well, and gets rid of the
gm45/delay.c file.
Change-Id: I7970bb5205f4aa10b38172ab5b9f8bcd6766c4e7
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/17330
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Instead of hardcoding pci_mmio_size in the raminit code,
this makes it a parameter in the devicetree.
A safe minimum of 768M is also defined since using anything
less causes problems (if 4G of ram is used).
Change-Id: If004c861464162d5dbbc61836a3a205d1619dfd5
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/16856
Tested-by: build bot (Jenkins)
Reviewed-by: Nico Huber <nico.h@gmx.de>
Don't use scratchpad registers when we have romstage_handoff
to pass S3 resume flag. Also fixes console log from reporting
early in ramstage "Normal boot" while on S3 resume path.
Change-Id: I2f1f05ef4fc640face3d9dc92d12cfe4ba852566
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17676
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Don't use scratchpad registers when we have romstage_handoff
to pass S3 resume flag. Scratchpad register was read too
late in ramstage so acpi_is_wakeup_s3() did not evaluate
correctly.
This fixes low memory corruption at 0x1000-0x102c and the lack
of coreboot tables (util/cbmem not working) after S3 resume.
This also fixes console log from reporting early in ramstage
"Normal boot" while on "S3 resume" path.
Change-Id: I2922a15a90d2f8272c3482579bdd96f8f33e9705
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17675
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Don't use scratchpad registers when we have romstage_handoff
to pass S3 resume flag. Also fixes console log from reporting
early in ramstage "Normal boot" while on S3 resume path.
Change-Id: I4e2eabc59ff87b7ed40cfc9885bbe0256fe4a695
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17674
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Adapt implementation from skylake to prepare for removal of
HIGH_MEMORY_SAVE and moving on to RELOCATABLE_RAMSTAGE.
With this change, CBMEM region is set early-on as WRBACK
with MTRRs and romstage ram stack is moved to CBMEM.
Change-Id: Idee5072fd499aa3815b0d78f54308c273e756fd1
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/15791
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This makes CHIPSEC happy. We don't enable PAVP, but it shouldn't hurt
to lock it nevertheless.
Change-Id: I9428f0b6e8868832eb79f7aea24cbc7961c2aa8f
Signed-off-by: Dennis Wassenberg <dennis.wassenberg@secunet.com>
Reviewed-on: https://review.coreboot.org/17352
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Resource is actually stored even before read_resources, but
that's where we currently log this resource.
For Intel, use PCI config register offset as the resource
index, while AMD side uses MSR address.
Change-Id: I6eeef1883c5d1ee5bbcebd1731c0e356af3fd781
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17696
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Doing PCI config operations via MMIO window by default is a
requirement, if supported by the platform. This means chipset
or CPU code must enable MMCONF operations early in bootblock
already, or before platform-specific romstage entry.
Platforms are allowed to have NO_MMCONF_SUPPORT only in the
case it is actually not implemented in the silicon.
Change-Id: Id4d9029dec2fe195f09373320de800fcdf88c15d
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17693
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Resource allocator and 64-bit PCI BARs will need it and
PCI use is not really restricted to x86.
Change-Id: Ie97f0f73380118f43ec6271aed5617d62a4f5532
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17733
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
MMCONF was explicitly used here to avoid races of 0xcf8/0xcfc access
being non-atomic and/or need to access 4kiB of PCI config space.
All these platforms now have MMCONF_SUPPORT_DEFAULT.
I liked the style of code in pci_mmio_cfg.h more, and used those to
replace the ones in io.h.
Change-Id: Ib5e6a451866c95d1edb9060c7f94070830b90e92
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17689
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
MMCONF was explicitly used here to avoid races of 0xcf8/0xcfc access
being non-atomic and/or need to access 4kiB of PCI config space.
All these platforms now have MMCONF_SUPPORT_DEFAULT.
Change-Id: I943e354af0403e61263f1c780f02c7b463b3fe11
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17529
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Boards with this chipset do not have any reference of
MMCONF_BASE_ADDRESS being written to chipset registers.
Either board support is already broken or FSP takes
care of this early and Kconfig lacks the notice that
this parameter must match with the chosen FSP binary.
CPU bootblock associated with this chipset uses
exclusive PCI IO access already.
Untested.
Change-Id: I07d20d81266ff6aaa6384d20a806d52fd4568e08
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/17547
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Split raminit.c into smaller parts. Move all functions that will
be used by chip-specific code into raminit_common.c.
The chip-specific changes includes new configuration values
for IvyBridge and 100Mhz reference clock support, including new
frequencies.
No functionality is changed.
Tested on Lenovo T420.
Change-Id: If7bb5949f4b771430f3dba1b754ad241a7e8426b
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/17604
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The same pattern was being used throughout the code base
for initializing the romstage handoff structure. Provide
a helper function to initialize the structure with the S3
resume state then utilize it at all the existing call sites.
Change-Id: I1e9d588ab6b9ace67757387dbb5963ae31ceb252
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/17646
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Building an image for the Lenovo X60 on Debian 8.5 (jessie) with GCC 4.9.2,
compilation fails with the error below.
```
$ gcc --version
gcc (Debian 4.9.2-10) 4.9.2
[…]
$ make # lenovo/x60 with native graphics initialization
[…]
CC ramstage/northbridge/intel/i945/gma.o
src/northbridge/intel/i945/gma.c: In function 'probe_edid':
src/northbridge/intel/i945/gma.c:570:2: error: 'for' loop initial declarations are only allowed in C99 or C11 mode
for (int i = 0; i < 8; i++) {
^
src/northbridge/intel/i945/gma.c:570:2: note: use option -std=c99, -std=gnu99, -std=c11 or -std=gnu11 to compile your code
Makefile:316: recipe for target 'build/ramstage/northbridge/intel/i945/gma.o' failed
make: *** [build/ramstage/northbridge/intel/i945/gma.o] Error 1
```
Fix this by declaring the count variable outside the 'for' loop.
Change-Id: Icf69337ee46c86bafc4e1320fd99f8f8f5155bfe
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: https://review.coreboot.org/17623
Tested-by: build bot (Jenkins)
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
The code won't allow anything beyond CL11 due to short
CAS Latency mask and a bug in mr0 which had the wrong
bit set for CL > 11.
Increase the CAS bitmask, fix the mr0 reg to allow CAS Latencies
from CL 5 to CL 18.
Use defines instead of hardcoding min and max CAS latencies.
Tested on X220 with two 1866 MHz, CL13 memories
Tested-By: Nicola Corna <nicola@corna.info>
Change-Id: I576ee20a923fd63d360a6a8e86c675dd069d53d6
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/17502
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
This allows to set the backlight PWM frequency and the
duty cycle in the devicetree instead of using a plain BLC_PWM_CTL
value.
Change-Id: I4d9a555ac7ea5605712c1fcda994a6fcabf9acf3
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/17597
Tested-by: build bot (Jenkins)
Reviewed-by: Nico Huber <nico.h@gmx.de>
Fix-up for 696abfc
nb/intel/x4x: Fix and deflate `dimm_config` in raminit
It didn't fix the channel-number shifting issue as intended.
The channel index is either 0 or 1. DIMMs are counted from 0
to 3 where 0..1 covers channel 0, and 2..3 covers channel 1.
Since we have two DIMMs per channel, we have to multiply the
channel index by 2 (or shift it left by 1) to get the index
of the first DIMM in the channel. Finally, to get the offset
of a DIMM in the channel we take its index modulo 2 (again,
the number of DIMMs per channel).
Change-Id: I2784b0cb655bfe823bf5fa48b722623dfca1ddc3
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/17612
Tested-by: build bot (Jenkins)
Reviewed-by: Damien Zammit <damien@zamaudio.com>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
We kept this value at it's default on the native graphics init path.
Maybe the Video BIOS path, too, I don't know if the VBIOS sets it.
The panel power sequencer uses the core display clock (CDCLK). It's
based on the HPLLVCO and a frequency selection we made during raminit.
The value written is the (actual divisor/2)-1 for a 100us timer.
v2: Fix unaligned mmio access inherited from Linux.
v3: Use MCHBAR8() instead. Also, the unaligned access might have
worked after all.
Change-Id: I877d229865981fb0f96c864bc79e404f6743fd05
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/17619
Tested-by: build bot (Jenkins)
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Some methods like discover_402x assume an clear state.
Should fix fallback attempt raminit failures.
Change-Id: I7a6fe044c17f5e0dbfa0e9b9d2aed0c3b6ae3972
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/17471
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
