The correct id string for gm45 is "$VBT CANTIGA ".
This can be found in the gm45 option rom:
"strings vbios.bin | grep VBT".
Change-Id: Icd67a87dac774b4b3c211511c784c4fb4e2ea97c
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/16551
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
This implements "Keep VESA framebuffer" behavior on VGA output of gm45.
This patch reuses Linux code to compute vga divisors.
Change-Id: I2db5dd9bb1a7e309ca763b1559b89f7f5c8e6d3d
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/16338
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The intel x4x and gm45 have very similar integrated graphic devices.
Currently the x4x native graphic init enables VGA, while gm45 can output
on LVDS.
This patch reuses the x4x graphic initialisation code
to enable output on VGA in gm45 in a way that the behavior is similar to vbios:
If no VGA display is connected the internal LVDS screen is used.
If an external screen is detected on the VGA port it will be used instead.
Change-Id: I7e9ff793a5384ad8b4220fb1c0d9b28e6cee8391
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/16295
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Replace the use of the old device_t definition inside
northbridge/intel/nehalem.
The patch has been tested both with the arch/io.h definition of device_t
enabled and disabled in order to ensure compatibility while the
transaction takes place.
Change-Id: I6da4e0a9ef21b3285f4a369c8ddfbdb32a7a3801
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16406
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Replace the use of the old device_t definition inside
northbridge/intel/gm45.
The patch has been tested both with the arch/io.h definition of device_t
enabled and disabled in order to ensure compatibility while the
transaction takes place.
Change-Id: I87754799f922cf241fb456071bac04e6fe1eab34
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16402
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Replace the use of the old device_t definition inside
northbridge/amd/amdht.
Change-Id: I7dfb8f001504c691aeddf1bfbc3be05cc7d31ce4
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16468
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Replace the use of the old device_t definition inside
northbridge/amd/amdk8.
Change-Id: I5209dd309f0685f83d8a468c50309d5fda77973a
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16467
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Replace the use of the old device_t definition inside
northbridge/amd/amdfam10.
Change-Id: I5037feb31c51d06ccc672b0771d5d6e8c0dac949
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16466
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Move the funtion to find most significant bit set(fms)
and function to find least significant bit set(fls) to a common
place. And remove the duplicates.
Change-Id: Ia821038b622d93e7f719c18e5ee3e8112de66a53
Signed-off-by: Rizwan Qureshi <rizwan.qureshi@intel.com>
Reviewed-on: https://review.coreboot.org/16525
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Replace the use of the old device_t definition inside
northbridge/intel/x4x.
Change-Id: I65cd02eacf57cb41ded434582ca6e9d9f655e6ea
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16472
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Replace the use of the old device_t definition inside
northbridge/intel/i5000.
Change-Id: Ic049d882ef22f117ee52ba497351f548e2355193
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16471
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Replace the use of the old device_t definition inside
northbridge/intel/e7505.
Change-Id: Ie819f380ec06667e11bcff3e9e993126a86b2c89
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16469
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Replace the use of the old device_t definition inside
northbridge/intel/fsp_rangeley.
Change-Id: I4c1e6af64fe70211db2fafdba9f39182dfea66fc
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16470
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Previously the ME PCI interface (HECI) was being reported as present in
the DMAR ACPI table even when ME firmware was missing or the PCI device
was hidden and HECI would be unresponsive.
Now we check via the PCI config space itself to verify if the HECI
is present or not.
Note that this test could fail if ME firmware is present but
HECI is disabled in devicetree, because it would not advertise that the HECI
exists even though there is a running ME. Perhaps this behaviour is desirable
because in this case you won't see the HECI in the lspci tree anyway.
Change-Id: Ib692d476d85236b4886ecf3d6e6814229f441de0
Signed-off-by: Damien Zammit <damien@zamaudio.com>
Reviewed-on: https://review.coreboot.org/16330
Tested-by: build bot (Jenkins)
Reviewed-by: Swift Geek <swiftgeek@gmail.com>
Reviewed-by: Nico Huber <nico.h@gmx.de>
This is necessary for PCI express graphics card add-ons,
otherwise the pci allocator cannot fit the mmio for the
add on card into the space it has available and the OS
turns off the card. Old value was 1GiB.
Change-Id: I606994501b15e636fe209d1ed4b3d3f73b42bf5c
Signed-off-by: Damien Zammit <damien@zamaudio.com>
Reviewed-on: https://review.coreboot.org/16494
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
No more hang on DMI init when wait for DMI is re-enabled.
Previously the virtual channel arbitration table was not being
set up in the south/north bridges causing invalid DMI state.
This has been tested on GA-G41M-ES2L with patches following.
An NVIDIA GT218 card was detected by the OS and displayed using
the nouveau driver with no blobs.
Change-Id: I35e03c40f5f7aa4915afd5d26db7ab053abcf0cd
Signed-off-by: Damien Zammit <damien@zamaudio.com>
Reviewed-on: https://review.coreboot.org/16491
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
According to "G45: Volume 3: Display Register
Intel ® 965G Express Chipset Family and Intel ®
G35 Express Chipset Graphics Controller" some registries
are set incorrectly in gm45/gma.c.
Some values are changed after comparing them with the values
the i915 linux kernel (3.13 was used) module sets while modesetting.
The values were obtained using 'intel_reg' from intel-gpu-tools,
during a normal boot and with 'nomodeset' as a kernel argument.
Some registers that don't exist on gm45 are set in gma.c, which is
probably the result of copying code from a more recent intel
northbridge.
The result is that that gm45 laptops with wxga displays still work as
before. gm45 laptops with wxga+ or higher resolution now just work,
where previously a black screen was shown.
TEST: build with native graphic init and flash on a gm45 target, like
lenovo x200.
Change-Id: If66b60c7189997c558270f9e474851fe7e2219f1
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/16217
Tested-by: build bot (Jenkins)
Reviewed-by: Nico Huber <nico.h@gmx.de>
Remove an unusued function declaration that caused problems while
compiling the target.
Change-Id: Idfd73693e9b0e1777cafa4706113fde394e95795
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16435
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Replace the use of the old device_t definition inside
northbridge/intel/i945.
The patch has been tested both with the arch/io.h definition of
device_t enabled and disabled in order to ensure compatibility
while the transaction takes place.
Change-Id: I041c150a7b50261e26955ad9287ef05b9a06e412
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16371
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Replace the use of the old device_t definition inside
northbridge/intel/sandybridge.
The patch has been tested both with the arch/io.h definition of device_t
enabled and disabled in order to ensure compatibility while the
transaction takes place.
Change-Id: I35cc76ec7b6baa216666d06f6f325f43ac69067e
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/16409
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
On i945 the vram size is the default 8mb. It is also possible
to set it 1mb or 0mb hardcoding the GGC register in early_init.c
The intel documentation on i945, "Mobile Intel® 945 Express Chipset
Family datasheet june 2008" only documents those three options.
They are set using 3 bits. The documententation also makes mention
of 4mb, 16mb, 32mb, 48mb, 64mb but not how to set it.
The other non documented (straight forward) bit combinations allow
to change the VRAM size to those other states.
What this patch does is:
- add those undocumented registers with their respective vram size to
the i945 NB code;
- make this a cmos option on targets that have this northbridge.
TEST: build, flash to target, set cmos as desired and boot linux.
On Debian it can be found using "dmesg | grep stolen".
NOTE: dmesg message about reserved vram are quite different depending
on linux version
Change-Id: Ia71367ae3efb51bd64affd728407b8386e74594f
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/14819
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
This allows to set the preallocated memory for the IGD on x4x
using a cmos option.
If no cmos option is found a default value of 64M is used.
TESTED most options on ga-g41m-es2l with 2G dimm in one slot and 2x2G.
352M also works in contrast with gm45 where it is known to cause issues
with certain ram combinations.
Change-Id: I9051d080be82f6dfab37d353252e29b2ed1fca7f
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/15492
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The Intel documtentation, "Intel ® 4 Series Chipset Family"
mentions the possibility of 1, 4, 8 and 16M of preallocated
memory for the IGD, but does not document this.
This allows to set those undocumented values.
TESTED on ga-g41m-es2l with 2G dimm in one slot and 2x2G.
Change-Id: I92beb8d78907d4514a5aaf69248dd607dcf227c0
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/15491
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
With VBOOT_VERIFY_FIRMWARE separated from CHROMEOS, move recovery and
developer mode check functions to vboot. Thus, get rid of the
BOOTMODE_STRAPS option which controlled these functions under src/lib.
BUG=chrome-os-partner:55639
Change-Id: Ia2571026ce8976856add01095cc6be415d2be22e
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/15868
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Now hardcode maximum memory frequency capability to 800MHz, as
all chipsets in x4x family support PC2-6400 according to the datasheet.
CAS latency detection also relies on this, and has been cleaned up.
Ram initialization does not work with FSB 1333MHz / DDR2 800MHz combination,
so disable this combination for now, and reduce to 667MHz instead.
Still don't know why this is the case, but FSB1333/667 works.
These changes should now allow existing configurations to continue working,
while providing support for previously unworking configurations, due to
previous buggy CAS latency detection code.
TESTED: on GA-G41M-ES2L
CPU: E5200 @ 2.50GHz (FSB 800MHz)
2x 1GB 667MHz hynix worked @ 667
1x 2GB 800Mhz ARAM worked @ 800
1x 1GB 667Mhz StarRam worked @ 667
2x 2GB 800Mhz (generic) worked @ 800
Change-Id: I1ddd7827ee6fe3d4162ba0546f738a8f9decdf93
Signed-off-by: Damien Zammit <damien@zamaudio.com>
Reviewed-on: https://review.coreboot.org/15818
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Match the definition and use of these variable with haswell, such that
DCACHE_RAM_MRC_VAR_SIZE is not included in DCACHE_RAM_SIZE.
Change-Id: I5af20f63cd0cb631d39f7c7fe0e2a99ebd3ce986
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/15761
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Fix and use the failsafe CAS detection logic rather than
recalulating the values from raw SPDs.
Tested on GA-G41M-ES2L with 2x2GB DDR2-800 DIMMs
(which worked before and still work)
Change-Id: I6af0f1705d099f7bcbff8c9baa94a68dae689e01
Signed-off-by: Damien Zammit <damien@zamaudio.com>
Reviewed-on: https://review.coreboot.org/15726
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr>
If S3 support was implemented for this platform later on, use
romstage handoff structure instead.
Change-Id: I03c1e07a7fcc17c27203d0c4e32e3958f2ba5273
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/15716
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Damien Zammit <damien@zamaudio.com>
If S3 support was implemented for this platform later on, use
romstage handoff structure instead.
Change-Id: Ib0cf3ad41753baee26354c5ed19294048e7fb533
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/15715
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Damien Zammit <damien@zamaudio.com>
Previously, any 800MHz DIMMs were being slowed to 667MHz
for no reason other than there was a bug in the maximum
frequency detection code for the MCH.
Change-Id: Id6c6c88c4a40631f6caf52f536a939a43cb3faf1
Signed-off-by: Damien Zammit <damien@zamaudio.com>
Reviewed-on: https://review.coreboot.org/15257
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Broken with commit:
5c10abe nb/intel/sandybridge: increase MMCONF_BASE_ADDRESS
Available sandybridge/systemagent-r6.bin has MMCONF hard-coded
at some places and samsung/lumpy fails at boot here:
CBFS: Locating 'mrc.bin'
CBFS: Found @ offset 9fec0 size 2fc94
System Agent: Starting up...
System Agent: Initializing
These are the last lines as captured over USB debug.
Change-Id: I441847f0e71a5e1be9c8ef6a04a81eb7bdd8a6d9
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/15328
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Without RELOCATABLE_RAMSTAGE have WB cache large enough
to cover the greatest ramstage needs, as there is no benefit
of trying to accurately match the actual need. Choose
this to be bottom 16MiB.
With RELOCATABLE_RAMSTAGE write-back cache of low ram is
only useful for bottom 1MiB of RAM as a small part of this gets used
during SMP initialisation before proper MTRR setup.
Change-Id: Icd5f8461f81ed0e671130f1142641a48d1304f30
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/15249
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Instead of programming unsupported CAS use the highest supported
value. Start at DDR3 maximum of CAS 18T.
Increase error message verbosity level.
Useful for overclocking.
Tested on Lenovo T520 and DDR3-1600 DIMM (RMT3170eb86e9w16).
Allows to run a DDR3-1600 DIMM at 933Mhz.
Change-Id: I2e8aadd541f06fa032ad7095c9a2d5e3bb7613f3
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/15217
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Alexander Couzens <lynxis@fe80.eu>
Calculate the value from current DDR frequency.
Tested on Lenovo T520 and DDR3-1600 DIMM (RMT3170eb86e9w16).
Change-Id: I57ffbfeb291fc2fede278d18527993e7432e9bd8
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/15184
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Alexander Couzens <lynxis@fe80.eu>
Set max_mem_clock_mhz in devicetree to 933Mhz.
Allows to run the memory at up to DDR3-1866.
The same frequency was allowed within the first vendor bios,
but Lenovo than decided to limit it to DDR3-1333.
Tested on Lenovo T520 and DDR3-1600 DIMM (RMT3170eb86e9w16).
The RAM is now running at DDR3-1600 instead of DDR3-1333.
This gives about 4% performance increase in glmark2 using the
Intel GPU.
Change-Id: If15be497402d84a2778f0434b6381a64eda832d6
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/15158
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
This is more of ACPI S3 resume and x86 definition than CBMEM.
Change-Id: Iffbfb2e30ab5ea0b736e5626f51c86c7452f3129
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/15190
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Instead of hardcoding the PCI mmio size read it from devicetree.
Set a default value of 2048 MiB and 1024MiB for laptops without
discrete graphics.
Tested on Sandybridge Lenovo T520.
Change-Id: I791ebd6897c5ba4e2e18bd307d320568b1378a13
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/15140
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Remove code duplication and use the common function
store_current_mrc_cache instead.
No functionality is changed.
Tested on Sandybridge Lenovo T520.
Change-Id: I4aa5463f1b1d5e1afbe44b4bfc659524d86204db
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/15074
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
Previously, 0x0 was the value being used for an unpopulated dimm
on spd[62], however some DDR2 dimms have 0x0 as a valid value.
Now use 0xff which is an unused value even on DDR2/DDR3.
Change-Id: I55a91a6c3fe3733a7bb2abc45ca352c955c07c99
Signed-off-by: Damien Zammit <damien@zamaudio.com>
Reviewed-on: https://review.coreboot.org/15058
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
Setting the size of the preallocated memory for the igd is done
using a cmos parameter, gfx_uma_size. This was limited to a subset of
all available sizes, that were already implemented elsewhere
in the northbridge code.
What this does is change the cmos parameter to 4 bits instead
of 3 bits to accomodate all vram sizes.
It also adds a sane default of 32mb that already was in place.
The northbridge code that reads this cmos parameter is
also changed for this new cmos settings.
352M is disabled since it causes issues on systems with 4GB or more ram.
TEST: Build, flash target. Clear cmos by corrupting
the checksum (nvramtool -c something).
Set a desired value in gfx_uma_size using nvramtool.
"dmesg | grep stolen" to see what is actually allocated.
Change-Id: Ia6479d03f1abe6d0c94bd7264365505e8f8eaeec
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/14900
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
The declarations of CFG_ evaluate to correct values only when
included after the definitions of BLDCFG_ in buildOpts.c.
So we never have CFG_PLAT_NUM_IO_APICS defined here.
Change-Id: I94b3dee5a3207b37921eb24a0bcd73b5a217b2d3
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/14887
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The values were obtained from vendor bios at runtime.
I am not 100% sure of the sequence required to initiate them,
but guessed from the gm45 code. There may be some status bytes
needed to be polled during the sequence that is missing,
but as I don't have bios writer's datasheet it's very hard
for me to know.
Change-Id: Idd205e0bab5f75e01c6e3a5dc320c08639f52db8
Signed-off-by: Damien Zammit <damien@zamaudio.com>
Reviewed-on: https://review.coreboot.org/14925
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The existing DIMM size calculation for DDR3 was incorrect. Use
the recommended calculation from the DDR3 SPD specification.
Change-Id: Id6a39e2b38b5d9f483341ebef8f2960ae52bda6c
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14739
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
While some stubs existed before this patch to handle non-ECC
memory initialization, there were a number of ECC detect unaware
sections of code. Add ECC support detection to those sections.
Change-Id: I56dad8a0f6833b2f42796212afb9777e9cc73d6d
Tested-On: ASUS KGPE-D16
Tested-With: 1x Opteron 6262
Tested-With: 1x SuperTalent 4G non-ECC DIMM in slot A2
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14737
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Damien Zammit <damien@zamaudio.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Mask out the bit that doesn't fit in 32bits, so gcc 6.1 is happy
Change-Id: I13e2b41742206b8d86b90314b80cc324c00ae637
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: https://review.coreboot.org/14639
Reviewed-by: Damien Zammit <damien@zamaudio.com>
Tested-by: build bot (Jenkins)
gcc 6.1 complains that SMM_OFFSET << 8 is larger than the register
it is assigned to (rightly so):
src/northbridge/amd/gx2/northbridgeinit.c:196:23: error: result of
'1077936128 << 8' requires 40 bits to represent, but 'int' only
has 32 bits [-Werror=shift-overflow=]
msr.lo = (SMM_OFFSET << 8) & 0xfff00000;
^~
Change-Id: Ib0d669268202d222574abee335a6a65c8a255cc7
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: https://review.coreboot.org/14617
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The PLL will never lock if the requested frequency is already set.
As the fallback may request the same frequency again exit early
to prevent a hang.
Test system:
* Gigabyte GA-B75M-D3H
* Intel Pentium CPU G2130
Change-Id: I625b2956346d8c50cca84def6190c076bf99dbec
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/14174
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
Add the following fallbacks:
* Try decreasing clock frequency.
In case of DDR1600 the next possible value of DDR1333 is being used.
* Try decreasing clock frequency.
In case of DDR1333 the next possible value of DDR1066 is being used.
* Disable failing channel.
The system may be able to boot with a single channel enabled.
The fallbacks are untested.
Change-Id: I3be7034ad25312b3ebf47a54f335a3893f8d7cc1
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/14173
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The LVDS port is configured to accept data from pipe A, but the panel
fitter and VGA were attached to pipe B.
Changes to VGACNTRL:
- select pipe A instead of pipe B.
- disable VGA centering to fix jitter.
TEST=Build and run on Thinkpad X200 in both text and framebuffer modes.
Change-Id: I2356f264580d8b021952c217de3477291d866f98
Signed-off-by: Nick High <nhigh@openmailbox.org>
Reviewed-on: https://review.coreboot.org/14524
Tested-by: build bot (Jenkins)
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
The ECC check bits of all ECC DIMMS were inadvertently initialized
twice in the same routine, significantly delaying startup. Part
of this was related to an obsolete MCA workaround that has been
fixed through multiple commits, therefore the workaround is no
longer needed.
Only initialize the ECC check bits once.
Change-Id: I90ac1147d9b006794d29b866a9cb5b7ead8f01e7
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14503
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
During receiver enable cycle training on Family 15h the entire range
of possible delays is searched, even though the single passing window
is often found nearly immediately. Skip the remainder of the delay
range after the passing window has been located.
Change-Id: If98217fa8e7de77366762d3c7bb01049a1dc080f
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14544
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Martin Roth <martinroth@google.com>
During DQS receiver enable cycle training on Family 15h platforms the
read data timing registers were inadvertently set to zero on every
lane training attempt.
Ensure that the read data timing registers are correctly set after
each lane is trained in receiver enable cycle training. This allows
more than one RDIMM to function on a given DCT channel.
Change-Id: I87d732f0383e9785a73b57e6f48855f3e872f1f9
Tested-On: ASUS KGPE-D16
Tested-With: 1x Opteron 6262HE
Tested-With: 4x Crucial 36KSF1G72PZ-1G6M1 (slots A2 / A1 / B2 / B1)
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14543
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Martin Roth <martinroth@google.com>
The existing Family 15h receiver enable training code stored
temporary delay values in the wrong variables, leading to
the requisite averaging of delays across nibbles not being
applied. This in turn made x4 DIMMs less stable than they
should have been.
Store temporary nibble delay values in a dedicated array.
Change-Id: Ic5da898af7d689db4110211f89b886ccdbb5f78f
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14541
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Martin Roth <martinroth@google.com>
Fix regression introduced by:
Ib48fe8380446846df17d37b22968f7d4fd6b9b13
Don't run channel_test on S3 resume as it overrides memory
that might be in use.
Fixes MCE events reported by the GNU/Linux kernel that
low memory has been modified.
Reset on failed s3 resume.
Change-Id: Ibadea286619c7906225f86a93aaa0b4caf26cabe
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/14439
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
DIMM training can sporadically fail due to external influences or various
errata. In these cases, restarting to retry training is a more appropriate
response than halting the system and requiring manual intervention.
Change-Id: Id49f7419f56e0640a84448cc06ecbaf62bed145e
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14529
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
The wrong DIMM number was used in the initial non-target MRS
setup routines. This had no functional impact other than to
print the wrong DIMM number in the DDR3 verbose debug output.
Change-Id: I480118ed00e1786a06e641a56f0fb19cd87f92eb
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14501
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Martin Roth <martinroth@google.com>
The existing RDIMM RC control word send routines were a hodgepodge
of various AGESA chunks with different ways of handling the same
task. Unify the control word chip select setup, use precise timing
routines on Family 15h, fix a couple of incorrect masks, and add
additional debugging statements.
It is believed that this patch is cosmetic and does not significantly
alter existing functionality.
Change-Id: Ie4ec7b6a7be7fce09e89f9eec146cc98b15b6160
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14500
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Martin Roth <martinroth@google.com>
When more than one DIMM is installed on a DCT, only the first DIMM
delay values are scaled to the new memory clock frequency after a
memory clock change during write leveling.
Store the previous memory clock of each DIMM during write leveling
to ensure that every DIMM has its delay values rescaled.
Change-Id: I56e816d3d3256925598219d92783246f5f4ab567
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14479
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Martin Roth <martinroth@google.com>
After substantial testing it has been determined that it is neither
required nor safe to disable the DRAM MCA during initial startup.
This (mostly) reverts commit c094d99611.
The minor debugging enhancements from that commit were left in place.
Tested-On: ASUS KGPE-D16
Config-CPU: 1x Opteron 6262HE
Config-RAM: 4x Crucial 36KSF1G72PZ-1G6M1
Config-RAM: 1x Kingston 9965516-483.A00LF
Change-Id: I58fcc296b8c45ecaedf540951c365e4ce52baaf5
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14446
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Martin Roth <martinroth@google.com>
Certain RDIMMs have inherently large write levelling delays,
in some cases exceeding 1.5 MEMCLK. When these DIMMs are
utilized, the phase recovery system requires special handling
due to the resultant offset exceeding the phase recovery reporting
capabilities.
Fix an old error where delays > 1.5 MEMCLK were not being programmed
(gross delay high bit was not in set range), and restore special
delay handling for delays greater than 1.5 MEMCLK.
Also enhance debugging for x4 DIMMs around the affected code.
Tested-On: ASUS KGPE-D16
Config-CPU: 1x Opteron 6262HE
Config-RAM: 4x Crucial 36KSF1G72PZ-1G6M1
Change-Id: I0fb5454c4d5a9f308cc735597607f095fe9188db
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14441
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The BKDG requires phy fences to be re-trained after a memory clock change.
Memory training on the ASUS KGPE-D16 and KCMA-D8 somehow "mostly" worked
-- without actually following this requirement -- !
Fix the single typo that caused several weeks of delay in putting
servers with Kingston RAM (and others) into production...
Tested-On: ASUS KGPE-D16
Config-CPU: 1x Opteron 6262HE
Config-RAM: 4x Crucial 36KSF1G72PZ-1G6M1
Change-Id: I197e6728d2b0ac8c1535740599459d080b17af33
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: https://review.coreboot.org/14445
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Martin Roth <martinroth@google.com>
We never define B1_IMAGE or B2_IMAGE. These are about building
CIMx as separate binary modules, while coreboot builds these into
same romstage or ramstage module.
Change-Id: I9cfa3f0bff8332aff4b661d56d0e7b340a992992
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/14393
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Kerry Sheh <shekairui@gmail.com>
