The existing code did not set the northbridge throttle
values on Family 15h, leading to sporadic and random
deadlocks in the crossbar per AMD notes.
Properly set the northbridge throttle values on Family 15h.
Change-Id: I6304b63708c65fedb9c2d46b8c862b7f0adf1102
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/12025
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
The existing HyperTransport register configuration values were incorrect
in many spots. Apply the correct values from the BKDG on Family 10h and
Family 15h processors.
Change-Id: I009b6f478340e2dbfcda2b4534473d4397f9ecef
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/12022
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
This fixes Family 15h multiple package support; the previous code
hung in CAR setup and romstage when more than one CPU package was
installed for a variety of loosely related reasons.
TEST: Booted ASUS KGPE-D16 with two Opteron 6328 processors
and several different RDIMM configurations.
Change-Id: I171197c90f72d3496a385465937b7666cbf7e308
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/12020
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
On some multi-socket AMD platforms there are too many cores for all
APs to start up without stack collisions with either each other or
the BSP. On such platforms a larger amount of CAR memory is also
available.
Allow the maximum DCACHE size to be increased via a mainboard-
specific Kconfig flag.
Change-Id: I72ae8f7abeb9a83b57505469922818f9ec5bdf3f
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/12015
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
There were numerous issues surrounding AMD ECC initialization on
Family 15h processors due to the incomplete derivation from Family
10h MCT code. Bring the Family 15h ECC initialization and supporting
setup code in line with the BKDG recommendations.
Change-Id: I7f009b655f8500aeb22981f7020f1db74cdd6925
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/12003
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
This patch adds CC6 power save support to the AMD Family 15h
support code. As CC6 is a complex power saving state that
relies heavily on CPU, northbridge, and southbridge cooperation,
this patch alters significant amounts of code throughout the
tree simultaneously.
Allowing the CPU to enter CC6 allows the second level of turbo
boost to be reached, and also provides significant power savings
when the system is idle due to the complete core shutdown.
Change-Id: I44ce157cda97fb85f3e8f3d7262d4712b5410670
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/11979
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
NOTE: This commit switches CacheBase in CAR to use the DCACHE_RAM_BASE
Kconfig variable. There should be no functional difference between
the existing code and the new code, however hardware verfication is
encouraged on lesser used architectures such as AMD Geode.
Change-Id: Ia2e8f99be9df388e492a633c49df21ca1c57ba13
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/11970
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The additional local data storage requirements of the full DDR3
DRAM training algorithm make a BSP stack overrun a distint
possibility. Increase the BSP stack size to compensate.
Change-Id: I51af31442f2b77cb64a4b788751ccc7186acb283
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/11972
Tested-by: build bot (Jenkins)
Reviewed-by: Peter Stuge <peter@stuge.se>
The K8 PowerNow! state generator does not generate _PSS objects
for nodes other than the first CPU package. This patch backports
the PowerNow! core count fixes for Family 10h to the K8 CPUs.
Change-Id: I7b411ab75155dfb4bf51ae04301aa16fb2ae89f3
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/12286
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
TEST: Booted ASUS KGPE-D16 with single Opteron 6380
* Unbuffered DDR3 DIMMs tested and working
* Suspend to RAM (S3) tested and working
Change-Id: Idffd2ce36ce183fbfa087e5ba69a9148f084b45e
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/11966
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
It encourages users from writing to the FSF without giving an address.
Linux also prefers to drop that and their checkpatch.pl (that we
imported) looks out for that.
This is the result of util/scripts/no-fsf-addresses.sh with no further
editing.
Change-Id: Ie96faea295fe001911d77dbc51e9a6789558fbd6
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: http://review.coreboot.org/11888
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The existing Kconfig option for FIDVID was permanently
set to "no" due to Kconfig stopping at the first matching
value set when parsing the file. This patch moves the
conditional set above the unconditional set, resolving
the issue.
Change-Id: Ic19f68f6b17943f9133ff32a9b6538f0bf942eca
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/12224
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Backport a handful of debugging routines and the extended APIC
initialization code from Family 10h support to K8 support.
Change-Id: I08cc5c8bc65635ce09a69e32940dd7edd8d3be87
TEST: Booted ASUS KFSN4-DRE with 1x Opteron 8222
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/12251
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The recommendation to set DisFillP during CAR initialization
on K8 NPT CPUs was ignored. The consequences of this are
largely unknown; fix up coreboot to follow the recommendations.
Change-Id: Ide512bbc1d9aa284179628e2aa598ef5475e8eeb
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/12249
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
We use UNDERSCORE_CASE. For the MTRR macros that refer to an MSR,
we also remove the _MSR suffix, as they are, by definition, MSRs.
Change-Id: Id4483a75d62cf1b478a9105ee98a8f55140ce0ef
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/11761
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
The existing microcode update system used custom, manually generated microcode
blob files. This made updates very difficult. Update parser to use stock
microcode update files as provided by AMD.
Change-Id: I772b264ad167f2a5d629dab5d64d9b0ccab3a053
Signed-off-by: Audrey Pearson <apearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/11829
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
Using a copiler to compile something that's already a binary is pretty
stupid. Now that Stefan converted most microcode in blobs to a plain
binary, use the binary version.
Change-Id: Iecf1f0cdf7bbeb7a61f46a0cd984ba341af787ce
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/11607
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Current code written in C is calling a function implemented
in assembly. However, the symbol's visibility is not set
for such usage. Of course this works because MAINBOARDDIR/romstage.c
is being processed into an assembly file currently.
BUG=chrome-os-partner:44827
BRANCH=None
TEST=Built digitallogic/msm800sev while not changing romstage.c
into an assembly file.
Change-Id: I84c3af0026f3f98bc64af007aa7cc196429f4e5f
Signed-off-by: Aaron Durbin <adubin@chromium.org>
Reviewed-on: http://review.coreboot.org/11511
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
When building up which files to include in romstage there
were both 'cpu_incs' and 'cpu_incs-y' which were used to
generate crt0.S. Remove the former to settle on cpu_incs-y
as the way to be included.
BUG=chrome-os-partner:44827
BRANCH=None
TEST=Built rambi. No include file changes.
Change-Id: I8dc0631f8253c21c670f2f02928225ed5b869ce6
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11494
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
In the wake of the recent Intel "Memoy Sinkhole" exploit a code review
of the AMD SMM code was undertaken. While native Family 10h support
does not appear to be affected by the same SMM flaw, it also does not
require SMM to function. Therefore, the SMM memory range initialization
should only be executed if SMM will be used on the target platform.
Change-Id: I6531908a7724933e4ba5a2bbefeb89356197e8fd
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/11211
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
The sysinfo object within the k8 ram init is used
to communicate progess/status from all the nodes in the
system. However, the code was assuming where the sysinfo
object lived in cache-as-ram. The layout of cache-as-ram
is dynamic so one needs to do the lookup of the correct
address at runtime. The way the amd code is compiled
by #include'ing .c files makes the solution a little
more complex in that some cache-as-ram support code
needed to be refactored.
Change-Id: I6500fa7b005dc082c4c0b3382ee2c3a138d9ac31
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/10961
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Moves the K8 CPU_ADDR_BITS definition from socket to model.
Previously socket_F was not setting CPU_ADDR_BITS correctly.
Tested on Sun Ultra 40 M2 with two 2nd-gen Opterons w/ 2x4x2GiB DIMMs.
Most if not all K8-based chips support 40-bit physical addresses, with
possible exception of IA32-only K8-based Athlon XP-M chips.
Probably irrelevant, unless your machine has enough memory (at least 60 to
64GiB before MMIO hoisting) to exceed the CPU_ADDR_BITS default of 36 from
src/cpu/x86/Kconfig.
Change-Id: I01a2a59fa902280171840c36ca2e631476d3d603
Signed-off-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-on: http://review.coreboot.org/10963
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
Relevant for systems having processors that only have two (the minimum
and maximum) P-states, such as the Opteron 2210 at 1.0 and 1.8GHz.
Change-Id: Ic66fe6d10ce495c1bf21796cb7e1eb4e11e85283
Signed-off-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-on: http://review.coreboot.org/10910
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The prior ACPI _PSD generator committed in ef33db01 incorrectly assumed the active
link count of each processor was identical. Detect the link count on each node
when generating the _PSD objects.
Change-Id: Ic8aaa0728a43936cd4c6e1ed590e01ba8f0fbf9b
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/10158
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Caching SPD data during startup requires additional CAR space.
There was a large chunk of free space between the AP stack top and
the BSP stack bottom; moving the AP stacks below the BSP stack
allows this space to be utilized.
TEST: Booted ASUS KGPE-D16 with dual Opteron 6129 processors (16 cores)
and 120k of CAR.
Change-Id: I370ff368affde7061d6547527bda058b9016e977
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/10404
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
When increasing the number of supported CPUs on AMD Family 10h/15h
systems there is a relatively high chance of causing a collision
between the CAR global variable region and the AP stack space.
Such collision was noted when increasing the number of supported
CPUs to 32 on the ASUS KGPE-D16.
Detect collision at runtime and print a warning if collision is
present.
Change-Id: Ib5c32f868b1dfffb3b840bb1b1df5f55b5a25f8d
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/10401
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
