CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode.
Factor out some common expressions.
Add an error message when coreboots hangs waiting for a pstate
that never comes (it happened to me), and throw some
paranoia at it for good mesure.
If I understood BKDG fam10 CPUs never need a software initiated vid transition,
because the hardware knows what to do when you just request
a Pstate change if the cpu is properly configured. In fact
unifying a little what PVI and SVI do was better for my board (SVI).
So I drop transitionVid, which I didn't understand either (why
did it have a case for PVI if it is never called for PVI ?
Why did the PVI case distinguigh cpu or nb when PVI is
theoretically single voltage plane ? ).
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6401 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode.
Contemplate the possibility of nbCofVidUpdate not being
defined, trying to get closer to BKDG
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6399 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode.
Configuration of F3x[84:80] was hardcoded for rev B.
I change that for some code that checks for revision
and configures according to BKDG. Unfinished but
hopefully better than it was.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6398 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode.
BKDG says nbSynPtrAdj may also be 6 sometimes.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6397 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode.
I didn't understand quite why it did that iwth F3xA0 (Power
Control Misc Register) so I moved Pll Lock time to rules in defaults.h
and reimplemented F3xA0 programming. A later patch will remove
a part I don't know what's mean to do.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6396 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode.
Bring F3xD4 (Clock/Power Control Register 0) more in line
with BKDG i more cases. It requires looking at the CPU package type
so I add a function for that (in the wrong place?) and some
new constants
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6395 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode).
No change of behaviour intended.
Refactor FAM10 fidvid . Factor out the decision whether
to update northbridge frequency and voltage because there
was the same code in 3 places and so we can later modify it
in one place.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6394 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode).
No change of behaviour intended.
Refactor FAM10 fidvid . Factor out the decision whether
to update northbridge frequency and voltage because there
was the same code in 3 places and so we can later modify it
in one place.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6393 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode).
No change of behaviour intended.
Refactor FAM10 fidvid. Factor out a little common code.
Also, our earlier config_clk_power_ctrl_reg0
was still too long and it'd get longer with forthcoming patches.
We now take apart F3xD4[PowerStepUp,PowerStepDown]
to its own function.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6392 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode).
No change of behaviour intended.
Refactor FAM10 fidvid . prep_fid_change was already long and it'd
get longer with forthcoming patches. We now take apart F3x[84:80],
ACPI Power State Control Registers, to its own function.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6391 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode).
No change of behaviour intended.
Refactor FAM10 fidvid . prep_fid_change was already long and it'd
get longer with forthcoming patches. We now take apart F3xDC[NbsynPtrAdj],
Northbridge/core synchronization FIFO pointer adjust, to its own function.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6390 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode).
No change of behaviour intended.
Refactor FAM10 fidvid . prep_fid_change was already long and it'd
get longer with forthcoming patches. We now take apart F3xA0,
Power Control Misc Register to its own function.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6389 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode).
No change of behaviour intended.
Refactor FAM10 fidvid . prep_fid_change was already long and it'd
get longer with forthcoming patches. We now take apart F3xD4,
Clock Power/Timing Control 0 to its own function.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6388 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
CPU and northbridge frequency and voltage
handling for Fam 10 in SVI mode). No change of behaviour intended.
Refactor FAM10 fidvid . prep_fid_change was already long and it'd
get longer with forthcoming patches. We now take apart VSRamp in step b
of 2.4.1.7 BKDG to its own function.
Signed-off-by: Xavi Drudis Ferran <xdrudis@tinet.cat>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6387 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Also it enables the FID/VID changes in SB. Jakllsch had some troubles with that too but on am2 CPU. Those bits are only documented in SB600. They arent in RRG RPR and BDG.
Signed-off-by: Rudolf Marek <r.marek@asssembler.cz>
Acked-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6381 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
This affects the CMOS options iommu, ECC_memory, max_mem_clock,
hw_scrubber, interleave_chip_selects.
If they're absent in cmos.layout, a Kconfig value is used if it exists,
or a hardcoded default otherwise.
[Patrick: I changed the ramstage CMOS handling a bit, and dropped the
reliance of hw_scrubber on ECC RAM, as it has nothing to do with it -
it's the cache that's being scrubbed here.]
Signed-off-by: Josef Kellermann <seppk@arcor.de>
Acked-by: Patrick Georgi <patrick.georgi@secunet.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6380 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
This code provides cpu early initialization for Family 14h cpus. It is dependent on the AMD Agesa code.
Signed-off-by: Frank Vibrans <frank.vibrans@amd.com>
Acked-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6347 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
execution from flash memory. Coreboot uses WB. While there is no
noticeable performance difference between the two settings, use
of WB can cause a problem for a jtag debugger. The attached
patch changes AMD cache as ram setting for flash execution from
WB to WP.
Signed-off-by: Scott Duplichan <scott@notabs.org>
Acked-by: Marc Jones <marcj303@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6342 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Workaround for 131 removed.
Changed workaround for erratum 110 to only include pre-revision-F
processors.
For details, check AMD publications:
#25759 (Errata for Fam F pre-revision F processors)
#33610 (Errata for Fam F revision F and later processor)
Based on work and previous patches by:
Rudolf Marek <r.marek@assembler.cz>
Josef Kellermann <seppk@arcor.de>
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Acked-by: Patrick Georgi <patrick.georgi@secunet.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6340 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
a bit later.
Signed-off-by: Josef Kellermann <seppk@arcor.de>
Acked-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6339 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
have this go away again.
Signed-off-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Rudolf Marek <r.marek@assembler.cz>
Acked-by: Kevin O'Connor <kevin@koconnor.net>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6273 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
The main CIMx code is in a src/vendorcode directory and should not be
changed with regard to coding style etc. in order to remain easily syncable
with the "upstream" AMD code.
Signed-off-by: Kerry She <Kerry.she@amd.com>
Acked-by: Stefan Reinauer <stepan@coreboot.org>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6229 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Signed-off-by: Nils Jacobs <njacobs8@hetnet.nl>
Acked-by: Patrick Georgi <patrick@georgi-clan.de>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6210 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Signed-off-by: Nils Jacobs <njacobs8@hetnet.nl>
Acked-by: Patrick Georgi <patrick@georgi-clan.de>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6209 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Signed-off-by: Nils Jacobs <njacobs8@hetnet.nl>
Acked-by: Patrick Georgi <patrick@georgi-clan.de>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6208 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
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@6202 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
All K8/Fam10h boards use CAR, so move the "select CACHE_AS_RAM"
into the socket directories, and remove it from the individual boards.
Do the same for Intel CPUs/sockets where all boards use CAR.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Stefan Reinauer <stepan@coreboot.org>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6151 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
2) the patch implements get_cbmem_toc in chipset specific way if defined.
On Intel targets it should be unchanged. On K8T890 the the cbmem_toc is read from NVRAM. Why you ask? Because we cannot do it as on intel, because the framebuffer might be there making it hard to look for it in memory (and remember we need it so early that everying is uncached)
3) The patch removes hardcoded limits for suspend/resume save area (it was 1MB) on intel. Now it computes right numbers itself.
4) it impelements saving the memory during CAR to reserved range in sane way. First the sysinfo area (CAR data) is copied, then the rest after car is disabled (cached copy is used). I changed bit also the the copy of CAR area is now done uncached for target which I feel is more right.
I think I did not change the Intel suspend/resume behaviour but best would be if someone can test it. Please note this patch was unfinished on my drive since ages and it would be very nice to get it in to prevent bit rotten it again.
Now I feel it is done good way and should not break anything. I did a test with abuild and it seems fine.
Signed-off-by: Rudolf Marek <r.marek@assembler.cz>
Acked-by: Tobias Diedrich <ranma+coreboot@tdiedrich.de>
Acked-by: Stefan Reinauer <stepan@coreboot.org>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6117 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
deselected it, and very likely there won't ever be any
hardware that requires it deselected.
Keep the "selected" code path around, leading to no
functional change.
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Acked-by: Scott Duplichan <scott@notabs.org>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6086 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
make their defaults more obvious.
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Acked-by: Stefan Reinauer <stepan@coreboot.org>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6077 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
-- When building for UMA, reduce the limit for DRAM below 4GB
from E0000000 to C0000000. This is needed to accomodate the
UMA frame buffer.
-- Correct problem where msr C0010010 bits 21 and 22 (MtrrTom2En
and Tom2ForceMemTypeWB) are not set consistently across cores.
-- Enable TOM2 only if DRAM is present above 4GB.
-- Use AMD Tom2ForceMemTypeWB feature to avoid the need for
variable MTRR ranges above 4GB.
-- Add above4gb flag argument to function x86_setup_var_mtrrs. Clearing
this flag causes x86_setup_var_mtrrs() to omit MTRR ranges for
DRAM above 4GB. AMD systems use this option to conserve MTRRs.
-- Northbridge.c change to deduct UMA memory from DRAM size reported
by ram_resource. This corrects a problem where mtrr.c generates an
unexpected variable MTRR range.
-- Correct problem causing build failure when CONFIG_GFXUMA=1 and
CONFIG_VAR_MTRR_HOLE=0.
-- Reserve the UMA DRAM range for AMD K8 as is already done for AMD
family 10h.
Tested with mahogany on ECS A780G-GM with 2GB and 4GB.
Tested with mahogany_fam10 on ECS A780G-GM with 2GB and 4GB.
Signed-off-by: Scott Duplichan <scott@notabs.org>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6067 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Linux kernel, which was previously complaining that the MTRR setup
is wrong, if the cpu supports more than CONFIG_CPU_ADDR_BITS bits of
address space.
Shamelessly copied from Linux arch/x86/kernel/cpu/mtrr/main.c
Signed-off-by: Tobias Diedrich <ranma+coreboot@tdiedrich.de>
Acked-by: Scott Duplichan <scott@notabs.org>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6052 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
This is Abuild and boot tested.
Signed-off-by: Nils Jacobs <njacobs8@hetnet.nl>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@6015 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
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@5989 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
for 256 buses, even if fewer are configured. This patch lets msr
c0010058 programming use the configured bus count, CONFIG_MMCONF_BUS_NUMBER.
Signed-off-by: Scott Duplichan <scott@notabs.org>
Acked-by: Myles Watson <mylesgw@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@5976 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Also, update the board that uses this socket to match.
Signed-off-by: Jonathan Kollasch <jakllsch@kollasch.net>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@5969 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
is sometimes logged. The reason is that the AP first sets a completion value
such as 0x13, which is what function wait_cpu_state() is waiting for. Then a
short time later, the AP calls function init_fidvid_ap(). This function sets
a completion value of 01. When logging is off, wait_cpu_state is fast enough
to see the initial completion value for each of the APs. But with logging
enabled, one or more APs may go on to complete function init_fidvid_ap, which
sets the completion value to 01. While mostly harmless, the timeout does
increase boot time. This patch eliminates the timeout by making function
wait_cpu_state recognize 01 as an additional valid AP completion value.
Signed-off-by: Scott Duplichan <scott@notabs.org>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@5966 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
As both ioapic.h and acpi.h define a macro named "NMI", rename one
of them (NMI -> NMIType in acpi.h).
Abuild-tested.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@5943 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Macros for the register addresses for the MTRR MSRs are already defined
in include/cpu/x86/car.h. This patch uses those macros instead of
creating a second instance of that same data.
I also added a few macros to the amd mtrr.h to make the MSR naming more
consistent.
Signed-off-by: Warren Turkal <wt@penguintechs.org>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@5942 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
The following functions are moved to devices/device_util.c:
- ram_resource()
- tolm_test()
- find_pci_tolm()
There are only two tolm_test() / find_pci_tolm() which differ from the
defaults, one of them can easily be eliminated in a follow-up patch,
maybe even both, but for now keep it simple and only eliminate the majority.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@5937 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
This makes the CAR implementations a lot more readable, shorter and
easier to follow, and also reduces the amount of uselessly duplicated code.
For example there are more than 12 open-coded "enable cache" instances
spread all over the place (and 12 "disable cache" ones), multiple
"enable mtrr", "save BIST", "restore BIST", etc. etc.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Peter Stuge <peter@stuge.se>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@5902 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
