| Age | Commit message (Collapse) | Author |
|---|
|
This makes the intermediate rule visible so BL1 gets automatically
placed in the final image.
Change-Id: Iffb0268e5bbcbe135f2d39863ed64fa302409a22
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3141
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
|
|
The "wakeup" procedure will be shared by bootblock and romstage for different
types of resume processes.
Note, this commit does not include changes in romstage/bootblock to enable
suspend/resume feature. Simply adding functions to handle suspend/resume.
Verified by successfully building and booting Google/Snow firmware image.
Change-Id: I17a256afb99f2f8b5e0eac3393cdf6959b239341
Signed-off-by: Hung-Te Lin <hungte@chromium.org>
Reviewed-on: http://review.coreboot.org/3129
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
content.
To support suspend/resume, PHY control must be reset only on normal boot
path. So add a new param "mem_reset" to specify that.
Verified to boot successfully on Google/Snow.
Change-Id: Id49bc6c6239cf71a67ba091092dd3ebf18e83e33
Signed-off-by: Hung-Te Lin <hungte@chromium.org>
Reviewed-on: http://review.coreboot.org/3128
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This adds support for display bring-up on Snow. It
includes framebuffer initialization and LCD enable functions.
Change-Id: I16e711c97e9d02c916824f621e2313297448732b
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3116
Tested-by: build bot (Jenkins)
|
|
Add a safety check in function `intel_update_microcode` to return when
accidentally `NULL` is passed as `microcode_updates`, which would lead
to a null pointer dereference later on.
for (c = microcode_updates; m->hdrver; m = (const struct microcode *)c) {
While at it, use `return NULL` for clarity in function
`intel_microcode_find` and include the header file `stddef.h`. for it.
The review of this patch had some more discussion on adding more
comments and more detailed error messages. But this should be done in
a separate patch.
For clarity here some history, on how this was found and what caused
the discussion and confusion.
Originally when Vladimir made this improvement, selecting
`CPU_MICROCODE_IN_CBFS` in Kconfig but not having the microcode blob
`cpu_microcode_blob.bin` in CBFS resulted in a null pointer dereference
later on causing a crash.
for (c = microcode_updates; m->hdrver; m = (const struct microcode *)c) {
Vladimir fixed this by returning if `microcode_updates` is `NULL`,
that means no file is found and successfully tested this on his
Lenovo X201.
When pushing the patch to Gerrit for review, the code was rewritten
though by Aaron in commit »intel microcode: split up microcode loading
stages« (98ffb426) [1], which also returns when no file is found. So
the other parts of the code were checked and the safety check as
described above is added.
[1] http://review.coreboot.org/2778
Change-Id: I6e18fd37256910bf047061e4633a66cf29ad7b69
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: http://review.coreboot.org/2990
Reviewed-by: Aaron Durbin <adurbin@google.com>
Tested-by: build bot (Jenkins)
|
|
This makes sure that the product ID (PRO_ID) register can be read
when the OS kernel is figuring out what kind of CPU it's running on.
For historical reference, the original U-Boot code seems to have
worked basically by accident here. The hardware has a quirk where by
reading the value before gating the IP block keeps the value
persistent. U-Boot reads the chip ID early on to distinguish between
chip family, but we do not mix code the same way so we do not read
the chip ID. Since the value has been read before the clock gating
happens, the value remains available for the kernel to use during the
decompression stage. We don't want to rely on that behavior when using
coreboot. Instead the kernel should gate unused IPs.
(credit to Gabe for finding symptom in the kernel)
Change-Id: Iaa21e6e718b9000b5558f568020f393779fd208e
Signed-off-by: Gabe Black <gabeblack@chromium.org>
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3121
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
|
|
This simple error led to corrupted graphics.
How annoying.
Change-Id: I2295c0df0f1d16014a603dc5d66bd4d72f3fb7c9
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/3120
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Tested-by: build bot (Jenkins)
|
|
The original imported code used "lcdbase" and "lcd_base" which quite
predictably caused confusion and bugs. Let's put an end to this little
bit of insanity.
Change-Id: I4f995482cfbff5f23bb296a1e6d35beccf5f8a91
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3114
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
|
|
descriptor
We neglected to copy xres and yres out; now we do.
Change-Id: Icc4a8eb35799d156b11274f71bcfb4a1d10e01e3
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/3111
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Tested-by: build bot (Jenkins)
|
|
This updates the Exynos TMU code for coreboot:
- Remove dependency on device tree
- Add Makefile entries
Change-Id: I55e1b624d7c7b695b1253ec55f6ae3de8dc671bc
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3107
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
|
|
This simply imports the Exynos TMU driver from u-boot. It is not
built and thus should not break anything.
Change-Id: I7861132fbf97f864e4250ffbda1ef3843f296ddc
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3106
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
|
|
This moves the prototype for power_enable_hw_thermal_trip() to
a generic location so it can be used by generalized thermal
management code. The implementation will still be CPU-specific.
Change-Id: Iae449cb8c72c8441dedaf65b73db9898b4730cef
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3105
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
|
|
This gets rid of the clock-tick based sdelay in favor of udelay().
udelay() is more consistent and easier to work with, and this allows
us to carry one less variation of timers (and headers and sources...).
Every 1 unit in the sdelay() argument was assumed to cause a delay of
2 clock ticks (@1.7GHz). So the conversion factor is roughly:
sdelay(N) = udelay(((N * 2) / 1.7 * 10^9) * 10^6)
= udelay((N * 2) / (1.7 * 10^3))
The sdelay() periods used were:
sdelay(100) --> udelay(1)
sdelay(0x10000) --> udelay(78) (rounded up to udelay(100))
There was one instance of sdelay(10000), which looked like sort of a
typo since sdelay(0x10000) was used elsewhere. sdelay(10000) should
approximate to about 12us, so we'll stick with that for now and leave
a note.
Change-Id: I5e7407865ceafa701eea1d613bbe50cf4734f33e
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3079
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The types are (esp. int) are confusing at times as to size.
Make them definite as to size.
Change-Id: Id7808f1f61649ec0a3403c1afc3c2c3d4302b7fb
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/3103
Tested-by: build bot (Jenkins)
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
|
|
This removes the wait_ms argument from the dp_controller_init(). The
only delay involved is a constant 60ms delay that happens if
everything else goes well. This delay is derived from the LCD spec
so there's no reason it should be baked into the controller code.
(This patch also has the side-effect of fixing a bug where we were
delaying on an undefined value for wait_ms).
Change-Id: I03aa19f2ac2f720524fcb7c795e10cc57f0a226e
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3078
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Add a microsecond timer, its declaration, the function to start it,
and its usage. To start it, one calls timer_start(). From that point
on, one can call timer_us() to find microseconds since the timer was
started.
We show its use in the bootblock. You want it started very early.
Finally, the delay.h change having been (ironically) delayed, we
create time.h and have it hold one declaration, for the timer_us() and
timer_start() prototype.
We feel that these two functions should become the hardware specific
functions, allowing us to finally move udelay() into src/lib where it
belongs.
Change-Id: I19cbc2bb0089a3de88cfb94276266af38b9363c5
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/3073
Tested-by: build bot (Jenkins)
|
|
Split the Persimmon DSDT into common code areas.
For example, split the Southbridge specific code into
the Southbridge directory and CPU specific code into
the CPU directory. Also adding the superio.asl file
to the Persimmon DSDT tree. This file is empty for
the moment but will be necessary in the future. I have
also emptied the thermal.asl file in the mainboard
directory because it does not seem to perform as
intended (fan control does not change when it is
brought back into the code base) and it has been
inside a '#if 0' statement for a long time. Removing
it until it is decided that it is actually necessary.
This change was verified in three different ways:
1. Visual comparison of the compiled DSDT pulled from the
Persimmon after booting into Linux using the ACPI tools
acpidump, acpixtract, and iasl. The comparison was done
between the DSDT before and after doing the split work.
This test is somewhat difficult considering the expanse
of the changes. Blocks of code have been moved, and
others changed.
2. Linux logs were dumped before and after the DSDT split.
Logs dumped and compared include dmesg and lspci -tv.
Neither log changed significantly between the two compare
points.
3. The test suite FWTS was run on the Coreboot build both
before and after doing the DSDT split with the command
'sudo fwts -b -P -u'. The flag -b specifies all batch jobs,
-P specifies all power tests, and -u specifies utilities.
Interactive jobs were not run as most of them consist of
laptop checks. Again, there were no significant changes
between the two endpoints.
These tests lead me to believe that there was no change in
the functionality of the ACPI tables apart from what is
known and expected.
This patch is the first of a series of patches to split the DSDT.
The ASRock patch was merged before this one and breaks the ASROCK
E350M1 build (patch 8d80a3fb: http://review.coreboot.org/#/c/3050/).
Please be aware of this dependency when pulling these patches.
Other patches that depend on this patch are
'AMD Fam14: Split out the AMD Fam14 DSDT'
(http://review.coreboot.org/#/c/3051/)
and 'Fam14 DSDT: Also return for unrecognized UUID in _OSC'
(http://review.coreboot.org/#/c/3052/)
Change-Id: I53ff59909cceb30a08e8eab3d59b30b97c802726
Signed-off-by: Mike Loptien <mike.loptien@se-eng.com>
Reviewed-on: http://review.coreboot.org/3048
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martin.roth@se-eng.com>
|
|
Properly use the chip settings when configuring the CPU,
at this point being purely graphics.
Change-Id: I9bc2d32c1037653837937b314e4041abc0024835
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/3054
Tested-by: build bot (Jenkins)
Reviewed-by: David Hendricks <dhendrix@chromium.org>
|
|
Add basic edp support to the ramstage. Not working.
Change-Id: I15086e03417edca7426c214e67b51719d8ed9341
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/3055
Tested-by: build bot (Jenkins)
|
|
Basic cleanup, this code still does not work.
Change-Id: I84ed9f08fd04cd8eb74cd860e0775d8c602f42d6
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/3049
Tested-by: build bot (Jenkins)
|
|
This enables type checking for safety as to help prevent errors like
http://review.coreboot.org/#/c/3038/ . Now compilation fails if the
wrong type is passed into readb/readw/readl/writeb/writew/writel
or other macros in io.h.
This also deprecates readw/writew. The previous definition was 16-bits
which is incorrect since wordsize on ARMv7 is 32-bits and there was
only 1 instance of writew (#if 0'd anyway). Going forward we should
always use read{8,16,32} and write{8,16,32} where N specifies the
exact length rather than relying on ambiguous definition of wordsize.
Since many macros relied on __raw_*, which were basically the same
(minus data memory barrier instructions), this patch also gets rid
of __raw_*. There were parts of the code which ended up using these
macros consecutively, for example:
setbits_le32(®s->ch_cfg, SPI_CH_RST);
clrbits_le32(®s->ch_cfg, SPI_CH_RST);
In such cases the safe versions of readl() and writel() should be
used anyway.
Note: This also fixes two dubious casts as to avoid breaking
compilation.
Change-Id: I8850933f68ea3a9b615d00ebd422f7c242268f1c
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3045
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This re-factors the Exynos5 I2C code to be simpler and use the
new API, and updates users accordingly.
- i2c_read() and i2c_write() functions updated to take bus number
as an argument.
- Get rid of the EEPROM_ADDR_OVERFLOW stuff in i2c_read() and
i2c_write(). If a chip needs special handling we should take care
of it elsewhere, not in every low-level i2c driver.
- All the confusing bus config functions eliminated. No more
i2c_set_early_config() or i2c_set_bus() or i2c_get_bus(). All this
is handled automatically when the caller does a transaction and
specifies the desired bus number.
- i2c_probe() eliminated. We're not a command-line utility.
- Let the compiler place static variables automatically. We don't need
any of this fancy manual data placement.
- Remove dead code while we're at it. This stuff was ported early on
and much of it was left commented out in case we needed it. Some
also includes nested macros which caused gcc to complain.
- Clean up #includes (no more common.h, woohoo!), replace debug() with
printk().
Change-Id: I8e1f974ea4c6c7db9f33b77bbc4fb16008ed0d2a
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3044
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The existing header was imported along with the Exynos code and left
mostly unchanged. This is the first patch in a series intended to
replace the imported u-boot I2C API with a much simpler and cleaner
interface:
- We only need to expose i2c_read() and i2c_write() in our public API.
Everything else is board/chip-dependent and should remain hidden
away.
- i2c_read and i2c_write functions will take bus number as an arg
and we'll eliminate i2c_get_bus and i2c_set_bus. Those are prone to
error and end up cluttering the code since the user needs to save
the old bus number, set the new one, do the read/write, and restore
the old value (3 added steps to do a simple transaction).
- Stop setting default values for board-specific things like SPD
and RTC bus numbers (as if we always have an SPD or RTC on I2C).
- Death to all the trivial inline wrappers. And in case there was any
doubt, we really don't care about the MPC8xx. Though if we did then
we would not pollute the public API with its idiosyncrasies.
Change-Id: I4410a3c82ed5a6b2e80e3d8c0163464a9ca7c3b0
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3043
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This adds a missing address-of operator. This was a subtle bug that
didn't seem to cause problems at first since the serial console
appeared to work. However it caused an imprecise external abort which
became apparent later on when aborts were unmasked in the kernel via
the CPSR_A bit.
(credit goes to Gabe Black for finding this)
Signed-off-by: Gabe Black <gabeblack@chromium.org>
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Change-Id: I80a33b147d92d559fa8fefbe7d5642235deb9aea
Reviewed-on: http://review.coreboot.org/3038
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This moves highly board-specific code out from the Exynos5250
power_init() into Snow's romstage.c. There's no reason the CPU-
specific code should care about which PMIC we are using and
which bus it is on.
Change-Id: I52313177395519cddcab11225fc23d5e50c4c4e3
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/3034
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Display hardware is part of this SOC, and we need to be able
to set certain variables in devicetree.cb. This chip file
contains the initial things we think we need to set.
Change-Id: I16f2d4228c87116dbeb53a3c9f3f359a6444f552
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/3031
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Tested-by: build bot (Jenkins)
|
|
This was a first pass at display port support, we have
realized that it was ultimately a bad path. The display
hardware is intimately tied into a specific cpu and
mainboard combination, and the code has to be elsewhere.
The devicetree formatting is ugly, but it matters not:
it's changing soon.
Change-Id: Iddce54f9e7219a7569315565fac65afbbe0edd29
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/3029
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
Based on comments in cpu/x86/msr.h for wrmsr/rdmsr, and for symmetry,
I have added __attribute__((always_inline)) for these.
Change-Id: Ia0a34c15241f9fbc8c78763386028ddcbe6690b1
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/2898
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
If ROM caching is selected the haswell CPU initialization code
will enable ROM caching after all other CPU threads are brought
up.
Change-Id: I75424bb75174bfeca001468c3272e6375e925122
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/3016
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
The MP code on haswell was mirroring the BSPs MTRRs. In addition it
was cleaning up the ROM cache so that the MTRR register values were
the same once the OS was booted. Since the hyperthread sibling of
the BSP was going through this path the ROM cache was getting torn
down once the hyperthread was brought up.
That said, there was no differnce in observed boot time keeping the
ROM cache enabled.
Change-Id: I2a59988fcfeea9291202c961636ea761c2538837
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/3008
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
The haswell code was using the old assumption of which MTRR
was used for the ROM cache. Now that there is an API for doing
this use it as the old assumption is no longer valid.
Change-Id: I59ef897becfc9834d36d28840da6dc4f1145b0c7
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/3007
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
On certain architectures such as x86 the bootstrap processor
does most of the work. When CACHE_ROM is employed it's appropriate
to ensure that the caching enablement of the ROM is disabled so that
the caching settings are symmetric before booting the payload or OS.
Tested this on an x86 machine that turned on ROM caching. Linux did not
complain about asymmetric MTRR settings nor did the ROM show up as
cached in the MTRR settings.
Change-Id: Ia32ff9fdb1608667a0e9a5f23b9c8af27d589047
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2980
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
I did not check what was once after the 'and'.
Change-Id: I9f3f725bec281a94abdb2eeb692a96fecdebcc0c
Signed-off-by: Stefan Tauner <stefan.tauner@gmx.at>
Reviewed-on: http://review.coreboot.org/2999
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
Previously southbridge_smm_init() was provided that did both
the clearing of the SMM state and enabling SMIs. This is
troublesome in how haswell machines bring up the APs. The BSP
enters SMM once to determine if parallel SMM relocation is possible.
If it is possible the BSP releases the APs to do SMM relocation.
Normally, after the APs complete the SMM relocation, the BSP would then
re-enter the relocation handler to relocate its own SMM space.
However, because SMIs were previously enabled it is possible for an SMI
event to occur before the APs are complete or have entered the
relocation handler. This is bad because the BSP will turn off parallel
SMM save state. Additionally, this is a problem because the relocation
handler is not written to handle regular SMIs which can cause an
SMI storm which effectively looks like a hung machine. Correct these
issues by turning on SMIs after all the SMM relocation has occurred.
Change-Id: Id4f07553b110b9664d51d2e670a14e6617591500
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2977
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This adds a new function to configure L2 cache for the
exynos5250 and deprecates the old function.
Change-Id: I9562f3301aa1e2911dae3856ab57bb6beec2e224
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2949
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-by: Gabe Black <gabe.black@gmail.com>
Tested-by: build bot (Jenkins)
|
|
There is an optimization that can take place when hole
carving in ranges above 4GiB. If the range is the last
range then there is no need to carve UC holes out from
the larger WB range.
This optimization also has the same assumption of choosing
WB as the default MTRR type: the OS needs to properly
handle accessing realloacted MMIO resources with PAT so
that the MTRR type can be overidden.
Below are results using a combination of options. The
board this was tested on has 10 variable MTRRs at its
disposal. It has 4GiB of RAM.
IO hole config #1: hole starts at 0xad800000
No CACHE_ROM and no WRCOMB resources (takes 4 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x00000000ad800000 size 0xad740000 type 6
0x00000000ad800000 - 0x0000000100000000 size 0x52800000 type 0
0x0000000100000000 - 0x000000014f600000 size 0x4f600000 type 6
MTRR: default type WB/UC MTRR counts: 4/6.
MTRR: WB selected as default type.
MTRR: 0 base 0x00000000ad800000 mask 0x0000007fff800000 type 0
MTRR: 1 base 0x00000000ae000000 mask 0x0000007ffe000000 type 0
MTRR: 2 base 0x00000000b0000000 mask 0x0000007ff0000000 type 0
MTRR: 3 base 0x00000000c0000000 mask 0x0000007fc0000000 type 0
No CACHE_ROM and 1 WRCOMB resource (takes 6 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x00000000ad800000 size 0xad740000 type 6
0x00000000ad800000 - 0x00000000d0000000 size 0x22800000 type 0
0x00000000d0000000 - 0x00000000e0000000 size 0x10000000 type 1
0x00000000e0000000 - 0x0000000100000000 size 0x20000000 type 0
0x0000000100000000 - 0x000000014f600000 size 0x4f600000 type 6
MTRR: default type WB/UC MTRR counts: 6/7.
MTRR: WB selected as default type.
MTRR: 0 base 0x00000000ad800000 mask 0x0000007fff800000 type 0
MTRR: 1 base 0x00000000ae000000 mask 0x0000007ffe000000 type 0
MTRR: 2 base 0x00000000b0000000 mask 0x0000007ff0000000 type 0
MTRR: 3 base 0x00000000c0000000 mask 0x0000007ff0000000 type 0
MTRR: 4 base 0x00000000d0000000 mask 0x0000007ff0000000 type 1
MTRR: 5 base 0x00000000e0000000 mask 0x0000007fe0000000 type 0
CACHE_ROM and no WRCOMB resources (takes 7 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x00000000ad800000 size 0xad740000 type 6
0x00000000ad800000 - 0x00000000ff800000 size 0x52000000 type 0
0x00000000ff800000 - 0x0000000100000000 size 0x00800000 type 5
0x0000000100000000 - 0x000000014f600000 size 0x4f600000 type 6
MTRR: default type WB/UC MTRR counts: 11/7.
MTRR: UC selected as default type.
MTRR: 0 base 0x0000000000000000 mask 0x0000007f80000000 type 6
MTRR: 1 base 0x0000000080000000 mask 0x0000007fe0000000 type 6
MTRR: 2 base 0x00000000a0000000 mask 0x0000007ff0000000 type 6
MTRR: 3 base 0x00000000ad800000 mask 0x0000007fff800000 type 0
MTRR: 4 base 0x00000000ae000000 mask 0x0000007ffe000000 type 0
MTRR: 5 base 0x00000000ff800000 mask 0x0000007fff800000 type 0
MTRR: 6 base 0x0000000100000000 mask 0x0000007f00000000 type 6
CACHE_ROM and 1 WRCOMB resource (takes 8 MTRRs):
Previously this combination was impossible without the optimization.
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x00000000ad800000 size 0xad740000 type 6
0x00000000ad800000 - 0x00000000d0000000 size 0x22800000 type 0
0x00000000d0000000 - 0x00000000e0000000 size 0x10000000 type 1
0x00000000e0000000 - 0x00000000ff800000 size 0x1f800000 type 0
0x00000000ff800000 - 0x0000000100000000 size 0x00800000 type 5
0x0000000100000000 - 0x000000014f600000 size 0x4f600000 type 6
MTRR: default type WB/UC MTRR counts: 12/8.
MTRR: UC selected as default type.
MTRR: 0 base 0x0000000000000000 mask 0x0000007f80000000 type 6
MTRR: 1 base 0x0000000080000000 mask 0x0000007fe0000000 type 6
MTRR: 2 base 0x00000000a0000000 mask 0x0000007ff0000000 type 6
MTRR: 3 base 0x00000000ad800000 mask 0x0000007fff800000 type 0
MTRR: 4 base 0x00000000ae000000 mask 0x0000007ffe000000 type 0
MTRR: 5 base 0x00000000d0000000 mask 0x0000007ff0000000 type 1
MTRR: 6 base 0x00000000ff800000 mask 0x0000007fff800000 type 0
MTRR: 7 base 0x0000000100000000 mask 0x0000007f00000000 type 6
IO hole config #1: hole starts at 0x80000000
No CACHE_ROM and no WRCOMB resources (takes 1 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x0000000080000000 size 0x7ff40000 type 6
0x0000000080000000 - 0x0000000100000000 size 0x80000000 type 0
0x0000000100000000 - 0x000000017ce00000 size 0x7ce00000 type 6
MTRR: default type WB/UC MTRR counts: 1/2.
MTRR: WB selected as default type.
MTRR: 0 base 0x0000000080000000 mask 0x0000007f80000000 type 0
No CACHE_ROM and 1 WRCOMB resource (takes 3 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x0000000080000000 size 0x7ff40000 type 6
0x0000000080000000 - 0x00000000d0000000 size 0x50000000 type 0
0x00000000d0000000 - 0x00000000e0000000 size 0x10000000 type 1
0x00000000e0000000 - 0x0000000100000000 size 0x20000000 type 0
0x0000000100000000 - 0x000000017ce00000 size 0x7ce00000 type 6
MTRR: default type WB/UC MTRR counts: 4/3.
MTRR: UC selected as default type.
MTRR: 0 base 0x0000000000000000 mask 0x0000007f80000000 type 6
MTRR: 1 base 0x00000000d0000000 mask 0x0000007ff0000000 type 1
MTRR: 2 base 0x0000000100000000 mask 0x0000007f00000000 type 6
CACHE_ROM and no WRCOMB resources (takes 3 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x0000000080000000 size 0x7ff40000 type 6
0x0000000080000000 - 0x00000000ff800000 size 0x7f800000 type 0
0x00000000ff800000 - 0x0000000100000000 size 0x00800000 type 5
0x0000000100000000 - 0x000000017ce00000 size 0x7ce00000 type 6
MTRR: default type WB/UC MTRR counts: 9/3.
MTRR: UC selected as default type.
MTRR: 0 base 0x0000000000000000 mask 0x0000007f80000000 type 6
MTRR: 1 base 0x00000000ff800000 mask 0x0000007fff800000 type 0
MTRR: 2 base 0x0000000100000000 mask 0x0000007f00000000 type 6
CACHE_ROM and 1 WRCOMB resource (takes 4 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x0000000080000000 size 0x7ff40000 type 6
0x0000000080000000 - 0x00000000d0000000 size 0x50000000 type 0
0x00000000d0000000 - 0x00000000e0000000 size 0x10000000 type 1
0x00000000e0000000 - 0x00000000ff800000 size 0x1f800000 type 0
0x00000000ff800000 - 0x0000000100000000 size 0x00800000 type 5
0x0000000100000000 - 0x000000017ce00000 size 0x7ce00000 type 6
MTRR: default type WB/UC MTRR counts: 10/4.
MTRR: UC selected as default type.
MTRR: 0 base 0x0000000000000000 mask 0x0000007f80000000 type 6
MTRR: 1 base 0x00000000d0000000 mask 0x0000007ff0000000 type 1
MTRR: 2 base 0x00000000ff800000 mask 0x0000007fff800000 type 0
MTRR: 3 base 0x0000000100000000 mask 0x0000007f00000000 type 6
Change-Id: Ia3195af686c3f0603b21f713cfb2d9075eb02806
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2959
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
Some ranges would use less variable MTRRs if an UC area
can be carved off the top of larger WB range. Implement this
approach by doing 3 passes over each region in the addres space:
1. UC default type. Cover non-UC and non-WB regions with respectie type.
Punch UC hole at upper end of larger WB regions with WB type.
2. UC default type. Cover non-UC regions with respective type.
3. WB default type. Cover non-WB regions with respective type.
The hole at upper end of a region uses the same min alignment of 64MiB.
Below are results using a combination of options. The board this was
tested on has 10 variable MTRRs at its disposal. It has 4GiB of RAM.
IO hole config #1: hole starts at 0xad800000
No CACHE_ROM or WRCOMB resources (takes 4 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x00000000ad800000 size 0xad740000 type 6
0x00000000ad800000 - 0x0000000100000000 size 0x52800000 type 0
0x0000000100000000 - 0x000000014f600000 size 0x4f600000 type 6
MTRR: default type WB/UC MTRR counts: 4/9.
MTRR: WB selected as default type.
MTRR: 0 base 0x00000000ad800000 mask 0x0000007fff800000 type 0
MTRR: 1 base 0x00000000ae000000 mask 0x0000007ffe000000 type 0
MTRR: 2 base 0x00000000b0000000 mask 0x0000007ff0000000 type 0
MTRR: 3 base 0x00000000c0000000 mask 0x0000007fc0000000 type 0
No CACHE_ROM. 1 WRCOMB resource (takes 6 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x00000000ad800000 size 0xad740000 type 6
0x00000000ad800000 - 0x00000000d0000000 size 0x22800000 type 0
0x00000000d0000000 - 0x00000000e0000000 size 0x10000000 type 1
0x00000000e0000000 - 0x0000000100000000 size 0x20000000 type 0
0x0000000100000000 - 0x000000014f600000 size 0x4f600000 type 6
MTRR: default type WB/UC MTRR counts: 6/10.
MTRR: WB selected as default type.
MTRR: 0 base 0x00000000ad800000 mask 0x0000007fff800000 type 0
MTRR: 1 base 0x00000000ae000000 mask 0x0000007ffe000000 type 0
MTRR: 2 base 0x00000000b0000000 mask 0x0000007ff0000000 type 0
MTRR: 3 base 0x00000000c0000000 mask 0x0000007ff0000000 type 0
MTRR: 4 base 0x00000000d0000000 mask 0x0000007ff0000000 type 1
MTRR: 5 base 0x00000000e0000000 mask 0x0000007fe0000000 type 0
CACHE_ROM and no WRCOMB resources (taks 10 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x00000000ad800000 size 0xad740000 type 6
0x00000000ad800000 - 0x00000000ff800000 size 0x52000000 type 0
0x00000000ff800000 - 0x0000000100000000 size 0x00800000 type 5
0x0000000100000000 - 0x000000014f600000 size 0x4f600000 type 6
MTRR: default type WB/UC MTRR counts: 11/10.
MTRR: UC selected as default type.
MTRR: 0 base 0x0000000000000000 mask 0x0000007f80000000 type 6
MTRR: 1 base 0x0000000080000000 mask 0x0000007fe0000000 type 6
MTRR: 2 base 0x00000000a0000000 mask 0x0000007ff0000000 type 6
MTRR: 3 base 0x00000000ad800000 mask 0x0000007fff800000 type 0
MTRR: 4 base 0x00000000ae000000 mask 0x0000007ffe000000 type 0
MTRR: 5 base 0x00000000ff800000 mask 0x0000007fff800000 type 0
MTRR: 6 base 0x0000000100000000 mask 0x0000007fc0000000 type 6
MTRR: 7 base 0x0000000140000000 mask 0x0000007ff0000000 type 6
Taking a reserved OS MTRR.
MTRR: 8 base 0x000000014f600000 mask 0x0000007fffe00000 type 0
Taking a reserved OS MTRR.
MTRR: 9 base 0x000000014f800000 mask 0x0000007fff800000 type 0
A combination of CACHE_ROM and WRCOMB just won't work.
IO hole config #2: hole starts at 0x80000000:
No CACHE_ROM or WRCOMB resources (takes 1 MTRR):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x0000000080000000 size 0x7ff40000 type 6
0x0000000080000000 - 0x0000000100000000 size 0x80000000 type 0
0x0000000100000000 - 0x000000017ce00000 size 0x7ce00000 type 6
MTRR: default type WB/UC MTRR counts: 1/5.
MTRR: WB selected as default type.
MTRR: 0 base 0x0000000080000000 mask 0x0000007f80000000 type 0
No CACHE_ROM. 1 WRCOMB resource (takes 4 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x0000000080000000 size 0x7ff40000 type 6
0x0000000080000000 - 0x00000000d0000000 size 0x50000000 type 0
0x00000000d0000000 - 0x00000000e0000000 size 0x10000000 type 1
0x00000000e0000000 - 0x0000000100000000 size 0x20000000 type 0
0x0000000100000000 - 0x000000017ce00000 size 0x7ce00000 type 6
MTRR: default type WB/UC MTRR counts: 4/6.
MTRR: WB selected as default type.
MTRR: 0 base 0x0000000080000000 mask 0x0000007fc0000000 type 0
MTRR: 1 base 0x00000000c0000000 mask 0x0000007ff0000000 type 0
MTRR: 2 base 0x00000000d0000000 mask 0x0000007ff0000000 type 1
MTRR: 3 base 0x00000000e0000000 mask 0x0000007fe0000000 type 0
CACHE_ROM and no WRCOMB resources (takes 6 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x0000000080000000 size 0x7ff40000 type 6
0x0000000080000000 - 0x00000000ff800000 size 0x7f800000 type 0
0x00000000ff800000 - 0x0000000100000000 size 0x00800000 type 5
0x0000000100000000 - 0x000000017ce00000 size 0x7ce00000 type 6
MTRR: default type WB/UC MTRR counts: 9/6.
MTRR: UC selected as default type.
MTRR: 0 base 0x0000000000000000 mask 0x0000007f80000000 type 6
MTRR: 1 base 0x00000000ff800000 mask 0x0000007fff800000 type 0
MTRR: 2 base 0x0000000100000000 mask 0x0000007f80000000 type 6
MTRR: 3 base 0x000000017ce00000 mask 0x0000007fffe00000 type 0
MTRR: 4 base 0x000000017d000000 mask 0x0000007fff000000 type 0
MTRR: 5 base 0x000000017e000000 mask 0x0000007ffe000000 type 0
CACHE_ROM and 1 WRCOMB resource (takes 7 MTRRs):
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x0000000080000000 size 0x7ff40000 type 6
0x0000000080000000 - 0x00000000d0000000 size 0x50000000 type 0
0x00000000d0000000 - 0x00000000e0000000 size 0x10000000 type 1
0x00000000e0000000 - 0x00000000ff800000 size 0x1f800000 type 0
0x00000000ff800000 - 0x0000000100000000 size 0x00800000 type 5
0x0000000100000000 - 0x000000017ce00000 size 0x7ce00000 type 6
MTRR: default type WB/UC MTRR counts: 10/7.
MTRR: UC selected as default type.
MTRR: 0 base 0x0000000000000000 mask 0x0000007f80000000 type 6
MTRR: 1 base 0x00000000d0000000 mask 0x0000007ff0000000 type 1
MTRR: 2 base 0x00000000ff800000 mask 0x0000007fff800000 type 0
MTRR: 3 base 0x0000000100000000 mask 0x0000007f80000000 type 6
MTRR: 4 base 0x000000017ce00000 mask 0x0000007fffe00000 type 0
MTRR: 5 base 0x000000017d000000 mask 0x0000007fff000000 type 0
MTRR: 6 base 0x000000017e000000 mask 0x0000007ffe000000 type 0
Change-Id: Iceb9b64991accf558caae2e7b0205951e9bcde44
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2925
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
Downstream payloads may need to take advantage of caching the
ROM for performance reasons. Add the ability to communicate the
variable range MTRR index to use to perform the caching enablement.
An example usage implementation would be to obtain the variable MTRR
index that covers the ROM from the coreboot tables. Then one would
disable caching and change the MTRR type from uncacheable to
write-protect and enable caching. The opposite sequence is required
to tearn down the caching.
Change-Id: I4d486cfb986629247ab2da7818486973c6720ef5
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2919
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
The CONFIG_CACHE_ROM support in the MTRR code allocates an MTRR
specifically for setting up write-protect cachine of the ROM. It is
assumed that CONFIG_ROM_SIZE is the size of the ROM and the whole
area should be cached just under 4GiB. If enabled, the MTRR code
will allocate but not enable rom caching. It is up to the callers
of the MTRR code to explicitly enable (and disable afterwards) through
the use of 2 new functions:
- x86_mtrr_enable_rom_caching()
- x86_mtrr_disable_rom_caching()
Additionally, the CACHE_ROM option is exposed to the config menu so
that it is not just selected by the chipset or board. The reasoning
is that through a multitude of options CACHE_ROM may not be appropriate
for enabling.
Change-Id: I4483df850f442bdcef969ffeaf7608ed70b88085
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2918
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
All resources that set the IORESOURCE_WRCOMB attribute which are
also marked as IORESOURCE_PREFETCH will have a MTRR set up that
is of the write-combining cacheable type. The only resources on
x86 that can be set to write-combining are prefetchable ones.
Change-Id: Iba7452cff3677e07d7e263b79982a49c93be9c54
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2892
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The old MTRR code had issues using too many variable
MTRRs depending on the physical address space layout dictated
by the device resources. This new implementation calculates
the default MTRR type by comparing the number of variable MTRRs
used for each type. This avoids the need for IORESOURE_UMA_FB
because in many of those situations setting the default type to WB
frees up the variable MTTRs to set that space to UC.
Additionally, it removes the need for IORESOURCE_IGNORE_MTRR
becuase the new mtrr uses the memrange library which does merging
of resources.
Lastly, the sandybridge gma has its speedup optimization removed
for the graphics memory by writing a pre-determined MTRR index.
That will be fixed in an upcoming patch once write-combining support
is added to the resources.
Slight differences from previous MTRR code:
- The number of reserved OS MTRRs is not a hard limit. It's now advisory
as PAT can be used by the OS to setup the regions to the caching
policy desired.
- The memory types are calculated once by the first CPU to run the code.
After that all other CPUs use that value.
- CONFIG_CACHE_ROM support was dropped. It will be added back in its own
change.
A pathological case that was previously fixed by changing vendor code
to adjust the IO hole location looked like the following:
MTRR: Physical address space:
0x0000000000000000 - 0x00000000000a0000 size 0x000a0000 type 6
0x00000000000a0000 - 0x00000000000c0000 size 0x00020000 type 0
0x00000000000c0000 - 0x00000000ad800000 size 0xad740000 type 6
0x00000000ad800000 - 0x00000000d0000000 size 0x22800000 type 0
0x00000000d0000000 - 0x00000000e0000000 size 0x10000000 type 1
0x00000000e0000000 - 0x0000000100000000 size 0x20000000 type 0
0x0000000100000000 - 0x000000014f600000 size 0x4f600000 type 6
As noted by the output below it's impossible to accomodate those
ranges even with 10 variable MTRRS. However, because the code
can select WB as the default MTRR type it can be done in 6 MTRRs:
MTRR: default type WB/UC MTRR counts: 6/14.
MTRR: WB selected as default type.
MTRR: 0 base 0x00000000ad800000 mask 0x0000007fff800000 type 0
MTRR: 1 base 0x00000000ae000000 mask 0x0000007ffe000000 type 0
MTRR: 2 base 0x00000000b0000000 mask 0x0000007ff0000000 type 0
MTRR: 3 base 0x00000000c0000000 mask 0x0000007ff0000000 type 0
MTRR: 4 base 0x00000000d0000000 mask 0x0000007ff0000000 type 1
MTRR: 5 base 0x00000000e0000000 mask 0x0000007fe0000000 type 0
Change-Id: Idfcc78d9afef9d44c769a676716aae3ff2bd79de
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2889
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
This moves the ram resource allocation into cpu_init() so that we
no longer rely on declaring a domain in devicetree.cb (which is kind
of weird for this platform). This does not cause any actual changes
to the coreboot memory table, and paves the way for further updates
to Snow's devicetree.
Change-Id: I141277f59b5d48288f409257bf556a1cfa7a8463
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2923
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This reverts commit 9427ca151e44644238b1b52138894195a9f5175f
Looks like we were a bit too anxious to see this one get in. The devicetree.cb change seems to have broken things.
coreboot memory table:
0. 0000000050000000-000000005000ffff: RESERVED
1. 00000000bff00000-00000000bfffffff: CONFIGURATION TABLES
2. 0000014004000000-00000140044007ff: RESERVED
Before this patch:
coreboot memory table:
0. 0000000040000000-00000000bfefffff: RAM
1. 00000000bff00000-00000000bfffffff: CONFIGURATION TABLES
Change-Id: I618e4f1976265d56cfd6a61d0c5736c55a0f3cec
Reviewed-on: http://review.coreboot.org/2914
Tested-by: build bot (Jenkins)
Reviewed-by: David Hendricks <dhendrix@chromium.org>
|
|
This does NOT turn on the graphics.
The device tree has been changed enough so that, at the very least, the correct
functions are called at the correct time, with the correct paramaters. We
decided to yank the I2C entries as they did not obvious function and might
not even have been correct.
Not working, seemingly, but we need to add a 4M resource for
memory, and it seems it needs to be fixed at the address shown.
This address was chosen from current hardware.
We realized that the display code should be part of the cpu -- that's how
the hardware works!
Change-Id: Ied65a554f833566be817540702f79a02e7b6cb6e
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/2615
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
|
|
The amd_mtrr.c file contains a copy of the fixed MTRR algorithm.
However, the AMD code needs to handle the RdMem and WrMem attribute
bits in the fixed MTRR MSRs. Instead of duplicating the code
with the one slight change introduce a Kconfig option,
X86_AMD_FIXED_MTRRS, which indicates that the RdMem and WrMem fields
need to be handled for writeback fixed MTRR ranges.
The order of how the AMD MTRR setup routine is maintained by providing
a x86_setup_fixed_mtrrs_no_enable() function which does not enable
the fixed MTRRs after setting them up. All Kconfig files which had a
Makefile that included amd/mtrr in the subdirs-y now have a default
X86_AMD_FIXED_MTRRS selection. There may be some overlap with the
agesa and socket code, but I didn't know the best way to tease out
the interdependency.
Change-Id: I256d0210d1eb3004e2043b46374dcc0337432767
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2866
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
|
|
Change-Id: I53982d88f94255abdbb38ca18f9d891d4bc161b0
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/2858
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Take the vboot path in romstage. This will complete the haswell
support for vboot firmware selection.
Built and booted. Noted firmware select worked on an image with
RW firmware support. Also checked that recovery mode worked as
well by choosing the RO path.
Change-Id: Ie2b0a34e6c5c45e6f0d25f77a5fdbaef0324cb09
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2856
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Convert the existing haswell code to support reloctable ramstage
to use dynamic cbmem. This patch always selects DYNAMIC_CBMEM as
this option is a hard requirement for relocatable ramstage.
Aside from converting a few new API calls, a cbmem_top()
implementation is added which is defined to be at the begining of the
TSEG region. Also, use the dynamic cbmem library for allocating a
stack in ram for romstage after CAR is torn down.
Utilizing dynamic cbmem does mean that the cmem field in the gnvs
chromeos acpi table is now 0. Also, the memconsole driver in the kernel
won't be able to find the memconsole because the cbmem structure
changed.
Change-Id: I7cf98d15b97ad82abacfb36ec37b004ce4605c38
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2850
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Here's the great news: From now on you don't have to worry about
hitting the right io.h include anymore. Just forget about romcc_io.h
and use io.h instead. This cleanup has a number of advantages, like
you don't have to guard device/ includes for SMM and pre RAM
anymore. This allows to get rid of a number of ifdefs and will
generally make the code more readable and understandable.
Potentially in the future some of the code in the io.h __PRE_RAM__
path should move to device.h or other device/ includes instead,
but that's another incremental change.
Change-Id: I356f06110e2e355e9a5b4b08c132591f36fec7d9
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2872
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Using the CPU microcode update script and
Intel's Linux* Processor Microcode Data File
from 2013-02-22
Change-Id: I9bb60bdc46f69db85487ba923e62315f6e5352f9
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2845
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Using the CPU microcode update script and
Intel's Linux* Processor Microcode Data File
from 2013-02-22
Change-Id: Icaf0e39978daa9308cc2f0c4856d99fb6b7fdffa
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2844
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
for latest URL of their microcode tar ball
Change-Id: I3da2bdac4b2ca7d3f48b20ed389f6a47275d24fe
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2842
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
These base addresses are used in several places and it
is helpful to have one location that is reading it.
Change-Id: Ibf589247f37771f06c18e3e58f92aaf3f0d11271
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/2812
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The RESET_ON_INVALID_RAMSTAGE_CACHE option indicates what to do
when the ramstage cache is found to be invalid on a S3 wake. If
selected the system will perform a system reset on S3 wake when the
ramstage cache is invalid. Otherwise it will signal to load the
ramstage from cbfs.
Change-Id: I8f21fcfc7f95fb3377ed2932868aa49a68904803
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2807
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Cache the relocated ramstage into the SMM region. There is
a reserved region within the final SMM region (TSEG). Use that
space to cache the relocated ramstage program. That way, on S3 resume
there is a copy that can be loaded quickly instead of accessing the
flash. Caching the ramstage in the SMM space is also helpful in that
it prevents the OS from tampering with the ramstage program.
Change-Id: Ifa695ad1c350d5b504b14cc29d3e83c79b317a62
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2806
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The SMM region is available for multipurpose use before the SMM
handler is relocated. Provide a configurable sized region in the
TSEG for use before the SMM handler is relocated. This feature is
implemented by making the reserved size a Kconfig option. Also
make the IED region a Kconfig option as well. Lastly add some sanity
checking on the Kconfig options.
Change-Id: Idd7fccf925a8787146906ac766b7878845c75935
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2804
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The TSEG region is accessible until the SMM handler is relocated
to that region. Set the region as cacheable in romstage so that it
can be used for other purposes with fast access.
Change-Id: I92b83896e40bc26a54c2930e05c02492918e0874
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2803
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The haswell processors support the ability to save their SMM state
into MSR space instead of the memory. This feaure allows for parallel
SMM relocation handlers as well as setting the same SMBASE for each
CPU since the save state memory area is not used.
The catch is that in order determine if this feature is available the
CPU needs to be in SMM context. In order to implement parallel SMM
relocation the BSP enters the relocation handler twice. The first time
is to determine if that feature is available. If it is, then that
feature is enabled the BSP exits the relocation handler without
relocating SMBASE. It then releases the APs to run the SMM relocation
handler. After the APs have completed the relocation the BSP will
re-enter the SMM relocation handler to relocate its own SMBASE to the
final location. If the parallel SMM feature is not available the BSP
relocates its SMBASE as it did before.
This change also introduces the BSP waiting for the APs to relocate
their SMBASE before proceeding with the remainder of the boot process.
Ensured both the parallel path and the serial path still continue
to work on cold, warm, and S3 resume paths.
Change-Id: Iea24fd8f9561f1b194393cdb77c79adb48039ea2
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2801
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Now that there is a way to disseminate the presence of s3 wake more
formally use that instead of hard coded pointers in memory and stashing
magic values in device registers. The northbridge code picks up the
field's presence in the romstage_handoff structure and sets up the
acpi_slp_type variable accordingly.
Change-Id: Ida786728ce2950bd64610a99b7ad4f1ca6917a99
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2799
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Some of the functions called from assembly assume the standard
x86 32-bit ABI of passing all arguments on the stack. However,
that calling ABI can be changed by compiler flags. In order to
protect against the current implicit calling convention annotate
the functions called from assembly with the cdecl function
attribute. That tells the compiler to use the stack based parameter
calling convention.
Change-Id: I83625e1f92c6821a664b191b6ce1250977cf037a
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2794
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Now that CONFIG_RELOCTABLE_RAMSTAGE is available support it on
Haswell-based systems. This patch is comprised of the following changes:
1. Ensure that memory is not preserved when a relocatable ramstage is
enabled. There is no need.
2. Pick the proper stack to use after cache-as-ram is torn down. When
the ramstage is relocatable, finding a stack to use before vectoring
into ramstage is impossible since the ramstage is a black box with an
unknown layout.
Change-Id: I2a07a497f52375569bae9c994432a8e7e7a40224
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2793
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The current ramstage code contains uses of symbols that cause issues
when the ramstage is relocatable. There are 2 scenarios resolved by this
patch:
1. Absolute symbols that are actually sizes/limits. The symbols are
problematic when relocating a program because there is no way to
distinguish a symbol that shouldn't be relocated and one that can.
The only way to handle these symbols is to write a program to post
process the relocations and keep a whitelist of ones that shouldn't
be relocated. I don't believe that is a route that should be taken
so fix the users of these sizes/limits encoded as absolute symbols
to calculate the size at runtime or dereference a variable in memory
containing the size/limit.
2. Absoulte symbols that were relocated to a fixed address. These
absolute symbols are generated by assembly files to be placed at a
fixed location. Again, these symbols are problematic because one
can't distinguish a symbol that can't be relocated. The symbols
are again resolved at runtime to allow for proper relocation.
For the symbols defining a size either use 2 symbols and calculate the
difference or provide a variable in memory containing the size.
Change-Id: I1ef2bfe6fd531308218bcaac5dcccabf8edf932c
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2789
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
Using the CPU microcode update script and
Intel's Linux* Processor Microcode Data File
from 2013-02-22
Change-Id: I853e381240b539b204c653404ca3d46369109219
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2846
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Using the CPU microcode update script and
Intel's Linux* Processor Microcode Data File
from 2013-02-22
Change-Id: I4585288905cf7374e671894ab37f125220ae535e
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2843
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This adds a new API for cache maintenance operations. The idea is
to be more explicit about operations that are going on so it's easier
to manage branch predictor, cache, and TLB cleans and invalidations.
Also, this adds some operations that were missing but required early
on, such as branch predictor invalidation. Instruction and sync
barriers were wrong earlier as well since the imported API assumed
we compield with -march=armv5 (which we don't) and was missing
wrappers for the native ARMv7 ISB/DSB/DMB instructions.
For now, this is a start and it gives us something we can easily use
in libpayload for doing things like cleaning and invalidating dcache
when doing DMA transfers.
TODO:
- Set cache policy explicitly before re-enabling. Right now it's left
at default.
- Finish deprecating old cache maintenance API.
- We do an extra icache/dcache flush when going from bootblock to
romstage.
Change-Id: I7390981190e3213f4e1431f8e56746545c5cc7c9
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2729
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
There should be a fixed 10ms wait after sending an INIT IPI. The
previous implementation was just waiting up to 10ms for the IPI to
complete the send. That is not correct. The 10ms is unconditional
according to the documentation. No ill effects were observed with the
previous behavior, but it's important to follow the documentation.
Change-Id: Ib31d49ac74808f6eb512310e9f54a8f4abc3bfd7
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2780
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This patch parallelizes the AP startup for Haswell-based devices. It
does not touch the generic secondary startup code. Instead it provides
its own MP support matching up with the Haswell BWG. It seemed to be too
much trouble to support the old startup way and this new way. Because of
that parallel loading is the only thing supported.
A couple of things to note:
1. Micrcode needs to be loaded twice. Once before MTRR and caching is
enabled. And a second time after SMM relocation.
2. The sipi_vector is entirely self-contained. Once it is loaded and
written back to RAM the APs do not access memory outside of the
sipi_vector load location until a sync up in ramstage.
3. SMM relocation is kicked off by an IPI to self w/ SMI set as the
destination mode.
The following are timings from cbmem with dev mode disabled and recovery mode
enabled to boot directly into the kernel. This was done on the
baskingridge CRB with a 4-core 8-thread CPU and 2 DIMMs 1GiB each. The
kernel has console enabled on the serial port. Entry 70 is the device
initialization, and that is where the APs are brought up. With these two
examples it looks to shave off ~200 ms of boot time.
Before:
1:55,382
2:57,606 (2,223)
3:3,108,983 (3,051,377)
4:3,110,084 (1,101)
8:3,113,109 (3,024)
9:3,156,694 (43,585)
10:3,156,815 (120)
30:3,157,110 (295)
40:3,158,180 (1,069)
50:3,160,157 (1,977)
60:3,160,366 (208)
70:4,221,044 (1,060,677)
75:4,221,062 (18)
80:4,227,185 (6,122)
90:4,227,669 (484)
99:4,265,596 (37,927)
1000:4,267,822 (2,225)
1001:4,268,507 (685)
1002:4,268,780 (272)
1003:4,398,676 (129,896)
1004:4,398,979 (303)
1100:7,477,601 (3,078,621)
1101:7,480,210 (2,608)
After:
1:49,518
2:51,778 (2,259)
3:3,081,186 (3,029,407)
4:3,082,252 (1,066)
8:3,085,137 (2,884)
9:3,130,339 (45,202)
10:3,130,518 (178)
30:3,130,544 (26)
40:3,131,125 (580)
50:3,133,023 (1,897)
60:3,133,278 (255)
70:4,009,259 (875,980)
75:4,009,273 (13)
80:4,015,947 (6,674)
90:4,016,430 (482)
99:4,056,265 (39,835)
1000:4,058,492 (2,226)
1001:4,059,176 (684)
1002:4,059,450 (273)
1003:4,189,333 (129,883)
1004:4,189,770 (436)
1100:7,262,358 (3,072,588)
1101:7,263,926 (1,567)
Booted the baskingridge board as noted above. Also analyzed serial
messages with pcserial enabled.
Change-Id: Ifedc7f787953647c228b11afdb725686e38c4098
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2779
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This patch only applies to CONFIG_MICROCODE_IN_CBFS. The intel microcode
update routine would always walk the CBFS for the microcode file. Then
it would loop through the whole file looking for a match then load the
microcode. This process was maintained for intel_update_microcode_from_cbfs(),
however 2 new functions were exported:
1. const void *intel_microcode_find(void)
2. void intel_microcode_load_unlocked(const void *microcode_patch)
The first locates a matching microcode while the second loads that
mircocode. These new functions can then be used to cache the found
microcode blob w/o having to re-walk the CBFS.
Booted baskingridge board to Linux and noted that all microcode
revisions match on all the CPUs.
Change-Id: Ifde3f3e5c100911c4f984dd56d36664a8acdf7d5
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2778
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The non-relocatable SMM code was changed to link against libgcc a while back
so that printk could use built-in division instead of a hand crafted div()
function. However, the relocatable SMM code was not adapted by mistake.
This patch links the relocatable SMM against libgcc, too, so we can enable it
for Haswell.
Change-Id: Ia64a78e2e62348d115ae4ded52d1a02c74c5cea4
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2727
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
There are changes coming to perform more complex tasks after cache-as-ram
has been torn down but before ramstage is loaded. Therefore, add the
romstage_after_car() function to call after cache-as-ram is torn down.
Its responsibility is for loading the ramstage and any other complex
tasks. For example, the saving of OS-controlled memory in the resume
path has now been moved into C instead of assembly.
Change-Id: Ie0c229cf83a9271c8995b31c534c8e5a696b164e
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2757
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The save_mrc_data() was previously called conditionally
in the raminit code. The save_mrc_data() function was called
in the non-S3 wake paths. However, the common romstage_common()
code was checking cbmem initialization things on s3 wake. Between
the two callers cbmem_initialize() was being called twice in the
non-s3 wake paths. Moreover, saving of the mrc data was not allowed
when CONFIG_EARLY_CBMEM_INIT wasn't enabled.
Therefore, move the save_mrc_data() to romstage_common. It already has
the knowledge of the wake path. Also remove the CONFIG_EARLY_CBMEM_INIT
requirement from save_mrc_data() as well as the call to cbmem_initialize().
Change-Id: I7f0e4d752c92d9d5eedb8fa56133ec190caf77da
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2756
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Instead of hard coding the policy for the stack and MTRR values after
the cache-as-ram is torn down, allow for the C code to pass those
policies back to the cache-as-ram assembly file. That way, ramstage
relocation can use a different stack as well as different MTRR policies.
Change-Id: Ied024d933f96a12ed0703c51c506586f4b50bd14
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2755
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This commit pulls in all the common logic for romstage into
the Haswell cpu directory. The bits specific to the mainboard
still reside under their respective directories. The calling
sequence bounces from the cpu directory to mainboard then back
to the cpu directory. The reasoning is that Haswell systems use
cache-as-ram for backing memory in romstage. The stack is used to
allocate structures. However, now changes can be made to the
romstage for Haswell and apply to all boards.
Change-Id: I2bf08013c46a99235ffe4bde88a935c3378eb341
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2754
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
It was found that the Haswell reference code was smashing through the
stack into the reference code's heap implementation. The reason for this
is because our current CAR allocation is too small. Moreover there are
quite a few things to coordinate between 2 code bases to get correct.
This commit separates the CAR into 2 parts:
1. MRC CAR usage.
2. Coreboot CAR usage.
Pointers from one region can be passed between the 2 modules, but one
should not be able to affect the others as checking has been put into
place in both modules.
The CAR size has effectively been doubled from 0x20000 (128 KiB) to
0x40000 (256KiB). Not all of that increase was needed, but enforcing
a power of 2 size only utilizes 1 MTRR.
Old CAR layout with a single contiguous stack with the region starting
at CONFIG_DCACHE_RAM_BASE:
+---------------------------------------+ Offset CONFIG_DCACHE_RAM_SIZE
| MRC global variables |
| CONFIG_DCACHE_RAM_MRC_VAR_SIZE bytes |
+---------------------------------------+
| ROM stage stack |
| |
| |
+---------------------------------------+
| MRC Heap 30000 bytes |
+---------------------------------------+
| ROM stage console |
| CONFIG_CONSOLE_CAR_BUFFER_SIZE bytes |
+---------------------------------------+
| ROM stage CAR_GLOBAL variables |
+---------------------------------------+ Offset 0
There was some hard coded offsets in the reference code wrapper to start
the heap past the console buffer. Even with this commit the console
can smash into the following region depending on what size
CONFIG_CONSOLE_CAR_BUFFER_SIZE is.
As noted above This change splits the CAR region into 2 parts starting
at CONFIG_DCACHE_RAM_BASE:
+---------------------------------------+
| MRC Region |
| CONFIG_DCACHE_RAM_MRC_VAR_SIZE bytes |
+---------------------------------------+ Offset CONFIG_DCACHE_RAM_SIZE
| ROM stage stack |
| |
| |
+---------------------------------------+
| ROM stage console |
| CONFIG_CONSOLE_CAR_BUFFER_SIZE bytes |
+---------------------------------------+
| ROM stage CAR_GLOBAL variables |
+---------------------------------------+ Offset 0
Another variable was add, CONFIG_DCACHE_RAM_ROMSTAGE_STACK_SIZE,
which represents the expected stack usage for the romstage. A marker
is checked at the base of the stack to determine if either the stack
was smashed or the console encroached on the stack.
Change-Id: Id76f2fe4a5cf1c776c8f0019f406593f68e443a7
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2752
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Haswell's BCLK is fised at 100MHz like Sandy/Ivy. Add Haswell's model
to the switch statement.
Change-Id: Ib9e2afc04eba940bfcee92a6ee5402759b21cc45
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2747
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The SMM handler resides in the TSEG region which is far above
CONFIG_RAM_TOP (which is the highest cacheable address) before
MTRRs are setup. This means that calling initialize_cpus() before
performing MTRR setup on the BSP means the SMM handler is copied
using uncacheable accesses.
Improve the SMM handler setup path by enabling performing MTRR setup on
for the BSP before the call to initialize_cpus(). In order to do this
the haswell_init() function was split into 2 paths: BSP & AP paths.
There is a cpu_common_init() that both call to perform similar
functionality. The BSP path in haswell_init() then starts the APs using
intel_cores_init(). The AP path in haswell_init() loads microcode and
sets up MTRRs.
This split will be leveraged for future support of bringing up APs in
parallel as well as adhering to the Haswell MP initialization
requirements.
Change-Id: Id8e17af149e68d708f3d4765e38b1c61f7ebb470
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2746
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The configure_mca() function was hard coding the number of
banks the cpu supported. Query this dynamically so that it
no longer clears only 7 banks.
Change-Id: I33fce8fadc0facd1016b3295faaf3ae90e490a71
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2745
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This just moves the definiton of CORE_THREAD_COUNT_MSR so
that future code can utilize it.
Change-Id: I15a381090f21ff758288f55dc964b6694feb6064
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2744
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This commit adds support for using the SMM modules for haswell-based
boards. The SMI handling was also refactored to put the relocation
handler and permanent SMM handler loading in the cpu directory. All
tseg adjustment support is dropped by relying on the SMM module support
to perform the necessary relocations.
Change-Id: I8dd23610772fc4408567d9f4adf339596eac7b1f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2728
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Haswell CPUs require a FIT table in the firmware. This commit
adds rudimentary support for a FIT table. The number of entries
in the table is based on a configuration option. The code only
generates a type 0 entry. A follow-on tool will need to be developed
to populate the FIT entries as well as checksumming the table.
Verified image has a FIT pointer and table when option is selected.
Change-Id: I3a314016a09a1cc26bf1fb5d17aa50853d2ef4f8
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2642
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The Link native graphics commit 49428d84 [1]
Add support for Google's Chromebook Pixel
was missing some of the higher level bits, and hence could not be
used. This is not new code -- it has been working since last
August -- so the effort now is to get it into the tree and structure
it in a way compatible with upstream coreboot.
1. Add options to src/device/Kconfig to enable native graphics.
2. Export the MTRR function for setting variable MTRRs.
3. Clean up some of the comments and white space.
While I realize that the product name is Pixel, the mainboard in the
coreboot tree is called Link, and that name is what we will use
in our commits.
[1] http://review.coreboot.org/2482
Change-Id: Ie4db21f245cf5062fe3a8ee913d05dd79030e3e8
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/2531
Tested-by: build bot (Jenkins)
|
|
This adds microcode ffff000a and the CPUIDs for ULT.
Change-Id: I341c1148a355d8373b31032b9f209232bd03230a
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/2647
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
|
|
The IA32_ENERGY_PERFORMANCE_BIAS MSR can only be read or written
to if the CPU supports it. The support is indicated by ECX[3] for
cpuid(6). Without this guard, some Haswell parts would GP# fault
in this routine.
No more GP# while running on haswell CRBs.
Change-Id: If41e1e133e5faebb3ed578cba60743ce7e1c196f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2639
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martin.roth@se-eng.com>
|
|
Add support for SMM modules by leveraging the RMODULE lib. This allows
for easier dynamic SMM handler placement. The SMM module support
consists of a common stub which puts the executing CPU into protected
mode and calls into a pre-defined handler. This stub can then be used
for SMM relocation as well as the real SMM handler. For the relocation
one can call back into coreboot ramstage code to perform relocation in
C code.
The handler is essentially a copy of smihandler.c, but it drops the TSEG
differences. It also doesn't rely on the SMM revision as the cpu code
should know what processor it is supported.
Ideally the CONFIG_SMM_TSEG option could be removed once the existing
users of that option transitioned away from tseg_relocate() and
smi_get_tseg_base().
The generic SMI callbacks are now not marked as weak in the
declaration so that there aren't unlinked references. The handler
has default implementations of the generic SMI callbacks which are
marked as weak. If an external compilation module has a strong symbol
the linker will use that instead of the link one.
Additionally, the parameters to the generic callbacks are dropped as
they don't seem to be used directly. The SMM runtime can provide the
necessary support if needed.
Change-Id: I1e2fed71a40b2eb03197697d29e9c4b246e3b25e
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2693
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The Haswell parts use a PCH code named Lynx Point (Series 8). Therefore,
the southbridge support is included as well. The basis for this code is
the Sandybridge code. Management Engine, IRQ routing, and ACPI still requires
more attention, but this is a good starting point.
This code partially gets up through the romstage just before training
memory on a Haswell reference board.
Change-Id: If572d6c21ca051b486b82a924ca0ffe05c4d0ad4
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2616
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The original code attempted to reserve a space in RAM for coreboot to
remain resident. This turns out not to be needed, and breaks things
for the kernel since the exynos5250-smdk5250 kernel device tree starts
RAM at 0x40000000.
(This patch was originally by Gabe, I'm just uploading it)
Change-Id: I4536edaf8785d81a3ea008216a2d57549ce5edfb
Signed-off-by: Gabe Black <gabeblack@chromium.org>
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2698
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
|
|
This enables branch prediction. We can probably find a better place
to do this, but for now we'll do it in snow's romstage main().
Change-Id: I86c7b6bc9e897a7a432c490fb96a126e81b8ce72
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2701
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
PS_HOLD gets set in exynos' power_init().
Change-Id: Ib08e0afcad23cbd07dc7e3727fd958a1bc868b5a
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2700
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Call the power_init() function. We appear to have forgotten about it
when deprecating lowlevel_init_subsystems(), but it didn't seem to
cause problems until we got to doing more interesting stuff recently.
There are some clean-ups to do from the original code, such as not
attempting to configure I2C from PMIC code, which we'll get around
to in follow-up patches.
(Credit to Gabe for spotting this)
Change-Id: I6a59379e9323277d0b61469de9abe6d651ac5bfb
Signed-off-by: Gabe Black <gabeblack@chromium.org>
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2699
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This eliminates the use of do_div() in favor of using libgcc
functions.
This was tested by building and booting on Google Snow (ARMv7)
and Qemu (x86). printk()s which use division in vtxprintf() look good.
Change-Id: Icad001d84a3c05bfbf77098f3d644816280b4a4d
Signed-off-by: Gabe Black <gabeblack@chromium.org>
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2606
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
This adds an enum for GPIO ports on the Exynos5. To make them
useful, they are assigned the absolute MMIO address where a
s5p_gpio_bank struct can point to.
Change-Id: Ia539ba52d7393501d434ba8fecde01da37b0d8aa
Signed-off-by: David Hendricks <dhendrix@chromium.org>
Reviewed-on: http://review.coreboot.org/2602
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Models 6ex and 6fx select UDELAY_LAPIC so cannot select
contradicting UDELAY_TSC here.
Model 1067x requires speedstep.
Change-Id: I69d3ec8085912dfbe5fe31c81fa0a437228fa48f
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/2525
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
|
|
These are essential functions for setting up the display port and
framebuffer, and also enable such things as aux channel
communications. We do some very simple initialization in romstage,
mainly set a GPIO so that the graphics is powering up, but the complex
parts are done in the ramstage. This mirrors the way in which graphics
is done in the x86 size.
I've added a first pass at a real device, and put it in the mainboard
Kconfig, hoping for corrections. Because startup is so complex,
depending on device type, I've created a 'displayport' device that
removes some of the complexity and makes the flow *much* clearer. You
can actually follow the flow by looking at the code, which is not true
on other implementations. Since display port is perhaps the main port
used on these chips, that's a reasonable compromise. All parameters of
importance are now in the device tree.
Change-Id: I56400ec9016ecb8716ec5a5dae41fdfbfff4817a
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/2570
Tested-by: build bot (Jenkins)
|
|
OK, this is tl;dr. But I need to write this in hopes we make
sure we don't put code like this into coreboot. Ever.
Our excuse in this case is that it was imported, not obviously wrong,
and easily changed. It made sense to get it in, make it work, then
do a cleanup pass, because changing everything up front is almost
impossible to debug.
The exynos code has bunch of base register values, e.g.
These are base addresses of things that look like a memory-mapped
struct. To get these to a pointer, they created the following macro,
which creates an inline function.
static inline unsigned int samsung_get_base_##device(void) \
{ \
return cpu_is_exynos5() ? EXYNOS5_##base : 0; \
}
And then invoke it 31 times in a .h file, e.g.:
SAMSUNG_BASE(clock, CLOCK_BASE)
to create 31 functions.
And then use it:
struct exynos5_clock *clk =
(struct exynos5_clock *)samsung_get_base_clock();
OK, what's wrong with this? It's easier to ask what's right with it. Answer: nothing.
I have a long list of what's wrong, and I may leave some things out,
but here goes:
1. the "function" can return a NULL if we're not on exynos5. Most uses of the code
don't check the return value.
2. And why would this function be running, if we're not on an exynos5? Why compile it in?
3. Note the cast everywhere a samsung_get_base_xxx is used.
The function returns an untyped variable, requiring the *user* to get two
things right: the cast, and the function invocation. One can replace that _clock(); with
_power(); in the code above, and they will be referencing the wrong registers, and
they'll never get an error!
We have a C compiler; use it to type data.
4. You're generating 31 functions using cpp each and every time the file is included.
The C compiler has to parse these each time. It's not at all like a simple cpp
macro which is only generated on use.
5. You can't tags or etags this code
6. In fact, any kind of analysis tool will be unable to do anything with this cpp magic.
That's only a partial list.
So what's the right way to do it? Just make typed constants, viz:
Or, since I expect people will want the lower case function syntax, I've left
it that way:
Now we've got something that is efficient, and we don't even need to protect with
any more.
Hence this change. We've got something that is type checked, does not require users to
cast on each use, will catch simple programming errors, can be analyzed with standard tools,
and builds faster.
So if we make a mistake:
struct exynos5_clock *clk =
samsung_get_base_adc();
We'll see it:
src/cpu/samsung/exynos5250/clock.c: In function 'get_pll_clk':
src/cpu/samsung/exynos5250/clock.c:183:3: error: initialization from incompatible pointer type [-Werror]
which we would not have seen before.
As a minor benefit, it shaves most of a second off the compilation.
Change-Id: Ie67bc4bc038a8dd1837b977d07332d7d7fd6be1f
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-on: http://review.coreboot.org/2582
Tested-by: build bot (Jenkins)
|
|
In the file `COPYING` in the coreboot repository and upstream [1]
just one space is used.
The following command was used to convert all files.
$ git grep -l 'MA 02' | xargs sed -i 's/MA 02/MA 02/'
[1] http://www.gnu.org/licenses/gpl-2.0.txt
Change-Id: Ic956dab2820a9e2ccb7841cab66966ba168f305f
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: http://review.coreboot.org/2490
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
|
|
It's been on for all boards per default since several years now
and the old code path probably doesn't even work anymore. Let's
just have one consistent way of doing things.
Change-Id: I58da7fe9b89a648d9a7165d37e0e35c88c06ac7e
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2547
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
The ARM CPUs we know of don't have CPU microcode updates,
so don't show the selection in Kconfig.
Also simplify (and fix) the microcode selection in the Makefile
that would try to include microcode even though none is available.
Change-Id: I502d9b48d4449c1a759b5e90478ad37eef866406
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2540
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
|
|
According to both Haswell and the SandyBridge/Ivybridge
BWGs the save state area actually starts at 0x7c00 offset
from 0x8000. Update the em64t101_smm_state_save_area_t
structure and introduce a define for the offset.
Note: I have no idea what eptp is. It's just listed in the
haswell BWG. The offsets should not be changed.
Change-Id: I38d1d1469e30628a83f10b188ab2fe53d5a50e5a
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/2515
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
|
|
Commit »AMD S3: Introduce Kconfig variable 'S3_DATA_SIZE'« (22ec9f9a) [1]
introduced a check throwing an error if S3_DATA_SIZE isn't big enough.
However without CONFIG_HAVE_ACPI_RESUME the variable S3_DATA_SIZE
isn't defined at all and compilation will fail if s3_resume.h is
included.
This patch makes it again possible turn off HAVE_ACPI_RESUME relatively
easily in Parmer/Thatcher/Persimmon's Kconfig if you don't care about S3
and don't want flash writes on every boot.
[1] http://review.coreboot.org/2383
Change-Id: I999e4b7634bf172d8380fd14cba6f7f03468fee3
Signed-off-by: Jens Rottmann <JRottmann@LiPPERTembedded.de>
Reviewed-by: Dave Frodin <dave.frodin@se-eng.com>
Reviewed-by: Marc Jones <marcj303@gmail.com>
Reviewed-on: http://review.coreboot.org/2528
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
|
|
This reverts commit ca6e1f6c04c96c435bdbf30a1b88cab0e5be330b.
The packet size changes ends up corrupting the flash when booting
Persimmon. I did figure out that the maximum number of bytes that
can be sent is actually 8 bytes according to the sb800 spec. There
must be additional problems beyond that since setting the packet
size to 8 still causes problems.
Change-Id: Ieb24247cf79e95bb0e548c83601dfddffbf6be59
Signed-off-by: Dave Frodin <dave.frodin@se-eng.com>
Reviewed-on: http://review.coreboot.org/2509
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martin.roth@se-eng.com>
Reviewed-by: Zheng Bao <zheng.bao@amd.com>
|
