If the firwmare is flashed and the MRC cache is blown away
then it is not possible to resume.
Right now this can be inferred from the event log but it can
be made very clear by adding a unique post code for this event.
1) boot falco
2) flash firmware
3) suspend and then resume
4) check for post code 0xef in log
0 | 2013-08-08 16:27:47 | Log area cleared | 4096
1 | 2013-08-08 16:27:47 | ACPI Enter | S3
2 | 2013-08-08 16:27:55 | System boot | 48
3 | 2013-08-08 16:27:55 | Last post code in previous boot | 0xef | Resume Failure
4 | 2013-08-08 16:27:55 | System Reset
5 | 2013-08-08 16:27:55 | ACPI Wake | S5
Change-Id: I7602d9eef85d3b764781990249ae32b84fe84134
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/65259
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4458
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The PCIe root port has ASPM settings/workarounds that are only applied
based on the value of an undocumented bit in PCI config register 0x32C.
If that bit is not set for some reason then the settings are not applied.
This devicetree config option will force the ASPM settings for each port
based on the bit map.
Change-Id: I40b08ca9a0ef52742609bac72fb821454a373799
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/65314
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4453
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The default ME output is quite verbose and not all that useful
unless you are actively debugging the ME and then you can enable
the CONFIG_DEBUG_INTEL_ME option.
This commit silences the firmware capabilities and the MBP output.
Change-Id: I2b8abcb34ae0d00d9a38d029979e84ee0d0ca287
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/65252
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4452
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
CLKOUT for PCIE ports 1-5 and CLKOUT_XDP are not used
and can be disabled.
I couldn't test this directly without a scope so instead I
used a modified commit that also disabled PCIe Port 0 and
saw that that correctly disabled the WLAN port.
Change-Id: I0f996e90f0ae42780de3a0c8dc5db00ec600748b
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/65251
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4451
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The original ME code was assuming that the only type of messages
it would send were MKHI type and so it had some embedded checks
for that header and that type of message.
In order to support ICC messages this needs to change to handle
different header types, so now the header will be sent first
and then the data will follow, rather than the two both being
sent in the same low-level function.
This change has no real affect on the system, subsequent commit
will add new ICC messages.
Change-Id: I52848581e49b88c0a79e8bb6bda2a179419808a3
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/65249
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4449
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
When the EC requests the host to throttle (for charging or thermal
related reasons) the package power consumption will be limited.
Right now this is set at 12W but that is somewhat arbitrary and may
need tuning.
1) define the THRT method in \_TZ scope for EC to call
2) enable SCI events for throttle start and stop
3) define the power limit at 12W and set it in NVS
1) Enable CONFIG_ACPI_DEBUG=y in the kernel
2) Enable the Debug object event in acpi module
acpi.debug_layer=0x7f acpi.debug_level=0x2f
3) Using EC console generate host event for throttle start
> hostevent set 0x20000
4) Check dmesg for throttle start events
ACPI: Execute Method [\_SB_.PCI0.LPCB.EC0_._Q12] (Node ffff8801002c5988)
[ACPI Debug] String [0x12] "EC: THROTTLE START"
[ACPI Debug] String [0x10] "Enable PL1 Limit"
5) Using EC console generate host event for throttle stop
> hostevent set 0x40000
6) Check dmesg for throttle stop events
ACPI: Execute Method [\_SB_.PCI0.LPCB.EC0_._Q13] (Node ffff8801002c59b0)
[ACPI Debug] String [0x11] "EC: THROTTLE STOP"
[ACPI Debug] String [0x11] "Disable PL1 Limit"
Change-Id: I39b53a5e8abc2892846bcd214a333fe204c6da9b
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/63989
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4416
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Two new events possible from the EC for starting and stopping throttle.
These are handled in a per-board method that is defined under the
thermal zone. This is not quite where I wanted it but the scoping
rules in ACPI don't let me have a defined external object in the
same scope.
Change-Id: I766f07b4365b29df3daa8e45e88f7c38c645c287
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/63988
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4415
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
An issue was observed using a specific vendor's TPM in that it
chokes on access to registers that are not explicitly defined in the
PC client specification. The previous driver used generic access
functions for reading and writing registers. However, issues come
to play when reading from the status register. It read it as a 32-bit
value, but that read address 0x1b which is not defined in the spec.
Instead of using generic access functions for the tpm registers
provide explicit ones. To that end provide more high level wrapper
functions to perform the semantic access required.
Change-Id: I781b31723f819e1387d7aa25512c83780ea0877f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/63243
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/4388
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Now that the rtd2132 device has the full settings the
panel timings need to be implemented. Sadly, the Tx timings
in the rtd2132 aren't 1:1 with the panel's Tx timings. Below
is the table equivalent:
RTD2132 | Falco Panel
--------+------------
T1 | T2
--------+------------
T2 | T8+T10+T12
--------+------------
T3 | T14
--------+------------
T4 | T15
--------+------------
T5 | T9+T11+T13
--------+------------
T6 | T3
--------+------------
T7 | T4
--------+------------
Change-Id: I10a3ad475d6b9485a707eb49e31afd197fc8d24d
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/65858
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4472
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
It has been disseminated that the RTD2132 chip
needs to be fully programmed for settings to take affect.
Most of the settings are note documented very well and
present themselves as magic values. Also, the wait time
for starting the sequence needs to be bumped from 2ms to 60ms.
Lastly, expose all the known settings through devicetree.
Change-Id: I9eeea9c4a13ec20b8ce1c5297e43c4dd793d90e5
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/65857
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4471
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
When we go through the resume path, there shouldn't ever be a need to
initialize the PS/2 keyboard. The OS is going to reinitialize it
anyway, and it just slows the resume.
Verified Code flow in normal boot/S3 resume with print statements.
Verified Keyboard was correctly disabled and flushed by booting
to recovery mode screen while pressing keys on the integrated
keyboard.
Change-Id: I48bdca2fa2cc0c965401d10fef75cadb09d2e1e9
Signed-off-by: Martin Roth <martin.roth@se-eng.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/63648
Reviewed-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: Shawn Nematbakhsh <shawnn@chromium.org>
Commit-Queue: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: http://review.coreboot.org/4396
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
When the board is in S3 and S5 the WLAN_DISABLE_L signal
can leak power into the WLAN power well since the GPIO
controlling WLAN_DISABLE_L is in the suspend well. Therefore,
drive WLAN_DISABLE_L low to avoid the power leak.
This is a clone of a Falco change:
I1a0df80dd47fdbd535aca7a9d49253794c480606.
Change-Id: I625dfbb228d1f293b880a52dfe552842d55a17d1
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/63220
Reviewed-by: Dave Parker <dparker@chromium.org>
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
Reviewed-on: http://review.coreboot.org/4383
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
... based on the EDID detailed timing values for
pixel_clock and link_clock.
Two undocumented registers 0x6f040 and 0x6f044 correspond to link_m and link_n
respectively. Other two undocumented registers 0x6f030 and 0x6f034 correspond
to data_m and data_n respectively.
Calculations are based on the intel_link_compute_m_n from linux kernel.
Currently, the value for 0x6f030 does not come up right with our calculations.
Hence, set to hard-coded value.
Change-Id: I40ff411729d0a61759164c3c1098504973f9cf5e
Reviewed-on: https://gerrit.chromium.org/gerrit/62915
Reviewed-by: Ronald G. Minnich <rminnich@chromium.org>
Tested-by: Furquan Shaikh <furquan@chromium.org>
Commit-Queue: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: http://review.coreboot.org/4381
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The linux kernel will unconditionally route all USB
ports to the XCHI controller at boot. The EHCI controller
can then be disabled, and it should be left disabled
by the reference code when this is done.
However not all OS may do this unconditional route,
so provide an option to the reference code binary to
enable this behavior.
Change-Id: Iedf5af54182bf109cd1119c1999e46300665d41e
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/63797
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4403
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The management engine is occasionally hanging the system on resume
when it is accessed. Since we actually don't need to do anything
with it on resume it can be disabled early in the resume path and
avoid assigning resources just to remove them later.
suspend/resume on falco and check /sys/firmware/log
to ensure that device 00:16.0 is disabled early and that no
resources are probed or assigned and that the device init path
does not execute.
Change-Id: I35573681e3a1d43d816d24954842cbe9c61f3484
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/62897
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4376
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The management engine is slow, requiring at least 500ms between
when the Dram Init Done message is sent (right after memory training)
to when the MBP will report that it is successfully cleared and
that the ME can finally be sent the EOP message.
Currently this is adding 100-150ms to the boot time. If we defer
waiting for the MBP Clear indicator until the finalize step we
can gain back that lost time.
boot on falco with SMI debugging enabled to
ensure that the ME is locked down in the finalize step:
Finalizing Coreboot
SMI# #0
SMI_STS: PM1 APM
ME: MBP cleared
ME: mkhi_end_of_post
ME: END OF POST message successful (0)
Change-Id: Icab4c8c8e00eea67bed5e8154d91a1eb48a492d1
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/62633
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4375
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
There are specific programming requirements for the usb3 ports
on all LynxPoint chipsets when transitioning to D0 or D3.
LynxPoint-LP has additional workaround steps needed involving
resetting the disconnected ports when transitioning to D0.
The workarounds are implemented in ACPI code so the controller
can transition properly into D3 at runtime.
Change-Id: I3b428562f48c9cb250b97779a3b2753ed4f81509
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/62632
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4374
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The intel_ddi.c change I thought should be in but I don't see it. It just adds two functions back
that we need.
There are two new files for slippy annotated with comments about how it needs to evolve.
That said, this code has been tested on 3 different panels. Both dev and non-dev usages work.
physbase initialization to static value removed.
Moved spin calls to intel_dp_*
Change-Id: I0480af45c21c7dedcaff7e8be729f0eb554ec78a
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/61136
Commit-Queue: Ronald G. Minnich <rminnich@chromium.org>
Reviewed-by: Ronald G. Minnich <rminnich@chromium.org>
Tested-by: Ronald G. Minnich <rminnich@chromium.org>
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/4370
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Tested-by: build bot (Jenkins)
Peppy SPD table has 4GB configurations followed by 2GB configurations.
Current implementation does remapping to point 2GB configuration to the
same SPD index as the 4GB. This is different than Falco, which simply
duplicates the SPD data for all configurations. To simplify probing in
mosys, copy the Falco implementation of duplicating SPD data.
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Change-Id: Idb185a437f3cf4f40d2dae1ae59c30235df8f489
Reviewed-on: https://gerrit.chromium.org/gerrit/61847
Reviewed-by: Dave Parker <dparker@chromium.org>
Reviewed-by: Jay Kim <yongjaek@chromium.org>
Commit-Queue: Shawn Nematbakhsh <shawnn@chromium.org>
Tested-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: http://review.coreboot.org/4369
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Tested-by: build bot (Jenkins)
When using RW firmware path the proper recovery reason can
be retrieved from the shared data region. This will result
in the actual reason being logged instead of the default
"recovery button pressed" reason.
1) build and boot on falco
2) crossystem recovery_request=193
3) reboot into recovery mode, check reason with <TAB>
4) reboot back into chromeos
5) check event log entry for previous recovery mode:
25 | 2013-07-15 10:34:23 | Chrome OS Recovery Mode | Test from User Mode
Change-Id: I6f9dfed501f06881e9cf4392724ad28b97521305
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/61906
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/4368
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Tested-by: build bot (Jenkins)
The EC temperature sensors were renumbered and now PECI
is at index 0.
1) boot on falco
2) check /sys/class/thermal/thermal_zone0/temp
3) check 'temps' on ec console
Change-Id: Idde1457c42c80850b5b8ac22781060ed9b224d13
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/61896
Reviewed-on: http://review.coreboot.org/4367
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Tested-by: build bot (Jenkins)
This may need further tuning but will start at 1.0%.
boot on falco and check /sys/firmware/log
localhost ~ # grep RTD2132 /sys/firmware/log
RTD2132: Enable 1.0% Spread Spectrum
I2C: 01:35 (Realtek RTD2132 LVDS Bridge)
Change-Id: I96e1c14dbc6a7bfaf1c8deb1806c48bf2fd3e32a
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/61895
Reviewed-on: http://review.coreboot.org/4366
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Tested-by: build bot (Jenkins)
