summaryrefslogtreecommitdiff
path: root/src/soc/intel/apollolake/pmutil.c
AgeCommit message (Collapse)Author
2018-02-05soc/intel/apollolake: Clear RTC failure bit after reading itFurquan Shaikh
This change ensures that the RTC failure bit is cleared in PMCON1 after cmos_init checks for it. Before this change, RPS was cleared in dev init phase. If any reboot occurred before dev init stage (e.g. FSP reset) then RPS won't be cleared and cmos_init will re-initialize CMOS data. This resulted in any information like VBNV flags stored in CMOS after first cmos_init to be lost. BUG=b:72879807 BRANCH=coral TEST=Verified that recovery request is preserved when recovery is requested without battery on coral. Change-Id: Ib23b1fcd5c3624bad6ab83dce17a469b2f5b5ba8 Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/23578 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2018-01-17soc/intel/apollolake: Fix prev_sleep_state on G3 exitHannah Williams
If waking up from S5, then prev_sleep_state was correct but not when waking up from G3. Change-Id: I39011a0846f042d224a7cd65f736e749acc8ec75 Signed-off-by: Hannah Williams <hannah.williams@intel.com> Reviewed-on: https://review.coreboot.org/23221 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-10-22security/vboot: Move vboot2 to security kconfig sectionPhilipp Deppenwiese
This commit just moves the vboot sources into the security directory and fixes kconfig/makefile paths. Fix vboot2 headers Change-Id: Icd87f95640186f7a625242a3937e1dd13347eb60 Signed-off-by: Philipp Deppenwiese <zaolin@das-labor.org> Reviewed-on: https://review.coreboot.org/22074 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Martin Roth <martinroth@google.com>
2017-10-19soc/intel/apollolake: Use newly added pmc_read_pm1_controlFurquan Shaikh
BUG=b:67874513 Change-Id: I6d5a76122b7d6e508e5ff3f4099e5d706fb48f9d Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/22082 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-10-12soc/intel/common: Clean up PMC library GPE handling APIFurquan Shaikh
1. Update gpe handling function names to explicitly mention if they are operating on: a. STD GPE events b. GPIO GPE events c. Both 2. Update comment block in pmclib.h to use generic names for STD and GPIO GPE registers instead of using any one platform specific names. BUG=b:67712608 Change-Id: I03349fe85ac31d4215418b884afd8c4b531e68d3 Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/21968 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2017-09-20vboot: reset vbnv in cmos when cmos failure occursAaron Durbin
There's an occasional issue on machines which use CMOS for their vbnv storage. The machine that just powers up from complete G3 would have had their RTC rail not held up. The contents of vbnv in CMOS could pass the crc8 though the values could be bad. In order to fix this introduce two functions: 1. vbnv_init_cmos() 2. vbnv_cmos_failed() At the start of vboot the CMOS is queried for failure. If there is a failure indicated then the vbnv data is restored from flash backup or reset to known values when there is no flash backup. BUG=b:63054105 Change-Id: I8bd6f28f64a116b84a08ce4779cd4dc73c0f2f3d Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/21560 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Julius Werner <jwerner@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-09-20soc/intel/apollolake: refactor rtc failure checkingAaron Durbin
In order to prepare for checking RTC failure in the early boot paths move the rtc failure calculation to pmutil.c and add a helper function to determine if failure occurred. BUG=b:63054105 Change-Id: I1b02028a1830ff9b28b23da7a4a1fd343f329f0d Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/21551 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-08-08soc/intel/apollolake: Use common PMC for apollolakeShaunak Saha
With this patch apollolake uses the common PMC util code.No regression observed on a APL platform. Change-Id: I322a25a8b608d7fe98bec626c6696e723357a9d2 Signed-off-by: Shaunak Saha <shaunak.saha@intel.com> Reviewed-on: https://review.coreboot.org/19375 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-06-09soc/intel/apollolake: Use CPU common library codeBarnali Sarkar
This patch makes SOC files to use common/block/cpu/cpulib.c file's helper functions. Change-Id: I529c67cf20253cf819d1c13849300788104b083c Signed-off-by: Barnali Sarkar <barnali.sarkar@intel.com> Reviewed-on: https://review.coreboot.org/19827 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Subrata Banik <subrata.banik@intel.com> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2017-06-09soc/intel/apollolake: Rename ACPI Base Address and Size MacroBarnali Sarkar
Rename these two Macros to help use Common Code - ACPI_PMIO_BASE --> ACPI_BASE_ADDRESS ACPI_PMIO_SIZE --> ACPI_BASE_SIZE Change-Id: I21125b7206c241692cfdf1cdb10b8b3dee62b24a Signed-off-by: Barnali Sarkar <barnali.sarkar@intel.com> Reviewed-on: https://review.coreboot.org/20038 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Subrata Banik <subrata.banik@intel.com> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2017-04-25lib: provide clearer devicetree semanticsAaron Durbin
The devicetree data structures have been available in more than just ramstage and romstage. In order to provide clearer and consistent semantics two new macros are provided: 1. DEVTREE_EARLY which is true when !ENV_RAMSTAGE 2. DEVTREE_CONST as a replacment for ROMSTAGE_CONST The ROMSTAGE_CONST attribute is used in the source code to mark the devicetree data structures as const in early stages even though it's not just romstage. Therefore, rename the attribute to DEVTREE_CONST as that's the actual usage. The only place where the usage was not devicetree related is console_loglevel, but the same name was used for consistency. Any stage that is not ramstage has the const C attribute applied when DEVTREE_CONST is used. Change-Id: Ibd51c2628dc8f68e0896974f7e4e7c8588d333ed Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/19333 Tested-by: build bot (Jenkins) Reviewed-by: Philippe Mathieu-Daudé <philippe.mathieu.daude@gmail.com> Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-03-29soc/intel/apollolake: Fix debug build booting issueSubrata Banik
This patch fix apollolake devices unable to boot with coreboot debug image issue. Change-Id: I28943100ba19dec1e540fdbba1c1e110c6af1488 Signed-off-by: Subrata Banik <subrata.banik@intel.com> Reviewed-on: https://review.coreboot.org/19036 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Martin Roth <martinroth@google.com> Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-03-28soc/pci_devs.h: Use consistent naming in soc/pci_devs.hSubrata Banik
This patch to make common PCI device name between APL and SKL. Change-Id: I5e4c7502e9678c0a367e9c7a96cf848d5b24f68e Signed-off-by: Barnali Sarkar <barnali.sarkar@intel.com> Signed-off-by: Subrata Banik <subrata.banik@intel.com> Reviewed-on: https://review.coreboot.org/18576 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Martin Roth <martinroth@google.com>
2017-03-13soc/intel/apollolake: Fix spacing around operators and commasLee Leahy
Fix the following errors detected by checkpatch.pl: ERROR: spaces required around that '==' (ctx:VxO) ERROR: space required before that '-' (ctx:OxV) ERROR: spaces required around that '=' (ctx:VxW) ERROR: spaces required around that '=' (ctx:WxV) ERROR: spaces required around that '=' (ctx:VxV) ERROR: need consistent spacing around '+' (ctx:VxW) ERROR: space prohibited before that '++' (ctx:WxB) ERROR: space prohibited before that ',' (ctx:WxW) ERROR: space required after that ',' (ctx:VxV) TEST=Build for reef Change-Id: I37265a69fcb14fbf7c182ef29d823f70a5748ad8 Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com> Reviewed-on: https://review.coreboot.org/18720 Tested-by: build bot (Jenkins) Reviewed-by: Martin Roth <martinroth@google.com>
2016-12-02soc/intel/apollolake: Enable ACPI PM timer emulation on all CPUsAndrey Petrov
Currently we enable ACPI PM timer emulation only on BSP. So the timer doesn't work on other cores and that breaks OSes that use it. Also, microcode uses this information to figure out ACPI IO base, and that is used for other features. This patch enables ACPI timer emulation on all the logical CPUs. BUG=chrome-os-partner:60011 TEST=iotools rdmsr x 0x121, x={0..3}, make sure it is set Change-Id: I0d6cb8761c1c25d3a2fcf59a49c1eda9e4ccc70c Signed-off-by: Andrey Petrov <andrey.petrov@intel.com> Reviewed-on: https://review.coreboot.org/17663 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2016-11-07soc/intel/{sky,apollo}lake: Wait until GPE is clear when readingDuncan Laurie
When reading+clearing a GPE for use as an interrupt we need to re-read the status register and keep setting the clear bit until it actually reads back clear. Also add a 1ms timeout in case the status never clears. This is needed if a device sends a longer interrupt pulse and it is still asserted when the "ISR" goes to clear the status. BUG=chrome-os-partner:59299 TEST=test cr50 TPM with 20us pulse to ensure it can successfully communicate with the TPM and does not get confused due to seeing interrupts that it should not. Change-Id: I384f484a1728038d3a355586146deee089b22dd9 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/17212 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2016-09-28soc/intel/apollolake: Use fixed resource for SRAM and IPC1Lijian Zhao
Intel telemetry support will require PMC IPC1 and SRAM devices to be operated in ACPI mode. Then using fixed resources on BAR0, BAR1 and BAR2 (PMC only) for those two devices will help the resource assignment in DSDT stage. BUG=chrome-os-partner:57364 BRANCH=None TEST=Boot up into Chrome OS successfully and check with dmesg to see the driver has been loaded successfully without errors. Change-Id: I8f0983a90728b9148a124ae3443ec29cd7b344ce Signed-off-by: Zhao, Lijian <lijian.zhao@intel.com> Reviewed-on: https://review.coreboot.org/16648 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-09-21soc/intel/apollolake: Add function to read and clear GPE statusDuncan Laurie
Implement the generic acpi_get_gpe() function to read and clear the GPE status for a specific GPE. Tested by watching GPE status in a loop while generating interrupts manually from the EC console. BUG=chrome-os-partner:53336 Change-Id: I482ff52051a48441333b573f1cd0fa7f7579a6ab Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/16671 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-09-21soc/intel/apollolake: Initialize GPEs in bootblockDuncan Laurie
Initialize the GPEs from mainboard config in bootblock, so they can be used in verstage to query latched interrupt status. I still left it called in ramstage just to be sure that the configuration was not overwritten in FSP stages. Tested by reading and reporting GPE status in a loop in verstage and manually triggering an interrupt on EC console. BUG=chrome-os-partner:53336 Change-Id: Iacd0483e4b3229aca602bb5bb40586eedf35a6ea Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/16670 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-09-19Revert "soc/intel/apollolake: Initialize GPEs in bootblock"Duncan Laurie
This reverts commit 5e3dad66227bba4be9365ee76d00231bb5577b56.
2016-09-19Revert "soc/intel/apollolake: Add function to read and clear GPE status"Duncan Laurie
This reverts commit 3d43a7c111d00be246160a04023fe438ae0cac57.
2016-09-19soc/intel/apollolake: Add function to read and clear GPE statusDuncan Laurie
Implement the generic acpi_get_gpe() function to read and clear the GPE status for a specific GPE. Tested by watching GPE status in a loop while generating interrupts manually from the EC console. BUG=chrome-os-partner:53336 Change-Id: Id885e98d48c2133a868da19eca3360e2dfb82e84 Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
2016-09-19soc/intel/apollolake: Initialize GPEs in bootblockDuncan Laurie
Initialize the GPEs from mainboard config in bootblock, so they can be used in verstage to query latched interrupt status. I still left it called in ramstage just to be sure that the configuration was not overwritten in FSP stages. Tested by reading and reporting GPE status in a loop in verstage and manually triggering an interrupt on EC console. BUG=chrome-os-partner:53336 Change-Id: I1af3e9ac1e5c59b9ebb5c6dd1599309c1f036581 Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
2016-08-25intel/apollolake: Add power management utility functionShaunak Saha
This patch adds a power management utility function to clear wake status bits in ACPI GPE0 registers. We need to call this function before going to sleep from common smi handler function. BUG=chrome-os-partner:55583 BRANCH=None TEST=Verified that system goes to sleep on lidclose and powerd_dbus_suspend command issued from built-in keyboard. Change-Id: Icd095d377c82f2e154f2e2db773f737aa49cda64 Signed-off-by: Shaunak Saha <shaunak.saha@intel.com> Reviewed-on: https://review.coreboot.org/16298 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-08-20intel/apollolake: Do not halt in poweroff if in SMMFurquan Shaikh
Calling halt in poweroff when in SMM prevents SLP_SMI to be triggered preventing the system from entering sleep state. Fix this by calling halt only if ENV_SMM is not true. BUG=chrome-os-partner:56395 BRANCH=None TEST=Verified lidclose behavior on reef. Change-Id: If116c8f4e867543abdc2ff235457c167b5073767 Signed-off-by: Furquan Shaikh <furquan@google.com> Reviewed-on: https://review.coreboot.org/16257 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-08-04soc/apollolake: Return correct wake status in _SWSShaunak Saha
Wake status is calculated from the four pairs of gpe0 in cbmem CBMEM_ID_POWER_STATE which is filled very early in romstage and depends on the routing information in PMC GPE_CFG register. Coreboot sets the proper value of routing based on devicetree from pmc_init. But when system goes to S3 on waking up PMC is writing default values again in GPE_CFG which results in returning wrong wake status in _SWS. This patch corrects that behaviour by correcting the gpe0 pairs in cbmem after PMC sets the routing table in resume path. BUG=chrome-os-partner:54876 TEST=On resume through powerbtn, lidopen, keyboard press, etc. we are getting proper wake status. Change-Id: I5942d5c20d8c6aef73468dc611190bb7c49c7c7a Signed-off-by: Shaunak Saha <shaunak.saha@intel.com> Reviewed-on: https://review.coreboot.org/16040 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Brandon Breitenstein <brandon.breitenstein@intel.com>
2016-07-28vboot: Separate vboot from chromeosFurquan Shaikh
VBOOT_VERIFY_FIRMWARE should be independent of CHROMEOS. This allows use of verified boot library without having to stick to CHROMEOS. BUG=chrome-os-partner:55639 Change-Id: Ia2c328712caedd230ab295b8a613e3c1ed1532d9 Signed-off-by: Furquan Shaikh <furquan@google.com> Reviewed-on: https://review.coreboot.org/15867 Tested-by: build bot (Jenkins) Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
2016-07-15soc/intel/apollolake: provide poweroff() implementationAaron Durbin
Implement poweroff() by putting the chipset into ACPI S5 state. BUG=chrome-os-partner:54977 Change-Id: I4ee269f03afd252d4bce909a8cc7c64d6270b16e Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/15686 Reviewed-by: Furquan Shaikh <furquan@google.com> Tested-by: build bot (Jenkins)
2016-07-15soc/intel/apollolake: use common Intel ACPI hardware definitionsAaron Durbin
Transition to using the common Intel ACPI hardware definitions generic ACPI definitions. BUG=chrome-os-partner:54977 Change-Id: Icaca9367b526999f0475b21dd968724baa32e3f6 Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/15667 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com> Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
2016-06-30soc/apollolake: Expose a function to read pmc barShaunak Saha
This patch exposes a function to read pmc bar. PMC bar is read in function read_pmc_mmio_bar which is defined static in file pmutil.c. This patch exposes that functionality to call it from other files. BUG=chrome-os-partner:53438 TEST= Read the PMC bar value properly from outside pmutil file. Change-Id: I26ee13e6ab95d3a8991c7f8ea4b3856ceb015d10 Signed-off-by: Shaunak Saha <shaunak.saha@intel.com> Reviewed-on: https://review.coreboot.org/15460 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-06-27intel/apollolake: Set sleep type to S5 on vboot reboot requestFurquan Shaikh
Add support for vboot_platform_prepare_reboot which is called whenever vboot requests reboot of the platform. SLP_TYPE needs to be set to S5 in such conditions since the platform would no longer be in a resuming state after reset. Change-Id: I01392bfda90c9274cd52c1004555d250b1d539b7 Signed-off-by: Furquan Shaikh <furquan@google.com> Reviewed-on: https://review.coreboot.org/15340 Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-06-24soc/apollolake: Clear SLP_TYP in PM1_CNTHannah Williams
Change-Id: Id49319ec6b52648b03eaeddfdd1580dd82110fb9 Signed-off-by: Hannah Williams <hannah.williams@intel.com> Reviewed-on: https://review.coreboot.org/15336 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-06-24soc/intel/apollolake: Add utility functions for global resetAndrey Petrov
Apollolake defines Global Reset where Host, TXE and PMC are reset. During boot we may need to trigger a global reset as part of platform initialization (or for error handling). Add functions to trigger global reset, enable/disable it and lock global reset bit. BUG=chrome-os-partner:54149 BRANCH=none TEST=none Change-Id: I84296cd1560a0740f33ef6b488f15f99d397998d Signed-off-by: Andrey Petrov <andrey.petrov@intel.com> Reviewed-on: https://review.coreboot.org/15198 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-06-12soc/intel/apollolake: provide fake PM1 SMI status bitAaron Durbin
It appears that PM1 is not wired up to the SMI status register, but it does definitely cause SMIs to trigger. Therefore, provide a fake PM1 status bit by checking the power button status when SMI status is indicating no status as well as the PM1 control indicating that SCI mode is not enabled. BUG=chrome-os-partner:54262 TEST=Smashed power button on reef to cause SMI in firmware. No longer loops infinitely with constant SMIs firing. Change-Id: I9aa1b5f79b651cbc19a2d3353d9ef65429386889 Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/15155 Tested-by: build bot (Jenkins) Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2016-06-12soc/intel/apollolake: add SMI status bit definitons and use themAaron Durbin
Provide the bit definitions for the SMI status register. Also, utilize them which means deleting some of the handlers that can't exist because there are no status bits. BUG=chrome-os-partner:54262 Change-Id: I389c7cb3cad01ba0eca52a337ffee352a2010bfa Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/15154 Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com> Tested-by: build bot (Jenkins)
2016-05-27soc/intel/apollolake: add support for verstageAaron Durbin
There previously was no support for building verstage on apollolake. Add that suport by linking in the appropriate modules as well as providing vboot_platform_is_resuming(). The link address for verstage is the same as FSP-M because they would never be in CAR along side each other. Additionally, program the ACPI I/O BAR and enable decoding so sleep state can be determined for early firmware verification. Change-Id: I1a0baab342ac55fd82dbed476abe0063787e3491 Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/14972 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2016-05-09soc/apollolake/pmutil: Get PMC base address dynamicallyAlexandru Gagniuc
Instead of using a hardcoded address for PMC device BAR0, read it dynamically. This allows the allocator to move the BAR without needing a fixed resource. Note that we cannot do the same for the ACPI BAR (index 0x20), as it cannot be read back. Change-Id: If43e1ccb693ffb17b78bdd76140a0849493a0010 Signed-off-by: Alexandru Gagniuc <alexandrux.gagniuc@intel.com> Reviewed-on: https://review.coreboot.org/14633 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-04-15soc/apollolake: Add helper functions to access Power Management RegistersHannah Williams
Change-Id: I928efea33030e03cbbaead6812c617d20446f7c9 Signed-off-by: Hannah Williams <hannah.williams@intel.com> Reviewed-on: https://review.coreboot.org/14289 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins)