A few of the brackets and bold text asterisks in the markdown links were
missing their corresponding closing symbol.
Change-Id: I9bfab1d2c83bdc12586bd31b1939bd241df2e932
Signed-off-by: Nicholas Chin <nic.c3.14@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65371
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin L Roth <gaumless@tutanota.com>
Add a new board 'Tentacruel', and enable SDCARD_INIT for it.
BUG=b:234409654
BRANCH=corsola
TEST=none
Signed-off-by: Kane Chen <kane_chen@pegatron.corp-partner.google.com>
Change-Id: Ia10efeead575b4e193a73562275a78839415a706
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65192
Reviewed-by: Yu-Ping Wu <yupingso@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change makes mdss configuration common for both sc7180 & sc7280
to avoid code duplicacy.
Changes in v2:
- Move soc related mdss changes to soc specific disp.c
BUG=b:182963902,b:216687885
TEST=Validated on qualcomm sc7280 development board.
Monitor name: LQ140M1JW49
Change-Id: Ibc43ab6ee5ced08e34625e1485febd2f4717d6a0
Signed-off-by: Vinod Polimera <quic_vpolimer@quicinc.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/64886
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Shelley Chen <shchen@google.com>
In order to improve gpio merge mechanism. Change iteration override
to padbased table override. And the following patch will change fw
config override with ramstage gpio table override.
BUG=b:231690996
TEST=check gpios in pinctrl are the same.
Signed-off-by: Eric Lai <eric_lai@quanta.corp-partner.google.com>
Change-Id: I3d0beabc2c185405cb0af31e5506b6df94e9522c
Reviewed-on: https://review.coreboot.org/c/coreboot/+/64713
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Introduce three functions:
- new_padbased_table:
Returns the gpio pad number based table
- gpio_padbased_override:
Must pass the table with padbased table
- gpio_configure_pads_with_padbased:
Must pass the table with padbased table, will skip configures the
unmapped pins by check pad and DW0 are 0.
Some boards may have complex, SKU-based GPIO programming. This
patch provides for a simpler pattern of controlling overrides of
GPIO programming by providing a table of pad configuration indexed
by pad number. Thus, pad state can be overwritten over multiple
overrides until the final takes place, and then all GPIO
programming is performed at once.
Signed-off-by: Eric Lai <eric_lai@quanta.corp-partner.google.com>
Change-Id: I8b99127b73701b50a7f2e051dee9d12c9da9b741
Reviewed-on: https://review.coreboot.org/c/coreboot/+/64712
Reviewed-by: Kangheui Won <khwon@chromium.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The only callsites in intel/xeon_sp were replaced with calls to
log_resource() and functionality is provided with LOG_RESOURCE()
now.
Change-Id: Ie44694f7a0b119d10f1bef9158fa30e71c312a55
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55478
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Replace xx_resource() calls with calls that take the base
and size arguments as-is, without dividing by KiB (or >> 10).
With replacement of the allocator/constructor function
caller can use log_resource() instead.
Change-Id: I7e4e1e5a779c418f369dd2dab8c811f67ad1399f
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55477
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
These should help to make the reviews as platforms
remove KiB scaling.
Change-Id: I40644f873c0ea993353753c0ef40df4c83233355
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55474
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Function fixed_io_resource() and alias io_resource() were
previously unused. Unlike previously, IORESOURCE_STORED flag
needs to be set by the caller, when necessary.
For fixed resources, fields alignment, granularity and
limit need not be initialised, as the resource cannot
be moved. It is assumed the caller provides valid base
and size parameters.
Change-Id: I8fb4cf2dee4f5193e5652648b63c0ecba7b8bab2
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55458
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Unlike fixed_mem_resource_kb() the arguments are not in KiB.
This allows coccinelle script to assign the base and size
without applying the KiB division or 10 bit right-shift.
Unlike with fixed_mem_resource_kb() the IORESOURCE_STORED flag is
passed in the flags parameter until some inconsistencies in the tree
get resolved.
Change-Id: I2cc9ef94b60d62aaf4374f400b7e05b86e4664d2
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55436
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
This patch changes the serial message type to BIOS_WARNING as sometimes
it may raise a wrong signal when microcode resides inside other part
of the IFWI instead /CBFS.
Signed-off-by: Subrata Banik <subratabanik@google.com>
Change-Id: I714bf74a91c2d783982c5e5ca76a70deed872473
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65316
Reviewed-by: Werner Zeh <werner.zeh@siemens.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This patch drops FSP Debug interface selection as coreboot now decides
the UART inerface to redirect the debug msg.
BUG=none
TEST=Able to see all coreboot and FSP debug log with and without this
patch.
Signed-off-by: Subrata Banik <subratabanik@google.com>
Change-Id: If8c07d7e63c5d445fdb77ac38b99217bf015e15f
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65359
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tarun Tuli <taruntuli@google.com>
Reviewed-by: Lean Sheng Tan <sheng.tan@9elements.com>
Enable caching of memory training data for recovery as well as normal
mode. We had HAS_RECOVERY_MRC_CACHE selected in the sc7280 Kconfig,
but never allocated a RECOVERY_MRC_CACHE in the herobrine fmap so it
never worked. Adding RECOVERY_MRC_CACHE and also removing
RO_DDR_TRAINING, RO_LIMITS_CFG, RW_LIMITS_CFG entries which have been
deprecated.
BUG=b:236995289
BRANCH=None
TEST=run dut-control power_state:rec twice and make sure that
DDR training doesn't run on the second boot.
Change-Id: I39ac7eca4ae94075874324b13c69eef59522e3c5
Signed-off-by: Shelley Chen <shchen@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65370
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Having PTT means mocking secdata, so saving/reading the hash always
succeeds, but there is no data stored/read from/to TPM. The code
comparing MRC hashes did not care if secdata mocking was enabled
and failed during hash comparison with invalid data. This broke the
fastboot even if the MRC cache data was filled and correctly
checksummed. If mocking is enabled simply fallback to checksum
computing to proceed with fastboot.
TEST=Boot MSI PRO Z690-A WIFI DDR4 in fastboot mode with PTT and vboot
enabled.
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Change-Id: Ic0cf04b129fe1c5e94cd8a803bb21aa350c3f8da
Reviewed-on: https://review.coreboot.org/c/coreboot/+/64221
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
With CB:65012, google_chromeec_vbnv_context() is no longer used. Remove
it from the codebase.
BUG=b:178689388
TEST=./util/abuild/abuild -t GOOGLE_STOUT -a -x
Change-Id: I717f600f0f73c3ca932b6a442a9d5b90c35c8f3b
Signed-off-by: Yu-Ping Wu <yupingso@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65326
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
We have clarified the powerdown sequence with Nvidia, and the EEs have
come up with this modified sequence which still meets the requirements
from the hardware design guide.
BUG=b:233959099
TEST=Verified by ODM and EE
Signed-off-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Change-Id: I37715165ab488f994c825fb9ff532ebf8d7f4cb0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65182
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Robert Zieba <robertzieba@google.com>
Reviewed-by: Eric Lai <eric_lai@quanta.corp-partner.google.com>
Reviewed-by: Subrata Banik <subratabanik@google.com>
These changes made my crude pattern matching work with
coccinelle simpler.
Change-Id: I83f3ef38b8663640594b4d726838f7a6f96a58a2
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/64698
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Commit 0cc56a2848 (nb/intel/gm45/dsdt: Fix number of PCI busses) derives
the maximum PCI bus number at runtime. However, IASL complains about the
initial 0 in the resource template, which rendered the PB00 definition
self-contradictory at build time (maximum was lower than minimum +
length - 1).
Let's return to the old default values (min: 0, max: 255, length: 256)
and adapt max and length at runtime. Also fix some surrounding whites-
pace.
NB. The issue wasn't detected before merging commit 0cc56a2848 because
of broken IASL versions that can't count errors.
Change-Id: I359d357f276feda8fe04383080d51dc492c3f2e8
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/64347
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-by: Stefan Ott <coreboot@desire.ch>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Return the correct processor family code for smbios per System
Management BIOS (SMBIOS) Reference Specification DSP0134 revision 3.5.0.
BUG=b:234409052
TEST=Boot chausie to chromeos and verify "dmidecode -t processor"
outputs the correct processor family.
Signed-off-by: Fred Reitberger <reitbergerfred@gmail.com>
Change-Id: I617ce3e23f4b28a197034756d285339595d3b53b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65364
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The patch disables PCH USB2 Phy power gating to prevent possible display
flicker issue for osiris board. Please refer Intel doc#723158 for
more information.
BUG=None
TEST=Verify the build for osiris board
Signed-off-by: David Wu <david_wu@quanta.corp-partner.google.com>
Change-Id: Ia30a7b915df14c91a2526dca3e374436da286b7a
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65324
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Ren Kuo <ren.kuo@quanta.corp-partner.google.com>
The tutorial documents were updated from the wiki very early in the
transition to markdown, and the style has changed over time. This
updates the markdown style to match documents that are being created
now.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I619c04f420042f530335482c30070436f9190865
Reviewed-on: https://review.coreboot.org/c/coreboot/+/64966
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Singer <felixsinger@posteo.net>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
The previous two patches removed all of the soc/Kconfig files, so there
is nothing to include anymore. Get rid of the 'source' command that
includes them.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I95067c4702ef25a8a6db4d480c089f06986ce9b8
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65329
Reviewed-by: Felix Singer <felixsinger@posteo.net>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Most of the src/soc/Kconfig files are only there for AMD and Intel to
load the main SoC Kconfig files before any common files. That can be
done in src/Kconfig instead. Moving the loads to the lower level allows
the removal of all but the Intel soc/Kconfig file, which can be removed
in a follow-on patch.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I5061191fe23e0b7c745e90874bd7b390806bbcfa
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65327
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Singer <felixsinger@posteo.net>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Clean up Makefile.inc by sorting entries and moving common entries to
all-y. In this way it is more clear to know what drivers have been
involved in each stage and the hardware differences between each SoC.
BUG=none
TEST=emerge-corsola coreboot
TEST=emerge-asurada coreboot
TEST=emerge-cherry coreboot
Change-Id: Idfc7de36ebf36650f7c6bd1584ef77e2a540cde9
Signed-off-by: Yidi Lin <yidilin@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65315
Reviewed-by: Yu-Ping Wu <yupingso@google.com>
Reviewed-by: Rex-BC Chen <rex-bc.chen@mediatek.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The `send` and `recv` API functions currently print error messages
if a timeout occurs while polling the EC, but they perform the I/O
transaction regardless. This can put the EC in a bad state or
otherwise invoke undefined hardware behavior.
Most callers ignore the return value currently, but for callers
which do not, we should make sure our behavior is correct.
Signed-off-by: Abel Briggs <abelbriggs1@hotmail.com>
Change-Id: Ifb87dd1ac869807fd08463bd8fef36d0389b325e
Reviewed-on: https://review.coreboot.org/c/coreboot/+/64350
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Reviewed-by: Lean Sheng Tan <sheng.tan@9elements.com>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Adjust sensor trigger point and fan duty according to thermal team
tuning results.
BRANCH=brya
BUG=b:215033682
TEST=Built and tested on taniks board
Signed-off-by: Joey Peng <joey.peng@lcfc.corp-partner.google.com>
Change-Id: I8135684d471fdcfdbbe2f1bc5455902d56bb71de
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65287
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
The patch disables PCH USB2 Phy power gating to prevent possible display
flicker issue for kano board. Please refer Intel doc#723158 for
more information.
BUG=None
TEST=Verify the build for volmar board
Signed-off-by: Ren Kuo <ren.kuo@quanta.corp-partner.google.com>
Change-Id: I4d12f7214a306ded54b4536a27fe0fb7f3c33b8b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65307
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
This CL fixes my previous CL (commit ca741055e)
which introduced a couple of issues found by Coverity (see below).
The Coverity explanation is: "Potentially overflowing expression "size_field * 1048576U" with type "unsigned int" (32 bits, unsigned) is evaluated using 32-bit arithmetic, and then used in a context that expects an expression of type "uint64_t" (64 bits, unsigned)."
*** CID 1490122: Integer handling issues (OVERFLOW_BEFORE_WIDEN)
/src/soc/intel/alderlake/systemagent.c: 305 in get_dpr_size()
*** CID 1490121: Integer handling issues (OVERFLOW_BEFORE_WIDEN)
/src/soc/intel/alderlake/systemagent.c: 254 in get_dsm_size()
BUG=b:149830546
BRANCH=firmware-brya-14505.B
TEST='emerge-brya coreboot chromeos-bootimage' builds correctly.
Tested on an Anahera device which successfully boots to ChromeOS
with kernel version 5.10.109-15688-g857e654d1705.
Change-Id: Ib2d66ad24a5ad67b51036ad376a6938f698134c3
Signed-off-by: Eran Mitrani <mitrani@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65212
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This patch fixes an issue introduced with commit ca741055e
(soc/intel/adl: Add missing claimed memory regions) where PRMRR base
should be read using MSR 0x2a0 and mask from MSR 0x1f5 instead
System Agent PCI configuration space.
With this change, coreboot is able to read PRMRR base when the
PRMRR size > 0.
TEST=Able to read PRMRR base MSR 0x2a0 in proper with this CL.
Signed-off-by: Subrata Banik <subratabanik@google.com>
Change-Id: I3770b1a92dbd2552cf1b9764522c9cac9f29c13c
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65263
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-by: Eran Mitrani <mitrani@google.com>
This patch adds a new mainboard variant called mc_apl7 which is based
on mc_apl4. So far only the names have been adjusted with no further
changes. Following commits will introduce the needed changes for this
mainboard variant.
Test: build mc_apl7, flash to mc_apl4 and compare log level 8 output
Change-Id: Ie9f2f5c29d071de442f8f3e3eaf4b3c2a6b8920f
Signed-off-by: Uwe Poeche <uwe.poeche@siemens.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65283
Reviewed-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The common AMD ACPI GPIO access code is verified to be correct for
Sabrina.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I834076c0a1d1784a272896f2d8f082ebfb86a383
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65317
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
The MCA banks were updated in commit 736d68c0b3 ("soc/amd/sabrina/mca:
update MCA bank names to match the hardware"), but seems that I forgot
to remove the TODO about checking if this is still correct for Sabrina.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ifd86113ccb9eeab704679afab0b985f9febed13b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65314
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
The common microcode update mechanism is verified to be correct and work
on Sabrina.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I5c41674299a829507438beb3ea597a71a0c5a972
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65313
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Current implementation only supports glibc (by looking for __GLIBC__)
and fails to build on systems with alternative libc implementations,
such as musl; sys/io.h is never included, there are no outb/inb
functions which results in undefined references at linking stage.
Using __linux__ instead of __GLIBC__ to test whether the system is Linux
seems to be a more proper way to detect Linux and it also fixes
nvramtool compilation on musl systems.
Tested on Gentoo Linux with musl 1.2.2 (builds and works fine) and Void
Linux with glibc (still builds and works fine).
Change-Id: Idcdc3a033b40f16a6053209813f1e06209ee459a
Signed-off-by: Evgeny Zinoviev <me@ch1p.io>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/48757
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
There is an existing issue for nissa boards where wake up from
RTC wake is not working during suspend_stress_test.
This issue was root caused to the patch which was setting GPE_EN
bits for the GPIOs before locking.
Reference: https://review.coreboot.org/c/coreboot/+/64089
Later issue was found to be with GPP_F14 configuration for nissa
boards. When coreboot skips setting GPE_EN bit for GPP_F14, RTC
wake works properly. Another way to make it work is to skip locking
GPP_F14 GPIO to allow kernel to configure it properly.
This patch skips the locking for GPP_F14 to allow kernel to
configure it later. This fixes the issue of RTC wake not working.
Note: This patch provides workaround for the existing issue and
BUG will be closed once actual reason is identified and proper
fix is available.
BUG=b:234097956
BRANCH=None
TEST=RTC wake works on Nivviks board with the patch.
Change-Id: Ie8091ab8acf2b3f064cb79bdf4700f6b4c1674a5
Signed-off-by: Maulik V Vaghela <maulik.v.vaghela@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/65086
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Eric Lai <eric_lai@quanta.corp-partner.google.com>
Reviewed-by: Reka Norman <rekanorman@chromium.org>
Reviewed-by: Kangheui Won <khwon@chromium.org>