It is for crossgcc.
Change-Id: Ia1d676adfea340b6b80858215459491c9338d614
Signed-off-by: Andrew Wu <arw@dmp.com.tw>
Reviewed-on: http://review.coreboot.org/3955
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
It is expected this will always be a casted get_top_of_ram() call
on x86, no reason to do that under chipset.
Change-Id: I3a49abe13ca44bf4ca1e26d1b3baf954bc5a29b7
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3972
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
As boards without EARLY_CBMEM_INIT do not initialize CBMEM in romstage,
and have no CAR migration, these features are available for ramstage only.
Change-Id: Ic3f77ccdedd4e71ba693619c02c9b98b328a0882
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3970
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
Dummy get_top_of_ram() is removed from romstage to fail already at
build-time for cases where cbmem_initialize() would not complete.
The mechanisms behind CAR_GLOBAL migration only work correctly when
romstage can succesfully make the cbmem_initialize() call.
Change-Id: I359820fb196ef187b9aa2e8a3e8f658a0550f237
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3969
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
This is needed to apply a rule that get_top_of_ram() in romstage is
required to select HAVE_ACPI_RESUME, otherwise chipset/board has no
means to backup low memory to CBMEM on s3 resume.
Only board affected is asus/p2b.
Change-Id: Ia5cbf4e5e40af25f52a19de584d8bc5370487154
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3971
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
Timestamps from cbfs_and_run, TS_START_COPYRAM and TS_END_COPYRAM,
were lost with commit b766b1c7.
Reason is variable ts_table was referencing CAR storage after CAR
is torn doesn. Add use of car_get_var() / car_set_var() so the
references go to migrated storage in CBMEM.
Change-Id: I5a942ad7fd59a04e3a5255f4a3636d37dcfc1591
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3967
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
Ubuntu's HDMI audio has noise and echo. Disable NoSnoopEnable can
resolve this issue.
I have tested on Ubuntu 13.04 with AMD Catalyst 13.4 Proprietary
Linux Display Driver[1].
[1]. http://support.amd.com/us/gpudownload/linux/Pages/radeon_linux.aspx
Change-Id: I5d2dddb1b7469d56cd64e3c1e0f4c6c6f095b4ab
Signed-off-by: WANG Siyuan <SiYuan.Wang@amd.com>
Signed-off-by: WANG Siyuan <wangsiyuanbuaa@gmail.com>
Reviewed-on: http://review.coreboot.org/3934
Tested-by: build bot (Jenkins)
Reviewed-by: Dave Frodin <dave.frodin@se-eng.com>
Ubuntu's HDMI audio has noise and echo. Disable NoSnoopEnable can
resolve this issue.
I have tested on Ubuntu 13.04 with latest graphic driver.
Change-Id: I09c19b8925eedee03cfb1d8c0831a84e8aeeba4f
Signed-off-by: WANG Siyuan <SiYuan.Wang@amd.com>
Signed-off-by: WANG Siyuan <wangsiyuanbuaa@gmail.com>
Reviewed-on: http://review.coreboot.org/3937
Tested-by: build bot (Jenkins)
Reviewed-by: Dave Frodin <dave.frodin@se-eng.com>
Windows 7 cannot find HDMI audio device because of acpi setting.
I have tested on Windows 7. I can play music.
Change-Id: I90ade7e7be79f65783922333c2cbb2d3cc6557ea
Signed-off-by: WANG Siyuan <SiYuan.Wang@amd.com>
Signed-off-by: WANG Siyuan <wangsiyuanbuaa@gmail.com>
Reviewed-on: http://review.coreboot.org/3933
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The platform initialization (PI) code v1.0.0.7 for Kabini has some
enhancements like ECC DIMM support, new CPU microcode rev 0700010B, FCH
bug fix (RTC) and so on.
Use the name Kabini instead of Kerala everywhere.
Note, the former PI code was indeed version v1.0.0.0 instead of v0.0.1.0
as used in `AGESA_VERSION_STRING`.
Change-Id: I186de1aef222cd35ea69efa93967a3ffb8da7248
Signed-off-by: WANG Siyuan <SiYuan.Wang@amd.com>
Signed-off-by: WANG Siyuan <wangsiyuanbuaa@gmail.com>
Reviewed-on: http://review.coreboot.org/3935
Tested-by: build bot (Jenkins)
Reviewed-by: Bruce Griffith <Bruce.Griffith@se-eng.com>
This reverts commit de1fe7f655.
While things appeared to work, there were actually invalid references
to CAR storage after CAR was torn down on boards without
EARLY_CBMEM_INIT. It was discussed use of CAR_GLOBAL should be
restricted to boards that handle CAR migration properly.
Change-Id: I9969d2ea79c334a7f95a0dbb7c78065720e6ccae
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3968
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
Linux unhelpfully "fixes" the value in PCI_BASE_ADDRESS_1 when it is
0xfec00000 (that is, outside the range of bus 0 address space). This
causes IOAPIC interrupts to fail to work under Linux. This issue was
originally unnoticed by me when testing as sanity checking such as
this is not done by NetBSD.
Hiding the IOAPIC BAR is done by the OEM BIOS on the ck804 boards I've
checked.
Change-Id: I736db163750f709d68c988fac075597a50b29ab7
Signed-off-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-on: http://review.coreboot.org/3963
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Change-Id: Ibdd438455a545aa9266b0fd893d5ff27124ab22c
Signed-off-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-on: http://review.coreboot.org/3961
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Change-Id: I67192c8ae99e396ea4b17e03c658f31dbb5c1800
Signed-off-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-on: http://review.coreboot.org/3960
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Call pc_keyboard_init function in southbridge. It makes PS/2
keyboard work in coreinfo payload.
Change-Id: Idb79f87b09eeeade94e966fb8769dec7578e2cf5
Signed-off-by: Andrew Wu <arw@dmp.com.tw>
Reviewed-on: http://review.coreboot.org/3957
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Change-Id: I7fc7819c329c058472031e82237be5c170b277f4
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: http://review.coreboot.org/3965
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
A late for loop may reference over the current array allocation
and corrupt an unrelated global variable. As a quick fix bumb the
size of the array allocation uniformly to 6.
We missed these boards for commit 9c7d73ca because the arrays
had been renamed.
Change-Id: Iff2f2a0090d9302576bc72195d2a3f6fa37ce29a
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3954
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-by: Bruce Griffith <Bruce.Griffith@se-eng.com>
When building coreboot with the Clang static analyzer scan-build,
it reports »Value stored to 'type_index' is never read«. Indeed,
in `memranges_each_entry()` `type_index` is assigned a value
before being read. So remove that line.
Change-Id: I6da2fb8be7157bb98c57281babd4a08ca0d9f7a7
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: http://review.coreboot.org/3953
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
Qemu has the fw_cfg interface at 0x510, which conflicts with
power management base address in coreboot. Move the pmbase to a
non-conflicting address. No need to worry about speedstep, it
is not supported by qemu and isn't enabled in the qemu config.
Change-Id: I3e87d8301988028ca0ea7d96c08b4e26ac15a7c2
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-on: http://review.coreboot.org/3938
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
On Intel's Panther Point the xHCI ports are shared with an EHCI
controller. Our xHCI driver switches them to xHCI, naturally. But
we forgot to switch them back on shutdown, which left them
unusable by a non-xHCI aware operating system.
Change-Id: I70ef08655a603b42ee939935d50cf77ea97878a3
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: http://review.coreboot.org/3791
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Fix "set but not used" variable warning with gcc 4.7.3
Change-Id: Ia27291ecb4f993c4ba6f29b134167dc23a449bf5
Signed-off-by: Allen Martin <amartin@nvidia.com>
Reviewed-on: http://review.coreboot.org/3949
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Windows 7 cannot find HDMI audio device because of acpi setting.
I have tested on Windows 7. I can play music.
Change-Id: I53177ce00b676824a903a3397d69338e8c1a38af
Signed-off-by: WANG Siyuan <SiYuan.Wang@amd.com>
Signed-off-by: WANG Siyuan <wangsiyuanbuaa@gmail.com>
Reviewed-on: http://review.coreboot.org/3936
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
`util/lint/lint-stable-002-build-dir-handling` always overwrites your
current `config.h` and `auto.conf` when the pre-commit hook is run. It
can be very confusing when your configuration is suddenly broken. So fix
it by not using the default build directory.
Change-Id: If2bbc97ac2f12a8203a3769d813386a023f93dd6
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/3593
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Chipsets sb700 and sb800/hudson have more than one USB EHCI controller,
implement the selection logic using already existing Kconfig option.
Change-Id: I9e0df1669d73863c95c36a3a7fee40d58f6f097e
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3928
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
keyboard_init attempts to read the existing mode register, set the
'XLATE' bit, and write it back. The implementation is buggy because the
keyboard may be active at the time we read the mode, and we can
misinterpret scancode data as the reply to our command. It leads to
problems where the KB gets disabled in firmware.
In fact, setting the 'XLATE' bit is completely unnecessary, even if we
desire QEMU keyboard support. We already set this bit when we initialize
the keyboard in pc_keyboard_init. Basically, this code does nothing
(or worse), so just remove it.
Change-Id: Iab23f03fa8bced74842c33a7d263de5f449bb983
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: http://review.coreboot.org/3883
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Northbridge code includes these headers, so they all need to
have the same name to allow different combinations of northbridge
and southbridge. This changes the sb900 names to match sb700 &
sb800, and points agesa/family12 and amd/torpedo to the new file
names.
Change-Id: I7a654ce9ae591a636a56177f64fb8cb953b4b04f
Signed-off-by: Corey Osgood <corey.osgood@gmail.com>
Reviewed-on: http://review.coreboot.org/3825
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
If romstage does not make cbmem_initialize() call, linker should
optimize the code for CAR migration away.
This simplifies design of CBMEM console by a considerable amount.
As console buffer is now migrated within cbmem_initialize() call there
is no longer need for cbmemc_reinit() call made at end of romstage.
Change-Id: I8675ecaafb641fa02675e9ba3f374caa8e240f1d
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3916
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
Assume EARLY_CBMEM_INIT=y everywhere and remove option from Kconfig.
If romstage does not make the cbmem_initialize() call, features like
COLLECT_TIMESTAMPS and early CBMEM_CONSOLE will execute during
romstage, but that data will get lost as no CAR migration is
executed.
Change-Id: I5615645ed0f5fd78fbc372cf5c3da71a3134dd85
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3917
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
These features depend on CAR_GLOBAL region, which is not available
when romstage is built with ROMCC. Exclude these from romstage, keep
them available for ramstage.
A follow-up patch will fix the dependencies and allows enabling these
features in menuconfig.
Change-Id: I9de5ad41ea733655a3fbdc734646f818e39cc471
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3919
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@google.com>
It is not compulsory to have CBMEM console initialised in romstage,
so try add the CBMEM table entry again in ramstage, if not found.
Change-Id: I96ab502df7f05d6bf1d6e6fa84d395ef6306b525
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3915
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@google.com>
We only have one table to collect timestamps into.
Change-Id: I80180fe9a05226f0351c3e66eacaf2d0cb82c924
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3912
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-by: Aaron Durbin <adurbin@google.com>
This retrieves back the value stored with store_initial_timestamp()
in the bootblock for southbridge.
Change-Id: I377c823706c33ed65af023d20d2e4323edd31199
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3908
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-by: Aaron Durbin <adurbin@google.com>
gcc 4.8.x has issues with using ebp, which broke some builds,
so downgrade. The problem also manifested elsewhere, so it's
not necessarily our fault.
While at it, gcc complained about "armv7a" where it seems to
expect "armv7-a".
Change-Id: I6f0c35f49709cb41022475bb47116c12ab1c7ee3
Signed-off-by: Patrick Georgi <patrick.georgi@secunet.com>
Reviewed-on: http://review.coreboot.org/3930
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Some development kits with USB 2.0 HS OTG have an USB hub instead
of being directly connected to the USB host/device controller.
Send the necessary initialisation sequence, using HUB CLASS requests
of PORT_POWER and PORT_RESET to enable a pre-selected port number
where a device supporting debug descriptor is located.
This also adds the Kconfig option for BeagleBone.
Change-Id: I7a5d0ba0962a9ca06bf3196232ed4a03bdfb2b06
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3925
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Assign the lanes correctly to the physical slots
on the motherboard in `PlatformGnbPcie.c`.
• UMI is connected to SB via 4x PCIe bridge 8.
• The blue x16 slot is not shared with DDI and is routed
through PCIe bridge 2.
• The black x8 slot is in fact a x4 slot and uses all 4 GPPs
from the CPU.
• Assume that DDI is on out-of-PCIe-band lanes.
Change-Id: I44c4c83e6a8e31d6150a602a0993972ac63105bd
Signed-off-by: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: http://review.coreboot.org/3194
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martin.roth@se-eng.com>
Reviewed-by: David Hubbard <david.c.hubbard+coreboot@gmail.com>
Without this coreboot may (depends on the amount of memory) place the
pci bars below 0xb0000000, then the linux kernel goes move them around
so they are inside the window declared in the acpi tables.
This breaks vesafb as the vga framebuffer gets moved after vgabios
initialization. It's also not exactly nice to expect the OS fix our
mess ;)
Change-Id: If6b50ea863958eea71b567ccb7a06c6a28076111
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-on: http://review.coreboot.org/3927
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
This simplifies debugging and also fixes an issue when build directories
are kept between buildgcc runs for different architectures.
Change-Id: I5badccd3368e3014680da3eedb607119fff8fa7f
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/3929
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Clean whitespace errors that have gotten past lint-stable-003-whitespace
and gerrit review.
Change-Id: Id76fc68e9d32d1b2b672d519b75cdc80cc4f1ad9
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3920
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Bruce Griffith <Bruce.Griffith@se-eng.com>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The rule "-perm +111 -prune" matched any searchable directory
and did not recursively find files in them. The use of "+mode"
for -perm is deprecated.
Change-Id: I1b43f89ee9ab37928e56104b0f07241ff84b84c0
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3921
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The console has already been initialized in the generic bootblock code, and
reinitializing it causes the same banner line to be printed twice and lots of
artifacts in the actual output. This same change had been made to the other
ARM boards but not for beaglebone.
Change-Id: I72e3be1326b1a52b7ec438a44e4fd5f87e4ec717
Signed-off-by: Gabe Black <gabeblack@chromium.org>
Reviewed-on: http://review.coreboot.org/3924
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
These dependencies came indirectly through kconfig.h which was included
automatically with a -include option which was either part of INCLUDES or
specified directly. With this change, I'm able to build for beaglebone with
make -j 48.
Change-Id: Ib57d0c6a755b747165b235c2328c3c30bd6dd67d
Signed-off-by: Gabe Black <gabeblack@chromium.org>
Reviewed-on: http://review.coreboot.org/3922
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The way those variables work has changed twice since this file was last
changed, and console output was no longer working. Now that they're up to
date there's serial output from beaglebone again.
Change-Id: I5167fd8c0a8c33438d7f056fdf5951bd054010ed
Signed-off-by: Gabe Black <gabeblack@chromium.org>
Reviewed-on: http://review.coreboot.org/3923
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Old name was too much x86.
All external references have been removed.
Change-Id: I982b9abfcee57a7ea421c245dadb84342949efae
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3906
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
The parameters can be dropped as initialisation always happens for
the region resolved with cbmem_locate_table().
This is no longer referenced externally, make it static.
Change-Id: Ia40350a5232dcbf30aca7b5998e7995114c44551
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/3565
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@google.com>
Andrew Wu