Age | Commit message (Collapse) | Author |
|
Change-Id: I30baee4084399b8078232f31296c4d891a903d47
Signed-off-by: Prashant Gaikwad <pgaikwad@nvidia.com>
Reviewed-on: http://git-master/r/395123
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
Tested-by: Bharat Nihalani <bnihalani@nvidia.com>
|
|
Move the core EMC header to a common location. This allows different
sources to include it without the annoying "../../" type stuff.
Change-Id: I57d2d35e1961eca23d7dd69c2634e8e9af52d4b1
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/376787
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bo Yan <byan@nvidia.com>
|
|
regulator_get() returns either error pointer or valid pointers.
It never returns NULL. Hence it is not a correct to check
the return pointer as IS_ERR_OR_NULL(). It should be only IS_ERR().
Bug 1410234
Change-Id: I8cf668ca0bced49ee4f47ccf9b02d70fd0263cd6
Signed-off-by: Mallikarjun Kasoju <mkasoju@nvidia.com>
Reviewed-on: http://git-master/r/334471
GVS: Gerrit_Virtual_Submit
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
|
|
So that the upstream common clk infrastructure can live side by side.
Signed-off-by: Dan Willemsen <dwillemsen@nvidia.com>
Change-Id: I0fe0ef8cd207d27b707821eed838c75b8ec04025
|
|
Code cleanup for tegra_bb driver:
- use _set_drvdata() and _get_drvdata() functions to avoid
playing with pointers directly.
- shorting hw register definitions to improve readability of
driver
Bug 1341563
Change-Id: I5c827f27f809d48f65651de6224d080e01488f7e
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/251494
(cherry picked from commit 9d7df9f6fd96f2e4737baab9f0b9e32664620120)
Reviewed-on: http://git-master/r/262048
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
If an IRQ is fired as we are going to suspend then we want to abort
and make sure voltage and frequency remain high enough for BB to
operate.
The suspend function must not ack the IRQ as the handler shall be
fired and do it.
Bug 1294872
Change-Id: I20e3872b58814402e5ee5945bb6d91c2651eaacb
Signed-off-by: Hervé Fache <hfache@nvidia.com>
Reviewed-on: http://git-master/r/239765
(cherry picked from commit 022b42d27f9dcd3614cd44d297326c8778c66840)
Reviewed-on: http://git-master/r/256176
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
|
|
Use mem_req_soon signal instead of mem_req signal to
remove EMC clock request. This prevent EMC drops for
short BB hibernate periods.
Bug 1323192
Change-Id: I6dc2e6a4069e5fcbfad0481860b8a9a691e3c0ab
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/249419
(cherry picked from commit a23b66a54bc335e426cebf2441ec8382b5b50e06)
Reviewed-on: http://git-master/r/254496
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
Inform LP1BB entry part about the current BBC emc floor and
the corresponding voltage required. These EMC parameters can
be used by LP1 entry low power routine to set optimal values
before entering in LP1BB hw state. The EMC parameter should
also be used by lp1bb exit routine to set the required EMC
frequency and voltage to minimize suspend-resume latency.
Bug 1270116
Bug 1301005
Change-Id: I1a979ae92fd4d579cd5ecc293f5c40203b440e4d
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/238239
(cherry picked from commit 6b083a4101c7d30dd14b391b2381ed91b938306f)
Reviewed-on: http://git-master/r/254495
Reviewed-by: Mandar Padmawar <mpadmawar@nvidia.com>
Tested-by: Mandar Padmawar <mpadmawar@nvidia.com>
|
|
There is window when BBC driver may not be aware of upcoming
paging event. During LP1BB suspend the EMC rate is used whatever
the system was using before suspend. That EMC rate may not be well
suited for BBC to operate in LP1BB state. Changing the emc clock
rate to set at the maximum floor for BBC. The LP1BB entry part
should reduce it down to the actual frequency floor.
Also for RPC request during LP1BB this new minimum floor gives a
safeguard against the resume time to raise EMC frequency.
Bug 1294872
Change-Id: I3b98cfac174f9dac12307923a1ef54c2c9430bad
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/238238
(cherry picked from commit 36a3bffaf90209f1a6b607a2a3bf581e9731ae49)
Reviewed-on: http://git-master/r/254494
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
If BBC has asserted IPC interrupt then abort the system
suspend. This is to reduce the wakeup latency for BBC
requests.
Also, add function to check for pending BB IPC interrupt.
Bug 1304608
Change-Id: Ie3766ab82c9e7f359cc9866dcf7ee163ab1aabc6
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/237796
(cherry picked from commit 7d65d264cae1810faa3e8193125a0da61da8dfce)
Reviewed-on: http://git-master/r/254492
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
Bug 1304124
Change-Id: Iad4fe7df0c1749c462cd7de155fa5ecde23f5e28
Signed-off-by: Prashant Gaikwad <pgaikwad@nvidia.com>
Reviewed-on: http://git-master/r/242469
(cherry picked from commit 85ca483b895cfe7c91fc4ff49d17d0adb11fcfcf)
Reviewed-on: http://git-master/r/252267
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
|
|
When modem is active device can not enter SoC idle
state. Add stub runtime PM to notify the BB state
to MC clock domain.
Bug 1304124
Change-Id: I37ad5236706d9b765c5ca64b79bc6c3153e98908
Signed-off-by: Prashant Gaikwad <pgaikwad@nvidia.com>
Reviewed-on: http://git-master/r/241947
(cherry picked from commit 65da4905194c6ecdd91e7cd1fa7e25b68ea01f99)
Reviewed-on: http://git-master/r/252266
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
|
|
The statistics from the BBC are placed in the begining of the shared
memory and need to be accessed uncached as they potentially keep
changing.
Bug 1273958
Signed-off-by: Hervé Fache <hfache@nvidia.com>
Change-Id: I0c1e7bca11821418597c861af192d5df3d904d5a
Reviewed-on: http://git-master/r/247247
(cherry picked from commit 01be97eee50b1960ba1ecdca8bc0d7b2d3a70afb)
Reviewed-on: http://git-master/r/250663
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
|
|
When BBC crash is detected via IPC mailbox, prepare EMC
so that the coredump can be collected by FILD.
Restore it back to normal when fild turns off BBC.
Change-Id: I47ff8dee93e714aa2d8a4d4c46e1cfa1cac7cf03
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/230027
Reviewed-by: Steve Lin <stlin@nvidia.com>
Reviewed-by: Neil Patel <neilp@nvidia.com>
|
|
Make EMC floor change effective immediately from RPC call.
This is needed for modem mode of operation change on fly,
instead of having to wait for BBC to go through inactive
to active cycle.
Bug 1284210
Change-Id: Iec6169b4525c609955b10e67d1fab1156b719cfb
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/231702
Reviewed-by: Neil Patel <neilp@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Stephane Dion <sdion@nvidia.com>
Tested-by: Stephane Dion <sdion@nvidia.com>
Reviewed-by: Herve Fache <hfache@nvidia.com>
Reviewed-by: Peter Hsiang <phsiang@nvidia.com>
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
The ISO request from BBC must be cleared when BBC is hibernated.
And the ISO bandwidth should be restored back when BBC wakes up.
Adding APIs in bbc proxy driver to support above need to be
controlled from bb power driver.
Bug 1284833
Change-Id: I8313bb1fd529a2f5a0c6f32bac5a5e51bf78509b
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/226866
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Stephane Dion <sdion@nvidia.com>
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
The mem_req and mem_req_soon interrupt are being missed if they
are set trigger on rising edge. The transition for mem_req_soon HIGH
to mem_req LOW is happening quicker than software can reenable
those signals. Therefore the mem_req LOW edge will be missed resulting
in no interrupt.
Fix this by making those interrupts as level HIGH interrupts.
And after handing IRQ make them as edge again to mark them as clear
otherwise the respective signal remains at their state until
next BBC transition.
Re-enable mem_req and mem_req_soon interrupts at resume.
Bug 1288726
Bug 1287648
Change-Id: Id191f56b8e4c4a2e97843ff4853ac888f4b545d0
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/228610
Reviewed-by: Neil Patel <neilp@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Steve Lin <stlin@nvidia.com>
Reviewed-by: Peter Hsiang <phsiang@nvidia.com>
|
|
The bbc pll voltage should be 1.1V.
Bug 1257234
Change-Id: I0b2e8f6bdf53527a87978d8f82eca1e5a4fb9931
Signed-off-by: Neil Patel <neilp@nvidia.com>
Reviewed-on: http://git-master/r/220711
Reviewed-by: Mrutyunjay Sawant <msawant@nvidia.com>
Tested-by: Mrutyunjay Sawant <msawant@nvidia.com>
|
|
Added new function interface to set EMC floor. BBC can change
EMC floor dynamically by RPC via bbc_proxy driver. This new
floor will be used when modem wakes-up next time.
Also added flag to control DSR setting in EMC.
Bug 1284210
Bug 1249082
Change-Id: I74bda10d3f40f233d4fbfd14edff31dc320ebf2d
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/227266
Reviewed-by: Simone Willett <swillett@nvidia.com>
Tested-by: Simone Willett <swillett@nvidia.com>
|
|
Two EMC clocks are assigned to BBC, namely emc_bw
and emc_fl. The emc_fl should be used to set the
floor frequency and emc_bw is used by ISO manager.
Remove old unused definition from the clocks table.
Bug 1284210
Change-Id: I9926bead2dcaee2b1c5d50d21b7ed074f7fa5a78
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/228609
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Neil Patel <neilp@nvidia.com>
Reviewed-by: Steve Lin <stlin@nvidia.com>
Reviewed-by: Aleksandr Frid <afrid@nvidia.com>
|
|
BBC crash is observed with low memory floor. Increase the
emc floor to be at 600Mhz at least.
Bug 1283678
Change-Id: Ic364d91d8364205e422e63f9e65e110580abb120
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/226462
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Neil Patel <neilp@nvidia.com>
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
Dynamic Self Refresh (DSR) feature in EMC affects BBC's latency.
Override DSR when min EMC frequency is set and enable it
when BBC becomes idle.
Change-Id: I8554910c62bc35a2887e6ed2e49b3240944028dc
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/224045
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
fix coverity issue of dereference before NULL check
Coverity id : 22858
Bug 1046331
Change-Id: I58e52754945e2bbdee3a596543548fe562791bea
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/217268
Reviewed-by: Sachin Nikam <snikam@nvidia.com>
|
|
The BBC signal mem_req_soon is delivered from flow controller
on request. Enable mem_req_soon to notify when BBC is active.
mem_req soon is delivered to PMC, so add handler for PMC_WAKE.
This signal indicates when BBC is becoming inactive.
Adding functionality to trigger emc frequency floor rates in
BB active and inactive states.
Change-Id: Id87da95d0720a6f45cdcd4b584a3be16041a1f0a
Signed-off-by: Michael Hsu <mhsu@nvidia.com>
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/207829
Reviewed-by: Simone Willett <swillett@nvidia.com>
Tested-by: Simone Willett <swillett@nvidia.com>
|
|
Add support to read LP0 to LP0BB transitions via debugfs.
Also read number of LP0 to Active transitions. This will be
utilized by test application.
Bug 1253997
Change-Id: I30d2467d33509189283b92173022d1829a20d0d4
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/209821
Reviewed-by: Neil Patel <neilp@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
Bug 1254387
Change-Id: I4ddbc8e2f1a706bdd1522a6619609fdb29be01a8
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-on: http://git-master/r/210051
Reviewed-by: Stephane Dion <sdion@nvidia.com>
Reviewed-by: Bo Yan <byan@nvidia.com>
GVS: Gerrit_Virtual_Submit
|
|
BBC interrupt BB2AP_INT0 is set as wake source. This
IRQ is used for IPC. Also, make this interrupt level-high
triggered.
Bug 1236920
Change-Id: If0dc5c44ae79a9d7275b6e66ac0e3b3a34874d4e
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/206866
Reviewed-by: Prashant Malani <pmalani@nvidia.com>
Tested-by: Prashant Malani <pmalani@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bo Yan <byan@nvidia.com>
|
|
BBC power rails are turned on before it is released
from reset. By default it should not be turned on
in the probe function.
Changed the regulator names to make it more relevant
to bbc.
Change-Id: I7e7a5b6303b8840931e465ea1f630d7dccb83b80
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/206078
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-by: Neil Patel <neilp@nvidia.com>
Reviewed-by: Steve Lin <stlin@nvidia.com>
|
|
BBC private region is no more mapped into kernel
This saves 64MB of vmalloc space
Bug 1233224
Change-Id: I523ba3d936c28e94a39e23daf30bb1e80cff7951
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-on: http://git-master/r/203817
(cherry picked from commit 9505df4e024ba45c11e3fd9ba373356be13d19d2)
Reviewed-on: http://git-master/r/204574
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Greg Heinrich <gheinrich@nvidia.com>
Reviewed-by: Thomas Cherry <tcherry@nvidia.com>
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
If compatible SHM config version, BBC serial number (PCID) is
populated in tegra_bb sysfs
Bug 1226888
Change-Id: Ic2ee91547b0454ecbefac5aee580f247b1e081d4
Signed-off-by: Alexandre Berdery <aberdery@nvidia.com>
Reviewed-on: http://git-master/r/200765
(cherry picked from commit 35f75aa0191e05d6c8d3b9dac8316e93bb877d8e)
Reviewed-on: http://git-master/r/204573
Reviewed-by: Automatic_Commit_Validation_User
Tested-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-by: Neil Patel <neilp@nvidia.com>
Reviewed-by: Thomas Cherry <tcherry@nvidia.com>
|
|
Now irq request is aborted after 1s if not cleared
Bug 1234170
Change-Id: I10fb354f08813ee38c700f86c174ff931b64315c
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-on: http://git-master/r/199158
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-by: Seshendra Gadagottu <sgadagottu@nvidia.com>
|
|
BBC clock enables on tegra_bb.
remove BBC clock setting from clk init table.
Change-Id: I002cc2e4408a03a1065dcd6dd563628e7c6a096c
Signed-off-by: shawn joo <sjoo@nvidia.com>
Reviewed-on: http://git-master/r/197324
Reviewed-by: Thomas Cherry <tcherry@nvidia.com>
Tested-by: Thomas Cherry <tcherry@nvidia.com>
|
|
add buck4 and ldo8 variable on tegra bb structure
Change-Id: I1be04acbd4f190c4397b8d260ab3b6900d74c7e5
Signed-off-by: shawn joo <sjoo@nvidia.com>
Reviewed-on: http://git-master/r/196641
Reviewed-by: Thomas Cherry <tcherry@nvidia.com>
Tested-by: Thomas Cherry <tcherry@nvidia.com>
|
|
Add regulator enable on proble and disable on remove for buck4 and ldo8.
once the regulator is enabled, en3 will be enabled so that bb_pwr_req can control them.
buck4 is set to 1.1v and ldo8 to 0.9v.
Enable clock for pll_p_bbc and mc_bbc.
Bug 1177376
Bug 1226031
Change-Id: I7cdb4421e8cdebeee398767b7749c7888103b008
Signed-off-by: shawn joo <sjoo@nvidia.com>
Reviewed-on: http://git-master/r/194722
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Pradeep Goudagunta <pgoudagunta@nvidia.com>
Reviewed-by: Bitan Biswas <bbiswas@nvidia.com>
|
|
The sysfs node "state" will show state of BB2AP_MEM_REQ signal.
This essentially shows BBC's current power state as hibernate/active.
Bug 1160000
Change-Id: I40451552cbe4c8c13e8855788e59c15a821fad87
Signed-off-by: Vinayak Pane <vpane@nvidia.com>
Reviewed-on: http://git-master/r/194842
Reviewed-by: Seshendra Gadagottu <sgadagottu@nvidia.com>
Tested-by: Seshendra Gadagottu <sgadagottu@nvidia.com>
|
|
Fixed the issue with unused functions/variables
BUG 1164711
Change-Id: I72c3fa6f3db0d4681a3df05e86f07b5bfd04647b
Signed-off-by: Seshendra Gadagottu <sgadagottu@nvidia.com>
Reviewed-on: http://git-master/r/194998
|
|
Bug 1052795
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Change-Id: I4aea1fec5d081209de04e6d74a3c66a1f12554a3
Reviewed-on: http://git-master/r/193063
Reviewed-by: Simone Willett <swillett@nvidia.com>
Tested-by: Simone Willett <swillett@nvidia.com>
|
|
Change-Id: I29fab3b08a059089976abdf6d8036ad4d6b959b4
Signed-off-by: Kaz Fukuoka <kfukuoka@nvidia.com>
Reviewed-on: http://git-master/r/193981
Reviewed-by: Simone Willett <swillett@nvidia.com>
Tested-by: Simone Willett <swillett@nvidia.com>
|
|
bug 1213479
Change-Id: Ic9318aee5c37d19668c302bd4ee8d12116f86f30
Signed-off-by: Kaz Fukuoka <kfukuoka@nvidia.com>
Reviewed-on: http://git-master/r/189408
Reviewed-by: Simone Willett <swillett@nvidia.com>
Tested-by: Simone Willett <swillett@nvidia.com>
|
|
ipc size was wrongly taken from private size field
Bug 1217721
Change-Id: I18356c25639dbbd354e68e1d56beecdb3d19509d
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-on: http://git-master/r/193001
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bo Yan <byan@nvidia.com>
Reviewed-by: Jean-Marc Guiraudet <jguiraudet@nvidia.com>
|
|
Memory regions are not shown anymore on carveout summary
Bug 1180526
Bug 1050961
Change-Id: I9b6bd9b79e26680e95d702ab784c11431bd78d7f
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-on: http://git-master/r/188229
Reviewed-by: Simone Willett <swillett@nvidia.com>
Tested-by: Simone Willett <swillett@nvidia.com>
|
|
This was causing a kernel crash in bb_boot_test test suite
Bug 1156600
Change-Id: If56a95590cb4d8fd5c0f6f498a426723ebda9e4c
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-on: http://git-master/r/143280
Reviewed-by: Bo Yan <byan@nvidia.com>
|
|
Bug 1050322
Change-Id: I31b6a8b62e9dd75322614f9cab1299b8fedb87a4
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-on: http://git-master/r/135137
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Greg Heinrich <gheinrich@nvidia.com>
Reviewed-by: Bo Yan <byan@nvidia.com>
|
|
Move IPC region to uncached to avoid corruption
between AP and BB.
Bug 1050322
Change-Id: Ifcd55560dd98d53f68396191d82799c63c78ad1b
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-on: http://git-master/r/133570
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Peter Hsiang <phsiang@nvidia.com>
Reviewed-by: Bo Yan <byan@nvidia.com>
|
|
This driver implement integrated icera baseband support
use CONFIG_TEGRA_BASEBAND and optional CONFIG_TEGRA_BASEBAND_SIMU
Bug 957942
Change-Id: Icf995ee2a74e01bd83f7ea352ea12bbd62472fcd
Signed-off-by: Martin Chabot <mchabot@nvidia.com>
Reviewed-on: http://git-master/r/120944
Reviewed-by: Bo Yan <byan@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
|