oak/elm/gru/kevin-release: chromeos-chrome-57.0.2956.0_rc-r1 fails to build: CAN_USE_ARMV7_INSTRUCTIONS should match CAN_USE_VFP3_INSTRUCTIONS |
||||||||||
Issue descriptionoak/elm/gru/kevin-release builders fail to build Chrome since the bump to 57.0.2956.0. These 2 families of boards have in common to use aarch32 (aka armv8) compiler flags, which are not tested by the PFQ. I suggest pinning back Chrome to 57.0.2955.0. https://uberchromegw.corp.google.com/i/chromeos/builders/elm-release/builds/683/steps/BuildPackages%20%5Bafdo_use%5D/logs/stdio https://uberchromegw.corp.google.com/i/chromeos/builders/elm-release/builds/684/steps/BuildPackages%20%5Bafdo_use%5D/logs/stdio https://uberchromegw.corp.google.com/i/chromeos/builders/kevin-release/builds/691/steps/BuildPackages%20%5Bafdo_use%5D/logs/stdio chromeos-chrome-57.0.2956.0_rc-r1: FAILED: v8_snapshot/obj/v8/v8_base/assembler-arm.o chromeos-chrome-57.0.2956.0_rc-r1: x86_64-pc-linux-gnu-g++ -MMD -MF v8_snapshot/obj/v8/v8_base/assembler-arm.o.d -DV8_DEPRECATION_WARNINGS -DUSE_UDEV -DUI_COMPOSITOR_IMAGE_TRANSPORT -DUSE_AURA=1 -DUSE_NSS_CERTS=1 -DUSE_OZONE=1 -DUSE_PROPRIETARY_CODECS -DFULL_SAFE_BROWSING -DSAFE_BROWSING_CSD -DSAFE_BROWSING_DB_LOCAL -DOFFICIAL_BUILD -DGOOGLE_CHROME_BUILD -DENABLE_MEDIA_ROUTER=1 -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -DNO_UNWIND_TABLES -DNDEBUG -DNVALGRIND -DDYNAMIC_ANNOTATIONS_ENABLED=0 -DV8_I18N_SUPPORT -DENABLE_HANDLE_ZAPPING -DV8_USE_EXTERNAL_STARTUP_DATA -DV8_TARGET_ARCH_ARM -DCAN_USE_ARMV7_INSTRUCTIONS -DUSE_EABI_HARDFLOAT=1 -DU_USING_ICU_NAMESPACE=0 -DU_ENABLE_DYLOAD=0 -DU_STATIC_IMPLEMENTATION -DICU_UTIL_DATA_IMPL=ICU_UTIL_DATA_FILE -I../../../../../../../home/chrome-bot/chrome_root/src -Iv8_snapshot/gen -I../../../../../../../home/chrome-bot/chrome_root/src/v8 -I../../../../../../../home/chrome-bot/chrome_root/src/v8/include -I../../../../../../../home/chrome-bot/chrome_root/src/third_party/icu/source/common -I../../../../../../../home/chrome-bot/chrome_root/src/third_party/icu/source/i18n -fno-strict-aliasing --param=ssp-buffer-size=4 -fstack-protector -fno-unwind-tables -fno-asynchronous-unwind-tables -fPIC -pipe -m32 -msse2 -mfpmath=sse -mmmx -pthread -g2 -gsplit-dwarf -fvisibility=hidden -Wno-unused-local-typedefs -Wno-maybe-uninitialized -Wno-missing-field-initializers -Wno-unused-parameter -O3 -fno-ident -fdata-sections -ffunction-sections -fno-threadsafe-statics -fvisibility-inlines-hidden -std=gnu++11 -fno-rtti -fno-exceptions -Wno-deprecated -Wno-narrowing -c ../../../../../../../home/chrome-bot/chrome_root/src/v8/src/arm/assembler-arm.cc -o v8_snapshot/obj/v8/v8_base/assembler-arm.o chromeos-chrome-57.0.2956.0_rc-r1: ../../../../../../../home/chrome-bot/chrome_root/src/v8/src/arm/assembler-arm.cc:173:2: error: #error "CAN_USE_ARMV7_INSTRUCTIONS should match CAN_USE_VFP3_INSTRUCTIONS" chromeos-chrome-57.0.2956.0_rc-r1: #error "CAN_USE_ARMV7_INSTRUCTIONS should match CAN_USE_VFP3_INSTRUCTIONS" chromeos-chrome-57.0.2956.0_rc-r1: ^ chromeos-chrome-57.0.2956.0_rc-r1: chromeos-chrome-57.0.2956.0_rc-r1: ninja: build stopped: subcommand failed. chromeos-chrome-57.0.2956.0_rc-r1: * ERROR: chromeos-base/chromeos-chrome-57.0.2956.0_rc-r1::chromiumos failed (compile phase): chromeos-chrome-57.0.2956.0_rc-r1: * (no error message)
,
Dec 20 2016
,
Dec 20 2016
If I had to guess, I'd blame: chromium:670541
,
Dec 20 2016
hashimoto@ was looking into this
,
Dec 20 2016
hashimoto: ETA? I'm happy to do the revert on the suspect CL if it helps. It is certanly in the blame list: https://crosland.corp.google.com/log/9103.0.0..9104.0.0
,
Dec 20 2016
Given that it's 2am in Tokyo where hashimoto@ is located and he doesn't appear to be responding to IMs (I wouldn't if it were 2am where I was, either), I'm proposing we revert. Revert posted at: https://chromium-review.googlesource.com/#/c/422416/ Trybots kicked off: cbuildbot --remote -g 422416 kevin-release https://uberchromegw.corp.google.com/i/chromiumos.tryserver/builders/release/builds/7898 cbuildbot --remote -g 422416 elm-release https://uberchromegw.corp.google.com/i/chromiumos.tryserver/builders/release/builds/7899 If someone has already confirmed that the revert fixes the problems then we can chump it in. Otherwise I'm inclined to see evidence. I'm not setup myself to build Chrome efficiently and I'd be inclined to wait for the trybots, but if someone is setup to build Chrome efficiently or thinks it's worth it to setup, please let me know.
,
Dec 20 2016
Ack. Comment 7 was intended for bug #676013 . Deleting it and moving to comment 5 there.
,
Dec 20 2016
OK, well I can't test terribly well given the InitSDK issue ( bug #676013 ), but the revert seems like the right move. Chumping it.
,
Dec 20 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/b0f4c9fd293889da2bb4a3dd55743a9654fecbec commit b0f4c9fd293889da2bb4a3dd55743a9654fecbec Author: Douglas Anderson <dianders@chromium.org> Date: Tue Dec 20 17:15:55 2016 Revert "chromeos-chrome: Explicitly set arm_fpu and arm_tune" This reverts commit 4341948bf829e486fdc319733b563ff37a00d431. This appears to be causing problems. See: chromium:675978 BUG= chromium:675978 , chromium:670541 TEST=Compile Chrome Change-Id: Idce4779b8f3061c71ea6bd1ddfce1b92122b2664 Reviewed-on: https://chromium-review.googlesource.com/422416 Tested-by: Douglas Anderson <dianders@chromium.org> Reviewed-by: Mike Frysinger <vapier@chromium.org> [modify] https://crrev.com/b0f4c9fd293889da2bb4a3dd55743a9654fecbec/chromeos-base/chromeos-chrome/chromeos-chrome-9999.ebuild
,
Dec 20 2016
I will propose a "Fixed" tag here, but hard to tell for now with the tree closed.
,
Dec 21 2016
Thank you for handling the revert. Seems V8's GN logic doesn't accept -mfpu flag values other than vfpv3-d16, vfpv3, and neon, while kevin uses crypto-neon-fp-armv8. I'll address this before relanding the change after coming back from vacation.
,
Dec 21 2016
Issue 676319 has been merged into this issue.
,
Dec 21 2016
Argh. Change didn't really take effect in recent elm and kevin builds because we didn't have a chrome uprev, so the -9999 build changes didn't have any effects on the real builders. For instance, elm 687 failed because it tried to build: [ebuild N ] chromeos-base/chromeos-chrome-57.0.2957.0_rc-r1 ...and that _didn't_ have the changes needed. It looks like the uprev has happened now, so future builds should be OK. Leaving this as proposed "Fixed".
,
Mar 4 2017
,
Apr 17 2017
,
May 30 2017
,
Aug 1 2017
,
Oct 14 2017
|
||||||||||
►
Sign in to add a comment |
||||||||||
Comment 1 by diand...@chromium.org
, Dec 20 2016