BuildTime.InThePast fails in base_unittests on fuchsia arm64 |
||
Issue descriptionTime::Now is 0, possibly due to running on non-kvm qemu. [00007.244] 02590.02617> [ RUN ] BuildTime.InThePast [00007.245] 02590.02617> ../../base/build_time_unittest.cc:29: Failure [00007.245] 02590.02617> Expected: (base::GetBuildTime()) < (base::Time::Now()), actual: 2017-10-01 05:00:00.000 UTC vs 1970-01-01 00:00:07.111 UTC [00007.245] 02590.02617> ../../base/build_time_unittest.cc:30: Failure [00007.245] 02590.02617> Expected: (base::GetBuildTime()) < (base::Time::NowFromSystemTime()), actual: 2017-10-01 05:00:00.000 UTC vs 1970-01-01 00:00:07.116 UTC [00007.245] 02590.02617> [ FAILED ] BuildTime.InThePast (10 ms) [00007.246] 02590.02617> [1/1] BuildTime.InThePast (10 ms) [00007.248] 02590.02617> 1 test failed: [00007.248] 02590.02617> BuildTime.InThePast (../../base/build_time_unittest.cc:28)
,
Oct 30 2017
I think it's relatively not-so-bad. TimeTicks is working fine (i.e. it's 7.111s, which is consistent with a boot time of zero) but the initial retrieval is wrong.
,
Oct 30 2017
Re #2: The two are different clocks, though; TimeTicks is ticks-since-boot, while Time is wall-clock time. Are you saying that Time starts at zero but then increases in line w/ TimeTicks, though?
,
Oct 30 2017
Booting zircon arm64 qemu (in Fuchsia tree), I see this: [00000.000] 00000.00000> arm generic timer freq 62500000 Hz [00000.011] 00000.00000> cntpct_per_ns: 00000000.1000000000000000 [00000.011] 00000.00000> ns_per_cntpct: 00000010.0000000000000000 [00000.011] 00000.00000> test_time_conversion_check_result:265: FAIL, off by 72057594037927936 which seems likely related.
,
Oct 30 2017
Hm, that might not be relevant after all. It's trying to see if converting a zx_time of 1ULL<<60 can be converted to a cntpct value, but it's getting truncated due to the size. I think it this case the test is wrong. But it is a little concerning since it probably means we're using a codepath that's not much used.
,
Oct 30 2017
ZX-1317 upstream.
,
Oct 30 2017
The summary of the upstream bug is that there's no RTC integration on real or qemu ARM yet, so zx_time_get(ZX_CLOCK_UTC) is equivalent to zx_time_get(ZX_CLOCK_MONOTONIC) (i.e. Time::Now() == TimeTicks::Now()). Other than the test, I imagine this would affect TLS for cert validation (?), and potentially other protocols that need to coordinate with a server.
,
Oct 30 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/7c466ad1b46ec7b3be6d33cb5280a996c3741d47 commit 7c466ad1b46ec7b3be6d33cb5280a996c3741d47 Author: Wez <wez@chromium.org> Date: Mon Oct 30 21:55:33 2017 Enable base_unittests on Fuchsia/ARM64 bots. This requires filtering out some tests: - Some PartitionAllocTests which fail under ARM64, and were just being skipped under x64. - BuildTime.InThePast which current fails under ARM64. Bug: 772031 , 779645, 779668 Change-Id: Ic764f276d8298bf309c815af6b60eed4ea5d3c4f Reviewed-on: https://chromium-review.googlesource.com/744315 Commit-Queue: John Budorick <jbudorick@chromium.org> Reviewed-by: John Budorick <jbudorick@chromium.org> Reviewed-by: Scott Graham <scottmg@chromium.org> Cr-Commit-Position: refs/heads/master@{#512637} [modify] https://crrev.com/7c466ad1b46ec7b3be6d33cb5280a996c3741d47/testing/buildbot/chromium.fyi.json [modify] https://crrev.com/7c466ad1b46ec7b3be6d33cb5280a996c3741d47/testing/buildbot/filters/fuchsia.base_unittests.filter
,
Oct 31 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/37843401d0f4b5bad0ce6b3bf337f32bcd4d5de4 commit 37843401d0f4b5bad0ce6b3bf337f32bcd4d5de4 Author: Scott Graham <scottmg@chromium.org> Date: Tue Oct 31 07:23:50 2017 Roll Fuchsia SDK to f6dffb2fee82a21900fc3a00261dc5844901ea9e Includes https://fuchsia-review.googlesource.com/c/zircon/+/80476 which should make time work on ARM64. Bug: 779668 Change-Id: I1488b2e595f6f2f0159ed6662f31dc6f8ef8f2d1 TBR: kmarshall@chromium.org Reviewed-on: https://chromium-review.googlesource.com/746027 Commit-Queue: Scott Graham <scottmg@chromium.org> Reviewed-by: Scott Graham <scottmg@chromium.org> Cr-Commit-Position: refs/heads/master@{#512770} [modify] https://crrev.com/37843401d0f4b5bad0ce6b3bf337f32bcd4d5de4/DEPS
,
Oct 31 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/d9d096e359a07d6bb087aa0e3743eb1a21c31fcb commit d9d096e359a07d6bb087aa0e3743eb1a21c31fcb Author: Scott Graham <scottmg@chromium.org> Date: Tue Oct 31 18:24:18 2017 fuchsia: Re-enable BuildTime.InThePast, works on arm64 now Bug: 779668 Change-Id: If23ef86c79f73581b3330f8fa5d020212f74c518 Reviewed-on: https://chromium-review.googlesource.com/747201 Commit-Queue: Scott Graham <scottmg@chromium.org> Commit-Queue: Wez <wez@chromium.org> Reviewed-by: Wez <wez@chromium.org> Cr-Commit-Position: refs/heads/master@{#512879} [modify] https://crrev.com/d9d096e359a07d6bb087aa0e3743eb1a21c31fcb/testing/buildbot/filters/fuchsia.base_unittests.filter
,
Nov 1 2017
|
||
►
Sign in to add a comment |
||
Comment 1 by w...@chromium.org
, Oct 30 2017