rockot@: Can you please have a look or help route?

There are four flakes over the past two days.

The test output is:
===============================================================================================================


[ RUN      ] SharedBufferTest.PassHandleBetweenCousins
[26913:2021895383:0620/045956.807445:19453733:WARNING:test_suite.cc(240)] Test launcher output path /tmp/.org.chromium.Chromium.bGHNlj/test_results.xml exists. Not adding test launcher result printer.
[26982:66235161:0620/045956.847834:19494123:WARNING:test_suite.cc(240)] Test launcher output path /tmp/.org.chromium.Chromium.bGHNlj/test_results.xml exists. Not adding test launcher result printer.
[27253:640808943:0620/045956.988934:19635222:WARNING:test_suite.cc(240)] Test launcher output path /tmp/.org.chromium.Chromium.bGHNlj/test_results.xml exists. Not adding test launcher result printer.
[27307:1498481971:0620/045956.994367:19640655:WARNING:test_suite.cc(240)] Test launcher output path /tmp/.org.chromium.Chromium.bGHNlj/test_results.xml exists. Not adding test launcher result printer.
[25299:175598433:0620/045957.060869:19707158:FATAL:platform_handle.cc(81)] Check failed: (0) == result (0 vs. -11)CloseIfNecessary(zx_handle_close): ZX_ERR_BAD_HANDLE
bt#00: pc 0x38e78a81ffa1 (app:/pkg/bin/app,0x563fa1)
bt#01: pc 0x38e78a7c84f0 (app:/pkg/bin/app,0x50c4f0)
bt#02: pc 0x38e78a6baf2c (app:/pkg/bin/app,0x3fef2c)
bt#03: pc 0x38e78a6e412c (app:/pkg/bin/app,0x42812c)
bt#04: pc 0x38e78a6cd3cb (app:/pkg/bin/app,0x4113cb)
bt#05: pc 0x38e78a6cddf2 (app:/pkg/bin/app,0x411df2)
bt#06: pc 0x38e78a6d6734 (app:/pkg/bin/app,0x41a734)
bt#07: pc 0x38e78a6ce4cf (app:/pkg/bin/app,0x4124cf)
bt#08: pc 0x38e78a6bc663 (app:/pkg/bin/app,0x400663)
bt#09: pc 0x38e78a6e5015 (app:/pkg/bin/app,0x429015)
bt#10: pc 0x38e78a8274e6 (app:/pkg/bin/app,0x56b4e6)
bt#11: pc 0x38e78a828379 (app:/pkg/bin/app,0x56c379)
bt#12: pc 0x38e78a827990 (app:/pkg/bin/app,0x56b990)
bt#13: pc 0x38e78a7ced69 (app:/pkg/bin/app,0x512d69)
bt#14: pc 0x38e78a7e7865 (app:/pkg/bin/app,0x52b865)
bt#15: pc 0x38e78a805901 (app:/pkg/bin/app,0x549901)
bt#16: pc 0x38e78a805bb8 (app:/pkg/bin/app,0x549bb8)
bt#17: pc 0x38e78a82cff2 (app:/pkg/bin/app,0x570ff2)
bt#18: pc 0x7c2b99dd59c6 (libasync-default.so,0x2b8b83519c6)
bt#19: pc 0x7c2b99e49de9 (libasync-default.so,0x2b8b83c5de9)
bt#20: end
[26913:2021895383:0620/045957.101519:19747806:FATAL:mojo_test_base.cc(128)] Check failed: WaitForSignals(mp, MOJO_HANDLE_SIGNAL_READABLE) == MOJO_RESULT_OK (9 vs. 0)
bt#00: pc 0x19ad17cb2fa1 (app:/pkg/bin/app,0x563fa1)
bt#01: pc 0x19ad17c5b4f0 (app:/pkg/bin/app,0x50c4f0)
bt#02: pc 0x19ad17d47521 (app:/pkg/bin/app,0x5f8521)
bt#03: pc 0x19ad17d47c7d (app:/pkg/bin/app,0x5f8c7d)
bt#04: pc 0x19ad178f4cf7 (app:/pkg/bin/app,0x1a5cf7)
bt#05: pc 0x19ad17d4a535 (app:/pkg/bin/app,0x5fb535)
bt#06: pc 0x19ad178f2f9d (app:/pkg/bin/app,0x1a3f9d)
bt#07: pc 0x19ad17cd28ba (app:/pkg/bin/app,0x5838ba)
bt#08: pc 0x19ad17cd618b (app:/pkg/bin/app,0x58718b)
bt#09: pc 0x19ad17cd5fcc (app:/pkg/bin/app,0x586fcc)
bt#10: pc 0x19ad17967882 (app:/pkg/bin/app,0x218882)
bt#11: pc 0x646e33997f6e (libfdio.so,0x2dd28ca0f6e)
bt#12: end
[27253:640808943:0620/045957.114293:19760579:FATAL:mojo_test_base.cc(128)] Check failed: WaitForSignals(mp, MOJO_HANDLE_SIGNAL_READABLE) == MOJO_RESULT_OK (9 vs. 0)
bt#00: pc 0x4169b0bc9fa1 (app:/pkg/bin/app,0x563fa1)
bt#01: pc 0x4169b0b724f0 (app:/pkg/bin/app,0x50c4f0)
bt#02: pc 0x4169b0c5e521 (app:/pkg/bin/app,0x5f8521)
bt#03: pc 0x4169b0c5ec7d (app:/pkg/bin/app,0x5f8c7d)
bt#04: pc 0x4169b080b100 (app:/pkg/bin/app,0x1a5100)
bt#05: pc 0x4169b0c61535 (app:/pkg/bin/app,0x5fb535)
bt#06: pc 0x4169b0809e3d (app:/pkg/bin/app,0x1a3e3d)
bt#07: pc 0x4169b0be98ba (app:/pkg/bin/app,0x5838ba)
bt#08: pc 0x4169b0bed18b (app:/pkg/bin/app,0x58718b)
bt#09: pc 0x4169b0becfcc (app:/pkg/bin/app,0x586fcc)
bt#10: pc 0x4169b087e882 (app:/pkg/bin/app,0x218882)
bt#11: pc 0x6bede407bf6e (libasync-default.so,0x8c86be59f6e)
bt#12: end

The same test had also been previously disabled due to flakiness on Android.

Nothing relevant in mojo has changed near the regression. +wez - any reason you could think of for a double-close being introduced in Fuchsia recently? All the check failures here are likely a result of an unexpected process death, with the unexpected process death being a result of a double handle closure assertion failure.

Only affects Fuchsia and only appeared yesterday, assuming prior flake (months ago) is unrelated.

s/All the check failures/The last two check failures/

Thanks Ken! I think this is a symptom ofa change in the handle-passing semantics of the Fuchsia IPC write API - it now consumes all the supplied handles whether or not it succeeds.

We'll need to update the Channel implementation for Fuchsia to release passed handles unconditionally, after write()ing them, similarly to the change we had to make for process-launch handle transfer.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/676b5fae19bd233e4cce1479d631e421fec9e7b1

commit 676b5fae19bd233e4cce1479d631e421fec9e7b1
Author: Wez <wez@chromium.org>
Date: Wed Jun 20 22:34:08 2018

[mojo] Release() sent handles regardless of zx_channel_write() result.

Fuchsia recently changed the zx_channel_write() API to always consume
the supplied handles, by default, regardless of whether the call
succeeds in transferring them to the peer process.

Update the mojo::edk::Channel::Write() implementation to account for the
new semantics.

Bug:  854458 , 754084,  740791 
Change-Id: I27f78a09ff8bf2aaf94ac571678d4c2259b48b10
Reviewed-on: https://chromium-review.googlesource.com/1108970
Reviewed-by: Wez <wez@chromium.org>
Reviewed-by: Ken Rockot <rockot@chromium.org>
Commit-Queue: Wez <wez@chromium.org>
Cr-Commit-Position: refs/heads/master@{#569053}
[modify] https://crrev.com/676b5fae19bd233e4cce1479d631e421fec9e7b1/mojo/edk/system/channel_fuchsia.cc