New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 817241 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Mar 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Fuchsia
Pri: 2
Type: Bug



Sign in to add a comment

IPCSendFdsTest.DescriptorTest fails under Fuchsia

Project Member Reported by w...@chromium.org, Feb 28 2018

Issue description

Our latest SDK roll broke IPCSendFdsTest.DescriptorTest e.g. in https://ci.chromium.org/buildbot/chromium.fyi/Fuchsia/14419.

The file-descriptors passed in the test are being un-wrapped to type PA_FDIO_REMOTE, and two underlying handles, while the fdio_create_fd() call we use to re-wrap them into a file-descriptor in the receiving process expects a single underlying handle to be passed.
 

Comment 1 by w...@chromium.org, Feb 28 2018

Status: Started (was: Assigned)
Project Member

Comment 2 by bugdroid1@chromium.org, Feb 28 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/acf081fce6f21b2573639df45609a12f3eaaba68

commit acf081fce6f21b2573639df45609a12f3eaaba68
Author: Wez <wez@chromium.org>
Date: Wed Feb 28 08:18:05 2018

Roll back Fuchsia SDK temporarily.

The most recent SDK roll has broken an IPC file-descriptor passing test,
which appears to be due to a mismatch between the semantics of
fdio_transfer_fd() and fdio_create_fd().

TBR: scottmg
Bug:  817241 
Change-Id: I020425fa1a948c67f8cd8cdb2b0b64cd052c7479
Reviewed-on: https://chromium-review.googlesource.com/940164
Reviewed-by: Wez <wez@chromium.org>
Reviewed-by: Scott Graham <scottmg@chromium.org>
Commit-Queue: Wez <wez@chromium.org>
Cr-Commit-Position: refs/heads/master@{#539756}
[modify] https://crrev.com/acf081fce6f21b2573639df45609a12f3eaaba68/build/fuchsia/update_sdk.py

Comment 3 by w...@chromium.org, Feb 28 2018

Cc: smklein@chromium.org
smklein: It looks like the special file '/dev/null' is being mis-translated by fdio_transfer_fd/fdio_create_fd, from a "file" remote-IO fd to a "service" one, at a first glance.

Comment 4 by w...@chromium.org, Feb 28 2018

Re #3: Correction; the special file in question is /dev/zero. :)

Comment 5 by w...@chromium.org, Feb 28 2018

Cc: jam...@chromium.org
Should be fixed by https://fuchsia-review.googlesource.com/c/zircon/+/127900

Let me know if that doesn't work.

Comment 7 by w...@chromium.org, Mar 1 2018

Re #6: Confirmed latest SDK builds work again & have uploaded a roll to the CQ.
Project Member

Comment 8 by bugdroid1@chromium.org, Mar 1 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/26f4cdcf5fea1eb642e3cabd35c1bd1aaf502303

commit 26f4cdcf5fea1eb642e3cabd35c1bd1aaf502303
Author: Wez <wez@chromium.org>
Date: Thu Mar 01 07:39:27 2018

Roll Fuchsia SDK to 9d4016533477903c796470e7ab46c2e1dad31761.

- Removes sys/resource.h, requiring some additional conditionals around
  includes of that header under OS_POSIX.
- Improves the system PRNG implementation to reduce the potential for
  lock contention.

TBR: scottmg
Bug: 707030,  801353 ,  809660 , 738275,  817241 ,  817586 
Change-Id: I4e01a0b23ad66c060ac7a6776a45329775117886
Reviewed-on: https://chromium-review.googlesource.com/942685
Reviewed-by: Wez <wez@chromium.org>
Commit-Queue: Wez <wez@chromium.org>
Cr-Commit-Position: refs/heads/master@{#540093}
[modify] https://crrev.com/26f4cdcf5fea1eb642e3cabd35c1bd1aaf502303/build/fuchsia/update_sdk.py

Comment 9 by w...@chromium.org, Mar 1 2018

Status: Fixed (was: Started)

Sign in to add a comment