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

Issue 622724 link

Starred by 2 users

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

DisplayTest.CreateLinuxDMABufBuffer crash

Project Member Reported by osh...@chromium.org, Jun 23 2016

Issue description

This test crashes as of 401602. Assigning to dcastagna@ who updated the test recently.


/usr/lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/debug/vector:346:
    error: attempt to subscript container with out-of-bounds index 0, but     
    container only holds 0 elements.

Objects involved in the operation:
sequence "this" @ 0x0x7ffc1f7111c0 {
}
Received signal 6
#0 0x7f505c7f667e base::debug::StackTrace::StackTrace()
#1 0x7f505c7f61bf base::debug::(anonymous namespace)::StackDumpSignalHandler()
#2 0x7f5048c2c330 <unknown>
#3 0x7f504888dc37 gsignal
#4 0x7f5048891028 abort
#5 0x7f5049100fe5 __gnu_debug::_Error_formatter::_M_error()
#6 0x00000050a1d0 std::__debug::vector<>::operator[]()
#7 0x000000506b11 exo::(anonymous namespace)::DisplayTest_CreateLinuxDMABufBuffer_Test::TestBody()
#8 0x0000008815da _ZN7testing8internal12InvokeHelperIjNSt3tr15tupleIJEEEE12InvokeMethodIN2gl13GPUTimingFakeEMS8_FjvEEEjPT_T0_RKS4_
#9 0x0000013fdf1e testing::internal::HandleExceptionsInMethodIfSupported<>()
#10 0x0000013f3055 testing::Test::Run()
#11 0x0000013f3778 testing::TestInfo::Run()
#12 0x0000013f3d1a testing::TestCase::Run()
#13 0x0000013f906c testing::internal::UnitTestImpl::RunAllTests()
#14 0x0000014107ba testing::internal::HandleSehExceptionsInMethodIfSupported<>()
#15 0x0000013ff5ce testing::internal::HandleExceptionsInMethodIfSupported<>()
#16 0x0000013f8d0f testing::UnitTest::Run()

 
Is this a new regression?
AFAIK that test never passed nor ran on a real device. 

They're expected to fail if you run them on linux with ozone since you can't really use ozone drm on linux.

I'm assuming you ran them on linux/ozone. Should we just disable that test on that device?

I tried to make it run on a real cros device but it's not trivial, since that test depends on AuraShellTestSuite that has many dependencies.

Comment 2 by osh...@chromium.org, Jun 24 2016

I don't quite understand. Where does this suppose to pass then?


It's supposed to pass on a real cros device.
But as far as I know it has never happened so far.
crrev.com/2091223002 is a draft of what's needed to run exo_unittests on a cros device.

The problem with that patch is that when the test initializes Aura, it basically start the whole window system and tries to load many files that have to be copied manually to the devices.

Comment 5 by osh...@chromium.org, Jun 24 2016

tests on chromium tree is supposed to run locally and on waterfall.

If that's supposed to run on a device, you need to move it to chromiumos test suite.

https://www.chromium.org/chromium-os/testing/adding-unit-tests-to-the-build

Or if this is just for your testing, then it be in a separate command/test binary.


Project Member

Comment 6 by sheriffbot@chromium.org, Jul 3 2016

Labels: -M-53 M-54 MovedFrom-53
Moving this nonessential bug to the next milestone.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Cc: j.iso...@samsung.com
Components: Tests OS>Kernel>Graphics

Sign in to add a comment