New issue
Advanced search Search tips

Issue 711433 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 710492
Owner: ----
Closed: Apr 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Bug



Sign in to add a comment

VMTest failed in eve-pre-cq because cryptohomed hangs

Project Member Reported by xiaoh...@chromium.org, Apr 13 2017

Issue description

This happens with https://chromium-review.googlesource.com/c/465426/
which triggers a eve-pre-cq test
https://luci-milo.appspot.com/buildbot/chromiumos.tryserver/pre_cq/27592

We can also repro locally with steps below.
What steps will reproduce the problem?
(1) Build eve qemu vm image
(2) Run the vm, sudo kvm -m 1024 -vga cirrus -pidfile /tmp/kvm.pid -net nic,model=virtio -net user,hostfwd=tcp:127.0.0.1:9222-:22 -hda [path to chromiumos_qemu_image.bin]
(3) Run test: test_that --board=eve localhost:9222 'login_OwnershipTaken'

What is the expected result?
test pass

What happens instead?
test times out

Looking at the logs it shows the cryptohomed hangs.  Manually trying to login in the vm using a gaia test account also hangs.

2017-04-13T19:56:22.282387+00:00 INFO attestationd[7981]: Attestation Daemon Started.
2017-04-13T19:56:22.283729+00:00 WARNING attestationd[7981]: libminijail[7981]: allowing syscall: connect
2017-04-13T19:56:22.283808+00:00 WARNING attestationd[7981]: libminijail[7981]: allowing syscall: sendto
2017-04-13T19:56:22.284058+00:00 WARNING attestationd[7981]: libminijail[7981]: logging seccomp filter failures
2017-04-13T19:56:22.285077+00:00 INFO attestationd[7981]: Attestation service started.
2017-04-13T19:56:30.775168+00:00 WARNING cryptohomed[5940]: Failed to initialize the trunks IPC proxy; trunksd is not ready.
2017-04-13T19:56:30.775219+00:00 ERR cryptohomed[5940]: Failed to initialize trunks factory.
2017-04-13T19:56:31.705742+00:00 WARNING kernel: [ 1630.330080] hwrng: no data available
2017-04-13T19:56:34.039765+00:00 WARNING cryptohomed[5940]: Failed to initialize the trunks IPC proxy; trunksd is not ready.
2017-04-13T19:56:34.039786+00:00 ERR cryptohomed[5940]: Failed to initialize trunks factory.
2017-04-13T19:56:34.039851+00:00 WARNING cryptohomed[5940]: Failed to finalize boot lockbox.
2017-04-13T19:56:34.039954+00:00 WARNING cryptohomed[5940]: Could not load the device policy file.
2017-04-13T19:56:41.706876+00:00 WARNING kernel: [ 1640.331120] hwrng: no data available
2017-04-13T19:56:50.800321+00:00 WARNING tpm_managerd[7971]: Failed to initialize the trunks IPC proxy; trunksd is not ready.
2017-04-13T19:56:50.801828+00:00 ERR tpm_managerd[7971]: Error TrunksDBusProxy cannot connect to trunksd.
2017-04-13T19:56:50.801869+00:00 ERR tpm_managerd[7971]: GetCapability: SAPI_RC_NO_CONNECTION
2017-04-13T19:56:50.801877+00:00 ERR tpm_managerd[7971]: Failed to query TPM properties: SAPI_RC_NO_CONNECTION
2017-04-13T19:56:50.801892+00:00 WARNING tpm_managerd[7971]: Error initializing trunks tpm state: SAPI_RC_NO_CONNECTION
2017-04-13T19:56:50.802210+00:00 CRIT tpm_managerd[7971]: Check failed: initialized_. #012/usr/lib64/libbase-core-395517.so(base::debug::StackTrace::StackTrace()+0x13) [0x7a283e60eda3]#012
2017-04-13T19:56:50.820301+00:00 WARNING crash_reporter[8038]: Could not load the device policy file.
2017-04-13T19:56:50.820585+00:00 WARNING crash_reporter[8038]: [user] Received crash notification for tpm_managerd[7971] sig 6, user 0 (developer build - not testing - always dumping)
2017-04-13T19:56:50.821739+00:00 INFO crash_reporter[8038]: State of crashed process [7971]: D (disk sleep)
2017-04-13T19:56:50.822203+00:00 WARNING crash_reporter[8038]: Crash directory /var/spool/crash already full with 32 pending reports
2017-04-13T19:56:50.822219+00:00 ERR crash_reporter[8038]: Could not create crash directory
2017-04-13T19:56:50.822227+00:00 ERR crash_reporter[8038]: Unable to find/create process-specific crash path
2017-04-13T19:56:50.826600+00:00 WARNING kernel: [ 1649.450700] init: tpm_managerd main process (7971) killed by ABRT signal
2017-04-13T19:56:50.826611+00:00 WARNING kernel: [ 1649.450721] init: tpm_managerd main process ended, respawning





 

Comment 1 by xiy...@chromium.org, Apr 13 2017

Cc: apronin@chromium.org
apronin@, would cryptohome with USE="tpm2" work in a VM ?

It seems cryptohme has problems talking to trunksd, tpm_managerd etc in such setup, which might be the cause of the failed mount/login request.
Yes, trunksd, tpm_managerd and attestationd expect to find tpm, and die if it's not there (or trunksd is not there).
cryptohomed is at least slowed down by constant attempts to communicate with trunksd, tpm_managerd and attestationd over dbus.
It was decided to remove VMTests from eve-pre-cq - see  issue 710492 .

Mergedinto: 710492
Status: Duplicate (was: Untriaged)

Sign in to add a comment