Closed Bug 956841 Opened 12 years ago Closed 12 years ago

[B2G][Dialer][Bluetooth] Navigation perf occurs while within a bluetooth call

Categories

(Firefox OS Graveyard :: Bluetooth, defect, P2)

ARM
Gonk (Firefox OS)
defect

Tracking

(blocking-b2g:1.4+, firefox28 wontfix, firefox29 fixed, b2g-v1.3 unaffected, b2g-v1.3T unaffected, b2g-v1.4 verified)

VERIFIED FIXED
1.3 C2/1.4 S2(17jan)
blocking-b2g 1.4+
Tracking Status
firefox28 --- wontfix
firefox29 --- fixed
b2g-v1.3 --- unaffected
b2g-v1.3T --- unaffected
b2g-v1.4 --- verified

People

(Reporter: gbennett, Assigned: tzimmermann)

References

()

Details

(Keywords: perf, regression, Whiteboard: [c=handeye p= s=2014.01.17 u=1.4])

Attachments

(2 files, 1 obsolete file)

Attached file BluetoothPerf_LOG.txt
Description: While within a Bluetooth call if you press the home button and navigate around the phone the framerate will drop significantly and become unresponsive at times for 3-8 seconds. Repro Steps: 1) Updated Buri to BuildID: 20140106040201 2) Connect a bluetooth headset device 3) Make a phone call and have the other phone answer 4) Press the home button and attempt to go into the settings app Actual: Severe perf occurs and navigating the phone is quite difficult. Expected: User is able to navigate anywhere on the phone while in a bluetooth call. Environmental Variables: Device: Buri 1.4 MOZ BuildID: 20140106040201 Gaia: 9a222ac02db176e47299bb37112ae40aeadbeca7 Gecko: 14ac61461f2a Version: 29.0a1 Firmware Version: v1.2_20131115 Notes: This does not occur on 1.3. There is a crash shown in the video due to the perf a Bluetooth call creates. Let me know if that should be written separately. Repro frequency: 3/3, 100% See attached: http://www.youtube.com/watch?v=sUTptGN3Zms, BluetoothPerf_LOG.txt
Link to the crash that occured. It seems this bug causes an older crash. https://crash-stats.mozilla.com/report/index/e525fa38-54af-4dfe-98bd-37ee72140106
blocking-b2g: --- → 1.4?
I'll take this one. Since this doesn't occur on v1.3, I'm not going to do it this week. Will start to investigate next Monday.
Assignee: nobody → echou
Status: NEW → ASSIGNED
Priority: -- → P2
Whiteboard: [c=handeye p= s= u=]
QA Contact: sparsons
This issue started to occur on the Buri 1.4 Build ID: 20131213040203 Gaia 390b313a254a947d12e3cdbcde19d7d1619ff63c SourceStamp 8b5875dc7e31 BuildID 20131213040203 Version 29.0a1 Last working Buri 1.4 Build ID: 20131212040203 Gaia 8952898bbc98dd31e25b647203791cf129867ff1 SourceStamp 1ad9af3a2ab8 BuildID 20131212040203 Version 29.0a1
Update: I can reproduce. Still trying to find the root cause. > Repro Steps: > 1) Updated Buri to BuildID: 20140106040201 > 2) Connect a bluetooth headset device Actually this step is not necessary. The issue could be easily reproduced as long as Bluetooth is on. > 3) Make a phone call and have the other phone answer > 4) Press the home button and attempt to go into the settings app
This is a regression of bug 932728. Still figuring out the root cause.
Hi Thomas, After bisect I found out that this bug seems to be caused by bug 932728. Would you mind taking a look?
Flags: needinfo?(tzimmermann)
Blocks: 932738
Blocks: 932728
No longer blocks: 932738
Ok, I'll have a look.
Assignee: echou → tzimmermann
Flags: needinfo?(tzimmermann)
triage: regression 1.4+ to make sure this gets fixed in 1.4
blocking-b2g: 1.4? → 1.4+
I was able to reproduce the issue. It's caused by constantly reporting DBus sockets to be writable by calling OnFileCanWriteWithoutBlocking in the I/O thread. The attached patch fixes the problem for me.
Attachment #8359796 - Flags: review?(kyle)
Comment on attachment 8359796 [details] [diff] [review] [01] Bug 956841: Respect DBusWatch flags in I/O-thread watchers Review of attachment 8359796 [details] [diff] [review]: ----------------------------------------------------------------- ::: ipc/dbus/DBusThread.cpp @@ +124,5 @@ > MessageLoopForIO* ioLoop = MessageLoopForIO::current(); > + > + unsigned int flags = dbus_watch_get_flags(mWatch); > + > + if (flags&DBUS_WATCH_READABLE) { supernit: Spaces around the & might be more readable, here and below.
Attachment #8359796 - Flags: review?(kyle) → review+
Status: ASSIGNED → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Whiteboard: [c=handeye p= s= u=] → [c=handeye p= s=2014.01.17 u=1.4]
Target Milestone: --- → 1.3 C2/1.4 S2(17jan)
Keywords: verifyme
Unable to reproduce on todays Buri v1.4. Verified fixed. v1.4 Environmental Variables: Device: Buri v1.4 MOZ BuildID: 20140502000201 Gaia: 7b2b82d72cbdd1c7e0f4542cb3390802e65f473e Gecko: 50be03cea340 Version: 30.0 Firmware Version: v1.2-device.cfg Also unable to reproduce on todays Open C v1.4 v1.4 Environmental Variables: Device: Open C v1.4 BuildID: 20140502000201 Gaia: 7b2b82d72cbdd1c7e0f4542cb3390802e65f473e Gecko: 50be03cea340 Version: 30.0 Firmware Version: P821A10_ENG_20140410
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: