Closed Bug 1137091 Opened 10 years ago Closed 10 years ago

[FFOS2.0][Woodduck][HOMO][Orange AMEA] Receive Incomming call during play video under Dailymotion

Categories

(Firefox OS Graveyard :: Gaia::Dialer, defect, P2)

defect

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: sync-1, Unassigned, NeedInfo)

References

Details

Attachments

(3 files)

Created an attachment (id=1153824) ADB log DEFECT DESCRIPTION: During play of video under Dailymotion, the screen of the incomming call take time to display, and sometimes the phone block during the test. It can be due to the memory, in attachement you will find the ADB logs and image about the memory usage during the test. REPRODUCING PROCEDURES: EXPECTED BEHAVIOUR: ASSOCIATE SPECIFICATION: TEST PLAN REFERENCE: TOOLS AND PLATFORMS USED: USER IMPACT: REPRODUCING RATE: For FT PR, Please list reference mobile's behavior:
Attached image Memory usage (deleted) —
Attached file ADB log (deleted) —
The log doesn't show any LMK activity so I doubt this is a memory problem. Is the device you've tested on using a zram swap?
Hi reporter, could you please answer the question in comment 4?
Flags: needinfo?(sync-1)
Blocks: Woodduck
Flags: needinfo?(xingming.yin.hz)
Blocks: Woodduck_P2
Hi Jason, Do you know the answer for comment 4? Thanks!
Flags: needinfo?(chien-hao.li)
(In reply to Gabriele Svelto [:gsvelto] from comment #4) > The log doesn't show any LMK activity so I doubt this is a memory problem. > Is the device you've tested on using a zram swap? Hi: How can I confirm whether the device is using a zram swap or not?Thanks。
Flags: needinfo?(xingming.yin.hz)
(In reply to Gabriele Svelto [:gsvelto] from comment #4) > The log doesn't show any LMK activity so I doubt this is a memory problem. > Is the device you've tested on using a zram swap? Our config: CONFIG_ZRAM=y CONFIG_ZRAM_DEBUG=y CONFIG_ZSMALLOC=y Hi,I think we are using a zram。
(In reply to xingming.yin from comment #9) > Our config: > > CONFIG_ZRAM=y > CONFIG_ZRAM_DEBUG=y > CONFIG_ZSMALLOC=y > > Hi,I think we are using a zram。 OK, then my guess is that you're running into excessive swapping; we've seen this issue in the past on devices similarly configured. What might be happening is the following: under normal conditions when the device is running low on memory we start killing the least recently used applications to make room for the foreground ones, in this case the dialer. When zram is enabled however the LMK - which is responsible for killing background apps - is confused by the presence of the zram swap which it considers free memory. What happens is that instead of being killed the video application is being swapped in and out continuously, consuming a large amount of CPU time. We already have a solution in master for this in the form of bug 1081871 and bug 1082290. You might want to try uplifting them on your device and see if they help. Note that you will need cgroup cpu and memory controllers enabled on the device, these correspond to the following kernel parameters: CONFIG_CGROUP_SCHED=y CONFIG_CGROUP_MEM_RES_CTLR=y CONFIG_CGROUP_MEM_RES_CTLR_SWAP=y
(In reply to Gabriele Svelto [:gsvelto] from comment #10) > (In reply to xingming.yin from comment #9) > > Our config: > > > > CONFIG_ZRAM=y > > CONFIG_ZRAM_DEBUG=y > > CONFIG_ZSMALLOC=y > > > > Hi,I think we are using a zram。 > > OK, then my guess is that you're running into excessive swapping; we've seen > this issue in the past on devices similarly configured. > > What might be happening is the following: under normal conditions when the > device is running low on memory we start killing the least recently used > applications to make room for the foreground ones, in this case the dialer. > When zram is enabled however the LMK - which is responsible for killing > background apps - is confused by the presence of the zram swap which it > considers free memory. What happens is that instead of being killed the > video application is being swapped in and out continuously, consuming a > large amount of CPU time. > > We already have a solution in master for this in the form of bug 1081871 and > bug 1082290. You might want to try uplifting them on your device and see if > they help. Note that you will need cgroup cpu and memory controllers enabled > on the device, these correspond to the following kernel parameters: > > CONFIG_CGROUP_SCHED=y > CONFIG_CGROUP_MEM_RES_CTLR=y > CONFIG_CGROUP_MEM_RES_CTLR_SWAP=y Hi: Our development base on the Firefox OS 2.0M.May I ask whether these configs is useful as like as the Firefox OS master.
(In reply to xingming.yin from comment #11) > Our development base on the Firefox OS 2.0M.May I ask whether these > configs is useful as like as the Firefox OS master. Yes, that configuration can be useful on zram swap-based devices, provided you also take the patches for bug 1081871 and bug 1082290. We (Mozilla) haven't uplifted those patches to 2.0, they are only in master and 2.2 at the moment so you'll have to try by yourself and see if they help.
xingming, could you please try out the patches mentioned in comment 12?
Flags: needinfo?(xingming.yin.hz)
Hi XingMing, Can you update your test result per comment 12?
Status: NEW → RESOLVED
Closed: 10 years ago
Flags: needinfo?(xingming.yin.hz)
Resolution: --- → FIXED
Flags: needinfo?(chien-hao.li)
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: