L2_hsic (latest update 12/12/13)

Search This thread

becosemsaida

Senior Member
Jul 29, 2008
311
38
Tried to do a test and had some results.......

Did a reset on my note (normal one not factory), plugged in the charger with screen on, disabled all data connections (wi-fi and 3g) and charged it for a night.

Took it off charger unlocked, did not enable any data connection, very hard but i resisted, during all day. Some sms and calls, some music etc. etc. ALWAYS closed the app after using it.

At night did a check: no L2_hsic wakelock, battery at 70% (never happened before with data on) so did the same but this time with data connection on......

Next day: l2_hsic for a period of time. It must be related with the data connection, now i want to check if it occurs with wi-fi or 3g.

I will try some other stuff but i just can't find the time to try it out....

Stock note (no rooting) on LPY and yes i have the "bug".....

Sorry no screens, forgot completly and did a refresh on cpu spy and bbs.

becosemsaida
 
  • Like
Reactions: suhasn46

exrm

Member
Apr 25, 2011
5
0
I think I found a fix

I had this annoying l2_hsic wakelock problem.
I had to reboot every 3-4 hours to stop the l2_hsic wakelock

I'm running stock.

I disabled fast dormancy (dial *#9900# and disable) and rebooted.
This was my last reboot in 36 hours...
It looks like it fixed the problem, now 16 hours since unplugged and I have only 1 hour of l2_hsic wakelock and the phone keeps deep sleeping.

hope it helps you all
 

gaddy888

Senior Member
Dec 18, 2011
57
3
I had this annoying l2_hsic wakelock problem.
I had to reboot every 3-4 hours to stop the l2_hsic wakelock

I'm running stock.

I disabled fast dormancy (dial *#9900# and disable) and rebooted.
This was my last reboot in 36 hours...
It looks like it fixed the problem, now 16 hours since unplugged and I have only 1 hour of l2_hsic wakelock and the phone keeps deep sleeping.

hope it helps you all

This did nothing for me.

However, I signed out of the Samsung dive account. ( A certain "mobile tracker " App crashed and notified me. After which my battery life was nothing like I've seen before on ICS. I'm assuming this mobile tracker is related to the Samsung dive service)

Testing now. So far it seems to have helped.



Sent from my GT-N7000 using XDA
 

Ast2001

Senior Member
May 17, 2006
92
13
Paris
Xiaomi Mi Pad 4
Google Pixel 6
This did nothing for me.

However, I signed out of the Samsung dive account. ( A certain "mobile tracker " App crashed and notified me. After which my battery life was nothing like I've seen before on ICS. I'm assuming this mobile tracker is related to the Samsung dive service)

Testing now. So far it seems to have helped.

Sent from my GT-N7000 using XDA

I signed out to my Samsung Dive account three hours ago. So far so good :)

I will check every hour and report any problem.
 

Ast2001

Senior Member
May 17, 2006
92
13
Paris
Xiaomi Mi Pad 4
Google Pixel 6
This did nothing for me.

However, I signed out of the Samsung dive account. ( A certain "mobile tracker " App crashed and notified me. After which my battery life was nothing like I've seen before on ICS. I'm assuming this mobile tracker is related to the Samsung dive service)

Testing now. So far it seems to have helped.

I'm sorry but it's not the solution. Four hours later, I have observed the return of the infamous l2_hsic wakelock. :-(
 

priyanv

Senior Member
Jan 8, 2012
259
46
Chennai
So is this wakelock raised right from the leaked or AOSP version ICS or from stock alone??
It seems thread started only after official release..

Sent from my GT-N7000 using xda premium
 
Last edited:

jbeef86

Senior Member
May 18, 2012
192
56
Oakland, CA
So is this wakelock raised right from the leaked or AOSP version ICS or from stock alone??
It seems thread started only after official release..

Sent from my GT-N7000 using xda premium

Stock LPY on mine has it. I get l2_hsic and tons of deleted_wake_locks on stock. Its an issue.

EDIT...and then just like that l2_hsic gets a wakelock for the past 2 hours after doing nothing else with my phone other than locking it. Phone hasnt been in deep sleep since then. Nothing but issues with this phone and its bad if you are not paying attention because the battery can drain quickly at 200mhz.
 
Last edited:

frasch42

Member
Jul 14, 2011
28
3
This is strange. My device was l2_hsic-free for days after I went to RocketROM V5. Yesterday I used a friends charger (for Motorola Defy) and I got the wakelock.
This is reproducable. As soon as I use this charger and take off the phone in locked state, l2_hsic kicks in and stays until reboot. When I use the SammyCharger everything is fine.

Sent from my GT-N7000 using xda premium
 

Ast2001

Senior Member
May 17, 2006
92
13
Paris
Xiaomi Mi Pad 4
Google Pixel 6
Ok. I have chosen another approach. I was on LPY and I've just changed my firmware to LQ3. I will keep you informed to the status of l2_hsic wakelock now.

Envoyé depuis mon GT-N7000 avec Tapatalk

18 hours after the migration to LQ3 and only a fair amount of h2_lsic for the moment. I don't know if the problem will be resolved by this firmware but I can assure that LQ3 is a far better ROM than LPY in terms of battery drain...
 

shaftenberg

Senior Member
Jul 31, 2010
702
742
Cologne
I am on LQ3 and had the l2_hsic issue for the last 6 hours, sorry guys :)
I tested LPT, LR1 and LB2 modems, no solution.

Now I'll try with fast dormancy disabled.
 

Ast2001

Senior Member
May 17, 2006
92
13
Paris
Xiaomi Mi Pad 4
Google Pixel 6
I am on LQ3 and had the l2_hsic issue for the last 6 hours, sorry guys :)
I tested LPT, LR1 and LB2 modems, no solution.

Now I'll try with fast dormancy disabled.

24 hours later and still no big l2_hsic 'flood' in my case :) I still have deep sleep.... I have never experienced this kind of result on ICQ

For your information, I have fast dormancy disabled for weeks and it never help to avoid l2_hsic :-(
 

shaftenberg

Senior Member
Jul 31, 2010
702
742
Cologne
So, if disabling fast dormancy doesn't help either (which I'm still testing), then we can sort out these things which do not affect the l2_hsic wakelock:

- equal which modem we are using (GB or ICS)
- all available kernels are affected
- all ICS-ROMs are affected
- disable fast dormancy doesn't help(?)
- Wifi settings aren't responsible
- AC charger is equal (maybe a random thing for this user)
- turning airplane mode or data on/off does not help to kill the service

One guy said, killing the settings service helped to kill the wakelock, ok, fine.
I saw one time, that l2_hsic started just in that moment going wild, when I activated airplane mode(!).
One time it went wild when I changed from HSPA to Wifi.
One member of a german forum discovered, that l2_hsic is bound to a data transmission change of the RIL, i.e. changing cells, changing airplane mode, changing HSPA to Wifi or such things.

So, to sum everything up: which setting could be the culprit? Location service over data? Background data? USB debug? GSM or HSPA? Maybe we should seek in this direction...

[edit]
To add: l2_hsic is not bad in itself when it shows up in BBS, but under some conditions it starts to hang and doesn't shutdown itself. Let's watch out for the conditions :)
 
Last edited:
  • Like
Reactions: frasch42

Braxos

Senior Member
Nov 18, 2011
693
145
Can a rooted guy please somehow disable the stock watchdog aka data volume tracker app and check how l2_hsic is going.

Thanks

Sent from my GT-N7000 using Tapatalk 2
 

liutszho

Senior Member
Dec 9, 2010
157
14
So, if disabling fast dormancy doesn't help either (which I'm still testing), then we can sort out these things which do not affect the l2_hsic wakelock:

- equal which modem we are using (GB or ICS)
- all available kernels are affected
- all ICS-ROMs are affected
- disable fast dormancy doesn't help(?)
- Wifi settings aren't responsible
- AC charger is equal (maybe a random thing for this user)
- turning airplane mode or data on/off does not help to kill the service

One guy said, killing the settings service helped to kill the wakelock, ok, fine.
I saw one time, that l2_hsic started just in that moment going wild, when I activated airplane mode(!).
One time it went wild when I changed from HSPA to Wifi.
One member of a german forum discovered, that l2_hsic is bound to a data transmission change of the RIL, i.e. changing cells, changing airplane mode, changing HSPA to Wifi or such things.

So, to sum everything up: which setting could be the culprit? Location service over data? Background data? USB debug? GSM or HSPA? Maybe we should seek in this direction...

[edit]
To add: l2_hsic is not bad in itself when it shows up in BBS, but under some conditions it starts to hang and doesn't shutdown itself. Let's watch out for the conditions :)

Fast dormancy off does not fix it. Tested.

Sent from my GT-N7000 using xda premium
 

Top Liked Posts

  • There are no posts matching your filters.
  • 6
    Yesterday I had this problem with l2_hsic......

    Read the rest here

    I moved my post because I will need the space for the findings, as I don't see the light at the end of the tunnel.

    Here are the findings till now from some users.
    I will update this when new ideas/solutions have posted.

    what is l2_hsic read the rest here

    HSIC is technology for inter processor comminication. Our phone has two main chipsets AP (here snapdragon, thus the name MSM ) and modem chipset (again by qualcomm). They both are connected via hsic, ...........

    Yes, HSIC is an hardware along with its device driver used for Inter Processor Communication.

    Main principle is Android RIL communicates with help of hsic driver to communicate to the CP (modem) chip. And vice-versa. Modem chip also has the same hsic at there side also.

    .............

    That L2 does not stand for L2 of OSI layer................ There are 4 states (L0(active),L1,L2,L3) state of power saving in USB link also , hsic also uses the same convention.
    ................


    kernel related? read the rest here


    Hello
    I've been encountering L2_hsic and a few other related wakelocks for months (even years), ever since updating my I9100T to ICS (which I did the minute custom ICS roms hit web).......

    I searched around and found out the l2_hsic wakelocks are generated by a kernel file "/drivers/misc/modem_if/modem_link_device_hsic.c" which is also responsible for rpm_hsic and tx_hsic wakelocks.
    From there I went to CM git and with help from their IRC channel I found the source code for their version of this file:
    https://github.com/CyanogenMod/andr...rivers/misc/modem_if/modem_link_device_hsic.c

    A explanation from anirudhgargi what hsic is! (09/06/12) post 264 :

    HISC is the interface that connected the main Soc processor ( Also called Applciation Processor AP) to the modem ( 3G or 4G if there are two modems separately )

    HSIC is used the standard USB core library. And it is stanrd given by usb.org. The HSIC by itself consumes very less power and is one of the very power efficient and space efficient ways of interconnecting chips. Currently common to Qualcomm and Sammy Exynos architectures.

    The reason you seeing too many HSIC wake lock and other wake us's is due to the fact that the HSIC interface is waken up by modem (Normal scenarios are also so many) .


    A find about hsic from becosemsaida (09/06/12) post #107 :

    Well i have the wake-lock too and after some digging around i found this:

    http://www.synopsys.com/dw/ipdir.php?ds=dwc_usb_lpm_hsic

    A article about wakelocks from unthinkable apps(18/06/12) post #126 :



    A sume of what we know from Shaftenberg(20/06/12) post #136

    - equal which modem we are using (GB or ICS)
    - all available kernels are affected
    - all ICS-ROMs are affected
    - disable fast dormancy doesn't help(?)
    - Wifi settings aren't responsible
    - AC charger is equal (maybe a random thing for this user)
    - turning airplane mode or data on/off does not help to kill the service

    a found on the kernel code from funky81 (25/06/12) post #179

    After search codes in the kernel related with this problem, I've found out the conclusion is that this wild uncontrolled wake lock related with data communication (modem)

    Code:
    ./drivers/misc/modem_if/modem_link_device_hsic.c:wake_lock_init(&pm_data->l2_wake, WAKE_LOCK_SUSPEND, "l2_hsic");

    I will try to update the code w/ Samsung Galaxy S3 source code in my kernel...

    I will inform you if there's any progress with this

    Could a app with it's traffic causing it? (26/01/13) post #273


    ....... I got suspicious. I then forced notifications from Asphalt 7 off in Jelly Beans App Administration. After doing that, I have had no l2_hsic wake lock. ......... Problem solved for me:)

    Till now 4.0.4 dbt is not giving high numbers of wakelocks with l2_hsic, 11min in 24 hours cycle


    Dates are using DD/MM/UT format

    Sent from my GT-N7000 using Tapatalk 2
    3
    L2_HISC Wake lock explanation summary

    This is getting ridiculous. Nobody knows what triggers l2_hsic, nobody knows how to kill it without rebooting. C'mon man... Somebody do something already! This is so frustrating! Please please please!!! Samsung, Android, xda-developers... Anyone! For the love of everything sacred! *end of rant*

    HISC is the interface that connected the main Soc processor ( Also called Applciation Processor AP) to the modem ( 3G or 4G if there are two modems separately )

    HSIC is used the standard USB core library. And it is stanrd given by usb.org. The HSIC by itself consumes very less power and is one of the very power efficient and space efficient ways of interconnecting chips. Currently common to Qualcomm and Sammy Exynos architectures.

    The reason you seeing too many HSIC wake lock and other wake us's is due to the fact that the HSIC interface is waken up by modem (Normal scenarios are also so many) .

    Also one imp reason is EFS sync. A concept : that the new modem architectures doesn't support a flash storage.(Saves lot of cost to manufacturer) .So every time modem needs to store something , it wakes up hisc interface and ask the AP (Main processor) to store it for him.

    So this CP storing via AP is a fairy frequent activity.

    I see the concern over this L2_hsic wake lock. I can say at this point that it is quite naturally and a necessary evil for smooth working of your modem( network 2g,3g,4g ) subsystem.

    Hope this is of some help. :good:
    1
    I got a 6h wakelock from l2-hsic today and I never installed neither nova or chrome. So I assume they can be ruled out

    Main reason is unplugging the charger without unlocking lockscreen (swipe to unlock) or unplugging when screen is black.

    Sent from my GT-N7000 using Tapatalk 2
    1
    How do you kill the settings service?

    Sent from my GT-N7000 using XDA

    I have a german Phone. Don't know if the english Names of the Items are right. Perhaps someone who has a english GT-N7000 can translate ?:

    Kill the Settings (german: Einstellungen) Service:

    Go to Settings (german: Einstellungen)
    => Applications (german: Anwendungen)
    => Running (german: Ausführen, Third Tab !)
    => Choose The "Settings" Entry (german: Einstellungen)
    => Hit the "Stopp" Button
    1
    http://xdaforums.com/showthread.php?p=41338478#post41338478

    l2_hsic root cause
    Analysis of kmsg files provide a consistent pattern. Function if_usb_suspend(..) from modem_link_device_hsic.c does not call wake_lock_timeout(..) in wakelock.c in case of l2_hsic is running amok.
    The missed call to wake_lock_timeout is obviously caused by a non cancellable USB-connection.
    It is a single non expiring wakelock, what makes trouble.