A few tips and bug fixing workarounds - Sony Xperia M2

If like me after installing Lollipop you're experiencing the GPS going nut, a bad battery drain or the WiFi showing an exclamation mark dropping the connections these could be your solutions!
GPS LOSING SIGNAL (it needs root):
OBSOLETE - GPS is now fixed in 18.6.A.0.182
After reading some complaints about the GPS not working in Lollipop I checked it in my phone and I discovered it was indeed behaving poorly. In my case it does get the signals from the satellites, computes the position and other value and then after just few seconds it starts all over again. Comparing the 5.1.1 /etc/gps.conf file with the the one in KitKat and some suggested in the internet I tested few changes and I managed to make it fully operational just changing a single parameter:
Actually it seems the gps.conf file from KitKat does fix the issue. It happened that I made a mess with the file permissions (umask is set by default to 0077). @nunu00 made me notice that it worked for him. So I tested it again, and after fixing the file permissions it does seem to do the job.
You can find a copy of the KK file in http://forum.xda-developers.com/showpost.php?p=63285443&postcount=2.
As root make a backup of the original /etc/gps.conf and copy it over (as gps.conf. Remove the ".txt" postfix). Be sure the file is readable by everyone. Then reboot the phone and the GPS will be back to normal.
Additionally (but not necessarily) one may want to use an NTP server for his or her geographical region, like for example, if one is in Europe should comment out the stock NTP server and set instead the European one:
Code:
#Test
# NTP_SERVER=time.gpsonextra.net
#Asia
# NTP_SERVER=asia.pool.ntp.org
#Europe
[B]NTP_SERVER=europe.pool.ntp.org[/B]
#North America
# NTP_SERVER=north-america.pool.ntp.org
Even better on can add the NTP server, if available, for the Country, like I did for Italy:
Code:
#Italy
NTP_SERVER=it.pool.ntp.org
All other servers must obviously be commented out.
I/O WAIT ALWAYS AT 50% (it needs root):
Lollipop added in the Xperia M2 the compressed virtual memory, also known as zram. It is a feature that I personally like (I implemented it in the kernel for my crappy low range/low memory/low quality Samsung phone I had previously) as long as it works as it must. At the cost of little CPU time one can benefit from few hundreds megabytes of more usable RAM. Unfortunately in Lollipop it seems bugged, since the zram makes the process kswapd0 in constant I/O Wait (about 50%), which may degrade the max performances and also keeps at least two cores running at 1190MHz, draining the battery. One can see it with the "busybox top" command (busybox is installable in /data/local/tmp by anyone - no root needed):
Code:
Mem: 733016K used, 152060K free, 0K shrd, 40788K buff, 180220K cached
CPU: 1.2% usr 7.4% sys 0.0% nic [B][COLOR="red"]47.3% idle[/COLOR] [COLOR="Blue"]43.8% io[/COLOR][/B] 0.0% irq 0.0% sirq
- Solution 1: enable the "laptop mode" and change the swappiness value
It has been reported by @semil and confirmed by @linuxct that changing the swappiness value to something between 40-60 improves it (with zram enabled). Obviously you need root. Just make a /etc/init.d script with the following line
Code:
sysctl -w vm.swappiness=60
(you can change 60 to 40 or else while testing) or add it to /etc/init.d/00stop_ric. It changed nothing in my phone, though, so I tried tampering with other parameters and eventually I've been successful. Appending after the command above the following two lines
Code:
sysctl -w vm.laptop_mode=1
sysctl -w vm.drop_caches=1
finally the kernel moved out of the I/O wait. In my phone with 130-140 apps installed it always takes a about 10 minutes after boot, so be patient while testing. The two parameters above are non destructive and they get rewritten after each boot, so they can't make things worse than what they already are.
After the update 18.6.A.0.182 the IO Wait is still there. I did some some tests and I could raise the swappiness to 90. The IO Wait still goes down, but theoretically giving a slightly better "multitasking" than with a 40 or 60 value.
- Solution 2: disable the zram
Well, the trick is simple: disable the zram and the iowait has gone. To disable it as root one needs to execute
Code:
swapoff /dev/block/zram0
At any time. But it is obviously suggested to put the command in a script in /etc/init.d (just add it at the end of /etc/init.d/00stop_ric). And this is how the idle and iowait times are now:
Code:
Mem: 780028K used, 105048K free, 1000K shrd, 2284K buff, 149080K cached
CPU: 3.9% usr 4.9% sys 0.0% nic [B][COLOR="Red"]91.0% idle[/COLOR] [COLOR="RoyalBlue"]0.0% io[/COLOR][/B] 0.0% irq 0.0% sirq
WIFI DROPPING CONNECTIONS SHOWING AN EXCLAMATION MARK:
If you see an exclamation mark on the right of the WiFi status icon it means that Lollipop can't access the Internet still being connect to a WiFi AP. This is apparently due to some incompatibility between the new drivers and the Captive Networks, the Hot Spots showing as open but that indeed require a web login to be used.
- Solution 1: Disable the Capitve Networks all together:
The good news is that as regular users (non root) we can still switch it off. From a terminal emulator app or via "adb shell" (google to know how to install and play with the latter) simply type:
Code:
settings put global captive_portal_detection_enabled 0
and reboot. To re-enable it give the command:
Code:
settings put global captive_portal_detection_enabled 1
To check the current state (I'd suggest to do it before setting it. It should report 1) type:
Code:
settings get global captive_portal_detection_enabled
To run it via a Terminal emulator on the phone you need to have UID shell. Just execute first:
Code:
su shell
and then one of the commands above. But one needs the phone to be rooted for that.
UPDATE: In 18.6.A.0.182 as soon as I enable the Global Captive Support I get the exclamation mark. It doesn't drop the connection anymore but it is quite annoying.
- Solution 2: Disable the check for open/public networks:
This is still under investigation, but it seems to do the job. Simply go to the WiFi Settings/Advanced and disable the open networks notification. UPDATE: I still get the exclamation mark but it doesn't drop the connection.
GOOGLE SEARCH BAR ANIMATIONS AND VOICE SEARCH (requires root):
In the Xperia Home we can now have the Google search bar (to hide it just "swipe in" two fingers in the screen as when we add widgets and touch the hide button on the right of the bar. Which is my most favorite solution). The ugly thing is that when touching the bar Google Now opens with no animation. Also the option to enable "voice search" from any screen isn't available anymore. The value of "ro.config.low_ram " in /system/buil.prop is set to "true". An user suggested that changing (as root) it to "false" as in below
Code:
ro.config.low_ram=false
should fix it. Always make a backup of build.prop file before modifying it!!
I tested it myself and the option to have the voice search in all screens did appear indeed. I didn't see any changes in the animations, though.
LITTLE ICONS IN XPERIA HOME:
If like me you have been shocked by the new huge icons in the Xperia Home, so much that you were already thinking about switching the home launcher, don't fret: the small icons are still there!! Simply "swipe in" two fingers in the screen and together with the ability of adding widgets, changing themes and else on the bottom right there is the new "Home Settings". Click on it and you get the option to have the new rotating Xperia Home and.. yes, the smaller lovely icons back!
ENABLE MULTI USER (it needs root):
Due to the low amount of RAM (see the need for the zram above) and internal memory Sony doesn't support the multi user option in Lolliopop 5.1.1 for devices with 1GB. To enable it just add to the /system/build.prop file the following two lines:
Code:
fw.show_multiuserui=1
fw.max_users=3
where 3 users mean the main account, the guest account and a second account. Always remember to make a backup of the stock build.prop before any modification! If one just wants to test it without messing up with build.prop, the same behavior is achievable withthe following commands:
Code:
setprop fw.show_multiuserui 1
setprop fw.max_users 3
Unfortunately it is missing the user icon in the top right of the notification panel and frankly it is a lag fest. Which should be the reason why Sony removed it.
To disable it first delete any account you have eventually created (otherwise I'm afraid you end up wasting precious internal memory for nothing), restore the build.prop file in case it has been modified and reboot. Everytthing will be back to normal.
[CONTINUE]
--------------------
I hope this helps. If you have any other suggestion/solution, please write about them in the thread and I will add them in the main post. Also this is not a thread about getting help, use the forum for it. Still questions about the mentioned tips are welcome

RAISE ALL MAX VOLUMES (it needs root):
Many complains that in Lollipop 5.1.1 the volumes went lower. There seems to be a solution for it. As suggested by @Ciptraandoni in the files
Code:
/etc/mixer_paths.xml
/etc/mixer_paths_VY55.xml
/etc/mixer_paths_VY58.xml
Look for the lines:
Code:
<ctl name="RX1 Digital Volume" value="84" />
<ctl name="RX2 Digital Volume" value="84" />
<ctl name="RX3 Digital Volume" value="84" />
<ctl name="RX4 Digital Volume" value="84" />
Raising the value 84 by a little bit in all files improves the max volumes. I personally tested it. With the value 88 it was fine, then raised them to the value 91 I have got some distortion. So do it at your own risk. Also always remember to backup your system files first!!
TRIM YOUR NAND (it needs root):
Since Android 4.3 Google introduced the Trim (see link) in the kernel, which should restore the performances on SSD/NAND after a long use. It should be executed regularly by Android (not sure when), still after manually trimming my ext4 partition in the SD card it reported more than 9 GB cleaned. To execute it yourself first check if fstrim is available on your busybox. As root type in the terminal:
Code:
busybox fstrim
If present you'll get the list of options, otherwise an error. If fstrim is available type ("-v" is the verbose option):
Code:
busybox fstrim -v /system
busybox fstrim -v /data
busybox fstrim -v /cache
then, as usual, you can put the last three lines in a /etc/init.d script (they don't take very long. less if recently trimmed).
ONE SHOT KITKAT ROOTING:
@linuxct made a modified rootkit script which exploits one of the latest KK vulnerabilities, allowing to root KK without the need to downgrade the kernel. With my help he also modified the script to disable the RIC in one shot.
You can find the instructions in here.
This isn't directly Lollipop related, but since we need to go through a rooted KK to have a rooted Lollipop i thought to add it here.
IMPROVE THE APPEARANCE OF YOUR ROM
Click on the Google Play button on my signature and download one of my themes. Most are free! Also don't forget to rate them 5 stars.

Great! I thought it was my router problem (WRT160NL + DDwrt), but I'll disable that wifi option. Thanks!

linuxct said:
Great! I thought it was my router problem (WRT160NL + DDwrt), but I'll disable that wifi option. Thanks!
Click to expand...
Click to collapse
You're very welcome, amigo.
Also you may want to try to mess up with the router's QoS settings. But if working (I can't test it myself) that will fix it only for your router. The solution above is much more general. Not sure what are the downsides, but in case, on a weak signal, just manually disable the WiFi yourself and use 3G/4G (the radio?).

Miche1asso said:
You're very welcome, amigo.
Also you may want to try to mess up with the router's QoS settings. But if working (I can't test it myself) that will fix it only for your router. The solution above is much more general. Not sure what are the downsides, but in case, on a weak signal, just manually disable the WiFi yourself and use 3G/4G (the radio?).
Click to expand...
Click to collapse
I have QoS disabled, but I noticed that in both routers I have (160NL and Archer C20i), the exclamation mark only appears when I'm far from them, in another room for example. So it's because of the poor signal, even though some times it shows all bars (max. signal) and the warning. :silly:
I'm going to stay with the setting on 0, and if it still shows the exclamation mark, then I'll try configuring my routers. Grazie

linuxct said:
I have QoS disabled, but I noticed that in both routers I have (160NL and Archer C20i), the exclamation mark only appears when I'm far from them, in another room for example. So it's because of the poor signal, even though some times it shows all bars (max. signal) and the warning. :silly:
I'm going to stay with the setting on 0, and if it still shows the exclamation mark, then I'll try configuring my routers. Grazie
Click to expand...
Click to collapse
Just out of curiosity, when showing the exclamation mark does it switch to 3G/4G? In my case it just drops all connections, and to add insult to injury it happens even few cm from the "router" (which is actually the internet shared via MacBook since the WiFi fried in my router).

Miche1asso said:
Just out of curiosity, when showing the exclamation mark does it switch to 3G/4G? In my case it just drops all connections, and to add insult to injury it happens even few cm from the "router" (which is actually the internet shared via MacBook since the WiFi fried in my router).
Click to expand...
Click to collapse
It drops all connections. I noticed once, because Whatsapp web stopped working. Then I check the phone and the exclamation mark is there. Turn it off and on and it works again.
And yes, I was very near my router, let's say 50cm?

linuxct said:
It drops all connections. I noticed once, because Whatsapp web stopped working. Then I check the phone and the exclamation mark is there. Turn it off and on and it works again.
And yes, I was very near my router, let's say 50cm?
Click to expand...
Click to collapse
Ok, same experience then. So it's better staying off and good riddance!!

Not had any WiFi problems all I know is I have min 50mps home broadband with 5ghz and 2.5ghz networks my data connection is running fine on 4G/3.5G battery is definitely worse but I was running amplify, greenify with a de-bloated ftf and getting 6hrs plus sot regularly it's gonna take me a while to get used to normal battery life again
D2303 5.1 lollipop customised NCB rooted de-bloated LB t2u nut dual recovery x-reality+supervivid

Hello
I have also this wifi problem but i don't understand how to fix it.
I tried, as it's written, to write the command in a terminal app, but it said that i don't have the permission to execute it.
My phone is rooted and i tried also somehow to have the sudo but it didn't work either.
Someone could explain to me what should i do ?
Thank you.

[email protected] said:
Not had any WiFi problems all I know is I have min 50mps home broadband with 5ghz and 2.5ghz networks my data connection is running fine on 4G/3.5G battery is definitely worse but I was running amplify, greenify with a de-bloated ftf and getting 6hrs plus sot regularly it's gonna take me a while to get used to normal battery life again
D2303 5.1 lollipop customised NCB rooted de-bloated LB t2u nut dual recovery x-reality+supervivid
Click to expand...
Click to collapse
Well, you probably remember that I was complaining about KK battery life. Now staying with WiFi and applying the tweaks above, in about 17h I had a drop of 24%. about 70min usage. In other words it was less than 1.5%!!
I'll test it in 4G, but my signal is crap, so I don't expect much . Still the Stamina mode (completely enabled, with the usual apps in white list, like Whatsapp) seems to do a great job, finally.
Niklaus89 said:
Hello
I have also this wifi problem but i don't understand how to fix it.
I tried, as it's written, to write the command in a terminal app, but it said that i don't have the permission to execute it.
My phone is rooted and i tried also somehow to have the sudo but it didn't work either.
Someone could explain to me what should i do ?
Thank you.
Click to expand...
Click to collapse
That's weird. I just doubled checked and it works with both the user and root accounts. i mean, there is little to explain, you just have to type this command:
settings put global captive_portal_detection_enabled 0
I am not sure either what you mean with "i tried also somehow to have the sudo ". To "sudo" in a rooted android you just type "su". "su -c <command>" to run a command (like "su -c id". That should report user 0 and group 0).

Okay thank you, finaly i did it.
I still couldn't do it without typing "su". Thank you.

I dont know if its ok this, but i Just use my celphone today on facebook and whatsapp and a hour of videos on YouTube, without you tip (i Just used with wifi)

I got roughly 5.5hrs screen on time yesterday with no changes to settings. Still not sure about how fast the battery drains though
D2303 5.1 lollipop customised NCB rooted de-bloated LB t2u nut dual recovery x-reality+supervivid

Is there appreciable difference in performance between zram on and off ? Antutu Benchmark shows 19421 points for ON zram and 19889 points for OFF

dark90 said:
I dont know if its ok this, but i Just use my celphone today on facebook and whatsapp and a hour of videos on YouTube, without you tip (i Just used with wifi)
Click to expand...
Click to collapse
Pantalla is screen? And you have got 9+ hours? You're more than fine then. Indeed I am not sure if that iowait thing is a general problem. Can anyone with the zram on and possibly busysbox installed, check the iowait average values using the command
Code:
busybox iostat
I wonder if it is anything wrong with my installatiom, but I had the issue even just after wiping the data.
[email protected] said:
I got roughly 5.5hrs screen on time yesterday with no changes to settings. Still not sure about how fast the battery drains though
Click to expand...
Click to collapse
Yeah, I can't manage to make a real battery evaluation myself. I am still connected many times with adb and that charges the phone, ruining all battery statistics.
~Patryk said:
Is there appreciable difference in performance between zram on and off ? Antutu Benchmark shows 19421 points for ON zram and 19889 points for OFF
Click to expand...
Click to collapse
Well, technically on small jobs the zram does reduce slightly the performances. You see the benefits when you open a previous application, it should have the app screen as you left it more often than when it is turned off.
BTW any chance you can do the test mentioned above? The one about iostat.

Miche1asso said:
Can anyone with the zram on and possibly busysbox installed, check the iowait average values using the command
Code:
busybox iostat
Click to expand...
Click to collapse
Average of 14%. Bouncing from 13.3% to 14.5% in my case.
Oh, BTW, yes, CE1 has zRAM, but Algerian FTF was working worse than this one.
EDIT: Now it went up to 25.04%. Well, sh*t. The values I took before were after a reboot, so, when the device is fresh, just restarted, it only uses a 14~15%, but after opening apps it goes up

linuxct said:
Average of 14%. Bouncing from 13.3% to 14.5% in my case.
Oh, BTW, yes, CE1 has zRAM, but Algerian FTF was working worse than this one.
EDIT: Now it went up to 25.04%. Well, sh*t. The values I took before were after a reboot, so, when the device is fresh, just restarted, it only uses a 14~15%, but after opening apps it goes up
Click to expand...
Click to collapse
Indeed. Mine is stuck at 30-40%. Even higher as you can see from the "top" output in the first post (but that's for a single moment). This is what I get now with zram disabled about one night after reboot :
Code:
avg-cpu: %user %nice %system %iowait %steal %idle
14.00 1.79 10.87 4.13 0.00 69.21
And now I reboot the phone not touching the zram. here now just after boot:
Code:
avg-cpu: %user %nice %system %iowait %steal %idle
24.30 3.89 31.94 14.59 0.00 25.29
Which would be ok, but it is raising. Indeed now, just after few minutes..
Code:
avg-cpu: %user %nice %system %iowait %steal %idle
13.67 2.82 17.22 31.39 0.00 34.91
31.39% and it keeps raising. Nope, disabled it stays.
Just to be clear, since a picture it worth 100 words, this is the effect of having an high iowait:
The CPU stays most of the time at full frequency. Often just for doing nothing but waiting for kswapd0 trying to access the RAM.

Miche1asso said:
Indeed. Mine is stuck at 30-40%. Even higher as you can see from the "top" output in the first post (but that's for a single moment). This is what I get now with zram disabled about one night after reboot :
Code:
avg-cpu: %user %nice %system %iowait %steal %idle
14.00 1.79 10.87 4.13 0.00 69.21
And now I reboot the phone not touching the zram. here now just after boot:
Code:
avg-cpu: %user %nice %system %iowait %steal %idle
24.30 3.89 31.94 14.59 0.00 25.29
Which would be ok, but it is raising. Indeed now, just after few minutes..
Code:
avg-cpu: %user %nice %system %iowait %steal %idle
13.67 2.82 17.22 31.39 0.00 34.91
31.39% and it keeps raising. Nope, disabled it stays.
The CPU stays most of the time at full frequency. Often just for doing nothing but waiting for kswapd0 trying to access the RAM.
Click to expand...
Click to collapse
Clear enough. Just added swapoff to my 00stopric. Now let's see how multitasking goes

Code:
su
swapoff /dev/block/zram0
fails for me... Central Europe prerooted, flashed with twrp. My %iowait is around 23% and counting...
Edit: After an restart, it executed the command just fine, and my idle time is normal.
But now there is another problem - I cannot open 00stop_ric as a text, to add the command... tried both es explorer and total commander... any idea?

Related

I fixed my gps!

If you use a program like visualGPSce, you may notice that you can see lots of satellites, but you can't get a fix. I managed to fix it by:
1) Setting my log file to 3200bytes (using advanced config)
2) Disabling aGPS
3) I then turned off the phone, took out the battery for 10-20 seconds and replaced it.
4) Ran the quickgps program
5) Managed to get a fix and have it working quite fast!
I don't think all of this is necessary, but I have a theory:
I have a holux gpsslim 236B and sometimes it exhibits similar behaviour if you send it screwed up commands. The way to fix that was to remove the battery and then open the case to short out the backup battery (Yep, they were so helpful in enabling it to "remember" your corrupt settings). Reassemble and the gps device works fine. I get a feeling that the raphael may also benefit from such treatment (a forced reset on the gps module).
The only other theory I have is that the gps log file was corrupt (I couldn't delete it so I just made another one) and maybe that fixed it.
I've only given it limited testing, but if it breaks, I'll let you know so you don't waste your time!
I noticed my GPS performance was much better when I set my gpslogfile to 32KB as opposed to setting it's size to 0.
Just my 2c.
Isn't the quickgps tool for agps?
So either the disabling of agps or the quickgps is not necessary
ledg3 said:
I noticed my GPS performance was much better when I set my gpslogfile to 32KB as opposed to setting it's size to 0.
Just my 2c.
Click to expand...
Click to collapse
what's the logic/reasoning behind adjusting he gpslog file? i occasionally use GPS/tomtom though i would probably consider using it more frequently if getting a signal lock didn't take anywhere from 5-15 minutes.
i currently have disabled agps (no data plan so it's useless anyway) and have no logfile set to 0 as per the suggested settings i saw on fuzemobility.com post.
Can you post the file you used to load QuickGPS? I tried to download one but could not get it to run on my XDA Serra.
@kimbja98
how long has it been working? get back to us after a month or so of solid performance... the GPS issues seem to be so random that doing almost anything can seem like it fixes it... but i hope you're right
Just used it on the train on the way into work and it was fine, took about 2 minutes to get a location.
One thing I think may have been the problem is that it thinks it should see X satellites but obviously can't and gets confused (even if there are others in clear view), which may be down to agps, quickgps (I'm not sure if they're the same thing) or corrupted log file but now those satellites it can't see get dropped from the list (you can see this happening in VisualGPSce) and it gets a location after about 5 or 6 have been found. The holux seems to get positions at 4 (which theoretically, is the minimum but could be done with 3 if altitude is negligable) but that uses the sirfstar III chipset...I don't know what the raphael uses though.
I should mention I use ROMeOS 1.40 and that may have something to do with it. Either way, I'll be giving it some fairly rigorous testing. If it does break, I'll know how long it took and it may be that the log file got full and if it that, some regular deleting may be called for.
kimbja98 said:
If you use a program like visualGPSce, you may notice that you can see lots of satellites, but you can't get a fix. I managed to fix it by:
1) Setting my log file to 3200bytes (using advanced config)
2) Disabling aGPS
3) I then turned off the phone, took out the battery for 10-20 seconds and replaced it.
4) Ran the quickgps program
5) Managed to get a fix and have it working quite fast!
I don't think all of this is necessary, but I have a theory:
I have a holux gpsslim 236B and sometimes it exhibits similar behaviour if you send it screwed up commands. The way to fix that was to remove the battery and then open the case to short out the backup battery (Yep, they were so helpful in enabling it to "remember" your corrupt settings). Reassemble and the gps device works fine. I get a feeling that the raphael may also benefit from such treatment (a forced reset on the gps module).
The only other theory I have is that the gps log file was corrupt (I couldn't delete it so I just made another one) and maybe that fixed it.
I've only given it limited testing, but if it breaks, I'll let you know so you don't waste your time!
Click to expand...
Click to collapse
YES, your method seems to be one of working the best.
Congratulation.
I use the mix of some methods in the forum, and my GPS reception is quite good, even in my room inside!!
MY CONCLUSION:
1. to get rid of useless stuffs from the system (which also bad for the sat reception), use
[ROM] [WWE]RAPHAEL WM6.1 V05 - Build 20758.1.4.1 - pdaviet.net (http://forum.xda-developers.com/showthread.php?t=438488&highlight=20758.1.4.1) and
[RADIO] Raphael Radio 1.02.25.28 (http://forum.xda-developers.com/showthread.php?t=437219)
(before installing these ROM, the first step should be installing the HARD-SPL to save your device (http://forum.xda-developers.com/showthread.php?t=410150)
2. to get a really quick fix,
take the battery out, as mentioned above, for a few minutes, to clear the GPS memory. (You can try for example igo now, you will see nothing on the sat view at this time...).
After that download QuickGPS data (the ROM I mentioned has a correct quickGPS data download!) If you see iGO sat view this time, you will see a lot of red dots immediately, and after a not too long waiting, you'll have fix)
3. to make the GPS even faster,
adjust the baudrate to 38400 (with comport4) in your navigation program. (But leave port4 and baud4800 settings in the Windows Mobile settings!!)
4. there could be some navi software, which couldn't get fix (iGO yes, for example Googlemapsmobile not every time), use GPStest (http://www.chartcross.co.uk/products/WM00002_details.asp), select com4 and baud 38400 also, minimize, and run the other navi. It WILL work!
5. after getting a fast signal, in order to get rid of lag while navigating, use this excellent method from the forum:
A. With Advanced config 3.2 http://www.touchxperience.com/fr/out...ol-32-cab.html
1. disable A-GPS
2. disable GPS logging
3. logfile name must be empty
4. old logfile name must be empty
5. maximum size of logfile must be 0
6 delete the files : \windows\GPSLogFile.txt and \windows\GPSLogFileBack.txt
Under: HKLM\SYSTEM\CurrentControlSet\GPS Intermediate Driver\
- Drivers\GpsOneDevice\PollInterval -> 100 (default is 1000)
- Drivers\InputBufferSize -> 256 (EDITED!!) (default is 4096)
- Drivers\OutputBufferSize -> 256 (EDITED!!) (default is 4096)
- Drivers\SleepOnNoData -> 100 (default is 1000)
- Multiplexer\MaxBufferSize -> 256 (EDITED!!) (by default not present, you have to create it)
AND: from
[HKEY_LOCAL_MACHINE\System\CurrentControlSet\GPS Intermediate Driver] leave blank the followings (delete the data from inside):
"CurrentLogFile"=" "
"OldLogFile"=" "
With these settings, I have a really cool reception!!
One more thing: if you feel that the reception going worse, do the battery-taking-quickgps-using trick again!
FOR ALL THESE INFOS THANKS FOR THE DIFFERENT MEMBERS OF THIS FORUM!!!!!
(ps: some series ((series "2" ?!)) have no any gps reception problem. If you have this series, be happy, and don't do anything...)
ps2: I've tried other methods also from the forum (for example Nadavi HTC GPS tool, etc.), and made hard reset on my phone many many times to try other methods, but the mix of these methods seems to work, the other ones are not really necessary. So this mix has no contains all the possible solutions from the forum, and this was not the way of selection without thinking..
g2tl said:
Isn't the quickgps tool for agps?
So either the disabling of agps or the quickgps is not necessary
Click to expand...
Click to collapse
as I know quickgps and agps are quite different
-quickgps is downloading the latest sat positions to get faster fix
-agps is using phone cell informations also for positioning, besides the satellites
you NEED quickGPS, DO NOT USE aGPS!!!

[KERNEL] data streaming/VOIP in standby, wifi module for HTC's OTA 2.6.35.10 kernel

Edit 31/07/2011
Updated diff files for gingerbread kernels (HTC and AOSP), and a module for the more recent HTC kernel (2.6.35.10-g308ba26) can be found later in the thread Here.
The current issue...
Note: This isn't related to the sleep mode wifi disconnections caused by MAC addresses that don't begin 00, so far as i'm aware that's been patched for a while now. Wifi remains connected fine in standby mode, but the throughput is only suitable for very light traffic.
Since compiling the cifs module for the HTC official 2.2 kernel on monday i hit a pretty massive problem when listening to music over wifi from a network share - when the screen enters standby, bandwidth drops massively, well below 100kbps. This makes music you're playing from a cifs share over wifi stutter and cut out and basically become pointless.
A little googling and searching these forums i came to find this has been a fairly massive bugbear of the nexus one and htc desire? Any kind of normal bandwidth streaming over wifi is practically useless once that screen goes into standby. Ditto for high speed transfer of large files.
That means difficult VOIP calls, and no internet radio at a decent bitrate, or cifs access to your music library.
Finding a solution...
As luck would have it, the HTC Incredible 2.6.32.15 source code was released on tuesday, and as it's basically the non-GSM variant of the Desire, i was fairly certain the HTC customised bcm4329 source code was the same as the Desire's, and it seems like it is. So armed with HTC's customised bcm4329 module sourcecode and the output of dmesg when the phone was in various states, i had a poke around.
First of all i tried editing the DTIM update interval, so the phone would wait fewer beacons to collect incoming packets while the wifi was in sleep mode, and this improved the situation, barely. In the end i just locked out the phone's deepest wifi sleep mode, PM_MAX, in favour of PM_FAST, the low power mode for wifi used when the screen is active but the phone is relying on the battery alone (There is a third mode, PM_OFF, used when the phone is powered by a usb connection, but as the name implies, all power management is disabled by it, so forget that one).
PM_MAX appears to have a very fixed latency depending upon the DTIM update interval (300ms with DTIM as the default 3), and severely limited bandwidth. PM_FAST appears to be a listen mode or active rx mode, with latency scaling to the bandwidth used. An initial ping at idle will be up to 1000ms (or even no reply) before steadying at 100-300ms. During light streaming it's more of a steady 50-70ms, and during a file transfer it's a very low 5ms. So it looks like PM_FAST keeps the wifi receiver powered up to some extent, and only powers up the transmitter when needed.
You can guess the edit i've made: Essentially my new kernel module tells the phone to keep wifi in low power mode when the screen is switched off, rather than in a full sleep mode.
Just give me the damn kernel module...
Downloads are attached to the bottom of this post.
Files are unsigned, and install with clockworkmod recovery's 'Install zip from sd card' option.
Your wifi driver must match your phone's kernel version. To check your kernel version go to Settings - About Phone - Software Information and look at Kernel version.
Files including _N_ in the name have wifi mode N support for HTC kernels (non HTC kernel versions should include mode N by default).
If you don't see a version of the module that matches your kernel version, post a request in this thread.
The results..
Well obviously now the wifi bandwidth when the screen sleeps is the same as when the screen is active, so streaming data during sleep mode is perfect, wonderful even! However, i was a little concerned with the effect on battery usage when the connection is alive but fully or partially idle.
So i just did some practical testing. Note i'm connected using 802.11g, my router doesn't support mode n, and my signal is a less than ideal -80dBm so i'm probably using a little extra power.
With the original HTC wifi module, my phone lost about 1% charge an hour with wifi permanently enabled, push gmail enabled, and sync for weather and contacts every 15 minutes.
With the new wifi module, and the same conditions (wifi always on, gmail push, sync 15 mins), i lose around 1% an hour still. Excellent, no difference.
60 minutes of 200kbit VBR mp3 playback through the speaker with the screen off used about 8% of my battery.
60 minutes of 800kbit FLAC playback through the speaker with the screen off used up 10% of my battery.
volume was set at roughly 75% for both tests.
I didn't do any testing with the earphones, but i suspect they'd be marginally lower on the power usage.
To conclude, from my brief testing this afternoon i couldn't discern a difference in battery usage when the wifi connection is left on but mostly idle, and when data is streaming it doesn't eat the entire battery charge in minutes.
Note to ROM authors
Any rom authors using 2.6.32.15 who would like to include this module in their rom, you're quite welcome to. You don't need to credit me either, although it would be nice.
This looks like it could be interesting, but I downloaded and tried to install the b/g/n version and Clockwork spits it out, aborting it as a bad zip.
Anyone else succeed in applying this mod?
Quist said:
This looks like it could be interesting, but I downloaded and tried to install the b/g/n version and Clockwork spits it out, aborting it as a bad zip.
Anyone else succeed in applying this mod?
Click to expand...
Click to collapse
Disable sign check?
xmoo said:
Disable sign check?
Click to expand...
Click to collapse
Signature check was disabled. I'm running an unsigned version of Paul's v8 WIP rom (lightning fast by the way).
Anyway, double checked and manually toggled the sign check on and off again, still same results:
Can't open...
(bad)
Installation aborted.
I just downloaded the b/g/n file and rebooted to clockworkmod and it applied without error.
Possibly you have a bad download? does the zip file open on your computer?
I'd recommend you download again, open the zip on your pc to check it downloaded properly, then copy it to your phone. Let me know how you get on.
Edit:
Just reuploaded without the signature (and erroneous boot.img stuffed in each zip by Kitchen), i downloaded and installed both versions to be sure they work. Links are updated in first post.
Works great.
I use "Twonky Server" to view pictures on my TV and the connection was lost as soon as the phone went to sleep mode. No problem now!
Thanks.
Anymore information about g or n mode?
Im still unsure if it would be good to enable n mode... there were some reports about significant higher battery drain.
Can someone compare the g and n versions running on a stock-near rom?
Thanks in advance!
Myshkinbob said:
I just downloaded the b/g/n file and rebooted to clockworkmod and it applied without error.
Possibly you have a bad download? does the zip file open on your computer?
I'd recommend you download again, open the zip on your pc to check it downloaded properly, then copy it to your phone. Let me know how you get on.
Edit:
Just reuploaded without the signature (and erroneous boot.img stuffed in each zip by Kitchen), i downloaded and installed both versions to be sure they work. Links are updated in first post.
Click to expand...
Click to collapse
Before you re-uploaded it, I noticed that when I downloaded it directly to the phone, it was clipping it at 47 kb (both the standard browser and Dolphin). Apparently Rapidshare is doing something strange through the phone browsers.
Anyway, it worked fine even before you re-uploaded. Good one!
Now I just have to figure out why I keep dropping the wifi connection after a while (didn't do that before).
May I ask you such you are using to have those incredible values for battery?
1% with WiFi and Sync (over 3g?)is a dream for me !
Thanks.
Sent from my HTC Desire using XDA App
sjauquet said:
May I ask you such you are using to have those incredible values for battery?
1% with WiFi and Sync (over 3g?)is a dream for me !
Thanks.
Sent from my HTC Desire using XDA App
Click to expand...
Click to collapse
The wifi signal is only -80dBm, but barely used at idle. My cell signal is amazing though (the cell tower is 100 yards from my flat), and i run the phone in 2G mode with no mobile data. My sync is done over the wifi, not the cell network.
Nice. This is exactly what I want.
Have you looked at 2.6.34.1 kernel by the way? It is the one that most AOSP ROMs in this forum build on and I do not think I have met the latency problems with those ROMs.
tested and works great my friend
A bit more info on the battery usage in this wifi power mode...
I've been out all day and took my other phone, i left the Desire on the desk, 2G 5 bar signal, wifi always on, sync enabled for gmail and weather. Battery was charged to 100 then unplugged.
After 16 hours, battery is at 93, and battery usage breaks down to:
cell standby: 34%
wifi: 33%
phone idle: 33%
That's roughly 2.5% battery usage over 16 hours, from an idling wifi always on connection just doing 15 minute syncs all day, on a fairly poor -80dBm wireless-g network.
I think it's safe to say the power usage differences are negligable compared to the stock wifi module.
Regarding the 2.6.34.2 kernels, i'm not entirely sure if i'd look at them, since they don't include HTC's own kernel changes, i'm a little wary of them working properly with all the features. Part of it is i like the sense-ui too, it seems responsive, and i'm not lacking for free ram, so i'm not sure why people dislike it so much?
I just tried to reproduce the initial effect with the original WiFi module and some music streams. Switching off the screen actually sets my ROM into powersave mode at 245MHz only. Still, I did get smooth streaming without any problems.
Is there any good way to recreate and test this issue? I simply seem not to have it.
Thanks!
I don't suppose this will work on 2.6.34 kernels, will it? I'm trying to find a fix for OpenDesire (CM6 based)...
bemymonkey said:
I don't suppose this will work on 2.6.34 kernels, will it? I'm trying to find a fix for OpenDesire (CM6 based)...
Click to expand...
Click to collapse
I'm afraid not, since the module kernel version has to match the running kernel version.
The HTC roms also use a modified wifi driver with some of their own custom code added for VoIP and a few other things, so i'm not entirely sure the sleep functions are the same in the stock bcm4329 (wifi chipset) driver.
I'll check out the opendesire kernel source and post a patch file to apply to the source if it's possible though.
Myshkinbob said:
I'll check out the opendesire kernel source and post a patch file to apply to the source if it's possible though.
Click to expand...
Click to collapse
That would be AWESOME. Might as well patch Cyanogenmod as you're at it, as all the other AOSP ROMs (including OpenDesire) will receive those changes down the line
I should warn you though: My request has already been met with resistance... some people seem to think that this is actually the way WiFi should work in standby (i.e. it should be connected, but shouldn't actually be able to do anything ). Is a toggle setting (between PSP and CAM) in any way possible?
I thought about a toggle when i first made the patch, but then i'd have to learn to add a /proc interface to the driver to do that. I'm far from a kernel guru, and hadn't touched kernel code in years till i got my Desire last month.
I had trouble locating the new git for opendesire, the old one appears to be closed. So i got hold of the teamdesire 2.6.34.5 CM6 source instead, i expect the two are identical or very close to being so.
While i was looking for the opendesire kernel source git in the rom thread, i noticed the hostility toward altering the wifi power management mode, but at least for the HTC-modified driver it's not a problem. I leave my wifi connected whenever i'm at home, and my results for overnight standby are hard to tell apart. I lose between 5-8% battery over 12 hours of standby depending on txts and pushed emails received.
Here's a patch for the CM6 stock bcm4329 driver. It's power management code is a little different to the HTC-written driver, but it should work fine. I don't really want to wipe my phone to test it on opendesire, so one thing you'll want to test is standby battery usage. I get 6-8mA standby according to CurrentWidget.
To apply it, cd into <kernel source root dir>/drivers/net/wireless/ and copy the .diff there, then do
Code:
patch -p0 < bcm4329_pmfast.diff
then it just needs a make modules with the crosscompile toolchain etc.
Let me know how you get on.
Myshkinbob said:
Here's a patch for the CM6 stock bcm4329 driver. It's power management code is a little different to the HTC-written driver, but it should work fine. I don't really want to wipe my phone to test it on opendesire, so one thing you'll want to test is standby battery usage. I get 6-8mA standby according to CurrentWidget.
To apply it, cd into <kernel source root dir>/drivers/net/wireless/ and copy the .diff there, then do
Code:
patch -p0 < bcm4329_pmfast.diff
then it just needs a make modules with the crosscompile toolchain etc.
Let me know how you get on.
Click to expand...
Click to collapse
I guess I didn't read your previous post carefully enough - thought you were talking about just posting a patch in the form of an update.zip, or a replacement driver file to push via ADB in Recovery...
I think this is a bit over my head . You may not be a kernel guru, but you're talking to an average Windows user here
Let's see if I understood this correctly: This actually patches the source code of the wireless driver by adding/removing code, before compilation? Hmmm, I'd need to learn how to compile Android first, I guess (my experience with compilation is limited to hitting the "Compile & Run" button in Visual Studio)...
Thanks for the hard work though. I'll hit up AdamG and Kali and see if they feel like integrating this...
Oh, sorry
if you're familiar with adb, do an adb shell and then in the console do...
Code:
cat /proc/config.gz > /sdcard/config.gz
then attach that to a reply here. I need the 2.6.34 config to build the module here, then i can just post a zip to apply in the recovery console.

wifi sleep policy improvement

Hi all
Thinking about the Wifi Sleep Policy settings:
I could not find the exact details of what the different options mean, but with a little common sense:
a) When screen turns off :
- if you have a wifi connection established : When the screen times out, it will turn off wifi and swith to mobile data.
- if you do not have a wifi connection established : The phone is constantly scanning for known wifi networks. When the screen times out, it will turn off wifi and stop all scanning.
b) Never :
- if you have a wifi connection established : when the screen times out the phone will keep that connection.
- if you do not have a wifi connection established : The phone is scanning. When the screen times out, it will keep wifi ON and continue to scan permanently for known networks.
c) Never when plugged in: it is not very clear, but would say its b) if the phone is charging, and a) if the phone is not charging.
Thinking a bit about this, it appears to me that none of this options is the ideal situation.
I would say that the ideal situation is that if I have a wifi connection established, then keep wifi on and the connection for as long as the network is there.
But if I do not have a connection established, I would like the phone to turn off the wifi interface, and turn off all scanning so as not to drain the battery while I am on the streets away from home. Then only begin scanning when I turn the screen back on, and if it no wifi network to connect, then turn off wifi again when the screen times out.
In other words, I think the best situation is a mix of behaviours of a) when there is no wifi connection established, and b) if there is a wifi connection established.
Does this make sense ? or am I crazy? I'd like your opinions please.
Anybody thought about this before? Is there any app or mod that will implement this behavior ?
Thanks
Gonzalo
gbalerdi said:
Hi all
Thinking about the Wifi Sleep Policy settings:
I could not find the exact details of what the different options mean, but with a little common sense:
a) When screen turns off :
- if you have a wifi connection established : When the screen times out, it will turn off wifi and swith to mobile data.
- if you do not have a wifi connection established : The phone is constantly scanning for known wifi networks. When the screen times out, it will turn off wifi and stop all scanning.
b) Never :
- if you have a wifi connection established : when the screen times out the phone will keep that connection.
- if you do not have a wifi connection established : The phone is scanning. When the screen times out, it will keep wifi ON and continue to scan permanently for known networks.
c) Never when plugged in: it is not very clear, but would say its b) if the phone is charging, and a) if the phone is not charging.
Thinking a bit about this, it appears to me that none of this options is the ideal situation.
I would say that the ideal situation is that if I have a wifi connection established, then keep wifi on and the connection for as long as the network is there.
But if I do not have a connection established, I would like the phone to turn off the wifi interface, and turn off all scanning so as not to drain the battery while I am on the streets away from home. Then only begin scanning when I turn the screen back on, and if it no wifi network to connect, then turn off wifi again when the screen times out.
In other words, I think the best situation is a mix of behaviours of a) when there is no wifi connection established, and b) if there is a wifi connection established.
Does this make sense ? or am I crazy? I'd like your opinions please.
Anybody thought about this before? Is there any app or mod that will implement this behavior ?
Thanks
Gonzalo
Click to expand...
Click to collapse
Sleep Policy is pretty simple:
Never means the wifi radio is never turned off. This is default now on Honeycomb and up, with fairly good reason.
When screen is off means.. the radio is turned off when the screen goes off regardless of whether you have a connection.
Only when plugged in means that the wifi radio is only turned off when the screen is off and it's not plugged in.
Sounds like you want Tasker http://tasker.dinglisch.net/
It will let you do what you describe.
Juice Defender will, as well, but it's terrible terrible software.
As to your problem, on most devices, scanning is usually reduced quite a bit when the screen is off. It's not as much of a battery drain as people think.
I'm in the habit of disabling wifi when I'm not at home or in the office. There's really no reason to leave it on when you're not going to be using it unless you want open network notification.
Thank you , I didn't know that the default is now Never. I'll set it this way.
And I guess I don't really have a real life problem, it's just that it got me thinking and came to the conclusions above.
And thanks for the link to the app, I will check it out.
Gonza
edit: oh I see you are the wifi fixer dev. Thanks so much for both replies.
Is your phone rooted? Do you have root explorer and sqlite installed. Then you can have more control over the time wifi goes to sleep after screen is turned off.
BUT: it remains on for 15 minutes by default.
Now start root explorer. Navigate to
/data/data/com.android.providers.settings/databases and open settings.db
If you have sqlite installed you seesome tables. Open secure table.
Add New record. name is wifi_idle_ms and give it a value in milliseconds you want.
Default (if wifi_idle_ms is not present) is 15*60*1000 ms. I have 60000 (1 minute)
Cheers
I use juice defender. Why is it such a terrible software? I've not used others, so I have no means of comparison. Perhaps you can enlighten me?
ZanshinG1 said:
Sleep Policy is pretty simple:
Never means the wifi radio is never turned off. This is default now on Honeycomb and up, with fairly good reason.
When screen is off means.. the radio is turned off when the screen goes off regardless of whether you have a connection.
Only when plugged in means that the wifi radio is only turned off when the screen is off and it's not plugged in.
Sounds like you want Tasker http://tasker.dinglisch.net/
It will let you do what you describe.
Juice Defender will, as well, but it's terrible terrible software.
As to your problem, on most devices, scanning is usually reduced quite a bit when the screen is off. It's not as much of a battery drain as people think.
I'm in the habit of disabling wifi when I'm not at home or in the office. There's really no reason to leave it on when you're not going to be using it unless you want open network notification.
Click to expand...
Click to collapse
androidapk.net said:
I use juice defender. Why is it such a terrible software? I've not used others, so I have no means of comparison. Perhaps you can enlighten me?
Click to expand...
Click to collapse
I'd like to hear that opinion too
JD is not terrible terrible. It is terrific terrific. Honest, it really work flawlessy and it does pretty well what the OP wants. Just for the sake of it, I'm trying to replicate with tasker what JD does, but, you know... it just does it better.
tweakradje said:
Is your phone rooted? Do you have root explorer and sqlite installed. Then you can have more control over the time wifi goes to sleep after screen is turned off.
BUT: it remains on for 15 minutes by default.
Now start root explorer. Navigate to
/data/data/com.android.providers.settings/databases and open settings.db
If you have sqlite installed you seesome tables. Open secure table.
Add New record. name is wifi_idle_ms and give it a value in milliseconds you want.
Default (if wifi_idle_ms is not present) is 15*60*1000 ms. I have 60000 (1 minute)
Cheers
Click to expand...
Click to collapse
Thank you! Yes I am rooted, I am running CM7 latest kang, but I did not know all this.
Gonna give it a try
Interesting. It works indeed! If one knows a bit about Unix shell, I did the modification via adb shell. Obviously the phone must be rooted and have sqlite3 (the binary) installed.
This is how to do it:
- connect the phone via USB to the PC
- execute "adb shell" (adb must be installed) from the DOS prompt. You are into the phone
Then execute the following steps:
Code:
su
cd /data/data/com.android.providers.settings/databases
cp settings.db settings.db.ORIG
sqlite3 settings.db
Now you are "inside" the db. Type ("sqlite>" is just the prompt):
Code:
sqlite> select * from secure;
and check if wifi_idle_ms is not present. If it isn't, as it shouldn't, type:
Code:
sqlite> insert into secure values(60,'wifi_idle_ms',30000);
where instead of 60 you can write any number that is not used already in the list, being that the unique key in the table.
I used 30000 (30000 microseconds = 30 ms) but one can put the value s/he prefers.
Miche1asso said:
Interesting. It works indeed! If one knows a bit about Unix shell, I did the modification via adb shell. Obviously the phone must be rooted and have sqlite3 (the binary) installed.
This is how to do it:
- connect the phone via USB to the PC
- execute "adb shell" (adb must be installed) from the DOS prompt. You are into the phone
Then execute the following steps:
Code:
su
cd /data/data/com.android.providers.settings/databases
cp settings.db settings.db.ORIG
sqlite3 settings.db
Now you are "inside" the db. Type ("sqlite>" is just the prompt):
Code:
sqlite> select * from secure;
and check if wifi_idle_ms is not present. If it isn't, as it shouldn't, type:
Code:
sqlite> insert into secure values(60,'wifi_idle_ms',30000);
where instead of 60 you can write any number that is not used already in the list, being that the unique key in the table.
I used 30000 (30000 microseconds = 30 ms) but one can put the value s/he prefers.
Click to expand...
Click to collapse
Thanks to Miche1asso for the commands, you've made my day :good::fingers-crossed:
This thread is good, useful, it did help up reduce the default wifi idle timeout :good: Just tested on xperia sola, perfect on my custom idea idle timeout (to 120000ms)

Very Slow/Inconsistent Notifications

Hi, so the simplest way of stating my problem is when my device is either unlocked (AKA browsing apps and whatnot with the screen on), or locked (sleeping), notifications are slow to arrive. Right now I've been testing with Skype, and Gmail, both of which are inconsistent. My OPO does not currently have a SIM card (haven't setup the plan yet), and I've tried with both LTE and Radio off (using *#*#4636#*#* on the phone info page)
Info about my OPO/Setup:
Flashed CyanogenMod 12.1 as stated in the Sultan thread before flashing Sultanxda Rom
Using Sultanxda Unofficial CyanogenMod 12.1 Sep 14, 2015 (http://forum.xda-developers.com/one...m-kernel-unofficial-cyanogenmod-12-1-t3120259)
Using AK v.020 CAF Kernel (http://forum.xda-developers.com/oneplus-one/orig-development/kernel-ak-t3189886)
Using StyloKing v1 Synapse profile for CAF (http://forum.xda-developers.com/one...ad-oneplus-t3163211/post62600798#post62600798)
Using build.prop mods (http://forum.xda-developers.com/oneplus-one/general/build-prop-tweaks-device-t3168207)
Using TK Gapps Pico (http://forum.xda-developers.com/android/software/tk-gapps-t3116347)
Using Oxygen OS Modem ONLY (http://forum.xda-developers.com/one...dio-updated-modem-radio-files-oxygen-t3072516)
Using Xposed for Lolipop (http://forum.xda-developers.com/showthread.php?t=3034811)
Using Greenify and Amplify with settings from both of these threads (http://forum.xda-developers.com/android/general/guide-0-0-hour-idle-battery-drain-stock-t2973588 & http://forum.xda-developers.com/android/general/guide-extreme-battery-life-t3095884)
My OPO does not currently have a SIM card (haven't setup the plan yet)
Things I've tried:
With both LTE and Radio off (using *#*#4636#*#* on the phone info page)
Older Skype versions
Push Notification Fixer (https://play.google.com/store/apps/details?id=com.andqlimax.pushfixer&hl=en)
Disabling Amplify by disabling the module in Xposed, and rebooting
Double checked that Greenify isn't touching Skype or anything related
I also have a Samsung Galaxy S3 that is using the stock ROM (Android 4.1.2) with root permission and a custom recovery. That is the device I've been using to test with. If I send a Skype or Gmail message from my OPO to the Galaxy, it takes 2-3 seconds EVERY TIME consistently.
If I send a Skype or Gmail message from the Galaxy to the OPO, sometimes I'll get the message within 3-5 seconds, other times it takes 1 minute+ and even a few messages don't arrive until I unlock the device and open Skype or Gmail.
When I leave Skype open, and lock my OPO, it seems to get the messages very quickly almost every time.
Also, it appears that when my phone sleeps for a certain period of time, seems inconsistent, the WiFi will disconnect - and I DO have keep WiFi on while phone is sleeping.
I've run out of ideas short of trying a full wipe and different rom/kernel/who knows what at this point...=\
Any/all help is greatly appreciated. Thanks.
Edit/Update: Found this post (http://forum.cyanogenmod.org/topic/107023-cm12121-wifi-connectivity-issuesdrops-workaround-included/) on another forum, and that seems to have fixed it for me. Doing further testing, but so far positive results.
The delayed response is most probablye due to the Greenify/Amplify tweaks that you've made. Remove everything related to Gmail, Skype, Google Play Services from Greenify & Amplify and then test. Add back the tweaks one by one and find the culprit.
I've already tried that. Fully disabling Amplify, and double checking that I've touched nothing close to Skype or Gmail or Google Services in Greenify's settings.
As I put in the edit in my first post, I found out going to /system/etc/wifi/WCNSS_qcom_cfg.ini and changing the line
Code:
McastBcastFilter=3
to
Code:
McastBcastFilter=0
appears to have solved the issue for me.
But when I did that, I also noticed something else, and I wish I hadn't unplugged my device (AKA wiping battery data so I could show a clear screenshot of what I mean), because I saw when I was looking at the battery usage chart, all day long while I had McastBcastFilter=3, my WiFi signal was turning on and off CONSTANTLY throughout the day, and sometimes for decently long periods. Now that I've changed McastBcastFilter=0, my WiFi signal shows that it's on 24/7 in my battery screen.
I'm worried that this is going to cost me a fair bit more battery, but at the same time I need my notifications.
I found another setting in that file with this comment:
Code:
# Enable suspend or not
# 1: Enable standby, 2: Enable Deep sleep, 3: Enable Mcast/Bcast Filter
gEnableSuspend=3
Before I made any changes to this file, it was set to 3, which makes sense with the McastBcastFilter changes I made. But attempting to get some more battery life so my WiFi goes into some sort of partial power save mode for battery longevity, I'm going to try my OPO with Enable Deep Sleep on all day. We'll see how the battery fairs and if notifications are still delayed or not.
Edit: I was able to get the data from my secondary battery app so screenshots for all!
This is what my Wifi/Battery was like before I changed McastBcastFilter at all:
Code:
imgur.com/ZXmFYhI
This is after I changed it from 3 to 0:
Code:
imgur.com/H5ZlXc2
And this is what it's been like since I rebooted last night with gEnableSuspend=2 (AKA Deep Sleep):
Code:
imgur.com/2O6GuVE
Obviously when the WiFi is off, I'm not getting my notifications. >=\
(Gave the URLs CODE tags so I could link the images...)
Well so far I haven't seem to have come across any issues with leaving BcastMcast at 0 and using Deep Sleep for WiFi. I'm still getting all my notifications, and as far as I can tell the battery life is give or take about the same. Possibly a little more battery drain than with BcastMcast on, but I haven't done any sleep tests yet, just initial observations with using OPO in a daily setting.

Apps slowly loading or disconnecting in custom roms

My device is redmi note 8 pro, and I have flashed corvus, havoc, and evolution x. All three load app which depend on internet slow to load and/or even disconnecting. Especially for call of duty mobile, was the app that loads longer, reconnects and even disconnects (leading to me having to restart and login again). I'm comparing this to stock miui 12 rom, global version. I'm not sure is it related to kernel, but all roms have different kernels and all have this problem. Maybe its a related to mediatek chips? Or is there some thing i can install that improves connectivity? Hope someone can shed light on this. Thank you
Try to increase the time between Wi-Fi network scans
By default, Android scans for nearby Wi-Fi networks every 15 seconds when you have Wi-Fi enabled. This helps when you're moving around between multiple access points, but for the most part, the high interval only serves to drain your battery and slow down over-all performance. If you'd rather have your phone or tablet wait for 5 minutes before it looks for nearby Wi-Fi networks again, add the following line to the end of your Android's system file called build.prop:
Code:
wifi.supplicant_scan_interval=300
Try to increase system's TCP buffers size
That can help increase the speed of a slow Internet connection, especially when using mobile networks. You do so by adding these lines to build.prop file:
Code:
net.tcp.buffersize.default = 4096,87380,256960,4096,16384,256960
net.tcp.buffersize.wifi = 4096.87380,256960,4096,16384,256960
net.tcp.buffersize.umts = 4096,87380,256960,4096,16384,256960
net.tcp.buffersize.gprs = 4096.87380,256960,4096,16384,256960
net.tcp.buffersize.edge = 4096,87380,256960,4096,16384,256960
Also, prescribing Cloudflare’s DNS servers in some cases reduces the ping time. You again alter build.prop accordingly:
Code:
net.rmnet0.dns1 = 1.1.1.1
net.rmnet0.dns2 = 1.0.0.1
net.dns1 = 1.1.1.1
net.dns2 = 1.0.0.1
Wow a reply! Thanks a lot! Wells I'm back on my stock rom now, but basing on your recommendations, are you saying custom roms have settings that are slow/standard? I'm still new to the custom rom thing and a noob at entering codes... But yea I used call of duty mobile as my benchmark since thats my go to game. For miui stock I'd say loads the game in 30 seconds, where as custom roms loads it in 1min plus... I'm using home internet all by myself and speed is good.
Also one more point is that when I turn off the screen, then when turn it back on again (in like a few minutes) and re enter the app, on custom roms, its like its completely disconnected from the internet, it cannot load anymore unless i restart the app... but on miui it loads for around 5 seconds and its reconnected. I think I'm repeating myself here. but yea. thanks!

Categories

Resources