New (very fast) lag fix using internal memory and playlogo1 hack

[Guest heredavid]

This is copied from my post over on XDA dev. It makes the phone as smooth as butter. It's faster than any of the previous lag fixes plus it doesn't have any of the issues that the other ones had like not being able to remove the memory card, or running out of space on the new partition, etc. My SGS feels as smooth as an iPhone 4 now. My only contribution is that I made this into a complete guide. It was the hard work of others that discovered this fix.

Here are the instructions:

First copy the two files in the attached .zip to the root of your internal memory then run the following commands after enabling debugging mode. Make sure you have the latest busybox installed. The first command takes a while since it creates a 1 GB file on the internal memory. I also recommend starting in stock configuration (any firmware version is fine) I don't know what will happen if other fixes/kernels have been applied.

adb shell


su


busybox dd if=/dev/zero of=/data/dataimg bs=1024 count=1048576

busybox mknod /dev/loop0 b 7 0

busybox losetup /dev/loop0 /data/dataimg

busybox mkfs.ext2 /dev/loop0

mkdir /data/data1


busybox mount -o remount,rw -t rootfs /

mv /system/bin/playlogos1 /system/bin/playlogosnow

cp /sdcard/playlogos1 /system/bin/playlogos1

cp /sdcard/userinit.sh /system/bin/userinit.sh

chmod 755 /system/bin/playlogos1

chmod 755 /system/bin/userinit.sh

busybox mount -o remount,ro -t rootfs /


reboot

IF you just start with version 6 here is what was posted to fix when it does not install:

"...yeah, you had lagfix already installed once....

adb shell

su

mv /system/bin/userinit.sh /system/bin/used

reboot

Follow all the original install commands & then once everything in the original is done..

mv /system/bin/used /system/bin/userinit.sh

reboot.."

Edited August 22, 2010 by heredavid

[Guest Chapieee]

looks like you created a large file that you haven't deleted yet.

If you reflash from Odin, and select repartition, that will certainly get you back to a clean phone. Doing a factory wipe/reset from the phone should as well. Of course you'll lose all data on the internal ROM space (not your internal sdcard).

I think you really just want to delete that very large file (/data/dataimg) you created. If you want to do it without reflash or wipe, reboot the phone, then from the adb shell, run the following commands (paste the output here if you need help interpreting):

su

ls - l /data

if /data/dataimg exists, you should just be able to:

rm /data/dataimg

rm - r /data/dataimg if the above one didn't work.

CAUTION - messing around with rm command can really screw things up if you get the syntax wrong. Of course you're only a wipe/reflash away from fixing things.

In the future PLEASE get clockwork rom manager from the Market and make a nandroid backup of your phone before trying stuff, it's so easy to just go back to the exact point you were at with that nandroid restore!

Keep in mind that the One Click authors are recommending you do it from a fresh Odin flash anyways...

Good luck.

I ran the commands you gave me and it worked like a charm! Extra space was created and the One Click app was able to perform the fix. Thank you very much!

I will get clockwork rom manager before I'll do anything like this again.

Thanks again!

[Guest VomitOmelette]

at the risk of sounding like an idiot, i am running into the following error and not sure what im doing wrong:

>adb shell

$ su

su

# busybox dd if=/dev/zero of=/data/dataimg bs=1024 count=1048576

busybox dd if=/dev/zero of=/data/dataimg bs=1024 count=1048576

1048576+0 records in

1048576+0 records out

# busybox mknod /dev/loop0 b 7 0

busybox mknod /dev/loop0 b 7 0

mknod: /dev/loop0: File exists

# busybox losetup /dev/loop0 /data/dataimg

busybox losetup /dev/loop0 /data/dataimg

# busybox mkfs.ext2 /dev/loop0

busybox mkfs.ext2 /dev/loop0

mkfs.ext2: can't format mounted filesystem

#[/codebox]

Can anyone help?

[Guest crezaee]

Great, thanks!

Quadrant 2288. :-)

I followed these instructions to a "T" after all of those commands used from busybox, etc i rebooted like it says to do at the end.

I opened up Quadrant following this lag fix and ran the full benchmark test.... I scored exactly a 900 :huh:

Please help me to understand why i only have a 900 and others are getting in the 2000's

[Guest Compuse]

Like i said earlier in a lagfix thread, it's prolly cheating on his memory in the benchmark, 2000 points here too, lag is less, much less =/

[Guest Iain51]

Just applied and Quadrant has gone from high 800's to mid 2200's. In terms of how it feels: chuffing fast! :huh:

[Guest DistortedLoop]

I followed these instructions to a "T" after all of those commands used from busybox, etc i rebooted like it says to do at the end.

I opened up Quadrant following this lag fix and ran the full benchmark test.... I scored exactly a 900 :huh:

Please help me to understand why i only have a 900 and others are getting in the 2000's

You obviously made a mistake. There's no way you successfully did the fix and get that kind of score.

You might want to just head over to xda and check out one of the one click lag fixes that automate the whole process.

An ls -l command on your /data directory might be revealing.

[Guest allareconnected]

Can anyone confirm that performing a nandroid restore from before the lag fix does indeed erase the lag fix?? Also is there a fool proof way to tell for sure if you have a lag fix installed besides quadrant speeds?? Thanks in advance!

Edited August 26, 2010 by allareconnected

[Guest DistortedLoop]

Can anyone confirm that performing a nandroid restore from before the lag fix does indeed erase the lag fix?? Also is there a fool proof way to tell for sure if you have a lag fix installed besides quadrant speeds?? Thanks in advance!

Yes, clockwork nandroid restore put my phone back to a pre-lag-fix state when I borked my first attempt.

ls - l command on the appropriate directories an easy way to tell. If you have symlinks pointing your /data/data and other impacted directories under /data instead of the actual directories, you're still on the fix.

mount command also tell if you're loop mount is in place.

[Guest Fykick]

at the risk of sounding like an idiot, i am running into the following error and not sure what im doing wrong:

>adb shell

$ su

su

# busybox dd if=/dev/zero of=/data/dataimg bs=1024 count=1048576

busybox dd if=/dev/zero of=/data/dataimg bs=1024 count=1048576

1048576+0 records in

1048576+0 records out

# busybox mknod /dev/loop0 b 7 0

busybox mknod /dev/loop0 b 7 0

mknod: /dev/loop0: File exists

# busybox losetup /dev/loop0 /data/dataimg

busybox losetup /dev/loop0 /data/dataimg

# busybox mkfs.ext2 /dev/loop0

busybox mkfs.ext2 /dev/loop0

mkfs.ext2: can't format mounted filesystem

#[/codebox]

Can anyone help?

I second this...I also get this problem the other day. I was able to finally get it after a bunch of messing around one day, but that was the one and only time I haven't gotten the error of "can't format mounted filesystem". I prefer this fix cuz it seems to be faster and less buggy. I scored a quadrant of 2812 with this fix where as the others are around 2300(with OC kernel that is for both scores). But regardless if anyone can help answer this problem, I and I'm sure the other person will be happy too.

Thanks in advance.

[Guest Pauldavid77]

Can someone tell me how to remove this fix?

[Guest septic]

wow.... i finally believe that there is a lag on the phone....

http://galaxyss.blogspot.com/2010/08/20-in...-click-fix.html

[Guest Kaiser Chief]

yah already trying it :huh:

Will see how it goes....

update: looking great, faster quadrant scores than the ext4 fix, am getting around 2,300 with this, was about 1700 with the ext4 fix. (i also did it iwth the ext4 fix kernel) but that was over a stock rom, no issues there.

So just their crappy filesystems are at the root of the problem, nice to know its not hardware...

Am hoping the same for the GPS too (well the gps hardware rocks) so hopefully they can fix that with some good firmware

Getting better and better

Nice to be able to use the internal storage!!!

How does this affect wear and tear of the internal memory?

and they set aside 2 gigs right? any reason why we only made this file 1 gig?

I prefer to avoid the internal memory and make al this read-write action on a cheap class 6 card :(

Maybe a bit quicker 2300 instead of 1950 but it´s safe and working very fine.

If somebody can give me the assurance that it does not matter I go for the extra 350 quadrant I love speed but safety first ;)

[Guest allareconnected]

Can someone provide a clear "easy" way to remove this lagfix??? I know people have asked many times in this thread and I have not seen any clear answer other than "I think"... I wiped my a handful of times now but I still seem to find the files in place and apparently the lagfix is there but not active. When I attempt to install the one click lag fix for example it states that it already sees a lag fix installed... any help would be greatly appareciated thanks in advance!!!!

Edited August 31, 2010 by allareconnected

[Guest gunquixtar]

cp /sdcard/playlogos1 /system/bin/playlogos1

It's say

Not found

Can some one help me?

Edited September 1, 2010 by gunquixtar

[Guest ragflan]

Does anyone know how I can get Stock Android Contacts, Calendar and Dialer apps onto Galaxy S?

[Guest floepie]

What differentiates this fix from RyanZA's and Tatuyama's update.zips?

The former sticks mainly with ext2, and the latter is based on the former but uses modules with the stock kernel to support ext4. Does this fix also employ an ext2 loopback file too? It seems these all do pretty much the same thing, although the OP seems to indicate that this fix is better and faster than all previous ones. Furthermore, the first page of posts seems to indicate that ext2 is NOT used, but later on, many posts do speak of using ext2.

[Guest DistortedLoop]

What differentiates this fix from RyanZA's and Tatuyama's update.zips?

The former sticks mainly with ext2, and the latter is based on the former but uses modules with the stock kernel to support ext4. Does this fix also employ an ext2 loopback file too? It seems these all do pretty much the same thing, although the OP seems to indicate that this fix is better and faster than all previous ones. Furthermore, the first page of posts seems to indicate that ext2 is NOT used, but later on, many posts do speak of using ext2.

This one is the one that RyanZA's is based on.

This one first came to most people's attention when ykk_five brought it up, but apparently curio aka supercurio had proposed this method on some French forum a few weeks before that. The French folks decided after some time that lag came back eventually with this method (I agree) and gave up on it.

Speaking of supercurio - keep an eye on this guy. First (that I know of) to come up with the loop fix, first to find the redbhend method to actually let you flash a kernel on SGS via update.zip, first to find a software fix for headphone audio improvement, and now he's first to find the real fix for lag, which is to just reformat the /data partition with ext4. He's got a semi-private beta of it that I'm using right now, and it works great after 36 hours uptime still. Should be public beta real soon.

[Guest floepie]

This one is the one that RyanZA's is based on.

This one first came to most people's attention when ykk_five brought it up, but apparently curio aka supercurio had proposed this method on some French forum a few weeks before that. The French folks decided after some time that lag came back eventually with this method (I agree) and gave up on it.

Speaking of supercurio - keep an eye on this guy. First (that I know of) to come up with the loop fix, first to find the redbhend method to actually let you flash a kernel on SGS via update.zip, first to find a software fix for headphone audio improvement, and now he's first to find the real fix for lag, which is to just reformat the /data partition with ext4. He's got a semi-private beta of it that I'm using right now, and it works great after 36 hours uptime still. Should be public beta real soon.

Great, thanks for putting that into context. I like to understand how these fixes evolve from one to the next. I'm almost wondering if Tayutama's update.zips (based on RyanZA's) aren't getting recognized. It seems his fixes have evolved from a simple re-packaging of RyanZA's batch file method into an update.zip to one which sets up a /data ext4 partition, but one that does NOT require a reflash of the kernel simply to support ext4. He does it by means of modules that run at boot.

[Guest DistortedLoop]

Great, thanks for putting that into context. I like to understand how these fixes evolve from one to the next. I'm almost wondering if Tayutama's update.zips (based on RyanZA's) aren't getting recognized. It seems his fixes have evolved from a simple re-packaging of RyanZA's batch file method into an update.zip to one which sets up a /data ext4 partition, but one that does NOT require a reflash of the kernel simply to support ext4. He does it by means of modules that run at boot.

Funny, because Tat's were supposed to simplify for the user at first, but then he added options. Sounds too complicated.

The REAL fix is supercurio voodoo as far as I can tell. Will be exciting to see it go public beta, then final. Once final, he says he'll release the source code and then all the devs can build on it.

Rasdeep is in the chat room right now talking about trying to incorporate voodoo into his overclock kernels...

[Guest loaded.check]

I seemed to have applied the lag fix correctly. (needed to use "busybox cp /sdcard...") I was able to apply all the commands and reboot without having any errors. Unfortunately I don't see any major performance improvements. Benchmark pre-lagfix 860. Benchmark post-lagfix 893. Any ideas on what may have gone wrong or how I can get the scores up into the 2000's?

[Guest jimmius]

hi all i wanted to ask if this fix is better overall including battery consumption than mimocan's fix ?

[Guest Shaggsdope]

The download doesn't seem to be working. Am I doing something wrong ¿

[Guest jpvaneijk]

Can someone plz tel me if this fix can be removerd? The question has a beard but an answer is not been given?

Used this fix on stock firmware...

Thanx,

JP

[Guest DistortedLoop]

Can someone plz tel me if this fix can be removerd? The question has a beard but an answer is not been given?

Used this fix on stock firmware...

Thanx,

JP

Yes, it can be removed. Not sure the easiest way, though. I think on xda there's probably an undo script. When I decided this one was too kludgy, I just backed up my apps with titanium, then did a factory reset and flashed a new pda with ODIN (I wanted to try a new firmware at the time anyways).