Post Reply

[DEVS ONLY][BCM21553 series] CyanogenMod 11 for BCM21553 Development Discussion

OP psyke83

10th October 2014, 03:21 AM   |  #181  
cleverior.ipul's Avatar
Senior Member
Flag Bogor
Thanks Meter: 349
 
420 posts
Join Date:Joined: Nov 2012
More
Quote:
Originally Posted by psyke83

Your kernel will break OTA updates - changing even one byte of of the boot image will guarantee this. Your kernel is compiled on a different machine, so it will have a different build date and machine hostname compared to the kernel built on Jenkins. That change alone will change the md5sum of the image, and delta updates also process partitions as a delta patch.

Remember that due to the GPLv2 license of the kernel, now that you've released a binary compilation, you're obliged to release your source modifications, no matter how minor.

Ok Sorry. i'll remove the link

If so, then the kernel cm11 totoro only from jenkins? I think if there is more than 1 cm11 kernel then the user will be able to choose to use a kernel which one they like best. This is just IMHO
The Following User Says Thank You to cleverior.ipul For This Useful Post: [ View ]
10th October 2014, 03:52 AM   |  #182  
OP Recognized Developer
Thanks Meter: 3,586
 
1,169 posts
Join Date:Joined: Mar 2011
Donate to Me
Quote:
Originally Posted by cleverior.ipul

Ok Sorry. i'll remove the link

If so, then the kernel cm11 totoro only from jenkins? I think if there is more than 1 cm11 kernel then the user will be able to choose to use a kernel which one they like best. This is just IMHO

My CM11 builds are OTA enabled, but using the OTA update function is optional. If a user wants to install a custom kernel, I'm not going to stop them... as long as they are fully aware that it breaks OTA updates.

I don't need more headaches from people (who fail to read warnings/instructions) spamming my PM inbox with dumb questions like "HELLLP I CANT UPDATE STATUS 5 WHYYY???".

If you're posting a kernel for CM11 users, you should warn in big red letters that it breaks OTA updates, and that all further support requests are not to be directed to me.
The Following 10 Users Say Thank You to psyke83 For This Useful Post: [ View ]
10th October 2014, 09:05 AM   |  #183  
cleverior.ipul's Avatar
Senior Member
Flag Bogor
Thanks Meter: 349
 
420 posts
Join Date:Joined: Nov 2012
More
Quote:
Originally Posted by psyke83

My CM11 builds are OTA enabled, but using the OTA update function is optional. If a user wants to install a custom kernel, I'm not going to stop them... as long as they are fully aware that it breaks OTA updates.

I don't need more headaches from people (who fail to read warnings/instructions) spamming my PM inbox with dumb questions like "HELLLP I CANT UPDATE STATUS 5 WHYYY???".

If you're posting a kernel for CM11 users, you should warn in big red letters that it breaks OTA updates, and that all further support requests are not to be directed to me.

What if you disable the OTA Update? So, the users will be able to use a custom kernel for cm11 according to what they want. As the development of CM11 in other devices. They have a variety of custom kernel for CM11.
The Following User Says Thank You to cleverior.ipul For This Useful Post: [ View ]
10th October 2014, 09:16 AM   |  #184  
Bhargav97's Avatar
Senior Member
Gandhinagar - INDIA
Thanks Meter: 485
 
520 posts
Join Date:Joined: Mar 2013
Donate to Me
More
Quote:
Originally Posted by cleverior.ipul

What if you disable the OTA Update? So, the users will be able to use a custom kernel for cm11 according to what they want. As the development of CM11 in other devices. They have a variety of custom kernel for CM11.

Please read the previous post by psyke83 once again, he said that OTA updates are OPTIONAL. There's no point in disabling them. If a user doesn't wish to update via OTA, he would not update. It's that simple.

But if somebody wants to update via OTA and if he flash kernel built on your local machine, it will break OTA updates for him. To regain OTA support, he would have to flash the full build from Jenkins. But, not all users know about the above fact and they would spam discussion thread and PM Inbox of psyke83 saying that they can't get OTA update to work, that's why you should write in BIG, RED words on your kernel's thread that your kernel will break their OTA updates. You can always go ahead and make your custom kernel and share it but do follow what psyke83 said.
The Following 2 Users Say Thank You to Bhargav97 For This Useful Post: [ View ]
10th October 2014, 09:30 AM   |  #185  
cleverior.ipul's Avatar
Senior Member
Flag Bogor
Thanks Meter: 349
 
420 posts
Join Date:Joined: Nov 2012
More
Quote:
Originally Posted by Bhargav97

Please read the previous post by psyke83 once again, he said that OTA updates are OPTIONAL. There's no point in disabling them. If a user doesn't wish to update via OTA, he would not update. It's that simple.

But if somebody wants to update via OTA and if he flash kernel built on your local machine, it will break OTA updates for him. To regain OTA support, he would have to flash the full build from Jenkins. But, not all users know about the above fact and they would spam discussion thread and PM Inbox of psyke83 saying that they can't get OTA update to work, that's why you should write in BIG, RED words on your kernel's thread that your kernel will break their OTA updates. You can always go ahead and make your custom kernel and share it but do follow what psyke83 said.

The Following User Says Thank You to cleverior.ipul For This Useful Post: [ View ]
11th October 2014, 07:48 AM   |  #186  
savie's Avatar
Senior Member
Thanks Meter: 621
 
846 posts
Join Date:Joined: Feb 2012
kernel bcm21553-common issue
@psyke83 i see ur kernel source and find bugs on cpufreq as @irfanbagus mention on this thread http://forum.xda-developers.com/show....php?t=1572030 i thinks thats the problem cpufreq cant running well.

i hope u can fix it on next build.

for personal im patch and build from ur source . since we still have some space (5mb limit kernel) so im build with philz recovery.

patch: https://db.tt/uTreoOj7

thanks before.
The Following 12 Users Say Thank You to savie For This Useful Post: [ View ]
12th October 2014, 09:33 AM   |  #187  
OP Recognized Developer
Thanks Meter: 3,586
 
1,169 posts
Join Date:Joined: Mar 2011
Donate to Me
Quote:
Originally Posted by savie

@psyke83 i see ur kernel source and find bugs on cpufreq as @irfanbagus mention on this thread http://forum.xda-developers.com/show....php?t=1572030 i thinks thats the problem cpufreq cant running well.

i hope u can fix it on next build.

for personal im patch and build from ur source . since we still have some space (5mb limit kernel) so im build with philz recovery.

patch: https://db.tt/uTreoOj7

thanks before.

Thanks savie. I've tried the patch and it works. Although the ondemand scheduler seems to make the device feel a bit more sluggish (probably due to the latency when switching frequencies), it's probably a good feature to have, and it's good to give users more choices.

I assume that you don't have the full androidarmv6 source downloaded, so I can submit the patch to gerrit for you. I just need to know whose name I should set in the commit field of the patch - you, @irfanbagus, or someone else I'm not aware of. I took a quick look at @irfanbagus's patch in the linked thread and it's not the same as your patch. The author field requires an e-mail, so if necessary, you can PM me those details.
Last edited by psyke83; 12th October 2014 at 09:41 AM.
The Following 7 Users Say Thank You to psyke83 For This Useful Post: [ View ]
12th October 2014, 11:09 AM   |  #188  
savie's Avatar
Senior Member
Thanks Meter: 621
 
846 posts
Join Date:Joined: Feb 2012
Quote:
Originally Posted by psyke83

Thanks savie. I've tried the patch and it works. Although the ondemand scheduler seems to make the device feel a bit more sluggish (probably due to the latency when switching frequencies), it's probably a good feature to have, and it's good to give users more choices.

I assume that you don't have the full androidarmv6 source downloaded, so I can submit the patch to gerrit for you. I just need to know whose name I should set in the commit field of the patch - you, @irfanbagus, or someone else I'm not aware of. I took a quick look at @irfanbagus's patch in the linked thread and it's not the same as your patch. The author field requires an e-mail, so if necessary, you can PM me those details.

Ondemand still buggy i thinks @kurotsugi mod that governor on his thread.

Actually i have all androidarmv6 source but too hard to me playing on huge source

Original patch founder is @irfanbagus but he forget to add 1000000 on his patch so im add it and add 1 more freq since only thats freq working well. Overclock is still not work till now.

Actually @irfanbagus have one more work its call bacemtweek to undervolt with add module. But i have other method to undervolt but still have some bugs.
The Following 5 Users Say Thank You to savie For This Useful Post: [ View ]
12th October 2014, 08:40 PM   |  #189  
kurotsugi's Avatar
Recognized Contributor
Flag yogyakarta
Thanks Meter: 1,789
 
2,523 posts
Join Date:Joined: Feb 2012
More
it's just a small modification on the parameter
Code:
*/it should be this one/*
#define DEF_FREQUENCY_UP_THRESHOLD		(80)
*/lower value means your device will use high frequency more frequent./*
you might also want to check the sampling rate. I haven't see your source code but I saw several source code have 100.000. that's too high.

The Following 5 Users Say Thank You to kurotsugi For This Useful Post: [ View ]
Post Reply Subscribe to Thread
Previous Thread Next Thread
Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes


Top Threads in Galaxy Y GT-S5360 Android Development by ThreadRank