• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Root [KERNEL] MiRaGe - for HTC stock ICS ROM - 3.0.101 - 10222013

The new build of MiRaGe is available
- Synced with Linux 3.0.77
- Enabled hardware assisted crypto engine in the kernel
- minor updates from CAF
- kernel_based mpdecision is NOT included after testing results

The download link is in the OP.
Enjoy!
 
I am about to download 77. Tell me about the cryoto engine. How much and which data is processed by it?

edit: flashed and rebooted without any problem. time to enjoy.
 
I have linked the short document for qcrypto below.
https://github.com/mrg666/android_kernel_shooter/blob/master/Documentation/crypto/msm/qcrypto.txt

The following is from our actual dmesg that lists the functions of qcrypto
<6>[ 5.419555] qcrypto qcrypto.0: qcrypto-ecb-aes
<6>[ 5.419891] qcrypto qcrypto.0: qcrypto-cbc-aes
<6>[ 5.420349] qcrypto qcrypto.0: qcrypto-ctr-aes
<6>[ 5.420684] qcrypto qcrypto.0: qcrypto-ecb-des
<6>[ 5.426483] qcrypto qcrypto.0: qcrypto-cbc-des
<6>[ 5.426666] qcrypto qcrypto.0: qcrypto-ecb-3des
<6>[ 5.431854] qcrypto qcrypto.0: qcrypto-cbc-3des
<6>[ 5.432037] qcrypto qcrypto.0: qcrypto-sha1
<6>[ 5.432189] qcrypto qcrypto.0: qcrypto-sha256

Regarding S2W, we had a discussion on this at the XDA thread. The feedback about S2W was mixed and I am one of the naysayers :) One reason I am developing my own kernel is not to have a bloated kernel with so many features, governors, schedulers that I don't need. I want to keep it that way. I wish I could make everyone happy but I will not include S2W.

I can try to help you with S2W addiction if you want :D
 
mrg666 said:
I have linked the short document for qcrypto below.
https://github.com/mrg666/android_kernel_shooter/blob/master/Documentation/crypto/msm/qcrypto.txt

The following is from our actual dmesg that lists the functions of qcrypto
<6>[ 5.419555] qcrypto qcrypto.0: qcrypto-ecb-aes
<6>[ 5.419891] qcrypto qcrypto.0: qcrypto-cbc-aes
<6>[ 5.420349] qcrypto qcrypto.0: qcrypto-ctr-aes
<6>[ 5.420684] qcrypto qcrypto.0: qcrypto-ecb-des
<6>[ 5.426483] qcrypto qcrypto.0: qcrypto-cbc-des
<6>[ 5.426666] qcrypto qcrypto.0: qcrypto-ecb-3des
<6>[ 5.431854] qcrypto qcrypto.0: qcrypto-cbc-3des
<6>[ 5.432037] qcrypto qcrypto.0: qcrypto-sha1
<6>[ 5.432189] qcrypto qcrypto.0: qcrypto-sha256

Regarding S2W, we had a discussion on this at the XDA thread. The feedback about S2W was mixed and I am one of the naysayers :) One reason I am developing my own kernel is not to have a bloated kernel with so many features, governors, schedulers that I don't need. I want to keep it that way. I wish I could make everyone happy but I will not include S2W.

I can try to help you with S2W addiction if you want :D
Click to expand...

Sounds good. I respect your decision not to add S2W to your kernel. But my phone power button isn't easy to use so that's where S2W comes in to save me.

I will just have to dream about your kernel with S2W then. :p
 
LakersDroid said:
Sounds good. I respect your decision not to add S2W to your kernel. But my phone power button isn't easy to use so that's where S2W comes in to save me.

I will just have to dream about your kernel with S2W then. :p
Click to expand...

Did you try this electrical tape trick on the power button. I totally fixed mine. Just use even smaller than shown in the video, just enough to cover the tip of the switch. And the thin strip should not be bridging over the tip of the switch (one side connected to the phone only). I have used two layers several months ago. It is still there.
 
mrg666 said:
Did you try this electrical tape trick on the power button. I totally fixed mine. Just use even smaller than shown in the video, just enough to cover the tip of the switch. And the thin strip should not be bridging over the tip of the switch (one side connected to the phone only). I have used two layers several months ago. It is still there.
Click to expand...

Thanks! I will consider that.
 
ok so quadrant standard crashes with 1.728 performance in harmonia, mobsters hd and unlock potental why is that
 
Piznick64 said:
i tried everything idk just want to see my score lol
Click to expand...

EVO 3D CPU originally works at 1.188 GHz max. Your CPU apparently is not able to run at 1.7 GHz. Overclocking is never guaranteed and we are talking about 45% overclock at 1.728 GHz. Just set it back to 1.512.
 
The hardware crypto in the latest version really makes a difference. Battery depletion when surfing the web on wireless (network or encrypted WiFi) is greatly reduced. IDK if there is a way to make this kernel any better because it is damn good right now.

Mrg will probably find a way if there is one.

Please add a clean up the kitchen feature :)
 
I just took a peek at the github and see mrg is hard at work. LED control would be a good addition although I am setting them dim with this script:

#!/system/bin/sh
echo 2 > /sys/devices/platform/leds-pm8058/leds/button-backlight/currents
 
The new 05112013 build is available in the OP
- Synced with 3.0.78-rc1. Since the patch is tiny, I didn't wait for the final release which will be the same
- added msm thermal management to decrease overclock risks and eliminate errors in the kernel log from HTC source - "Thermal: failed to read out thermal zone 1"
- backported msm random number generator fixes from CAF msm-3.4

Enjoy!

PS: I have also tried the new Linaro gcc 4.8 toolchain. It was not ready yet. I will keep testing.

ron scubadiver said:
I just took a peek at the github and see mrg is hard at work. LED control would be a good addition although I am setting them dim with this script:

#!/system/bin/sh
echo 2 > /sys/devices/platform/leds-pm8058/leds/button-backlight/currents
Click to expand...


Edit: I have reverted the dynamic button backlight patch now since it doesn't seem to be working. If you have downloaded before this edit please download the kernel again. I forgot testing that patch. Sorry.
 
ron scubadiver said:
I flashed the latest and nothing bad happened, what else can I say?
Click to expand...

Well, that is excellent.

A kernel must be small and efficient; it must work with minimal resources since its main purpose is distributing them efficiently to the OS and apps. And it should also be transparent and unintrusive; a user shouldn't need to "tweak" the kernel.

This is what I tried to do with MiRaGe.
 
mrg666 said:
Well, that is excellent.

A kernel must be small and efficient; it must work with minimal resources since its main purpose is distributing them efficiently to the OS and apps. And it should also be transparent and unintrusive; a user shouldn't need to "tweak" the kernel.

This is what I tried to do with MiRaGe.
Click to expand...

Ran the RUU a few days ago, running stock rom with few small changes, and latest mirage build and it's absolutely great with default clock settings. I retract what I said the day u released this lol its truly nice and I got about 12hrs battery with streaming music the entire time and 4g and tons of screen time. The minimalistic simplicity is the greatest. Really cannot complain at all. Awesome work Mirage.
 
xstefen said:
Ran the RUU a few days ago, running stock rom with few small changes, and latest mirage build and it's absolutely great with default clock settings. I retract what I said the day u released this lol its truly nice and I got about 12hrs battery with streaming music the entire time and 4g and tons of screen time. The minimalistic simplicity is the greatest. Really cannot complain at all. Awesome work Mirage.
Click to expand...

I haven't noticed your post. Anyway, I am glad that MiRaGe works for you too.
 
I was reading recently about the debugfs exploit in Android and decided to test disabling the DEBUG_FS feature. This exploit is used for rooting but it is also a security risk for us. I could succeed after a lot of work due to loose coding practices of mainly CAF and Google developers. The kernel without the DEBUG_FS option is running fine on my phone now. I will keep testing but I wanted to ask if there is any feedback about the possible problems due to not having the DEBUG_FS option that I should look into. I would appreciate if you have any comments on this. Thanks!

If you also want to test, the test kernel is below.
shooter_test_kernel_782.zip
 
You must log in or register to reply here.
Back
Top Bottom