Database Info

msm7500 series overclocking

crazybyte said:
I suggest to the rom cookers, to check about Qualcomm 7201A overclock, because in another forum, someone overclocked it to 768 mhz, and it seems to be true, because it's the real speed of the output of PLL generator. Our devices has a divisor.
Here is the link:
http://www.nuerom.com/BlogEngine/post/2009/09/06/Touch-Pro-running-at-800MHZ.aspx
I think that all the problem resides in the file:
http://android.git.kernel.org/?p=ke...3;hb=ebea985a6ec461a59afb86f0068cc1d79d7c8bf1
We think that the first step it's enable cpufreq to scale to 768mhz, and testing it in this freq.The info about scalling are in the lines 227 and 228:
227 CONFIG_MSM_CPU_FREQ_ONDEMAND_MAX=384000
228 CONFIG_MSM_CPU_FREQ_ONDEMAND_MIN=245760
Click to expand...
Click to collapse
i thought this might have better set up as its own thread.
edit
i have done some mild kernel work in winmo on a 7200 and do believe this is not only possible but a decent hack worth putting time into. also no2chem is to winmo as cyanogen is to android. so its from a good source

I posted in that thread too. I think the 65nm MSM7201A in the G1 could be easily pushed to 768MHz.
coolbho3000 said:
I was thinking about this but never could get any kernel to compile because I'm stupid. The other way would be to use PLL 2 to go all the way to 1056 (probably impossible ).
It's actually not that simple. This is the file you want to modify: http://android.git.kernel.org/?p=ke...13e3ef308591e8c0e5fe7ce0311d3f4d6b0b8;hb=HEAD
Add an entry for 768000 in static struct clkctl_acpu_speed acpu_freq_tbl[] for 768000.
{ 768000, ACPU_PLL_1, 1, 0, 128000, 5, VDD_7, 128000, 0, 5, -1 }
Those should be the correct values for 768MHz - all except for the last two 5 and -1. I don't know what those do, probably something to do with scaling.
Click to expand...
Click to collapse
Then add 768000 to static struct cpufreq_frequency_table freq_table[] and you should be able to echo those values into scaling_max_freq and scaling_min_freq

although this would kill the battery in 5 minuites it is great for hero and emulators as well

If someone could pull that off it would be killer.
Do you really think it would be that big of a drain on the battery if it was only climbing that high under heavy load situations rather than being set that high at all times?
I would be willing to take a hit in battery life to gain a performance increase of that type as long as the batter life doesn't end of 1/4 of what it was.

I've got an upgraded battery, so I don't care. Give me more speed!

One note please remember this was confirmed on 7500 not 7501a also it was a cdma phone not gsm but it could work with some work

If this makes a big different in performance, I'm willing to use the brick size battery I bought. It's been laying around unused since I bought it.

i am willing to try this but i dont know which file you edit. If someone can tell me ill give it a go

I don't think it'll be that easy though

You need to recompile the kernel with the added values in clkctl_acpu_speed acpu_freq_tbl[] in http://android.git.kernel.org...b0b8;hb=HEAD
I might give it a go if you want to try it out. I think we should determine if the last two 5 and -1 numbers are important first. Does look like scaling:
Code:
78 /* Index in acpu_freq_tbl[] for steppings. */
79 short down;
80 short up;
As far as I can tell the stepping values are used here:
Code:
332 while (cur_s != tgt_s) {
333 /*
334 * Always jump to target freq if within 256mhz, regulardless of
335 * PLL. If differnece is greater, use the predefinied
336 * steppings in the table.
337 */
338 int d = abs((int)(cur_s->a11clk_khz - tgt_s->a11clk_khz));
339 if (d > drv_state.max_speed_delta_khz) {
340 /* Step up or down depending on target vs current. */
341 int clk_index = tgt_s->a11clk_khz > cur_s->a11clk_khz ?
342 cur_s->up : cur_s->down;
343 if (clk_index < 0) { /* This should not happen. */
344 printk(KERN_ERR "cur:%u target: %u\n",
345 cur_s->a11clk_khz, tgt_s->a11clk_khz);
346 rc = -EINVAL;
347 goto out;
348 }
349 cur_s = &acpu_freq_tbl[clk_index];
350 } else {
351 cur_s = tgt_s;
352 }

daproy said:
You need to recompile the kernel with the added values in clkctl_acpu_speed acpu_freq_tbl[] in http://android.git.kernel.org...b0b8;hb=HEAD
I might give it a go if you want to try it out. I think we should determine if the last two 5 and -1 numbers are important first. Does look like scaling:
Code:
78 /* Index in acpu_freq_tbl[] for steppings. */
79 short down;
80 short up;
Click to expand...
Click to collapse
yep ill try it out if you compile it btw what is the worst hat could happen to my phone?

The code is optimized to stay on the same PLL so the last two numbers are the array index in the acpu_freq_tbl[] two go down to or up to from that entry. Since our new entry is ACPU_PLL_1 the next one down is entry 6 and there are none to step up to so it should be:
{ 768000, ACPU_PLL_1, 1, 0, 128000, 5, VDD_7, 128000, 0, 6, -1 }
The step up from 384000 should also be changed now that we're adding a higher frequency on the same PLL to:
{ 384000, ACPU_PLL_1, 1, 1, 128000, 2, VDD_6, 128000, 0, 2, 8 }
So the acpu_freq_tbl[] should be:
Code:
static struct clkctl_acpu_speed acpu_freq_tbl[] = {
{ 19200, ACPU_PLL_TCXO, 0, 0, 19200, 0, VDD_0, 30720, 0, 0, 4 },
{ 122880, ACPU_PLL_0, 4, 1, 61440, 1, VDD_3, 61440, 0, 0, 4 },
{ 128000, ACPU_PLL_1, 1, 5, 64000, 1, VDD_3, 61440, 0, 0, 6 },
{ 176000, ACPU_PLL_2, 2, 5, 88000, 1, VDD_3, 61440, 0, 0, 5 },
{ 245760, ACPU_PLL_0, 4, 0, 81920, 2, VDD_4, 61440, 0, 0, 5 },
{ 352000, ACPU_PLL_2, 2, 2, 88000, 3, VDD_5, 128000, 0, 3, 7 },
{ 384000, ACPU_PLL_1, 1, 1, 128000, 2, VDD_6, 128000, 0, 2, 8 },
{ 528000, ACPU_PLL_2, 2, 1, 132000, 3, VDD_7, 128000, 0, 5, -1 },
{ 768000, ACPU_PLL_1, 1, 0, 128000, [B]5[/B], VDD_7, 128000, 0, 6, -1 }
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
};
and
Code:
static struct cpufreq_frequency_table freq_table[] = {
{ 0, 19200 },
{ 1, 122880 },
{ 2, 128000 },
{ 3, 245760 },
{ 4, 384000 },
{ 5, 528000 },
{ 6, 768000 },
{ 7, CPUFREQ_TABLE_END },
};
coolbho3000, does this sound correct? You sure about the values for a11clk_src_div, ahbclk_khz, ahbclk_div at 768000? Sound like you've already started looking into this.
Is there a reason to change the AHB bus clock divider ahbclk_div to 5?

JJbdoggg, not sure what it'll do if it doesn't work. Probably freeze up the phone.

JJbdoggg said:
yep ill try it out if you compile it btw what is the worst hat could happen to my phone?
Click to expand...
Click to collapse
Your processor could get fried.

I'm 100% sure about a11clk_src_div (according to Qualcomm docs 0000 means you don't divide the PLL at all, leaving it at 768000, 0001 means you divide by 2, 0010 means you divide by 3, etc) but I'm not sure if ahbclk_khz is supposed to go above 132000 so I left it at 128000, making ahbclk_div 5 (768000/6 = 128000). Other possible values for the two might be 153600 and 4, and 192000 and 3. I'm digging through the Qualcomm documentation to see if I can find anything.

More info in a follow up from the original post:
http://www.nuerom.com/BlogEngine/post/2009/09/08/Yet-more-notes-on-overclocking.aspx

daproy said:
More info in a follow up from the original post:
http://www.nuerom.com/BlogEngine/post/2009/09/08/Yet-more-notes-on-overclocking.aspx
Click to expand...
Click to collapse
Seems unstable, but worth a shot. Seems like the worst that's happened to a WM user is a freeze.
3) There is a clock divider for the AHB that is separate from the ARM11 frequency. I’ve tried increasing the divider from the default of 0x4 to 0x5, with no real increase in stability.
It seems like multiple AHB clocks/dividers would work then, probably answering my above post... maybe best to try and see which is the most stable?

that would end getting really hot, if it is usable and perfected, hell yeah would i love to have some power like that out of my g1, i got 2 batteries

Well I played around with this a bit. Pulled cyanogen's latest kernel code with BFS from github. Patched acpuclock.c (patchfile attached). Compiled with the android toolchain.
First off there is definitely something wrong with the kernel i compiled because it was horrible slow to boot. I did use the .config from cyanogen 4.1.6 without edits so I'm not sure why. I can't imagine it was due to the patch since with default min and max scaling it should be running the same as the kernel compiled by cyanogen.
Regardless, I was able to try to overclocking:
Code:
# echo 528000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq
# echo 768000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq
# cat /proc/cpuinfo
Processor : ARMv6-compatible processor rev 2 (v6l)
BogoMIPS : 527.15
Features : swp half thumb fastmult edsp java
CPU implementer : 0x41
CPU architecture: 6TEJ
CPU variant : 0x1
CPU part : 0xb36
CPU revision : 2
Hardware : trout
Revision : 0080
Serial : 0000000000000000
# echo 768000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq
# cat /proc/cpuinfo
Processor : ARMv6-compatible processor rev 2 (v6l)
BogoMIPS : 766.77
Features : swp half thumb fastmult edsp java
CPU implementer : 0x41
CPU architecture: 6TEJ
CPU variant : 0x1
CPU part : 0xb36
CPU revision : 2
Hardware : trout
Revision : 0080
Serial : 0000000000000000
It reboot immediately following the cpuinfo but looks like it did scale up to 768000
Phone is fine after the reboot and back to default min max freq.
Before trying to toy around with different AHB clock dividers I should probably work out getting a good kernel to compile first, unless cyanogen has some spare time and feels like taking a look at this. I know I won't have any spare time in the next few days.

Amazing!!
Nice to see 766 in cpuinfo What do you think, was the reboot caused by cpu failure or "something wrong" in kernel?
If I understand well, we must modify kernel, and we cant easily install app to overclock..

[Q] Input capture device (mic)

Android 4.0.8, kernel 3.0.8, ICS.
The device has no mic, nor camera has it is a mediaplayer.
I plug in a PS3EYE, USB webcam if you will.
So with
Code:
# insmod gspca_main.ko
# insmod gspca_ov534.ko videomode=13
I get a /dev/video0 and indeed, I can use the camera in Skype. Hurrah.
Now, I'd like to get the mic to work. I believe the hardware is detected ok:
Code:
# cat /proc/asound/cards
0 [sun4icodec ]: sun4i-CODEC - sun4i-CODEC
sun4i-CODEC Audio Codec
1 [sun4isndhdmi ]: - sun4i-sndhdmi
sun4i-sndhdmi
2 [CameraB409241 ]: USB-Audio - USB Camera-B4.09.24.1
OmniVision Technologies, Inc. USB Camera-B4.09.24.1 at usb-sw-ehci-1.3, high sp
3 [sun4isndspdif ]: - sun4i-sndspdif
sun4i-sndspdif
# ls -l /dev/snd/
crw-rw---- system audio 116, 0 2012-07-03 14:12 controlC0
crw-rw---- system audio 116, 32 2012-07-03 14:12 controlC1
crw-rw---- system audio 116, 64 2012-07-03 14:12 controlC2
crw-rw---- system audio 116, 96 2012-07-03 14:12 controlC3
crwxrwxrwx system system 116, 24 2012-07-03 14:12 pcmC0D0c
crwxrwxrwx system system 116, 16 2012-07-03 14:12 pcmC0D0p
crw-rw---- system audio 116, 48 2012-07-03 14:12 pcmC1D0p
crw-rw---- system audio 116, 88 2012-07-03 14:12 pcmC2D0c
crw-rw---- system audio 116, 112 2012-07-03 14:12 pcmC3D0p
crw-rw---- system audio 116, 33 2012-07-03 14:12 timer
But I am missing the magic needed to tell Android to use the Mic as the default capture device. How is that generally done in Android? I think if done right, Any of the audio-memo apps should work, as well as "voice input".and of course, Skype..
I have tried
# setprop usb.audio.cap.device pcmC0D0c
# setprop media.audio.in.mode usb
but neither changes anything. Perhaps that is something CM9 specific...

You need a /etc/asound.conf
Here's mine for unrelated hardware, yours will be different:
Code:
defaults.ctl.card 1
defaults.pcm.card 1
defaults.timer.card 1
ctl.AndroidOut
{
type hw
card 1
}
ctl.AndroidIn
{
type hw
card 1
}
pcm.AndroidPlayback
{
type hooks
slave.pcm
{
type hw
card 1
device 0
}
hooks.0
{
type ctl_elems
hook_args
[
{ name 'Speaker Playback Switch' value on }
{ name 'Speaker Playback Volume' value [150 150] }
]
}
}
pcm.AndroidPlayback_Speaker
{
type hooks
slave.pcm
{
type hw
card 1
device 0
}
hooks.0
{
type ctl_elems
hook_args
[
{ name 'Speaker Playback Switch' value on }
{ name 'Speaker Playback Volume' value [150 150] }
]
}
}
pcm.AndroidPlayback_Speaker_normal
{
type hooks
slave.pcm
{
type hw
card 1
device 0
}
hooks.0
{
type ctl_elems
hook_args
[
{ name 'Speaker Playback Switch' value on }
{ name 'Speaker Playback Volume' value [150 150] }
]
}
}
pcm.AndroidCapture
{
type hooks
slave.pcm
{
type hw
card 1
device 0
}
hooks.0
{
type ctl_elems
hook_args
[
{ name 'Mic Capture Switch' value on }
{ name 'Mic Capture Volume' value 12 }
{ name 'Mic Playback Switch' value on }
{ name 'Mic Playback Volume' value 20 }
]
}
}

Ah, I was expected there to already be one, if it one was to be read. But you are suggesting I should just create it, and one of the libraries will read it?
I will try, thanks for the hints.

Well, I feel it is frustratingly close
Code:
1|[email protected]:/mnt/obb # alsa_arecord -l
**** List of CAPTURE Hardware Devices ****
card 0: sun4icodec [sun4i-CODEC], device 0: M1 PCM [sun4i PCM]
Subdevices: 1/1
Subdevice #0: subdevice #0
card 2: CameraB409241 [USB Camera-B4.09.24.1], device 0: USB Audio [USB Audio]
Subdevices: 1/1
Subdevice #0: subdevice #0
# alsa_arecord -f dat -D plughw:2,0 -d 10 out.wav
Recording WAVE 'out.wav' : Signed 16 bit Little Endian, Rate 48000 Hz, Stereo
and I can confirm that works. It records from the USB mic and I can hear it.
I assume you meant /system/etc/asound.conf; I added the relevant lines:
Code:
ctl.AndroidIn
{
type plug
card 2
}
pcm.AndroidCapture
{
type hooks
slave.pcm
{
type plughw
card 2
device 0
}
hooks.0
{
type ctl_elems
hook_args
[
{ name 'Speaker Playback Switch' value on }
{ name 'Speaker Playback Volume' value [150 150] }
]
}
}
and it appears to read that file, because when I attempt to load "Mic Visualizer" I get in logcat:
Code:
D/audio_hw_primary( 89): ~~~~~~~~~~~~~~ start_input_stream, card: 2
E/audio_hw_primary( 89): cannot open pcm_in driver: pcm_open_req cannot set hw params: Invalid argument
D/audio_hw_primary( 89): ~~~~~~~~~~~~~~ start_input_stream, card: 2
E/audio_hw_primary( 89): cannot open pcm_in driver: pcm_open_req cannot set hw params: Invalid argument
So, it is trying to read from card 2, so it most likely is reading the asound.conf file. But beyond that.. nothing.

[Q] How to modify kernel source?

Hi all
I'm interested in creating my own ROM and kernel for my GT2 and I downloaded source codes from Samsung open source.
I want to underclock my device so that I have to modify some codes of the kernel.
I figured out the frequency table is coded in arch/arm/mach-omap2/opp4xxx_data.c
Code:
static struct omap_opp_def __initdata omap443x_opp_def_list[] = {
/* MPU OPP1 - OPP50 */
OPP_INITIALIZER("mpu", "dpll_mpu_ck", "mpu", true,
300000000, OMAP4430_VDD_MPU_OPP50_UV),
/* MPU OPP2 - OPP100 */
OPP_INITIALIZER("mpu", "dpll_mpu_ck", "mpu", true,
600000000, OMAP4430_VDD_MPU_OPP100_UV),
/* MPU OPP3 - OPP-Turbo */
OPP_INITIALIZER("mpu", "dpll_mpu_ck", "mpu", true,
800000000, OMAP4430_VDD_MPU_OPPTURBO_UV),
/* MPU OPP4 - OPP-SB */
OPP_INITIALIZER("mpu", "dpll_mpu_ck", "mpu", true,
1008000000, OMAP4430_VDD_MPU_OPPNITRO_UV),
I'm almost sure that the numbers at each line represent a CPU frequency,
and the last parameters are macro defined at the beginning of the code:
Code:
/*
* Structures containing OMAP4430 voltage supported and various
* voltage dependent data for each VDD.
*/
#define OMAP4430_VDD_MPU_OPP50_UV 1025000
#define OMAP4430_VDD_MPU_OPP100_UV 1200000
#define OMAP4430_VDD_MPU_OPPTURBO_UV 1313000
#define OMAP4430_VDD_MPU_OPPNITRO_UV 1374000
#define OMAP4430_VDD_MPU_OPPNITROSB_UV 1375000
So, I guess an argument for a frequency should be:
Code:
OPP_INITIALIZER("mpu", "dpll_mpu_ck", "mpu", true, [U][b]Frequency in Hz[/b][/U], [U][b]Voltage[/b][/U]),
And the frequency and voltage should match each other...
Also, I want to add some governors to the kernel, I think I need to do the followings:
1. Get the source of the governors
2. Modify Makefile
3. Add the governors in Kconfig
I want to know if the above is correct... Thanks a lot:victory:
Ivan

[KERNEL][24.12.2013][X920D][JB 4.2.2] KillX Kernel v3.8.6 [Discontinued]

KillX Kernel
(version 3.8.6)​
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
/!\ DISCLAIMER: I AM NOT RESPONSIBLE IF YOU BRICK / RUIN YOUR PHONE IN ANY WAY /!\​
What is ...?
What is USB Force Fast Charge?
This allows one to force AC charging for any charger that is detected as USB (e.g. many car chargers) and pull the full current the charger can support.
It also provides additional security when connecting to public charging stations because by forcing AC charging, USB/adb data transfers are disabled, protecting your data.
Fast charge can be toggled by issuing:
echo 1 > /sys/kernel/fast_charge/force_fast_charge
and off:
echo 0 > /sys/kernel/fast_charge/force_fast_charge
Click to expand...
Click to collapse
What is I/O Scheduler?
Input/output (I/O) scheduling is a term used to describe the method computer operating systems decide the order that block I/O operations will be submitted to storage volumes.
I/O Scheduling is sometimes called 'disk scheduling'. I/O schedulers can have many purposes depending on the goal of the I/O scheduler, some common goals are:
+ To minimize time wasted by hard disk seeks.
+ To prioritize a certain processes' I/O requests.
+ To give a share of the disk bandwidth to each running process.
+ To guarantee that certain requests will be issued before a particular deadline.
Click to expand...
Click to collapse
What is msm_thermal?
Kernel based 3-phase thermal control.This replaces your /system/bin/thermald binary which is renamed by the installer to thermald_bkp
Check /sys/kernel/msm_thermal/conf/ for the thermal configuration
allowed_max_high = highest threshold (phase 3)
allowed_max_low = remove the throttling if we cooled down to this (clr_thrshold)
allowed_max_freq = max frequency if throttled (limit)
[...]mid[...] = same as above, just for phase 2
[...]low[...] = Lowest threshold (phase 1)
check_interval_ms = how often shall we check? (sampling rate)
Default: 1000ms = 1sec
Click to expand...
Click to collapse
What is msm_mpdecision?
Kernel based multi core decision
This replaces your /system/bin/mpdecision binary which is ignored in init.rc and renamed by the installer to mpdecision_bkp
+ cpu auto-hotplug/unplug based on system load for MSM multicore cpus [should cpu(x) be online or not?]
+ single core while screen is off
+ extensive sysfs tuneables
Check /sys/kernel/msm_mpdecision/conf/ for the configuration.
Code:
startdelay = time until mpdecision starts doing it's magic (20000)
delay = time between checks (70)
pause = if something else plugs in the cpu, fall asleep for 3000ms (3 secs)
scroff_single_core = if the screen is off, don't plug in cpu1/2/3. Additionally: Unplug all cpus except cpu0 when screen is turned off (1)
enabled = enable(1) or disable(0) mpdecision. This does not affect scroff_single_core!
min_cpus = min cpus to be online, cannot be < 1. Default: 1
max_cpus = max cpus to be online, cannot be > 4. (if you set it to 2 and min_cpus to 1 you will basically have a dualcore) Default: 4
idle_freq = a value against that will be checked if a core +/- is requested. (486000)
If cpu0 is below that value and a core up of another cpu is requested, nothing will happen.
If any other cpu is above that value and a core down of that cpu is requested, nothing will happen. (otherwise it would now put down that cpu even though it is still working, which isn't what we want)
Hot plug thresholds (aka now it gets 'complicated')
This small formula calculates which value will be used: (number_of_cpus_online - 1) * 2
The result of this formula will be the nwns_threshold where a new cpu is hotplugged.
The result of this formula + 1 will be the nwns_threshold where a cpu is unplugged.
nwns_threshold_x = runqueue threshold, if this is reached cpuX will be hot/unplugged
twts_threshold_x = time threshold, this amount of time must have passed for the related action to be taken (hot/unplug)
Example:
Code:
One cpu is online.
(1 - 1) * 2 = 0 ergo:
nwns_threshold_0 = cpu1 will be hotplugged at this value
((1 - 1) * 2) + 1 = 1
nwns_threshold_1 = cpu0 will be unplugged at this value
Since we can't unplug cpu0 this is '0'.
Two cpus are online.
(2 - 1) * 2 = 2 ergo:
nwns_threshold_2 = cpu2 will be hotplugged at this value
((2 - 1) * 2) + 1 = 3
nwns_threshold_3 = cpu1 will be unplugged at this value
etc...
The default values are:
NwNs_Threshold: 12, 0, 25, 20, 32, 28, 0, 35
TwTs_Threshold: 140, 0, 140, 190, 140, 190, 0, 190
Where the position and function of the number equals the result of the above explained formula
Click to expand...
Click to collapse
Features:
Code:
[FONT=Arial Black][SIZE=2]
* /system rw
* support init.d
* adb root
* swap on
* USB Force Fast Charging
* [B]Energy Efficiency[/B]
* XZ compressed kernel
* Using /dev/frandom
* allow WiFi networks with low signal to be seen !
* exFAT support
* intelli/msm_thermal
* msm_mpdecision
* [B]IO[/B]: noop, deadline, cfq, sio, fiops
* [B]CPU Governor[/B] : msm-dcvs, userspace, powersave, ondemand, performance, DanceDance & Lionheart
* Zram (wip)
* staging: zram: per-cpu support to Crypto
[/SIZE][/FONT][CENTER][B][FONT=Arial][SIZE=2][FONT=Arial]
Please read Change log for more info.[/FONT][/SIZE][/FONT][/B][/CENTER]
Installation:
[!] Unlocked Bootloader
[!] Custom Recovery Installed CWM or TWRP
[!] S-OFF + rooted + busybox Installed
Code:
1. Download the zip
2. flash in recovery ( copy to sdcard or use adb sideload )
Reboot n Enjoy!
Thank`s to:
Linus Torvalds, showp1984, faux123, Zarboz, dsb9938, elkay, crpalmer, chad0989, codeaurora.org, kernel.org, linux-foundation.org, xda-developers.com, HTC and all kernel and rom dev !!!
Click to expand...
Click to collapse
Please use and enjoy it, and if you can make a donation of any amount, I'd appreciate it immensely
​

Change Log
Previous Version:
KillX v1.0.1
Code:
Wed Dec 18 06:31:31 MYT 2013
gcc version 4.7.2 (Ubuntu/Linaro 4.7.2-1ubuntu1)
TWEAK:
+ io: deadline
+ io: cfq
+ force fast charge: battery 8960: unknown USB -> AC
ADD:
+ io: fiops
+ io: sio
+ force fast charge by chad0989
FIX:
+ compile warning TODO: return_address
+ compile error: adreno.c [-Wframe-larger-than=]
+ Makefile: Assembler messages
Compile:
+ linaro -O3 compliance #1
+ Test Built #1
Change:
+ Compiler flags
+ set version : This is KillX v1.0.1
Disable:
+ PERFLOCK #1
KillX v3.8.6
Code:
Tue Dec 24 01:18:20 MYT 2013
gcc version 4.7.2 (Ubuntu/Linaro 4.7.2-1ubuntu1)
ADD:
* /dev/frandom
* allow WiFi networks with low signal to be seen
* cpu gov: DanceDance & Lionheart
* exFAT support
* intelli/msm_thermal
* msm_mpdecision
* Zram (wip)
* crypto: lz4 Cryptographic API
* lib: lz4 compressor module
* arm: support for LZ4-compressed kernel
* lib: support for LZ4-compressed kernel
* decompressor: LZ4 decompressor module
* staging: zram: per-cpu support to Crypto
SET:
* b43: Increase number of RX DMA slots
* config: intelli/msm_thermal
* config: msm_mpdecision boost and CPU FREQ MIN MAX
* config: msm_mpdecision boost off
* config: UACCESS_WITH_MEMCPY = No
* Don't force RUN_QUEUE_STATS to be on
* lib/int_sqrt.c: optimize square root algorithm
* LOAD_FREQ: avoids loadavg Moire http://ripke.com/loadavg/moire
* Makefile: Snapdragon optimizations
* msm_hsic: decrease timeout, 500 to 120
* wifi: enabled PM_FAST by default
* config: ADAPTIVE_TUNING off
* softirq: reduce latencies
* zram_drv: switch default compressor to new lz4
* AIO: Don't plug the I/O queue in do_io_submit()
* zcache-main: switch to lz4 as default
* zcache-main.c: use MACRO define instead of hardcoded default of lzo
REMOVE:
* bonding: already created master sysfs link on failure
* code: __cpuinitdata and __cpuinit
* code: get rid of __devinit and __cpuinit, cpuinit into nops
* code: perflock.c
* inotify: remove broken mask checks causing unmount to be EINVAL
* net: WARN_ON() in net_enable_timestamp()
* rwsem: steal writing sem for better performance
* rwsem: Writer lock-stealing
* scheduler: compute time-average nr_running per run-queue #2
* video: msm:mdp pr_debug _mdp_histogram_crtl start false
* workqueue: HTC's incompatible workqueue debug
* drivers/gpu/msm/kgsl.c: silence annoying and useless dmesg log spam
* sched/rt: redundant nr_cpus_allowed test
* drivers/misc/pm8xxx-vibrator-pwm.c: disable meaningless dmesg log spam
* drivers/input/evbug.c: silence annoying dmesg log spam
UPDATE:
* ARM: 7011/1: Add ARM cpu topology definition #2
* efivars: update
* fs: sync
* ipv6: addrconf
* ipv6: ip6_input
* sysfs: dir.c
* zram_drv.c
* SELinux: include definition of new capabilities
FIX:
* ALSA: hda - Release assigned pin/cvt at error path of hdmi_pcm_open()
* aoe: reserve enough headroom on skbs
* ARM: PXA3xx: program the CSMSADRCFG register
* ARM: scheduling while atomic warning in alignment handling code
* ARM: VFP: emulation of second VFP instruction
* ARMv7: perf: EVTYPE_MASK to include NSH bit
* Bluetooth: not closing SCO sockets in the BT_CONNECT2 state
* bonding: disabling of arp_interval and miimon and arp_interval
* bridge: priority of STP packets
* Btrfs: don't drop path when printing out tree errors in scrub
* Btrfs: limit the global reserve to 512mb
* Btrfs: race between mmap writes and compression
* cgroup: exit() vs rmdir() race
* cifs: Allow passwords which begin with a delimitor
* cifs: ensure that cifs_get_root() only traverses dir
* cifs: handling of blank password option
* compat_rw_copy_check_uvector() misuse in aio, readv, writev
* cpufreq: Convert the cpufreq_driver_lock to a rwlock
* cpufreq: Convert the cpufreq_driver_lock to use RCU
* cpuset: cpuset_print_task_mems_allowed() vs rename() race
* dca: check against empty dca_domains list before unregister provider
* Driver core: treat unregistered bus_types as having no devices
* drivercore: ordering between deferred_probe and exiting initcalls
* drivers/video: fsl-diu-fb: fix pixel formats for 24 and 16 bpp
* drm/udl: disable fb_defio by default
* drm/udl: make usage as a console safer
* drm/usb: bind driver to correct device
* drm: don't add inferred modes for monitors that don't support them
* drm: Fill depth/bits_per_pixel for C8 format
* drm: Use C8 instead of RGB332 when determining the format from depth/bpp
* efivars: explicitly calculate length of VariableName
* fb: rework locking to fix lock ordering on takeover
* fb: Yet another band-aid for fixing lockdep mess
* fbcon: don't lose the console font across generic->chip driver switch
* force vsync in board file
* fs/block_dev: page cache wrongly left invalidated,revalidate_disk()
* fs: cifs: warning: 'midQ' may be used uninitialized in this function
* fs: proc: stat: not showing all cpus
* fuse: don't WARN when nlink is zero
* futex: Mark get_robust_list as deprecated
* HID: wiimote: nunchuck button parser
* hw_random: make buffer usable in scatterlist
* iommu/msm: the include guard in iommu.h
* ipv4: bug in ping_err()
* keys,race with concurrent install_user_keyrings()
* ks8851: interpretation of rxlen field
* loop: prevent bdev freeing while device in use
* memcopy -subroutines for memory copy functions.
* memory leak in cpufreq stats
* missing nf_reset() in tun_net_xmit()
* mm/fadvise: drain all pagevecs, POSIX_FADV_DONTNEED fails
* mmc: sdhci-esdhc-imx: host version read
* mmu_notifier: have mmu_notifiers use a global SRCU
* mmu_notifier: make the mmu_notifier srcu static
* mmu_notifier: unregister NULL Pointer deref and multiple
* msm: mdp: Send fake vsync only when enabled from userspace
* msm_fb: display: blt always enabled for 720p,1080p, secure buf
* msm_fb: display: Fix flickering when BLT is enabled/disabled
* net/irda: add missing error path release_sock call
* net/sunrpc: 'subbuf.[...]' may be used uninitialized
* net: *_DIAG_MAX constants
* net: compile error when SOCK_REFCNT_DEBUG is enabled
* net: infinite loop in __skb_recv_datagram()
* NFS: Don't allow NFS silly-renamed files to be deleted,no signal
* NLM: Ensure that we resend all pending blocking locks after a reclaim
* ocfs2: unlock super lock if lockres refresh failed
* perf tools: build with bison 2.3 and older
* pnfs-block: removing DM device maybe cause oops when call dev_remove
* posix-cpu-timers: nanosleep task_struct leak
* pstore: Avoid deadlock in panic and emergency-restart path
* qseecom: Fix issue with incomplete command exiting prematurely
* Re-apply all PM/Sleep
* resume drivers/video/backlight/adp88?0_bl.c
* sky2: Receive Overflows not counted
* sky2: Threshold for Pause Packet is set wrong
* sock_diag: out-of-bounds access to sock_diag_handlers
* staging:comedi:comedi_fops: update
* staging:comedi:ni_labpc: correct differential channel sequence for AI cmd
* staging:comedi:ni_labpc: set up command4 register *after* command3
* svcrpc: make svc_age_temp_xprts enqueue under sv_lock
* switch: do_fsync() to fget_light()
* sysctl: null checking in bin_dn_node_address()
* sysfs: handle failure path correctly for readdir()
* sysfs: race between readdir and lseek
* tcp: preserve ACK clocking in TSO
* tcp: undo spurious timeout after SACK reneging
* tmpfs: use-after-free of mempolicy object
* tty: Prevent deadlock in n_gsm driver
* tty: set_termios/set_termiox should not return -EINTR
* umount oops when remove blocklayoutdriver first
* unbreak automounter support on 64-bit kernel with 32-bit userspace (v2)
* unix: race condition in unix_release()
* USB: ehci-omap: autoloading of module
* usb: gadget: udc-core: regression during gadget driver unbinding
* USB: serial: null-pointer dereferences on disconnect
* USB: usb-storage: unusual_devs update for Super TOP SATA bridge
* usb: xhci: TRB transfer length macro used for Event TRB
* vfs,pipe counter breakage
* vgacon/vt: clear buffer attributes when we load a 512 char
* vhost: length for cross region descriptor
* vlan: adjust vlan_set_encap_proto() for its callers
* vt: synchronize_rcu() under spinlock is not nice
* xen-netback: cancel the credit timer when taking the vif down
* xen-netback: correctly return errors from netbk_count_requests()
* intelli/msm_therma: missing default parameters
* writeback: occasional slow sync(1)
* CHROMIUM: mm: calculation of dirtyable memory
* intelli-thermal: stats calculations
* htc_monitor: BAD coding from HTC using global vars
* lib/lz4/lz4_compress: macro usage error
* LZ4: compression/decompression signedness mismatch (v2)
Note:
- This kernel can be installed on any rom with Sense 5 Jelly Bean 4.2.2 except BOA v6 (will built another version to support this rom)
- Work great on hboot 1.54, JB 4.2.2, TW 2.28.709.3 !
- It proved, betterrrrrrrrrrrrr battery life !

Download Area
Code:
[1] [URL="http://forum.xda-developers.com/attachment.php?attachmentid=2460809&d=1387322910"]killx_v1.0.1.zip[/URL] [B]MD5[/B]: 8c31a3ff3bf1ccef12fc45a94f5d76f7
[2] [URL="http://forum.xda-developers.com/attachment.php?attachmentid=2471082&stc=1&d=1387837499"]killx_v3.8.6.zip[/URL] [B]MD5[/B]: 07f9ac0202ad46772566758346bee15f
My GitHub​

Great work!
Thanks man, really great work!
Just downloaded the modified hboot and your kernel, will flash it soon and report any issues.
Is there already a sweep2wake option, like in your older kernel version?

I will port all the futures from old version, lot of work to do now
Sent from my HTC Butterfly using xda premium

Hi, thanks for the hard work i flashed it just now and found that Wifi will not wake. It shows "Error" after long "Turning On"..
what can i do to revive wifi?
Update : I managed to wake Wifi after follow the instructions from this post, http://forum.xda-developers.com/showpost.php?p=41320412&postcount=302
thanks and will test it throughly tomorrow

Coz your device is S-ON.
Sent from my HTC Butterfly using xda premium

Will you make kernel for x920e European version too
Sent from my HTC Butterfly using xda app-developers app

Bassie39 said:
Will you make kernel for x920e European version too
Sent from my HTC Butterfly using xda app-developers app
Click to expand...
Click to collapse
if i had time

Woot. Love your work man. Does it work for GPE ROM? Has anyone tested? I'm on newtroot's GPE ROM.
sent from the world where no one knows...

newtroot ROM
I'm on newtroot's NOS v3 ROM and KillX v1.0.1 Kernel works like a charm! Gave me a bit more then 21k at AnTuTu benchmark.
Can't say anything about battery life yet ... But looks good so far.
http://forum.xda-developers.com/showthread.php?t=2193912

Hi
Can you add support to japan butterfly?

I dont see kernel source, android 4.2.2, j butterfly at www.htcdev.com.. only 4.1
Sent from my HTC Butterfly using xda app-developers app

anyone try on Boa ROM?
Sent from my HTC Butterfly using Tapatalk 4

Yup, it work on any android 4.2.2
Sent from my HTC Butterfly using xda app-developers app

chockchip said:
anyone try on Boa ROM?
Sent from my HTC Butterfly using Tapatalk 4
Click to expand...
Click to collapse
Yes, It runs smoothly without any trouble. Be sure to flash. New kernnel give me 20500 antutu benchmark score.:laugh::laugh::laugh:

I just flashed the kernel...it works well...I am running stock 2.33..Also I tried it on gpe still no luck...
Sent from my HTC Butterfly using XDA Premium 4 mobile app

New version still under testing..see changelog @ post #2
Sent from my HTC Butterfly using xda app-developers app

MIUI ROM
I flashed MIUI rom yesterday with your kernel....It runs really smooth and everythings works fine...

Thank you for your feedback!
Can we share the rom in here ?
Sent from my HTC Butterfly using xda app-developers app

preferred_network_mode and 2G prefered

Hello,
I wish to configure my rooted device to use 2G if available, else 3G/4G. The idea is to economise battery if 2G is available (2G is enough for me).
I do not know what to choose in the list of values, I think 7 is the correct value but I am not sure. Of course it's hard to test, having 2G, moving to an area without 2G but with 3/4G ....
Can you help me to choose the correct value please ?
typedef enum {
PREF_NET_TYPE_GSM_WCDMA = 0, /* GSM/WCDMA (WCDMA preferred) */
PREF_NET_TYPE_GSM_ONLY = 1, /* GSM only */
PREF_NET_TYPE_WCDMA = 2, /* WCDMA */
PREF_NET_TYPE_GSM_WCDMA_AUTO = 3, /* GSM/WCDMA (auto mode, according to PRL) */
PREF_NET_TYPE_CDMA_EVDO_AUTO = 4, /* CDMA and EvDo (auto mode, according to PRL) */
PREF_NET_TYPE_CDMA_ONLY = 5, /* CDMA only */
PREF_NET_TYPE_EVDO_ONLY = 6, /* EvDo only */
PREF_NET_TYPE_GSM_WCDMA_CDMA_EVDO_AUTO = 7, /* GSM/WCDMA, CDMA, and EvDo (auto mode, according to PRL) */
PREF_NET_TYPE_LTE_CDMA_EVDO = 8, /* LTE, CDMA and EvDo */
PREF_NET_TYPE_LTE_GSM_WCDMA = 9, /* LTE, GSM/WCDMA */
PREF_NET_TYPE_LTE_CMDA_EVDO_GSM_WCDMA = 10, /* LTE, CDMA, EvDo, GSM/WCDMA */
PREF_NET_TYPE_LTE_ONLY = 11, /* LTE only */
PREF_NET_TYPE_LTE_WCDMA = 12, /* LTE/WCDMA */
PREF_NET_TYPE_TD_SCDMA_ONLY = 13, /* TD-SCDMA only */
PREF_NET_TYPE_TD_SCDMA_WCDMA = 14, /* TD-SCDMA and WCDMA */
PREF_NET_TYPE_TD_SCDMA_LTE = 15, /* TD-SCDMA and LTE */
PREF_NET_TYPE_TD_SCDMA_GSM = 16, /* TD-SCDMA and GSM */
PREF_NET_TYPE_TD_SCDMA_GSM_LTE = 17, /* TD-SCDMA,GSM and LTE */
PREF_NET_TYPE_TD_SCDMA_GSM_WCDMA = 18, /* TD-SCDMA, GSM/WCDMA */
PREF_NET_TYPE_TD_SCDMA_WCDMA_LTE = 19, /* TD-SCDMA, WCDMA and LTE */
PREF_NET_TYPE_TD_SCDMA_GSM_WCDMA_LTE = 20, /* TD-SCDMA, GSM/WCDMA and LTE */
PREF_NET_TYPE_TD_SCDMA_GSM_WCDMA_CDMA_EVDO_AUTO = 21, /* TD-SCDMA, GSM/WCDMA, CDMA and EvDo */
PREF_NET_TYPE_TD_SCDMA_LTE_CDMA_EVDO_GSM_WCDMA = 22 /* TD-SCDMA, LTE, CDMA, EvDo GSM/WCDMA */
} RIL_PreferredNetworkType;