[KERNEL][GB/SENSE] 2.6.35.10-g3f43272 [v06] | OC1152 | LZMA | fixes +addtl modules

Search This thread

TomLeeDesire

Senior Member
Aug 3, 2010
1,058
147
Had a spontaneous reboot this morning, but as it was in my pocket, I'm not sure if it can be an unwanted long press on power button :S
Got the last_kmsg if it can help.

Then later, trying some gscript to test (had a freeze using it on a previous version), i had a freeze again (pull out, no last_kmsg). I checked vdd levels twice, went fine, then I checked the sd scheduler. Freeze.
I tried again after reboot. Check Sd scheduler, freeze. But maybe my script isn't correct anymore :S
cat /sys/block/mmcblk0/queue/scheduler

111003 on latest Lee HD port
 

rjvrijn

Senior Member
Sep 12, 2009
104
8
To those complaining about the usb-popup bug, snq- said a few messages ago that it was a bug with newer android versions (2.3.4 and 2.3.5) and he said he's working on it.

Ah ok. But for those of us with Android 2.3.3. Fastboot oem eraseconfig should solve this. Just retest the whole thing.

Sent from my HTC Desire using XDA App
 

tobi01001

Senior Member
Dec 4, 2008
258
64
Munich
@snq-
Got another - for now maybe final - one:
Code:
static int get_ls_adc_level(uint8_t *data)
{
	struct microp_ls_info *li = ls_info;
	uint8_t i, adc_level = 0;
	uint16_t adc_value = 0;
	[COLOR="Blue"][B]static uint16_t old_value = 0;[/B][/COLOR]


	data[0] = 0x00;
	data[1] = li->ls_config->channel;
	if (microp_read_adc(data))
		return -1;

	read_spi_status();

	adc_value = [B][COLOR="Blue"]old_value = [/COLOR][/B]ls_info->lsauto ? (data[0]<<8 | data[1]) :[B][COLOR="Blue"] old_value;[/COLOR][/B]

Reasoning: If the screen uses the last value when it will be turned on, why shouldn't the light sensor report the last one...
 
  • Like
Reactions: snq-
Have been using the bravo-2.6.35.14-snq-111003-no2wcr kernel for a day now. Snappy! And excellent battery life:
1mA with airplane mode and screen off
3-9mA with 3G on and screen off
130-270mA with 3G on and browsing/Tapatalk/Twidroyd
170-320mA with WiFi on and browsing/Tapatalk/Twidroyd

So WiFi has a slightly higher drain than 3G on my phone.

And I don't have the USB option window issue.

Great job, snq-!
 

resadent

Senior Member
Nov 1, 2010
274
30
Málaga
Ah ok. But for those of us with Android 2.3.3. Fastboot oem eraseconfig should solve this. Just retest the whole thing.

Sent from my HTC Desire using XDA App

My bad, I tried to be too clever :p
BTW, have to say latest kernel ia the fastest yet on my desire. No stability problema for now, excellent batt life... great job snq-.
 
Last edited:

rabeel

Senior Member
Jan 11, 2011
115
5
London
Have been using the bravo-2.6.35.14-snq-111003-no2wcr kernel for a day now. Snappy! And excellent battery life:
1mA with airplane mode and screen off
3-9mA with 3G on and screen off
130-270mA with 3G on and browsing/Tapatalk/Twidroyd
170-320mA with WiFi on and browsing/Tapatalk/Twidroyd

So WiFi has a slightly higher drain than 3G on my phone.

And I don't have the USB option window issue.

Great job, snq-!

which app do u use to track battery usage?
 

Gr4bov5ky

Member
Jun 19, 2011
28
1
I've flashed the bravo-2.6.35.14-snq-111003 kernel, the wifi wasn't workin couldn't connect. Not even with the wifin update... About 200mA when phone was idle, only screen on.

Flashed kernel 110926....wifi isn't still working.. OMG

Any ideas?
 
Have been using the bravo-2.6.35.14-snq-111003-no2wcr kernel for a day now. Snappy! And excellent battery life:
1mA with airplane mode and screen off
3-9mA with 3G on and screen off
130-270mA with 3G on and browsing/Tapatalk/Twidroyd
170-320mA with WiFi on and browsing/Tapatalk/Twidroyd

So WiFi has a slightly higher drain than 3G on my phone.

And I don't have the USB option window issue.

Great job, snq-!

Spoke too soon... Total freeze! Had to pull the battery. Back to the 110926 kernel.

Sent from my HTC Desire using xda premium
 

gregsoir

Senior Member
Aug 1, 2011
515
103
North Pole
I've flashed the bravo-2.6.35.14-snq-111003 kernel, the wifi wasn't workin couldn't connect. Not even with the wifin update... About 200mA when phone was idle, only screen on.

Flashed kernel 110926....wifi isn't still working.. OMG

Any ideas?
full wipe and reflash everything again?

btw, i just had one freeze, had to pull the battery out. apart from that kernel runs smooth and battery is fine after full wipe
 
  • Like
Reactions: Gr4bov5ky

Gr4bov5ky

Member
Jun 19, 2011
28
1
full wipe and reflash everything again?

btw, i just had one freeze, had to pull the battery out. apart from that kernel runs smooth and battery is fine after full wipe

I've wiped everything that could be wiped,reflashed everything:confused:..WiFi isn't working...

Update: Don't know why...it's working

Now with 111003, fast and stable. Let's see the battery life.
 
Last edited:

snq-

Retired Recognized Developer
Aug 13, 2010
581
853
A little ot, but can you please make a froyo kernel with the autobrightness mod?

hey, sorry for the delay in releasing this, but i found some upstream code that caused reproducible (\o/) insta-crash (radio!), still investigating whether they have anything to do with the gb issues (looking at it this way, your req wasn't o/t :))


again, this is for froyo
bravo-2.6.32.46-snq-111004.zip
(contains tobi's latest one, all credits to him for the lightsensor mod!, smartass updated to v2, compiled w/ gcc 4.6)

modified sources here
 
  • Like
Reactions: anderl78

snq-

Retired Recognized Developer
Aug 13, 2010
581
853
Could I ask why are you using 1.0.9 and not latest one? Isn't 1.1.3-1.1.4 the best avaiable?

it's just i haven't had the time to upgrade recently
i'm pretty sure the newer versions are better (afaik baadnwz fixed the reboot-while-taking-photos issue, etc)
 
  • Like
Reactions: gambit3rd

snq-

Retired Recognized Developer
Aug 13, 2010
581
853
Reasoning: If the screen uses the last value when it will be turned on, why shouldn't the light sensor report the last one...

love your static local
you can test it by fastbooting the froyo build's zImage for now (works ok on a gb rom)
 

tobi01001

Senior Member
Dec 4, 2008
258
64
Munich
love your static local
you can test it by fastbooting the froyo build's zImage for now (works ok on a gb rom)

Already tested it with tristan202's source and a fastboot zImage... Works well... Just waiting for your GB git. But it's good to see that the froyo guy's get a taste of 10 sensor levels... I wonder how services.jar is implemented on froyo.

Sent from my HTC Desire using xda premium
 

anderl78

Senior Member
Jul 18, 2008
2,044
965
hey, sorry for the delay in releasing this, but i found some upstream code that caused reproducible (\o/) insta-crash (radio!), still investigating whether they have anything to do with the gb issues (looking at it this way, your req wasn't o/t :))


again, this is for froyo
bravo-2.6.32.46-snq-111004.zip
(contains tobi's latest one, all credits to him for the lightsensor mod!, smartass updated to v2, compiled w/ gcc 4.6)

modified sources here

Wow! Thank you very much for this! The frojo user will have a lot of fun with this! :D

Sent from my HTC Desire using XDA App
 

anderl78

Senior Member
Jul 18, 2008
2,044
965
Already tested it with tristan202's source and a fastboot zImage... Works well... Just waiting for your GB git. But it's good to see that the froyo guy's get a taste of 10 sensor levels... I wonder how services.jar is implemented on froyo.

Sent from my HTC Desire using xda premium

I guess, HTC made the same in GB as in frojo... Let's see... If I find a little bit more time I will look into it... ;-)

The now available 10 levels is for sure a big improvement for the frojo guys!

Sent from my HTC Desire using XDA App
 

Top Liked Posts

  • There are no posts matching your filters.
  • 106
    (modifications/addons for the _closed_source_ 2.6.35.10 htc kernel - for rom maintainers)

    PATCHED KERNEL IMAGE

    v01:
    • lzma-compressed kernel
    • o/c: 128MHz, 245MHz, 384MHz, 576-1152MHz (stepping: 38400kHz)
    • u/v: svs (@128000kHz: 950mV, @245000kHz: 975mV, @384000kHz: 1000mV, @576000kHz: 1075mV, @614400kHz: 1100mV, @652800kHz: 1100mV, @691200kHz: 1125mV, @729600kHz: 1150mV, @768000kHz: 1175mV, @806400kHz: 1200mV, @844800kHz: 1225mV, @883200kHz: 1250mV, @921600kHz: 1250mV, @960000kHz: 1275mV, @998400kHz: 1275mV, @1036800kHz: 1275mV, @1075200kHz: 1275mV, @1113600kHz: 1300mV, @1152000kHz: 1300mV)
    • cpu vdd regulator constraints expanded (800..1350mV)
    • patched out "Take the fastest speed available at the specified VDD level" (=full freq tab)
    • pre-cracked perflock (no need for perflock disabler module)

    v02:
    • patched cpufreq code: set default cpu frequency constraints: 245..998MHz (with v01 cpu freq jumped to 1152 during the boot process)

    v03:
    • trackpad-to-wake support (oj will not choke after a while when screen is off)
    • workaround for non-standard (extended) batteries (devices w/ stock htc battery are not affected)

    v04:
    • 2.6.35.10-g308ba26 base (new leak) with the same mods as v03

    v05:
    • fixed htc akm8973 ("compass") driver (read more)

    v06:
    • 2.6.35.10-g3f43272 base (official gb update)

    download: zImage-v06 (flashable v05-to-v06 kernel image+modules update package here)


    --

    MODDED WIFI DRIVER
    • min rssi threshold for scan results lowered (-88dBm -> -108dBm)
    • don't send "rssi low" events via wireless_send_event()
    if you have messed up your framework/calibration file/etc with other "fixes", revert them

    download:
    2.6.35.10-gdd56551: bcm4329.ko
    2.6.35.10-g308ba26: bcm4329.ko


    ORIGINAL WIFI DRIVER

    download:
    2.6.35.10-g3f43272: bcm4329.ko


    CIFS
    # insmod /system/lib/modules/nls_utf8.ko
    # insmod /system/lib/modules/cifs.ko
    # busybox mount -t cifs -o user=lolwut,password=hackme,uid=1000,gid=1015,file_mode=0775,dir_mode=0775,iocharset=utf8,nodfs //192.168.40.10/public /sdcard/smb/public
    (removed slow_work_* dependencies and patched with 2.6.31 oplock code)

    download:
    2.6.35.10-gdd56551: nls_utf8.ko, cifs.ko
    2.6.35.10-g308ba26: nls_utf8.ko, cifs.ko
    2.6.35.10-g3f43272: nls_utf8.ko, cifs.ko


    EXT4
    # insmod /system/lib/modules/jbd2.ko
    # insmod /system/lib/modules/ext4.ko
    # busybox mount -t ext4 /dev/block/mmcblk0p3 /cache/lolwut
    # busybox mount | grep mmcblk0p3
    /dev/block/mmcblk0p3 on /cache/lolwut type ext4 (rw,relatime,barrier=1,data=ordered)

    download:
    2.6.35.10-gdd56551: jbd2.ko(dependecy), ext4.ko
    2.6.35.10-g308ba26: jbd2.ko(dependecy), ext4.ko
    2.6.35.10-g3f43272: jbd2.ko(dependecy), ext4.ko


    TUN
    # insmod /system/lib/modules/tun.ko
    # ls -l /dev/tun
    crw-r----- vpn vpn 10, 200 2011-04-17 14:06 tun

    download:
    2.6.35.10-gdd56551: tun.ko
    2.6.35.10-g308ba26: tun.ko
    2.6.35.10-g3f43272: tun.ko


    CPUFREQ: POWERSAVE, INTERACTIVE & SMARTASS
    # insmod /system/lib/modules/cpufreq_interactive.ko
    # insmod /system/lib/modules/cpufreq_smartass.ko
    # insmod /system/lib/modules/cpufreq_powersave.ko
    # cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors
    interactive smartass powersave userspace ondemand performance
    (you don't have to pass symbol addresses)

    download:
    2.6.35.10-gdd56551: cpufreq_interactive.ko, cpufreq_smartass.ko, cpufreq_powersave.ko
    2.6.35.10-g308ba26: cpufreq_interactive.ko, cpufreq_smartass.ko, cpufreq_powersave.ko
    2.6.35.10-g3f43272: cpufreq_interactive.ko, cpufreq_smartass.ko, cpufreq_powersave.ko


    PERFLOCK DISABLER (c) Michael Huang <[email protected]>
    # insmod /system/lib/modules/perflock_disable.ko
    (you don't have to pass symbol addresses)

    download:
    2.6.35.10-gdd56551: perflock_disable.ko (you don't need this with the pre-cracked kernel image above)
    2.6.35.10-g308ba26: perflock_disable.ko (you don't need this with the pre-cracked kernel image above)
    2.6.35.10-g3f43272: perflock_disable.ko (you don't need this with the pre-cracked kernel image above)


    BTRFS

    download:
    2.6.35.10-gdd56551: btrfs.ko
    2.6.35.10-g308ba26: btrfs.ko
    2.6.35.10-g3f43272: btrfs.ko
    29
    hey, sorry for the lack of input, i'm busy fixing the vdd issue
    there will be a separate thread for this serie of kernels w/detailed features/info +modded sources
    27
    new kernel image (v05):

    fixed off-by-one bug (missing value for register [email protected]) in latest htc akm8973 driver that caused calibration failure in akmd
    (that funny HOFFSET*.z values in /data/misc/AK8973Prms.txt)


    original (objdump output, comments by me)
    AKECS_GetData() inlined in akm_work_func() @0xc024d3d4:
    c024d3d4: mov ip, sp
    c024d3d8: push {r4, r5, fp, ip, lr, pc}
    c024d3dc: sub fp, ip, #4
    c024d3e0: sub sp, sp, #8
    c024d3e4: sub r5, fp, #28 @ r5: buffer
    c024d3e8: mov r1, #4 @ r1: RBUFF_SIZE
    c024d3ec: mov r0, r5 @ r0: buffer
    c024d3f0: bl c01ab3a0 @ < memset(buffer, 0, RBUFF_SIZE) < one too few actually, should be RBUFF_SIZE+1
    c024d3f4: mov r0, r5 @ r0: buffer
    c024d3f8: mov r1, #4 @ r1: RBUFF_SIZE
    c024d3fc: mvn r3, #63 @ same as mov r3, 0xffffffc0, and 0xc0 = AKECS_REG_ST
    c024d400: strb r3, [fp, #-28] @ buffer[0] = AKECS_REG_ST
    c024d404: bl c024d2b4 @ < AKI2C_RxData(buffer, RBUFF_SIZE) < one too few actually, should be RBUFF_SIZE+1
    c024d408: cmp r0, #0 @ checking ret val
    c024d40c: blt c024d450 @ goto fail
    c024d410: ldr r4, [pc, #88]
    c024d414: add r0, r4, #44 @ r0: &sense_data_mutex
    c024d418: bl c03a7ad4 @ mutex_lock(&sense_data_mutex)
    c024d41c: ldm r5, {r0, r1} @ copy buffer into r0,r1
    c024d420: mov ip, #1
    c024d424: mov r3, #0
    c024d428: mov r2, ip
    c024d42c: str ip, [r4, #76] @ atomic_set(&data_ready, 1)
    c024d430: str r0, [r4, #68] @ r0 (contents of buffer) into sense_data (part 1)
    c024d434: strb r1, [r4, #72] @ r1 (contents of buffer) into sense_data (part 2)
    c024d438: mov r1, #3
    c024d43c: ldr r0, [pc, #48] @ &data_ready_wq
    c024d440: bl c0072038 @ wake_up(&data_ready_wq)
    c024d444: add r0, r4, #44 @ &sense_data_mutex
    c024d448: bl c03a6f58 @ mutex_unlock(&sense_data_mutex);
    c024d44c: b c024d458 @ ...proceed with enable_irq, etc...

    modified
    ...
    c024d3e8: mov r1, #5
    ...
    c024d3f8: mov r1, #5
    ...
    c024d42c: str r0, [r4, #68]
    c024d430: str r1, [r4, #72]
    c024d434: str ip, [r4, #76] @ data_ready: 1
    ...


    remove b0rked calibration file (might be optional)
    recalibrate your compass
    outdoor calibration highly recommended (magnetic interference)
    21
    bravo-2.6.35.14-snq-110919.zip

    353w6f.jpg


    (to test w/o flashing it unpack zImage from the archive, adb reboot bootloader, fastboot boot zImage (you can't use wifi until the next reboot though))


    optional wifi-n addon
    20
    wow i really need to catch up on comments here

    in the meantime:
    bravo-kernel/bravo-2.6.35.14-snq-111011.zip
    wifin_update-2.6.35.14-snq-111011.zip (optional wifi-n addon)

    bravo-2.6.35.14-snq-111011-no2wcr.zip (w/o callrec support)
    wifin_update-2.6.35.14-snq-111011-no2wcr.zip (optional wifi-n addon for the version w/o callrec support)
Our Apps
Get our official app!
The best way to access XDA on your phone
Nav Gestures
Add swipe gestures to any Android
One Handed Mode
Eases uses one hand with your phone