3.4 kernel bring-up

kabaldan · Jun 12, 2014

By KitKat update for xt897/asanti being canceled, we're missing at least:
updated bootladers (sbl1,sbl2,sbl3 and appsboot), updated trustzone (not sure if that is an issue at this point), updated modem firmware.

Fortunately, at least the device tree for asanti seems to be ready and working fine, apart from not being handled by JB bootloader.

As a base, I'm using https://github.com/CyanogenMod/android_kernel_motorola_msm8960dt-common/tree/cm-11.0 that has been updated by dhacker for support of KitKat updated Razr HD/M.

To circumvent dt not being handled, I'm using appended dtb:
CONFIG_ARM_APPENDED_DTB=y
CONFIG_ARM_ATAG_DTB_COMPAT=y

Device tree partition check revealed that my device is p2 or p2b revision (https://github.com/MotorolaMobility...razrm/arch/arm/boot/dts/msm8960-asanti-p2.dts and https://github.com/MotorolaMobility...azrm/arch/arm/boot/dts/msm8960-asanti-p2b.dts are identical), so I'm using msm8960-asanti-p2.dtb appended to zImage.

Next issue is that bootloader is supposed to add dynamic data on runtime in addition to static ones loaded from dtb.

To get the display working, I've added:

Code:

	chosen {
		/* mipi_mot_cmd_auo_qhd_430 */
		mmi,panel_name = [6d6970695f6d6f745f636d645f61756f5f7168645f34333000];
	};

to msm8960-asanti-p2.dts.

That's where I'm currently. The next step is to add additional entries we're missing there (e.g. "mmi,mbmprotocol" etc.).

Regarding bootloaders, I see another issue - memory configuration.
There's failing shared mem allocation in mmi_unit_info_init:
https://github.com/CyanogenMod/andr...ob/cm-11.0/arch/arm/mach-msm/board-mmi.c#L433
That memory block is supposed to be reserved by sbl3, as indicated by https://github.com/MotorolaMobilityLLC/kernel-msm/commit/585b70b3fcd1af84edcb8748d1fdf190b32784af

Also, this commit https://github.com/MotorolaMobilityLLC/kernel-msm/commit/e3b9a040ca408d217a873df7994b9452ece04d16 indicates that Moto has done some tests also with older modem firmware, at least at some point...

skeevydude · Jun 12, 2014

If you need any logs or whatever from something not a Q, hit me up. Maybe something from my Atrix HD will help you out, I dunno -- AHD is stuck with JB until Q3 this year, so we're interested in this too.

mrvek · Jun 17, 2014

Just 2 questions:

1) that moto test for older modem firmware with newer kernels, patch for asanti device tree (in which they state that device mostly works) might indicate that moto actually has updated bootloaders, etc, which we will never see?
2) this thread in development section is best left for dev disscusson? Ie, no potentially irritating stupid suggestions/questions/ideas/thoughts? (I have to ask)

fatihx · Jun 17, 2014

mrvek said:
Just 2 questions:

1) that moto test for older modem firmware with newer kernels, patch for asanti device tree (in which they state that device mostly works) might indicate that moto actually has updated bootloaders, etc, which we will never see?
2) this thread in development section is best left for dev disscusson? Ie, no potentially irritating stupid suggestions/questions/ideas/thoughts? (I have to ask)

http://xdaforums.com/showthread.php?t=2778330

kabaldan · Jun 23, 2014

mrvek said:
Just 2 questions:

1) that moto test for older modem firmware with newer kernels, patch for asanti device tree (in which they state that device mostly works) might indicate that moto actually has updated bootloaders, etc, which we will never see?
2) this thread in development section is best left for dev disscusson? Ie, no potentially irritating stupid suggestions/questions/ideas/thoughts? (I have to ask)

1 - yes, I'm pretty sure that test builds for asanti exist somewhere in moto labs from the time before 4.4 upgrade for Photon Q cancellation, with updated bootloader images and stuff, that we will unfortunately never see (I'm not very good in social hacking, so I don't expect myself to be able to get access to that internal stuff).

I also noticed this commit - https://github.com/MotorolaMobilityLLC/kernel-msm/commit/14e5887c3db194ec1a8aa6145910f7c7b64f95dc - they did not have enough asanti p2 hw revision devices for testing, so they needed some fixes to be able to test on an older p1 revision of the hardware...

VijayKumar Gn surely wasted some of his time on 4.4 bring-up for asanti before it got cancelled...

2 - It depends on the level of stupidity

Even some very bright ideas may appear stupid at first sight, but can get proved to be great in the end.

At this point, I'm still not able get the BP going, it goes to crash and restart loop as soon as it's brought out of reset by the PIL (peripheral image loader).

I've pushed the kernel commits I use so far here in cm-11.0-3.4kJBbl branch:
https://github.com/nadlabak/android_kernel_motorola_msm8960dt-common/commits/cm-11.0-3.4kJBbl

It should go with something like this kernel setting in BoardConfig.mk:

Code:

# Inline kernel building
TARGET_KERNEL_SOURCE := kernel/motorola/msm8960dt-common
TARGET_KERNEL_CONFIG := msm8960_mmi_defconfig
TARGET_KERNEL_VARIANT_CONFIG := msm8960_mmi_xt897_defconfig
TARGET_KERNEL_SELINUX_CONFIG := msm8960_mmi_selinux_defconfig
BOARD_KERNEL_IMAGE_NAME := zImage-dtb
BOARD_KERNEL_CMDLINE := console=ttyHSL0,115200,n8 androidboot.hardware=qcom user_debug=31 msm_rtb.filter=0x3F ehci-hcd.park=3 maxcpus=2 vmalloc=400M androidboot.write_protect=0 zcache androidboot.selinux=permissive androidboot.modelno=XT897 product=asanti_c androidboot.device=asanti_c androidboot.secure_hardware=1
BOARD_KERNEL_BASE := 0x80200000
BOARD_KERNEL_PAGESIZE := 2048
BOARD_MKBOOTIMG_ARGS := --ramdisk_offset 0x02200000
BOARD_USERDATAIMAGE_PARTITION_SIZE := 12884901888

kabaldan · Jun 23, 2014

OK, good news, I've got the BP running fine now, under 3.4 kernel.

Loader009 · Jun 23, 2014

Just a little off-topic:
Do we have some indicators to differentiate between p1 and p2 devices?

niko99 · Jun 23, 2014

kabaldan said:
OK, good news, I've got the BP running fine now, under 3.4 kernel.

AAAAAAAAAAAAAMAAAAAAAAZINNNNNNGGGGGGGGGGGG

cant wait for some first beta/test/kernel for a try ! thanks very good news very !

kabaldan · Jun 23, 2014

Loader009 said:
Just a little off-topic:
Do we have some indicators to differentiate between p1 and p2 devices?

You can check e.g. /proc/device-tree/System@0/EMUDetection@0 on your device.
If it contains Gpio-dmb-ppd@0, your device is P2.
If Gpio-dmb-ppd@0 is missing, but Gpio-id-en@0 and Gpio-id@0 are there, your device is P1.

It would be useful for me to know if there's actually a P1 device running CM11 out there somewhere.

Loader009 · Jun 23, 2014

Ok, P2 also here.

Is it possible that P1 devices are prototypes?
Someone here had a prototype, I wouldn't wonder if it was a P1 device.

JoseBerga · Jun 26, 2014

@kabaldan this would work also with xt925 with JB bl.. Because Motorola will not release kk for us until Q3

If you need logs or something for xt925 count with me..

Sent from my RAZR HD using Forum Fiend v1.2.8.

CornholioGSM · Jun 27, 2014

kabaldan said:
You can check e.g. /proc/device-tree/System@0/EMUDetection@0 on your device.
If it contains Gpio-dmb-ppd@0, your device is P2.
If Gpio-dmb-ppd@0 is missing, but Gpio-id-en@0 and Gpio-id@0 are there, your device is P1.

It would be useful for me to know if there's actually a P1 device running CM11 out there somewhere.

...my photon is one of first and it is P2

comstyle · Jul 1, 2014

p2 here too

palmwangja · Jul 2, 2014

P2 too...

kabaldan · Jul 4, 2014

Just to give a 3.4 kernel on Q current status report:
It's not stable yet. Occasionally, there's a kernel panic, usually when going to suspend state.
Will need to figure out the cause.
Otherwise, everything is basically working: radio, gps, nfc, wlan (while using the new wcnss firmware), bt, audio, camera, usb...

fatihx · Jul 4, 2014

can this commit may help for our kernel-panic problem .?

https://github.com/edoko/Air_Kernel-NOTE/commit/5cd3f27ce5373fa3e84b7916480cebaa4ebc9df1

gehstock · Jul 9, 2014

Kernel 3.4 working?

http://review.cyanogenmod.org/#/c/67589/

palmbeach05 · Jul 14, 2014

gehstock said:
Kernel 3.4 working?

http://review.cyanogenmod.org/#/c/67589/

If the Atrix HD is not running a 3.4 kernel yet, then no, it does not work. Be patient until something is announced.

kabaldan · Jul 18, 2014

@stargo , @Skrilax_CZ , @mrvek , @arrrghhh and others:
Sorry for the delay. I've been too busy at work and home again (surprise!).

Anyway, I've finally managed to push my current local mess to razrqcom, reusing the old xt897 and msm8960-common repos for tentative 3.4 kernel on JB firmware builds.
Local manifest:

Code:

<?xml version="1.0" encoding="UTF-8"?>
<manifest>
  <project name="razrqcom-dev-team/android_device_motorola_xt897" path="device/motorola/xt897" remote="github" revision="cm-11.0-3.4" />
  <project name="razrqcom-dev-team/android_device_motorola_msm8960-common" path="device/motorola/msm8960-common" remote="github" revision="cm-11.0-3.4" />
  <project name="razrqcom-dev-team/proprietary_vendor_motorola" path="vendor/motorola" remote="github" revision="cm-11.0-3.4" />
</manifest>

The main issues remain the same:
- while the modem subsystem can occasionally work stable for tens of minutes, it can also get being restarted by watchdog bite in matter of minutes or even seconds (=too often).
- occasional kernel panic (NULL pointer dereference) at process_one_work coming immediately after hitting this warning https://github.com/CyanogenMod/andr...t-common/blob/cm-11.0/kernel/workqueue.c#L550

Skrilax_CZ · Jul 20, 2014

No worries, I'm extremely busy lately too with practically no time for this stuff. I'm in China atm. And it's very hot in here (real feel: 44° C).

3.4 kernel bring-up

Inactive Recognized Developer

Inactive Recognized Contributor

Senior Member

Member

Inactive Recognized Developer

Inactive Recognized Developer

Attachments

Senior Member

Senior Member

Inactive Recognized Developer

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Inactive Recognized Developer

Member

Senior Member

Senior Member

Inactive Recognized Developer

Inactive Recognized Developer

Similar threads

Top Liked Posts