HP Omi 10 (baytrail), Android 4.4 & 5.0 bcm43241 not detected

[JJ Meijer]

Hello everybody,

First of all thank you for all your hard work!

I am trying to get the wifi working on my HP Omni 10 (baytrail) but the sdio card is not detected. I tried Android 4.4 and 5.0 with kernels 3.18.0-android-x86 and 3.19-rc2-android-x86 with defconfig.

The bus is setup like this:
mmc0: storage (80860F14:00)
mmc1: bcm43241 (INT33BB:00)
mmc2: sd card (80860F14:01)

In Android bcm43241 fails to load(Storage and SD card work fine) :

<6>[   31.077863] sdhci: Secure Digital Host Controller Interface driver
<6>[   31.083536] sdhci: Copyright(c) Pierre Ossman
<6>[   31.089867] sdhci-acpi 80860F14:00: No vmmc regulator found
<6>[   31.094721] sdhci-acpi 80860F14:00: No vqmmc regulator found
<6>[   31.102326] usb 1-2: new high-speed USB device number 2 using xhci_hcd
<6>[   31.108186] mmc0: SDHCI controller on ACPI [80860F14:00] using ADMA
<6>[   31.113938] sdhci-acpi INT33BB:00: No vmmc regulator found
<6>[   31.118521] sdhci-acpi INT33BB:00: No vqmmc regulator found
<6>[   31.128537] mmc1: SDHCI controller on ACPI [INT33BB:00] using ADMA
<4>[   31.135984] sdhci-acpi 80860F14:01: failed to setup card detect gpio
<6>[   31.142421] sdhci-acpi 80860F14:01: No vmmc regulator found
<6>[   31.147970] sdhci-acpi 80860F14:01: No vqmmc regulator found
<6>[   31.158181] mmc2: SDHCI controller on ACPI [80860F14:01] using ADMA

The only error message that I can find is: failed to setup card detect gpio


In Ubuntu 14.10 (3.18.1 mainline) wifi is detected:

Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.202544] sdhci: Secure Digital Host Controller Interface driver
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.202545] sdhci: Copyright(c) Pierre Ossman
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.204056] sdhci-acpi 80860F14:00: No vmmc regulator found
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.204059] sdhci-acpi 80860F14:00: No vqmmc regulator found
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.207306] mmc0: SDHCI controller on ACPI [80860F14:00] using ADMA
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.208441] sdhci-acpi INT33BB:00: No vmmc regulator found
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.208443] sdhci-acpi INT33BB:00: No vqmmc regulator found
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.211597] mmc1: SDHCI controller on ACPI [INT33BB:00] using ADMA
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.212778] sdhci-acpi 80860F14:01: No vmmc regulator found
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.212780] sdhci-acpi 80860F14:01: No vqmmc regulator found
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.215985] mmc2: SDHCI controller on ACPI [80860F14:01] using ADMA
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.369074] mmc0: BKOPS_EN bit is not set
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.398039] mmc0: new HS200 MMC card at address 0001
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.407662] mmcblk0: mmc0:0001 HBG4e 29.1 GiB
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.413884] mmcblk0boot0: mmc0:0001 HBG4e partition 1 4.00 MiB
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.419465] mmcblk0boot1: mmc0:0001 HBG4e partition 2 4.00 MiB
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.439030]  mmcblk0: p1 p2 p3
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.496934] mmc1: queuing unknown CIS tuple 0x80 (2 bytes)
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.502904] mmc1: queuing unknown CIS tuple 0x80 (3 bytes)
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.508802] mmc1: queuing unknown CIS tuple 0x80 (3 bytes)
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.515812] mmc1: queuing unknown CIS tuple 0x80 (7 bytes)
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.590071] mmc1: new ultra high speed DDR50 SDIO card at address 0001
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.681457] EXT4-fs (mmcblk0p2): mounted filesystem with ordered data mode. Opts: (null)
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.829991] mmc2: new high speed SDHC card at address aaaa
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.835792] mmcblk1: mmc2:aaaa SL32G 29.7 GiB (ro)
Jan  6 18:23:38 jjm-HP-Omni10 kernel: [   32.853317]  mmcblk1: p1

I tried merging a lot of kernel config settings from ubuntu to android (sdhci & mmc related) but still the same result. Could somebody please point me in the right direction?

[Chih-Wei Huang]

Thank you for the detailed report.

I believe there are some fixes either
in kernel 3.18.1 (while we use 3.18),
or in ubuntu's kernel package.

To test kernel 3.18.1 (or newer stable release),
try to merge it from upstream. You can do this:
(on the android-x86 source dir)

cd kernel
git remote add stable
http://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git
git pull stable refs/tags/v3.18.1
(fortunately there should be no conflict, just commit it)

Then build the android-x86 image and test it.

If the fixes are not in 3.18.1, check what patches
that ubuntu kernel deb have and try to apply them
to the android-x86 kernel.

> --
> You received this message because you are subscribed to the Google Groups
> "Android-x86" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to android-x86...@googlegroups.com.
> To post to this group, send email to andro...@googlegroups.com.
> Visit this group at http://groups.google.com/group/android-x86.
> For more options, visit https://groups.google.com/d/optout.



--
Chih-Wei
Android-x86 project
http://www.android-x86.org

[JJ Meijer]

Thank you for your fast response! I added Ubuntu syslog and android dmesg as I may have missed something, maybe you can take a look at it?

Some more information:

ls /sys/bus/mmc/devices
mmc0:0001
mmc1:0001 (missing in android)
mmc2:aaaa

ls /sys/bus/sdio/devices
empty in android

I blacklisted brcmfmac in Ubuntu to rule out firmware issues and mmc1:0001 will still be detected (but no wifi of course).


Also here is the merge report but I dont see changes that might help but I will build the image and try the extra Ubuntu patches tomorrow:

Makefile                                        |  2 +-
 arch/mips/net/bpf_jit.c                         |  4 +-
 arch/powerpc/platforms/cell/spufs/inode.c       |  2 +-
 drivers/net/ethernet/marvell/mvneta.c           |  5 +-
 drivers/net/vxlan.c                             | 10 +---
 drivers/net/wireless/rtlwifi/rtl8192ce/hw.c     |  3 +-
 drivers/net/wireless/rtlwifi/rtl8192ce/sw.c     |  3 +
 drivers/net/wireless/rtlwifi/rtl8192ce/trx.c    | 24 ++++++--
 drivers/net/wireless/rtlwifi/rtl8192ce/trx.h    |  2 +
 drivers/net/xen-netfront.c                      |  3 +
 drivers/staging/lustre/lustre/llite/dcache.c    |  2 +-
 drivers/staging/lustre/lustre/llite/llite_lib.c |  2 +-
 drivers/staging/lustre/lustre/llite/namei.c     |  8 +--
 fs/affs/amigaffs.c                              |  2 +-
 fs/autofs4/expire.c                             | 12 ++--
 fs/autofs4/root.c                               |  2 +-
 fs/ceph/dir.c                                   |  8 +--
 fs/ceph/inode.c                                 |  2 +-
 fs/cifs/inode.c                                 |  2 +-
 fs/coda/cache.c                                 |  2 +-
 fs/dcache.c                                     | 80 ++++++++++++-------------
 fs/debugfs/inode.c                              |  2 +-
 fs/exportfs/expfs.c                             |  2 +-
 fs/libfs.c                                      | 12 ++--
 fs/ncpfs/dir.c                                  |  2 +-
 fs/ncpfs/ncplib_kernel.h                        |  4 +-
 fs/nfs/getroot.c                                |  2 +-
 fs/notify/fsnotify.c                            |  4 +-
 fs/ocfs2/dcache.c                               |  2 +-
 include/linux/dcache.h                          |  8 +--
 kernel/trace/trace.c                            |  4 +-
 kernel/trace/trace_events.c                     |  2 +-
 net/ipv4/gre_offload.c                          |  3 +
 net/ipv4/tcp_ipv4.c                             |  4 +-
 net/ipv6/tcp_ipv6.c                             | 28 ++++-----
 net/netlink/af_netlink.c                        |  2 +-
 net/openvswitch/flow_netlink.c                  |  4 +-
 net/sched/sch_fq_codel.c                        |  2 +-
 net/sctp/output.c                               |  4 +-
 security/selinux/selinuxfs.c                    |  6 +-
 sound/pci/hda/patch_analog.c                    |  1 +
 sound/pci/hda/patch_realtek.c                   |  1 +
 sound/usb/midi.c                                |  2 +
 43 files changed, 155 insertions(+), 126 deletions(-)

[JJ Meijer]

A small update, the sdio card is detected now!

Merging 3.18.1 didn't help, the problem appears to be in the kernel config. I took config-3.18.1-031801-generic and added the Android and security settings from android-x86_defconfig also I made the mmc drivers compile into the kernel.

Sadly now HW acceleration is broken and Android only starts with nomodeset, I'm not sure if it's the new config or the merge that caused this. I attached the new config file.

Op woensdag 7 januari 2015 04:55:22 UTC+1 schreef Chih-Wei Huang:

[Jocky Wilson]

Hi JJ, I have similar problems where sdio device tree is not populated on my vayo a1 mini baytrail tablet. I have built a 3.18.1 kernel with your attached config. When I boot it of a fat32 usb stick it gets stuck in a boot loop: constantly showing something like character support for iso-8859 not found. I tried to fix this by changing back values to original values, but to no avail. Can you post your kernel?

Cheers,
JockyW

[JJ Meijer]

No problem! I tried to address your problems and also HW acceleration is working again (forgot to re-enable the GPU authentication hack) and  home & power buttons are working now. Used this patch to make the home button usable: https://github.com/AdamWill/baytrail-m/blob/master/kernel/soc_button_use_leftmeta.patch


https://drive.google.com/folderview?id=0B9vJ9JSuyPijaVNZZHNuaDJDWUE&usp=sharing

Cheers,

Op zaterdag 10 januari 2015 11:59:25 UTC+1 schreef Jocky Wilson:

[Vaidotas]

This is amazing. Finally accelerated hardware is working on Lenovo Miix 2 8 tablet. "Home" and volume buttons work too. Power button responds, but its function is to restart(?) android.
Could you explain what is "GPU authentication hack"?

[pstglia]

Guess he's talking about one of the patches applied by Chih-Wei to enable drm working on Android-x86

"

http://git.android-x86.org/?p=kernel/common.git;a=commit;h=a14b6b3c8d2e4de989954987a33bb73082b5a45d

HACK: drm: disable GPU authentication

Signed-off-by: Chih-Wei Huang <cwh...@linux.org.tw>

[JJ Meijer]

Yes your right, sorry for being a bit unclear about that :)

Regarding the power button, Android (or something else...) is actually crashing trying to suspend :P but you can see it actually saved and restored it's state properly, not yet sure whats wrong... But at least you can now properly shut your tablet down is you keep it pressed till the power menu comes up :)

Audio is working now as well after adding the missing firmware file and needed to change the sample rate to 48kHz in the android HAL driver.

Op zaterdag 10 januari 2015 16:04:23 UTC+1 schreef pstglia:

[pstglia]

Thanks for your nice work. Lots of baytrail people will be happier now :)

Can you create a patch with your changes, so Chih-Wei could review and merge on Android-x86 git?

Audio is working now as well after adding the missing firmware file and needed to change the sample rate to 48kHz in the android HAL driver.

 

Can you change sampling rate by adding some parameter to audio driver? I know some drivers accept it (like snd_atiixp - you can change sampling rate by setting ac97_clock parameter):

ex: modprobe snd_atiixp ac97_clock=44100

[JJ Meijer]

No problem, happy to be able to contribute to the project! At the moment I'm not sure yet which setting actually made the MMC stuff to work properly, maybe you have an idea? Also the keyboard is broken now during booting when starting Android with DEBUG=1(or 2) parameter.

The firmware being loaded is fw_sst_0f28.bin-48kHz_i2s_master from Intel, the filename suggests it's 48kHz only but I have no clue about that either (yet). I tried changing defaults.pcm.dmix.rate to 48000 in alsa.conf but this seems to have no effect. For now only changing the hardcoded value in audio_hw.c helped.

Op zaterdag 10 januari 2015 16:25:12 UTC+1 schreef pstglia:

[Chih-Wei Huang]

2015-01-11 0:05 GMT+08:00 JJ Meijer <jjmei...@gmail.com>:
> No problem, happy to be able to contribute to the project! At the moment I'm
> not sure yet which setting actually made the MMC stuff to work properly,
> maybe you have an idea? Also the keyboard is broken now during booting when
> starting Android with DEBUG=1(or 2) parameter.
>
> The firmware being loaded is fw_sst_0f28.bin-48kHz_i2s_master from Intel,
> the filename suggests it's 48kHz only but I have no clue about that either
> (yet). I tried changing defaults.pcm.dmix.rate to 48000 in alsa.conf but
> this seems to have no effect. For now only changing the hardcoded value in
> audio_hw.c helped.

Do you build the lollipop-x86 branch?

The 48000 fix has already been in the kitkat-x86 branch.
Just found it's not in lollipop-x86 yet.
(I've merged it but forgot to push it to the server)

Now it's pushed.
You can sync to test again.
Thanks!

[Chih-Wei Huang]

2015-01-09 4:01 GMT+08:00 JJ Meijer <jjmei...@gmail.com>:
> A small update, the sdio card is detected now!
>
> Merging 3.18.1 didn't help, the problem appears to be in the kernel config.
> I took config-3.18.1-031801-generic and added the Android and security
> settings from android-x86_defconfig also I made the mmc drivers compile into
> the kernel.
>
> Sadly now HW acceleration is broken and Android only starts with nomodeset,
> I'm not sure if it's the new config or the merge that caused this. I
> attached the new config file.

The newconfig differs from ours largely.
Could you check which one (or more) causes the difference?
I suggest check the ACPI part first.

[JJ Meijer]

Ah yes, I see now, thanks! I indeed build the lollipop-x86 branch. Should the value also be changed here? I changed both at the same time, so not sure > http://git.android-x86.org/?p=platform/hardware/libhardware.git;a=blob;f=modules/audio/audio_hw.c;h=18c0e59404b423ff0127835720e565e9cfd51bae;hb=HEAD

43 static uint32_t out_get_sample_rate(const struct audio_stream *stream)
44 {
45     return 44100;
46 }
47

Op zaterdag 10 januari 2015 18:05:17 UTC+1 schreef Chih-Wei Huang:

[Chih-Wei Huang]

2015-01-11 1:18 GMT+08:00 JJ Meijer <jjmei...@gmail.com>:
> Ah yes, I see now, thanks! I indeed build the lollipop-x86 branch. Should
> the value also be changed here? I changed both at the same time, so not sure
>>
> http://git.android-x86.org/?p=platform/hardware/libhardware.git;a=blob;f=modules/audio/audio_hw.c;h=18c0e59404b423ff0127835720e565e9cfd51bae;hb=HEAD

This is the AOSP default dummy audio HAL
which is not used by android-x86.
(Actually audio.primary.default.so should be removed
from the installed list. I'll check that later)

[pstglia]

No problem, happy to be able to contribute to the project! At the moment I'm not sure yet which setting actually made the MMC stuff to work properly, maybe you have an idea? Also the keyboard is broken now during booting when starting Android with DEBUG=1(or 2) parameter.

As Chih-Wei pointed, lots of params are different from your kernel config to Android-x86 one.

I've listed them in the attached file (options_not_set_android_x86_kernel_config.txt). As the failure seems to occurs on mmc_gpiod_request_cd I suspect these ones (Not enabled on Android-x86):

CONFIG_MTD_NAND_GPIO ( Device Drivers > Memory Technology Device (MTD) support > "NAND Device Support / GPIO NAND Flash driver" option )

CONFIG_SPI_GPIO ( Device Drivers > SPI support -> "GPIO-based bitbanging SPI Master" option)

Could you enable these options on Android-x86 kernel config and test please?

Thank you!

pstglia

[JJ Meijer]

Building right now, thanks for the suggestions! I saw some nice pushes at git, this will become a nice Lollipop tablet! Sadly the tablet still freezes a lot and needs reset due to mmc timeout error -110.

I also noticed the config difference was quite big ;-)

Op zaterdag 10 januari 2015 20:39:04 UTC+1 schreef pstglia:

[JJ Meijer]

I tested the new kernel and the "sdhci-acpi 80860F14:01: failed to setup card detect gpio" message is gone but the sdio card is not detected.

Op zaterdag 10 januari 2015 20:39:04 UTC+1 schreef pstglia:

[Jocky Wilson]

I can confirm that. Now building a kernel with JJ's newconfig in order to find out if that can detect the sdio card.

--
You received this message because you are subscribed to a topic in the Google Groups "Android-x86" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/android-x86/-i-D2C7hdZk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to android-x86...@googlegroups.com.

[pstglia]

I tested the new kernel and the "sdhci-acpi 80860F14:01: failed to setup card detect gpio" message is gone but the sdio card is not detected. 

Try enabling these also:

CONFIG_MMC_RICOH_MMC ( Device Drivers > MMC/SD/SDIO card support / "Ricoh MMC Controller Disabler"  )

CONFIG_MTD_NAND_ECC_BCH (Device Drivers > Memory Technology Device (MTD) support > NAND Device Support / "Support software BCH ECC")

CONFIG_MTD_NAND_DENALI (Device Drivers > Memory Technology Device (MTD) support > NAND Device Support / "Support Denali NAND controller" )

CONFIG_MTD_NAND_RICOH ( Device Drivers > Memory Technology Device (MTD) support > NAND Device Support / "Ricoh xD card reader")

CONFIG_MTD_NAND_DISKONCHIP (Device Drivers > Memory Technology Device (MTD) support > NAND Device Support / "Device Drivers > Memory Technology Device (MTD) support > NAND Device Support ")

CONFIG_MTD_NAND_DOCG4 (Device Drivers > Memory Technology Device (MTD) support > NAND Device Support / "Support for DiskOnChip G4")

CONFIG_MTD_NAND_CAFE (Device Drivers > Memory Technology Device (MTD) support > NAND Device Support / "NAND support for OLPC CAFÃ~I chip")

CONFIG_MTD_NAND_CS553X (Device Drivers > Memory Technology Device (MTD) support > NAND Device Support / "NAND support for CS5535/CS5536 (AMD Geode companion chip)")

CONFIG_MTD_NAND_NANDSIM (Device Drivers > Memory Technology Device (MTD) support > NAND Device Support / "Support for NAND Flash Simulator")

CONFIG_MTD_NAND_PLATFORM (Device Drivers > Memory Technology Device (MTD) support > NAND Device Support / "Support for generic platform NAND driver")

CONFIG_MTD_PMC551 (Device Drivers > Memory Technology Device (MTD) support > Self-contained MTD device drivers / "Ramix PMC551 PCI Mezzanine RAM card support")

CONFIG_MTD_DATAFLASH (Device Drivers > Memory Technology Device (MTD) support > Self-contained MTD device drivers / "Support for AT45xxx DataFlash")

CONFIG_MTD_SST25L (Device Drivers > Memory Technology Device (MTD) support > Self-contained MTD device drivers / "Support SST25L (non JEDEC) SPI Flash chips")

CONFIG_MTD_SLRAM (Device Drivers > Memory Technology Device (MTD) support > Self-contained MTD device drivers /  "Uncached system RAM")

CONFIG_MTD_PHRAM (Device Drivers > Memory Technology Device (MTD) support > Self-contained MTD device drivers / "Physical system RAM")

CONFIG_MTD_BLOCK2MTD (Device Drivers > Memory Technology Device (MTD) support > Self-contained MTD device drivers / "MTD using block device")

CONFIG_MTD_DOCG3 (Device Drivers > Memory Technology Device (MTD) support > Self-contained MTD device drivers / "M-Systems Disk-On-Chip G3")

CONFIG_MTD_CFI (Device Drivers > Memory Technology Device (MTD) support > RAM/ROM/Flash chip drivers / "Detect flash chips by Common Flash Interface (CFI) probe")

CONFIG_MTD_JEDECPROBE (Device Drivers > Memory Technology Device (MTD) support > RAM/ROM/Flash chip drivers / "Detect non-CFI AMD/JEDEC-compatible flash chips")

CONFIG_MTD_RAM (Device Drivers > Memory Technology Device (MTD) support > RAM/ROM/Flash chip drivers / "Support for RAM chips in bus mapping")

CONFIG_MTD_ROM (Device Drivers > Memory Technology Device (MTD) support > RAM/ROM/Flash chip drivers / "Support for ROM chips in bus mapping")

CONFIG_MTD_ABSENT (Device Drivers > Memory Technology Device (MTD) support > RAM/ROM/Flash chip drivers / "Support for absent chips in bus mapping")

CONFIG_MTD_COMPLEX_MAPPINGS (Device Drivers > Memory Technology Device (MTD) support > Mapping drivers for chip access / "Support non-linear mappings of flash chips")

CONFIG_MTD_PHYSMAP (Device Drivers > Memory Technology Device (MTD) support > Mapping drivers for chip access / "Flash device in physical memory map")

CONFIG_MTD_INTEL_VR_NOR (Device Drivers > Memory Technology Device (MTD) support > Mapping drivers for chip access / "NOR flash on Intel Vermilion Range Expansion Bus CS0")

CONFIG_MTD_PLATRAM (Device Drivers > Memory Technology Device (MTD) support > Mapping drivers for chip access / "Map driver for platform device RAM (mtd-ram)")

# Not sure about these 2 ones... maybe you could enable then on a 2nd round if necessary...

CONFIG_USB_EHCI_HCD_PLATFORM ( Device Drivers > USB support / "Generic EHCI driver for a platform device" )

CONFIG_USB_OHCI_HCD_PLATFORM ( Device Drivers > USB support / "Generic OHCI driver for a platform device" )

[Jocky Wilson]

Tried on my Voyo A1 mini, but no joy. Perhaps some options not set on JJ Meijer's kernel can make a difference. Bit hard to find out.

[JJ Meijer]

Finally found it, it was a combination of just two settings:
CONFIG_PINCTRL_BAYTRAIL=y
CONFIG_SPI_GPIO=m

Now with the latest commits and these config additions a lot is working out of the box:
wifi
HW acceleration
audio
backlight
HW buttons (windows + power)
battery status
multi touch
...

One strange thing: HW acceleration broke once but after cleaning data it was working again...


Thank you all!


Up next(if I can pull it of): volume buttons, stand by(or screen of), sensors, ...

Op zondag 11 januari 2015 01:08:08 UTC+1 schreef pstglia:

[pstglia]

Great job!!! Congrats!

[Chih-Wei Huang]

2015-01-13 7:10 GMT+08:00 JJ Meijer <jjmei...@gmail.com>:
> Finally found it, it was a combination of just two settings:
> CONFIG_PINCTRL_BAYTRAIL=y
> CONFIG_SPI_GPIO=m

Unfortunately I removed CONFIG_PINCTRL_BAYTRAIL=y
on purpose since it causes several strange issues on my T100.
For example, the camera led is always on,
can't detect usb stick unless re-plugin...
There are bugs to be fixed.
But I'm not the expert...

[Nathan Schmidt]

Have you tried leaving it enabled? I am wondering if we just have a seperate USB issue. Almost sou D's like this might fix more than it breaks.

Anyone have the omni kernel compiled with those changes hosted anywhere?

[Fajar A. Nugraha]

I tried kitkat default kernel & config + pull 3.8.1 + CONFIG_PINCTRL_BAYTRAIL=y, CONFIG_SPI_GPIO=y on T100. 

LED is always on (I can live with that), SD card is detected (awesome), but touch screen stopped working :(

That's weird since touch screen works on ubuntu (which also has LED always on and can detect SD card)

Oh well, back to testing and waiting.

I wonder if this mmc fix is applicable to 3.8.1 as well?

https://plus.google.com/115582184723609367547/posts/hroVAmAywBH

-- 

Fajar

[JJ Meijer]

I'm getting some strange results as well. At some boots the sdio card is not detected and hw acceleration is not working. With Ubuntu this is not the case so we probably have some more configuring to do :(

Also very curious if the patch from brainwreck works for me, those I/O errors are very annoying..

[Fajar A. Nugraha]

Update on my T100: CONFIG_PINCTRL_BAYTRAIL=y, CONFIG_SPI_GPIO=m (NOT y, which I used earlier) does indeed allow to detect sd card while still keeping touch screen functionality (which is gone with CONFIG_SPI_GPIO=y). Thanks for the tests, JJ.

No luck with the mmc fix so far, as even with the edited source it still hangs after some time both in Ubuntu (3.19-rc4) and Android (3.18.1)

-- 

Fajar

[JJ Meijer]

I ran into some stability issues like HW acceleration suddenly stopped working and a lot of times the brcm sdio device was not detected during boot. I then took the ACPI and  Pin controller settings from the Ubuntu config and this solved the issues. I attached the config if anybody wants to test it.
Also a prebuild iso-img with the brain wreck patch which seems to make it freeze a bit less: https://drive.google.com/open?id=0B9vJ9JSuyPijaVNZZHNuaDJDWUE&authuser=1

[JJ Meijer]

Even all sensors are working now, this is getting better and better! (just need to swap some axes :) ) Thanks to all the devs for making this possible.

[Chih-Wei Huang]

2015-01-14 5:38 GMT+08:00 JJ Meijer <jjmei...@gmail.com>:
> Even all sensors are working now, this is getting better and better! (just
> need to swap some axes :) ) Thanks to all the devs for making this possible.

What did you do to make it work?
Or it just work?

What's the result of ;getprop hal.sensors'?
(or getprop ro.hardware.sensors in lollipop-x86)

[JJ Meijer]

Sorry I was cheering to soon. Although the sdio card is now always detected, HW acceleration is only working 25% of the time after boot. Also only when HW acceleration is working the sensors are working as well.

Sensors where just working without changing anything so this was a big surprise :).

getprop ro.hardware.sensors
when not working: kbd
when working: hsb

I added dmesg and lsmod when working and when not. Maybe some conflicting modules? I must admit I have not looked into them yet.


Op woensdag 14 januari 2015 04:30:05 UTC+1 schreef Chih-Wei Huang:

[Chih-Wei Huang]

2015-01-14 16:20 GMT+08:00 JJ Meijer <jjmei...@gmail.com>:
> Sorry I was cheering to soon. Although the sdio card is now always detected,
> HW acceleration is only working 25% of the time after boot. Also only when
> HW acceleration is working the sensors are working as well.
>
> Sensors where just working without changing anything so this was a big
> surprise :).

The tablet use intel sensorhub.
We have the hal. Good.

> getprop ro.hardware.sensors
> when not working: kbd
> when working: hsb
>
> I added dmesg and lsmod when working and when not. Maybe some conflicting
> modules? I must admit I have not looked into them yet.

Seems a timing issue to me after diff the two files...

What's getprop ro.hardware.gralloc in both cases?

[JJ Meijer]

Great stuff the HAL is already in.

In both cases I get drm from ro.hardware.gralloc. Do you have suggestions how I can debug this timing issue further?

[JJ Meijer]

After looking at the log files I also could not find any obvious difference what could explain my problems. I tried modifying the config some more (using ubuntu config as a guide) and synced to android-x86 5.0.2, kernel 3.18.2, now HW acceleration, sensors and wifi are always working and the tablet seems to not freeze anymore, as long as I start the tablet without charger.

When the charger is connected HW acceleration almost never works. Maybe this could be a hint where the actual problem is?

I must admit I did not test without charger and the default config as I do not have that much time to spare, maybe somebody else could check this?

I attached the new config and also a iso_img at: https://drive.google.com/open?id=0B9vJ9JSuyPijaVNZZHNuaDJDWUE&authuser=1

[Chih-Wei Huang]

Have I said the HW acceleration also stop working
on the T100 sometimes?
Possibly a general issue of baytrail platform.

Actually I think it doesn't really stop working.
Each evidence shows the mesa loads and works,
but it just runs very slow...
I usually test without charger. I don't see the difference.

I can only guess it's a kernel (emmc?) or mesa issue.
Probably update to newer version may be helpful.

[JJ Meijer]

Your problem sounds to be a little different, my tablet seems to decide during boot it it will run fast or not. Without charger it will always start to run fast, very strange. The only times it got slow during running is because of mmc I/O timeout error and it will just stop working, I hope somebody will find a solution for this.

I will try to update mesa but this will probably take me some time :)


Op maandag 19 januari 2015 04:43:47 UTC+1 schreef Chih-Wei Huang:

[Vaidotas]

@JJ Meijer, could you post your dmesg.log from successful boot (accelerated HW) with drm.debug=14 kernel parameter? I'd like to compare it to my own log, to see what goes wrong on my tablet.

[JJ Meijer]

No problem! Please note I just build a new kernel using the full Ubuntu config with minimal changes (mmc, security, gpu authentication hack, networking). It appears that HW acceleration is now always working so maybe it really is a config problem? (or timing?)

Op maandag 19 januari 2015 19:38:57 UTC+1 schreef Vaidotas:

[JJ Meijer]

It seems I broke audio with my previous config, here is a new one. Also the modified Ubuntu config build is now on my drive if anyone wants to test it, HW acceleration and sensors are still working after quite a few reboots (w/wo charger). It bothers me that I can't explain why...

https://drive.google.com/open?id=0B9vJ9JSuyPijaVNZZHNuaDJDWUE&authuser=1

[Jocky Wilson]

Hi JJ, thanks for all your hard testing and building work. I tried your image and , the kernel boots up and then it hangs forever in "Detecting Android-x86 ..........". I tried it with two different usb 4GB sticks. Any ideas?

[JJ Meijer]

Sorry, I forgot to add the drivers to detect the filesystem on usb flash drives. All should be good now, at least here it now works nicely with a fat32 boot and ext4 data partition :).

Op dinsdag 20 januari 2015 12:10:37 UTC+1 schreef Jocky Wilson:

[Jocky Wilson]

Many thanks. I have built kitkat and it works fine on my Voyo A1 Mini tablet.

[JJ Meijer]

No problem and great tot hear it also works for you. Maybe the devs can add the config to the default configs as a "large" kernel config or something :)

[Jocky Wilson]

Some more feedback: the biggest advantage is that HW acceleration works stable. It would be nice to find out which kernel config definitions are responsible for that.

On my Voyo tablet the orientation and accel sensors are loaded and getprop ro.hardware.sensors=hsb, but display orientation change doesn't work.

On my Lenovo tablet not even the drivers are loaded and getprop ro.hardware.sensors=kbd. I modified /etc/init.sh and force loaded drivers and setprop hardware.sensors to hsb, but still no luck.

[marko poljansek]

Hello JJ Meijer
I tried using your "android_x86_5.0.2_3.18.2_baytrail_gapps.iso" build on lenovo miix 2 10 and at first it did not even boot then I added "androidboot.hardware=android_x86" in boot parameters and by surprise it booted with hw acceleration I am booting it from internal storage, dual boot with Windows using rEFInd. "android_x86_5.0.2_3.18.2_ubuntu_gapps.iso" build doesn't boot for me. I tested some functionality:

Working:
HW acceleration
WI-FI (finds and connects to networks)
Apps (mostly working)
touchscreen
HW keyboard
volume controls (but no sound)
battery (shows procents) + charging

partialy working:
autorotate (not every time it boots)

Not working:
brightness (always on max)
audio
bluetooth
camera
youtube (fails to play video)
home + power button
suspend fails (restarts android)

[JJ Meijer]

Good to hear your experiences. I also noticed sometimes after boot the sensors don't work. Maybe you could do a simple sensor check using a sensor app from the market?

Still a lot of work to do but at least it is fast and quite stable now.

[marko poljansek]

Ok I tested couple of reboots and i found that if I forcefully turn off tablet when in Android (presing power button for 30seconds) and turn on again I have autorotate functioning and sensors loaded if I type lsmod in terminal. Installed AnTuTu Benchmark and it shows these sensors: Direction; G-sensor; Acceleration; Light; E-Compass; Gyro; Linear Acceleration; Rotation Vector.
On tester Linear Acceleration and Rotation Vector don't work all other seems working.

Score on AnTuTu: 35054 (not bad for an unsupported device)