S805/S812 Device Tree upstreaming (was: Testing.)

S805/S812 Device Tree upstreaming (was: Testing.)

[Martin Blumenstingl]

Hi Oleg,

On Sat, May 13, 2017 at 9:25 PM, Oleg <balbes-150@yandex.ru> wrote:
>
>> you have to play around and see whether you need "enable-active-high;"
>> or not (depends on the original .dts / Amlogic kernel implementation)
>>
>> Additionally you need to add "phy-supply = <&usb_vbus>;" to the
>> usb0_phy node - otherwise the PHY does not know that it has to enable
>> the usb_vbus regulator.
>>
>>
>> Regards,
>> Martin
>
> Hi.
> I tried several different options.
> The result is about the same.
>
>
> [    5.714478] tegra-ehci: Tegra EHCI driver
> [    5.718461] ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver
> [    5.724554] ohci-pci: OHCI PCI platform driver
> [    5.728986] ohci-platform: OHCI generic platform driver
> [    5.734256] ohci-omap3: OHCI OMAP3 driver
> [    5.738217] SPEAr-ohci: OHCI SPEAr driver
> [    5.742204] ohci-st: OHCI STMicroelectronics driver
> [    5.747077] ohci-atmel: OHCI Atmel driver
> [    5.751365] usbcore: registered new interface driver usb-storage
> [    5.759402] i2c /dev entries driver
> [    5.765797] meson_wdt c1109900.watchdog: Watchdog enabled (timeout=8 sec,
> nowayout=0)
> [    5.770002] sdhci: Secure Digital Host Controller Interface driver
> [    5.774398] sdhci: Copyright(c) Pierre Ossman
> [    5.779584] Synopsys Designware Multimedia Card Interface Driver
> [    5.786645] sdhci-pltfm: SDHCI platform and OF driver helper
> [    5.791544] of_pwm_get(): PWM chip not found
> [    5.794619] leds_pwm pwmleds: unable to request PWM for m8s:blue:power:
> -517
> [    5.801821] ledtrig-cpu: registered to indicate activity on CPUs
> [    5.807847] usbcore: registered new interface driver usbhid
> [    5.813133] usbhid: USB HID core driver
> [    5.874177] NET: Registered protocol family 10
> [    5.874892] Segment Routing with IPv6
> [    5.876682] sit: IPv6, IPv4 and MPLS over IPv4 tunneling driver
> [    5.883038] NET: Registered protocol family 17
> [    5.886958] can: controller area network core (rev 20120528 abi 9)
> [    5.893167] NET: Registered protocol family 29
> [    5.897529] can: raw protocol (rev 20120528)
> [    5.901775] can: broadcast manager protocol (rev 20161123 t)
> [    5.907424] can: netlink gateway (rev 20130117) max_hops=1
> [    5.913163] Key type dns_resolver registered
> [    5.917401] ThumbEE CPU extension supported.
> [    5.921369] Registering SWP/SWPB emulation handler
> [    5.927806] Btrfs loaded, crc32c=crc32c-generic
> [    5.935944] dwc2 c9040000.usb: c9040000.usb supply vusb_d not found,
> using dummy regulator
> [    5.938931] dwc2 c9040000.usb: c9040000.usb supply vusb_a not found,
> using dummy regulator
> [    5.948296] phy phy-c1108800.phy.0: USB ID detect failed!
> [    5.952506] phy phy-c1108800.phy.0: phy poweron failed --> -22
> [    5.958424] dwc2: probe of c9040000.usb failed with error -22
> [    5.965482] dwc2 c90c0000.usb: c90c0000.usb supply vusb_d not found,
> using dummy regulator
> [    5.972288] dwc2 c90c0000.usb: c90c0000.usb supply vusb_a not found,
> using dummy regulator
> [    6.112633] dwc2 c90c0000.usb: DWC OTG Controller
> [    6.112703] dwc2 c90c0000.usb: new USB bus registered, assigned bus
> number 1
> [    6.118756] dwc2 c90c0000.usb: irq 23, io mem 0xc90c0000
> [    6.124706] hub 1-0:1.0: USB hub found
> [    6.127775] hub 1-0:1.0: 1 port detected
> [    6.136994] of_pwm_get(): PWM chip not found
> [    6.137048] leds_pwm pwmleds: unable to request PWM for m8s:blue:power:
> -517
> [    6.143362] hctosys: unable to open rtc device (rtc0)
> [    6.148218] VCC3V3: disabling
> [    6.150894] USB0_VBUS: disabling
> [    6.156075] Freeing unused kernel memory: 2048K
> [    6.552575] usb 1-1: new high-speed USB device number 2 using dwc2
> [    6.762552] usb 1-1: device descriptor read/64, error -32
> [    7.092974] usb 1-1: device descriptor read/64, error -32
> [    7.422513] usb 1-1: new high-speed USB device number 3 using dwc2
> [    7.632505] usb 1-1: device descriptor read/64, error -32
> [    7.962503] usb 1-1: device descriptor read/64, error -32
> [    8.292515] usb 1-1: new high-speed USB device number 4 using dwc2
> [    8.732504] usb 1-1: device not accepting address 4, error -32
> [    8.942503] usb 1-1: new high-speed USB device number 5 using dwc2
> [    9.382504] usb 1-1: device not accepting address 5, error -32
> [    9.382735] usb usb1-port1: unable to enumerate USB device
> [  133.322520] random: crng init done
can you please boot the Android system that came with the device and
get a copy of:
- dmesg output
- /sys/kernel/debug/gpio
- /proc/interrupts
- /proc/iomem

> I tried to run the image without a INITRD. Got a kernel panic.
>
>
>     5.587299] ThumbEE CPU extension supported.
> [    5.591346] Registering SWP/SWPB emulation handler
> [    5.597785] Btrfs loaded, crc32c=crc32c-generic
> [    5.605922] dwc2 c9040000.usb: c9040000.usb supply vusb_d not found,
> using dummy regulator
> [    5.608891] dwc2 c9040000.usb: c9040000.usb supply vusb_a not found,
> using dummy regulator
> [    5.618271] phy phy-c1108800.phy.0: USB ID detect failed!
> [    5.622465] phy phy-c1108800.phy.0: phy poweron failed --> -22
> [    5.628412] dwc2: probe of c9040000.usb failed with error -22
> [    5.635451] dwc2 c90c0000.usb: c90c0000.usb supply vusb_d not found,
> using dummy regulator
> [    5.642271] dwc2 c90c0000.usb: c90c0000.usb supply vusb_a not found,
> using dummy regulator
> [    5.782627] dwc2 c90c0000.usb: DWC OTG Controller
> [    5.782696] dwc2 c90c0000.usb: new USB bus registered, assigned bus
> number 1
> [    5.788748] dwc2 c90c0000.usb: irq 23, io mem 0xc90c0000
> [    5.794694] hub 1-0:1.0: USB hub found
> [    5.797768] hub 1-0:1.0: 1 port detected
> [    5.806231] hctosys: unable to open rtc device (rtc0)
> [    5.806943] VCC3V3: disabling
> [    5.809619] USB0_VBUS: disabling
> [    5.813139] VFS: Cannot open root device "LABEL=ROOTFS" or
> unknown-block(0,0): error -6
> [    5.820800] Please append a correct "root=" boot option; here are the
> available partitions:
> [    5.829154] 0100           65536 ram0
> [    5.829157]  (driver?)
> [    5.835194] 0101           65536 ram1
> [    5.835197]  (driver?)
> [    5.841260] 0102           65536 ram2
> [    5.841262]  (driver?)
> [    5.847328] 0103           65536 ram3
> [    5.847330]  (driver?)
> [    5.853404] 0104           65536 ram4
> [    5.853407]  (driver?)
> [    5.859463] 0105           65536 ram5
> [    5.859465]  (driver?)
> [    5.865530] 0106           65536 ram6
> [    5.865533]  (driver?)
> [    5.871598] 0107           65536 ram7
> [    5.871600]  (driver?)
> [    5.877666] 0108           65536 ram8
> [    5.877669]  (driver?)
> [    5.883742] 0109           65536 ram9
> [    5.883744]  (driver?)
> [    5.889802] 010a           65536 ram10
> [    5.889804]  (driver?)
> [    5.895956] 010b           65536 ram11
> [    5.895959]  (driver?)
> [    5.902110] 010c           65536 ram12
> [    5.902113]  (driver?)
> [    5.908265] 010d           65536 ram13
> [    5.908268]  (driver?)
> [    5.914427] 010e           65536 ram14
> [    5.914430]  (driver?)
> [    5.920574] 010f           65536 ram15
> [    5.920576]  (driver?)
> [    5.926735] Kernel panic - not syncing: VFS: Unable to mount root fs on
> unknown-block(0,0)
> [    5.934965] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.11.0+ #6
> [    5.940941] Hardware name: Amlogic Meson platform
> [    5.945644] [<c03105b8>] (unwind_backtrace) from [<c030bb10>]
> (show_stack+0x10/0x14)
> [    5.953346] [<c030bb10>] (show_stack) from [<c07f562c>]
> (dump_stack+0x88/0x9c)
> [    5.960542] [<c07f562c>] (dump_stack) from [<c03d3c28>]
> (panic+0xdc/0x268)
> [    5.967392] [<c03d3c28>] (panic) from [<c1401218>]
> (mount_block_root+0x198/0x270)
> [    5.974842] [<c1401218>] (mount_block_root) from [<c1401500>]
> (mount_root+0x124/0x12c)
> [    5.982729] [<c1401500>] (mount_root) from [<c1401684>]
> (prepare_namespace+0x17c/0x1c4)
> [    5.990704] [<c1401684>] (prepare_namespace) from [<c1400e50>]
> (kernel_init_freeable+0x1dc/0x1ec)
> [    5.999548] [<c1400e50>] (kernel_init_freeable) from [<c0e80254>]
> (kernel_init+0x8/0x110)
> [    6.007697] [<c0e80254>] (kernel_init) from [<c0307db8>]
> (ret_from_fork+0x14/0x3c)
> [    6.015242] ---[ end Kernel panic - not syncing: VFS: Unable to mount
> root fs on unknown-block(0,0)
>
>
>
> I have a suggestion that the reason for the stop run mine the not correct
> Assembly of the image.
> I can see the log of system start-up do You have ?
> What image are You using ?
I'm using an initrd from Arch Linux ARM (from their ARMv7 multi-platform image).
however, I found another missing kernel config option yesterday: you
also need to enable CONFIG_COMMON_CLK_PWM
once you enable that you should be able to see /dev/mmcblk1p[0-9] with
all the partitions from your SD card

I was also able to reproduce the kernel crash you reported last week.
it seems to go away when enabling CONFIG_DEBUG_SPINLOCK in the kernel
config - but I will create a separate thread about this (as I suspect
that there may be some memory corruption along with a race condition
somewhere)


Regards,
Martin

S805/S812 Device Tree upstreaming (was: Testing.)

[Oleg]

> get a copy of:
> - dmesg output
> - /sys/kernel/debug/gpio
> - /proc/interrupts
> - /proc/iomem

I received the necessary data from LebreELEC (this system now I have eMMC).
Let me know if you need to receive this data always from Android.

https://yadi.sk/d/deis2Z023JAiyE

>
> I'm using an initrd from Arch Linux ARM (from their ARMv7 multi-platform image).
> however, I found another missing kernel config option yesterday: you
> also need to enable CONFIG_COMMON_CLK_PWM
> once you enable that you should be able to see /dev/mmcblk1p[0-9] with
> all the partitions from your SD card
>
> I was also able to reproduce the kernel crash you reported last week.
> it seems to go away when enabling CONFIG_DEBUG_SPINLOCK in the kernel
> config - but I will create a separate thread about this (as I suspect
> that there may be some memory corruption along with a race condition
> somewhere)
>
>
> Regards,
> Martin
Hi Martin

Good news. Using the sources Your latest kernel (20170513) and amendments
in the config of the kernel I was able to run on Tronsmart MXIII Plus 
(s812).
By the way, in the source for Your kernel I had to make one small change 
to comment
out the in /arch/arm/mach-meson/Makefile a build of platsmp.
Thanks for the support. Now I have the ability
to run basic system on the device and test will be easier.

S805/S812 Device Tree upstreaming (was: Testing.)

[Oleg]

Log system start-up.

https://yadi.sk/i/0kvkEnct3JB7QZ

S805/S812 Device Tree upstreaming (was: Testing.)

[Martin Blumenstingl]

Hi Oleg,

On Mon, May 15, 2017 at 4:45 PM, Oleg <balbes-150@yandex.ru> wrote:
> Log system start-up.
>
> https://yadi.sk/i/0kvkEnct3JB7QZ
this looks pretty good - the SD card is found!

have you tried enabling the brcmfmac driver (CONFIG_BRCMFMAC and
CONFIG_BRCMFMAC_SDIO) to see whether your wireless chip is detected as
well?
testing if wifi works would help me a lot, since my board comes with a
RTL8723BS wifi chip and I couldn't get it working on first try (but I
don't trust the driver there, while I know that the brcmfmac driver
works fine based on my experience with other boards)

S805/S812 Device Tree upstreaming (was: Testing.)

[Oleg]

> this looks pretty good - the SD card is found!
>
> have you tried enabling the brcmfmac driver (CONFIG_BRCMFMAC and
> CONFIG_BRCMFMAC_SDIO) to see whether your wireless chip is detected as
> well?
> testing if wifi works would help me a lot, since my board comes with a
> RTL8723BS wifi chip and I couldn't get it working on first try (but I
> don't trust the driver there, while I know that the brcmfmac driver
> works fine based on my experience with other boards)
Hi Martin

I added modules to the kernel. Put together a new way s812_wifi.txt.
The result of loading the modules in the application.
I guess I need to add settings in the DTB file ?

https://mega.nz/#F!yw5kVCrR!zNvlE8f6bcpOR2XZbJdCog

S805/S812 Device Tree upstreaming (was: Testing.)

[Martin Blumenstingl]

Hi Oleg,

On Tue, May 16, 2017 at 11:05 AM, Oleg <balbes-150@yandex.ru> wrote:
>
>> this looks pretty good - the SD card is found!
>>
>> have you tried enabling the brcmfmac driver (CONFIG_BRCMFMAC and
>> CONFIG_BRCMFMAC_SDIO) to see whether your wireless chip is detected as
>> well?
>> testing if wifi works would help me a lot, since my board comes with a
>> RTL8723BS wifi chip and I couldn't get it working on first try (but I
>> don't trust the driver there, while I know that the brcmfmac driver
>> works fine based on my experience with other boards)
>
> Hi Martin
>
> I added modules to the kernel. Put together a new way s812_wifi.txt.
> The result of loading the modules in the application.
> I guess I need to add settings in the DTB file ?
>
> https://mega.nz/#F!yw5kVCrR!zNvlE8f6bcpOR2XZbJdCog
thanks for that - along with your dmesg output from the vendor kernel
this gives us some more details

your dmesg indicates that SDIO port/slot 0 is used for the wifi chip,
see all these aml_sdio_print_err messages with "pdata->port=0"
it also confirms that GPIOAO_6 and GPIOX_11 are used to power on the
wifi chip and that GPIOX_10 is used as pin for the LPO (32.768 kHz)
clock

unfortunately the kernel log when booting my kernel does not show
anything about the wifi chip.
I would have either expected error messages from mmc0 or a short
message that the device was detected.
can you please show me the output of /sys/kernel/debug/mmc[0-2]/* with
my kernel and the original Android/LibreELEC kernel?

regarding your USB problem: unfortunately there doesn't seem to be
anything relevant in the Android/LibreELEC kernel logs.
can you please provide a patch with the .dts changes you made so I can
have a look at these (maybe there's a typo somewhere)?


Regards,
Martin

S805/S812 Device Tree upstreaming

[Oleg]

Hi Martin
In Lebreelec I did not find the specified directories.
in the file ls_debug_lebreelec.txt the contents of the
directory from Lebreelec.
Output content Linux in appropriate files.
https://yadi.sk/d/X-26JMEn3JNFbD

I use this dtb file.
https://github.com/150balbes/Amlogic_s8xx-kernel/blob/test/arch/arm/boot/dts/meson8m2-mxiii-plus.dts 


Question, wath that me need to add or change data in the dtb, what would 
I have to run a wired network ?

Thank You

S805/S812 Device Tree upstreaming

[Martin Blumenstingl]

Hi Oleg,

On Sun, May 21, 2017 at 10:57 AM, Oleg <balbes-150@yandex.ru> wrote:
> Hi Martin
> In Lebreelec I did not find the specified directories.
> in the file ls_debug_lebreelec.txt the contents of the
> directory from Lebreelec.
> Output content Linux in appropriate files.
> https://yadi.sk/d/X-26JMEn3JNFbD
ah, in Amlogic's kernel the directories are probably
/sys/kernel/debug/{emmc,sd,sdio}

> I use this dtb file.
> https://github.com/150balbes/Amlogic_s8xx-kernel/blob/test/arch/arm/boot/dts/meson8m2-mxiii-plus.dts
this looks good to me. so you have tried:
gpio = <&gpio GPIOZ_1 GPIO_ACTIVE_HIGH>;
enable-active-high;

as well as:
gpio = <&gpio GPIOZ_1 GPIO_ACTIVE_LOW>;
// enable-active-high; not set

> Question, wath that me need to add or change data in the dtb, what would I
> have to run a wired network ?
you already have phy-mode = "rgmii"; in your .dts so this looks fine.
maybe the pinctrl driver is missing some bits for rgmii mode, can you
please run "md 0xc11080b0 0x28" in u-boot and paste the output (this
dumps the pinctrl mux mode registers from u-boot, where I assume that
ethernet is working)?
do you see any dwmac/stmmac related errors in your kernel log?


Regards,
Martin

S805/S812 Device Tree upstreaming

[Oleg]

Hi Martin
1. Updated the source code to the latest version of meson-mx 
integration-4.12-20170520
2. Collected test image with the new kernel. 2. Checked both settings in 
dtb (HIGH and LOW).
While the result is the same.
3. The output from the console log and with the command "md".
4 Contents of directories emmc sd sdio (from the working image Armbian 
with kernel 3.14 ,
which features Wi-Fi and wired network with USB and HDMI)

https://yadi.sk/d/KvSjVEK-3JQF7c .


Is it possible to transfer the wired network settings of dtb data from 
s905 (p200)
  in the dtb settings for MXIII-Plus or are they not compatible ?

S805/S812 Device Tree upstreaming

[Martin Blumenstingl]

Hi Oleg,

On Mon, May 22, 2017 at 2:12 PM, Oleg <balbes-150@yandex.ru> wrote:
> Hi Martin
> 1. Updated the source code to the latest version of meson-mx
> integration-4.12-20170520
> 2. Collected test image with the new kernel. 2. Checked both settings in dtb
> (HIGH and LOW).
> While the result is the same.
> 3. The output from the console log and with the command "md".
> 4 Contents of directories emmc sd sdio (from the working image Armbian with
> kernel 3.14 ,
> which features Wi-Fi and wired network with USB and HDMI)
>
> https://yadi.sk/d/KvSjVEK-3JQF7c .
regarding USB: I don't see a reason why it's reporting this error :(
Kevin's Odroid-C2 board (used by kernelci) reports the same error [0]
- I cannot reproduce this with any of my devices unfortunately (so I'm
a bit lost here)

regarding wifi:
I would like to see this fixed because it would be great to have
someone with Broadcom wifi (to confirm that the MMC driver works
fine).
I'm a bit puzzled though why it doesn't seem to find your SDIO wifi,
even though it's at slot 0 (confirmed by looking at your sdio.tar.gz).
maybe you can enable the debug logs in drivers/mmc/core.c (just add
#define DEBUG at the beginning of the file and prepare for a HUGE
log-spam!)

I also compared your pinmux register dump to see if there's anything
interesting regarding ethernet.
all eth_* related bits are set just like the kernel would do (REG6
bits 5-8 and bits 10-14)
a similar error was reported for another platform [1]
however, I must also say that I am sure that I need to switch from the
"meson6-dwmac" driver to "meson8b-dwmac" - but that will require
additional work in the "meson8b-dwmac" driver.

> Is it possible to transfer the wired network settings of dtb data from s905
> (p200)
>  in the dtb settings for MXIII-Plus or are they not compatible ?
with my current .dts they are not compatible - the goal is to make
them compatible (more work is required for that though)


Regards,
Martin


[0] https://storage.kernelci.org/mainline/master/v4.12-rc1-268-g6fe1de43c5f6/arm64/defconfig+CONFIG_LKDTM=y/lab-baylibre-seattle/boot-meson-gxbb-odroidc2.html
[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2017-March/492508.html

S805/S812 Device Tree upstreaming

[Oleg]

Hi Martin
> regarding USB: I don't see a reason why it's reporting this error:(
> Kevin's Odroid-C2 board (used by kernelci) reports the same error [0]
> - I cannot reproduce this with any of my devices unfortunately (so I'm
> a bit lost here)
I will try to perform a number of tests and changes in the dtb.
When I have results, I'll let You know. Thanks for the support.

> regarding wifi:
> I would like to see this fixed because it would be great to have
> someone with Broadcom wifi (to confirm that the MMC driver works
> fine).
> I'm a bit puzzled though why it doesn't seem to find your SDIO wifi,
> even though it's at slot 0 (confirmed by looking at your sdio.tar.gz).
> maybe you can enable the debug logs in drivers/mmc/core.c (just add
> #define DEBUG at the beginning of the file and prepare for a HUGE
> log-spam!)
I will try to obtain this data in the coming days.

> a similar error was reported for another platform [1]
> however, I must also say that I am sure that I need to switch from the
> "meson6-dwmac" driver to "meson8b-dwmac" - but that will require
> additional work in the "meson8b-dwmac" driver.
I'll wait for Your results. I'm not in a hurry  :) .

S805/S812 Device Tree upstreaming

[Oleg]

Hi Martin

I regularly collect the test images using your GIT and check them on my 
MXIII Plus.
Since the upgrade to version 4.15 I have stopped, not start kernel.
It does not even starts.
The same behavior I see on other kernels if i'm use not correct dtb or 
if the dtb is missing.
The latest version that I have started kernel 4.13_20170814.
I compared the dts files, they match the version 4.15.
You've changed something in the source code 4.15 ?
You kernel 4.15 normal start on your model TV box ?

S805/S812 Device Tree upstreaming

[Martin Blumenstingl]

Hi Oleg,

On Sun, Oct 1, 2017 at 9:22 AM, Oleg <balbes-150@yandex.ru> wrote:
> Hi Martin
>
> I regularly collect the test images using your GIT and check them on my
> MXIII Plus.
> Since the upgrade to version 4.15 I have stopped, not start kernel.
> It does not even starts.
> The same behavior I see on other kernels if i'm use not correct dtb or if
> the dtb is missing.
> The latest version that I have started kernel 4.13_20170814.
> I compared the dts files, they match the version 4.15.
> You've changed something in the source code 4.15 ?
> You kernel 4.15 normal start on your model TV box ?
I have started experimenting with CPU clock changes some time ago -
unfortunately I didn't get it working properly so far.
could you please re-try with branch meson-mx-integration-4.15-20171001
where I made all the CPU-clocks read-only again?

Thanks for testing by the way!


Regards,
Martin

S805/S812 Device Tree upstreaming

[Oleg]

Tested, this works.