From mboxrd@z Thu Jan 1 00:00:00 1970 From: martin.blumenstingl@googlemail.com (Martin Blumenstingl) Date: Sat, 27 May 2017 20:18:02 +0200 Subject: Testing. In-Reply-To: References: <743c82c3-52af-bcdb-b45a-d392cda08fb6@o2online.de> <4ce994ac-9272-ccec-8dd0-6698ce35f289@googlemail.com> Message-ID: To: linus-amlogic@lists.infradead.org List-Id: linus-amlogic.lists.infradead.org On Sat, May 27, 2017 at 8:06 PM, raptorsds wrote: > > > Am 27.05.2017 um 18:08 schrieb Martin Blumenstingl: >> >> On Sat, May 27, 2017 at 4:57 PM, raptorsds >> wrote: >>> >>> Hi Martin >>> >>>> Hi Tobias, >>>> >>>> On Thu, May 25, 2017 at 8:35 PM, Tobias Baumann >>>> <017623705678@o2online.de> wrote: >>>>> >>>>> hi martin >>>>> >>>>> oleg (balbes150) from armbian forum told me you could need some help >>>>> >>>>> iam new to linux but i can test images and files for you >>>>> >>>>> my system OTT M8S Clone TVBox with S812 8GBRom 2GB RAM RTL8723BS WIFI >>>>> BT >>>>> Combo HDMI 2xUSB >>>>> >>>>> ( run as server for USB Printer and Sane server for canon scanner) >>>>> >>>>> at the moment i use the armbian unbuntu rom from oleg with 3.10.99 >>>>> kernel >>>>> (14 04 2017 ) with working wifi usb and internet also i could run a >>>>> successful test the 4.11 kernel >>>> >>>> disclaimer: I cannot really help with Amlogic's 3.10/3.14 kernel >>> >>> >>> But maybe than i can help by 4.11 kernel test for older devices >> >> you need to be aware of the fact that my kernel patches are experimental. >> what you get is: >> - one CPU core at the frequency set by the bootloader >> - maybe USB (Oleg has problems with it while it works for me) >> - maybe ethernet (again, Oleg has problems while it works for me) >> - SD card >> - front LED >> >> basically everything else is broken (wifi, CPU cores 2-4, cpu frequency >> scaling, etc.). I even managed to corrupt u-boot when I was testing my >> NAND driver. >> >> what "older" devices do you have? and do you have access to the serial >> console of these devices? > > > Yes Serial console only my s812 box , >>>>> >>>>> my project now is to control the front LED >>>> >>>> what are you trying to control exactly? >>> >>> >>> red = stop >>> blue = boot/start >>> red/blue (maybe dimming) = run >>> blinking = update (maybe) >> >> let me explain how where the LED is controlled currently: >> - red LED turns on automatically (probably because it's wired to VCC) >> when the device is powered on >> - u-boot turns on the blue LED >> - (kernel boots) >> - the kernel can now control the LED partially (as explained in my >> previous mail, you can only set the "mix" between red and blue - you >> can't turn it off) >> >> so with another kernel you can only change the LED color from the >> point on where the kernel is started. you'll have no control over it >> (without additional work on the u-boot side) for the first seconds >> after powering on the device. > > ok thank you for your explanation , it would be great when after kernelk > boot (maybe inside rc.local) the LED will have a color change > > >>>>> but there is a problem oleg send my the dts for led but without gpio . >>>> >>>> my m8s.dts doesn't use a GPIO because the LED is connected to >>>> GPIO_TEST_N which can also be controlled by one of the PWM controllers >>>> (this allows dimming the LED to a certain level, but in my case the >>>> LED is a dual-color LED with blue and red: the further you reduce the >>>> brightness of the (blue) LED the more you see of the red one -> if you >>>> turn off the blue LED it turns red) >>> >>> >>> I though " GPIO_TEST_N " was a placeholder . I havent any trigger at >>> /sys/class/leds , how can i controll PWM for LED >> >> are you testing this on a 3.10/3.14 kernel or on a kernel with my >> patches? I think the Amlogic kernel does not use the kernel's LED >> framework, so the LEDs are not exposed in /sys/class/leds >> if it's one of my kernels (I think Oleg has compiled some images with >> them) then please post the kernel log (dmesg) > > > yes i use 3.10.99 kernel from the image from oleg with a dtb files with your > Led infos about "gpio test n" like I said in a previous mail, I can't say much about the 3.10/3.14 kernel - it contains a lot of "custom" code which duplicates functionality that it already provided by various frameworks inside the Linux kernel. this is also the reason why your LED doesn't show up in /sys/class/leds > i also found a older info about system LED > > https://lists.freebsd.org/pipermail/freebsd-arm/2015-February/010126.html > > there is a info about gpio ao 13 the LED pin is board specific - GPIOAO_13 is correct for the Odroid-C1 > i send you a private mail with my dmesg with oleg dts file with led line that is a kernel log and dts files from a 3.10 kernel. you can have a look at the "aml_sysled" code from their 3.10 kernel, see [0] but I'm a afraid this is all I can do for you here if you switch to one of Oleg's images with my patches then the LED will show up in /sys/class/leds, but like I mentioned previously all this is pretty experimental Regards, Martin [0] https://github.com/endlessm/linux-meson/blob/master/drivers/amlogic/led/aml_sysled.c