Nick

Hi Zador, I've just tried your dtb file, sadly it doesn't seem to have worked. It was a clean kernel source with none of my patches installed. If was however the dev kernel branch would that be a problem? Thanks for the patch notes by the way, I am at least learning alot working through this.

Sorry for the basic questions, after a close inspection of the build process I've realised that my patches aren't being applied anyway. I'm guessing it's a problem with the file paths at the top of the patch file, but despite testing a few different options I can't figure out what they should be? --- a/arch/arm/boot//sun7i-a20-bananapro.dts 2016-01-30 14:22:26.000000000 +0000 +++ sun7i-a20-bananapro-nick.dts 2016-01-30 17:47:29.651279300 +0000 @@ -242,6 +242,15 @@ <&spi0_cs0_pins_a>, <&spi0_cs1_pins_a>; status = "okay"; + spi0_0 { + #address-cells = <1>; + #size-cells = <0>; + + compatible = "spidev"; + + reg = <0>; + spi-max-frequency = <50000000>; + }; }; &uart0 {

Oh wow, I missed the nesting! Well spotted. I've created a new patch, so we'll see how that goes. Edit. As for the dtsi file, I agree that it doesn't appear to need patching as even without the patch the spi block is marked as enabled, I'm just not sure how / where it's being changed from disabled to okay?

Great to hear that you managed to get it to run. Which file did you patch with those lines? I've created two patches but they don't seem to help. --- ./sun7i-a20-bananapro.dts 2016-01-30 14:22:26.000000000 +0000 +++ sun7i-a20-bananapro-nick.dts 2016-01-30 15:22:50.134387022 +0000 @@ -243,6 +243,15 @@ <&spi0_cs1_pins_a>; status = "okay"; }; +spi0_0 { + #address-cells = <1>; + #size-cells = <0>; + + compatible = "spidev"; + + reg = <0>; + spi-max-frequency = <50000000>; + }; &uart0 { pinctrl-names = "default"; and... --- ./sun7i-a20.dtsi 2016-01-30 14:28:09.000000000 +0000 +++ sun7i-a20-nick.dtsi 2016-01-30 15:17:46.082850154 +0000 @@ -679,7 +679,7 @@ dmas = <&dma SUN4I_DMA_DEDICATED 27>, <&dma SUN4I_DMA_DEDICATED 26>; dma-names = "rx", "tx"; - status = "disabled"; + status = "okay"; #address-cells = <1>; #size-cells = <0>; };

Ok, i have tried creating my own patch for the kernel to add the extra info, the patch was placed in the userpatchs area. It didn't solve the problem, but then I'm not overly confident regarding patching or exactly how Igor's scripts compile the kernel etc. It's possible that my changes were either not being applied or just being overwritten. Attempt number two was to modify the dtb file directly on the device. This approach I'm more confident with, however sadly still no spidev.

Ah does that explain the boot.cmd hack that is required to get the wifi working? I hadn't made the connection between there not being a dts/dtb file for the m1+ at all, but that would explain why the variable expansion in boot.cmd fails.

Thanks for the explanation, I'll look at that now I'm guessing that I'll have to decompile the running dtb file and edit that as I still can't build the dev kernel from source at the moment. As an aside, I've just quickly tested the wifi with the deb packages from your dropbox account. Your kernel image with your dtb package appears to break the wifi, however if you use your kernel image with the armbian dtb-next package the wifi works just fine. Sorry for the slight hijack, I just didn't want you to waste your time looking at the kernel image when it appears to be a dtb issue.

Zador, The kernel in your dropbox link appears to break wifi as well, I'm not sure if that helps you at all? According to /var/log/syslog wpa_supplicant is trying to play with nl80211 I'm not sure if that makes any sense? Using the same SD card image if I remove your kernel and instead install linux-image-next-sunxi then the WiFi works again. As for the dtb changes, I'm looking at that now. If I'm honest I'm struggling to get my head around the link that you have posted but I'm sure it will all make sense when I investigate it. I'll keep you posted.

Hazard, if you haven't already try this: http://forum.armbian.com/index.php/topic/340-banana-pi-pro-m1plus-no-wifi-armbian-45/?p=2531

No problem, if you have anything else that you would like me to check just shout. I have reserved a spare SD card for testing this.

Ok, so I built an image of Jessie / next using Igor's compile script, booted that on the banana pi and using apt I removed linux image and linux dtb, then using dpkg I installed the new linux image and dtb from your dropbox link. The pi booted fine, no issues at all however still no spidev in /dev and no mention of spi in dmesg sadly. Edit. On the plus side, the new image booted incredibly quickly even the first time when it was resizing the SD card etc. Logins seem to take a little longer though, with a gap of around a few seconds between entering the username and being asked for a password.

Brilliant, thanks. I'm just waiting for a few things to finish and then I'll try it and report back...

Sorry for the silly question but how would I install that? Is it just a case of installing linux-image-dev-sunxi_4.83_armhf.deb into a running system via dpkg? If so, that I can do :-)

Sorry for the late reply, I was working on site yesterday. Sadly there was nothing at all mentioned in dmesg regarding SPI. I tried compiling the dev branch but the kernel didn't compile for some reason. I didn't have time to look into why but I can do if it helps? For the moment I think I'll just stick to the legacy kernel as I don't need the latest version for the moment but I do need the SPI driver. If any testing etc. is needed however then just shout and I will be happy to try out new bug fixes, suggestions etc. Thanks, Nick

Hi zador.blood.stained, Thanks for the info. I have started looking through the dts file, decompiling it as mentioned in the thread that you provided. I have various sections with the pins defined, I also have this: spi@01c05000 { compatible = "allwinner,sun4i-a10-spi"; reg = <0x1c05000 0x1000>; interrupts = <0x0 0xa 0x4>; clocks = <0x3 0x14 0x13>; clock-names = "ahb", "mod"; dmas = <0x14 0x1 0x1b 0x14 0x1 0x1a>; dma-names = "rx", "tx"; status = "okay"; #address-cells = <0x1>; #size-cells = <0x0>; pinctrl-names = "default"; pinctrl-0 = <0x15 0x16 0x17>; }; spi@01c06000 { compatible = "allwinner,sun4i-a10-spi"; reg = <0x1c06000 0x1000>; interrupts = <0x0 0xb 0x4>; clocks = <0x3 0x15 0x18>; clock-names = "ahb", "mod"; dmas = <0x14 0x1 0x9 0x14 0x1 0x8>; dma-names = "rx", "tx"; status = "disabled"; #address-cells = <0x1>; #size-cells = <0x0>; }; I'm assuming that the device at 01c05000 is SPI0 and the device at ..6000 is SPI1, if that's the case then I'm happy with SPI0 enabled and SPI1 disabled for the moment. However looking at the compatible line, it would appear that it is trying to load the sun4i-a10 driver, I'm assuming this should be sun7i-a20? Nick Edit... I have just tried modifying the compatible line to sun7i-a20, recompiled and rebooted but still no spidev :-(

Hi all, traditionally I have used Bananian on the Bannan Pi M1+, but I thought that I would give Armbian a go as there seems to be more activity here and there are tools for compiling an entire image from scratch :-) So far I love Ambian, thank you for the work. I have tried the vanilla Jessie image (both downloaded from here and compiled myself) on the Banana Pi M1+ and with the exception of the WiFi issuet which I solved from this forum, all ran well. However I can't seem to find an SPI driver. My apologies if there is another forum post about this, I did try searching but the search function removed SPI from the query due to it being less than 4 chars :-( When I built my own image using the tools from this site, I did check the kernel configuration which was set to build in SPI support, but still no entry in /dev or a module to load. Any help would be appreciated. Thanks, Nick