mboehmer Posted April 10, 2019 Posted April 10, 2019 Armbianmonitor: http://ix.io/1FOc Hi, I compiled some images for C2 yesterday (mainline 5.77, jessie), and had a nice console as usual. Today, the new images still have u-boot console, and give some information about kernel starting, but no login by console (USB serial) is possible. I tried to set "console=serial" in /boot/armbianEnv.txt, but that gives me an endless reboot loop. Were there some changes? Maybe in baudrate? I get u-boot messages at 115200 8N1 as expected.. Michael
martinayotte Posted April 10, 2019 Posted April 10, 2019 1 hour ago, mboehmer said: I tried to set "console=serial" in /boot/armbianEnv.txt, but that gives me an endless reboot loop. Did you also change "verbosity=7" ? What else are you seeing after "starting kernel" ?
mboehmer Posted April 10, 2019 Author Posted April 10, 2019 20 minutes ago, martinayotte said: Did you also change "verbosity=7" ? What else are you seeing after "starting kernel" ? I tried "verbosity=2". With that I don't see any further messagew after "starting kernel...".
Igor Posted April 10, 2019 Posted April 10, 2019 8 minutes ago, mboehmer said: I tried "verbosity=2". With that I don't see any further messagew after "starting kernel...". I just build and boot http://ix.io/1FOQ and everything looks normal. Login from console works, when adding console=serial to /boot/armbianEnv.txt boot messages go to serial console.
martinayotte Posted April 10, 2019 Posted April 10, 2019 31 minutes ago, mboehmer said: I don't see any further messagew after "starting kernel...". Maybe corruption while writting the image to SDCard ... Did you used "etcher" ?
mboehmer Posted April 10, 2019 Author Posted April 10, 2019 5 hours ago, martinayotte said: Maybe corruption while writting the image to SDCard ... Did you used "etcher" ? I use only eMMC card, no SD cards. I tried with two images, and yes, I use etcher (on Windows, sorry, it's my main machine here). Even tried a different USB serial converter, and checked output of console header with scope. No edges after "starting kernel..." Just for understanding: /dev/tty1 should be routed to the J2 Rxd/TxD pins, so "echo BLA > /dev/tty1" should give some edges on J2 UART?
martinayotte Posted April 10, 2019 Posted April 10, 2019 12 minutes ago, mboehmer said: Just for understanding: /dev/tty1 should be routed to the J2 Rxd/TxD pins, so "echo BLA > /dev/tty1" should give some edges on J2 UART? Not on AmLogic ... AmLogic is using /dev/ttyAML0 for debug port , at least on my NanoPiK2 ... For pin 8/10 on J2, that would be /dev/ttyAML1 (I've done an UartA overlay recently) ...
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 Sorry, double post. should pick coffee first
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 1 hour ago, mboehmer said: I just changed to "verbosity=7" and don't get any messages after GXBB:BL1:08dafd:0a8993;FEAT:EDFC318C;POC:3;RCY:0;EMMC:0;READ:0;CHK:0; TE: 376888 no sdio debug board detected BL2 Built : 11:44:26, Nov 25 2015. gxb gfb13a3b-c2 - jcao@wonton Board ID = 8 set vcck to 1100 mv set vddee to 1050 mv CPU clk: 1536MHz DDR channel setting: DDR0 Rank0+1 same DDR0: 2048MB(auto) @ 912MHz(2T)-13 DataBus test pass! AddrBus test pass! Load fip header from eMMC, src: 0x0000c200, des: 0x01400000, size: 0x000000b0 Load bl30 from eMMC, src: 0x00010200, des: 0x01000000, size: 0x00009ef0 Sending bl30........................................OK. Run bl30... Load bl301 from eMMC, src: 0x0001c200, des: 0x01000000, size: 0x000018c0 Wait bl30...Done Sending bl301.......OK. Run bl301... 00, des: 0x10100000, size: 0x00011130 --- UART initialized after reboot --- [Reset cause: unknown] [Image: unknown, amlogic_v1.1.3046-00db630-dirty 2016-08-31 09:24:14 tao.zeng@droid04] bl30: check_permit, count is 1 bl30: check_permit: ok! chipid: ef be ad de d f0Load bl33 from eMMC, src: 0x00034200, des: 0x01000000, size: 0x00078fa0 ad ba ef be ad de not ES chip [0.492576 Inits done] secure task start! high task start! low task start! NOTICE: BL3-1: v1.0(debug):4d2e34d NOTICE: BL3-1: Built : 17:08:35, Oct 29 2015 INFO: BL3-1: Initializing runtime services INFO: BL3-1: Preparing for EL3 exit to normal world INFO: BL3-1: Next image address = 0x1000000 INFO: BL3-1: Next image spsr = 0x3c9 U-Boot 2019.01-armbian (Apr 10 2019 - 13:22:25 +0200) odroid-c2 DRAM: 2 GiB MMC: mmc@72000: 0, mmc@74000: 1 Loading Environment from EXT4... Card did not respond to voltage select! In: serial@4c0 Out: serial@4c0 Err: serial@4c0 Net: eth0: ethernet@c9410000 Hit any key to stop autoboot: 0 Card did not respond to voltage select! switch to partitions #0, OK mmc1(part 0) is current device Scanning mmc 1:1... Found U-Boot script /boot/boot.scr 3048 bytes read in 2 ms (1.5 MiB/s) ## Executing script at 08000000 172 bytes read in 1 ms (168 KiB/s) 4705956 bytes read in 116 ms (38.7 MiB/s) 15253512 bytes read in 370 ms (39.3 MiB/s) 35547 bytes read in 9 ms (3.8 MiB/s) 687 bytes read in 3 ms (223.6 KiB/s) Applying kernel provided DT overlay meson-w1-gpio.dtbo 901 bytes read in 3 ms (293 KiB/s) Applying kernel provided DT overlay meson-ina219.dtbo 232 bytes read in 3 ms (75.2 KiB/s) Applying kernel provided DT fixup script (meson-fixup.scr) ## Executing script at 34000000 ## Loading init Ramdisk from Legacy Image at 13000000 ... Image Name: uInitrd Image Type: AArch64 Linux RAMDisk Image (gzip compressed) Data Size: 4705892 Bytes = 4.5 MiB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK ## Flattened Device Tree blob at 08008000 Booting using the fdt blob at 0x8008000 Loading Ramdisk to 7dad8000, end 7df54e64 ... OK reserving fdt memory region: addr=8008000 size=6f000 Loading Device Tree to 000000007da66000, end 000000007dad7fff ... OK Starting kernel ... Will try with a different Odroid C2, and also check with a freshly compiled image. As I can enter the u-boot menu and see that messages, at least I can exclude wiring problems with RxD and TxD Edit: removing the overlays doesn't change anything on console. Edit: second C2 shows same behaviour with compiled image from yesterday. Edit: freshly compiled kernel shows same behaviour. "verbosity=7" doesn't give any information after "kernel starting..."
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 I have no /dev/ttyAML0 and /dev/ttyAML1 on my Odroid C2.
Igor Posted April 11, 2019 Posted April 11, 2019 This is mine. root@odroidc2:~# w 06:11:08 up 16:13, 1 user, load average: 0.00, 0.00, 0.00 USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT root ttyAML0 - Wed14 2.00s 0.12s 0.00s w Try to build the exact image: bionic - default. I didn't try Jessie since its EOL.
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 1 hour ago, Igor said: This is mine. root@odroidc2:~# w 06:11:08 up 16:13, 1 user, load average: 0.00, 0.00, 0.00 USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT root ttyAML0 - Wed14 2.00s 0.12s 0.00s w Try to build the exact image: bionic - default. I didn't try Jessie since its EOL. Sir, yes, sir! Edit: I use "stretch", not "jessie". Welcome to ARMBIAN 5.78 user-built Debian GNU/Linux 9 (stretch) 4.19.34-meson64 System load: 0.67 0.28 0.10 Up time: 1 min Memory usage: 4 % of 1976MB IP: 10.162.208.175 CPU temp: 33°C Usage of /: 8% of 15G Edit: bionic compiled. Same result, no console. Welcome to ARMBIAN 5.78 user-built Ubuntu 18.04.2 LTS 4.19.34-meson64 System load: 0.46 0.11 0.04 Up time: 0 min Memory usage: 4 % of 1976MB IP: 10.162.208.175 CPU temp: 27°C Usage of /: 6% of 15G
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 Found something strange in both bionic and jessie dmesg output: [ 1.677916] GPIO line 501 (usb-hub-reset) hogged as output/high [ 1.684494] gxbb-aoclkc c8100000.sys-ctrl:clock-controller: Clock registration failed [ 1.684533] gxbb-aoclkc: probe of c8100000.sys-ctrl:clock-controller failed with error -17 [ 1.685152] soc soc0: Amlogic Meson GXBB (S905) Revision 1f:c (0:1) Detected [ 1.687516] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled According to C2 schematics, UART_AO_A is on AO bank. Maybe we should check on that part? Edit: seems to be located in drivers/clk/meson/meson-aoclk.c, according to the error message meson_aoclkc_probe() failed. Edit: I'm lost in tracing back the error number. Could be KDB_BADREG. Seems to happen in clk_hw_register() or clk_create_clk().
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 Another observation, which indicates that the AO bank may have problems: one pin AO_BIT2 is used to control the main DC/DC regulator (VCCK), and switching off the Odroid C2 leads to a reboot instead of shutdown. So my best guess is that something in bank AO goes wrong, and both effects have the same root.
Igor Posted April 11, 2019 Posted April 11, 2019 With 4.19.y (NEXT) there is something wrong while default 4.14.y must work. I will check later on.
martinayotte Posted April 11, 2019 Posted April 11, 2019 5 hours ago, mboehmer said: [ 1.685152] soc soc0: Amlogic Meson GXBB (S905) Revision 1f:c (0:1) Detected [ 1.687516] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled On my NanoPiK2, I'm getting : [ 1.690179] soc soc0: Amlogic Meson GXBB (Unknown) Revision 1f:c (13:1) Detected [ 1.692586] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled [ 1.693676] c81004c0.serial: ttyAML0 at MMIO 0xc81004c0 (irq = 12, base_baud = 1500000) is a meson_uart [ 2.504039] printk: console [ttyAML0] enabled In my case, I'm running Stretch Server 5.0.4 Mainline DEV.
Igor Posted April 11, 2019 Posted April 11, 2019 3 minutes ago, martinayotte said: in my case, I'm running Stretch Server 5.0.4 Mainline DEV. It looks like 4.19.y is broken. Will do more tests when possible.
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 38 minutes ago, Igor said: It looks like 4.19.y is broken. Will do more tests when possible. Indeed. Seems to be bank AO related (while that could also be somehow quirky in 5.77, as switching off failed there, too, but that's just my feeling). Is there any way to force the Armbian build system to use an older version? I had 5.77 running at least, so while you look at it (much appreciated!) I could continue with the GPIO stuff.
Igor Posted April 11, 2019 Posted April 11, 2019 50 minutes ago, mboehmer said: Is there any way to force the Armbian build system to use an older version? https://docs.armbian.com/Developer-Guide_User-Configurations/#user-provided-image-customization-script Add KERNELBRANCH="tag:v4.19.6" or which kernel version you want to build to userpatches/lib.config
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 1 hour ago, Igor said: Add KERNELBRANCH="tag:v4.19.6" or which kernel version you want to build to userpatches/lib.config Did I tell you already that the build system is really a miracle? Will try, had problems building Jessie today, apparently packages couldn't be downloaded. Starting new attempt right now.
martinayotte Posted April 11, 2019 Posted April 11, 2019 7 minutes ago, mboehmer said: apparently packages couldn't be downloaded This is because Jessie is EoL since June 6, 2018 ...
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 32 minutes ago, martinayotte said: This is because Jessie is EoL since June 6, 2018 ... Args. More coffee... Need a stretch image. Hope you guys are sometimes around Munich, I owe you a beer or two BTW, where do I find the available version numbers for kernels? Or better, the kernel sources? Maybe I can try to see if there were changes in code to touch AO banks?
martinayotte Posted April 11, 2019 Posted April 11, 2019 4 minutes ago, mboehmer said: Hope you guys are sometimes around Munich, I owe you a beer or two I'm a bit too far, I'm in Quebec ... But I like Hefeweizen ... 5 minutes ago, mboehmer said: BTW, where do I find the available version numbers for kernels? Or better, the kernel sources? Maybe I can try to see if there were changes in code to touch AO banks? I'm suggest to try Mainline Dev 5.0.y first, since A0 bank is working on my NanoPiK2, which is currently on 5.0.4. Otherwise, all tags can been seen in this github : https://github.com/torvalds/linux/ , there is a "branch" dropdown with two tabs, one for branch and one for tags.
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 53 minutes ago, martinayotte said: I'm a bit too far, I'm in Quebec ... But I like Hefeweizen ... I'm suggest to try Mainline Dev 5.0.y first, since A0 bank is working on my NanoPiK2, which is currently on 5.0.4. Otherwise, all tags can been seen in this github : https://github.com/torvalds/linux/ , there is a "branch" dropdown with two tabs, one for branch and one for tags. I may show up sooner or later again in Victoria. Too far from Quebec, I assume - but the Odroids will go to Canadian shore Hefeweizen... yes. We have such kind here. And many more styles, like the Starkbier. Will start some new compilations tomorrow, just finished a test with stretch, but the setup is at university.
martinayotte Posted April 11, 2019 Posted April 11, 2019 12 minutes ago, mboehmer said: And many more styles, like the Starkbier. Of course ! Starkbier can be dangerous if drank too fast ... I've maybe one hidden somewhere here : 1
mboehmer Posted April 11, 2019 Author Posted April 11, 2019 40 minutes ago, martinayotte said: Of course ! Starkbier can be dangerous if drank too fast ... I see, you are taking the issue "beer" very serious. But all that bottles seem to share a problem: they are empty For me, my bottles are stored in the cellar - not empty (but opened). But Vodka doesn't expire really while beer does (not in your place, dude, I guess). Starkbier is something which should be enjoyed with care. I remember some American students visiting us when I was a young boy - they made archeological excavations of some Germanic site nearby. We had some barbecue, and my dad served Starkbier instead of the normal "beer" they were used to from the States (and he warned them to take care). It was hell of a fun to watch them after the first bottle...
martinayotte Posted April 11, 2019 Posted April 11, 2019 5 minutes ago, mboehmer said: It was hell of a fun to watch them after the first bottle... 6 minutes ago, mboehmer said: not in your place, dude, I guess Not really, except when I used to brew my own ones, I try to keep some to taste them after months or even years. But, unfortunately, I didn't brew since almost 2 years, as I says often, "time is always the missing ingredient" !
TonyMac32 Posted April 12, 2019 Posted April 12, 2019 7 hours ago, mboehmer said: Hefeweizen... yes. We have such kind here. And many more styles, like the Starkbier. I rather enjoy a Dunkel myself, I've limited my exposure to Starkbier out of self-preservation, we over-industrialized beer in this country, thankfully we're coming back from it. 7 hours ago, mboehmer said: We had some barbecue, and my dad served Starkbier instead of the normal "beer" they were used to from the States (and he warned them to take care). It was hell of a fun to watch them after the first bottle... How many of them were even allowed to drink in their home country? I had an ugly beer that @martinayotte may be able to be blamed for, they named it perfectly : Now, for the AO issue, @Neil Armstrong, I did a quick scan of the mailing list but didn't come across anything, is there something kernel-wise that has been fixed?
mboehmer Posted April 12, 2019 Author Posted April 12, 2019 Hi, just could test now the 4.19.6 stretch compilation - that one works as expected, and gives me a console on /dev/ttyAML0. Console works again, I can log in. GPIO problem with I2c interference is still there, i.e. exporting and setting GPIO 475 to out kills I2C access. Michael
mboehmer Posted May 15, 2019 Author Posted May 15, 2019 HI guys, sorry for not giving earlier feedback (been busy with payload for our sounding rocket, no C2 included this time, guess it doesn't like 15g)... I just created a new image today (ARMBIAN 5.86 user-built Debian GNU/Linux 9 (stretch) 4.19.43-meson64), and console seems to be operational again. Thanks to everybody involved to fix that issue! Michael
Recommended Posts