Werner

Never tried that myself on H3. You can also try to install mesa from scratch. That did the trick for me with Panfrost. https://gitlab.freedesktop.org/lima/web/blob/master/README.md

https://wiki.debian.org/PanfrostLima

Ah okay, so an upstream "issue". Then never mind but keep an eye for

Linux kernel version.

A no-go. Python2 is EOL.

If this is a generic question: I don't know. If you actually have this board then use the nand-sata-install script and it should do the trick

Because it is not permanent. You could use rc.local for example to turn off the leds once the boot has been finished.

May have been removed while merging upstream config. I'll take a look. Aight there are two options for you: - Replace the legacy with the current kernel which contains this module or - Use the build script to create your own legacy kernel package which then will be up-to-date as well

Neat. Hm so no raid possible?

20.05.3 Known issue. Either start over with the most recent image or fix it by changing the compression algorythm of you ramdisk to gzip.

drivers/video/backlight/pwm_bl.c: In function ‘pwm_backlight_probe’: drivers/video/backlight/pwm_bl.c:639:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] u64 level = (u64)state.duty_cycle * pb->scale; ^~~ drivers/gpu/drm/sun4i/sun8i_ui_layer.c: In function ‘sun8i_ui_layer_enable’: drivers/gpu/drm/sun4i/sun8i_ui_layer.c:105:2: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] unsigned tmp; ^~~~~~~~ drivers/gpu/drm/sun4i/sun8i_ui_layer.c: In function ‘sun8i_ui_layer_update_formats’: drivers/gpu/drm/sun4i/sun8i_ui_layer.c:224:2: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] unsigned tmp; ^~~~~~~~ drivers/gpu/drm/sun4i/sun8i_vi_layer.c: In function ‘sun8i_vi_layer_enable’: drivers/gpu/drm/sun4i/sun8i_vi_layer.c:98:2: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] unsigned tmp; ^~~~~~~~ drivers/gpu/drm/sun4i/sun8i_vi_layer.c: In function ‘sun8i_vi_layer_update_formats’: drivers/gpu/drm/sun4i/sun8i_vi_layer.c:264:2: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] unsigned tmp; ^~~~~~~~ drivers/mtd/nand/raw/nand_base.c: In function ‘nand_block_isbad’: drivers/mtd/nand/raw/nand_base.c:4292:6: warning: unused variable ‘ret’ [-Wunused-variable] int ret; ^~~ drivers/mtd/nand/raw/nand_base.c:4291:6: warning: unused variable ‘chipnr’ [-Wunused-variable] int chipnr = (int)(offs >> chip->chip_shift); ^~~~~~ drivers/net/ethernet/stmicro/stmmac/dwmac-sun8i.c: In function ‘sun8i_dwmac_shutdown’: drivers/net/ethernet/stmicro/stmmac/dwmac-sun8i.c:1277:2: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] struct stmmac_priv *priv = netdev_priv(ndev); ^~~~~~ drivers/firmware/arm_scpi.c:815:12: warning: ‘scpi_init_versions’ defined but not used [-Wunused-function] static int scpi_init_versions(struct scpi_drvinfo *info) ^~~~~~~~~~~~~~~~~~ drivers/gnss/ubx.c: In function ‘ubx_set_standby’: drivers/gnss/ubx.c:49:24: warning: unused variable ‘serdev’ [-Wunused-variable] struct serdev_device *serdev = gserial->serdev; ^~~~~~ ERROR: modpost: "of_chosen" [drivers/gpu/drm/sun4i/sun6i_mipi_dsi.ko] undefined! ERROR: modpost: "of_chosen" [drivers/phy/allwinner/phy-sun6i-mipi-dphy.ko] undefined! make[1]: *** [__modpost] Error 1 make: *** [modules] Error 2 Well, that is unfortunate but my own problem since I customized the config by throwing out tons of unneeded stuff. Default config builds fine. Thanks for the hint. This actually fixes this. For everybody else to try: diff --git a/drivers/of/base.c b/drivers/of/base.c index ae03b1218..9d42a28bd 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -36,6 +36,7 @@ LIST_HEAD(aliases_lookup); struct device_node *of_root; EXPORT_SYMBOL(of_root); struct device_node *of_chosen; +EXPORT_SYMBOL(of_chosen); struct device_node *of_aliases; struct device_node *of_stdout; static const char *of_stdout_options;

Nice job!

Ethernet on all H6 boards should work as expected. Connect an USB serial adapter to get an idea what is happening inside.

It is a better idea to start from scratch.especially on Focal. There is a regression that causes the OS after upgrading to be unable to boot. This can be fixed manually though when changing the compression algorythm for the initramfs from lz4 to gzip. Check /etc/initramfs/initramfs.conf (I think it was) file.

Please retry with the most recent version.

You're welcome. Let me know if you have further questions. Also feel free to chit chat in our IRC channel. Link down below

That cannot be answered by a simple yes or no. It depends. Let's take the 5.4 example. If you follow the current branch it leads us to megi's orangepi-5.4 branch. If you check the Makefile (https://github.com/megous/linux/blob/orange-pi-5.4/Makefile) you will notice that the actual version of this branch is 5.4.18. But when you compile the kernel it is actually 5.4.47 or something like that. This means these patches come from Armbian. If you check the patch directory for your board family (https://github.com/armbian/build/tree/master/patch/kernel/sunxi-current) you can see that the upstream patches are added here. If you need a very specific kernel version remove or rename the patches you do not need. That was a quite easy example to get what you want. On other sources, vanilla for example, it is a bit tougher. You would need to specify a commit at which point you want to use the sources. Sometimes a board is fixed to a specific commit or kernel version because it is known that newer version of the same kernel branch are known to be broken and (as most of the times) nobody has time or resources to deal with it.

You can get some additional space back when clearing the rootfs and sources in cache directory. Also note that ccache is used to accelerate building further images. Check /root/.ccache

I am not sure what the actual issue is. When you build a bootable image for the Bananipi M2 zero does it boot as expected? This SBC is handled as a part of the sunxi family and you might notice that there are many more device tree blobs under /boot/dtb.

Not just because of that. Also ccache will make the next build much faster.

Um... you mean the built image is using the wrong DTS to boot?

reserved

I could compile a kernel package for a randomly selected board and kernel so I assume everything works as expected now.

The minimal installation seems lacking gpg. apt-get install gpg This fixes it. About 2013: Some boards, especially older ones need older toolchain to compile older sources...sometimes stuff breaks in upstream ¯\_(ツ)_/¯

Confirmed. Okay, seems like a general issue, not torrent related. USE_TORRENT=no leads into the same issue.