xloem

Validating
  • Content Count

    4
  • Joined

  • Last visited

  1. I'm not familiar with the a64 graphics driver and some tty graphics software of mine is failing on Armbian Ubuntu Focal after working fine on the Bionic image. Earlier I was able to mmap() /dev/fb0 to display graphics in a tty. On Focal 20.04, running linux 5.4.43-sunxi64 , mmap() on /dev/fb0 is failing with EPERM (operation not permitted). Is this new error something anybody knows about or could give a link to find a possible driver change that's responsible? It seems there might be a few different video drivers on this system and I haven't looked at them before. /sys/class/graphics/fb0/name contains sun4i-drmdrmfb /sys/class/graphics/fb0/device/modalias contains of:Ndisplay-engineT(null)Callwinner,sun50i-a64-display-engine Thank you.
  2. To update, this kind of situation seems much eased on the newer images, where early boot output is visible on the screen.
  3. After much comparing, I think the issue is that I had set the default systemd target to "emergency" to launch a single-user shell to maintain the system. This seems to prevent booting.
  4. Hi, I can boot off an SD card, but my Teres boots off internal storage to a black screen and I'm trying to figure out why. It seems even when I set verbose=7 in /boot/armbianEnv.txt, there is little to no kernel diagnostic output, even on a working SD card. I read on the debian wiki that u-boot-menu works on the Teres, but I tried installing this on an SD card and it prevents the SD card from booting (similarly to my internal storage booting issue). I've actually compared the contents of the unallocated space in front of the filesystem between the working sdcard and the internal storage, and they differ only in the partition table. Additionally /lib/firmware, /lib/modules, /boot/dtb-5.4.28-sunxi64, /boot/boot.cmd and /boot/vmlinuz-5.4.28-sunxi64 all compare byte-for-byte identical. /boot/boot.scr differs in only 8 bytes (0x4-0xC) which could be a timestamp. /boot/armbianEnv.txt differs in only the value of rootdev. The initrd differs between the two in that the contents of blacklist-framebuffer.conf, libdns-export.so.1100, and libisc-export.so.169 differ. blacklist-framebuffer.conf differs only in that the internal storage has bochs-drm added to the blacklisting. I wasn't sure how to extract the uinitrd to see if it had corruption. I tried replacing my differing files with those from the working sdcard but the system still did not boot, although it could have been for a different reason as there is no diagnostic output. Do you know a way to see more diagnostics than verbose=7 provides, or another avenue to pursue? I am likely next to try reinstalling the system and comparing files and bytes between before and after. I don't have an easy mailing address at this time to order the serial cable.