Having to press Enter each time i load the program. about steamlink-launcher HOT 6 OPEN

Same issue with my RPi4 (4GB), changing gpu_mem does not help. If i set gpu_mem to 512, OSMC will start (i can connect to SSH), but has no video output (no HDMI signal at all). Disabling check helped to skip the memory error and it works, but ignoring issues is not what i want.

EDIT: What we found out: Looks like Steam link runs its own instance of linux (in a sandbox or something like that) and that is missing vcgencmd.

What did work: adding gpu_mem=384 to [all] changed the memory to 384MB. We checked after reboot with vcgencmd in terminal. It does not matter if its KMS or FKMS. The error with vcgencmd is still there (because its missing in steam link).

from steamlink-launcher.

@Gaffers2277 As a workaround you might as well run:

touch /home/osmc/.local/share/SteamLink/.ignore_gpumem

from the terminal. That disables the check.

from steamlink-launcher.

logs please

from steamlink-launcher.

Same here with RPi3 B+. With OSMC's default:

dtoverlay=vc4-kms-v3d,cma-512

vcgencmd get_mem gpu reports the following:

gpu=76M

and steamlink.sh does not seem to like that:

..
# Check to make sure there's enough GPU memory
if [ ! -f "$TOP/.ignore_gpumem" ]; then
  mem=$(vcgencmd get_mem gpu | sed 's/.*=\([0-9]*\)M/\1/')
  if [ "$mem" = "" ] || [ "$mem" -lt 128 ]; then
    show_message --warning $"You are running with less than 128 MB video memory, you may need to go to the Raspberry Pi Configuration and increase your GPU memory."

Not sure with CMA if GPU mem is still a thing or this warning could be removed altogether.

from steamlink-launcher.

Does setting gpu_mem in the boot config make a positive change?

from steamlink-launcher.

hey guys great find ill try and make time for it or if someone is up for making a patch and merge it

from steamlink-launcher.