XECDesign
XECDesign
> Same problem, pi 5 Same explanation as before. I may make the script fail more gracefully in this situation make it warn the user about what happened, but it...
I can reproduce the issue and have managed to get some useful output. I think what's happening is that the udev rule runs before the 'system' node shows up, so...
The right thing happens and then the symlink gets replaced. ``` Aug 30 16:31:08 dmatest (udev-worker)[686]: system: Successfully created symlink '/dev/dma_heap/vidbuf_cached' to '/dev/dma_heap/system' ... Aug 30 16:31:08 dmatest (udev-worker)[682]: linux,cma:...
I don't know what the right fix is right now, but to temporarily work around the issue you can edit `/lib/udev/rules.d/60-dma-heap.rules` and comment out this line: ``` SUBSYSTEM=="dma_heap", KERNEL=="linux,cma", SYMLINK+="dma_heap/vidbuf_cached"...
Testing with `, OPTIONS+="link_priority=-50"` added to the end of that rule. If I am understanding it right, then that should mean that when it notices there's already a symlink there...
I removed everything else because I don't think it's needed. What I pasted above should be the whole file.
I would expect the fix to ship some time next week
The examples should now build as expected under bookworm and trixie.
@ghollingworth should we add a link to PINN and archive the repo?
@procount would you prefer a link to github or sourceforge?