2026-06-27 11:57 UTC
#Linux UX root cause analysis
I cannot play the new album I bought on the living room speakers,
because mpd on Raspbian trixie crashes when updating the libnfs-mounted database,
because of some bug in libnfs14 or how mpd in stable uses libnfs14.
I cannot use a kernel-based nfs mountpoint,
because mpd on Raspbian stable gets stuck during playback when running off nfs,
because of unknown reasons.
I cannot use a proxy database,
because on the NFS server, mpd gets stuck scanning the library immediately,
because of a kernel bug involving io_uring (which mpd uses),
because of an ancient kernel running on the NAS,
because newer kernels fail to boot,
because of a buggy PCIe root cluster in the rk3399 SoC triggering kernel panics and/or making the SATA card not enumerate, rendering the NAS useless.
Replies (1)
-
@jssfr@zombofant.net 2026-06-27 12:07
To add insult to injury, the mpd/libnfs14 crash does not occur on vanilla Debian (so I cannot reportbug it). Well, on the plus side … I can just use nfs://127.0.0.1 on the NAS to sidestep the io_uring bug. That is crazily inefficient, but at least nothing crashes or gets stuck….