Re: [PATCH 23/32] fuse: split out filesystem part of request sending

David CARLIER <[email protected]> Sat, 27 Jun 2026 21:16:06 +0100
Newsgroups dev.linux.lists.fuse-devel,org.kernel.vger.linux-kernel
Message-ID <CA+XhMqye_NBihcadRTU=w4xAfrdY9rsMB3oHF5+bwHukkjHMoQ@mail.gmail.com>
Hi Miklos,

This change alters the pid sent on force && nocreds requests — intended?

Previously these allocated a zeroed request and skipped fuse_force_creds(),
so they went out with in.h.pid == 0. In fuse_fill_creds(), args->pid is
now set before the nocreds early return, and fuse_args_to_req() copies it
unconditionally — so writeback, release, and DESTROY requests now report the
triggering (often kernel-worker) pid instead of 0. uid/gid sit below the
return and are still left at 0.

Looks like the args->pid assignment was hoisted above the nocreds check by
accident. If keeping these requests credential-free is intended, setting pid
only where uid/gid are filled would restore it. If unconditional pid is
deliberate, disregard.
Cheers