Fix "vrefact: wrong use count 0" with DRM

[trasz]

Summary

Bump the vnode use count, not just its hold count. This fixes a panic
triggered by fstatat(..., AT_EMPTY_PATH) on DRM device nodes, which happens
to be what glxinfo(1) from Ubuntu Jammy is doing.

PR: kern/274538

[trasz]

This follows a somewhat similar fix in https://reviews.freebsd.org/D29323?id=86801. I can't say I have a proper understanding of what's going on here though.

Here's the test case from Alex S:

#define _GNU_SOURCE

#include <assert.h>
#include <stdio.h>
#include <fcntl.h>
#include <sys/stat.h>

int main() {
  int fd = open("/dev/dri/card0", O_RDWR);
  assert(fd != -1);

  struct stat st;
  fstatat(fd, "", &st, AT_EMPTY_PATH);

  return 0;
}

[kib]

This cannot be right.

If the file references vnode, vnode must be active and have the use count at least 1, because file vref-ed the vnode on open.

[trasz]

Second try.

[trasz]

In D47391#1080971, @kib wrote:

This cannot be right.

If the file references vnode, vnode must be active and have the use count at least 1, because file vref-ed the vnode on open.

Good point. Okay, so here's another attempt; this time in code specific to linuxkpi.

[olce]

If going that way, then you can remove the adjacent vhold() and vdrop() as they are redundant.

[kib]

It should be fine modulo the note about vhold() no longer needed.

[trasz]

Get rid of vhold/vdrop. I've been running it for a few days
and it appears to work fine.