Re: kern/60746: kernel diagnostic assertion "ino >= LFS_IFILE_INUM" failed: file "../../../../ufs/lfs/lfs_vfsops.c", line 1800

["Shinichi Doyashiki via gnats" <gnats-admin%NetBSD.org@localhost>]

The following reply was made to PR kern/60746; it has been noted by GNATS.

From: Shinichi Doyashiki <clare%csel.org@localhost>
To: Konrad Schroder <perseant%hhhh.org@localhost>, gnats-bugs%netbsd.org@localhost
Cc: 
Subject:  Re: kern/60746: kernel diagnostic assertion "ino >= LFS_IFILE_INUM" failed:
 file "../../../../ufs/lfs/lfs_vfsops.c", line 1800
Date: Sat, 26 Sep 2026 10:40:26 +0900

 > > [  86.8886575] ifile dirty lbn 0x9 flags 0x4000 cflags 0x0 oflags 0x80
 > > [  86.8986565] ifile dirty lbn 0xb flags 0x4000 cflags 0x0 oflags 0x80
 > > [  86.9086571] ifile dirty lbn 0xd flags 0x4000 cflags 0x0 oflags 0x80
 > > [  86.9186564] ifile dirty lbn 0xe flags 0x4000 cflags 0x0 oflags 0x80
 > > [  86.9286590] ifile dirty lbn 0x10 flags 0x4000 cflags 0x0 oflags 0x80
 > > [  86.9386578] ifile dirty lbn 0x12 flags 0x4000 cflags 0x0 oflags 0x80
 > > [  86.9486576] ifile dirty lbn 0x14 flags 0x4000 cflags 0x0 oflags 0x80
 > > [  86.9586575] ifile dirty lbn 0x16 flags 0x4000 cflags 0x0 oflags 0x80
 > > [  86.9686574] panic: lfs_writeinode: looping
 > 
 > Aha!  We'd been assuming that this looping behavior was caused by a 
 > buffer being held busy by another process, and before it could have 
 > been.  But here you can see from the timestamps that it is a different 
 > buffer every iteration.  If the Ifile blocks are spread out all over the 
 > disk, we might have to rewrite the entire segment table, one pass at a 
 > time; so 10 iterations might be too few to merit a panic.  Please try 
 > this patch instead (also attached):
 
 Thank you.
 The reproducer script has completed over 1200 mount/write/remove/umount/fsck cycles
 on the 32 MB filesystem without reproducing panic.
 both userland cleaner and in-kernel cleaner was tested, it seems okay.