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

["Konrad Schroder via gnats" <gnats-admin%NetBSD.org@localhost>]

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

From: Konrad Schroder <perseant%hhhh.org@localhost>
To: Shinichi Doyashiki <clare%csel.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: Tue, 22 Sep 2026 09:37:17 -0700

 On 9/21/2026 7:33 PM, Shinichi Doyashiki wrote:
 > [  81.0289078] panic: kernel diagnostic assertion "(lfs_fi_getnblocks(fs, fip) > 0 && ino >= LFS_IFILE_INUM)" failed: file "../../../../ufs/lfs/lfs_kclean.c", line 669 bad FINFO count=0 inum=0 at offset 0xb4 (seg 2)
 >
 > [...]
 >
 > Segment Summary Info at 0xaf
 >      next     0xc0       nfinfo   1      ninos    1      flags    D---
 >      sumsum   0xd589     datasum  0xaf8  create   Tue Sep 22 11:25:45 2026
 >      roll_id  76aa3189   serial   1225
 >      Inode addresses:
 >          0xbc {1v1}
 >      FINFO for inode: 1 version 1 nblocks 6 lastlength 1024
 >          0       1       23      24      26      -12
 
 This ... doesn't make sense to me.  We are either parsing the segment 
 headers wrong somehow, or we are not reading the partial-segment header 
 back from disk correctly. There *is* no FINFO with base address 0xb4: 
 that's in the middle of the one that stretches from 0xb0 to 0xbb.
 
 Would it be possible for you to send me a compressed disk image of this 
 filesystem?  If I'm reading right, the uncompressed size is only 32 MB, 
 and from context it's a test filesystem so shouldn't have any 
 confidential data in it; is that correct?  Please don't send it to 
 gnats-bugs@, though, I don't think we need quite that level of detail in 
 the ticket. If you can put it on a website and send the URL or, if the 
 compressed image is small enough, just email it to me that would be great.
 
 Thanks,
 
 --
 Konrad Schroder
 perseant%hhhh.org@localhost