Bug 206923 - lock order reversal, vfs_bio.c:3476 ufs_dirhash.c:281
Summary: lock order reversal, vfs_bio.c:3476 ufs_dirhash.c:281
Status: Closed Not A Bug
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: CURRENT
Hardware: amd64 Any
: --- Affects Only Me
Assignee: freebsd-bugs (Nobody)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-02-04 12:56 UTC by Tim Chase
Modified: 2020-04-03 16:48 UTC (History)
3 users (show)

See Also:


Attachments
Screen capture of error, traceback and output of "uname -a" (176.46 KB, image/jpeg)
2016-02-04 12:59 UTC, Tim Chase
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Tim Chase 2016-02-04 12:56:06 UTC
I don't have an easy way to capture the back-trace and pull it from the machine right now, so I've attached a screen-cap and hand-transcribed salient bits:

    witness_debugger+0x70
    witness_checkorder+0xe71
    _sx_xlock+0x72
    ufsdirhash_add+0x3d
    ufs_direnter+0x62f
    ufs_makeinode+0x5f3
    ufs_create+0x2d
    VOP_CREATE_APV+0xf1
    vn_open_cred+0x2f8
    kern_openat+0x25c
    amd64_syscall+0x2db
    Xfast_syscall+0xfb

The system was a fresh install of CURRENT downloaded last night, put on a Dell Inspiron i1420 with pretty default configuration (one whole disk GPT partition with a boot, / and swap configuration) mostly just to test some jail stuff.  Only new packages installed were zip/unzip/bunzip/bunzip2 so I could uncompress the Handbook (html-split).  It was triggered in /usr/share/doc/handbook/html (the last two subdirectories being subdirectories I'd created to house the handbook).  The error was triggered by executing "tar xf book.html-split.tar"
Comment 1 Tim Chase 2016-02-04 12:59:44 UTC
Created attachment 166558 [details]
Screen capture of error, traceback and output of "uname -a"

For some reason the original didn't attach when creating the case, so I shrank it to 30% to see if it was a size issue and am retrying the attachment.
Comment 2 Allan Jude freebsd_committer freebsd_triage 2019-01-04 02:26:43 UTC
This is a known 'false positive' in Witness. We do not currently have a good way to supress this message even when it is a known false positive.
Comment 3 Mark Johnston freebsd_committer freebsd_triage 2020-04-03 16:48:21 UTC
We now suppress reports of this false positive.