Bug 81450 - [patch] ATAPI support broken in -STABLE
Summary: [patch] ATAPI support broken in -STABLE
Status: Closed FIXED
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: 5.4-STABLE
Hardware: Any Any
: Normal Affects Only Me
Assignee: Søren Schmidt
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-05-24 21:40 UTC by Sean McNeil
Modified: 2005-06-27 20:20 UTC (History)
0 users

See Also:


Attachments
file.diff (540 bytes, patch)
2005-05-24 21:40 UTC, Sean McNeil
no flags Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Sean McNeil 2005-05-24 21:40:01 UTC
      ata-queue.c:ata_completed() will issue a sense request when it encounters an ATAPI error.  This request fails to zero out the donecount which causes corruption of memory.  On amd64, it overwrites the callback value and crashes the computer.

Fix: The following patch fixes the crash:
How-To-Repeat:       Try to burn a CD from nautilus.  Since nautilus just invokes cdrecord, it should cause the same result by running cdrecord from a command line.
Comment 1 Mark Linimon freebsd_committer freebsd_triage 2005-05-24 23:32:46 UTC
Responsible Changed
From-To: freebsd-bugs->sos

Over to maintainer.
Comment 2 Søren Schmidt freebsd_committer freebsd_triage 2005-06-27 13:34:42 UTC
State Changed
From-To: open->closed

This is solved (differently) in -current and the ATA mkIII patches for 5-stable.
Comment 3 Sean McNeil 2005-06-27 20:18:30 UTC
This bug is being closed in error.  Please reopen it until the actual
bug is fixed by:

a) Applying my patch to -current.
b) Applying the ATA mkIII patches to -current.

Neither of these things have occurred and thus the bug still exists!

I am actually upset with the bug tracking software that I cannot reopen
this bug myself as it is an obvious mistake closing something before an
actual fix has been implemented.  Especially since this is a verified
problem.

Cheers,
Sean