Looks like some form of regression in RELENG_8, between the dates of 2010/03/30 and 2010/04/22. My 2010/03/30 kernel (built from RELENG_8 source dated 2010/03/30 @ 10:30 PDT) doesn't have this problem. Symptom: shutdown -p no longer powers off either of my systems; instead, the box reboots (e.g. shutdown -p behaves like shutdown -r). Systems are a Supermicro X7SBL-LN2 board, and a Supermicro X7SBA board. These are server-class boards. I have not tried using the physical power switch yet. Looks like others have reported the same on a different board, though I haven't booted verbose to see if the behaviour is identical: http://lists.freebsd.org/pipermail/freebsd-acpi/2010-March/006345.html http://lists.freebsd.org/pipermail/freebsd-acpi/2010-April/006462.html I went digging through HEAD commits between the above dates and wasn't able to find much other than this, which appears to be the HEAD commit that was MFC'd to RELENG_8: http://freshbsd.org/2010/04/02/23/04/31 So I'm not sure where/what fixed this in HEAD, but it should probably be MFC'd sooner than the 4 week window. -- | Jeremy Chadwick jdc@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB |
on 22/04/2010 17:35 Jeremy Chadwick said the following:> I went digging through HEAD commits between the above dates and wasn't > able to find much other than this, which appears to be the HEAD commit > that was MFC'd to RELENG_8: > > http://freshbsd.org/2010/04/02/23/04/31I don't think this has been MFC-ed to 8. Most likely a red herring. AFAIK, there were no ACPI-related changes in 8 for the last couple of months, so the issue must be elsewhere. -- Andriy Gapon
On 22-04-2010 7:35, Jeremy Chadwick wrote:> Looks like some form of regression in RELENG_8, between the dates of > 2010/03/30 and 2010/04/22. My 2010/03/30 kernel (built from RELENG_8 > source dated 2010/03/30 @ 10:30 PDT) doesn't have this problem. > > Symptom: shutdown -p no longer powers off either of my systems; instead, > the box reboots (e.g. shutdown -p behaves like shutdown -r). Systems > are a Supermicro X7SBL-LN2 board, and a Supermicro X7SBA board. These > are server-class boards. I have not tried using the physical power > switch yet.For the record: I have the same problem on at least two boxes, both running 9-CURRENT. -- Joel
On Thu, 22 Apr 2010 07:35:42 -0700, Jeremy Chadwick wrote:> Looks like some form of regression in RELENG_8, between the dates of > 2010/03/30 and 2010/04/22. My 2010/03/30 kernel (built from RELENG_8 > source dated 2010/03/30 @ 10:30 PDT) doesn't have this problem.Does not work for me either since I updated my system on April 4.
On Thu, Apr 22, 2010 at 07:35:42AM -0700, Jeremy Chadwick wrote:> Looks like some form of regression in RELENG_8, between the dates of > 2010/03/30 and 2010/04/22. My 2010/03/30 kernel (built from RELENG_8 > source dated 2010/03/30 @ 10:30 PDT) doesn't have this problem. > > Symptom: shutdown -p no longer powers off either of my systems; instead, > the box reboots (e.g. shutdown -p behaves like shutdown -r). Systems > are a Supermicro X7SBL-LN2 board, and a Supermicro X7SBA board. These > are server-class boards. I have not tried using the physical power > switch yet. > > Looks like others have reported the same on a different board, though I > haven't booted verbose to see if the behaviour is identical:I'm having the same problem on an HP DC7800 Desktop since updating the system yesterday. I'm using 8-STABLE. -- Guido Falsi <mad@madpilot.net>
On Thu, Apr 22, 2010 at 07:35:42AM -0700, Jeremy Chadwick wrote:> Symptom: shutdown -p no longer powers off either of my systems; instead, > the box reboots (e.g. shutdown -p behaves like shutdown -r). Systems > are a Supermicro X7SBL-LN2 board, and a Supermicro X7SBA board. These > are server-class boards. I have not tried using the physical power > switch yet.Manual power switch behaves the same way (and given what I understand of ATX and ACPI, this is normal). The only way I found to shut the machine off is to press the power button shortly after POST.> I went digging through HEAD commits between the above dates ...I realised I'll need to find when the ACPICA code was originally committed to HEAD (probably much earlier than RELENG_8!), then work forwards to present day to try and find the committed fix. -- | Jeremy Chadwick jdc@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB |