amd64/189267: AMD Kabini (AM1/FM1Sb) Frequency Switching Causes Crash

[Brodey Dover]

The following reply was made to PR amd64/189267; it has been noted by GNATS.

From: Brodey Dover <doverosx at gmail.com>
To: Jeremias Blendin <jeremias at blendin.org>
Cc: bug-followup at freebsd.org
Subject: Re: amd64/189267: AMD Kabini (AM1/FM1Sb) Frequency Switching Causes Crash
Date: Sun, 4 May 2014 11:03:07 -0400

 --089e01493df60178f704f89450e9
 Content-Type: text/plain; charset=UTF-8
 
 I also noticed that HPET isn't an eventtimer option in spite of it being
 enabled in the BIOS.
 
 [brodey at freenas] /# sysctl -a | grep eventtimer
 kern.eventtimer.choice: LAPIC(400) i8254(100) RTC(0)
 kern.eventtimer.et.LAPIC.flags: 15
 kern.eventtimer.et.LAPIC.frequency: 50000961
 kern.eventtimer.et.LAPIC.quality: 400
 kern.eventtimer.et.i8254.flags: 1
 kern.eventtimer.et.i8254.frequency: 1193182
 kern.eventtimer.et.i8254.quality: 100
 kern.eventtimer.et.RTC.flags: 17
 kern.eventtimer.et.RTC.frequency: 32768
 kern.eventtimer.et.RTC.quality: 0
 kern.eventtimer.periodic: 0
 kern.eventtimer.timer: LAPIC
 kern.eventtimer.activetick: 1
 kern.eventtimer.idletick: 0
 kern.eventtimer.singlemul: 2
 
 I also note that one core has a different supported state (maybe this is an
 AMD implementation):
 
 hw.acpi.cpu.cx_lowest: C1
 dev.cpu.0.cx_supported: C1/1/0 C2/2/400
 dev.cpu.0.cx_lowest: C1
 dev.cpu.0.cx_usage: 100.00% 0.00% last 98us
 dev.cpu.1.cx_supported: C1/1/0
 dev.cpu.1.cx_lowest: C1
 dev.cpu.1.cx_usage: 100.00% last 235us
 dev.cpu.2.cx_supported: C1/1/0
 dev.cpu.2.cx_lowest: C1
 dev.cpu.2.cx_usage: 100.00% last 17us
 dev.cpu.3.cx_supported: C1/1/0
 dev.cpu.3.cx_lowest: C1
 dev.cpu.3.cx_usage: 100.00% last 16us
 
 Seeing C1 state supported and it's usage makes me wonder where my other
 wattage is being used. I think the GPU isn't being switched down.
 
 
 On Sat, May 3, 2014 at 5:17 AM, Jeremias Blendin <jeremias at blendin.org>wrote:
 
 > I have the same problem with an AMD Athlon 5350 on Asrock AM1H-ITX. I did
 > not verify the powerd solution yet. However, the experience is similar,
 > shortly after booting the system would reset. I tried FreeBSD 10 release
 > and stable, in both cases I activated powerd in the installation process
 > and had the described reboot issue.
 > I suspected a driver issue and switched to a FreeBSD 11 snapshot which
 > works perfectly. I guess the reason is that powerd is not enabled on these
 > snapshots.
 >
 
 --089e01493df60178f704f89450e9
 Content-Type: text/html; charset=UTF-8
 Content-Transfer-Encoding: quoted-printable
 
 <div dir=3D"ltr"><div><div>I also noticed that HPET isn't an eventtimer=
  option in spite of it being enabled in the BIOS. <br><br>[brodey at freenas] =
 /# sysctl -a | grep eventtimer<br>kern.eventtimer.choice: LAPIC(400) i8254(=
 100) RTC(0)<br>
 kern.eventtimer.et.LAPIC.flags: 15<br>kern.eventtimer.et.LAPIC.frequency: 5=
 0000961<br>kern.eventtimer.et.LAPIC.quality: 400<br>kern.eventtimer.et.i825=
 4.flags: 1<br>kern.eventtimer.et.i8254.frequency: 1193182<br>kern.eventtime=
 r.et.i8254.quality: 100<br>
 kern.eventtimer.et.RTC.flags: 17<br>kern.eventtimer.et.RTC.frequency: 32768=
 <br>kern.eventtimer.et.RTC.quality: 0<br>kern.eventtimer.periodic: 0<br>ker=
 n.eventtimer.timer: LAPIC<br>kern.eventtimer.activetick: 1<br>kern.eventtim=
 er.idletick: 0<br>
 kern.eventtimer.singlemul: 2<br><br></div>I also note that one core has a d=
 ifferent supported state (maybe this is an AMD implementation):<br><br>hw.a=
 cpi.cpu.cx_lowest: C1<br>dev.cpu.0.cx_supported: C1/1/0 C2/2/400<br>dev.cpu=
 .0.cx_lowest: C1<br>
 dev.cpu.0.cx_usage: 100.00% 0.00% last 98us<br>dev.cpu.1.cx_supported: C1/1=
 /0<br>dev.cpu.1.cx_lowest: C1<br>dev.cpu.1.cx_usage: 100.00% last 235us<br>=
 dev.cpu.2.cx_supported: C1/1/0<br>dev.cpu.2.cx_lowest: C1<br>dev.cpu.2.cx_u=
 sage: 100.00% last 17us<br>
 dev.cpu.3.cx_supported: C1/1/0<br>dev.cpu.3.cx_lowest: C1<br>dev.cpu.3.cx_u=
 sage: 100.00% last 16us<br><br></div>Seeing C1 state supported and it's=
  usage makes me wonder where my other wattage is being used. I think the GP=
 U isn't being switched down.<br>
 </div><div class=3D"gmail_extra"><br><br><div class=3D"gmail_quote">On Sat,=
  May 3, 2014 at 5:17 AM, Jeremias Blendin <span dir=3D"ltr"><<a href=3D"=
 mailto:jeremias at blendin.org" target=3D"_blank">jeremias at blendin.org</a>>=
 </span> wrote:<br>
 <blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
 x #ccc solid;padding-left:1ex"><div dir=3D"ltr"><div>I have the same proble=
 m with an AMD Athlon 5350 on Asrock AM1H-ITX. I did not verify the powerd s=
 olution yet. However, the experience is similar, shortly after booting the =
 system would reset. I tried FreeBSD 10 release and stable, in both cases I =
 activated powerd in the installation process and had the described reboot i=
 ssue.<br>
 
 </div><div>I suspected a driver issue and switched to a FreeBSD 11 snapshot=
  which works perfectly. I guess the reason is that powerd is not enabled on=
  these snapshots.</div></div>
 </blockquote></div><br></div>
 
 --089e01493df60178f704f89450e9--