Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 28 Oct 2018 18:49:28 -0700
From:      Mark Millard <>
To:        Michael Tuexen <>
Cc:        freebsd-ppc <>, Koakuma <>
Subject:   Re: G5 fans
Message-ID:  <>
In-Reply-To: <>
References:  <> <>

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help
[I looked on a 7,2 via ofwdump.]

On 2018-Oct-28, at 6:10 PM, Mark Millard <marklmi at> wrote:

> On 2018-Oct-28, at 5:31 PM, Koakuma via freebsd-ppc =
<> wrote:
>> Hi Michael,
>> The first command failed to run for me.
>> 0 > dev u3/i2c/fan can't find device u3/i2c/fan ok
>> Looking at the output of 'dev / ls', there's several devices named
>> "fan" under /smu@0,0/fans@4a00/*, but running the " .status" command
>> on those devices gave me this error:
>> 0 > " .status" " /smu@0,0/fans@4a00/fan@0" open-dev $call-method
>> Method <.status> not found; ihandle=3Dffb352c0 phandle=3Dff99ee00
>> ok
>> This is on a PowerMac11,2, 2 GHz "dual-core" model.
> A so-called G5 "Quad Core" suggests something similar
> for finding fans when looked at via ofwdump:

I probably should have listed:

# ofwdump -p /
Node 0xff89a7a8: device-tree
    50 6f 77 65 72 4d 61 63 31 31 2c 32 00=20
. . .

# ofwdump -r /cpus
Node 0xff89d3f0: cpus
  Node 0xff89d680: PowerPC,G5
    Node 0xff89db28: l2-cache
  Node 0xff89eb70: PowerPC,G5
    Node 0xff89f070: l2-cache
  Node 0xff89f248: PowerPC,G5
    Node 0xff89f748: l2-cache
  Node 0xff89f920: PowerPC,G5
    Node 0xff89fe20: l2-cache

# ofwdump -r /smu/fans | less
Node 0xff99f698: fans
  Node 0xff99fbb0: fan
  Node 0xff9a04a8: fan
  Node 0xff9a0d98: fan
  Node 0xff9a1688: fan
  Node 0xff9a1f78: fan
  Node 0xff9a2868: fan
  Node 0xff9a3158: fan
  Node 0xff9a3a48: fan

And also where fan is not (compared
to the 7,2):

# ofwdump -r /u4 | less
Node 0xff97c040: u4
  Node 0xff97c3b8: i2c
    Node 0xff97cfe0: i2c-bus
      Node 0xff97d370: temp-monitor
        Node 0xff97d7e8: temperature
      Node 0xff97dcb0: temp-monitor
        Node 0xff97e158: temperature
      Node 0xff97e448: temp-monitor
        Node 0xff97eff0: internal-temperature
        Node 0xff97f258: external-temperature
      Node 0xff97fbc8: temp-monitor
        Node 0xff980770: internal-temperature
        Node 0xff9809d8: external-temperature
  Node 0xff981348: dart
  Node 0xff981488: mpic

I show the analogous 7,2 /u3 later.

> # ofwdump -pr /smu/fans | less
> . . .
>> On Sun Oct 28 22:02:15 UTC 2018, Michael Tuexen <tuexen at =>
>> wrote:
>>> Dear all,
>>> if you have access to a G5 Powermac, could you provide the output of =
>>> following command from the OpenFirmware prompt:
>>> dev u3/i2c/fan
>>> " .status" " u3/i2c/fan" open-dev $call-method
>>> I'm interested in the RPM channels which are reported as =
>>> Please provide also some information what kind of PowerMac G5 it is.

As for what ofwdump reports for a 7,2
dual processor:

# ofwdump -p /
Node 0xff885330: device-tree
    50 6f 77 65 72 4d 61 63 37 2c 32 00=20
. . .

# ofwdump -r /cpus
Node 0xff887b40: cpus
  Node 0xff887e10: PowerPC,970
    Node 0xff888338: l2-cache
  Node 0xff889150: PowerPC,970
    Node 0xff889678: l2-cache

# ofwdump -r /u3
Node 0xff959968: u3
  Node 0xff959cb8: i2c
    Node 0xff95a7f8: i2c-hwclock
    Node 0xff95afb0: temp-monitor
    Node 0xff95b7c8: temp-monitor
    Node 0xff95c318: temp-monitor
    Node 0xff95ce70: i2c-hwclock
    Node 0xff95f170: fan
    Node 0xff9618e0: i2c-cpu-voltage
    Node 0xff962848: supply-monitor
    Node 0xff9640e8: cpuid
    Node 0xff966740: i2c-cpu-voltage
    Node 0xff9676a8: supply-monitor
    Node 0xff968f48: cpuid
    Node 0xff96b5a0: cereal
  Node 0xff96bd10: dart
  Node 0xff96beb0: mpic

The 11,2's seem to have no equivalent of the above fan
but have more like the other 7,2 fan info structure

# ofwdump -r /ht/pci@3/mac-io/fans
Node 0xff983720: fans
  Node 0xff9839a0: fan
  Node 0xff983b50: fan
  Node 0xff983d00: fan
  Node 0xff983eb0: fan


# ofwdump -pr /u3/i2c/fan | less

did not decode much, just name, device_type, compatible, and
version. The rest was hex bytes under titles.

I'll note that there is some risk in using ofwdump:
the CPUs are synchronized/stopped during the
use of OpenFirmware and this can time out and
crash the box. (The so-called Quad-Core did this
while I was looking around.)

Mark Millard
marklmi at
( went
away in early 2018-Mar)

Want to link to this message? Use this URL: <>