hard hang reading CPU temperature from /sys

[andrum99]

I have a python script that retrieves the Pi's CPU temperature from /sys/class/thermal/thermal_zone0/temp once a minute. I've seen a couple of hard hangs when reading this temperature. Here is an example from /var/log/messages:

Code: Select all

Jan 22 19:48:18 raspi2 temp_py2.py: Starting up
Jan 22 20:38:19 raspi2 temp_py2.py: Temperatures old#012Jan 22 20:38:19
Jan 22 20:39:44 raspi2 kernel: [ 3120.818067] python          D c044ab4c     0  2407      1 0x00000000
Jan 22 20:39:44 raspi2 kernel: [ 3120.818144] [<c044ab4c>] (__schedule+0x2a0/0x5ac) from [<c044903c>] (schedule_timeout+0x18c/0x21c)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818175] [<c044903c>] (schedule_timeout+0x18c/0x21c) from [<c044a574>] (__down+0x7c/0xb8)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818213] [<c044a574>] (__down+0x7c/0xb8) from [<c0040ea4>] (down+0x80/0x84)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818241] [<c0040ea4>] (down+0x80/0x84) from [<c001b4c0>] (dev_mbox_read+0x48/0x74)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818269] [<c001b4c0>] (dev_mbox_read+0x48/0x74) from [<c001b620>] (bcm_mailbox_property+0xa0/0x194)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818306] [<c001b620>] (bcm_mailbox_property+0xa0/0x194) from [<c033e548>] (bcm2835_get_temp_or_max.isra.0+0x50/0xc0)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818336] [<c033e548>] (bcm2835_get_temp_or_max.isra.0+0x50/0xc0) from [<c033be98>] (thermal_zone_get_temp+0x38/0x54)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818361] [<c033be98>] (thermal_zone_get_temp+0x38/0x54) from [<c033c010>] (temp_show+0x18/0x3c)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818405] [<c033c010>] (temp_show+0x18/0x3c) from [<c02aa728>] (dev_attr_show+0x1c/0x48)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818482] [<c02aa728>] (dev_attr_show+0x1c/0x48) from [<c014d110>] (sysfs_read_file+0x90/0x134)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818528] [<c014d110>] (sysfs_read_file+0x90/0x134) from [<c00e6bac>] (vfs_read+0x98/0x178)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818589] [<c00e6bac>] (vfs_read+0x98/0x178) from [<c00e7378>] (SyS_read+0x3c/0x78)
Jan 22 20:39:44 raspi2 kernel: [ 3120.818631] [<c00e7378>] (SyS_read+0x3c/0x78) from [<c000dec0>] (ret_fast_syscall+0x0/0x30)

When this hang occurs the python process cannot be killed, even using kill -9. Here is the relevant section of code from my script:

Code: Select all

file = open("/sys/class/thermal/thermal_zone0/temp", "r")
string = file.readline()
file.close()

Is this a known problem? Is there anything I can do about this?

[DougieLawson]

Here's my version

Code: Select all

#!/usr/bin/python
import RPi.GPIO as GPIO
GPIO.setmode(GPIO.BCM)
GPIO.setup(17, GPIO.OUT)

while True:
  try:
    tFile = open('/sys/class/thermal/thermal_zone0/temp')
    temp = float(tFile.read())
    tempC = temp/1000
    if tempC > 43.5:
      GPIO.output(17, 1)
      print "HOT"
    else:
      GPIO.output(17, 0)
      print "COLD"

  except:
    tFile.close()
    GPIO.cleanup()
    exit

I don't close the file each time round the loop. The except clause is only to catch a [CTRL]+[C] to terminate the program.

[andrum99]

That does seem more efficient. I put the code to read the temperature from /sys into a separate function to make the code a bit more readable - I don't really want to keep the file open between reads as that would complicate the code. Strangely, both times this problem occurred the script got the same process ID - it runs at startup from /etc/rc.local. I also had the whole Pi crash in the middle of the night last night - don't know why. Hopefully these are just glitches and there is no bug or anything like that. I think if there was someone would have come across it and it would probably have been fixed.

[andrum99]

I've tweaked my script to sleep for 0.1 seconds each time it goes round the button polling loop. So far it has not hung, so fingers crossed.