is magicdev evil?

[Tommy McNeely]

On Mon, 06 Oct 2003 17:14:25 -0400, Owen Taylor wrote:

> On Mon, 2003-10-06 at 16:57, Tommy McNeely wrote:
>> [quoted text muted]
> 
> Basically:
> 
>  A) If magicdev is eating 25% of your CPU, something is going wrong,
>     and you shoudl file a bug so we can fix it. It isn't supposed to do
>     that.

  PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME CPU COMMAND
16744 tommy     18   0  5664 5664  4852 S     7.5  0.5   0:01   0 magicdev
16749 tommy     17   0  1152 1152   888 R     0.3  0.1   0:00   0 top
13056 root      15   0  295M  39M  9320 S     0.1  3.9   2:50   0 X
13227 tommy     15   0 16628  16M  8428 S     0.1  1.6   0:11   0 gnome-termina



what about 7.5-11% ? I mean I cant get it to go back to 25% for some
reason :)



> 
>     There was a kernel bug recently in this area, but unless your
>     /var/log/messages is getting filled with lots and lots of junk, that's
>     not it.
> 

I am running severn-beta2 ... wihtout any updates (see my other up2date
message)

Oct  6 12:51:11 cookies /etc/hotplug/usb.agent: Setup usbcore for USB product 5e3/604/12
Oct  6 12:51:11 cookies /etc/hotplug/usb.agent: Setup usbcore for USB product 5e3/604/12
Oct  6 12:51:11 cookies /etc/hotplug/usb.agent: Setup hid for USB product 45e/47/300
Oct  6 12:51:11 cookies /etc/hotplug/usb.agent: Setup mousedev for USB product 45e/47/300
Oct  6 12:51:12 cookies devlabel: devlabel service started/restarted
Oct  6 12:51:12 cookies /etc/hotplug/usb.agent: Setup hid for USB product 430/5/102
Oct  6 12:51:12 cookies /etc/hotplug/usb.agent: Setup keybdev for USB product 430/5/102
Oct  6 12:51:13 cookies devlabel: devlabel service started/restarted
Oct  6 12:51:13 cookies devlabel: devlabel service started/restarted
Oct  6 12:56:59 cookies kernel: usb.c: USB disconnect on device 00:07.4-2 address 10
Oct  6 12:56:59 cookies kernel: usb.c: USB disconnect on device 00:07.4-2.1 address 11
Oct  6 12:56:59 cookies kernel: usb.c: USB disconnect on device 00:07.4-2.2 address 12
Oct  6 12:57:07 cookies devlabel: devlabel service started/restarted
Oct  6 12:57:07 cookies last message repeated 2 times
Oct  6 13:45:27 cookies smartd[4117]: Device: /dev/sda, Temperature changed -1 degrees to 31 degrees since last reading
Oct  6 14:28:56 cookies kernel: hub.c: new USB device 00:07.4-2, assigned address 13
Oct  6 14:28:56 cookies kernel: hub.c: USB hub found
Oct  6 14:28:56 cookies kernel: hub.c: 4 ports detected
Oct  6 14:28:57 cookies kernel: hub.c: new USB device 00:07.4-2.1, assigned address 14
Oct  6 14:28:57 cookies kernel: input: USB HID v1.10 Mouse [Microsoft Microsoft 5-Button Mouse with IntelliEye(TM)] on usb1:14.0
Oct  6 14:28:57 cookies kernel: hub.c: new USB device 00:07.4-2.2, assigned address 15
Oct  6 14:28:57 cookies kernel: input: USB HID v1.00 Keyboard [0430:0005] on usb1:15.0
Oct  6 14:28:59 cookies /etc/hotplug/usb.agent: Setup usbcore for USB product 5e3/604/12
Oct  6 14:28:59 cookies /etc/hotplug/usb.agent: Setup usbcore for USB product 5e3/604/12
Oct  6 14:29:00 cookies /etc/hotplug/usb.agent: Setup hid for USB product 45e/47/300
Oct  6 14:29:00 cookies /etc/hotplug/usb.agent: Setup mousedev for USB product 45e/47/300
Oct  6 14:29:00 cookies /etc/hotplug/usb.agent: Setup hid for USB product 430/5/102
Oct  6 14:29:00 cookies /etc/hotplug/usb.agent: Setup keybdev for USB product 430/5/102
Oct  6 14:29:00 cookies devlabel: devlabel service started/restarted
Oct  6 14:29:02 cookies last message repeated 2 times
Oct  6 14:29:18 cookies modprobe: modprobe: Can't locate module sound-slot-1
Oct  6 14:29:18 cookies modprobe: modprobe: Can't locate module sound-service-1-0
Oct  6 14:29:18 cookies modprobe: modprobe: Can't locate module sound-slot-1
Oct  6 14:29:18 cookies modprobe: modprobe: Can't locate module sound-service-1-0
Oct  6 14:46:04 cookies smartd[4117]: Device: /dev/sda, Temperature changed 1 degrees to 32 degrees since last reading
Oct  6 15:16:22 cookies smartd[4117]: Device: /dev/sda, Temperature changed -1 degrees to 31 degrees since last reading
Oct  6 15:34:36 cookies userhelper: pam_timestamp: timestamp file `/var/run/sudo/tommy/1:root' is only 14 seconds old, allowing access to up2date for UID 123
Oct  6 15:34:38 cookies userhelper: pam_timestamp: timestamp file `/var/run/sudo/tommy/1:root' is only 2 seconds old, allowing access to up2date for UID 123



>     You should file a bug with a) any messages that are appearing in
>     /var/log/messages b) a short snippet of the strace of magicdev
>  

waiting to see if 7-11% is still bad enough for a bug



>      killall magicdev
>      strace -o /tmp/magicdev.log magicdev
>      <wait 20 seconds>
>      <control-c>
>     
>     c) The contents of /proc/ide/hdc/model (modify as appropriate if your
>     cd-rom is elsewhere.)
>     
>  B) No, CDROMS do not reliably send a signal when they are opened
>     and closed. There are things in the spec for that, they are not
>     implemented uniformly.

its nice to have standards :)

-Tommy


> 
>     There may be better things to do than what magicdev does. They
>     probably would require full-time work for about a month for a pair of
>     a test engineer and a developer to get going and make reliable.
> 
>     (Basically, you have to have a big list of hardware with information
>     on what works where.)
> 
> Regards,
> 					Owen