Out of curiosity, when running "systemctl status usrp-hwd" I get a message 
stating "vendor preset=enabled". Is there anyway possible that mpm is still 
being started at bootup? 

    On Tuesday, October 12, 2021, 01:55:53 PM EDT, Tellrell White via 
USRP-users <usrp-users@lists.ettus.com> wrote:  
 
  I/O Expander(TCA6408A) on the daughterboard.

    On Tuesday, October 12, 2021, 01:50:19 PM EDT, Marcus D Leech 
<patchvonbr...@gmail.com> wrote:  
 
 Which i2C device are you trying to access?

Sent from my iPhone

On Oct 12, 2021, at 12:36 PM, Tellrell White <t_whit...@yahoo.com> wrote:



 
Philip I'm highlty doubtful that its a permissions issue. After moving the app 
over to the n310, I enable it for all users to run it permissions wise. 
However, the only "user" is root on the n310. 

Martin I followed your instruction to disable usrp-hwd and then reboot and 
after rebooting and running "systemctl status usrp-hwd" i get 
"Active:inactive(dead)" which leads me to believe that upon rebooting the MPM 
isn't running however, when re-running my app I get the same error. Is it 
anyway that MPM may still be holding these resources or is there something else 
going on here? 

    On Thursday, October 7, 2021, 03:02:09 AM EDT, Martin Braun 
<martin.br...@ettus.com> wrote:  
 
 Just to make sure we're talking abou the same things: To disable MPM, you run
    systemctl stop usrp-hwd
You can also run
    systemctl disable usrp-hwd
which does not stop MPM, but will avoid it being started on reboot.
It seems you know what you're doing, but for the record: when MPM is not 
running, your USRP is a heavy piece of metal running Linux (at least, until 
you've ported OpenCPI or some other custom software). That's why we have 
designed MPM to keep running at all times, even when some kind of fatal error 
occurred, so UHD can at least see some kind of device. We use systemd watchdogs 
to reboot it when it's gone.

--M

On Wed, Oct 6, 2021 at 11:45 PM Marcus D. Leech <patchvonbr...@gmail.com> wrote:

  On 2021-10-06 5:17 p.m., Tellrell White wrote:
  
  
  So, just to clarify, if its "required" is there no way of disabling it or 
removing?? 
  
   The project I'm working on is based on using the Open CPI framework, which 
is basically an opensource framework for developing and executing component 
based apps on embedded platforms. With that being said, my task isn't tightly 
bound to using UHD, therefore, I would like to disable or remove MPM for a 
custom application.  
    
   
 After killing MPM, you can use "ps" to check to see if it has come back, and 
if it hasn't and "something else" still owns the I2C devices, you can use
   "lsof" to see which processes have which devices open.  I think the i2C 
devices will manifest in the /dev filesystem.  For example, on my *laptop*, 
there's:
 
 rw------- 1 root root 89, 0 Sep 28 18:14 /dev/i2c-0
 crw------- 1 root root 89, 1 Sep 28 18:14 /dev/i2c-1
 crw------- 1 root root 89, 2 Sep 28 18:14 /dev/i2c-2
 crw------- 1 root root 89, 3 Sep 28 18:14 /dev/i2c-3
 crw------- 1 root root 89, 4 Sep 28 18:14 /dev/i2c-4
 crw------- 1 root root 89, 5 Sep 28 18:14 /dev/i2c-5
 
 
 
      
 
 
       
 
 _______________________________________________
USRP-users mailing list -- usrp-users@lists.ettus.com
To unsubscribe send an email to usrp-users-le...@lists.ettus.com

_______________________________________________
USRP-users mailing list -- usrp-users@lists.ettus.com
To unsubscribe send an email to usrp-users-le...@lists.ettus.com
  
  _______________________________________________
USRP-users mailing list -- usrp-users@lists.ettus.com
To unsubscribe send an email to usrp-users-le...@lists.ettus.com
  
_______________________________________________
USRP-users mailing list -- usrp-users@lists.ettus.com
To unsubscribe send an email to usrp-users-le...@lists.ettus.com

Reply via email to