Bug 383959 - memory leak when connecting to chip
Summary: memory leak when connecting to chip
Status: RESOLVED FIXED
Alias: None
Product: plasma-nm
Classification: Plasma
Component: applet (show other bugs)
Version: 5.10.4
Platform: Other Linux
: NOR normal
Target Milestone: ---
Assignee: Jan Grulich
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-08-24 14:59 UTC by Marcus Furlong
Modified: 2020-07-14 22:04 UTC (History)
1 user (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Marcus Furlong 2017-08-24 14:59:11 UTC
When connecting a chip board via USB (similar to RaspPI -> https://getchip.com/), plasma-nm continues to try connecting to it, and gets stuck on acquiring network address.

It never gets a network address (and I don't want it to, I just want to power it, or e.g. flash it) , however the plasma-nm applet becomes non-responsive. In this case, plasmashell is consuming all the memory and eventually all the swap on the machine, and eventually plasmashell gets killed by the OOM.

Each time I plug in one of these devices it shows up as a new wired connection, so I cannot get plasma-nm to remember the device settings and disable the connection.

The USB dmesg output is as follows:

[121719.097187] usb 3-7.2: new high-speed USB device number 13 using xhci_hcd
[121719.185904] usb 3-7.2: New USB device found, idVendor=0525, idProduct=a4aa
[121719.185907] usb 3-7.2: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[121719.185907] usb 3-7.2: Product: CDC Composite Gadget
[121719.185908] usb 3-7.2: Manufacturer: Linux 4.4.13-ntc-mlc with musb-hdrc
[121719.196430] cdc_ether 3-7.2:1.0 usb0: register 'cdc_ether' at usb-0000:00:14.0-7.2, CDC Ethernet Device, fa:f6:32:41:11:a5
[121719.196731] cdc_acm 3-7.2:1.2: ttyACM0: USB ACM device
Comment 1 Jan Grulich 2017-10-23 11:08:09 UTC
Does it work if you manually create a wired connection and restrict it to the MAC address of your chip board device?
Comment 2 Jan Grulich 2020-07-14 20:33:16 UTC
Is this still a problem with latest Plasma 5.19?
Comment 3 Marcus Furlong 2020-07-14 22:04:51 UTC
No, not a problem any more. Seems like the latest plasma deactivates the connection after a while. Closing.