Bug 1275 - Segfault on cleanup
Summary: Segfault on cleanup
Status: RESOLVED FIXED
Alias: None
Product: libsigrok
Classification: Unclassified
Component: Driver: scpi-pps (show other bugs)
Version: unreleased development snapshot
Hardware: All All
: Normal normal
Target Milestone: ---
Assignee: Nobody
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-09-07 16:37 CEST by Martin Ling
Modified: 2020-06-24 23:51 CEST (History)
2 users (show)


Attachments
valgrind output (2.99 KB, text/plain)
2018-09-07 16:37 CEST, Martin Ling 		Details
proposal for fix (2 bytes, patch)
2020-04-11 23:54 CEST, Wolfram Sang 		Details
proposal for fix (1.26 KB, patch)
2020-04-11 23:57 CEST, Wolfram Sang 		Details
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Martin Ling 2018-09-07 16:37:01 CEST 
Created attachment 453 [details]
valgrind output

With a Rigol DP821A connected over USB, running "sigrok-cli -d scpi-pps" with either --scan or --show works, but results in a segmentation fault during shutdown.

From valgrind, looks like a null pointer dereference in the std_cleanup->dev_close path.
Comment 1 Martin Ling 2018-09-11 11:27:16 CEST 
Just an additional note, this doesn't happen after sigrok-cli has run a capture, so it looks like the problem may arise with dev_close() being called when the device hadn't actually been open.
Comment 2 Wolfram Sang 2020-04-11 23:54:59 CEST 
Created attachment 618 [details]
proposal for fix

Martin, I attached a patch (developed independently) which might fix your issue. Could you check?
Comment 3 Wolfram Sang 2020-04-11 23:57:26 CEST 
Created attachment 619 [details]
proposal for fix

Hopefully now correctly uploaded...
Comment 4 Uwe Hermann 2020-06-24 23:51:37 CEST 
Should be fixed in 9b093606545ec3963a03a3cfac61954c62e93e10, thanks!

I've tested a similar device, Rigol DP712. Before the patch I didn't get a segfault but rather an error:

  sr: scpi_serial: Error while sending SCPI command 'SYST:LOCAL': -1.

After the patch that error is gone indeed. It's relatively likely that this was the same issue as bug 1151 as well, yeah.