Bug 2776 – MacLow::StartTransmission ends up in segmentation fault

Bug 2776 - MacLow::StartTransmission ends up in segmentation fault


Summary:	MacLow::StartTransmission ends up in segmentation fault

Status:	RESOLVED FIXED

Product:	ns-3
Classification:	Unclassified
Component:	wifi
Version:	ns-3-dev
Hardware:	All All

Importance:	P2 major
Assigned To:	sebastien.deronne

URL:
Keywords:

Depends on:
Blocks:
	Show dependency tree / graph

Reported:	2017-08-21 04:58 UTC by Akin Soysal
Modified:	2017-08-23 16:48 UTC (History)
CC List:	1 user (show)

See Also:

Attachments
proposed mac_low_fix. (827 bytes, patch) 2017-08-21 04:58 UTC, Akin Soysal	Details \| Diff
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 Akin Soysal 2017-08-21 04:58:29 UTC

Created attachment 2905 [details]
proposed mac_low_fix.

Hello,

I have discovered and fixed a nasty bug. It came up in changeset 12571.

When MacLow::StartTransmission is called, first m_currentDca = dca; assignment is done and then CancelAllEvents (); is called. However, if there is any active timer exists, CancelAllEvents() empties the member m_currentDca inside. Since the assignment of dca is done before CancelAllEvents(), new dca is also emptied and whenever MacLow() calls its member m_currentDca, segmentation fault is thrown. Bug can be fixed by moving the assignment after the CancelAllEvents() function call.

I am not sure about the frequency of this bug but its severity is high and I have encountered it when I have applied our new fading channel.

After this fix, our curves seemed a lot smoother, it is also possible that it does not cause segmentation fault but also other problems, so I think this should be fixed ASAP!

Bugfix is in the attachment.

Regards,
Akın

Comment 1 sebastien.deronne 2017-08-23 16:48:52 UTC

Thanks Akin for the fix.
Pushed in changeset 13050:969140f099bb.