Fix lock order reversal in lagg_get_counter
AbandonedPublic
Actions

Authored by asomers on Oct 20 2016, 10:04 PM.

Tags

None

Referenced Files

	Unknown Object (File)
	Wed, Sep 30, 4:09 AM

	Unknown Object (File)
	Wed, Sep 30, 3:04 AM

	Unknown Object (File)
	Tue, Sep 29, 9:51 PM

	Unknown Object (File)
	Tue, Sep 29, 6:01 PM

	Unknown Object (File)
	Thu, Sep 24, 10:22 PM

	Unknown Object (File)
	Thu, Sep 24, 10:10 PM

	Unknown Object (File)
	Thu, Sep 24, 10:10 PM

	Unknown Object (File)
	Mon, Sep 21, 9:06 AM

Subscribers

Details

Reviewers

melifaro
glebius

Summary

r272211 introduced a LOR in lagg_get_counter by acquiring the lagg softc
lock after previously acquiring the if_addr lock in sysctl_iflist. My
solution is to opportunistically acquire the lagg softc. If it can't be
locked, then return an old copy of the counters. The user will see what
looks like a lull in traffic, but it will "catch up" the next time that
he checks.

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Passed

Unit

No Test Coverage

Build Status

Buildable 5670
Build 5955: CI src build	Jenkins
Build 5954: arc lint + arc unit

Event Timeline

asomers updated this revision to Diff 21564.Oct 20 2016, 10:04 PM

asomers retitled this revision from to Fix lock order reversal in lagg_get_counter.

asomers updated this object.

asomers edited the test plan for this revision. (Show Details)

asomers added reviewers: glebius, melifaro.

asomers added a subscriber: ken.

Herald added a subscriber: imp. · View Herald TranscriptOct 20 2016, 10:04 PM

This is not the correct fix. The bug lies in a higher layer; the stack must never call into the ifnet layer with the if_addr_lock already held. I see how to fix the upper layer, so I'll take that.

Obsolete. This bug was fixed in a better way by 308580 by @rstone.