This series fixes the race condition we have on startup. Patch 3/4 describes the exact sequence so I'll skip this here. In short, the problem is that because we share the fd between several tools, we may get events for tools that haven't been initialised yet. This can cause NULL-pointer dereferences.
I'd prefer if someone could test this on server 1.7 - 1.9, because I'm not sure if the UnInit still works now. On 1.10 we would crash because patch 2/4 changes the device order but it doesn't matter since we can leave the UnInit to the server now anyway. So a simple ifdef in patch 1/4 fixes that crash without negative side-effects. Testing if a race-condition has disappeared is near-impossible (how do you prove that something does not exist? :) so I really recommend doing the gdb trick described in 3/4 to trigger the crash before and after applying the patchset. Cheers, Peter ------------------------------------------------------------------------------ Free Software Download: Index, Search & Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev _______________________________________________ Linuxwacom-devel mailing list Linuxwacom-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel
