Messages in this thread | | | Date | Thu, 7 Aug 2008 11:52:53 -0600 (MDT) | Subject | Re: [ANNOUNCE] mdb-2.6.27-rc2-ia32-08-07-08.patch | From | jmerkey@wolfmoun ... |
| |
> I wrote: >> jmerkey@wolfmountaingroup.com wrote: >>> check mdb-main.c -- I disable preemption before rspin_lock is >>> attempted. > ... >> Then it is recommendable that you document the call context requirements >> at the functions. > > An alternative would of course be to establish the non-preemptible > context from inside the lock accessors, either redundantly to what the > call site is doing, or instead of doing it at the call site if that can > be arranged. > -- > Stefan Richter > -=====-==--- =--- --=== > http://arcgraph.de/sr/ >
It needs to get turned off right away, well before we get to the locks.
Keff
| |