Messages in this thread | | | Subject | Re: [PATCH 7/9] locktorture: Add infrastructure for torturing read locks | From | Davidlohr Bueso <> | Date | Fri, 12 Sep 2014 11:02:27 -0700 |
| |
On Fri, 2014-09-12 at 09:06 -0700, Paul E. McKenney wrote: > On Thu, Sep 11, 2014 at 09:40:41PM -0700, Davidlohr Bueso wrote: > > In addition, introduce a new nreaders_stress module parameter. The > > default number of readers will be the same number of writers threads. > > Writer threads are interleaved with readers. Documentation is updated, > > respectively. > > Nice!!! > > Conditional fairness checks in the future? (As in verifying that if > the rwlock in question claims some degree of fairness, trying to break > that guarantee, and contrariwise, if the lock is unfair, making sure > to avoid starvation during the test?)
Yep, there are all sorts of interesting things we can measure in rw runs. In this set I'm only trying to establish a minimum infrastructure. Future work will be pretty trivial (at least code wise) to implement once this sort of thing is in.
> And one nit below. [...] > C initialization does this already, no need to add the NULL initializers.
Sure, I always tend to be explicit when initializing. I guess you won't like the context initialization in patch 9/9.
Thanks, Davidlohr
| |