On Fri, Jul 7, 2017 at 6:47 AM, Luc Van Oostenryck <luc.vanoostenryck@xxxxxxxxx> wrote: > > We have to do something more (or something less :)) otherwise > the refcount can only be used to detect a problem. In other words, > having a situation like "OK, I can't safely do XYZ here" is maybe > better than the current situation but doesn't help when you need > to do XYZ. Right. The real implementation will replace the die with some code to handle those situation. I am also considering a debug version of sparse which performance those extra test at a performance penalty. Just generate two executable. One for normal production use. The other one for developer to find out potential issue. Some of the code have to be compile differently(ptrlist), it is not easy to implement as turn on by a command line options. Chris -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
