|
From: Julian S. <js...@ac...> - 2012-06-19 17:04:27
|
Hi Philippe, I just wanted to follow up on a couple of things about this. ------ Bart said ... It looks like a good idea to me to read more about existing multithreaded dynamic analysis tools first. An example: Paul Sack e.a., Accurate and efficient filtering for the Intel thread checker race detector, ASID '06 Proceedings of the 1st workshop on Architectural and system support for improving software dependability, Pages 34 - 41, 2006. I had a look at this. Do you have the paper? If not I can send it. Anyway, it seems to me that this filtering is specific to race detection and we cannot use the techniques in Memcheck. Maybe you got something different from the paper? ------- IIUC, the tt-xor idea got moved to /dev/null and we need a new plan, is that correct? Does Kim Hazlewood's PhD thesis have anything useful to say? ------- I didn't study your message re Memcheck and locking experiments yet, but will. In particular you mentioned something about find grained locking at the per-SecMap level (iirc), and this was something I was also wondering about. J |