Re: [LAD] a *simple* ring buffer, comments pls?

Guys, if you're writing code for your own use, and don't care about portability or security, go ahead and use volatile as a synchronization primitive. But if the code is going to be accepted into an open source project that gets wide use, volatile is a bad idea. But don't take my word for it; here's what kernel.org, cert.org, and an intel researcher say about it: "Volatile considered harmful", http://www.kernel.org/doc/Documentation/volatile-considered-harmful.txt "The use of volatile is likely to be seen as a bug and will bring additional scrutiny to the code." "CERT Secure Coding Standards, recommendation POS03: Do not use volatile as a synchronization primitive" https://www.securecoding.cert.org/confluence/display/seccode/POS03-C.+Do+no… "A variable declared as volatile is not cached in a register, leading to this confusion that it can be used safely as a synchronization primitive. When declared volatile, the compiler does not re-order the sequence of reads and writes to that memory location. However, the compiler might re-order these reads and writes with those to other memory locations. This might result in non-atomic operations on the synchronization variable resulting in errors." "Ad Hoc Synchronization Considered Harmful", www.xiongww.com/papers/osdi10-xiong.pdf "Ad hoc synchronizations are error-prone. Significant percentages (22-67%) of these ad hoc synchronizations introduced bugs or severe performance issues." Volatile is a 1980's solution to a 2000's problem, and it just doesn't cut the mustard anymore. - Dan