> My understanding is that the new code, by passing shared memory > through fd is a lot better since [...]
In those respects, yes. But it's worse in that it requires write access to a filesystem - a filesystem which supports mmap - with space enough to hold the shared memory segments, which MIT-SHM doesn't. Tradeoffs, tradeoffs.... > it doesn't rely on file system permissions to control access, Neither does MIT-SHM, I thought. I thought it relied on shared memory segment permissions (which in some respects look and work like filesystem permissions, but actually have nothing to do with any filesystem). What I think this _really_ needs is either a way to pass access to a chunk of memory through sockets a la SCM_RIGHTS, or a way to create a file descriptor which is mmappable but doesn't depend on anything in any filesystem. (I'd prefer the former; after all, "access rights" covers a lot more than file descriptors - and I've wished for it often enough before; it would hardly be specific to this application.) /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML [email protected] / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
