I am working on an app that controls a special hardware card. The card is accessed across the PCI bus and it has several configuration registers, etc., that I need to control. I would like to keep most of the smarts in the Java code (resource usage, etc. coded in the Java classes that model the hardware). I plan to use native threads. We will have a driver (Linux module) installed for our special hardware, that can give access to read/write and other special funcs (via ioctl). I have 2 questions regarding this impl. 1) If I use native threads, does each thread need to perform an "open" before it can call ioctl, or can I cache the fd in the native code? (different processes can't share fd's, can they?) 2) What about mmap? Does each thread need is own call? any other helpful hints or suggestions would be welcome. thanks, [EMAIL PROTECTED] ---------------------------------------------------------------------- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
