Hi group, I've been browsing the target code and noticed that a thread (running transport_processing_thread) is spawned for each target device. Looking at the code, it seems that this thread is a major part of the I/O path as it processes write operations. I am afraid that such an approach will be bad on performance when the number of devices begins to grow (Consider for example 100 logical units). My question is why not to use a simple work queue which will create a thread per logical CPU? The only reason my suggested approach won't really scale is if this thread is doing blocking operations. Is that the case? Many thanks, Eran. -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
