There is a problem that I suspect is pretty widespread in the majority of driver code. The problem is the improper capturing of user mode pointers. I decided to write a blog about it and try to get a feel for if I am right or not. J I figure that if people...

KeMemoryBarrier is a kernel DDK support macro. There is also a WIN32 macro called MemoryBarrier that is implemented identically (there is an observance test hidden here!) - so we will just talk about KeMemoryBarrier here, but everything we say about it...

In a previous post on compiler and processor reordering, I said that for multi-threaded, lock-free code to be totally correct - you need a barrier on the read sides of the code - but that it was pretty complicated and wasn't required on any processors...

So I thought that a good first technical blog entry would be one about a common – but “hardly thought about by most programmers” problem called “reordering”. It is a subtle problem but very important to understand if you write lock-less multithreaded...

I am brand new to the world of blogs so I apologize in advance to any one that reads this blog. Please let me know if I am doing something rude or ignorant. :) "Allow myself to .. introduce ... myself ..." My name is Jonathan Morrison and I work...

Well - you have to do it don't you?