Re: [Qemu-devel] [PATCH for-2.0? 2/2] cpu-exec: Unlock tb_lock if we longjmp out of code generation

2014-04-03 Thread Andrei E. Warkentin
Hiya, Cool. Definitely more compact and less intrusive, and definitely should catch more issues than the original page->flags check. The only possible cost is maintenance and debugging (implicit state and all that)... so... How about adding a comment around the "if (have_tb_lock)" to explain how w

[Qemu-devel] Signal handling bugs (and proposed fixes)

2014-04-02 Thread Andrei E. Warkentin
Hiya, I found a couple of corner cases where signal handling fails in QEMU linux-user support. "Signal handling" here being just a symptom - actual problems are in TB / page management. Here are a couple of simple tests (https://github.com/andreiw/andreiw-wip/tree/master/qemu/tests). The test: 1)