----- On May 13, 2020, at 8:12 PM, Thomas Gleixner t...@linutronix.de wrote: [...] > >>> Mathieu Desnoyers <mathieu.desnoy...@efficios.com> wrote: > >>> Also, color me confused: is "do_signal()" actually running any user-space, >>> or just setting up the user-space stack for eventual return to signal >>> handler ? > > I'm surprised that you can't answer that question yourself. How did you > ever make rseq work and how did rseq_signal_deliver() end up in > setup_rt_frame()? > > Hint: Tracing might answer that question > > And to cut it short: > > Exit to user space happnes only through ONE channel, i.e. leaving > prepare_exit_to usermode(). >
[...] Yes, I'm very well aware of this. But the patch commit message states: "Make sure task_work runs before any kind of userspace -- very much including signals -- is invoked." which seems to imply that "userspace" can be "invoked" before the task_work runs. Which makes no sense whatsoever. Hence my confused state. >>> Also, it might be OK, but we're changing the order of two things which >>> have effects on each other: restartable sequences abort fixup for preemption >>> and do_signal(), which also have effects on rseq abort. >>> >>> Because those two will cause the abort to trigger, I suspect changing >>> the order might be OK, but we really need to think this through. > > That's a purely academic problem. The order is completely > irrelevant. You have to handle any order anyway: Yes indeed, whether either a signal handler frame fixup or return IP fixup fires first (clearing the rseq_cs pointer in the process) is irrelevant, because they will have the effect on the user-space program's flow. And as you say, given it is run in a loop and can be preempted, any order can happen here, so we have to be prepared for it. This loop has caused me tons of headaches when stress-testing on NUMA machines by the way. > That said, even for the case Andy and Peter were looking at (MCE) the > ordering is completely irrelevant. Not sure about that, see Andy's follow up. Thanks, Mathieu -- Mathieu Desnoyers EfficiOS Inc. http://www.efficios.com