On 8 July 2014 17:13, Al Viro <v...@zeniv.linux.org.uk> wrote: > On Tue, Jul 08, 2014 at 09:05:10AM +0100, Peter Maydell wrote: > >> The code we have currently may well be buggy, but the correct > > It is ;-/ We set TARGET_FPE_FLTINV unconditionally there. BTW, what's > the reason why all these cpu_loop() instances can't go into > linux-user/<arch>/something?
It's just ancient code nobody's cleaned up yet. I do have "move all this stuff into arch directories" on my "would like to do" list, but I just haven't got round to it yet, since it's not actually actively broken (unlike many other areas of our codebase :-/). > BTW, are there any more or less uptodate docs on qemu profiling? I mean, > things like perf/oprofile on the host obviously end up lumping all tcg > output together. Is there any way to get information beyond "~40% of time > is spent in generated code, ~15% - in tb_find_fast(), and the rest is very > much colder"? Alex, did you say you'd done something with profiling recently? > Incidentally, combination of --enable-gprof and (default) --enable-pie > won't build - it dies with ld(1) complaining about relocs in gcrt1.o. This sounds like a toolchain bug to me :-) -- PMM