On Mon, Sep 30, 2013 at 7:07 AM, Rainer Orth <r...@cebitec.uni-bielefeld.de> wrote: > Ian Lance Taylor <i...@google.com> writes: > >> On Mon, Sep 30, 2013 at 6:07 AM, Rainer Orth >> <r...@cebitec.uni-bielefeld.de> wrote: >>> Ian Lance Taylor <i...@google.com> writes: >>> >>>> Following up on my earlier patch, this patch implements the >>>> reflect.MakeFunc function for 386. >>>> >>>> Tom Tromey pointed out to me that the libffi closure support can >>>> probably be used for this. I was not aware of that support. It >>>> supports a lot more processors, and I should probably start using it. >>>> The approach I am using does have a couple of advantages: it's more >>>> efficient, and it doesn't require any type of writable executable >>>> memory. I can get away with that because indirect calls in Go always >>>> pass a closure value. So even when and if I do change to using libffi, >>>> I might still keep this code for amd64 and 386. >>> >>> Unfortunately, this patch (and undoubtedly the corresponding amd64 one) >>> break Solaris/x86 libgo bootstrap with native as: >> >> Unfortunately I think I'll have to somehow disable this functionality >> on systems with assemblers that do not understand the .cfi directives, >> as otherwise calling panic in a function created with MakeFunc will >> not work. > > Alternatively, one could hand-craft the .eh_frame section for such > systems along the lines of libffi/src/x86/sysv.S: ugly, but doable.
Yeah. I'm not going to do that myself. But I would be happy to approve a patch for that if somebody else wants to write it. Ian
