From: amc96@nospicedham.cam.ac.uk   
      
   On 29/08/2017 16:07, Alex wrote:   
   > On 64 bit Windows, stack alignment on a 16 byte boundary is required   
   > before calling all except a leaf function. In the called function, the   
   > stack is 8 mod 16.   
   >   
   > Now, I'm struggling to come up with a way of doing it beyond this code   
   > (which I didn't invent, but I can't for the life of me remember where I   
   > found it.)   
   >   
   >   push   rsp   
   >   push   [rsp]   
   >   and    spl $F0   
   >   call   funkychicken   
   >   pop    rsp   
   >   
   > It seems to be the only way of doing this without branches, flags or   
   > other expensive nonsense. But, as ever, there may be a better way. Any   
   > suggestions?   
      
   Why is the stack 8 mod 16?  That is the bug in this scenario.   
      
   If every function sets up an ABI compatible stack for its callees, all a   
   callee needs to do is ensure it pushes/adjusts the stack pointer by an   
   even number of words.   
      
   I see from other replies that you are doing this in some Forth situation   
   with multiple stacks, but at any point that you have the above scenario,   
   a higher caller has screwed up.  Things will definitely go wrong when   
   you call into a C library, but also with any signal handler which   
   intends to use the red zone.   
      
   Your above code will function correctly, but has a performance hit,   
   because a direct write to the stack pointer interrupts stack-engine   
   optimisations in the pipeline for adjacent pushes/pops/calls/rets.   
   Also, writes to 8-bit registers suffer a merge penalty back into the   
   register file.  `and $~0xf, %rsp` would be more efficient; It encodes in   
   the same number of bytes, but doesn't suffer from merging.   
      
   As a minor note, you should use `leave` rather than pop %rsp, as it   
   takes less instruction bandwidth to execute.   
      
   ~Andrew   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)