Sender: schwab@LS5.informatik.uni-dortmund.de To: Roman Hodek Cc: linux-m68k@lists.linux-m68k.org Subject: Re: SCC_ACCESS_INIT macro problem (was L68K: SCC bug with oops) References: <199805290850.KAA09174@faui22c.informatik.uni-erlangen.de> X-Yow: I want to TAKE IT HOME and DRESS IT UP in HOT PANTS!! From: Andreas Schwab Date: 02 Jun 1998 10:33:44 +0200 Roman Hodek writes: |> The asm is wrong here. The and insn overwrites the previous value of |> %d0, but %d0 is not listed in the outputs, so gcc doesn't know this. |> My fault... I've written that macro :-) The fix is to copy the value |> of _SCC_p to some tmp variable and declare that it in the outputs so |> gcc doesn't reuse the value. That should teach you that such micro optimizations are not worth the effort. Rewriting that in C gives just as good code, but is much better maintainable. Andreas. ---------------------------------------------------------------------- --- linux/drivers/char/atari_SCC.h.~2~ Sat May 30 12:21:59 1998 +++ linux/drivers/char/atari_SCC.h Sat May 30 12:49:07 1998 @@ -589,20 +589,11 @@ return rv; } -#define SCC_ACCESS_INIT(info) \ - volatile struct PARTIAL_SCC *_SCC_p = \ - (volatile struct PARTIAL_SCC *)info->port; \ - unsigned char *_SCC_shadow = \ - ({ unsigned char *_rv = (unsigned char *)SCC_shadow; \ - volatile struct PARTIAL_SCC *_tmp = _SCC_p; \ - /* gcc generates suboptimal code for this, so I use an asm */ \ - __asm__ ( "and.l #4,%1; lea %0@(%1:l:4),%0" \ - /* alternative: lea %za0@(%2,%1:l:4),%0 \ - * what's faster? */ \ - : "=&a" (_rv), "=&d" (_tmp) \ - : "1" (_tmp), "0" (_rv) ); \ - _rv; \ - }) +#define SCC_ACCESS_INIT(info) \ + volatile struct PARTIAL_SCC *_SCC_p = \ + (volatile struct PARTIAL_SCC *)info->port; \ + unsigned char *_SCC_shadow = \ + &SCC_shadow[(info->port >> 2) & 1][0] #define SCCwrite(reg,val) _SCCwrite(_SCC_p,_SCC_shadow,scc_del,(reg),(val),1) #define SCCwrite_NB(reg,val) _SCCwrite(_SCC_p,_SCC_shadow,scc_del,(reg),(val),0)