Re: [PATCH v2] RISC-V: zicond: Fix opt2 pattern

[Jeff Law <jeffreyalaw@gmail.com>]

On 9/1/23 13:53, Vineet Gupta wrote:
> This was tripping up gcc.c-torture/execute/pr60003.c at -O1 since in
> failing case, pattern's asm czero.nez gets both rs2 and rs1 as non zero.
> 
> We start with the following src code snippet:
> 
>        if (a == 0)
> 	return 0;
>        else
> 	return x;
>      }
> 
> which is equivalent to:  "x = (a != 0) ? x : a" where x is NOT 0.
>                                                  ^^^^^^^^^^^^^^^^
> 
> and matches define_insn "*czero.nez.<GPR:mode><X:mode>.opt2"
> 
> | (insn 41 20 38 3 (set (reg/v:DI 136 [ x ])
> |        (if_then_else:DI (ne (reg/v:DI 134 [ a ])
> |                (const_int 0 [0]))
> |            (reg/v:DI 136 [ x ])
> |            (reg/v:DI 134 [ a ]))) {*czero.nez.didi.opt2}
> 
> The corresponding asm pattern generates
>      czero.nez x, x, a   ; %0, %2, %1
> 
> which implies
>      "x = (a != 0) ? 0 : a"
> 
> clearly not what the pattern wants to do.
> 
> Essentially "(a != 0) ? x : a" cannot be expressed with CZERO.nez if X
> is not guaranteed to be 0.
> 
> However this can be fixed with a small tweak
> 
> "x = (a != 0) ? x : a"
> 
>     is same as
> 
> "x = (a == 0) ? a : x" since middle operand is 0 when a == 0.
> 
> which can be expressed with CZERO.eqz
> 
> before fix			after fix
> -----------------		-----------------
> li        a5,1		        li        a5,1
> ld        a4,8(sp)	        ld        a4,8(sp)   # a4 is runtime non zero
> czero.nez a0,a4,a5 # a0=0 NOK 	czero.eqz a0,a4,a5   # a0=a4!=0 OK
> 
> The issue only happens at -O1 as at higher optimization levels, the
> whole conditional move gets optimized away.
> 
> This fixes 4 testsuite failues in a zicond build:
> 
> FAIL: gcc.c-torture/execute/pr60003.c   -O1  execution test
> FAIL: gcc.dg/setjmp-3.c execution test
> FAIL: gcc.dg/torture/stackalign/setjmp-3.c   -O1  execution test
> FAIL: gcc.dg/torture/stackalign/setjmp-3.c   -O1 -fpic execution test
> 
> gcc/ChangeLog:
> 	* config/riscv/zicond.md: Fix op2 pattern.
> 
> Fixes: 1d5bc3285e8a ("[committed][RISC-V] Fix 20010221-1.c with zicond")
> Signed-off-by: Vineet Gupta <vineetg@rivosinc.com>
OK.

Still not sure why I didn't trip over it in my own testing (execute.exp 
runs pr60003.c), but regardless, glad to have it fixed.

jeff