qemu-riscv
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-riscv] [PATCH] target/riscv: Zero extend the inputs of divuw a


From: Kito Cheng
Subject: Re: [Qemu-riscv] [PATCH] target/riscv: Zero extend the inputs of divuw and remuw
Date: Thu, 21 Mar 2019 23:40:55 +0800

Verified with gcc testsuite on rv64gc, no new regression introduced, and get less fails.

Palmer Dabbelt <address@hidden>於 2019年3月21日 週四,22:59寫道:
While running the GCC test suite against 4.0.0-rc0, Kito found a
regression introduced by the decodetree conversion that caused divuw and
remuw to sign-extend their inputs.  The ISA manual says they are
supposed to be zero extended:

    DIVW and DIVUW instructions are only valid for RV64, and divide the
    lower 32 bits of rs1 by the lower 32 bits of rs2, treating them as
    signed and unsigned integers respectively, placing the 32-bit
    quotient in rd, sign-extended to 64 bits. REMW and REMUW
    instructions are only valid for RV64, and provide the corresponding
    signed and unsigned remainder operations respectively.  Both REMW
    and REMUW always sign-extend the 32-bit result to 64 bits, including
    on a divide by zero.

Here's Kito's reduced test case from the GCC test suite

    unsigned calc_mp(unsigned mod)
    {
         unsigned a,b,c;
         c=-1;
         a=c/mod;
         b=0-a*mod;
         if (b > mod) { a += 1; b-=mod; }
         return b;
    }

    int main(int argc, char *argv[])
    {
         unsigned x = 1234;
         unsigned y = calc_mp(x);

         if ((sizeof (y) == 4 && y != 680)
      || (sizeof (y) == 2 && y != 134))
    abort ();
         exit (0);
    }

I haven't done any other testing on this, but it does fix the test case.

Signed-off-by: Palmer Dabbelt <address@hidden>
---
 target/riscv/insn_trans/trans_rvm.inc.c |  4 ++--
 target/riscv/translate.c                | 21 +++++++++++++++++++++
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/target/riscv/insn_trans/trans_rvm.inc.c b/target/riscv/insn_trans/trans_rvm.inc.c
index 204af225f8f3..47cd6edc72a1 100644
--- a/target/riscv/insn_trans/trans_rvm.inc.c
+++ b/target/riscv/insn_trans/trans_rvm.inc.c
@@ -103,7 +103,7 @@ static bool trans_divw(DisasContext *ctx, arg_divw *a)
 static bool trans_divuw(DisasContext *ctx, arg_divuw *a)
 {
     REQUIRE_EXT(ctx, RVM);
-    return gen_arith_div_w(ctx, a, &gen_divu);
+    return gen_arith_div_uw(ctx, a, &gen_divu);
 }

 static bool trans_remw(DisasContext *ctx, arg_remw *a)
@@ -115,6 +115,6 @@ static bool trans_remw(DisasContext *ctx, arg_remw *a)
 static bool trans_remuw(DisasContext *ctx, arg_remuw *a)
 {
     REQUIRE_EXT(ctx, RVM);
-    return gen_arith_div_w(ctx, a, &gen_remu);
+    return gen_arith_div_uw(ctx, a, &gen_remu);
 }
 #endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 049fa65c6611..dd763647ea55 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -600,6 +600,27 @@ static bool gen_arith_div_w(DisasContext *ctx, arg_r *a,
     return true;
 }

+static bool gen_arith_div_uw(DisasContext *ctx, arg_r *a,
+                            void(*func)(TCGv, TCGv, TCGv))
+{
+    TCGv source1, source2;
+    source1 = tcg_temp_new();
+    source2 = tcg_temp_new();
+
+    gen_get_gpr(source1, a->rs1);
+    gen_get_gpr(source2, a->rs2);
+    tcg_gen_ext32u_tl(source1, source1);
+    tcg_gen_ext32u_tl(source2, source2);
+
+    (*func)(source1, source1, source2);
+
+    tcg_gen_ext32s_tl(source1, source1);
+    gen_set_gpr(a->rd, source1);
+    tcg_temp_free(source1);
+    tcg_temp_free(source2);
+    return true;
+}
+
 #endif

 static bool gen_arith(DisasContext *ctx, arg_r *a,
--
2.19.2


reply via email to

[Prev in Thread] Current Thread [Next in Thread]