On 27.02.2026 11:58, Edwin Török wrote:
> --- a/tools/tests/x86_emulator/test_x86_emulator.c
> +++ b/tools/tests/x86_emulator/test_x86_emulator.c
> @@ -1882,8 +1882,13 @@ int main(int argc, char **argv)
> #define decl_insn(which) extern const unsigned char which[], \
> which##_end[] asm ( ".L" #which "_end" )
> #define put_insn(which, insn) ".pushsection .test\n" \
> - #which ": " insn "\n" \
> + ".ifndef "#which"\n" \
> + #which ": \n" \
> + ".endif\n" \
> + insn "\n" \
> + ".ifndef .L"#which"_end\n" \
> ".L" #which "_end:\n" \
> + ".endif\n" \
> ".popsection"
Nice idea, but why multiple .ifndef, and why emitting the insn even if the
labels are already there (and hence won't be emitted a 2nd time)?
Further, if the compiler unrolls a loop and instantiates such a put_insn()
twice, it could pick different inputs (where flexibility is allowed). Most
present uses (including ones I have pending) meet this requirement (i.e.
permit only a single register per operand), but vmovdqu{32,16}_to_mem,
evex_vcvtph2ps, vpcompress{d,q,w}, vpdpwssd_{vex1,vex2,evex}, and
vmovsh_to_mem don't. {,v}movsd{,_masked}_to_mem, pcmp{e,i}str{i,m} and a
few more could also end up being problematic if different addressing was
used for the memory operand(s).
None of those sit in loops, I think, so we may be safe. But the constraints
need properly writing down in a comment, I think.
> @@ -5248,7 +5253,7 @@ int main(int argc, char **argv)
> memset(res + 3, ~0, 8);
> regs.eax = (unsigned long)res;
> regs.ecx = ~0;
> - for ( i = 0; i < 2; ++i )
> + for (i = 0; i < 2; ++i )
> {
> decl_insn(vmovsh_to_mem);
Excuse me?
Jan
