mikhail.ramalho added inline comments.
================
Comment at: libc/src/setjmp/riscv64/longjmp.cpp:54-55
+
+ LIBC_INLINE_ASM("seqz %0, %1" : "+r"(buf) : "r"(val) :);
+ LIBC_INLINE_ASM("add %0, %0, %1" : "+r"(buf) : "r"(val), "r"(buf) :);
+}
----------------
sivachandra wrote:
> Your comment is in the previous diff but thanks for the explanation. I think
> we have missed the `val` check for zero in the x86_64 case and should be
> fixed separately.
>
> For the above two instructions, in the interest of reducing the amount of
> logic in inline assembly, can we do:
>
> ```
> val = val == 0 ? 1 : val;
> LIBC_INLINE_ASM("add a0, %0, zero\n\t" : : "r"(val) :);
> ```
> Your comment is in the previous diff but thanks for the explanation. I think
> we have missed the `val` check for zero in the x86_64 case and should be
> fixed separately.
no problem, do you want me to fix the x86_64 version?
>
> For the above two instructions, in the interest of reducing the amount of
> logic in inline assembly, can we do:
>
> ```
> val = val == 0 ? 1 : val;
> LIBC_INLINE_ASM("add a0, %0, zero\n\t" : : "r"(val) :);
> ```
Done, I'll rerun the tests now.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D145584/new/
https://reviews.llvm.org/D145584
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
