Zc extension is enabled, the behavior of the .option

[bigmagic123]

I want to restrict the generation of Zc in a segment of assembly code using the .option.

When I write code like this：

.option push
.option arch, -zca
.option arch, -zcb
    add x3,x3,x27
.option pop

Passing the argument "arch=rv32ima_zca_zcb" during compilation.

zca remains enabled.

When I browse through the code, I notice that:

https://github.com/bminor/binutils-gdb/blob/master/bfd/elfxx-riscv.c

  {"zcf", "zca",	check_implicit_always},
  {"zcd", "zca",	check_implicit_always},
  {"zcb", "zca",	check_implicit_always},

When I enable zcb, zca will be included by default.

why it was designed this way?What would be the appropriate usage?

[jnk0le]

Zca is basically C extension without floating point load/stores

[KaiYG]

Based on the spec, all Zc* depend on Zca.
But aren't Zcb & Zca re-enabled by the .option pop?

[tovine]

I want to restrict the generation of Zc in a segment of assembly code using the .option.

When I write code like this：

.option push
.option arch, -zca
.option arch, -zcb
    add x3,x3,x27
.option pop

Passing the argument "arch=rv32ima_zca_zcb" during compilation.

zca remains enabled.

Maybe the issue is that you're trying to disable zca before disabling zcb so the tools don't allow you to do this because zca is a dependency of zcb (which is still enabled at that point)?
Does it work if you reverse the order and first disable zcb and then zca?