New review comment by ericonr on void-packages repository

https://github.com/void-linux/void-packages/pull/24528#discussion_r555772264

Comment:
From the generic glibc version:

```
    case _SC_LEVEL1_ICACHE_SIZE:
    case _SC_LEVEL1_ICACHE_ASSOC:
    case _SC_LEVEL1_ICACHE_LINESIZE:
    case _SC_LEVEL1_DCACHE_SIZE:
    case _SC_LEVEL1_DCACHE_ASSOC:
    case _SC_LEVEL1_DCACHE_LINESIZE:
    case _SC_LEVEL2_CACHE_SIZE:
    case _SC_LEVEL2_CACHE_ASSOC:
    case _SC_LEVEL2_CACHE_LINESIZE:
    case _SC_LEVEL3_CACHE_SIZE:
    case _SC_LEVEL3_CACHE_ASSOC:
    case _SC_LEVEL3_CACHE_LINESIZE:
    case _SC_LEVEL4_CACHE_SIZE:
    case _SC_LEVEL4_CACHE_ASSOC:
    case _SC_LEVEL4_CACHE_LINESIZE:
      /* In general we cannot determine these values.  Therefore we
	 return zero which indicates that no information is
	 available.  */
      return 0;
```

Then they have these files which for the most part implement the linesize detection for each arch:

```
sysdeps/unix/sysv/linux/aarch64/sysconf.c
sysdeps/unix/sysv/linux/alpha/sysconf.c
sysdeps/unix/sysv/linux/powerpc/sysconf.c
sysdeps/unix/sysv/linux/riscv/sysconf.c
sysdeps/unix/sysv/linux/s390/sysconf.c
sysdeps/unix/sysv/linux/x86/sysconf.c
```

It seems `x86_64` isn't even covered, unless they use `x86` in it?

Anyway, should be ok to add a patch with `#ifdef` guards around the block for one of those values, and return 0 otherwise.