MAP_FIXED_NOREPLACE support

[jgarvin]

A new flag introduced in kernel 4.17.

I was taking a look at trying to add this myself, but I don't know if it should use #[cfg(target_os = "linux")] or #[cfg(any(target_os = "android", target_os = "linux"))]. I also don't know if there is any cfg hackery that should be done for supporting people on older kernels?

#define MAP_FIXED_NOREPLACE 0x100000

   MAP_FIXED_NOREPLACE (since Linux 4.17)
          This  flag  provides  behavior  that  is  similar   to
          MAP_FIXED  with  respect  to the addr enforcement, but
          differs in that MAP_FIXED_NOREPLACE never  clobbers  a
          preexisting  mapped  range.   If  the  requested range
          would collide with an existing mapping, then this call
          fails  with the error EEXIST.  This flag can therefore
          be used as a way to atomically (with respect to  other
          threads)  attempt  to map an address range: one thread
          will succeed; all others will report failure.

          Note that older kernels which  do  not  recognize  the
          MAP_FIXED_NOREPLACE  flag will typically (upon detect‐
          ing a collision with a preexisting mapping) fall  back
          to  a  "non-MAP_FIXED" type of behavior: they will re‐
          turn an address that is different from  the  requested
          address.    Therefore,   backward-compatible  software
          should check the  returned  address  against  the  re‐
          quested address.

[asomers]

You don't need to do anything at build-time to support older Linux kernels. If an application tries to use an unsupported feature, it will simply fail at runtime. As for Android support, the best thing to do is to check the latest version of the Android headers. If it's in there, then you should include it here (and you may have to add it to https://github.com/rust-lang/libc first).

[rtzoeller]

#1636 addresses this for Linux. If it should be defined on Android as well, it'll need to be added to rust-lang/libc before it can be added to nix.