[Issue 21089] New: With vm.overcommit_memory=0, DMD can't link if it uses more than half the total (ram+swap) memory in the system.

[d-bugmail at puremagic.com]

https://issues.dlang.org/show_bug.cgi?id=21089

          Issue ID: 21089
           Summary: With vm.overcommit_memory=0, DMD can't link if it uses
                    more than half the total (ram+swap) memory in the
                    system.
           Product: D
           Version: D2
          Hardware: x86
                OS: Linux
            Status: NEW
          Severity: minor
          Priority: P1
         Component: dmd
          Assignee: nobody at puremagic.com
          Reporter: default_357-line at yahoo.de

This is because DMD fork()s to start the linker. Fork temporarily
overcommit-allocates, and the allocation is refused by the system if
`vm.overcommit_memory=0` and the forked DMD would push the total used address
space over the total of memory + swap.

This is a known issue, see the vfork() doc:

       *  On systems where memory is constrained, vfork() avoids the need to
          temporarily commit memory (see the description of
          /proc/sys/vm/overcommit_memory in proc(5)) in order to execute a
          new program.  (This can be especially beneficial where a large
          parent process wishes to execute a small helper program in a child
          process.)  By contrast, using fork(2) in this scenario requires
          either committing an amount of memory equal to the size of the
          parent process (if strict overcommitting is in force) or
          overcommitting memory with the risk that a process is terminated
          by the out-of-memory (OOM) killer.

But vfork() is itself deprecated.

The fork call should be replaced with posix_spawn(), which does not share
caller process memory.

--