On Wed, Jul 10, 2019 at 08:40:43PM -0700, Graham Percival wrote: > > Ah, but how do you ask malloc to exceed RLIMIT_DATA? Not by passing it > > a large number, or any number, since that doesn't change the size of > > any data segment. > > Aha! I misunderstood "this limit" to mean "the number returned by > getrlimit()", not "the size of the data segment". So I thought that second > sentence was about talking about what should happen if you passed malloc() a > large value.
Ah, so we were all talking past one another. Good thing I wrote that post, I guess :-) In practice (across arbitrary platforms) I wouldn't count on malloc necessarily respecting any of these limits, including the limit imposed by the amount of DRAM in the known universe. In addition to bugs and perhaps unexpected phenomena like NetBSD's non-sbrk malloc, AFAIK there's a faction out there (connected with C++? dunno) who think it's incorrect for malloc to fail, ever, so if they're in control it won't. Instead, at some unpredictable point you get SIGSEGV or SIGKILL, like if the system ran out of swap. > My sincere apologies for wasting everybody's time. Eh, no worries :-) -- David A. Holland dholl...@netbsd.org
