Re: malloc() exceeds RLIMIT_DATA

[Graham Percival <gperciva%tarsnap.com@localhost>]

On Wed, Jul 10, 2019 at 05:59:05AM +0000, David Holland wrote:
> On Tue, Jul 09, 2019 at 10:24:53AM -0700, Graham Percival wrote:
>  >     RLIMIT_DATA
>  >     This is the maximum size of a data segment of the process, in bytes. If
>  >     this limit is exceeded, the malloc() function shall fail with errno set to
>  >     [ENOMEM].
>  >     http://pubs.opengroup.org/onlinepubs/9699919799/functions/getrlimit.html
>  > 
>  > The relevant part is the second sentence.
> 
> 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.

My sincere apologies for wasting everybody's time.


> (Note: if you're trying to do something practical, you might find
> RLIMIT_AS or RLIMIT_RSS more helpful.)

We were checking all three, actually:
https://github.com/Tarsnap/scrypt/blob/master/lib/util/memlimit.c#L153
but since RLIMIT_DATA is not relevant, I've submitted a patch to remove it.
Using that patch allows the scrypt tests to complete successfully on NetBSD.

Thank you for the clarifications, and sorry again,
- Graham