mmap(2): Document guard size and related EINVAL.

[kib]

Summary

Based on submission by: emaste

[emaste]

lib/libc/sys/mmap.2

354

The size of the guard, in pages, ...

460

less than or equal to

[kib]

Handle Ed' notes.
Use .Dv for sysctl mib name.
Bump date.

[markj]

lib/libc/sys/mmap.2

456

Didn't the mailing list report mention ENOMEM?

[markj]

lib/libc/sys/mmap.2

456

Sorry, I did not read the full thread.

[alc]

head/lib/libc/sys/mmap.2

354–355 ↗

(On Diff #78571)

IMHO, we should be explicitly saying that the guard is carved out the specified stack size, reducing the effective size of the stack. Right now, I could infer that from reading the errors section, but I argue that I shouldn't have to infer it.

[emaste]

head/lib/libc/sys/mmap.2

354–355 ↗

(On Diff #78571)

good point, how's this:

Index: lib/libc/sys/mmap.2
===================================================================
--- lib/libc/sys/mmap.2 (revision 366945)
+++ lib/libc/sys/mmap.2 (working copy)
@@ -353,6 +353,8 @@
 stacks can grow without requiring mapping the whole stack in advance.
 The size of the guard, in pages, is specified by sysctl
 .Dv security.bsd.stack_guard_page .
+The guard region is placed at the starting address of the allocation,
+and the usable stack is reduced by the size of the guard.
 .El
 .Pp
 The

[emaste]

Oh I see @kib has D26908 for this.