Page MenuHomeFreeBSD

Return different error code for the guard page layout violation.
ClosedPublic

Authored by kib on Nov 21 2017, 10:01 PM.
Tags
None
Referenced Files
Unknown Object (File)
Fri, Mar 20, 5:27 AM
Unknown Object (File)
Thu, Mar 19, 5:29 PM
Unknown Object (File)
Sun, Mar 15, 10:54 AM
Unknown Object (File)
Sat, Mar 14, 11:13 PM
Unknown Object (File)
Tue, Mar 10, 11:35 AM
Unknown Object (File)
Mon, Mar 9, 9:54 PM
Unknown Object (File)
Thu, Mar 5, 3:52 AM
Unknown Object (File)
Wed, Mar 4, 11:25 PM
Subscribers

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

sys/vm/vm_map.c
3619–3621 ↗(On Diff #35568)

vm_map_insert() returns KERN_INVALID_ARGUMENT for all of these errors. I would argue that this function should do the same, both for consistency with vm_map_insert() and because KERN_INVALID_ARGUMENT better describes these errors.

Also, vm_map_insert() has a slighyly different "max" check. The equivalent would be "addrbos + max_ssize > vm_map_max(map)".

sys/vm/vm_map.c
3619–3621 ↗(On Diff #35568)

vm_map_insert()'s address wrap check also uses <= rather than <.

Duplicate the vm_map_insert() behavior for error checking WRT the address range checking.

I've skimmed the code for both this function and vm_map_insert(), and I believe that both the error checks and their corresponding return values are now consistent.

This revision is now accepted and ready to land.Nov 22 2017, 4:19 PM
This revision was automatically updated to reflect the committed changes.