This is the mail archive of the
mailing list for the GCC project.
Re: need help with Pointer Bounds Checking documentation
- From: Sandra Loosemore <sandra at codesourcery dot com>
- To: Ilya Enkovich <enkovich dot gnu at gmail dot com>
- Cc: Ilya Enkovich <ilya dot enkovich at intel dot com>, GCC Development <gcc at gcc dot gnu dot org>
- Date: Tue, 24 Feb 2015 09:47:17 -0700
- Subject: Re: need help with Pointer Bounds Checking documentation
- Authentication-results: sourceware.org; auth=none
- References: <54E68470 dot 90700 at codesourcery dot com> <CAMbmDYYoshG0-Lx6MOobAZyR1CMaSGwQxU_xmZxXrUyCcepyPA at mail dot gmail dot com> <54E76349 dot 7050508 at codesourcery dot com> <CAMbmDYZ9wRaZdS7Dy-QpNpjiuxTdi6pWSBbgNrW+m5kxMjXfEw at mail dot gmail dot com>
On 02/24/2015 01:14 AM, Ilya Enkovich wrote:
2015-02-20 19:39 GMT+03:00 Sandra Loosemore <firstname.lastname@example.org>:
On 02/20/2015 01:12 AM, Ilya Enkovich wrote:
2015-02-20 3:48 GMT+03:00 Sandra Loosemore <email@example.com>:
The section "Pointer Bounds Checker builtins" in extend.texi is on my
for being in need of copy-editing, but reading through the existing text,
am quite confused. In several places it refers to turning the Pointer
Bounds Checker off, but how do you do that? I don't see any documented
Pointer Bounds Checker is off by default. Patch adding options
documentation is a part of reviewed libmpx patch series which is
waiting for steering committee approval. I didn't want to commit
documentation separately because it has references to features not yet
committed to trunk. Meanwhile it is available on Wiki:
Well, hmmm. Isn't that true of the already-committed Pointer Bounds
Checking intrinsics documentation, then, too? That is, the intrinsics
aren't useful to users without the support to enable this feature that
hasn't been committed yet?
Intrinsics are fully implemented. Compiled programs may be run using
SDE and MPX runtime library provided with it.
Poking around, I see that the -fcheck-pointer-bounds and various
-fchkp-* options are listed in c-family/c.opt, but they are not listed
in the GCC manual. The section on intrinsics is not enough, by itself,
to tell users how to use this feature. Can we please try to get in the
documentation for the options that have been implemented so far? That's
a requirement of the GCC coding conventions.
I see Joseph already brought up the issue of missing documentation back
But I cannot find any proposed documentation patch for these options either.