This is the mail archive of the
mailing list for the GCC project.
RE: [PATCH] [i386] Enable Control-flow Enforcement Technology (CET).
- From: "Tsimbalist, Igor V" <igor dot v dot tsimbalist at intel dot com>
- To: Joseph Myers <joseph at codesourcery dot com>
- Cc: "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>, "Chupin, Pavel V" <pavel dot v dot chupin at intel dot com>, "Tsimbalist, Igor V" <igor dot v dot tsimbalist at intel dot com>
- Date: Wed, 21 Jun 2017 14:07:29 +0000
- Subject: RE: [PATCH] [i386] Enable Control-flow Enforcement Technology (CET).
- Authentication-results: sourceware.org; auth=none
- Dlp-product: dlpe-windows
- Dlp-reaction: no-action
- Dlp-version: 10.0.102.7
- References: <D511F25789BA7F4EBA64C8A63891A0027A9EA880@IRSMSX102.ger.corp.intel.com> <alpine.DEB.email@example.com>
Thanks for the feedback. I'll redo the patch according to your comments.
From: Joseph Myers [mailto:firstname.lastname@example.org]
Sent: Tuesday, June 20, 2017 11:30 PM
To: Tsimbalist, Igor V <email@example.com>
Cc: firstname.lastname@example.org; Chupin, Pavel V <email@example.com>
Subject: Re: [PATCH] [i386] Enable Control-flow Enforcement Technology (CET).
This patch is completely missing documentation (in *.texi files) of the new options, attribute, built-in functions etc.
You appear to be adding quite target-specific things to the architecture-independent compiler. If the attribute, for example, is to be architecture-independent, the documentation needs to define semantics for it that make sense on any architecture (or any architecture providing such features), not just on x86; the patch submission needs to justify the design choices of what is target-specific and what is target-independent.
(Cf. MPX where there are various architecture-independent features for which a software implementation would be logically possible, although actually the only implementation of those features in GCC is for MPX
hardware.) I don't think this patch would even build for non-x86 targets, because you're putting completely x86-specific references such as TARGET_CET and gen_nop_endbr in target-independent files.
Joseph S. Myers