This is the mail archive of the gcc-help@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: Relocation avoidance for function pointers

From: Florian Weimer <fweimer at redhat dot com>
To: Jeff Law <law at redhat dot com>, Martin Sebor <msebor at gmail dot com>
Cc: "gcc-help at gcc dot gnu dot org" <gcc-help at gcc dot gnu dot org>
Date: Mon, 7 Nov 2016 15:03:17 +0100
Subject: Re: Relocation avoidance for function pointers
Authentication-results: sourceware.org; auth=none
References: <c235e3e8-e320-90d0-6c91-1f76e221fe30@redhat.com> <39c2783e-bd96-892b-24ec-eb397fff1a64@gmail.com> <eab1e944-e51e-cadf-9b36-4240380526f5@redhat.com> <212c9e44-0e67-e405-b0bf-8014d537143b@gmail.com> <94436c40-5ffe-1557-90b5-5bcb44d73c96@redhat.com> <3863ba2e-3e83-5fce-8a4d-bc5773fe91e8@redhat.com>

On 11/05/2016 01:37 AM, Jeff Law wrote:

Quite a few targets support symbol differences in relocations, as long
as they are contained within the same module.  So even if some
relaxation is going on, function pointer differences would still work.

Really?  When I looked at this many years ago, none did.  Hacking it for
the PA (which had proper reloc support for difference of symbols) was
painful.

Uh-oh. Someone mentioned internally that there are other targets nowwhich allow this. I tried to generate some object files which use thislatitude, but I failed, mainly because I couldn't GCC to emit anyinstructions whose width changed after linking.

In any case, I still think this feature is sufficiently useful to exposeit to programmers even if it is not completely portable.


Florian

References:
- Re: Relocation avoidance for function pointers
  - From: Florian Weimer
- Re: Relocation avoidance for function pointers
  - From: Jeff Law

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]