[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YympLxofJVAJVyDj@arm.com>
Date: Tue, 20 Sep 2022 12:51:11 +0100
From: Szabolcs Nagy <szabolcs.nagy@....com>
To: Chris Kennelly <ckennelly@...gle.com>
Cc: libc-coord@...ts.openwall.com,
Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
"carlos@...hat.com" <carlos@...hat.com>,
libc-alpha <libc-alpha@...rceware.org>
Subject: Re: Re: RSEQ symbols: __rseq_size, __rseq_flags vs
__rseq_feature_size
The 09/16/2022 11:41, Chris Kennelly wrote:
> > >
> > > 2) Implement the entire function as IFUNC and select whether a rseq or
> > > non-rseq implementation should be used at C startup. The tradeoff
> > > here is code size vs speed, and using IFUNC for things like malloc
> > > may add additional constraints on the startup order.
> >
>
> IFUNC has significant performance overheads as well. For frequently used
> code (like memcpy), avoiding them has been an optimization for us (
> https://research.google/pubs/pub50338/) even if it leaves some nominal
> microbenchmark performance on the table.
only in static linked exe. which is not the important case for distros
that want the ability to update the libc with security fixes.
glibc is not optimized for static linking, if that's a requirement that
should be a separate discussion (it's not just about ifunc).
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
