[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <87v9embufl.fsf@oldenburg2.str.redhat.com>
Date: Tue, 03 Nov 2020 11:34:22 +0100
From: Florian Weimer <fweimer@...hat.com>
To: Szabolcs Nagy <szabolcs.nagy@....com>
Cc: libc-alpha@...rceware.org, Jeremy Linton <jeremy.linton@....com>,
Catalin Marinas <catalin.marinas@....com>,
Mark Rutland <mark.rutland@....com>,
Will Deacon <will.deacon@....com>,
Mark Brown <broonie@...nel.org>,
Kees Cook <keescook@...omium.org>,
Salvatore Mesoraca <s.mesoraca16@...il.com>,
Lennart Poettering <mzxreary@...inter.de>,
Topi Miettinen <toiwoton@...il.com>,
linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
kernel-hardening@...ts.openwall.com,
linux-hardening@...r.kernel.org
Subject: Re: [PATCH 3/4] aarch64: Use mmap to add PROT_BTI instead of
mprotect [BZ #26831]
* Szabolcs Nagy:
> Re-mmap executable segments if possible instead of using mprotect
> to add PROT_BTI. This allows using BTI protection with security
> policies that prevent mprotect with PROT_EXEC.
>
> If the fd of the ELF module is not available because it was kernel
> mapped then mprotect is used and failures are ignored. It is
> expected that linux kernel will add PROT_BTI when mapping a module
> (current linux as of version 5.9 does not do this).
>
> Computing the mapping parameters follows the logic of
> _dl_map_object_from_fd more closely now.
What's the performance of this on execve-heavy workloads, such as kernel
or glibc builds? Hopefully it's cheap because these mappings have not
been faulted in yet.
Thanks,
Florian
--
Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn,
Commercial register: Amtsgericht Muenchen, HRB 153243,
Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill
Powered by blists - more mailing lists