linux-kernel - [PATCH v2 0/4] wire up CPU features to udev based module loading

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <1385048552-12528-1-git-send-email-ard.biesheuvel@linaro.org>
Date:	Thu, 21 Nov 2013 16:42:28 +0100
From:	Ard Biesheuvel <ard.biesheuvel@...aro.org>
To:	x86@...nel.org, linux-arm-kernel@...ts.infradead.org,
	linux-kernel@...r.kernel.org
Cc:	akpm@...ux-foundation.org, gregkh@...uxfoundation.org,
	catalin.marinas@....com, hpa@...or.com, ak@...ux.intel.com,
	mingo@...hat.com, Ard Biesheuvel <ard.biesheuvel@...aro.org>
Subject: [PATCH v2 0/4] wire up CPU features to udev based module loading

X86 maintainer(s): if you have no objections to this approach, may I please have
your ack(s) on the x86 bits so I can work with Catalin to take the series
upstream through his tree?

This series implements automatic module loading based on optional CPU features,
and tries to do so in a generic way.

Typical usage would look like this:

static struct cpu_feature mod_cpu_feature[] = {
        { HWCAP_CRC32 },
        {}
};
MODULE_DEVICE_TABLE(cpu, mod_cpu_feature);

where (on the arm64 arch) the module in question would be loaded automatically
if the CPU has support for the optional CRC instructions. [Note that the example
is based on a pending patch by Steve Capper: 'arm64: Add hwcaps for crypto and
CRC32 extensions']

Changes wrt v1:
- use ELF_PLATFORM rather than hardcoded name for printing modalias on arm64

Changes wrt RFC version of this series:
- this time, instead of doing something entirely separate for non-x86 archs, the
  new generic modalias version allocates enough bits to cover all x86 features
  (320+), and the x86 version was tweaked to resemble the generic one more
  closely (i.e., both use 'cpu:type:...:feature:...' now)
- removed the weak and alias GCC attribute foo

Ard Biesheuvel (4):
  x86: move arch_cpu_uevent() to generic code
  cpu: advertise CPU features over udev in a generic way
  x86: align with generic cpu modalias
  arm64: advertise CPU features for modalias matching

 arch/arm64/Kconfig                |  3 +++
 arch/arm64/kernel/setup.c         | 20 ++++++++++++++++++++
 arch/x86/kernel/cpu/match.c       | 14 +-------------
 drivers/base/cpu.c                | 15 ++++++++++++++-
 include/linux/cpu.h               |  1 -
 include/linux/mod_devicetable.h   | 11 +++++++++++
 scripts/mod/devicetable-offsets.c |  3 +++
 scripts/mod/file2alias.c          | 20 +++++++++++++++-----
 8 files changed, 67 insertions(+), 20 deletions(-)

-- 
1.8.3.2

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/