linux-kernel - Re: [PATCH v2 1/7] arm64: Introduce system_capabilities

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

Message-ID: <20200110145019.GD8786@arrakis.emea.arm.com>
Date:   Fri, 10 Jan 2020 14:50:19 +0000
From:   Catalin Marinas <catalin.marinas@....com>
To:     Suzuki K Poulose <suzuki.poulose@....com>
Cc:     linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        will@...nel.org, maz@...nel.org, mark.rutland@....com,
        dave.martin@....com, ard.biesheuvel@...aro.org,
        christoffer.dall@....com, Will Deacon <will.deacon@....com>
Subject: Re: [PATCH v2 1/7] arm64: Introduce system_capabilities_finalized()
 marker

On Tue, Dec 17, 2019 at 06:33:56PM +0000, Suzuki K Poulose wrote:
> We finalize the system wide capabilities after the SMP CPUs
> are booted by the kernel. This is used as a marker for deciding
> various checks in the kernel. e.g, sanity check the hotplugged
> CPUs for missing mandatory features.
> 
> However there is no explicit helper available for this in the
> kernel. There is sys_caps_initialised, which is not exposed.
> The other closest one we have is the jump_label arm64_const_caps_ready
> which denotes that the capabilities are set and the capability checks
> could use the individual jump_labels for fast path. This is
> performed before setting the ELF Hwcaps, which must be checked
> against the new CPUs. We also perform some of the other initialization
> e.g, SVE setup, which is important for the use of FP/SIMD
> where SVE is supported. Normally userspace doesn't get to run
> before we finish this. However the in-kernel users may
> potentially start using the neon mode. So, we need to
> reject uses of neon mode before we are set. Instead of defining
> a new marker for the completion of SVE setup, we could simply
> reuse the arm64_const_caps_ready and enable it once we have
> finished all the setup. Also we could expose this to the
> various users as "system_capabilities_finalized()" to make
> it more meaningful than "const_caps_ready".
> 
> Cc: Ard Biesheuvel <ard.biesheuvel@...aro.org>
> Cc: Will Deacon <will.deacon@....com>
> Cc: Catalin Marinas <catalin.marinas@....com>
> Cc: Mark Rutland <mark.rutland@....com>
> Signed-off-by: Suzuki K Poulose <suzuki.poulose@....com>

Reviewed-by: Catalin Marinas <catalin.marinas@....com>