linux-kernel - Re: [PATCH 1/1] arm64: fix flush_cache

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] [day] [month] [year] [list]

Message-ID: <20160526164456.GE7456@e104818-lin.cambridge.arm.com>
Date:	Thu, 26 May 2016 17:44:57 +0100
From:	Catalin Marinas <catalin.marinas@....com>
To:	Russell King - ARM Linux <linux@...linux.org.uk>
Cc:	Mark Rutland <mark.rutland@....com>,
	Hanjun Guo <guohanjun@...wei.com>,
	Benjamin Herrenschmidt <benh@...nel.crashing.org>,
	"Leizhen (ThunderTown)" <thunder.leizhen@...wei.com>,
	Will Deacon <will.deacon@....com>,
	linux-kernel <linux-kernel@...r.kernel.org>,
	Xinwei Hu <huxinwei@...wei.com>, Zefan Li <lizefan@...wei.com>,
	Paul Mackerras <paulus@...ba.org>,
	Michael Ellerman <mpe@...erman.id.au>,
	Tianhong Ding <dingtianhong@...wei.com>,
	linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>
Subject: Re: [PATCH 1/1] arm64: fix flush_cache_range

On Wed, May 25, 2016 at 06:27:07PM +0100, Russell King - ARM Linux wrote:
> On Wed, May 25, 2016 at 04:22:55PM +0100, Catalin Marinas wrote:
> > That's when we realised that the CoW problem no longer exists for
> > non-aliasing VIPT caches. However, the I-cache counterpart 6060e8df5178
> > has not been reverted.
> 
> I think I mostly agree, except for reverting 6060e8df5178, which I don't
> think would be correct.  That reintroduces the possibility of flushing
> the I-cache twice in that path, once for aliasing vipt dcaches, and again
> for asid tagged vivt icaches.  I'd rather have the code structured so we
> only do this once.

__sync_icache_dcache() seems to take care of AIVIVT caches as well (i.e.
it invalidates the whole I-cache) so I think we could just remove the
__flush_icache_all() from flush_cache_range().

I can't find a scenario where we still need D-cache clean+invalidate in
flush_cache_range() for aliasing VIPT. As for the I-cache,
__sync_icache_dcache() takes care of all the aliases.

-- 
Catalin