linux-kernel - Re: [PATCH v1 11/13] arm64: mm: More flags for __flush_tlb

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: <dc9dec5c-f50d-4283-8a4b-609927f8a2bb@arm.com>
Date: Mon, 12 Jan 2026 12:00:16 +0000
From: Ryan Roberts <ryan.roberts@....com>
To: Linu Cherian <linu.cherian@....com>
Cc: Will Deacon <will@...nel.org>, Ard Biesheuvel <ardb@...nel.org>,
 Catalin Marinas <catalin.marinas@....com>,
 Mark Rutland <mark.rutland@....com>,
 Linus Torvalds <torvalds@...ux-foundation.org>,
 Oliver Upton <oliver.upton@...ux.dev>, Marc Zyngier <maz@...nel.org>,
 Dev Jain <dev.jain@....com>, linux-arm-kernel@...ts.infradead.org,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1 11/13] arm64: mm: More flags for __flush_tlb_range()

On 07/01/2026 03:21, Linu Cherian wrote:
> Ryan,
> 
> On Tue, Dec 16, 2025 at 02:45:56PM +0000, Ryan Roberts wrote:
>> Refactor function variants with "_nosync", "_local" and "_nonotify" into
>> a single __always_inline implementation that takes flags and rely on
>> constant folding to select the parts that are actually needed at any
>> given callsite, based on the provided flags.
>>
>> Flags all live in the tlbf_t (TLB flags) type; TLBF_NONE (0) continues
>> to provide the strongest semantics (i.e. evict from walk cache,
>> broadcast, synchronise and notify). Each flag reduces the strength in
>> some way; TLBF_NONOTIFY, TLBF_NOSYNC and TLBF_NOBROADCAST are added to
>> complement the existing TLBF_NOWALKCACHE.
> 
> It would be nice to have some notes added on the below for better clarity 
> * What a walk cache is and why we bother about them ?
> * Why and how should we invalidate the walk caches ?

There is a large comment block already in tlbflush.h which talks about which
operations affect the walk-cache and which don't. Although it never defines the
walk-cache; I'll add something vague there, although I don't want to be too
specific as it's a uarch thing really.

Thanks,
Ryan

> 
> --
> Linu Cherian.