linux-kernel - Re: [RFC PATCH 2/2] arm64, tlbflush: don't TLBI broadcast if page reused in write fault

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: <aMvbNvCsSxQjA2rq@arm.com>
Date: Thu, 18 Sep 2025 11:13:10 +0100
From: Catalin Marinas <catalin.marinas@....com>
To: "Huang, Ying" <ying.huang@...ux.alibaba.com>
Cc: Ryan Roberts <ryan.roberts@....com>, Will Deacon <will@...nel.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	David Hildenbrand <david@...hat.com>,
	Lorenzo Stoakes <lorenzo.stoakes@...cle.com>,
	Vlastimil Babka <vbabka@...e.cz>, Zi Yan <ziy@...dia.com>,
	Baolin Wang <baolin.wang@...ux.alibaba.com>,
	Yang Shi <yang@...amperecomputing.com>,
	"Christoph Lameter (Ampere)" <cl@...two.org>,
	Dev Jain <dev.jain@....com>, Barry Song <baohua@...nel.org>,
	Anshuman Khandual <anshuman.khandual@....com>,
	Yicong Yang <yangyicong@...ilicon.com>,
	Kefeng Wang <wangkefeng.wang@...wei.com>,
	Kevin Brodsky <kevin.brodsky@....com>,
	Yin Fengwei <fengwei_yin@...ux.alibaba.com>,
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
	linux-mm@...ck.org
Subject: Re: [RFC PATCH 2/2] arm64, tlbflush: don't TLBI broadcast if page
 reused in write fault

On Thu, Sep 18, 2025 at 10:18:49AM +0800, Huang, Ying wrote:
> Ryan Roberts <ryan.roberts@....com> writes:
> > contpte_ptep_set_access_flags() currently does a (broadcast) __flush_tlb_range()
> > on the (pte_write(orig_pte) == pte_write(entry)) path. I think that should be
> > changed to a local range invalidation to be consistent with this change.
> 
> Yes.  This should be changed too.  However, it means we need a local
> variant of __flush_tlb_range() and flush_tlb_mm().  Is it OK to
> introduce them first and tidy up later?

If it's just for contpte, we'd never take the flush_tlb_mm() path. So we
could instead add a specific local_flush_tlb_contpte_range().

-- 
Catalin