linux-kernel - Re: [PATCH v2 1/3] arm64/mm: Refactor PMD_PRESENT_INVALID and PTE_PROT

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]

Date: Tue, 30 Apr 2024 14:28:42 +0100
From: Catalin Marinas <catalin.marinas@....com>
To: Ryan Roberts <ryan.roberts@....com>
Cc: Will Deacon <will@...nel.org>, Joey Gouly <joey.gouly@....com>,
	Ard Biesheuvel <ardb@...nel.org>,
	Mark Rutland <mark.rutland@....com>,
	Anshuman Khandual <anshuman.khandual@....com>,
	David Hildenbrand <david@...hat.com>, Peter Xu <peterx@...hat.com>,
	Mike Rapoport <rppt@...ux.ibm.com>,
	Shivansh Vij <shivanshvij@...look.com>,
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 1/3] arm64/mm: Refactor PMD_PRESENT_INVALID and
 PTE_PROT_NONE bits

On Tue, Apr 30, 2024 at 12:35:49PM +0100, Ryan Roberts wrote:
> There is still one problem I need to resolve; During this work I discovered that
> core-mm can call pmd_mkinvalid() for swap pmds. On arm64 this will turn the swap
> pmd into a present pmd, and BadThings can happen in GUP-fast (and any other
> lockless SW table walkers). My original fix modified core-mm to only call
> pmd_mkinvalid() for present pmds. But discussion over there has shown that arm64
> is the only arch that cannot handle this. So I've been convinced that it's
> probably more robust to make arm64 handle it gracefully and add tests to
> debug_vm_pgtable.c to check for this. Patch incoming shortly, but it will cause
> a conflict with this series. So I'll send a v2 of this once that fix is accepted.

Sounds fine. I can queue the arm64 pmd_mkinvalid() fix for 6.9 and you
can base this series on top. But I have a preference for this patchset
to sit in -next for a bit anyway, so it might be 6.11 material.

-- 
Catalin