linux-kernel - Re: [PATCH v4 2/2] RISC-V: Implement sparsemem

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: <5506c875-9387-acc9-a7fe-5b7c10036c40@deltatee.com>
Date:   Tue, 13 Aug 2019 10:14:49 -0600
From:   Logan Gunthorpe <logang@...tatee.com>
To:     Greentime Hu <green.hu@...il.com>
Cc:     Rob Herring <robh@...nel.org>, Albert Ou <aou@...s.berkeley.edu>,
        Andrew Waterman <andrew@...ive.com>,
        Palmer Dabbelt <palmer@...ive.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Stephen Bates <sbates@...thlin.com>, paul.walmsley@...ive.com,
        Olof Johansson <olof@...om.net>, greentime.hu@...ive.com,
        linux-riscv@...ts.infradead.org,
        Michael Clark <michaeljclark@....com>,
        Christoph Hellwig <hch@....de>
Subject: Re: [PATCH v4 2/2] RISC-V: Implement sparsemem



On 2019-08-13 12:04 a.m., Greentime Hu wrote:
> I think  flat mem doesn't support memory-with-hole scenario.
> In mm/Kconfig, it says
> "
>           For systems that have holes in their physical address
>           spaces and for features like NUMA and memory hotplug,
>           choose "Sparse Memory"
> "
> IMHO, the memory-with-hole scenario should only be tested for sparse
> mem but flat mem.

Fair enough.

> The generic pfn_valid() is just for non-mmu arches. 

The generic pfn_valid() in asm-generic is only for non-mmu arches.

> Every architecture
> with mmu defines their own pfn_valid().

Not true. Arm64, for example just uses the generic implementation in
mmzone.h. My main question is whether we can just do that. If we can't
we should probably structure it like powerpc where they only use the
arch-specific helper for CONFIG_FLATMEM instead of when CONFIG_SPARSEMEM
isn't set.

Logan