linux-kernel - Re: [PATCH v1 1/1] ARM: Select DMA_DIRECT

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]

Date:   Mon, 23 Oct 2023 08:16:01 +0200
From:   Christoph Hellwig <hch@....de>
To:     Marek Szyprowski <m.szyprowski@...sung.com>
Cc:     Christoph Hellwig <hch@....de>,
        Jim Quinlan <james.quinlan@...adcom.com>,
        Linus Walleij <linus.walleij@...aro.org>,
        bcm-kernel-feedback-list@...adcom.com, jim2101024@...il.com,
        Russell King <linux@...linux.org.uk>,
        Arnd Bergmann <arnd@...db.de>,
        Geert Uytterhoeven <geert+renesas@...der.be>,
        "Russell King (Oracle)" <rmk+kernel@...linux.org.uk>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Jonathan Corbet <corbet@....net>,
        Thomas Gleixner <tglx@...utronix.de>,
        Sebastian Reichel <sebastian.reichel@...labora.com>,
        "Mike Rapoport (IBM)" <rppt@...nel.org>,
        Eric DeVolder <eric.devolder@...cle.com>,
        Nathan Chancellor <nathan@...nel.org>,
        "Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
        Christophe Leroy <christophe.leroy@...roup.eu>,
        "moderated list:ARM PORT" <linux-arm-kernel@...ts.infradead.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v1 1/1] ARM: Select DMA_DIRECT_REMAP to fix restricted
 DMA

On Fri, Oct 20, 2023 at 10:16:46AM +0200, Marek Szyprowski wrote:
> For historical reasons (performance and limitations of the pre-ARM v7 
> cores), on the 32bit ARM the whole kernel's direct mapping is done using 
> so called 'sections' (1MiB size afair). Those sections are created in 
> the per-process MMU page tables (there are no separate MMU table for the 
> kernel mappings), so altering those mappings requires updating bits in 
> all processes in the system. Practically this means that those mappings 
> has to be static once created during boot time.

That's actually the same on many architetures, and matches the
explanation I heard from Russell before.

> That's why when no CMA 
> is selected, the whole dma_alloc_coherent() allocations are limited to 
> rather small region, which is already remapped as non-cached during boot.

But this does not match my understanding of the code:

 - arch_dma_alloc calls __dma_alloc with is_coherent set to false
 - __dma_alloc then selects cma_allocator if CMA is supported for
   the device / allocation, else remap_allocator if the allocation
   is allowed to block and only if blocking is not allowed pool_allocator
   to allocate from the boot-time pool

This very match matches the dma-direct flow with DMA_DIRECT_REMAP
selected.  The major exception is the direct mapping of the CMA
allocations done by arm32.