linux-kernel - Re: [PATCH RFC 3/4] dma-direct: add offset to zone_dma

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: <871qaqr477.fsf@tarshish>
Date: Tue, 09 Jan 2024 15:54:13 +0200
From: Baruch Siach <baruch@...s.co.il>
To: Catalin Marinas <catalin.marinas@....com>
Cc: Christoph Hellwig <hch@....de>, Marek Szyprowski
 <m.szyprowski@...sung.com>, Rob Herring <robh+dt@...nel.org>, Frank Rowand
 <frowand.list@...il.com>, Will Deacon <will@...nel.org>, Robin Murphy
 <robin.murphy@....com>, iommu@...ts.linux.dev, devicetree@...r.kernel.org,
 linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org, Petr
 Tesařík <petr@...arici.cz>, Ramon Fried
 <ramon@...reality.ai>, Elad
 Nachman <enachman@...vell.com>
Subject: Re: [PATCH RFC 3/4] dma-direct: add offset to zone_dma_bits

Hi Catalin,

On Tue, Jan 09 2024, Catalin Marinas wrote:
> On Tue, Jan 09, 2024 at 12:03:43PM +0200, Baruch Siach wrote:
>> On Mon, Jan 08 2024, Catalin Marinas wrote:
>> > On Wed, Dec 27, 2023 at 05:04:27PM +0200, Baruch Siach wrote:
>> >> Current code using zone_dma_bits assume that all addresses range in the
>> >> bits mask are suitable for DMA. For some existing platforms this
>> >> assumption is not correct. DMA range might have non zero lower limit.
>> >> 
>> >> Add 'zone_dma_off' for platform code to set base address for DMA zone.
>> >> 
>> >> Rename the dma_direct_supported() local 'min_mask' variable to better
>> >> describe its use as limit.
>> >> 
>> >> Suggested-by: Catalin Marinas <catalin.marinas@....com>
>> >
>> > When I suggested taking the DMA offsets into account, that's not exactly
>> > what I meant. Based on patch 4, it looks like zone_dma_off is equivalent
>> > to the lower CPU address. Let's say a system has DRAM starting at 2GB
>> > and all 32-bit DMA-capable devices has a DMA offset of 0. We want
>> > ZONE_DMA32 to end at 4GB rather than 6GB.
>> 
>> Patch 4 sets zone_dma_off to the lower limit from 'dma-ranges' property
>> that determines zone_dma_bits. This is not necessarily equivalent to
>> start of DRAM, though it happens to be that way on my platform.
>
> A bit better but it still assumes that all devices have the same DMA
> offset which may not be the case.

Current code calculates zone_dma_bits based on the lowest high limit of
all 'dma-ranges' properties. The assumption appears to be that this
limit fits all devices. This series does not change this assumption. It
only extends the logic to the lower limit of the "winning" 'dma-ranges'
to set the base address for DMA zone.

Moving to dma_zone_limit would not change that logic. Unless I'm missing
something.

Breaking the "one DMA zone fits all devices" assumption as Petr
suggested is a much larger change.

baruch

-- 
                                                     ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch@...s.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -