[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 31 Jul 2022 16:04:11 +0900
From: Hector Martin <marcan@...can.st>
To: Robin Murphy <robin.murphy@....com>, Janne Grunau <j@...nau.net>,
iommu@...ts.linux-foundation.org
Cc: Konrad Dybcio <konrad.dybcio@...ainline.org>,
asahi@...ts.linux.dev, Sven Peter <sven@...npeter.dev>,
Alyssa Rosenzweig <alyssa@...enzweig.io>,
Georgi Djakov <quic_c_gdjako@...cinc.com>,
"Isaac J. Manjarres" <isaacm@...eaurora.org>,
Joerg Roedel <joro@...tes.org>, Will Deacon <will@...nel.org>,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 4/5] iommu/io-pgtable-dart: Add DART PTE support for
t6000
On 28/06/2022 00.13, Robin Murphy wrote:
> On 2022-06-21 08:18, Janne Grunau wrote:
>> From: Sven Peter <sven@...npeter.dev>
>>
>> The DARTs present in the M1 Pro/Max/Ultra SoC use a diffent PTE format.
>> They support a 42bit physical address space by shifting the paddr and
>> extending its mask inside the PTE.
>> They also come with mandatory sub-page protection now which we just
>> configure to always allow access to the entire page. This feature is
>> already present but optional on the previous DARTs which allows to
>> unconditionally configure it.
>>
>> Signed-off-by: Sven Peter <sven@...npeter.dev>
>> Co-developed-by: Janne Grunau <j@...nau.net>
>> Signed-off-by: Janne Grunau <j@...nau.net>
>>
>> ---
>>
>> Changes in v3:
>> - apply change to io-pgtable-dart.c
>> - handle pte <> paddr conversion based on the pte format instead of
>> the output address size
>>
>> Changes in v2:
>> - add APPLE_DART2 PTE format
>>
>> drivers/iommu/io-pgtable-dart.c | 51 +++++++++++++++++++++++++++------
>> drivers/iommu/io-pgtable.c | 1 +
>> include/linux/io-pgtable.h | 1 +
>> 3 files changed, 45 insertions(+), 8 deletions(-)
>>
> [...]
>> @@ -536,7 +571,7 @@ apple_dart_alloc_pgtable(struct io_pgtable_cfg *cfg, void *cookie)
>> if (!cfg->coherent_walk)
>> return NULL;
>>
>> - if (cfg->oas > 36)
>> + if (cfg->oas != 36 && cfg->oas != 42)
>> return NULL;
>
> Wouldn't it make sense to tie this to the format? Maybe 36-bit OAS is
> still valid with v2, but presumably 42-bit with v1 definitely isn't.
FWIW, 36-bit OAS with v2 is valid (this is the case on M2/t8112).
- Hector
Powered by blists - more mailing lists