Message-ID: <qputoki6kfvowm35ops5wytiezdixsgel3klaq2vj4ylpv44qy@r2khmqemg6jo>
Date: Tue, 9 Jul 2024 13:54:02 -0500
From: Bjorn Andersson <andersson@...nel.org>
To: Prakash Gupta <quic_guptap@...cinc.com>
Cc: Mukesh Ojha <quic_mojha@...cinc.com>, konrad.dybcio@...aro.org, 
	linux-arm-msm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] firmware: qcom_scm: Add a padded page to ensure DMA
 memory from lower 4GB

On Mon, Jul 08, 2024 at 03:33:57PM GMT, Prakash Gupta wrote:
> 
> 
> On 6/24/2024 8:34 PM, Mukesh Ojha wrote:
> > On Sun, Jun 23, 2024 at 07:25:23PM -0500, Bjorn Andersson wrote:
> >> On Wed, May 29, 2024 at 05:24:29PM GMT, Mukesh Ojha wrote:
> >>>
> >>>
> >>> On 5/27/2024 2:16 AM, Bjorn Andersson wrote:
> >>>> On Fri, May 24, 2024 at 09:01:45PM GMT, Mukesh Ojha wrote:
> >>>>> For SCM protection, memory allocation should be physically contiguous,
> >>>>> 4K aligned, and non-cacheable to avoid XPU violations. This granularity
> >>>>> of protection applies from the secure world. Additionally, it's possible
> >>>>> that a 32-bit secure peripheral will access memory in SoCs like
> >>>>> sm8{4|5|6}50 for some remote processors. Therefore, memory allocation
> >>>>> needs to be done in the lower 4 GB range. To achieve this, Linux's CMA
> >>>>> pool can be used with dma_alloc APIs.
> >>>>>
> >>>>> However, dma_alloc APIs will fall back to the buddy pool if the requested
> >>>>> size is less than or equal to PAGE_SIZE. It's also possible that the remote
> >>>>> processor's metadata blob size is less than a PAGE_SIZE. Even though the
> >>>>> DMA APIs align the requested memory size to PAGE_SIZE, they can still fall
> >>>>> back to the buddy allocator, which may fail if `CONFIG_ZONE_{DMA|DMA32}`
> >>>>> is disabled.
> >>>>
> >>>> Does "fail" here mean that the buddy heap returns a failure - in some
> >>>> case where dma_alloc would have succeeded, or that it does give you
> >>>> a PAGE_SIZE allocation which doesn't meeting your requirements?
> >>>
> >>> Yes, buddy will also try to allocate memory and may not get PAGE_SIZE memory
> >>> in lower 4GB(for 32bit capable device) if CONFIG_ZONE_{DMA|DMA32} is
> >>> disabled.
> >>
> >> Is that -ENOMEM or does "not get" mean that the buddy fallback will
> >> provide an allocation above 4GB?
> > 
> > dma_alloc_coherent() returns NULL in that situation.
> > 
> > https://elixir.bootlin.com/linux/v6.10-rc5/source/kernel/dma/direct.c#L142
> > 
> > -Mukesh
> > 
> 
> scm device is using DMA mask as 32b. With size <= PAGE_SIZE, call to 
> dma_alloc_contiguous() at [1] will return NULL.
> 
> With DMA mask as 32b and CONFIG_ZONE_{DMA|DMA32} disabled, the allocation 
> can't be guaranteed within 32b with buddy. Hence will return page as NULL.
> 
> Adding a padded page should allow allocation from CMA region.
> 

Thanks for the explanation, Prakash and Mukesh. It sounds wrong to me
that one has to trick the DMA API this way.

I assume this comes from the check [a], could we perhaps propose
modifying this? If nothing else it will give us the DMA maintainers'
input regarding this being the appropriate workaround.

[a] https://elixir.bootlin.com/linux/v6.10-rc5/source/kernel/dma/contiguous.c#L362

Regards,
Bjorn

> Thanks,
> Prakash
> 
> [1] https://elixir.bootlin.com/linux/v6.10-rc5/source/kernel/dma/direct.c#L131
> >>
> >> Regards,
> >> Bjorn
> >>
> >>> However, DMA memory would have successful such case if
> >>> padding is added to size to cross > PAGE_SIZE.
> >>>
> >>>>
> >>>>  From this I do find the behavior of dma_alloc unintuitive, do we know if
> >>>> there's a reason for the "equal to PAGE_SIZE" case you describe here?
> >>>
> >>> I am not a memory expert but the reason i can think of could be, <=
> >>> PAGE_SIZE can anyway possible to be requested outside DMA coherent api's
> >>> with kmalloc and friends api and that could be the reason it is falling
> >>> back to buddy pool in DMA api.
> >>>
> >>> -Mukesh
> > 
> 
> > 

Powered by blists - more mailing lists