[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <1591886384.28dd734zt4.none@localhost>
Date: Thu, 11 Jun 2020 10:49:53 -0400
From: "Alex Xu (Hello71)" <alex_y_xu@...oo.ca>
To: Christoph Hellwig <hch@....de>,
David Rientjes <rientjes@...gle.com>
Cc: alsa-devel@...a-project.org, bp@...en8.de, hch@...radead.org,
hpa@...or.com, linux-kernel@...r.kernel.org, mingo@...hat.com,
Pavel Machek <pavel@....cz>, perex@...ex.cz,
tglx@...utronix.de, tiwai@...e.com, x86@...nel.org
Subject: Re: [patch for-5.8] dma-pool: decouple DMA_REMAP from
DMA_COHERENT_POOL
Excerpts from David Rientjes's message of June 11, 2020 3:25 am:
> DMA_REMAP is an unnecessary requirement for AMD SEV, which requires
> DMA_COHERENT_POOL, so avoid selecting it when it is otherwise unnecessary.
>
> The only other requirement for DMA coherent pools is DMA_DIRECT_REMAP, so
> ensure that properly selects the config option when needed.
>
> Fixes: 82fef0ad811f ("x86/mm: unencrypted non-blocking DMA allocations use
> coherent pools")
> Suggested-by: Christoph Hellwig <hch@....de>
> Signed-off-by: David Rientjes <rientjes@...gle.com>
> ---
> kernel/dma/Kconfig | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
Works for me with SME on or off with af7b480103, and with SME off in
abfbb29297. There is some regression with amdgpu and SME between those
two points, I need to check that out too. I haven't tested either before
or after with SEV (which I'm not even sure my system supports).
Regardless, this is a definite improvement.
Tested-by: Alex Xu (Hello71) <alex_y_xu@...oo.ca>
Thanks,
Alex.
Powered by blists - more mailing lists