[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAN3W6UUs5TpJk7yvAcvnrH+5KgmnTfF9spQ=kCW13P2JL+rxcA@mail.gmail.com>
Date: Mon, 20 Jan 2025 12:09:20 +0530
From: Mukesh Ojha <mukesh.ojha@....qualcomm.com>
To: linux-remoteproc@...r.kernel.org
Cc: mathieu.poirier@...aro.org, andersson@...nel.org,
linux-kernel@...r.kernel.org
Subject: Re: Query: DMA device assigned to remoteproc usage by Linux
On Mon, Dec 23, 2024 at 11:27 PM Mukesh Ojha
<mukesh.ojha@....qualcomm.com> wrote:
>
> Hi All,
>
> Wanted to check if we have encountered remoteproc use case where a device
> with dma is assigned to a remoteproc with its own streamid and iommu
> translation context. This DMA device can have a selective DMA range
> within the remoteproc carveout memory that needs to be iommu mapped
> before the remoteproc is up.
>
> Do we have any example in remoteproc that handles such scenario where
> device dma (assigned to remoteproc) has its iommu setup by Linux (since
> the iommu is under its control) ?
>
> I was exploring some of the remoteproc drivers but did not find anything
> close to DMA use case and from where remoteproc should get this ddr range ?
>
> device tree ? or rsc table ? if rsc table then which among the below ?
Resurfacing this in case it was missed during the New Year vacation period.
-Mukesh
>
> enum fw_resource_type {
> RSC_CARVEOUT = 0,
> RSC_DEVMEM = 1,
> RSC_TRACE = 2,
> RSC_VDEV = 3,
> RSC_LAST = 4,
> RSC_VENDOR_START = 128,
> RSC_VENDOR_END = 512,
>
>
> -Mukesh
Powered by blists - more mailing lists