Linux Media Controller development
 help / color / mirror / Atom feed
From: Bryan O'Donoghue <bod@kernel.org>
To: Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
	libcamera-devel@lists.libcamera.org, linux-media@vger.kernel.org
Subject: Re: [ANNOUNCEMENT] Camera & ISP at Linux Plumbers Conference
Date: Thu, 17 Sep 2026 14:47:52 +0100	[thread overview]
Message-ID: <00cc46b2-0729-46d3-8e3b-ea717ec1675e@kernel.org> (raw)
In-Reply-To: <20260914233810.GA2796257@killaraus.ideasonboard.com>

On 15/09/2026 00:38, Laurent Pinchart wrote:
> We can consider scheduling a third topic. Proposals are welcome.

 From my LPC submission:

"Should offline ISPs be implemented in drm/accel"

Where does an offline Camera ISP belong in video4linux or in DRM and why ?

There are two main types of ISP - inline and offline. An inline ISP 
takes sensor input directly and produces some kind of processed image on 
the output. This is a natural fit for v4l2.

An offline ISP takes a memory buffer given to it and processes that 
buffer against supplied parameters. This model can be implemented as a 
video4linux m2m device but it could also be implemented as a DRM device 
and included in the DAG.

Taking Qualcomm SoCs as an example the Camera hardware provides two 
types of ISP. An inline processing engine IFE and an offline processing 
engine based on firmware ICP/HFI.

A use-case emerges to implement the inline ISP in v4l2 and the offline 
in drm/accel.

Some tension with this design would exist as user-space would have to be 
involved to update ISP parameters and process statistics. Another 
tension is including a firmware based system in dma-fencing including 
watchdogging and resetting if required.

There are clear benefits to having a loop of NPU/DSP to Offline ISP to 
Display that drm/accel seems to offer some interesting solutions to.

---
bod

  reply	other threads:[~2026-09-17 13:47 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <UzpC1s2YnBtZIaRCab_Y8AbCecyauDUO4OC7Nhd1-h4TE21aLSghbwDuOKgdwJQugvPMjM4foPHuL-UKnBLD6A==@protonmail.internalid>
2026-09-14 23:38 ` [ANNOUNCEMENT] Camera & ISP at Linux Plumbers Conference Laurent Pinchart
2026-09-17 13:47   ` Bryan O'Donoghue [this message]
2026-09-18  8:08     ` johannes.goede
2026-09-18  9:18     ` Laurent Pinchart
2026-09-19 21:45       ` Bryan O'Donoghue
2026-09-19 22:29         ` Laurent Pinchart
2026-09-21 11:52           ` Bryan O'Donoghue
2026-09-21 16:02             ` Nicolas Dufresne
2026-09-27  9:01   ` johannes.goede
2026-09-27  9:04   ` johannes.goede
2026-09-30 11:18   ` Stefan Klug
2026-09-30 22:01     ` Conor Dooley
2026-10-02 16:14       ` Stefan Klug
2026-10-03 14:56         ` Loic Poulain
2026-10-01 15:15     ` johannes.goede
2026-10-01 20:55       ` Laurent Pinchart
2026-10-02  6:56         ` Sakari Ailus
2026-10-02 18:31           ` Laurent Pinchart
2026-10-06 10:03             ` Sakari Ailus
     [not found]   ` <2e1d628f-8183-4446-9500-9886561f320b@mm-sol.com>
2026-09-30 23:23     ` Laurent Pinchart
2026-10-06  6:30       ` Gjorgji Rosikopulos
2026-10-01  6:19   ` Rishikesh Donadkar
2026-10-01 21:08     ` Laurent Pinchart

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=00cc46b2-0729-46d3-8e3b-ea717ec1675e@kernel.org \
    --to=bod@kernel.org \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=libcamera-devel@lists.libcamera.org \
    --cc=linux-media@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox