MCDE: Add frame buffer device driver
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest0 of 10 patches 0 repliesNew
Patches in v1 No applies recorded yet
None of the 10 patches the cover letter announces are in this archive.
Reviews trailers per patch
No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.
Cover letter [PATCH 00/10] MCDE: Add frame buffer device driver
Quoted 43 lines · Alex Deucher
Quoted 3 lines · Marcus LORENTZON
Hi Alex,
Do you have any idea of how we should use KMS without being a "real" drm 3D device? Do you mean that we should use the KMS ioctls on for display driver? Or do you mean that we should expose a/dev/drmXdevice only capable of KMS and no GEM?In this case I was only speaking of using the kms icotls and fbdev emulation for modesetting as your device seems to have a fairly complex display engine. As for 2D/3D/video accel, that's up to you. Each drm driver does it differently depending on how they handle command buffers. Intel and AMD have different sets of ioctls for submitting 2D/3D/video commands from userspace acceleration drivers and a different set of ioctls for memory management.
Quoted 2 lines
What if we were to add a drm driver for 3D later on. Is it possible to have a separate drm device for display and one for 3D, but still share "GEM" like buffers between these devices? It look like GEM handles are device relative. This is a vital use case for us. And we really don't like to entangle our MCDE display driver, memory manager and 3D driver without a good reason. Today they are maintained as independent drivers without code dependencies. Would this still be possible using drm? Or does drm require memory manager, 3D and display to be one driver? I can see the drm=graphics card on desktop machines. But embedded UMA systems doesn't really have this dependency. You can switch memory mamanger, 3D driver, display manager in menuconfig independently of the other drivers. Not that it's used like that on one particular HW, but for different HW you can use different parts. In drm it looks like all these pieces belong together.
No one has done anything like that, but I don't think it would be an issue, you'd just need some sort of way to get buffers in your display driver or your 3D driver, so I'd assume they would depend on your memory manager. Right now the userspace 2D/3D accel drivers all talk to the drm independently depending on what they need to do. Whatever your userspace stack looks like could do something similar, call into one set of ioctls for memory, another set for modesetting, and another for accel. As long as the kernel memory manager is common, you should be able to pass buffer handles between all of them. If you wanted separate memory managers for each, things get a bit trickier, but that's up to you.
Quoted 2 lines
Do you think the driver should live in the "gpu/drm" folder, even though it's not a gpu driver?
gpu is kind of a broad term. It encompasses, 3D, display, video, etc.
I don't think it really matters.Quoted 1 line
Do you know of any other driver that use DRM/KMS API but not being a PC-style graphics card that we could look at for inspiration?
Jordan Crouse submitted some patches for Qualcomm snapdragon a while back although it was mostly a shim for a userspace accel driver. He did implement platform support in the drm however:
http://git.kernel.org/?p=linux/kernel/git/airlied/drm-2.6.git;a=commit;h=dcdb167402cbdca1d021bdfa5f63995ee0a79317Quoted 1 line
Quoted 6 lines
Quoted 1 line
And GEM, is that the only way of exposing graphics buffers to user space in drm? Or is it possible (is it ok) to expose another similar API? You mentioned that there are TTM and GEM, do both expose user space APIs for things like sharing buffers between processes, security, cache management, defragmentation? Or are these type of features defined by DRM and not TTM/GEM?
GEM is not a requirement, it just happens that all the current drm drivers use variants of it for their external memory management interface. However, they are free to implement the memory manager however they like.
Actually, to reinforce your point, the vmwgfx driver doesn't use GEM at all but a TTM based userspace API.
Quoted 196 lines
Quoted 184 lines
Quoted 184 lines
-----Original Message-----
From: Alex Deucher [mailto:alexdeucher at gmail.com]
Sent: den 12 november 2010 16:53
To: Jimmy RUBIN
Cc: linux-fbdev at vger.kernel.org; linux-arm-kernel at lists.infradead.org; linux-media at vger.kernel.org; Linus WALLEIJ; Dan JOHANSSON; Marcus LORENTZON
Subject: Re: [PATCH 00/10] MCDE: Add frame buffer device driverQuoted 3 lines · Jimmy RUBIN
Hi Alex,
Good point, we are looking at this for possible future improvements
but for the moment we feel like
Quoted 1 line
the structure of drm does not add any simplifications for our driver.
We have the display manager (MCDE DSS = KMS) and the memory manager (HWMEM = GEM) that could be migrated to drm framework. But we do not have drm drivers for 3D hw and this also makes drm a less obvious choice at the moment.
Quoted 1 line
You don't have to use the drm strictly for 3D hardware. historically that's why it was written, but with kms, it also provides an interface for complex display systems. fbdev doesn't really deal properly with multiple display controllers or connectors that are dynamically re-routeable at runtime. I've seen a lot of gross hacks to fbdev to support this kind of stuff in the past, so it'd be nice to use the interface we now have for it if you need that functionality. Additionally, you can use the shared memory manager to both the display side and v4l side. While the current drm drivers use GEM externally, there's no requirement that a kms driver has to use GEM. radeon and nouveau use ttm internally for example. Something to consider. I just want to make sure people are aware of the interface and what it's capable of.
Alex
Quoted 7 lines
Jimmy
-----Original Message-----
From: Alex Deucher [mailto:alexdeucher at gmail.com]
Sent: den 10 november 2010 15:43
To: Jimmy RUBIN
Cc: linux-fbdev at vger.kernel.org; linux-arm-kernel at lists.infradead.org; linux-media at vger.kernel.org; Linus WALLEIJ; Dan JOHANSSON
Quoted 3 lines
Subject: Re: [PATCH 00/10] MCDE: Add frame buffer device driver
On Wed, Nov 10, 2010 at 7:04 AM, Jimmy Rubin
Quoted 1 line · <[email protected]> wrote:
Quoted 1 line
These set of patches contains a display sub system framework (DSS)
which is used to
Quoted 13 lines
Quoted 3 lines
implement the frame buffer device interface and a display device framework that is used to add support for different type of displays such as LCD, HDMI and so on.
For complex display hardware, you may want to consider using the drm kms infrastructure rather than the kernel fb interface. It provides an API for complex display hardware (multiple encoders, display controllers, etc.) and also provides a legacy kernel fb interface for compatibility. See:
Documentation/DocBook/drm.tmpldrivers/gpu/drm/
in the kernel tree.Quoted 12 lines
Alex
Quoted 10 lines
The current implementation supports DSI command mode displays.
Below is a short summary of the files in this patchset:
mcde_fb.c
Implements the frame buffer device driver.
mcde_dss.c
Contains the implementation of the display sub system framework(DSS).
Quoted 4 lines
Quoted 4 lines
This API is used by the frame buffer device driver.
mcde_display.c
Contains default implementations of the functions in the displaydriver
Quoted 1 line
Quoted 1 line
API. A display driver may override the necessary functions to
function
Quoted 1 line
Quoted 1 line
properly. A simple display driver is implemented in display-
generic_dsi.c.Quoted 6 lines
Quoted 6 lines
display-generic_dsi.c
Sample driver for a DSI command mode display.
mcde_bus.c
Implementation of the display bus. A display device is probed whenboth
Quoted 1 line
Quoted 1 line
the display driver and display configuration have been registered
with
Quoted 4 lines
Quoted 4 lines
the display bus.
mcde_hw.c
Hardware abstraction layer of MCDE. All code that communicatesdirectly
Quoted 4 lines
Quoted 4 lines
with the hardware resides in this file.
board-mop500-mcde.c
The configuration of the display and the frame buffer device ishandled
Quoted 3 lines
Quoted 3 lines
in this file
NOTE: These set of patches replaces the patches already sent out for
review.
Quoted 27 lines
Quoted 27 lines
RFC:[PATCH 1/2] Video: Add support for MCDE frame buffer driver RFC:[PATCH 2/2] Ux500: Add support for MCDE frame buffer driver
The old patchset was to large to be handled by the mailing lists.
Jimmy Rubin (10):
MCDE: Add hardware abstraction layer
MCDE: Add configuration registers
MCDE: Add pixel processing registers
MCDE: Add formatter registers
MCDE: Add dsi link registers
MCDE: Add generic display
MCDE: Add display subsystem framework
MCDE: Add frame buffer device driver
MCDE: Add build files and bus
ux500: MCDE: Add platform specific dataarch/arm/mach-ux500/Kconfig | 8 + arch/arm/mach-ux500/Makefile | 1 + arch/arm/mach-ux500/board-mop500-mcde.c | 209 ++ arch/arm/mach-ux500/board-mop500-regulators.c | 28 + arch/arm/mach-ux500/board-mop500.c | 3 + arch/arm/mach-ux500/devices-db8500.c | 68 + arch/arm/mach-ux500/include/mach/db8500-regs.h | 7 + arch/arm/mach-ux500/include/mach/devices.h | 1 + arch/arm/mach-ux500/include/mach/prcmu-regs.h | 1 +Quoted 8 lines
Quoted 8 lines
arch/arm/mach-ux500/include/mach/prcmu.h | 3 + arch/arm/mach-ux500/prcmu.c | 129 ++ drivers/video/Kconfig | 2 + drivers/video/Makefile | 1 + drivers/video/mcde/Kconfig | 39 + drivers/video/mcde/Makefile | 12 + drivers/video/mcde/display-generic_dsi.c | 152 ++ drivers/video/mcde/dsi_link_config.h | 1486++++++++++++++
Quoted 2 lines
Quoted 2 lines
drivers/video/mcde/mcde_bus.c | 259 +++ drivers/video/mcde/mcde_config.h | 2156++++++++++++++++++++
Quoted 5 lines
Quoted 5 lines
drivers/video/mcde/mcde_display.c | 427 ++++ drivers/video/mcde/mcde_dss.c | 353 ++++ drivers/video/mcde/mcde_fb.c | 697 +++++++ drivers/video/mcde/mcde_formatter.h | 782 ++++++++ drivers/video/mcde/mcde_hw.c | 2528++++++++++++++++++++++++
Quoted 28 lines
Quoted 28 lines
drivers/video/mcde/mcde_mod.c | 67 + drivers/video/mcde/mcde_pixelprocess.h | 1137 +++++++++++ include/video/mcde/mcde.h | 387 ++++ include/video/mcde/mcde_display-generic_dsi.h | 34 + include/video/mcde/mcde_display.h | 139 ++ include/video/mcde/mcde_dss.h | 78 + include/video/mcde/mcde_fb.h | 54 + 31 files changed, 11248 insertions(+), 0 deletions(-) create mode 100644 arch/arm/mach-ux500/board-mop500-mcde.c create mode 100644 drivers/video/mcde/Kconfig create mode 100644 drivers/video/mcde/Makefile create mode 100644 drivers/video/mcde/display-generic_dsi.c create mode 100644 drivers/video/mcde/dsi_link_config.h create mode 100644 drivers/video/mcde/mcde_bus.c create mode 100644 drivers/video/mcde/mcde_config.h create mode 100644 drivers/video/mcde/mcde_display.c create mode 100644 drivers/video/mcde/mcde_dss.c create mode 100644 drivers/video/mcde/mcde_fb.c create mode 100644 drivers/video/mcde/mcde_formatter.h create mode 100644 drivers/video/mcde/mcde_hw.c create mode 100644 drivers/video/mcde/mcde_mod.c create mode 100644 drivers/video/mcde/mcde_pixelprocess.h create mode 100644 include/video/mcde/mcde.h create mode 100644 include/video/mcde/mcde_display-generic_dsi.h create mode 100644 include/video/mcde/mcde_display.h create mode 100644 include/video/mcde/mcde_dss.h create mode 100644 include/video/mcde/mcde_fb.hQuoted 8 lines
Quoted 7 lines
Quoted 2 lines
Quoted 2 lines
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
Quoted 4 lines
Quoted 3 lines
the body of a message to majordomo at vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-fbdev" in the body of a message to majordomo at vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Signature · 2 lines
Earthling Michel D?nzer | http://www.vmware.com Libre software enthusiast | Debian, X and DRI developer