Broadcom VC5 DRM Driver Might Soon Be On Its Way To The Mainline Linux Kernel

[phoronix]

Phoronix: Broadcom VC5 DRM Driver Might Soon Be On Its Way To The Mainline Linux Kernel

Eric Anholt believes he is getting quite close to the stage of merging the Broadcom VC5 DRM driver into the mainline Linux kernel tree...

http://www.phoronix.com/scan.php?page=news_item&px=VC5-DRM-Might-Soon-Mainline

[microcode]

Broadcom is setting the example by showing that they can save their (and their customers'!) money by working with upstream, and in the open.

[DrYak]

On the other hand, Video Core runs currently a closed source firmware that handle all the graphics.

Saying that you have an opensource VC driver, is like saying that the openssh you use to connect to some Silicon Graphics render server is opensource :
- an opensource software like openSSH is by itself amazing (just like Eric has done amazing work getting his VC4 / VC5 drivers to work)
- but all the "magic" is still happening in some closed source software running on a separate CPU core (Broadcom has not opened the firmware running on the Video Core)

Currently there is an opensource firmware, though, that can work for some headless uses cases.
(It will initialise the hardware, but is still unable to do any rendering).

[slacka]

On the other hand, Video Core runs currently a closed source firmware that handle all the graphics.

Saying that you have an opensource VC driver, is like saying that the openssh you use to connect to some Silicon Graphics render server is opensource :
- an opensource software like openSSH is by itself amazing (just like Eric has done amazing work getting his VC4 / VC5 drivers to work)
- but all the "magic" is still happening in some closed source software running on a separate CPU core (Broadcom has not opened the firmware running on the Video Core)

Currently there is an opensource firmware, though, that can work for some headless uses cases.
(It will initialise the hardware, but is still unable to do any rendering).

I applaud your idealism, but I sincerely hope you wrote that message on a OpenSPARC or OpenRISC CPU because otherwise you are being a hypocrite. Intel CPUs haven't actually used x86 since the Pentium. From the P6 (Pentium Pro) onward, Intel used a closed source firmware to translate their complex x86 instructions into microcode. Nevermind the issue with CPUs' hidden secure processor code in AMD's PSP and Intel's IME. All of which are also running closed source firmware.

Broadcom is no saint here, but at least they are being somewhat open. I'll take Broadcom's approach any day over Arm with their Mali or PowerVR's.

[Nille]

On the other hand, Video Core runs currently a closed source firmware that handle all the graphics.

But the parts that the OS use is available as Source. Thats way better than what we get from Mali or IM.

[starshipeleven]

- but all the "magic" is still happening in some closed source software running on a separate CPU core (Broadcom has not opened the firmware running on the Video Core)

Lol no.
You make it look like this magic firmware running on the GPU provides anything resembling a graphics API like OpenGL or Vulkan, it's not the case.

The GPU runs its own microcode to do low-level stuff, which is normal for closed hardware. The actual driver (i.e. the thing actually providing a working API to the outside world) running in Linux is using Mesa.

[onicsis]

LoL YeS

From Video Core entry in Wikipedia

On 28 February 2014, on the day of the second anniversary of the Raspberry Pi, Broadcom, together with the Raspberry PI foundation, announced the release of full documentation for the VideoCore IV graphics core, and a complete source release of the graphics stack under a 3-clause BSD license.^[10][11][6]

• Official Broadcom VideoCore IV 3D Architecture Reference: VideoCoreIV-AG100-R.pdf

However, only a minor part of the driver was actually released as open source and all of the actual video acceleration is done using a firmware coded for its proprietary GPU, and which was not open sourced; the entire SoC itself is managed / initialized by a ThreadX-based RTOS that is loaded into the Videocore's VPU during bootup.^[12]

An architectural overview of the VideoCore based system was compiled (based on reverse engineering & patent research) by Herman Hermitage and is available on GitHub

From ThreadX own website

X-WARE IoT PLATFORM
powered by THREADX RTOS

The Industrial Grade solution for deeply embedded IoT devices based on THREADX RTOS, FILEX embedded file system, GUIX embedded GUI, NETX and NETX DUO embedded TCP/IP, and USBX embedded USB. Proprietary, fully supported, containing Absolutely NO open source. At the heart of the X-WARE IoT PLATFORM is the embedded industry's most deployed THREADX Real-time Operating System (RTOS), with well over billion deployments – and counting. 6,203,635,078

X-WARE IoT PLATFORM solutions serve all major embedded IoT markets, and are offered with full source code and no run-time royalties.

Proud to be full closed source not a bit opensource/free software. And we are very proud of that.


It's a matter of time before a complete open stack hardware will be available including CPU, GPU, firmware. Everything.

[onicsis]

Hmmm why :|

---

• Unapproved?

Phoronix do not support open source?

[Tomin]

Hmmm why :|

---

• Unapproved?

Phoronix do not support open source?

Probably they don't support being proud of not being open source.

The moderation system seems to be a bit random sometimes but I suppose the plethora of links in the quote was the cause here.

[Nille]

Proud to be full closed source not a bit opensource/free software. And we are very proud of that.

Sounds for me a bit different. Some Company's don't want FOSS-Software. For Example some Embedded Systems comes with Busybox. This add new requirements for your Deployment/Distribution. The "no open source" Statement is a nice thing for some products, because you know who has all rights on this software.