Group: Hardware/TODO/DUPLICATE INFO/Processors
(add HDCP section) |
m (→About HDMI, HDCP and DVI) |
||
| Line 208: | Line 208: | ||
The [http://en.wikipedia.org/wiki/High-bandwidth_Digital_Content_Protection Wikipedia Page on HDCP] shows that HDCP is completely pointless, with the master key having been cracked, based on flaws that were shown to exist as far back as 2001. HDCP is about control (failed attempts thereof) and its only effect has been to cause people to spend more money. | The [http://en.wikipedia.org/wiki/High-bandwidth_Digital_Content_Protection Wikipedia Page on HDCP] shows that HDCP is completely pointless, with the master key having been cracked, based on flaws that were shown to exist as far back as 2001. HDCP is about control (failed attempts thereof) and its only effect has been to cause people to spend more money. | ||
| − | Supporting HDCP outputs, such as DVI and HDMI, on a FSF Hardware-endorsed device, is therefore somewhat fraught. The benefits of digital output are clear, but it is only in combination with encryption keys and software (which the HDCP Cartel are never going to provide!) that it is possible to decrypt the HDCP streams. This really needs to be seriously discussed. | + | Supporting HDCP outputs, such as DVI and HDMI, on a FSF Hardware-endorsed device, is therefore somewhat fraught. The benefits of digital output are clear, but it is only in combination with encryption keys and software (which the HDCP Cartel are never going to provide!) that it is possible to decrypt the HDCP streams. So, on the one hand you have the problem that providing a DVI or HDMI interface appears to endorse HDCP Cartel business practices, but on the other, can you seriously blame the USB bus or the PCI bus for that matter if manufacturers place proprietary hardware with non-free firmware or libraries on it? |
| + | |||
| + | This really needs to be seriously discussed. | ||
Revision as of 17:13, 22 January 2011
Contents
Introduction
The purpose of this page is to allow CPUs to be evaluated for suitability for use in FSF-endorsed hardware. Presence or otherwise of a CPU on this page DOES NOT imply that it is an exhaustive list, nor does it imply that each and every CPU is or is not FSF-endorsed. Certain CPUs for example require proprietary libraries for their 3D Graphics, but over time a Free Software compliant alternative may become available, or may be reverse-engineered, in which case CPUs that previously did not qualify suddenly would.
So to be clear: this page is for the PURPOSES of evaluating CPUs, to store research about their present status, NOT to specifically or categorically state "this CPU is endorsed, period".
Processors, types and pricing (appx)
| Manufacturer | Model | Type | Speed | Price |
|---|---|---|---|---|
| Samsung | S5PV210 | ARM Cortex A8 | 1ghz | $14 |
| Samsung | Orion | ARM Cortex A9 | 1ghz | ?$45? |
| TI | AM3715 | ARM Cortex A8 | 1ghz | $24 |
| TI | AM3703 | ARM Cortex A8 | 1ghz | $17 |
| TI | AM3892 | ARM Cortex A8 | 1.5ghz | n/a |
| TI | AM3894 | ARM Cortex A8 | 1.5ghz | $43 |
| TI | OMAP4400 | ARM Cortex A9 | 1ghz | ?$45? |
| Ingenix | jz4760 | MIPS32 (x1) | 700mhz | $7 |
| NXP | PNX8490 | ARM Cortex A9 | 800mhz | n/a |
| NXP | PNX8491 | ARM Cortex A9 | 800mhz | n/a |
| TI | DM3730 | ARM Cortex A8 | 1ghz | $30 |
| TI | DM3725 | ARM Cortex A8 | 1ghz | $27 |
Processor, Compiler, Links, Status
| Model | Links | 3D | Treacherous | DVI or HDMI | Max LCD |
|---|---|---|---|---|---|
| S5PV210 | datasheet | PowerVR | e-fuse (*1) | HDMI | XGA (1024x768) |
| Orion | PowerVR | TBD | HDMI | not known | |
| AM3715 | PowerVR | No (*2) | neither (*4) | 2048x2048 (*5) | |
| AM3703 | SIMD (NEON) | No (*2) | neither (*4) | 2048x2048 (*5) | |
| AM3892 | SIMD (NEON) | No (*2) | neither (*4) | 2048x2048 (*5) | |
| AM3894 | PowerVR | No (*2) | neither (*4) | 2048x2048 (*5) | |
| OMAP4400 | PowerVR | No (*2) | HDMI (*3) | 2048x2048 (*5) | |
| jz4760 | linux-mips | SIMD FPU (XBurst) | No | neither | 1280x720 |
| PNX8490 | product brief | SIMD (NEON) | TBD | HDMI (*3) | none! |
| PNX8491 | product brief | PowerVR | TBD | HDMI (*3) | none! |
| DM3730 | datasheet | PowerVR | No | neither (*4) | 2048x2048 (*5) |
| DM3725 | datasheet | none | No | neither (*4) | 2048x2048 (*5) |
Notes
- (*1) It is possible to effectively disable any opportunity for the S5PC110 and S5PV210 from being "Treacherous", by blowing an e-fuse. Once "normal booting" has been selected, it is not possible to enable "secure" (Treacherous) booting:
6.1 OVERVIEW OF BOOTING SEQUENCE
S5PC110 consists of 64KB ROM and 96KB SRAM as internal memory. For booting, internal 64KB ROM and
internal 96KB SRAM regions can be used. S5PC110 boots from internal ROM to enable secure booting, which
ensures that the image cannot be altered by unauthorized users. To select secure booting or normal booting,
S5PC110 should use e-fuse information. This information cannot be altered after being programmed.
- (*2) TI CPUs with hardware motherboard configuration, to prevent Treachery. discussion here
- (*3) HDMI contains HDCP (encrypted HD content protection). This is Bad (tm).
- (*4) The AM37xx, AM38xx and DM37xx series can all have an external DVI-D IC added which provides at least a way to connect a Digital Monitor.
- (*5) The AM and DM series can all support 2048x2048 but only at really quite low refresh rates. Some more "standard" sizes include 1280x800 (60hz), 1400x1050 (50hz) and 1280x720 (60hz) etc. etc.
About Treacherous Computing
Treacherous Computing starts with the processor, by making it impossible, right from the startup of the CPU, to gain control over what is run on the device. This practice is lauded and loved by the Industry, as once you are running "secure" bootloader code, especially when blown into an on-board ROM, regaining control of the device that you have purchased is somewhat challenging.
Should any device with a Treacherous-capable CPU seek FSF Hardware Endorsement, it would be... unfortunate if that device was then purchased for the purposes of reprogramming the secure bootloader area (if the CPU had one), thus preventing and prohibiting users from being able to use that FSF Hardware-endorsed device! Thus, logically, it makes sense to ensure that the device can at least always have its Operating System replaced at any time, starting by ensuring that any Treacherous "Secure bootloader" capability is non-existent, impossible or can be permanently disabled.
At least, then, if the OS on that FSF Hardware Endorsed device is ever replaced with a Treacherous OS and the device re-sold, at least the OS can, again, be replaced with a Free Software Compliant OS of the new owner's choice.
About HDMI, HDCP and DVI
HDCP stands for "High-bandwith Digital Content Protection" (as if your "contentment" needs "protection" at high bandwidth...) [tv.about.com/od/hdtv/a/hdmidvihdcp.htm This article] gives a good layman's overview whilst also making it bluntly clear that it's a complete pain and that it is entirely driven by greed. Your inconvenience - and your useless purchase costing thousands - is not the problem or the concern of the Copyright Cartels.
The Wikipedia Page on HDCP shows that HDCP is completely pointless, with the master key having been cracked, based on flaws that were shown to exist as far back as 2001. HDCP is about control (failed attempts thereof) and its only effect has been to cause people to spend more money.
Supporting HDCP outputs, such as DVI and HDMI, on a FSF Hardware-endorsed device, is therefore somewhat fraught. The benefits of digital output are clear, but it is only in combination with encryption keys and software (which the HDCP Cartel are never going to provide!) that it is possible to decrypt the HDCP streams. So, on the one hand you have the problem that providing a DVI or HDMI interface appears to endorse HDCP Cartel business practices, but on the other, can you seriously blame the USB bus or the PCI bus for that matter if manufacturers place proprietary hardware with non-free firmware or libraries on it?
This really needs to be seriously discussed.
