On May 25, 2012, at 4:00 PM, Tom Metro wrote: > Federico Lucifredi wrote: >>> Also, on the topic of openness, you should note that the Raspberry Pi >>> isn't an open design, and uses a proprietary CPU, possibly requiring >>> proprietary binary-only kernel drivers. >> >> What do you mean by proprietary CPU here? It is a Broadcom SoC, but I >> don't see what's more proprietary there than a Freescale or Ti >> chip... > > It may not be any worse or better. I'm just extrapolating based on what > you hear about Broadcom chips used in routers. In the latter case it is > well documented that the APIs for various bits of the chipset are kept > proprietary, such that you need to use binary drivers provided by the > chipset vendor.
I did check, and indeed there is a binary blob (wrapped in a F/OSS driver shim) to access the video functionality of the SoC. I have to agree with you Tom… this is undesirable. Seems to be the usual evil of video, nothing new - but moved into the SoC as part of the integration. > (This is one of the reasons you can't run FreeBSD on > consumer routers.) Do you have a specific example? I wonder if there is more than video that is proprietary (and FreeBSD would run headless… I wonder where is the problem getting it to run on an ARM chip… probably more than the SoC in that case. > I'm assuming the situation will be the same with the Broadcom chip used > on the Raspberry Pi. They'll provide a binary driver that you can freely > copy, so it'll be possible to port over other Linux distributions (as > we've seen happen) that can work with the same kernel as the vendor > supplied distribution, but if you want to do something more obscure, > you'll be out of luck. Looks like they supply a blob to integrate with your distro. I did notice that they have published the schematics for the board, so it *may* qualify as Open Source Hardware… we shall see, I don't see the logo on their site, so I won't make hte claim for them. Best -F _________________________________________ -- "'Problem' is a bleak word for challenge" - Richard Fish (Federico L. Lucifredi) - flucifredi at acm.org - GnuPG 0x4A73884C _______________________________________________ Hardwarehacking mailing list [email protected] http://lists.blu.org/mailman/listinfo/hardwarehacking
