Hi Frank,
please do not drop the ML from your replies, thanks.
On 08/04/2014 12:44, Frank Ihle wrote:
Hello Stefano
Stefano Babic sbabic@denx.de 07.04.14 18.07 Uhr >>>
Hi Frank,
On 07/04/2014 15:49, Frank Ihle wrote:
Hi U-Boot,
you recently gave me the hint about the Falcon Mode for my Fast Booting investigations. But I think might got it wrong, could somebody correct me if so?:
- According to the "Falcon Mode" Pdf of October 2013
(http://www.denx.de/wiki/pub/U-Boot/MiniSummitELCE2013/2013-ELCE-U-Boot-Falco...)
The Falcon mode is just about to chose 2 different execution modes during startup - one that starts Linux directly (and skips u-boot) and the other for instance to make updates, or failsafe mode (and starts u-boot). Or does it even contain more functionality ?
The second starts U-Boot as you are used to.
agreed
- According to the "Minimizing boot times using U-Boot Falcon Mode"
Pdf of July 2012 (http://schedule2012.rmll.info/IMG/pdf/LSM2012_UbootFalconMode_Babic.pdf) it's mentioned that this can only run with boards, which are able to work with SPL. I guess here SPL means Second Programm Bootloader. I'm not an expert in this area, but wouldn't it be possible to run every board with an SPL, since it's just Software like a Bootloader?
Of course, but not all boards supported in mainline were converted to use SPL, and for this reason not all boards support out of the box Falcon Mode.
So this "part" of including more and more boards is up to the U-Boot team, not of the Board developer ? ( except one would like to do this by himself .. )
No. Everybody is free to add support if he needs sending patches to ML for review. Generally, SPL support is done on SOC-layer and the board developer decides himself how the board boots, in case there is more as one option. Most boards were converted to SPL, most not. Feel free to send a patch for your board.
And if not what is the "real" eye of the needle here ? I'm just asking because I didn't get the problem here.
It is simply a snapshot of the current U-Boot mainline. The switch between Linux/U-Boot is implemented in SPL, and this makes sense because SPL is much smaller that the whole U-Boot, dropping also the time that a SOC needs to load an image from the storage media.
That's what I understood too. I gotta say, this sounds like a working alternative to skipping U-Boot completely when it comes too keep the boot times at their lowest, I like that implementation.
Kind Regards,
Frank
Best regards, Stefano Babic