[U-Boot] Question: Full benifit of the .itb

Simon Glass sjg at chromium.org
Tue Sep 1 01:13:35 CEST 2015


+Lukasz and a few others

Hi Brian,

On 31 August 2015 at 16:38, <Brian_Brelsford at dell.com> wrote:
>
> Dell Customer Communication
>
>
>
> I have a question  with respect to the .itb as a complete bootable/execution entity.
>
>
>
> OUR PAST  DESIGNs:
>
> Our past home grown solutions,    much like fit update,    is a binary that contains the various components (meaning kernel, root file system, other stuff).
>
> 1)      They are copied into hard coded flash locations,   bootcmd then does mmc read of kernel,   then bootm,   kernel goes to hard coded location to get the rootfs, etc.
>
> 2)      We also have a  system that copy’s the kernel and rootfs as files into a ext4 file system,  bootcmd does ext4load of kernel,   then bootm,  kernel knows of file  /rootfs,  etc.
>
>
>
>
>
> FULL BENEFIT OF .itb
>
> It would seem that the benefit of having everything in the .itb    (Kernel, rootfs, dtb, etc).
>
> It can be copied as a single blob, so fwupdate is really just a single copy to your flash device.
>

Yes that's right.

>
>
>
>
> QUESTIONs:
>
> Is the full intent of the .itb  is to leave it all together,  kernel, dtb, rootfs, other stuff?
>
> Pull the itb into memory, boot the kernel, it knows how to find the rootfs, etc.
>
> OR
>
> Is the .itb a kernel/dtb ,  then via  bootargs we tell the kernel where rootfs is?
>
> Meaning it is not part of the .itb.

The second option. You can put a ramdisk in the .itb (FIT) but
typically that is just used to get the boot started, and you then use
a root disk on another device, with bootargs telling the kernel where
to find it.

>
>
>
> What the open source industry doing?
>
>
>
>
>
>
>
> If I am totally off base please feel free to correct me.
>
>
>
> Thank you for your time.
>
> Brian Brelsford
>

Regards,
Simon


More information about the U-Boot mailing list