Hi Tom,
On 31 March 2014 12:59, Tom Rini trini@ti.com wrote:
On Mon, Mar 31, 2014 at 12:38:20PM -0600, Simon Glass wrote:
Hi Tom,
On 31 March 2014 12:30, Tom Rini trini@ti.com wrote:
On Mon, Mar 24, 2014 at 09:51:05AM -0600, Simon Glass wrote:
A large chunk of U-Boot's executable size is the code to process and execute commands. This is reasonable, since commands and scripts are an important part of U-Boot's feature set and provide much of its flexibility.
However, for some applications only a very limited set of commands is required. Where image size is important, it is desirable to be able
to
easily remove unwanted code.
This series introduces a new board_run_command() function which can be used to run a small subset of commands as required by the board (typically load and bootm), thus allowing the rest of the commands to be automatically and reliably dropped from the image using
toolchain
dead code elimination.
This is an interesting concept certainly. Can you also provide a non-trivial example here?
I have a very non-trivial example which is a lot of code, and calls
things
like the SPI layer and do_bootm(). That might be more than you want.
Would it help if I wrote a simple FDT-based boot to a kernel in a
function?
How about both?
For the first case, here is a commit that adds a function
https://chromium.googlesource.com/chromiumos/third_party/u-boot/+/b9e8864294...
For example, this code set up and boots a kernel:
https://chromium.googlesource.com/chromiumos/third_party/u-boot/+/b9e8864294...
and this code deals with checking for a second-stage read-write U-Boot to run:
https://chromium.googlesource.com/chromiumos/third_party/u-boot/+/b9e8864294...
Does that help?
For the second case, I can write something.
Regards, Simon