[RFC] flashrom and coreboot tables

[Carl-Daniel Hailfinger]

I know that marking a mail as [RFC] is the best way to get no answers. I
wonder if that's the best way to get controversial patches in. Hmmm...

Richard Smith found a design limitation in flashrom the hard way: By
trying to work around it and getting a nice segfault.

Right now, flashrom assumes that if a coreboot table exists, the board
NEEDS a board specific enable function. If there is none (because it is
not needed), flashrom will complain.

Three options are available to solve this:
1. Leave the code as-is and confuse developers and users alike.
2. Add info to the coreboot table that the board doesn't need any
special flash enabling.
3. Provide dummy functions for every board supported by coreboot.

What do you prefer?

Regards,
Carl-Daniel

[Peter Stuge]

Carl-Daniel Hailfinger wrote:
> Right now, flashrom assumes that if a coreboot table exists, the
> board NEEDS a board specific enable function. If there is none
> (because it is not needed), flashrom will complain.
>
> Three options are available to solve this:
> Leave the code as-is and confuse developers and users alike.
> Add info to the coreboot table that the board doesn't need any
>
> special flash enabling.
> 3. Provide dummy functions for every board supported by coreboot.

4. Remove the table lookup completely.
5. Make flashrom continue silently instead of complaining.

I prefer either of 4 or 5 over 1-3.

//Peter

[Stefan Reinauer]

Peter Stuge schrieb:
> Carl-Daniel Hailfinger wrote:
> > Right now, flashrom assumes that if a coreboot table exists, the
> > board NEEDS a board specific enable function. If there is none
> > (because it is not needed), flashrom will complain.
> >
> > Three options are available to solve this:
> > Leave the code as-is and confuse developers and users alike.
> > Add info to the coreboot table that the board doesn't need any
> >
> > special flash enabling.
> > 3. Provide dummy functions for every board supported by coreboot.
> Remove the table lookup completely.
> Make flashrom continue silently instead of complaining.
>
> I prefer either of 4 or 5 over 1-3.

Removing the check completely is non-sense. Not complaining in case 
there is no entry sounds like the way to go though. I think we should 
focus on this once flashrom 1.0 has been released.

Stefan