Re: [PATCH v3 08/15] flattree: Handle unknown tags

[Herve Codina <herve.codina@bootlin.com>]

On Wed, 16 Sep 2026 18:27:53 +1000
David Gibson <david@gibson.dropbear.id.au> wrote:

> On Wed, Sep 16, 2026 at 08:31:30AM +0200, Herve Codina wrote:
> > Hi David,
> > 
> > On Tue, 15 Sep 2026 21:52:24 +1000
> > David Gibson <david@gibson.dropbear.id.au> wrote:  
> [snip]
> > > > This avoid the need for 2 dtbs generated by treegen and also avoid to compare
> > > > binary files which are difficult to analyze when the comparison detects a problem
> > > > due to something broken by some modifications.    
> > > 
> > > We _want_ to understand and test things at the binary byte level.
> > > Debugging differences is a little trickier, but it's really not that
> > > bad - -Odts or fdtdump or dtdiff can be used if/when there's a test
> > > failure.  I really think doing the comparison in binary is preferable
> > > - that's the level at which the behaviour is specified and should be
> > > tested.  
> > 
> > Right, will see what I can do. Probably generating using two specific dtbs
> > as suggested in your first proposal.
> > 
> > Worth noting that each time a dtb will change due to, for instance, a new
> > dtb version (or any other header field update), this test will need to be
> > updated. DTC will generate the dtb with new headers value.  
> 
> If you use the default call for the header in treegen, which always
> uses the latest version, then it should be fine.
> 

Yes indeed. Will give a try.

Best regards,
Hervé