On Sat, 16 Jul 2011 04:17:07 +0200 Stefan Tauner stefan.tauner@student.tuwien.ac.at wrote:
1/4 and 4/4 unchanged. the descriptor patches are now in a reviewable state i think. the only thing left to do from my point of view is OOB checking of offsets we derive from the flash contents (could be garbage).
another thing which could be done is reading the descriptor part of the flash on runtime and parse it then. this would give us access to the softstraps and to the table of supported flash chips. there are some drawbacks to be considered though: - we would need to access the flash in the init function of a programmer. this was probably not considered when initialization was designed. - access rights to the descriptor region must be considered.
if the descriptor region is readable and the layout option is fixed to work for reads too, there is an easy workaround to access the information: read it with flashrom, parse it with ich_descriptor_tool.
the implementation (of runtime dumping) should be straight forward though... should i work on that?