On Monday, June 07, 2010 18:55:18 Timur Tabi wrote: > Mike Frysinger wrote: > > On Monday, June 07, 2010 14:31:27 Timur Tabi wrote: > >> In general, if an function initializes only one device, then it should > >> return a negative number if there's an error. If it initializes more > >> than one device, then it should never return a negative number. This > >> is why these functions now return an unsigned integer. > > > > i dont think this is a good idea. either the init funcs should all be > > converted to unsigned int, or they should stay int. doing it piecemeal > > leads to confusion with zero upside. > > I don't want to change all of the functions. For most devices, there's no > way I can test them.
i dont think this is a big deal. plenty of tree wide changes are made and people try their best to break things without actually testing them. > Just because pci_eth_int() is incorrect, that doesn't mean that I can't > make tsec_eth_init() correct. except all the documentation and existing code says "use int", so "correct" is in the eye of the beholder here. i think consistency is more important at this point than an otherwise meaningless value. > > your fixes no way require these to be unsigned int funcs. > > What's the point of making the return value a signed integer if it can > never be a negative number? The reason I changed the type to unsigned int > is to make it very clear that it will never return an error code. the documentation currently states that a negative value is permissible and thus "int" is correct. as for the code that actually reads the result, that is by & large common code, so logic along those lines isnt terribly important. -mike
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot