* Felipe Balbi <me@xxxxxxxxxxxxxxx> [100309 06:27]: > Hi, > > On Mon, Mar 08, 2010 at 03:04:23PM -0800, Tony Lindgren wrote: > > Hmm yeah thanks for digging into this mystery. Looks like we're > > hitting some Linux initdata limitation. > > > > We can't mark everything as __initconst. Some platform data can > > get dynamically set based on the board type for resources. > > > > So to me it sounds like the only solution is to mark everything > > as __initdata and ignore the (correct) const for now for __initdata. > > > > Or am I missing something? > > sure. I'll cook something. But then it means all ehci_hcd_platform_data > won't be able to be declared as const or that might cause problems as > well later. But sounds like your fix to mark anything going to init.rodata as const is correct since it is rodata. I was originally thinking this is somehow related to __init vs __initconst, but it's for __initconst only, and we should use const there as it's read only. Let me take another look at your patches and make sure everything compiles OK, I don't think we're using __initconst in that many places. Regards, Tony -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
