On Saturday, October 15, 2011 10:56:54 AM Wolfgang Denk wrote: > Dear Prabhakar Lad, > > In message <CA+V-a8sYRZJDZojEpQ55ZGRZ6-- niq0thkvv8e_rtqrrush...@mail.gmail.com> you wrote: > > > I've explained this a number of times recently - there are actually > > > very, very few occasions where "volatile" actually makes sense. > > > > > > Agreed, but I see a piece of code where virtual address are > > > compared. > > > > For example in arch/arm/cpu/arm926ejs/davinci/cpu.c > > In this function static inline unsigned pll_prediv(unsigned pllbase) > > > > and > > > > also in this static inline unsigned pll_postdiv(unsigned pllbase) > > > > Any suggestion on this on how to tackle or let it remain stagnant? > > I cannot see a justification for any of the ""volatile" in this file. > > Of course, all these ugly REG() calls should be converted to proper > use of I/O accessors.
Definitelly ... but I'm not swiping this one, I have enough mess on my hands already ;-) Cheers _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
