On Sun, 13 May 2018, Andy Shevchenko wrote: > On Sun, May 13, 2018 at 9:09 PM, Thomas Gleixner <t...@linutronix.de> wrote: > > On Thu, 3 May 2018, Andy Shevchenko wrote: > >> #ifdef CONFIG_X86_64 > >> > >> -build_mmio_read(readq, "q", unsigned long, "=r", :"memory") > >> -build_mmio_read(__readq, "q", unsigned long, "=r", ) > >> -build_mmio_write(writeq, "q", unsigned long, "r", :"memory") > >> -build_mmio_write(__writeq, "q", unsigned long, "r", ) > >> +build_mmio_read(readq, "q", unsigned long long, "=r", :"memory") > >> +build_mmio_read(__readq, "q", unsigned long long, "=r", ) > >> +build_mmio_write(writeq, "q", unsigned long long, "r", :"memory") > >> +build_mmio_write(__writeq, "q", unsigned long long, "r", ) > > > > What's wrong with u64 which we use for expressing io access to a 64bit wide > > resource? > > Same answer as per v1, i.e. I would like to be consistent with other > types in this file (unsigned int for readl() and similar for the > rest). > If we would need them, we might change at once for all accessors.
I don;t think we need to fixup everything in one go. Having the patch which addresses the issue at hand first using u64 makes a lot of sense on its own. Changing the other instances can be done as a follow up patch. Having explicit with types for such kind of accessors makes a lot of sense. Thanks, tglx