On Tue, Dec 17, 2013 at 01:04:35PM +0000, Russell King - ARM Linux wrote: > On Tue, Dec 17, 2013 at 01:54:35PM +0100, Tomasz Figa wrote: > > On Tuesday 17 of December 2013 13:45:06 Thierry Reding wrote: > > > I fail to see how that would eliminate the problem with the types. That > > > said I don't actually see sparse complaining about any type mismatches. > > > That's probably because the various macros implicitly cast to u32. > > > > Well, in BE variant you would read the register using __raw_readl() into > > a __be32 and then get an u32 from be32_to_cpu() and return it. Similarly > > for writes > > __raw_readl() returns a u32, so you'll get a warning trying to assign a > u32 to a __be32. If sparse doesn't complain about the original code here, does that mean we have a bug that should be fixed? > We do have ioread32() and ioread32be() which do the appropriate conversion, > as well as the write versions too. They both include the barrier if you're > overly concerned about that. A few years ago GregKH commented in response to a patch that ioread*() weren't supposed to be used for memory-mapped only devices. The original purpose apparently was to allow drivers to work with both I/O and memory-mapped devices. Thierry
Attachment:
pgpcNu9EhRUcJ.pgp
Description: PGP signature
