On Mon, May 12, 2014 at 02:01:14PM -0500, Rob Herring wrote: > On Mon, May 12, 2014 at 1:06 AM, David Gibson > <david@xxxxxxxxxxxxxxxxxxxxx> wrote: > > Add the fdt_address_translate() function to process 'ranges' properties > > to translate addresses from one bus to another. > > Some trailing whitespace errors: > > Applying: libfdt: Add address translation functions > dtc/.git/rebase-apply/patch:87: trailing whitespace. > > dtc/.git/rebase-apply/patch:208: trailing whitespace. > * @inbusoffset: offset of node which defines the address space of the input > dtc/.git/rebase-apply/patch:342: trailing whitespace. > > warning: 3 lines add whitespace errors. Oops. Not sure how I missed those. > [...] > > > > +int fdt_address_translate(const void *fdt, > > + int inbusoffset, const fdt32_t *inaddr, > > + int outbusoffset, fdt32_t *outaddr) > > The translated address in native endian would be more useful, but more > importantly how does one determine the translated address size needed > to convert it? I don't see any reason for the caller to have the > outaddr in be32. While in theory you could need something larger than > a uint64_t, the reality is any >2 cell case is broken anyway. The output isn't necessarily representable in any simple type (e.g. translating from an ISA bus to a parent PCI bus - output is 3 cells). I'm certainly thinking helper functions to translate simple 1 and 2 cell addresses to native integers, but this function is intended to strictly translate between one device tree address representation and another device tree address representation. -- David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Attachment:
pgp7nyFkQoOLP.pgp
Description: PGP signature
