On Thu, Aug 4, 2011 at 8:13 AM, David Miller <davem@xxxxxxxxxxxxx> wrote: > In fact this points out that of_alias_scan() was only placed into one > of the two FT unflattening sites, it needs to go into both as well as > places that build the device tree from real OpenFirmware calls. And > that makes an even stronger argument for adding an allocator function > pointer arg to this thing. No argument from me here. However, I completely broke sparc by applying it so I needed to post an ASAP fix with the least possible impact. A followup patch for v3.2 would have be written to generalize alias parsing for v3.2 On Thu, Aug 4, 2011 at 9:17 AM, David Miller <davem@xxxxxxxxxxxxx> wrote: [...] > So even though it is implicitly creating properties, it doesn't update > the in-kernel device tree nor does it update the OpenFirmware > etc. copy. > > This behvaior is undesirable and makes no sense. This IMX serial > driver merely wants increasing incrementing indexes for the serial > nodes it finds, and there are much better ways to do that. Rolling the assigned ids into the alias list was a convenient & cheap way of making sure the id assigned would be unique, but fair enough. Shawn and I will talk about implementing a better solution. [...] > This stuff needs more work, I would like to suggest that you just > revert at this point. Very well, I'll prepare a revert branch and ask Linus to pull. g. -- Grant Likely, B.Sc., P.Eng. Secret Lab Technologies Ltd. -- To unsubscribe from this list: send the line "unsubscribe linux-next" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
