> On 10/30/2012 01:09 AM, Jayachandran C. wrote: > [...] > > > > > We had seen the same issue here, and worked around it the same way > > (i.e use dynamic allocation for the backing dev structures). > > > > I ran across a similar issue in using built-in DTB (basically, kernel > > data address does not work for virt_to_phys/phys_to_virt in 64-bit > > when the load address is in CKSEG0). There I did something like this: > > > > ptr = phys_to_virt(__pa(kernel_data_ptr)); > > > > This works since __pa knows about CKSEG0 addresses in 64bit. > > > > > > Really the proper fix is to make virt_to_phys() work. This isn't the only case where we have seen failures due to this issue: > > > http://www.linux-mips.org/archives/linux-mips/2011-09/msg00029.html > > I fixed it like this... > > In io.h: > > static inline unsigned long virt_to_phys(volatile const void *address) > { > return __pa(address); > } > > Really this needs to be pushed upstream by somebody. > Thanks, this makes sense and seems a good fix. I had some problems with twisty includes because page.h includes io.h. Did you move the definition of __pa also? Anyway, if this can be pushed to master and merged to linux-stable I'll pick it up. Jake
