[patch] Fix bfd_elf_bfd_from_remote_memory() loadbase

[Roland McGrath]

> What I'd like to see is:
> - remove the nonsense PF_R tests.
> - set loadbase using the first header with (p_offset & -p_align) == 0

I'd agree with those changes.  It is probably also appropriate to change
how the later part of the function works (the actual data reading).
Removing the PF_R check from the second loop in the existing code will
lead to a redundant second read of the first page in the ia64 vdso case.
But that is a further refinement.


Thanks,
Roland