Looking at the version of this that landed in git I don't think the
read-only handling is entirely clean after this.
- we now normally set the read_only flag from bdrv_open2 when we do
not have the O_RDWR flag set
- but the block drivers also mess with it:
o raw-posix superflously sets it when BDRV_O_RDWR is not in the
open flags
o bochs, cloop, dmg and parallels set it unconditionally given
that they do not support writing at all. But they do not
bother to reject opens without BDRV_O_RDWR
o vvfat as usual is a complete mess setting and clearing it in
various places
- in addition to that bdrv_open2 also sets it after calling itself for
the backing hd which seems superflous
- there also is a now unused bdrv_set_read_only helper to set it from
outside block.c
