On Fri, Feb 17, 2012 at 04:06:19PM +0100, Kai Bankett wrote:

- fixed broken mmi superblock handling, broken by the switch to pagemap,
root_node etc. stuff ;)

+       bh1 = sb_bread(s, s->s_blocksize);
in there looks rather fishy...

qnx6_lfile_checksum() doesn't use superblock argument anymore (used to
be used for conversion, but since you are passing length from the
caller...)

qnx6_longname() - since you are passing superblock anyway, sbi is
redundant...

+       inode->i_blocks = (inode->i_size >> sb->s_blocksize_bits)
+                       << (sb->s_blocksize_bits - 9);
looks too convoluted and if you really want to be pedantic, not 100%
correct - you end up rounding it down, not up.

FWIW, I would probably simply do (inode->i_size + 511) >> 9 and devil
take the rounding effects on big block sizes; it's an approximation
anyway, since you are not counting indirect blocks...

Other than that (and assuming it survives your testing), I'm fine with
that variant.