On Thu, 6 Dec 2007, Jörn Engel wrote:
>
> How about shifting and masking _before_ converting to host endianness?
>
> static inline u32 cramfs_offset(struct cramfs_inode *inode)
> {
> return le32_to_cpu(node->namelen_offset >> CRAMFS_NAMELEN_WIDTH);
> }
Stop this idiocy. The code I sent out was correct. You're just making it
worse.
"le32_to_cpu()" is a no-op on little-endian machines. So your change won't
make any difference there. But the point is, that if we want the disk
layout to be the *same* for both big-endian and little-endian, we need to
switch the word as it is loaded from memory, and not do *any* operations
on it before we've done that equalization.
This is not something unusual. It's bog-standard procedure for a lot of
filesystems. And sparse will help you (and would have complained about
your code that tries to shift a little-endian entity before changing it
into a CPU-endian one).
Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
