On Thu, Dec 30, 2021 at 03:07:25AM -0500, Hans Montero wrote:
> We noticed the FIXME comment and we weren't sure exactly what it meant so we
> kept tracing through older versions of `generic_write_checks()`, going as far as
> Linux 2.5.75, before it was implemented for `write_iter()` usage:
>
> if (!isblk) {
> /* FIXME: this is for backwards compatibility with 2.4 */
> if (file->f_flags & O_APPEND)
> *pos = inode->i_size;
> ...
> }
I was able to trace it back farther, to v2.4.14.8 -> v2.4.14.9. And it
looks like the mysterious FIXME is about the block device:
commit 1040c54c3b98ac4f8d91bc313cdc9d6669481da3
Author: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Date: Mon Feb 4 20:33:54 2002 -0800
...
@@ -2765,7 +2876,8 @@ generic_file_write(struct file *file,const char *buf,size_t count, loff_t *ppos)
written = 0;
- if (file->f_flags & O_APPEND)
+ /* FIXME: this is for backwards compatibility with 2.4 */
+ if (!S_ISBLK(inode->i_mode) && file->f_flags & O_APPEND)
pos = inode->i_size;
/*
But as Al Viro pointed out, O_APPEND really can't have any real
meaning for a block device. It would be neat if we could magically
make a 10TB HDD to become as 12TB HDD by writing 2TB using
O_APPEND, but reality doesn't work that way. :-)
It probably makes sense just to remove the FIXME from the current
kernel sources so that future people don't get confused, asking the
same questions you have.
Cheers,
- Ted
