Re: [PATCH 04/10] fhandle: create do_filp_path_open() helper

[Al Viro <viro@zeniv.linux.org.uk>]

On Wed, Sep 10, 2025 at 03:49:21PM -0600, Thomas Bertschinger wrote:
> This pulls the code for opening a file, after its handle has been
> converted to a struct path, into a new helper function.
> 
> This function will be used by io_uring once io_uring supports
> open_by_handle_at(2).

Not commenting on the rest of patchset, but...

Consider the choice of name NAKed with extreme prejudice.  "filp"
thing should die; please, do not introduce more instances.

It has crawled out of Minix guts, where AST had been tasteless enough
to call a structure the represents an open file (which is called struct
file on all Unices, Linux included) "struct filp" instead, the identifier
standing for "file and position", nevermind that he did include more
state than that - the damn thing (in OSD&I appendix) is
struct filp {
  mask_bits filp_mode;		/* RW bits, telling how file is opened */
  int filp_count;		/* how many file descriptors share this slot? */
  struct inode *filp_ino;	/* pointer to the inode */
  file_pos filp_pos;		/* file position */
}


Linus used "struct file" from the very beginning; unfortunately, if you
grep for filp in 0.01 you'll see a plenty of those - in form of
0.01:fs/file_dev.c:int file_write(struct m_inode * inode, struct file * filp, char * buf, int count)
as well as
0.01:fs/ioctl.c:        struct file * filp;
0.01:fs/ioctl.c:        if (fd >= NR_OPEN || !(filp = current->filp[fd]))
which was both inconsistent *and* resembling hungarian notation just
enough to confuse (note that in the original that 'p' does *NOT* stand for
"pointer" - it's "current IO position").  Unfortunately, it was confusing
enough to stick around; at some point it even leaked into function names
(filp_open(); that one is my fault - sorry for that brainfart).

Let's not propagate that wart any further, please.  If you are opening a file,
put "file" in the identifier.