Re: [PATCH v4 00/10] man2: document "new" mount API

[Aleksa Sarai <cyphar@cyphar.com>]

[-- Attachment #1: Type: text/plain, Size: 1821 bytes --]

On 2025-09-21, Askar Safin <safinaskar@gmail.com> wrote:
> Aleksa, thank you! Don't give up. We all need these manpages.
> 
> I see you didn't address some my previous notes.
> 
> * move_mount(2) still says "Mount objects cannot be attached beneath the filesystem root".
> I suggest saying "root directory" or "root" or "root directory of the process" or just "/"
> instead. But you may keep this phrase as is, of course.
> 
> * Docs for FSPICK_NO_AUTOMOUNT in fspick(2) are still wrong. They say that FSPICK_NO_AUTOMOUNT
> affects all components of path. Similar thing applies to mount_setattr(2) and move_mount(2)

Sorry, I last read through your review comments a month ago, I must've
forgotten to make these changes back then. I'll include them in v5.
(It seems I remembered to change the open_tree(2) automount one but
forgot to do it for the others, oops!)

> * open_tree(2) still says:
> > If flags does not contain OPEN_TREE_CLONE, open_tree() returns a file descriptor
> > that is exactly equivalent to one produced by openat(2) when called with the same dirfd and path.
> 
> This is not true if automounts are involved. I suggest adding "modulo automounts". But you may
> keep everything, of course.

Hmmm. As we discussed last time, this sentence is more intended to
indicate that the file descriptor is just a regular open file (with no
dissolve_on_fput() + FMODE_NEED_UNMOUNT magic) rather than the exact
behaviour you get with regards to path lookup.

I would honestly prefer to remove "when called with the same dirfd and
path" rather than add caveats, but I think it makes the sentence less
readable... I'll think about it and try to fix this wording up somehow
for v5.

-- 
Aleksa Sarai
Senior Software Engineer (Containers)
SUSE Linux GmbH
https://www.cyphar.com/

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 265 bytes --]