linux-kernel - Re: [PATCH] Rename get_unused_fd_flags to get_unused

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <aCpGePN3w0efNtpr@google.com>
Date: Sun, 18 May 2025 20:43:36 +0000
From: Carlos Llamas <cmllamas@...gle.com>
To: Anish Moorthy <amoorthy@...gle.com>
Cc: linux-kernel@...r.kernel.org
Subject: Re: [PATCH] Rename get_unused_fd_flags to get_unused_fd

On Thu, May 15, 2025 at 05:09:45PM +0000, Anish Moorthy wrote:
> The current name can be misread as having something to do with unused
> *flags*. And without a get_unused_fd() function already floating around,
> it's easy to resolve this by dropping the suffix.

I don't understand the "unused flags" argument. Did you interpret the
current naming as "get the flags not used by a certain fd"?

Note there _used_ to be a get_unused_fd() that would take no arguments
and thus the naming behind this _flags() version (just FYI).

> I'm not sure if there's any appetite for refactors like this: they're tedious
> for sure. I couldn't find any discouragement in the docs though, so I figured
> I'd just post the patch and let it find me.

If it helps, this kind of patches are usually tagged as "treewide:" and
are often implemented using coccinelle scripts.

Also, for this patch in particular I would:
Cc: Al Viro <viro@...iv.linux.org.uk>
Cc: Christian Brauner <brauner@...nel.org>

> diff --git a/drivers/android/binder.c b/drivers/android/binder.c
> index 76052006bd87..e162d92e8c1d 100644
> --- a/drivers/android/binder.c
> +++ b/drivers/android/binder.c
> @@ -4618,7 +4618,7 @@ static int binder_apply_fd_fixups(struct binder_proc *proc,
>  	int ret = 0;
>  
>  	list_for_each_entry(fixup, &t->fd_fixups, fixup_entry) {
> -		int fd = get_unused_fd_flags(O_CLOEXEC);
> +		int fd = get_unused_fd(O_CLOEXEC);
>  
>  		if (fd < 0) {
>  			binder_debug(BINDER_DEBUG_TRANSACTION,

This is the only reason I found this patch (binder), and fwiw the
renaming looks OK to me.

Cheers,
Carlos Llamas