Re: [PATCH v5] grantpt.3: no-op on modern glibc and other UNIXes

[Alejandro Colomar <alx@kernel.org>]

[-- Attachment #1.1: Type: text/plain, Size: 2877 bytes --]

On 2023-07-16 13:55, наб wrote:
> FreeBSD, OpenBSD, and Linux (/dev/ptmx) do all intialisation in open(2),
> and grantpt(3) is a no-op (that checks whether the fd is a pty, except on
> musl).
> 
> The illumos gate and NetBSD do a ioctl (and, indeed, illumos-gate commit
>  facf4a8d7b59fde89a8662b4f4c73a758e6c402c ("PSARC/2003/246 Filesystem
>   Driven Device Naming"), which kills pt_chmod, notes that it's been
>  "6464196 bfu should remove pt_chmod, obsoleted by /dev filesystem").
> 
> glibc 2.33 completely kills BSD PTY support on Linux
> (Debian hasn't configured with them on any architecture since 2007:
>    https://bugs.debian.org/338404
>  and even earlier on some arches; they're really just trivia under
>  Linux ‒ this may be better served stuffed into HISTORY as an explainer
>  for the SIGCHLD thing, since regardless of the "version", the behaviour
>  is well-defined and consistent).
> 
> Cc: Jakub Wilk <jwilk@jwilk.net>
> Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
> ---
> I read it but didn't really understand what you were saying, since
> you're on record as a text‒text‒text liker.
> You can trivially continue the lines with \c like the below, but
>   "no-op, with permissions ... on Linux, or an ioctl(2)."
> would probably also work just as well,
> and I leave that to your editorial sensibilities.
> 
>  man3/grantpt.3 | 18 ++++++++----------
>  1 file changed, 8 insertions(+), 10 deletions(-)
> 
> diff --git a/man3/grantpt.3 b/man3/grantpt.3
> index a19172a3e..363a7aebd 100644
> --- a/man3/grantpt.3
> +++ b/man3/grantpt.3
> @@ -84,17 +84,15 @@ .SH ATTRIBUTES
>  .ad
>  .sp 1
>  .SH VERSIONS
> -Many systems implement this function via a set-user-ID helper binary
> +Historical systems implemented this function via a set-user-ID helper binary
>  called "pt_chown".
> -On Linux systems with a devpts filesystem (present since Linux 2.2),
> -the kernel normally sets the correct ownership and permissions
> -for the pseudoterminal slave when the master is opened
> -.RB ( posix_openpt (3)),
> -so that nothing must be done by
> -.BR grantpt ().
> -Thus, no such helper binary is required
> -(and indeed it is configured to be absent during the
> -glibc build that is typical on many systems).
> +glibc on Linux before glibc 2.33 could do so as well,
> +in order to support configurations with only BSD pseudoterminals;
> +this support has been removed.
> +On modern systems this is either a no-op\c
> +\[em]with permissions configured on pty allotion, as is the case on Linux\[em]\c
> +or an
> +.BR ioctl (2).

I'm thinking that this paragraph deserves to be HISTORYzised, doesn't it?

>  .SH STANDARDS
>  POSIX.1-2008.
>  .SH HISTORY

-- 
<http://www.alejandro-colomar.es/>
GPG key fingerprint: A9348594CE31283A826FBDD8D57633D441E25BB5


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]