public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Dirk Gouders <dirk@gouders.net>
To: Al Viro <viro@zeniv.linux.org.uk>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Jiri Slaby <jslaby@suse.com>,
	linux-kernel@vger.kernel.org, linux-newbie@vger.kernel.org
Subject: Re: pty: childs don't always react on close(2)
Date: Fri, 05 Dec 2025 23:20:03 +0100	[thread overview]
Message-ID: <ghh5u4zi3w.fsf@gouders.net> (raw)
In-Reply-To: <20251205214852.GL1712166@ZenIV> (Al Viro's message of "Fri, 5 Dec 2025 21:48:52 +0000")

[-- Attachment #1: Type: text/plain, Size: 850 bytes --]

Al Viro <viro@zeniv.linux.org.uk> writes:

> On Fri, Dec 05, 2025 at 10:37:44PM +0100, Dirk Gouders wrote:
>
>> 	child_pid1 = forkpty(&pty_fd1, NULL, NULL, NULL);
>
> You do realize that it will inherit all your opened descriptors,
> including pty_fd, right?
>
> ...

>> 	close(pty_fd);
>
> ... which doesn't do anything to the second child's descriptor
> table, including the descriptor that refers to the same opened file.
> IOW, the IO channel (== opened file) is very much opened after
> that close() - descriptors refering to it still exist.

Oh yes, thank you very much!

I use FD_CLOEXEC nearly everywhere in the mentioned program but exactly
not for the pty file descriptors.  Oh well...

For completenes, I'll attach the modified test-case.
Would be interesting to hear if there are other possible fixes.

Thank you very much, again,

Dirk


[-- Attachment #2: PTY test program --]
[-- Type: text/plain, Size: 2138 bytes --]

/*
 * Test closing file descriptors opened via forkpty() when not all data has been
 * read.  A following waitpid() blocks, when we opened two childs and try to
 * close the file descriptor and then waitpid() for that child...
 */
#include <stdlib.h>
#include <stdio.h>
#include <pty.h>
#include <string.h>
#include <errno.h>
#include <unistd.h>
#include <sys/wait.h>
#include <stdint.h>
#include <fcntl.h>

#define READ_SIZE 4096

/*
 * Set PAGER variables and start a man(1) process.
 */
void do_child(void);
void do_child()
{
	char *e_argv[3] = {"man", "groff_ms", NULL};

	putenv("PAGER=cat");
	putenv("MANPAGER=cat");
	execvp("man", e_argv);
}

int main()
{
	unsigned char read_buffer[READ_SIZE];
	int pty_fd;
	int pty_fd1;
	int wstatus;
	pid_t child_pid;
	pid_t child_pid1;
	pid_t ret_pid;
	ssize_t ret;
	/*
	 * Start a child to send us a manual page.
	 */
	child_pid = forkpty(&pty_fd, NULL, NULL, NULL);

	if (child_pid == -1) {
		fprintf(stderr, "forkpty(): %s\n", strerror(errno));
		exit(EXIT_FAILURE);
	}

	ret = fcntl(pty_fd, F_SETFD, FD_CLOEXEC);

	if (child_pid == 0)
		do_child();

	printf("child_pid = %jd\n", (intmax_t) child_pid);

	memset(read_buffer, '\0', READ_SIZE);
	ret = read(pty_fd, read_buffer, READ_SIZE);

	if (ret == -1) {
		fprintf(stderr, "read(): %s\n", strerror(errno));
		exit(EXIT_FAILURE);
	}

	printf("%s\n", read_buffer);

	printf("%ld bytes read.\n", ret);


	/*
	 * Start another child to send us a manual page.
	 */
	child_pid1 = forkpty(&pty_fd1, NULL, NULL, NULL);

	if (child_pid1 == -1) {
		fprintf(stderr, "forkpty(): %s\n", strerror(errno));
		exit(EXIT_FAILURE);
	}

	ret = fcntl(pty_fd1, F_SETFD, FD_CLOEXEC);

	if (child_pid1 == 0)
		do_child();

	printf("child_pid1 = %jd\n", (intmax_t) child_pid1);

	memset(read_buffer, '\0', READ_SIZE);
	ret = read(pty_fd1, read_buffer, READ_SIZE);

	if (ret == -1) {
		fprintf(stderr, "read(): %s\n", strerror(errno));
		exit(EXIT_FAILURE);
	}

	printf("%s\n", read_buffer);

	printf("%ld bytes read.\n", ret);


	close(pty_fd);

	ret_pid = waitpid(child_pid, &wstatus, 0);

	printf("ret_pid = %jd\n", (intmax_t) ret_pid);

	exit(EXIT_SUCCESS);
}

      reply	other threads:[~2025-12-05 22:20 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-12-05 21:37 pty: childs don't always react on close(2) Dirk Gouders
2025-12-05 21:48 ` Al Viro
2025-12-05 22:20   ` Dirk Gouders [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ghh5u4zi3w.fsf@gouders.net \
    --to=dirk@gouders.net \
    --cc=gregkh@linuxfoundation.org \
    --cc=jslaby@suse.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-newbie@vger.kernel.org \
    --cc=viro@zeniv.linux.org.uk \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox