mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Alan Kennington <akenning@dog.topology.org>
To: Linux kernel <linux-kernel@vger.kernel.org>
Cc: Alan Kennington <akenning@dog.topology.org>
Subject: easy-to-fix bug in /dev/null driver
Date: Mon, 20 Nov 2000 16:06:38 +1030	[thread overview]
Message-ID: <20001120160638.A14325@dog.topology.org> (raw)

It seems to me that this code in linux/drivers/char/mem.c
is a bug:

===================================================
static ssize_t write_null(struct file * file, const char * buf,
                          size_t count, loff_t *ppos)
{
        return count;
}
===================================================

To activate the bug try this little program:

---------------------------------------------------
#include <stdio.h>
#include <errno.h>
#include <sys/fcntl.h>

main() {
    char buf[1];
    int fd = open("/dev/null", O_RDWR);
    int i;
    printf("fd = %d\n", fd);
    for (i = 1; i <= 10; ++i) {
        int ret = write(fd, buf, -i);
        if (ret < 0) {
            fprintf(stderr, "i = %d, errno = %d\n", i, errno);
            perror("write");
            }
        }
    } 
---------------------------------------------------

On my legacy 2.4.0-test1-ac21 system, I get this:

---------------------------------------------------
fd = 3
i = 1, errno = 1
write: Operation not permitted
i = 2, errno = 2
write: No such file or directory
i = 3, errno = 3
write: No such process
i = 4, errno = 4
write: Interrupted system call
i = 5, errno = 5
write: Input/output error
i = 6, errno = 6
write: Device not configured
i = 7, errno = 7
write: Argument list too long
i = 8, errno = 8
write: Exec format error
i = 9, errno = 9
write: Bad file descriptor
i = 10, errno = 10
write: No child processes
---------------------------------------------------

You could argue that user-space users shouldn't do such stupid
things, but in some contexts, the bug might be hard to find,
and having wrong error messages just makes such bugs hard to find.
Arguably, write() should not be defined to return the count of
written bytes when this is impossible for very large writes.
I.e. it is the definition of the user-space write() function which
is meaningless for large counts - so why bother to permit this
if a negative error code is returned when this is attempted?

Perhaps more correct code for the write_null function would be:

===================================================
static ssize_t write_null(struct file * file, const char * buf,
                          size_t count, loff_t *ppos)
{
	if ((ssize_t)count >= 0)
		return (ssize_t)count;
	else
        	return 0x7fffffff;
}
===================================================

....with preferably a symbol instead of 0x7fffffff.

Cheers,
Alan Kennington.

--------------------------------------------------------------------
    name: Dr. Alan Kennington
  e-mail: akenning@dog.topology.org
 website: http://topology.org/
    city: Adelaide, South Australia
  coords: 34.88051 S, 138.59334 E
timezone: UTC+1030 http://topology.org/timezone.html
 pgp-key: http://topology.org/key_ak2.asc
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

             reply	other threads:[~2000-11-20  6:12 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2000-11-20  5:36 Alan Kennington [this message]
2000-11-20 14:23 ` Alan Kennington
2000-11-20 19:33   ` Michal Jaegermann
2000-11-20 23:13     ` Abramo Bagnara

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=20001120160638.A14325@dog.topology.org \
    --to=akenning@dog.topology.org \
    --cc=linux-kernel@vger.kernel.org \
    /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

all inboxes | Powered by JetHome®