mailing list of musl libc
 help / color / mirror / code / Atom feed
From: Rich Felker <dalias@libc.org>
To: musl@lists.openwall.com
Subject: Re: [musl] [PATCH] stdio: Fix fdopen bug
Date: Wed, 19 Feb 2020 09:16:36 -0500	[thread overview]
Message-ID: <20200219141636.GG1663@brightrain.aerifal.cx> (raw)
In-Reply-To: <2dd69ee3939f469ab98a07d5337a8bf0@huawei.com>

On Wed, Feb 19, 2020 at 06:47:53AM +0000, zhangtianci wrote:
> > On Wed, Feb 19, 2020 at 10:37:29AM +0800, Zhang Tianci wrote:
> > > Currently, in musl the fdopen doesn't check the consistence between
> > > fd's mode and corresponding file's mode.
> > >
> > > For example,
> > >
> > > int fd = open("file1", O_RDONLY);
> > > FILE *f = fdopen(fd, "W")
> > >
> > > In musl, above code will be Okay.
> > > While according to POSIX, above code (fdopen) will return EINVAL.
> > >
> > > Signed-off-by: Zhang Tianci <zhangtianci1@huawei.com>
> > > ---
> > >  src/stdio/__fdopen.c | 10 ++++++++++
> > >  1 file changed, 10 insertions(+)
> > >
> > > diff --git a/src/stdio/__fdopen.c b/src/stdio/__fdopen.c index
> > > 116e78e..23c4ffd 100644
> > > --- a/src/stdio/__fdopen.c
> > > +++ b/src/stdio/__fdopen.c
> > > @@ -26,6 +26,16 @@ FILE *__fdopen(int fd, const char *mode)
> > >  	/* Impose mode restrictions */
> > >  	if (!strchr(mode, '+')) f->flags = (*mode == 'r') ? F_NOWR : F_NORD;
> > >
> > > +	int fd_flag = __syscall(SYS_fcntl, fd, F_GETFL);
> > > +
> > > +	if (fd_flag == -1) return 0;
> > > +
> > > +	if (((fd_flag & O_ACCMODE) == O_RDONLY && !(f->flags & F_NORD))
> > ||
> > > +	    ((fd_flag & O_ACCMODE) == O_WRONLY && !(f->flags &
> > F_NOWR))) {
> > > +		errno = EINVAL;
> > > +		return 0;
> > > +	}
> > > +
> > >  	/* Apply close-on-exec flag */
> > >  	if (strchr(mode, 'e')) __syscall(SYS_fcntl, fd, F_SETFD,
> > > FD_CLOEXEC);
> > >
> > > --
> > > 2.17.1
> > 
> > Per POSIX this is a "may fail" not a "shall fail". Testing for this is more costly
> > (see added code/syscalls in the patch) and serves no purpose, which is why
> > it's not done.
> > 
> > Rich
> 
> POSIX's require on fdopen:
> 
>      The application shall ensure that the mode of the stream as expressed by the
       ^^^^^^^^^^^^^^^
>      mode argument is allowed by the file access mode of the open file description 
>      to which fildes refers.
> 
> So I think the example above should return EINVAL.

The text you're quoting is placing a requirement on the application,
not the implementation. If the application fails to meet a "shall", it
has undefined behavior and there are no obligations whatsoever on the
implementation.

The error is clearly a "may fail" if you read the ERRORS section of
the specification.

Rich

  reply	other threads:[~2020-02-19 14:16 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-19  6:47 zhangtianci
2020-02-19 14:16 ` Rich Felker [this message]
2020-02-19 14:20 ` Jens Gustedt
  -- strict thread matches above, loose matches on Subject: below --
2020-02-19  2:37 Zhang Tianci
2020-02-19  3:44 ` Rich Felker

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=20200219141636.GG1663@brightrain.aerifal.cx \
    --to=dalias@libc.org \
    --cc=musl@lists.openwall.com \
    /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.
Code repositories for project(s) associated with this public inbox

	https://git.vuxu.org/mirror/musl/

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).