qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Eric Blake <eblake@redhat.com>
To: Keno Fischer <keno@juliacomputing.com>, qemu-devel@nongnu.org
Cc: groug@kaod.org
Subject: Re: [Qemu-devel] [PATCH v2 12/20] 9p: darwin: Explicitly cast comparisons of mode_t with -1
Date: Fri, 29 Jun 2018 15:32:23 -0500	[thread overview]
Message-ID: <dc97fc23-6e0f-e232-1ef8-3c562b28c9cd@redhat.com> (raw)
In-Reply-To: <dc7a7959808593c941fbf73f9f426f92632c0029.1527814874.git.keno@juliacomputing.com>

On 05/31/2018 08:26 PM, Keno Fischer wrote:
> Comparisons of mode_t with -1 require an explicit cast, since mode_t
> is unsigned on Darwin.

It's not JUST that mode_t is unsigned (an unsigned int compares just 
fine against -1), but ALSO that mode_t has unspecified width.  That is, 
you cannot portably assume whether mode_t is smaller, equivalent, or 
larger rank than int.  If it is smaller, then you can't use mode_t in 
va_arg(), and mode_t will promote to signed int, whether or not mode_t 
is unsigned; but '((mode_t)-1) == -1' is going to be false if mode_t is 
unsigned (because the cast truncates the sign extension bits into a 
positive value).  Conversely, since mode_t can be larger than int 
(although I know of no such platform that does so), blindly using 'int' 
when trying to parse a mode_t argument through va_arg() will truncate bits.

So, for example, to portably write a wrapper around open(), you HAVE to 
use hairy constructions like:

mode = (sizeof (mode_t) < sizeof (int)
         ? va_arg (ap, int)
         : va_arg (ap, mode_t));

or, to avoid spurious compiler warnings on the branch not taken, define 
a macro learned at configure time.  This is what gnulib does, for example:

   AC_CACHE_CHECK([for promoted mode_t type], [gl_cv_promoted_mode_t], [
     dnl Assume mode_t promotes to 'int' if and only if it is smaller 
than 'int',
     dnl and to itself otherwise. This assumption is not guaranteed by 
the ISO C
     dnl standard, but we don't know of any real-world counterexamples.
     AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[#include <sys/types.h>]],
       [[typedef int array[2 * (sizeof (mode_t) < sizeof (int)) - 1];]])],
       [gl_cv_promoted_mode_t='int'],
       [gl_cv_promoted_mode_t='mode_t'])
   ])
   AC_DEFINE_UNQUOTED([PROMOTED_MODE_T], [$gl_cv_promoted_mode_t],
     [Define to the type that is the result of default argument 
promotions of type mode_t.])

> +++ b/hw/9pfs/9p-local.c
> @@ -310,7 +310,7 @@ update_map_file:
>       if (credp->fc_gid != -1) {
>           gid = credp->fc_gid;
>       }
> -    if (credp->fc_mode != -1) {
> +    if (credp->fc_mode != (mode_t)-1) {

At any rate, this is the correct portability fix for this code.

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org

  reply	other threads:[~2018-06-29 20:32 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-06-01  1:25 [Qemu-devel] [PATCH v2 00/20] 9p: Add support for Darwin Keno Fischer
2018-06-01  1:25 ` [Qemu-devel] [PATCH v2 01/20] cutils: Provide strchrnul Keno Fischer
2018-06-01  8:15   ` Greg Kurz
2018-06-01  8:46     ` Dr. David Alan Gilbert
2018-06-01 14:15   ` Eric Blake
2018-06-01  1:25 ` [Qemu-devel] [PATCH v2 02/20] 9p: proxy: Fix size passed to `connect` Keno Fischer
2018-06-01  9:09   ` Greg Kurz
2018-06-01  1:25 ` [Qemu-devel] [PATCH v2 03/20] 9p: xattr: Fix crash due to free of uninitialized value Keno Fischer
2018-06-01  9:19   ` Greg Kurz
2018-06-01  1:25 ` [Qemu-devel] [PATCH v2 04/20] 9p: linux: Fix a couple Linux assumptions Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 05/20] 9p: Properly set errp in fstatfs error path Keno Fischer
2018-06-01  9:32   ` Greg Kurz
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 06/20] 9p: Avoid warning if FS_IOC_GETVERSION is not defined Keno Fischer
2018-06-01  9:57   ` Greg Kurz
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 07/20] 9p: Move a couple xattr functions to 9p-util Keno Fischer
2018-06-01 10:03   ` Greg Kurz
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 08/20] 9p: Rename 9p-util -> 9p-util-linux Keno Fischer
2018-06-01 10:07   ` Greg Kurz
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 09/20] 9p: Properly check/translate flags in unlinkat Keno Fischer
2018-06-01 10:13   ` Greg Kurz
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 10/20] 9p: darwin: Handle struct stat(fs) differences Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 11/20] 9p: darwin: Handle struct dirent differences Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 12/20] 9p: darwin: Explicitly cast comparisons of mode_t with -1 Keno Fischer
2018-06-29 20:32   ` Eric Blake [this message]
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 13/20] 9p: darwin: Ignore O_{NOATIME, DIRECT} Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 14/20] 9p: darwin: Provide a compatibility definition for XATTR_SIZE_MAX Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 15/20] 9p: darwin: *xattr_nofollow implementations Keno Fischer
2018-06-01 11:13   ` Greg Kurz
2018-06-02 20:01     ` Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 16/20] 9p: darwin: Compatibility for f/l*xattr Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 17/20] 9p: darwin: Provide a fallback implementation for utimensat Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 18/20] 9p: darwin: Implement compatibility for mknodat Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 19/20] 9p: darwin: virtfs-proxy: Implement setuid code for darwin Keno Fischer
2018-06-01  1:26 ` [Qemu-devel] [PATCH v2 20/20] 9p: darwin: configure: Allow VirtFS on Darwin Keno Fischer

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=dc97fc23-6e0f-e232-1ef8-3c562b28c9cd@redhat.com \
    --to=eblake@redhat.com \
    --cc=groug@kaod.org \
    --cc=keno@juliacomputing.com \
    --cc=qemu-devel@nongnu.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;
as well as URLs for NNTP newsgroup(s).