Date: Tue, 13 Nov 2012 09:52:21 +
From: Emmanuel Dreyfus m...@netbsd.org
New version of the patch for POSIX extended API set 2:
http://ftp.espci.fr/shadow/manu/openat3.patch
OK, thanks, this looks better. Some comments below, mostly very
minor. First, though -- are there any
Taylor R Campbell campbell+netbsd-tech-k...@mumble.net wrote:
OK, thanks, this looks better. Some comments below, mostly very
minor. First, though -- are there any changes to header files? I
didn't see any in the patch.
No, that have been done some time ago so that it could be already in
On Sun, Nov 11, 2012 at 08:17:26AM +0100, Emmanuel Dreyfus wrote:
I know this is a bike shed, and I'm sorry to be the one to bring it
up, but can we use the names chmodat, chownat, c., for our native
system calls, and just use libc aliases or _BLAH_SOURCE nonsense or
something for the
On Sun, Nov 11, 2012 at 05:55:23AM +0100, Emmanuel Dreyfus wrote:
+ /*
+ * openat() falls back to open() behavior if
+ * - path is absolute XXX check this.
+ * - fd is AT_FDCWD
+ */
Have you checked the XXX, and/or written automatic tests for it?
On Sun, 11 Nov 2012, Emmanuel Dreyfus wrote:
Taylor R Campbell campbell+netbsd-tech-k...@mumble.net wrote:
I know this is a bike shed, and I'm sorry to be the one to bring it
up, but can we use the names chmodat, chownat, c., for our native
system calls, and just use libc aliases or
On Sun, Nov 11, 2012 at 08:31:02AM +0100, Emmanuel Dreyfus wrote:
I don't think namei.h / vfs_lookup.c is the right place to be handling
file descriptors. Can you make these take vnodes, rather than file
descriptors, or move them into vfs_syscalls.c?
I made the change, but it makes
On Sun, Nov 11, 2012 at 10:39:06AM +0200, Alan Barrett wrote:
However, I also want the inconsistent POSIX names to be provided.
I don't know a good way of satisfying both goals.
#if defined(_POSIX_C_SOURCE) || defined(_XOPEN_SOURCE) /* or whatever */
int fchmodat(int, const char *, mode_t,
David Holland dholland-t...@netbsd.org wrote:
What is the goal? You want to write userland code using chmodat()
instead of fchmodat()?
The wrong names should be exposed in as few places as possible.
What do you call the wrong name? The non-standard chmodat() or the
odd-looking fchmodat()?
David Holland dholland-t...@netbsd.org wrote:
In order to check for absolute path, I need to copyin the path, and it
wil be copied in later again. I suspect that check should go at namei
level.
I don't understand. Why do you need to copy in the path twice? Don't
do that.
Yes, I
David Holland dholland-t...@netbsd.org wrote:
I think you'll leak pb here in the error branch above. I haven't gone
through the other error branches in the patch, but the proliferation
of new resource acquisitions requiring clean-up makes me nervous.
The problem is that we
On Nov 11, 2012, at 12:39 AM, Alan Barrett wrote:
I want the names to follow a clear and easily-documented pattern.
Takes a name Takes a fd, not a name Takes a name and an at fd
(prepend f) (append at)
--
On Sun, Nov 11, 2012 at 04:19:03AM -0800, Matt Thomas wrote:
On Nov 11, 2012, at 12:39 AM, Alan Barrett wrote:
I want the names to follow a clear and easily-documented pattern.
Takes a nameTakes a fd, not a name Takes a name and an at fd
(prepend f)
On Nov 11, 2012, at 4:13 AM, David Laight wrote:
On Sun, Nov 11, 2012 at 04:19:03AM -0800, Matt Thomas wrote:
On Nov 11, 2012, at 12:39 AM, Alan Barrett wrote:
I want the names to follow a clear and easily-documented pattern.
Takes a nameTakes a fd, not a name Takes a name and
Alan Barrett a...@cequrux.com wrote:
chdir fchdir chdirat
chdirat does not exist in standard.
--
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
m...@netbsd.org
I want the names to follow a clear and easily-documented pattern.
However, I also want the inconsistent POSIX names to be provided.
I see where you're coming from, but it seems like asking for trouble to
encourage any code to be written which uses function names other than
the ones defined
Greg Troxel g...@ir.bbn.com wrote:
I see where you're coming from, but it seems like asking for trouble to
encourage any code to be written which uses function names other than
the ones defined by POSIX. So I don't see that defining the names that
POSIX should have used as a net positive
Date: Sun, 11 Nov 2012 08:31:02 +0100
From: m...@netbsd.org (Emmanuel Dreyfus)
Taylor R Campbell campbell+netbsd-tech-k...@mumble.net wrote:
I don't think namei.h / vfs_lookup.c is the right place to be handling
file descriptors. Can you make these take vnodes, rather than file
Date: Sun, 11 Nov 2012 05:55:23 +0100
From: m...@netbsd.org (Emmanuel Dreyfus)
Taylor R Campbell campbell+netbsd-tech-k...@mumble.net wrote:
+ /*
+ * openat() falls back to open() behavior if
+ * - path is absolute XXX check this.
+ * - fd is AT_FDCWD
NetBSD does things right, but somehow encouraging the write of
nonportable code is not The Right Thing, IMO.
Like the way NetSBD does by providing extensions in libraries, like
glob(3) or regex(3)? By shipping with a compiler that supports (highly
useful, but nonstandard) extensions to C?
Taylor R Campbell campbell+netbsd-tech-k...@mumble.net wrote:
Sorry, my question wasn't clear. All I meant was `Have you confirmed
by code inspection that this is true
Yes.
--
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
m...@netbsd.org
Mouse mo...@rodents-montreal.org wrote:
NetBSD does things right, but somehow encouraging the write of
nonportable code is not The Right Thing, IMO.
Like the way NetSBD does by providing extensions in libraries, like
glob(3) or regex(3)?
[many other good BSD additions]
You cite a lot of
Date: Sat, 10 Nov 2012 07:44:18 +0100
From: m...@netbsd.org (Emmanuel Dreyfus)
This patch implementes openat(2), mkdirat(2), mknodat(2), mkfifoat(2)
fchmodat(2) and fchownat(2), and adds test cases for the new system
calls:
I know this is a bike shed, and I'm sorry to be the one
Some quick responses to the substance of the patch now that I have
skimmed it...
Date: Sat, 10 Nov 2012 07:44:18 +0100
From: m...@netbsd.org (Emmanuel Dreyfus)
Index: sys/sys/namei.h
+int nameiat_simple_kernel(int, const char *, namei_simple_flags_t,
+struct vnode **);
Taylor R Campbell campbell+netbsd-tech-k...@mumble.net wrote:
+ /*
+ * openat() falls back to open() behavior if
+ * - path is absolute XXX check this.
+ * - fd is AT_FDCWD
+ */
Have you checked the XXX, and/or written automatic tests for it?
In order to check for
Taylor R Campbell campbell+netbsd-tech-k...@mumble.net wrote:
I know this is a bike shed, and I'm sorry to be the one to bring it
up, but can we use the names chmodat, chownat, c., for our native
system calls, and just use libc aliases or _BLAH_SOURCE nonsense or
something for the ridiculous
Taylor R Campbell campbell+netbsd-tech-k...@mumble.net wrote:
I don't think namei.h / vfs_lookup.c is the right place to be handling
file descriptors. Can you make these take vnodes, rather than file
descriptors, or move them into vfs_syscalls.c?
I made the change, but it makes
Hi
Thanks to David Holland work on namei, I was able to start implementing
POSIX extended API set 2.
This patch implementes openat(2), mkdirat(2), mknodat(2), mkfifoat(2)
fchmodat(2) and fchownat(2), and adds test cases for the new system
calls:
http://ftp.espci.fr/shadow/manu/openat.patch
27 matches
Mail list logo