new snapshot available: coreutils-7.5.76-1c0ec
Here's another snapshot, prior to coreutils-7.6. We've inherited quite a few changes from gnulib, including one to fix the Solaris build failure. coreutils snapshot: http://meyering.net/cu/coreutils-ss.tar.gz 9.6 MB http://meyering.net/cu/coreutils-ss.tar.xz 4.0 MB http://meyering.net/cu/coreutils-ss.tar.gz.sig http://meyering.net/cu/coreutils-ss.tar.xz.sig aka http://meyering.net/cu/coreutils-7.5.76-1c0ec.tar.gz http://meyering.net/cu/coreutils-7.5.76-1c0ec.tar.xz Changes in coreutils since 7.5.65-61cc6: Jim Meyering (9): build: update gnulib submodule to latest tests: tail-2/infloop-1: avoid rare test failure on a busy system tests: ls/color-clear-to-eol: append NL to accommodate old sed tests: misc/cat-buf: clean up syntax tail: don't give up on inotify mode for an already-ignored "-" tail: ignore -f for piped-stdin, as POSIX requires tail: syntax-only: use "false", rather than equivalent, 0 tail: make the new piped-stdin test as portable as the old one * gnulib: Update submodule to latest. Pádraig Brady (3): tests: tail-2/flush-initial should not rely on stdbuf tests: address a race condition in misc/cat-buf tests: tail-2/assert: reduce risk of race condition Changes in gnulib since 7.5.65-61cc6: * gnulib c46ae39...2c90f1a (23): > fstatat: fix compilation on Solaris > rename: modernize replacement > stdio: sort witness names > getcwd: minor cleanups > openat: provide more convenience names > doc: fix comments in recent patches > faccessat, symlinkat: continue cleanup of previous patch > Remove dependency from unistd_h.m4, sys_stat_h.m4, fcntl_h.m4 to openat.m4. > Remove unneeded gl_MODULE_INDICATOR invocation. > Set errno to ENOSYS when a function is entirely unsupported. > Mention affected platforms. > Fix indentation. > For NULL, include , not . > Work around a glibc bug in strtok_r. > exclude: minor fix > bootstrap: improve error message > Update copyright year. > symlinkat: new module > test-openat-safer: add more checks * tests/test-openat-safer.c (main): Check more code paths. > syntax-check: detect unnecessary inclusion of openat.h > Support towlower and towupper. > Fix bug occurring when converting invalid input. > Test indirect conversion as well.
fsetxattr() fails on a writable descriptor?
fsetxattr() is failing for me when the permissions are readonly
but the descriptor is writable as demonstrated by the program below.
I've tried this on ext3 and ext4 with libattr-2.4.43-3.fc11.i586
and linux-2.6.30.5-43.fc11.i586
Is this expected?
cheers,
Pádraig.
#include
#include
#include
#include
#include
#include
int main(void)
{
int wfd;
if ((wfd=open("writable", O_CREAT|O_WRONLY|O_EXCL, S_IRUSR)) == -1)
fprintf(stderr, "open('writable') error [%m]\n");
if (write(wfd, "data", 1) == -1)
fprintf(stderr, "write() error [%m]\n");
if (fsetxattr(wfd, "user.test", "test", 4, 0) == -1)
fprintf(stderr, "fsetxattr() error [%m]\n");
}
Re: new snapshot available: coreutils-7.5.76-1c0ec
On 2009-09-08 13:55 +0200, Jim Meyering wrote: > Here's another snapshot, prior to coreutils-7.6. All tests passed on Debian sid i386, including the (very) expensive and root checks. :-) Cheers, Sven
Re: new snapshot available: coreutils-7.5.76-1c0ec
Sven Joachim wrote: > On 2009-09-08 13:55 +0200, Jim Meyering wrote: >> Here's another snapshot, prior to coreutils-7.6. > > All tests passed on Debian sid i386, including the (very) expensive and > root checks. :-) Nice. Thanks!
Re: new snapshot available: coreutils-7.5.76-1c0ec
Jim Meyering wrote: > Here's another snapshot, prior to coreutils-7.6. > We've inherited quite a few changes from gnulib, > including one to fix the Solaris build failure. Passed Skipped Failed \- Fedora core 5 x86 | 348 42 0 Fedora 11 x86 | 343 47 0 Solaris 10 x86| 328 61 1 Solaris 9 x86 | 314 61 15 Note the solaris systems were on nfs new solaris 10 failure in cp/into-self: > diff -u exp out --- exp Tue Sep 8 15:29:02 2009 +++ out Tue Sep 8 15:26:48 2009 @@ -4,5 +4,6 @@ 2 cp: cannot copy a directory, `dir', into itself, `dir/dir' 3 +cp: cannot create directory `dir/dir/dir/.../dir/dir': File name too long cp: cannot copy a directory, `dir', into itself, `dir/dir' 4 solaris 9 failures: misc/help-version install/basic-1 install/create-leading install/install-C install/strip-program ln/misc cp/backup-dir cp/cp-parents cp/file-perm-race cp/parent-perm cp/parent-perm-race cp/preserve-2 cp/reflink-perm cp/src-base-dot mv/part-symlink All due to: facl(4, SETACL, 3, 0x08047458) -> EOPNOTSUPP
Re: new snapshot available: coreutils-7.5.76-1c0ec
rerunning on local file systems: Passed Skipped Failed \- Fedora core 5 x86 | 348 42 0 Fedora 11 x86 | 343 47 0 Solaris 10 x86| 328 62 0 Solaris 9 x86 | 327 62 1 solaris 9 failure was: mv/part-symlink mv loc_reg rem_sl mv: preserving permissions for `rem_sl': Invalid argument facl(4, SETACL, 4, 0x08097FA0) -> Err#22 EINVAL cheers, Pádraig.
Bug#545721: info: No menu item `mktemp invocation' in node `(coreutils.info.gz)Top'.
X-debbugs-cc: bug-coreutils@gnu.org Package: coreutils Version: 7.5-3 Severity: minor $ man mktemp|col -b|grep -C 3 'info coreutils' The full documentation for mktemp is maintained as a Texinfo manual. If the info and mktemp programs are prop erly installed at your site, the command info coreutils 'mktemp invocation' should give you access to the complete manual. $ info coreutils 'mktemp invocation' 1>&- info: No menu item `mktemp invocation' in node `(coreutils.info.gz)Top'. P.S., the man page should say what directory the temp file ends up in if no options are given. Maybe it currently does but one has to look among the options to find out. Also mention what "Create a temporary file or directory, safely" means. Guaranteed not to overwrite any other existing file perhaps, and ...
wishlist: give cut(1) option to split fields on whitespace
It would be really nice if cut(1) could split fields using whitespace
as a separator (as many other text tools do). Currently I use
$ awk '{ print $3, $4, $5 }'
for this sort of thing, but it seems relatively klutzy, as opposed to (say)
$ cut -w -f 3-5
Mike
Re: fsetxattr() fails on a writable descriptor?
On Tue, Sep 08, 2009 at 12:29:49PM +0100, P??draig Brady wrote: > fsetxattr() is failing for me when the permissions are readonly > but the descriptor is writable as demonstrated by the program below. > I've tried this on ext3 and ext4 with libattr-2.4.43-3.fc11.i586 > and linux-2.6.30.5-43.fc11.i586 That's not XFS specifci behaviour. For all xattr operations, fd-based or not we always do the full permission checks, and have done so since day one. Check the routined xattr_permission in fs/xattr.c of the kernel tree. It seems a bit unusual, but I'm not sure changing it now is a good idea - the xattr support has been around for a long time.
extend inotify to support file descriptors in addition to paths
Hello,
at the moment inotify permits to add new files to be watched using their
path. There are situations where the file path is not know but a
descriptor is available. It would be desiderable to have the
possibility to use the inotify system even in these (rare) cases.
A concrete example of application that can take advantage of this
possibility is GNU tail. GNU tail, since version 7.5, is using inotify
to watch changes on files. This command will make `tail' sleeps until
some events are catched:
tail -qf foo bar baz
Inotify can't be used when stdin is watched too:
tail -qf boo bar - < baz
While the user gets the same result, internally it is a completely
different thing because polling is done on stdin instead of graciously
sleep until something is ready.
I don't know if it is possible to obtain the former result when the fd
is know but the path isn't.
My proposal is to extend inotify adding a new syscall:
long sys_inotify_add_watch_fd(int wd, int fd, u32 mask);
in this way the same mechanism can be used indifferently either a
descriptor or on a path is provided.
What do you think about this idea? Is there another way to get the same
result?
Regards,
Giuseppe
P.S.
In the attached patch I changed the syscalls table only for x86.
>From 6f3340347fb5d26000bec3a0a9cd8ef59c710922 Mon Sep 17 00:00:00 2001
From: Giuseppe Scrivano
Date: Tue, 8 Sep 2009 23:40:03 +0200
Subject: [PATCH] Add new syscall \`inotify_add_watch_fd'.
It makes possible to register a file to be watched by inotify using
its file descriptor instead of its path.
---
arch/x86/kernel/syscall_table_32.S |1 +
fs/notify/inotify/inotify_user.c | 54 ++--
include/linux/syscalls.h |3 ++
3 files changed, 55 insertions(+), 3 deletions(-)
diff --git a/arch/x86/kernel/syscall_table_32.S
b/arch/x86/kernel/syscall_table_32.S
index d51321d..46eb4ac 100644
--- a/arch/x86/kernel/syscall_table_32.S
+++ b/arch/x86/kernel/syscall_table_32.S
@@ -336,3 +336,4 @@ ENTRY(sys_call_table)
.long sys_pwritev
.long sys_rt_tgsigqueueinfo /* 335 */
.long sys_perf_counter_open
+ .long sys_inotify_add_watch_fd
diff --git a/fs/notify/inotify/inotify_user.c b/fs/notify/inotify/inotify_user.c
index dcd2040..53f3c41 100644
--- a/fs/notify/inotify/inotify_user.c
+++ b/fs/notify/inotify/inotify_user.c
@@ -747,16 +747,64 @@ SYSCALL_DEFINE3(inotify_add_watch, int, fd, const char
__user *, pathname,
/* create/update an inode mark */
ret = inotify_update_watch(group, inode, mask);
- if (unlikely(ret))
- goto path_put_and_out;
-path_put_and_out:
path_put(&path);
fput_and_out:
fput_light(filp, fput_needed);
return ret;
}
+SYSCALL_DEFINE3(inotify_add_watch_fd, int, wd, int, fd,u32, mask)
+{
+ struct fsnotify_group *group;
+ struct inode *inode;
+ struct path path;
+ struct file *filp;
+ struct file *filp2;
+ int ret, fput_needed, fput2_needed;
+ unsigned flags = 0;
+
+ filp = fget_light(wd, &fput_needed);
+ if (unlikely(!filp))
+ return -EBADF;
+
+ /* verify that this is indeed an inotify instance */
+ if (unlikely(filp->f_op != &inotify_fops)) {
+ ret = -EINVAL;
+ goto fput_and_out;
+ }
+
+ if (!(mask & IN_DONT_FOLLOW))
+ flags |= LOOKUP_FOLLOW;
+ if (mask & IN_ONLYDIR)
+ flags |= LOOKUP_DIRECTORY;
+
+ filp2 = fget_light(fd, &fput2_needed);
+
+ if (unlikely(!filp2)){
+ ret = -EBADF;
+ goto fput_and_out;
+ }
+
+ /* you can only watch an inode if you have read permissions on it */
+ ret = inode_permission(filp2->f_path.dentry->d_inode, MAY_READ);
+ if (ret)
+ goto fput2_and_out;
+
+ /* inode held in place by reference to path; group by fget on fd */
+ inode = filp2->f_path.dentry->d_inode;
+ group = filp->private_data;
+
+ /* create/update an inode mark */
+ ret = inotify_update_watch(group, inode, mask);
+
+fput2_and_out:
+ fput_light(filp2, fput2_needed);
+fput_and_out:
+ fput_light(filp, fput_needed);
+ return ret;
+}
+
SYSCALL_DEFINE2(inotify_rm_watch, int, fd, __s32, wd)
{
struct fsnotify_group *group;
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 80de700..0833b1d 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -761,4 +761,7 @@ int kernel_execve(const char *filename, char *const argv[],
char *const envp[]);
asmlinkage long sys_perf_counter_open(
struct perf_counter_attr __user *attr_uptr,
pid_t pid, int cpu, int group_fd, unsigned long flags);
+
+
+asmlinkage long sys_inotify_add_watch_fd(int wd, int fd, u32 mask);
#endif
--
1.6.3.3
