similar to: race condition bugs

There is a race in the setup of the ControlMaster socket in auto mode, as illustrated by the following command line: ssh -oControlMaster=auto -oControlPath=sock localhost 'sleep 1; echo 1' & ssh -oControlMaster=auto -oControlPath=sock localhost 'sleep 2; echo 2' & Both of the commands will try to start up as a control client, find that sock does not exist, and switch into

There are a couple of races in sftp-server as this patch shows: --- sftp-server.c 28 Jan 2003 18:06:53 -0000 1.1.1.2 +++ sftp-server.c 5 Feb 2003 19:19:42 -0000 @@ -832,19 +832,22 @@ process_rename(void) { u_int32_t id; - struct stat st; char *oldpath, *newpath; - int ret, status = SSH2_FX_FAILURE; + int status; id = get_int(); oldpath = get_string(NULL); newpath = get_string(NULL);

http://bugzilla.mindrot.org/show_bug.cgi?id=1349 Summary: race condition with ControlMaster=auto Product: Portable OpenSSH Version: 4.6p1 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: ssh AssignedTo: bitbucket at mindrot.org ReportedBy: dot at dotat.at

All callers of robust_rename() call copy_file() if EXDEV is received. This patch moves the copy_file() call into robust_rename(). Patch Summary: -12 +1 backup.c -15 +2 rsync.c -9 +33 util.c -------------- next part -------------- patchwork diff util.c --- util.c 2004-02-17 09:58:44.000000000 -0500 +++ util.c 2004-02-17 10:21:22.000000000 -0500 @@ -355,16 +355,40 @@

From: Randy Dunlap <randy.dunlap at oracle.com> scripts/headers_install.pl prevents "__user" from being exported to userspace headers, so just use compiler.h to make sure that __user is defined and avoid the error. unifdef: linux-next-20101112/xx64/usr/include/xen/privcmd.h.tmp: 79: Premature EOF (#if line 33 depth 1) Signed-off-by: Randy Dunlap <randy.dunlap at

From: Randy Dunlap <randy.dunlap at oracle.com> scripts/headers_install.pl prevents "__user" from being exported to userspace headers, so just use compiler.h to make sure that __user is defined and avoid the error. unifdef: linux-next-20101112/xx64/usr/include/xen/privcmd.h.tmp: 79: Premature EOF (#if line 33 depth 1) Signed-off-by: Randy Dunlap <randy.dunlap at

From: Randy Dunlap <randy.dunlap at oracle.com> scripts/headers_install.pl prevents "__user" from being exported to userspace headers, so just use compiler.h to make sure that __user is defined and avoid the error. unifdef: linux-next-20101112/xx64/usr/include/xen/privcmd.h.tmp: 79: Premature EOF (#if line 33 depth 1) Signed-off-by: Randy Dunlap <randy.dunlap at

We have a system on which users are given a very restricted environment (their shell is a menu) where they should not be able to run arbitrary commands. However, because their shell is not statically linked, ld.so provides a nice clutch of holes for them to exploit. The patch below adds a new configuration option to sshd which quashes their attempts to set LD_PRELOAD etc. using ~/.ssh/environment

There are a couple of niggles with the sandboxing of the unprivileged child in the privsep code: the empty directory causes namespace pollution, and it requires care to ensure that it is set up properly and remains set up properly. The patch below (against the portable OpenSSH, although the patch against the OpenBSD version is very similar) replaces the fixed empty directory with one that is

Here's another patch for people providing ssh access to restricted environments. We allow our users to use port forwarding when logging into our mail servers so that they can use it to fetch mail over an encrypted channel using clients that don't support TLS, for example fetchmail. (In fact, fetchmail has built-in ssh support.) However we don't want them connecting to other places

When the unprivileged child has chrooted it can no longer open /etc/resolv.conf, so if the resolver hasn't yet initialized itself then dns lookups will not be possible. This is unfortunately what normally happens, but sshd falls back gracefully. There are a couple of wrinkles: the resolver will typically try talking to a nameserver on the local host by default (using INADDR_ANY rather than

--- ssh-rand-helper.c.orig Mon Jun 18 16:48:13 2007 +++ ssh-rand-helper.c Mon Jun 18 16:47:32 2007 @@ -31,7 +31,7 @@ #include <sys/socket.h> #include <stdarg.h> -#include <stddef.h> +#include <string.h> #include <netinet/in.h> #include <arpa/inet.h> Tony. -- f.a.n.finch <dot at dotat.at> http://dotat.at/ SHANNON ROCKALL: EAST OR NORTHEAST

https://bugzilla.samba.org/show_bug.cgi?id=5147 Summary: Read-only files that have changed xattrs fail to allow xattr changes Product: rsync Version: 2.6.9 Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P3 Component: core AssignedTo: wayned@samba.org

subscribe openssh-unix-dev at mindrot.org > Send openssh-unix-dev mailing list submissions to > openssh-unix-dev at mindrot.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://www.mindrot.org/mailman/listinfo/openssh-unix-dev > or, via email, send a message with subject or body 'help' to > openssh-unix-dev-request at mindrot.org >

There is a small race condition in rsync 2.5.6. When the transfer is finished, and the file is moved into place, there is a short time period where the new file is in place with the wrong permissions. When using rsync on a busy email server to replace the exim config file with a new file, exim will produce several complaints in that short period. This small patch fixes the problem, by making

Greetings, cvsup.uk.FreeBSD.org is outdated. I know this is not the proper list, but which one is? -- Best Wishes, Stefan Lambrev ICQ# 24134177

https://bugzilla.samba.org/show_bug.cgi?id=3718 Summary: RSync should verify permission/time before commiting a change Product: rsync Version: 2.6.5 Platform: Sparc OS/Version: Solaris Status: NEW Severity: major Priority: P3 Component: core AssignedTo: wayned@samba.org

Hi There, This post is brought about due to the following two: http://www.mail-archive.com/rsync@lists.samba.org/msg10702.html http://www.mail-archive.com/rsync@lists.samba.org/msg10709.html I have a situation where I need to upload large files over an unstable link (resuming is a requirement) and only when they are complete can they be renamed and hence replace the original file. The

When you give sftp-server a bogus -l or -f parameter, it logs a useful error message then blunders on to call log_init() with the bogus value. log_init() then prints a less useful message to stderr and exits. The following patch tidies this up by only printing the more useful error to stderr and not blundering on afterwards. --- sftp-server.c.orig Mon Jun 18 16:37:46 2007 +++ sftp-server.c Mon

The attached patch provides an additional check for the checksumming mode to ensure that a file that is actually written out to disk can be read back and has the same MD4 sum as the file on at the originating location. Regards, Nick. -------------- next part -------------- *** rsync-2.6.6pre1/receiver.c 2005-04-14 02:42:13.000000000 +0100 --- rsync-new/receiver.c 2005-07-26