similar to: [PATCH v2] launch: Close file descriptors after fork (RHBZ#1123007).

This refactors existing code to close file descriptors in the recovery process, and also adds code to close file descriptors between the fork() and exec() of QEMU or User-Mode Linux. The reason is to avoid leaking main process file descriptors where the main process (or other libraries in the main process) are not setting O_CLOEXEC at all or not setting it atomically. Python is a particular

Hi Richard, I see this patch is included into libguestfs >= 1.26.6 & libguestfs >= 1.27.24. I feel that version is too high. Now RHEL 7.0 runs libguestfs-1.22.6, RHEL 6.6 alpha and 6.5 runs libguestfs-1.20.11. Is that possible to backport this patch, in order to make this problem fixed in RHEL 7.0, 6.6, and 6.5? Thanks & Best Regards Qin Zhao (赵钦) China System and Technology Lab,

v1 was here: https://www.redhat.com/archives/libguestfs/2013-August/msg00005.html This now works, to some extent. The main problem now is that devices are named /dev/ubd[a-] which of course confuses everything. I'm thinking it may be easier to add a udev rule to rename them. Rich.

This is a User-Mode Linux backend for libguestfs. You can select it by doing: export LIBGUESTFS_BACKEND=uml export LIBGUESTFS_QEMU=/path/to/vmlinux Note we're reusing the 'qemu' variable in the handle for convenience. QEmu is not involved when using the UML backend. This almost works. UML itself crashes when the daemon tries to connect to the serial port. I suspect it's

Add some warnings. Well, the first patch is a miscellaneous change, but patches 2-5 add some warnings. Rich.

If SIGTERM is blocked in the main program, then it ends up still being blocked in the subprocess after we fork. This means that we cannot kill qemu by sending SIGTERM to it. This commit fixes the problem by unblocking SIGTERM unconditionally after fork. Thanks: wtfuzz on IRC for reporting and analysis. --- lib/launch-direct.c | 15 +++++++++++++++ lib/launch-uml.c | 11 +++++++++++ 2 files

As the start of work to add remote support, I'm taking a close look at the protocol layer in the library. These are some small cleanups. Rich.

The last (only?) caller of guestfs_int_cmd_clear_close_files() disappeared in commit e4c396888056 ("lib/info: Remove /dev/fd hacking and pass a true filename to qemu-img info.", 2018-01-23), part of v1.37.36. Simplify the code by removing guestfs_int_cmd_clear_close_files(). Signed-off-by: Laszlo Ersek <lersek at redhat.com> --- lib/guestfs-internal.h | 1 - lib/command.c

This is the libguestfs companion patch to: https://www.redhat.com/archives/libguestfs/2013-September/msg00045.html Rich.

QEMU upstream has broken snapshot=on ... again. These two patches stop using it entirely. Instead we run 'qemu-img create' to create overlay disks as required. Note that the libvirt and UML backends were already doing this: The libvirt backend because <transient/> has never worked, and the UML backend was running uml_mkcow because the UML-equivalent syntax of snapshot=on was

Inspired by libvirt's virCommand* internal mini-library, this adds some internal APIs for running commands. The first patch contains the new APIs. The subsequent patches change various parts of the library over to use it. Rich.

As Eric Blake noted in: https://www.redhat.com/archives/libguestfs/2016-April/msg00154.html libguestfs doesn't correctly handle the case where waitpid receives a SIGCHLD signal and the main program has registered a non-restartable signal handler. In this case waitpid would return -EINTR and we would print an error, but actually we should retry this case. This adds two new internal functions,

Previously posted: https://www.redhat.com/archives/libguestfs/2017-October/msg00032.html This takes a completely different approach. It turns out that POSIX / the shell already defines a special exit code 127 for ‘command not found’. We can make a small adjustment to lib/command.c to return this exit code in that case. Then we just have to modify the db_dump code to test for this exit code. I

Hi all, we're suffering heavily from the "connection reset by peer" problem when running Rsync on WinNT. Great was our relief when seeing Randy's message in http://www.mail-archive.com/rsync@lists.samba.org/msg01918.html I immediately installed the full Cygwin to recompile Rsync, but since I'm far from a C guru, I'm a bit puzzled with the last 2 lines in Randy's

I have been trying to get rsync running correctly on cygwin for the past couple of days. I found a post on the cygwin list that said there was a bug in cygwin when using socketpair() but when I compiled the sample code: (http://jrepp.com/rsync/socketpair.c) both ways it works fine. Here's the article for reference: http://sources.redhat.com/ml/cygwin/2001-08/msg00357.html I am running the

The original implementation utilies select(). In Linux select() typically supports up to 1024 file descriptors. This can be a problem when user tries to boot up many guests. Switching to poll() has minimum impact on existing code and has better scalibility. Up to 8192 file descriptors are supported in the current implementation. Signed-off-by: Wei Liu <wei.liu2@citrix.com> ---

comments? alternatives: sigsetjmp(ugly) and pselect(not portable, available) drawback: additional filedescriptors. Index: serverloop.c =================================================================== RCS file: /home/markus/cvs/ssh/serverloop.c,v retrieving revision 1.82 diff -u -r1.82 serverloop.c --- serverloop.c 10 Oct 2001 22:18:47 -0000 1.82 +++ serverloop.c 11 Oct 2001 18:06:33 -0000 @@

We recently tried upgrading openssh from 2.5.2p2 to 2.9p1 and discovered that it no longer worked to feed the output from a remote command into a pipe, unless the output was short and the pipe was very fast at processing its input. Example 1: ssh remote_machine some_command | less (where "some_command" generates a lot of output) now fails after the first screenful, with a

From: "Richard W.M. Jones" <rjones at redhat.com> If the parent process uses a pipe (or any fd, but pipes are a particular problem), then the recovery process would hold open the file descriptor(s) of the pipe, meaning that it could not be fully closed in the parent. Because the recovery process doesn't use exec(2), this wasn't avoidable even using FD_CLOEXEC. Avoid this

It is just a wrapper around select(2). Signed-off-by: Wei Liu <wei.liu2@citrix.com> --- extras/mini-os/include/posix/poll.h | 1 + extras/mini-os/lib/sys.c | 90 ++++++++++++++++++++++++++++++++++- 2 files changed, 90 insertions(+), 1 deletion(-) create mode 100644 extras/mini-os/include/posix/poll.h diff --git a/extras/mini-os/include/posix/poll.h