This patch is in response to:
https://bugzilla.redhat.com/show_bug.cgi?id=818467
If a caller to virCommandRun doesn't ask for the exitstatus of the
program it's running, the virCommand functions assume that they should
log an error message and return failure if the exit code isn't
0. However, only the commandline and exit status are logged, while
potentially useful information sent by the program to stderr is
discarded.
Fortunately, virCommandRun is already checking if the caller had asked
for stderr to be saved and, if not, sets things up to save it in
*cmd->errbuf. This makes it fairly simple for virCommandWait to
include *cmd->errbuf in the error log (there are still other callers
that don't setup errbuf, and even virCommandRun won't set it up if the
command is being daemonized, so we have to check that it's non-zero).
Add function virCommandNewVAList which is equivalent to the
virCommandNewArgList but with va_list instead of a variable number
of arguments.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Per the FSF address could be changed from time to time, and GNU
recommends the following now: (http://www.gnu.org/licenses/gpl-howto.html)
You should have received a copy of the GNU General Public License
along with Foobar. If not, see <http://www.gnu.org/licenses/>.
This patch removes the explicit FSF address, and uses above instead
(of course, with inserting 'Lesser' before 'General').
Except a bunch of files for security driver, all others are changed
automatically, the copyright for securify files are not complete,
that's why to do it manually:
src/security/security_selinux.h
src/security/security_driver.h
src/security/security_selinux.c
src/security/security_apparmor.h
src/security/security_apparmor.c
src/security/security_driver.c
This removes nearly all the per-file error reporting macros
from the code in src/util/. A few custom macros remain for the
case, where the file needs to report errors with a variety of
different codes or parameters
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
There is a theoretical problem of an extreme bug where we can get
into deadlock due to command handshaking. Thanks to a pair of pipes,
we have a situation where the parent thinks the child reported an
error and is waiting for a message from the child to explain the
error; but at the same time the child thinks it reported success
and is waiting for the parent to acknowledge the success; so both
processes are now blocked.
Thankfully, I don't think this deadlock is possible without at
least one other bug in the code, but I did see exactly that sort
of situation prior to commit da831af - I saw a backtrace where a
double close bug in the parent caused the parent to read from the
wrong fd and assume the child failed, even though the child really
sent success.
This potential deadlock is not quite like commit 858c247 (a deadlock
due to multiple readers on one pipe preventing a write from completing),
although the solution is similar - always close unused pipe fds before
blocking, rather than after.
* src/util/command.c (virCommandHandshakeWait): Close unused fds
sooner.
It is possible to deadlock libvirt by having a domain with XML
longer than PIPE_BUF, and by writing a hook script that closes
stdin early. This is because libvirt was keeping a copy of the
child's stdin read fd open, which means the write fd in the
parent will never see EPIPE (remember, libvirt should always be
run with SIGPIPE ignored, so we should never get a SIGPIPE signal).
Since there is no error, libvirt blocks waiting for a write to
complete, even though the only reader is also libvirt. The
solution is to ensure that only the child can act as a reader
before the parent does any writes; and then dealing with the
fallout of dealing with EPIPE.
Thankfully, this is not a security hole - since the only way to
trigger the deadlock is to install a custom hook script, anyone
that already has privileges to install a hook script already has
privileges to do any number of other equally disruptive things
to libvirt; it would only be a security hole if an unprivileged
user could install a hook script to DoS a privileged user.
* src/util/command.c (virCommandRun): Close parent's copy of child
read fd earlier.
(virCommandProcessIO): Don't let EPIPE be fatal; the child may
be done parsing input.
* tests/commandhelper.c (main): Set up a SIGPIPE situation.
* tests/commandtest.c (test20): Trigger it.
* tests/commanddata/test20.log: New file.
EBADF errors are logged as warnings as they normally indicate a double
close bug. This patch also provides VIR_MASS_CLOSE helper to be user in
the only case of mass close after fork when EBADF should rather be
ignored.
KAMEZAWA Hiroyuki reported a nasty double-free bug when virCommand
is used to convert a string into input to a child command. The
problem is that the poll() loop of virCommandProcessIO would close()
the write end of the pipe in order to let the child see EOF, then
the caller virCommandRun() would also close the same fd number, with
the second close possibly nuking an fd opened by some other thread
in the meantime. This in turn can have all sorts of bad effects.
The bug has been present since the introduction of virCommand in
commit f16ad06f.
This is based on his first attempt at a patch, at
https://bugzilla.redhat.com/show_bug.cgi?id=823716
* src/util/command.c (_virCommand): Drop inpipe member.
(virCommandProcessIO): Add argument, to avoid closing caller's fd
without informing caller.
(virCommandRun, virCommandNewArgs): Adjust clients.
Currently, we are logging only one side of pipes we
create in virCommandRequireHandshake(); This is enough
in cases where pipe2() returns two consecutive FDs. However,
it is not guaranteed and it may return any FDs.
Therefore, it's wise to log the other ends as well.
The code is splattered with a mix of
sizeof foo
sizeof (foo)
sizeof(foo)
Standardize on sizeof(foo) and add a syntax check rule to
enforce it
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Nuke the last vestiges of printing pid_t values with the wrong
types, at least in code compiled on mingw64. There may be other
places, but for now they are only compiled on systems where the
existing %d doesn't trigger gcc warnings.
* src/rpc/virnetsocket.c (virNetSocketNew): Use %lld and casting,
rather than assuming any particular int type for pid_t.
* src/util/command.c (virCommandRunAsync, virPidWait)
(virPidAbort): Likewise.
(verify): Drop a now stale assertion.
Sometimes, its easier to run children with 2>&1 in shell notation,
and just deal with stdout and stderr interleaved. This was already
possible for fd handling; extend it to also work when doing string
capture of a child process.
* docs/internals/command.html.in: Document this.
* src/util/command.c (virCommandSetErrorBuffer): Likewise.
(virCommandRun, virExecWithHook): Implement it.
* tests/commandtest.c (test14): Test it.
* daemon/remote.c (remoteDispatchAuthPolkit): Use new command
feature.
If we are building not on a WIN32 architecture and without HAVE_CAPNG
virSetCapabilities has unused argument and virClearCapabilities
is unused as well.
This patch introduces virSetCapabilities() function and implements
virCommandAllowCap() function.
Existing virClearCapabilities() is function to clear all capabilities.
Instead virSetCapabilities() is function to set arbitrary capabilities.
Signed-off-by: Taku Izumi <izumi.taku@jp.fujitsu.com>
Signed-off-by: Shota Hirae <m11g1401@hibikino.ne.jp>
POLLIN and POLLHUP are not mutually exclusive. Currently the following
seems possible: the child writes 3K to its stdout or stderr pipe, and
immediately closes it. We get POLLIN|POLLHUP (I'm not sure that's possible
on Linux, but SUSv4 seems to allow it). We read 1K and throw away the
rest.
When poll() returns and we're about to check the /revents/ member in a
given array element, let's map all the revents bits to two (independent)
ideas: "let's attempt to read()", and "let's attempt to write()". This
should cover all errors, EOFs, and normal conditions; the read()/write()
call should report any pending error.
Under this approach, both POLLHUP and POLLERR are mapped to "needs read()"
if we're otherwise prepared for POLLIN. POLLERR also maps to "needs
write()" if we're otherwise prepared for POLLOUT. The rest of the mappings
(POLLPRI etc.) would be easy, but probably useless for pipes.
Additionally, SUSv4 doesn't appear to forbid POLLIN|POLLERR (or
POLLOUT|POLLERR) set simultaneously. One could argue that the read() or
write() call would return without blocking in these cases (with an error),
so POLLIN / POLLOUT would be justified beside POLLERR.
The code now penalizes POLLIN|POLLERR differently from plain POLLERR. The
former (ie. read() returning -1) is terminal and we jump to cleanup, while
plain POLLERR masks only the affected file descriptor for the future.
Let's unify those.
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Commit db371a2 mistakenly added new functions inside a #ifndef WIN32
guard, even though they are needed on all platforms.
* src/util/command.c (virCommandFDSet): Move outside WIN32
conditional.
Currently, virCommand implementation uses FD_ macros from
sys/select.h. However, those cannot handle more opened files
than FD_SETSIZE. Therefore switch to generalized implementation
based on array of integers.
It is a good practise to set revents to zero before doing any poll().
Moreover, we should check if event we waited for really occurred or
if any of fds we were polling on didn't encountered hangup.
virBufferContentAndReset (intentionally) returns NULL for a buffer
with no content, but it is feasible to invoke a command with an
explicit empty string.
* src/util/command.c (virCommandAddEnvBuffer): Reject empty string.
(virCommandAddArgBuffer): Allow explicit empty argument.
* tests/commandtest.c (test9): Test it.
* tests/commanddata/test9.log: Adjust.
virCommandTransferFD promises that the fd is no longer owned by
the caller. Normally, we want the fd to remain open until the
child runs, but in error situations, we must close it earlier.
* src/util/command.c (virCommandTransferFD): Close fd now if we
can't track it to close later.
(virCommandKeepFD): Adjust helper to make this easier.
The functions for manipulating pidfiles are in util/util.{c,h}.
We will shortly be adding some further pidfile related functions.
To avoid further growing util.c, this moves the pidfile related
functions into a dedicated virpidfile.{c,h}. The functions are
also all renamed to have 'virPidFile' as their name prefix
* util/util.h, util/util.c: Remove all pidfile code
* util/virpidfile.c, util/virpidfile.h: Add new APIs for pidfile
handling.
* lxc/lxc_controller.c, lxc/lxc_driver.c, network/bridge_driver.c,
qemu/qemu_process.c: Add virpidfile.h include and adapt for API
renames
Coverity complained that 395 out of 409 virAsprintf calls are
checked, and therefore assumed that the remaining cases are bugs
waiting to happen. But in each of these cases, a failed virAsprintf
will properly set the target string to NULL, and pass on that
failure to the caller, without wasting efforts to check the call.
Adding the ignore_value silences Coverity.
* src/conf/domain_audit.c (virDomainAuditGetRdev): Ignore
virAsprintf return value, when it behaves like we need.
* src/network/bridge_driver.c (networkDnsmasqLeaseFileNameDefault)
(networkRadvdConfigFileName, networkBridgeDummyNicName)
(networkRadvdPidfileBasename): Likewise.
* src/util/storage_file.c (absolutePathFromBaseFile): Likewise.
* src/openvz/openvz_driver.c (openvzGenerateContainerVethName):
Likewise.
* src/util/command.c (virCommandTranslateStatus): Likewise.
Although most functions in libvirt return 0 on success and < 0 on
failure, there are a few functions lingering around that return errno
(a positive value) on failure, and sometimes code calling those
functions incorrectly assumes the <0 standard. I noticed one of these
the other day when auditing networkStartDhcpDaemon after Guido Gunther
found a place where success was improperly returned on failure (that
patch has been acked and is pending a push). The problem was that it
expected the return value from virFileReadPid to be < 0 on failure,
but it was actually positive (it was also neglected to set the return
code in this case, similar to the bug found by Guido).
This all led to the fact that *all* of the virFile*Pid functions in
util.c are returning errno on failure. This patch remedies that
problem by changing them all to return -errno on failure, and makes
any necessary changes to callers of the functions. (In the meantime, I
also properly set the return code on failure of virFileReadPid in
networkStartDhcpDaemon).
Since libvirt is multi-threaded, we should use FD_CLOEXEC as much
as possible in the parent, and only relax fds to inherited after
forking, to avoid leaking an fd created in one thread to a fork
run in another thread. This gets us closer to that ideal, by
making virCommand automatically clear FD_CLOEXEC on fds intended
for the child, as well as avoiding a window of time with non-cloexec
pipes created for capturing output.
* src/util/command.c (virExecWithHook): Use CLOEXEC in parent. In
child, guarantee that all fds to pass to child are inheritable.
(getDevNull): Use CLOEXEC.
(prepareStdFd): New helper function.
(virCommandRun, virCommandRequireHandshake): Use pipe2.
* src/qemu/qemu_command.c (qemuBuildCommandLine): Simplify caller.
We already have a precedent of function documentation in C files,
where it is closer to the implementation (witness libvirt.h vs.
libvirt.c); maintaining docs in both files risks docs going stale.
While I was at it, I used consistent doxygen style on all comments.
* src/util/command.h: Remove duplicate docs, and move unique
documentation...
* src/util/command.c: ...here.
Suggested by Matthias Bolte.
Enforce the recent flags cleanups - we want to use 'unsigned int flags'
in any of our APIs (except where backwards compatibility is important,
in the public migration APIs), and that all flags are checked for
validity (except when there are stub functions that completely
ignore the flags argument).
There are a few minor tweaks done here to avoid false positives:
signed arguments passed to open() are renamed oflags, and flags
arguments that are legitimately ignored are renamed flags_unused.
* cfg.mk (sc_flags_usage): New rule.
(exclude_file_name_regexp--sc_flags_usage): And a few exemptions.
(sc_flags_debug): Tweak wording.
* src/util/iohelper.c (runIO, main): Rename variable.
* src/util/util.c (virSetInherit): Likewise.
* src/fdstream.h (virFDStreamOpenFile, virFDStreamCreateFile):
Likewise.
* src/fdstream.c (virFDStreamOpenFileInternal)
(virFDStreamOpenFile, virFDStreamCreateFile): Likewise.
* src/util/command.c (virExecWithHook) [WIN32]: Likewise.
* src/util/util.c (virFileOpenAs, virDirCreate) [WIN32]: Likewise.
* src/locking/lock_manager.c (virLockManagerPluginNew)
[!HAVE_DLFCN_H]: Likewise.
* src/locking/lock_driver_nop.c (virLockManagerNopNew)
(virLockManagerNopAddResource, virLockManagerNopAcquire)
(virLockManagerNopRelease, virLockManagerNopInquire): Likewise.
When using virCommandRunAsync and saving the pid for later, it
is useful to be able to reap that pid in the same way that it
would have been auto-reaped by virCommand if we had passed
NULL for the pid argument in the first place.
* src/util/command.c (virPidWait, virPidAbort): New functions,
created from...
(virCommandWait, virCommandAbort): ...bodies of these.
(includes): Drop duplicate <stdlib.h>. Ensure that our pid_t
assumptions hold.
(virCommandRunAsync): Improve documentation.
* src/util/command.h (virPidWait, virPidAbort): New prototypes.
* src/libvirt_private.syms: Export them.
* docs/internals/command.html.in: Document them.
Previously, the parent process opened 'null' to /dev/null, then
the child process closes 'null' as well as 'childout'. But if
childout was set to be null, then this is a double close. At
least the double close was confined to the child process after a
fork, and therefore there is no risk of another thread opening
an fd of the same value to be bitten by the double close, but it
is always better to avoid double-close to begin with.
Additionally, if all three fds were specified, then opening
'null' was wasted.
This patch fixes things to lazily open null on the first use,
then guarantees it gets closed exactly once.
* src/util/command.c (getDevNull): New helper function.
(virExecWithHook): Use it to avoid spurious opens and double close.
This also reduces malloc pressure for invoking a child when
VIR_DEBUG is enabled.
* src/util/command.c (virExecWithHook): Drop debug, since the only
caller (virCommandRunAsync) also prints debug info.
Similar in nature to commit fd21ecfd, which shut up valgrind.
sigaction is apparently a nasty interface for analyzer tools,
at least for how many false positives it generates.
* src/util/command.c (virExecWithHook): Initialize entire var, since
coverity gripes about the (unused and non-standard) sa_restorer.
Seems reasonable to have all command wrappers in the same place
v2:
Dont move SetInherit
v3:
Comment spelling fix
Adjust WARN0 comment
Remove spurious #include movement
Don't include sys/types.h
Combine virExec enums
Signed-off-by: Cole Robinson <crobinso@redhat.com>
Coverity complained that infd could be -1 at the point where it is
passed to write, when in reality, this code can only be reached if
infd is non-negative.
* src/util/command.c (virCommandProcessIO): Help out coverity.
Detected by Coverity. While it is possible on OOM condition, as
well as with bad code that passes binary == NULL, it is unlikely
to be encountered in the wild.
* src/util/command.c (virCommandNewArgList): Don't leak memory.
Allow the parent process to perform a bi-directional handshake
with the child process during fork/exec. The child process
will fork and do its initial setup. Immediately prior to the
exec(), it will stop & wait for a handshake from the parent
process. The parent process will spawn the child and wait
until the child reaches the handshake point. It will do
whatever extra setup work is required, before signalling the
child to continue.
The implementation of this is done using two pairs of blocking
pipes. The first pair is used to block the parent, until the
child writes a single byte. Then the second pair pair is used
to block the child, until the parent confirms with another
single byte.
* src/util/command.c, src/util/command.h,
src/libvirt_private.syms: Add APIs to perform a handshake
Mingw execve() has a broken signature. Disable this
function until gnulib fixes the signature, since we
don't really need this on Win32 anyway.
* src/util/command.c: Disable virCommandExec on Win32
These VIR_XXXX0 APIs make us confused, use the non-0-suffix APIs instead.
How do these coversions works? The magic is using the gcc extension of ##.
When __VA_ARGS__ is empty, "##" will swallow the "," in "fmt," to
avoid compile error.
example: origin after CPP
high_level_api("%d", a_int) low_level_api("%d", a_int)
high_level_api("a string") low_level_api("a string")
About 400 conversions.
8 special conversions:
VIR_XXXX0("") -> VIR_XXXX("msg") (avoid empty format) 2 conversions
VIR_XXXX0(string_literal_with_%) -> VIR_XXXX(%->%%) 0 conversions
VIR_XXXX0(non_string_literal) -> VIR_XXXX("%s", non_string_literal)
(for security) 6 conversions
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Actually execs the argv/env we've generated, replacing the current process.
Kind of has a limited usage, but allows us to use virCommand in LXC
driver to launch the 'init' process
