1:39 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
(2nd pass)
lxc-enter-namespace allows a process from outside a container to start a
process inside a container. One problem with the current code is the process
running within the container would run with the label of the process that
created it.
For example if the admin process is running as unconfined_t and executes the
following command
# virsh -c lxc:/// lxc-enter-namespace --nolabel dan -- /bin/ps -eZ
LABEL PID TTY TIME CMD
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 1 pts/0 00:00:00 systemd
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 3 pts/1 00:00:00 sh
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 24 ? 00:00:00 systemd-journal
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 29 ? 00:00:00 dhclient
staff_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 47 ? 00:00:00 ps
Note the ps command is running as unconfined_t, After this patch,
virsh -c lxc:/// lxc-enter-namespace dan -- /bin/ps -eZ
LABEL PID TTY TIME CMD
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 1 pts/0 00:00:00 systemd
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 3 pts/1 00:00:00 sh
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 24 ? 00:00:00 systemd-journal
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 32 ? 00:00:00 dhclient
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 38 ? 00:00:00 ps
I also add a --nolabel command to virsh, which can go back to the original
behaviour.
virsh -c lxc:/// lxc-enter-namespace --nolabel dan -- /bin/ps -eZ
LABEL PID TTY TIME CMD
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 1 pts/0 00:00:00 systemd
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 3 pts/1 00:00:00 sh
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 24 ? 00:00:00 systemd-journal
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 32 ? 00:00:00 dhclient
staff_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 37 ? 00:00:00 ps
Everything seems to be working perfectly now.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.13 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iEYEARECAAYFAlEC320ACgkQrlYvE4MpobPDjwCfTLjGarwDOLA333vE+XVp0Zj2
LYkAn3WGX3h309/kJejbE3uvXnIUCKJV
=rwfM
-----END PGP SIGNATURE-----
10:15 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 01/25/2013 02:39 PM, Daniel J Walsh wrote:
(2nd pass)
lxc-enter-namespace allows a process from outside a container to start a
process inside a container. One problem with the current code is the
process running within the container would run with the label of the
process that created it.
For example if the admin process is running as unconfined_t and executes
the following command
# virsh -c lxc:/// lxc-enter-namespace --nolabel dan -- /bin/ps -eZ LABEL
PID TTY TIME CMD system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 1
pts/0 00:00:00 systemd system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 3
pts/1 00:00:00 sh system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 24 ?
00:00:00 systemd-journal system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 29 ?
00:00:00 dhclient staff_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 47 ?
00:00:00 ps
Note the ps command is running as unconfined_t, After this patch,
virsh -c lxc:/// lxc-enter-namespace dan -- /bin/ps -eZ LABEL
PID TTY TIME CMD system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 1
pts/0 00:00:00 systemd system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 3
pts/1 00:00:00 sh system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 24 ?
00:00:00 systemd-journal system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 32 ?
00:00:00 dhclient system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 38 ?
00:00:00 ps
I also add a --nolabel command to virsh, which can go back to the original
behaviour.
virsh -c lxc:/// lxc-enter-namespace --nolabel dan -- /bin/ps -eZ LABEL
PID TTY TIME CMD system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 1
pts/0 00:00:00 systemd system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 3
pts/1 00:00:00 sh system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 24 ?
00:00:00 systemd-journal system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 32 ?
00:00:00 dhclient staff_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 37 ?
00:00:00 ps
Everything seems to be working perfectly now.
Any comment on this?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.13 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iEYEARECAAYFAlEL6iwACgkQrlYvE4MpobN4lACfZF6cBMngf7e9jJGuNkH9HfXC
tiAAoKNC7IuHy5yNrnwKmtS104FeryVl
=N0pN
-----END PGP SIGNATURE-----
10:20 a.m.
On Fri, Jan 25, 2013 at 02:39:25PM -0500, Daniel J Walsh wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
(2nd pass)
lxc-enter-namespace allows a process from outside a container to start a
process inside a container. One problem with the current code is the process
running within the container would run with the label of the process that
created it.
For example if the admin process is running as unconfined_t and executes the
following command
# virsh -c lxc:/// lxc-enter-namespace --nolabel dan -- /bin/ps -eZ
LABEL PID TTY TIME CMD
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 1 pts/0 00:00:00 systemd
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 3 pts/1 00:00:00 sh
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 24 ? 00:00:00 systemd-journal
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 29 ? 00:00:00 dhclient
staff_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 47 ? 00:00:00 ps
Note the ps command is running as unconfined_t, After this patch,
virsh -c lxc:/// lxc-enter-namespace dan -- /bin/ps -eZ
LABEL PID TTY TIME CMD
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 1 pts/0 00:00:00 systemd
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 3 pts/1 00:00:00 sh
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 24 ? 00:00:00 systemd-journal
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 32 ? 00:00:00 dhclient
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 38 ? 00:00:00 ps
I also add a --nolabel command to virsh, which can go back to the original
behaviour.
virsh -c lxc:/// lxc-enter-namespace --nolabel dan -- /bin/ps -eZ
LABEL PID TTY TIME CMD
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 1 pts/0 00:00:00 systemd
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 3 pts/1 00:00:00 sh
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 24 ? 00:00:00 systemd-journal
system_u:system_r:svirt_lxc_net_t:s0:c0.c1023 32 ? 00:00:00 dhclient
staff_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 37 ? 00:00:00 ps
diff --git a/include/libvirt/libvirt-lxc.h
b/include/libvirt/libvirt-lxc.h
index f2c87fb..257637b 100644
--- a/include/libvirt/libvirt-lxc.h
+++ b/include/libvirt/libvirt-lxc.h
@@ -43,6 +43,9 @@ int virDomainLxcEnterNamespace(virDomainPtr domain,
int **oldfdlist,
unsigned int flags);
+int virDomainLxcGetSecurityLabel(virDomainPtr domain,
+ virSecurityLabelPtr seclabel);
+
I'm not sure why you're adding this method - it is identical
to the existing virDomainGetSecurityLabel() method. Just
remove this.
# ifdef __cplusplus
}
# endif
diff --git a/python/generator.py b/python/generator.py
index f853d77..f98818e 100755
--- a/python/generator.py
+++ b/python/generator.py
@@ -551,6 +551,7 @@ skip_function = (
lxc_skip_function = (
"virDomainLxcEnterNamespace",
+ "virDomainLxcGetSecurityLabel",
And remove this.
diff --git a/src/libvirt-lxc.c b/src/libvirt-lxc.c
index f580c3c..a4aff59 100644
--- a/src/libvirt-lxc.c
+++ b/src/libvirt-lxc.c
@@ -41,6 +41,57 @@
__LINE__, info)
/**
+ * virDomainLxcGetSecurityLabel:
+ * @domain: a domain object
+ * @seclabel: pointer to a virSecurityLabel structure
+ *
+ * This API is LXC specific, so it will only work with hypervisor
+ * connections to the LXC driver.
+ *
+ * Get the security label associated with the container @domain.
+ *
+ * Returns 0 on success, or -1 on error
+ */
+int
+virDomainLxcGetSecurityLabel(virDomainPtr domain,
+ virSecurityLabelPtr seclabel)
+{
+ virConnectPtr conn;
+
+ VIR_DEBUG("domain=%p", domain);
+
+ virResetLastError();
+
+ if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
+ virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
+ virDispatchError(NULL);
+ return -1;
+ }
+
+ conn = domain->conn;
+
+ if (conn->flags & VIR_CONNECT_RO) {
+ virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
+ goto error;
+ }
+
+ if (conn->driver->domainGetSecurityLabel) {
+
+ if (conn->driver->domainGetSecurityLabel(domain,
+ seclabel) < 0)
+ goto error;
+
+ return 0;
+ }
+
+ virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
+
+error:
+ virDispatchError(conn);
+ return -1;
+}
And remove this.
diff --git a/src/lxc/lxc_driver.c b/src/lxc/lxc_driver.c
index 1fe8039..fd60712 100644
--- a/src/lxc/lxc_driver.c
+++ b/src/lxc/lxc_driver.c
@@ -1125,6 +1125,7 @@ static int lxcDomainGetSecurityLabel(virDomainPtr dom,
virSecurityLabelPtr secla
{
virLXCDriverPtr driver = dom->conn->privateData;
virDomainObjPtr vm;
+ virLXCDomainObjPrivatePtr priv;
int ret = -1;
lxcDriverLock(driver);
@@ -1162,8 +1163,14 @@ static int lxcDomainGetSecurityLabel(virDomainPtr dom,
virSecurityLabelPtr secla
* LXC monitor hasn't seen SIGHUP/ERR on poll().
*/
if (virDomainObjIsActive(vm)) {
+ priv = vm->privateData;
+ if (!priv->initpid) {
+ virReportError(VIR_ERR_OPERATION_INVALID, "%s",
+ _("Init pid is not yet available"));
+ goto cleanup;
+ }
Identation has gone a bit screwy here.
if
(virSecurityManagerGetProcessLabel(driver->securityManager,
- vm->def, vm->pid, seclabel) < 0)
{
+ vm->def, priv->initpid, seclabel)
< 0) {
Yep, this makes sense
diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c
index 026dac1..c53f817 100644
--- a/tools/virsh-domain.c
+++ b/tools/virsh-domain.c
@@ -56,6 +56,10 @@
#include "virtypedparam.h"
#include "virxml.h"
+#ifdef WITH_SELINUX
+#include <selinux/selinux.h>
+#endif
You need to use '# include' rather than "#include" in order
to get 'make syntax-check' pass
@@ -7678,6 +7682,7 @@ static const vshCmdInfo
info_lxc_enter_namespace[] = {
static const vshCmdOptDef opts_lxc_enter_namespace[] = {
{"domain", VSH_OT_DATA, VSH_OFLAG_REQ, N_("domain name, id or
uuid")},
+ {"nolabel", VSH_OT_BOOL, 0, N_("Do Not Change Process Label
")},
{"cmd", VSH_OT_ARGV, VSH_OFLAG_REQ, N_("namespace")},
{NULL, 0, 0, NULL}
};
@@ -7686,6 +7691,7 @@ static bool
cmdLxcEnterNamespace(vshControl *ctl, const vshCmd *cmd)
{
virDomainPtr dom = NULL;
+ virSecurityLabelPtr seclabel;
If you declare this as 'virSecurityLabel seclabel'...
bool ret = false;
const vshCmdOpt *opt = NULL;
char **cmdargv = NULL;
@@ -7694,6 +7700,7 @@ cmdLxcEnterNamespace(vshControl *ctl, const vshCmd *cmd)
int nfdlist;
int *fdlist;
size_t i;
+ int label = false;
dom = vshCommandOptDomain(ctl, cmd, NULL);
if (dom == NULL)
@@ -7715,12 +7722,30 @@ cmdLxcEnterNamespace(vshControl *ctl, const vshCmd *cmd)
if ((nfdlist = virDomainLxcOpenNamespace(dom, &fdlist, 0)) < 0)
goto cleanup;
+ if (! vshCommandOptBool(cmd, "nolabel")) {
+ if (VIR_ALLOC(seclabel) < 0)
+ goto cleanup;
...this allocation can be removed.
+ label = true;
+ if (virDomainLxcGetSecurityLabel(dom, seclabel) < 0)
Just replace this call with virDomainGetSecurityLabel()
cleanup:
+ if (label)
+ VIR_FREE(seclabel);
And then this can go too
Regards,
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
4311
days inactive
4318
days old
2 comments
2 participants
participants (2)
-
Daniel J Walsh -
Daniel P. Berrange