Bring in geli suspend/resume functionality (finally).

Before this change if you wanted to suspend your laptop and be sure that your
encryption keys are safe, you had to stop all processes that use file system
stored on encrypted device, unmount the file system and detach geli provider.

This isn't very handy. If you are a lucky user of a laptop where suspend/resume
actually works with FreeBSD (I'm not!) you most likely want to suspend your
laptop, because you don't want to start everything over again when you turn
your laptop back on.

And this is where geli suspend/resume steps in. When you execute:

	# geli suspend -a

geli will wait for all in-flight I/O requests, suspend new I/O requests, remove
all geli sensitive data from the kernel memory (like encryption keys) and will
wait for either 'geli resume' or 'geli detach'.

Now with no keys in memory you can suspend your laptop without stopping any
processes or unmounting any file systems.

When you resume your laptop you have to resume geli devices using 'geli resume'
command. You need to provide your passphrase, etc. again so the keys can be
restored and suspended I/O requests released.

Of course you need to remember that 'geli suspend' won't clear file system
cache and other places where data from your geli-encrypted file system might be
present. But to get rid of those stopping processes and unmounting file system
won't help either - you have to turn your laptop off. Be warned.

Also note, that suspending geli device which contains file system with geli
utility (or anything used by 'geli resume') is not very good idea, as you won't
be able to resume it - when you execute geli(8), the kernel will try to read it
and this read I/O request will be suspended.

1 files changed, 213 insertions, 2 deletions

diff --git a/sys/geom/eli/g_eli_ctl.c b/sys/geom/eli/g_eli_ctl.c
index 02ede13..7147b27 100644
--- a/sys/geom/eli/g_eli_ctl.c
+++ b/sys/geom/eli/g_eli_ctl.c
@@ -217,7 +217,7 @@ g_eli_ctl_detach(struct gctl_req *req, struct g_class *mp)
 			sc->sc_flags |= G_ELI_FLAG_RW_DETACH;
 			sc->sc_geom->access = g_eli_access;
 		} else {
-			error = g_eli_destroy(sc, *force);
+			error = g_eli_destroy(sc, *force ? TRUE : FALSE);
 			if (error != 0) {
 				gctl_error(req,
 				    "Cannot destroy device %s (error=%d).",
@@ -700,6 +700,213 @@ g_eli_ctl_delkey(struct gctl_req *req, struct g_class *mp)
 }
 
 static int
+g_eli_suspend_one(struct g_eli_softc *sc)
+{
+	struct g_eli_worker *wr;
+
+	g_topology_assert();
+
+	if (sc == NULL)
+		return (ENOENT);
+	if (sc->sc_flags & G_ELI_FLAG_ONETIME)
+		return (EOPNOTSUPP);
+
+	mtx_lock(&sc->sc_queue_mtx);
+	if (sc->sc_flags & G_ELI_FLAG_SUSPEND) {
+		mtx_unlock(&sc->sc_queue_mtx);
+		return (EALREADY);
+	}
+	sc->sc_flags |= G_ELI_FLAG_SUSPEND;
+	wakeup(sc);
+	for (;;) {
+		LIST_FOREACH(wr, &sc->sc_workers, w_next) {
+			if (wr->w_active)
+				break;
+		}
+		if (wr == NULL)
+			break;
+		/* Not all threads suspended. */
+		msleep(&sc->sc_workers, &sc->sc_queue_mtx, PRIBIO,
+		    "geli:suspend", 0);
+	}
+	/*
+	 * Clear sensitive data on suspend, they will be recovered on resume.
+	 */
+	bzero(sc->sc_mkey, sizeof(sc->sc_mkey));
+	bzero(sc->sc_ekeys,
+	    sc->sc_nekeys * (sizeof(uint8_t *) + G_ELI_DATAKEYLEN));
+	free(sc->sc_ekeys, M_ELI);
+	sc->sc_ekeys = NULL;
+	bzero(sc->sc_akey, sizeof(sc->sc_akey));
+	bzero(&sc->sc_akeyctx, sizeof(sc->sc_akeyctx));
+	bzero(sc->sc_ivkey, sizeof(sc->sc_ivkey));
+	bzero(&sc->sc_ivctx, sizeof(sc->sc_ivctx));
+	mtx_unlock(&sc->sc_queue_mtx);
+	G_ELI_DEBUG(0, "%s has been suspended.", sc->sc_name);
+	return (0);
+}
+
+static void
+g_eli_ctl_suspend(struct gctl_req *req, struct g_class *mp)
+{
+	struct g_eli_softc *sc;
+	int *all, *nargs;
+	int error;
+
+	g_topology_assert();
+
+	nargs = gctl_get_paraml(req, "nargs", sizeof(*nargs));
+	if (nargs == NULL) {
+		gctl_error(req, "No '%s' argument.", "nargs");
+		return;
+	}
+	all = gctl_get_paraml(req, "all", sizeof(*all));
+	if (all == NULL) {
+		gctl_error(req, "No '%s' argument.", "all");
+		return;
+	}
+	if (!*all && *nargs == 0) {
+		gctl_error(req, "Too few arguments.");
+		return;
+	}
+
+	if (*all) {
+		struct g_geom *gp, *gp2;
+
+		LIST_FOREACH_SAFE(gp, &mp->geom, geom, gp2) {
+			sc = gp->softc;
+			if (sc->sc_flags & G_ELI_FLAG_ONETIME)
+				continue;
+			error = g_eli_suspend_one(sc);
+			if (error != 0)
+				gctl_error(req, "Not fully done.");
+		}
+	} else {
+		const char *prov;
+		char param[16];
+		int i;
+
+		for (i = 0; i < *nargs; i++) {
+			snprintf(param, sizeof(param), "arg%d", i);
+			prov = gctl_get_asciiparam(req, param);
+			if (prov == NULL) {
+				G_ELI_DEBUG(0, "No 'arg%d' argument.", i);
+				continue;
+			}
+
+			sc = g_eli_find_device(mp, prov);
+			if (sc == NULL) {
+				G_ELI_DEBUG(0, "No such provider: %s.", prov);
+				continue;
+			}
+			error = g_eli_suspend_one(sc);
+			if (error != 0)
+				gctl_error(req, "Not fully done.");
+		}
+	}
+}
+
+static void
+g_eli_ctl_resume(struct gctl_req *req, struct g_class *mp)
+{
+	struct g_eli_metadata md;
+	struct g_eli_softc *sc;
+	struct g_provider *pp;
+	struct g_consumer *cp;
+	const char *name;
+	u_char *key, mkey[G_ELI_DATAIVKEYLEN];
+	int *nargs, keysize, error;
+	u_int nkey;
+
+	g_topology_assert();
+
+	nargs = gctl_get_paraml(req, "nargs", sizeof(*nargs));
+	if (nargs == NULL) {
+		gctl_error(req, "No '%s' argument.", "nargs");
+		return;
+	}
+	if (*nargs != 1) {
+		gctl_error(req, "Invalid number of arguments.");
+		return;
+	}
+
+	name = gctl_get_asciiparam(req, "arg0");
+	if (name == NULL) {
+		gctl_error(req, "No 'arg%u' argument.", 0);
+		return;
+	}
+	sc = g_eli_find_device(mp, name);
+	if (sc == NULL) {
+		gctl_error(req, "Provider %s is invalid.", name);
+		return;
+	}
+	if (!(sc->sc_flags & G_ELI_FLAG_SUSPEND)) {
+		gctl_error(req, "Provider %s not suspended.", name);
+		return;
+	}
+	cp = LIST_FIRST(&sc->sc_geom->consumer);
+	pp = cp->provider;
+	error = g_eli_read_metadata(mp, pp, &md);
+	if (error != 0) {
+		gctl_error(req, "Cannot read metadata from %s (error=%d).",
+		    name, error);
+		return;
+	}
+	if (md.md_keys == 0x00) {
+		bzero(&md, sizeof(md));
+		gctl_error(req, "No valid keys on %s.", pp->name);
+		return;
+	}
+
+	key = gctl_get_param(req, "key", &keysize);
+	if (key == NULL || keysize != G_ELI_USERKEYLEN) {
+		bzero(&md, sizeof(md));
+		gctl_error(req, "No '%s' argument.", "key");
+		return;
+	}
+
+	error = g_eli_mkey_decrypt(&md, key, mkey, &nkey);
+	bzero(key, keysize);
+	if (error == -1) {
+		bzero(&md, sizeof(md));
+		gctl_error(req, "Wrong key for %s.", pp->name);
+		return;
+	} else if (error > 0) {
+		bzero(&md, sizeof(md));
+		gctl_error(req, "Cannot decrypt Master Key for %s (error=%d).",
+		    pp->name, error);
+		return;
+	}
+	G_ELI_DEBUG(1, "Using Master Key %u for %s.", nkey, pp->name);
+
+	mtx_lock(&sc->sc_queue_mtx);
+	/* Restore sc_mkey, sc_ekeys, sc_akey and sc_ivkey. */
+	g_eli_mkey_propagate(sc, mkey);
+	bzero(mkey, sizeof(mkey));
+	bzero(&md, sizeof(md));
+	/* Restore sc_akeyctx. */
+	if (sc->sc_flags & G_ELI_FLAG_AUTH) {
+		SHA256_Init(&sc->sc_akeyctx);
+		SHA256_Update(&sc->sc_akeyctx, sc->sc_akey,
+		    sizeof(sc->sc_akey));
+	}
+	/* Restore sc_ivctx. */
+	switch (sc->sc_ealgo) {
+	case CRYPTO_AES_XTS:
+		break;
+	default:
+		SHA256_Init(&sc->sc_ivctx);
+		SHA256_Update(&sc->sc_ivctx, sc->sc_ivkey,
+		    sizeof(sc->sc_ivkey));
+		break;
+	}
+	sc->sc_flags &= ~G_ELI_FLAG_SUSPEND;
+	mtx_unlock(&sc->sc_queue_mtx);
+	G_ELI_DEBUG(1, "Resumed %s.", pp->name);
+	wakeup(sc);
+}
+
+static int
 g_eli_kill_one(struct g_eli_softc *sc)
 {
 	struct g_provider *pp;
@@ -749,7 +956,7 @@ g_eli_kill_one(struct g_eli_softc *sc)
 	}
 	if (error == 0)
 		G_ELI_DEBUG(0, "%s has been killed.", pp->name);
-	g_eli_destroy(sc, 1);
+	g_eli_destroy(sc, TRUE);
 	return (error);
 }
 
@@ -839,6 +1046,10 @@ g_eli_config(struct gctl_req *req, struct g_class *mp, const char *verb)
 		g_eli_ctl_setkey(req, mp);
 	else if (strcmp(verb, "delkey") == 0)
 		g_eli_ctl_delkey(req, mp);
+	else if (strcmp(verb, "suspend") == 0)
+		g_eli_ctl_suspend(req, mp);
+	else if (strcmp(verb, "resume") == 0)
+		g_eli_ctl_resume(req, mp);
 	else if (strcmp(verb, "kill") == 0)
 		g_eli_ctl_kill(req, mp);
 	else