roots: reimplement transid_max_nocache using extent tree root

ROOT_TREE contains the ROOT_ITEM for EXTENT_TREE.  Every modification
(that we care about) to a btrfs must go through EXTENT_TREE, and must
modify the page in ROOT_TREE pointing to the root of EXTENT_TREE...
which makes that a very good source for the filesystem transid.

Remove the loop and the root lookups, and just look at one item for
max_transid.

Also note that every caller of transid_max_nocache() immediately
feeds the return value to m_transid_re.update(), so don't do that
inside transid_max_nocache().

Signed-off-by: Zygo Blaxell <bees@furryterror.org>

src/bees-context.cc

@@ -773,11 +773,42 @@ BeesResolveAddrResult::BeesResolveAddrResult()
 {
 }
 
+void
+BeesContext::wait_for_balance()
+{
+	Timer balance_timer;
+	BEESNOTE("WORKAROUND: waiting for balance to stop");
+	while (true) {
+		btrfs_ioctl_balance_args args;
+		memset_zero<btrfs_ioctl_balance_args>(&args);
+		const int ret = ioctl(root_fd(), BTRFS_IOC_BALANCE_PROGRESS, &args);
+		if (ret < 0) {
+			// Either can't get balance status or not running, exit either way
+			break;
+		}
+
+		if (!(args.state & BTRFS_BALANCE_STATE_RUNNING)) {
+			// Balance not running, doesn't matter if paused or cancelled
+			break;
+		}
+
+		BEESLOGDEBUG("WORKAROUND: Waiting " << balance_timer << "s for balance to stop");
+		sleep(BEES_BALANCE_POLL_INTERVAL);
+	}
+}
+
 BeesResolveAddrResult
 BeesContext::resolve_addr_uncached(BeesAddress addr)
 {
 	THROW_CHECK1(invalid_argument, addr, !addr.is_magic());
 	THROW_CHECK0(invalid_argument, !!root_fd());
+
+	// Is there a bug where resolve and balance cause a crash (BUG_ON at fs/btrfs/ctree.c:1227)?
+	// Apparently yes, and more than one.
+	// Wait for the balance to finish before we run LOGICAL_INO
+	wait_for_balance();
+
+	// Time how long this takes
 	Timer resolve_timer;
 
 	// There is no performance benefit if we restrict the buffer size.

src/bees-roots.cc

@@ -207,43 +207,37 @@ uint64_t
 BeesRoots::transid_max_nocache()
 {
 	uint64_t rv = 0;
-	uint64_t root = BTRFS_FS_TREE_OBJECTID;
-	BEESNOTE("Calculating transid_max (" << rv << " as of root " << root << ")");
-	BEESTRACE("Calculating transid_max...");
-
-	rv = btrfs_get_root_transid(root);
-
-	// XXX:  Do we need any of this?  Or is
-	// m_transid_re.update(btrfs_get_root_transid(BTRFS_FS_TREE_OBJECTID)) good enough?
+	BEESNOTE("Calculating transid_max");
+	BEESTRACE("Calculating transid_max");
 
+	// We look for the root of the extent tree and read its transid.
+	// Should run in O(1) time and be fairly reliable.
 	BtrfsIoctlSearchKey sk;
 	sk.tree_id = BTRFS_ROOT_TREE_OBJECTID;
-	sk.min_type = sk.max_type = BTRFS_ROOT_BACKREF_KEY;
-	sk.min_objectid = root;
+	sk.min_type = sk.max_type = BTRFS_ROOT_ITEM_KEY;
+	sk.min_objectid = sk.max_objectid = BTRFS_EXTENT_TREE_OBJECTID;
 
 	while (true) {
 		sk.nr_items = 1024;
+		BEESTRACE("transid_max search sk " << sk);
 		sk.do_ioctl(m_ctx->root_fd());
 
 		if (sk.m_result.empty()) {
 			break;
 		}
 
+		// We are just looking for the highest transid on the filesystem.
+		// We don't care which object it comes from.
 		for (auto i : sk.m_result) {
 			sk.next_min(i);
-			if (i.type == BTRFS_ROOT_BACKREF_KEY) {
-				if (i.transid > rv) {
-					BEESLOGDEBUG("transid_max root " << i.objectid << " parent " << i.offset << " transid " << i.transid);
-					BEESCOUNT(transid_max_miss);
-				}
-				root = i.objectid;
-			}
 			if (i.transid > rv) {
 				rv = i.transid;
 			}
 		}
 	}
-	m_transid_re.update(rv);
+
+	// transid must be greater than zero, or we did something very wrong
+	THROW_CHECK1(runtime_error, rv, rv > 0);
 	return rv;
 }
 
@@ -422,13 +416,15 @@ BeesRoots::crawl_thread()
 	BEESNOTE("tracking transid");
 	auto last_count = m_transid_re.count();
 	while (true) {
-		// Measure current transid
+		BEESTRACE("Measure current transid");
 		catch_all([&]() {
+			BEESTRACE("calling transid_max_nocache");
 			m_transid_re.update(transid_max_nocache());
 		});
 
-		// Make sure we have a full complement of crawlers
+		BEESTRACE("Make sure we have a full complement of crawlers");
 		catch_all([&]() {
+			BEESTRACE("calling insert_new_crawl");
 			insert_new_crawl();
 		});
 
@@ -496,19 +492,24 @@ BeesRoots::insert_new_crawl()
 	unique_lock<mutex> lock(m_mutex);
 	set<uint64_t> excess_roots;
 	for (auto i : m_root_crawl_map) {
+		BEESTRACE("excess_roots.insert(" << i.first << ")");
 		excess_roots.insert(i.first);
 	}
 	lock.unlock();
 
 	while (new_bcs.m_root) {
+		BEESTRACE("excess_roots.erase(" << new_bcs.m_root << ")");
 		excess_roots.erase(new_bcs.m_root);
+		BEESTRACE("insert_root(" << new_bcs << ")");
 		insert_root(new_bcs);
 		BEESCOUNT(crawl_create);
+		BEESTRACE("next_root(" << new_bcs.m_root << ")");
 		new_bcs.m_root = next_root(new_bcs.m_root);
 	}
 
 	for (auto i : excess_roots) {
 		new_bcs.m_root = i;
+		BEESTRACE("crawl_state_erase(" << new_bcs << ")");
 		crawl_state_erase(new_bcs);
 	}
 }

src/bees.cc

@@ -667,6 +667,7 @@ bees_main(int argc, char *argv[])
 	unsigned thread_min = 0;
 	double load_target = 0;
 	bool workaround_btrfs_send = false;
+	BeesRoots::ScanMode root_scan_mode = BeesRoots::SCAN_MODE_ZERO;
 
 	// Configure getopt_long
 	static const struct option long_options[] = {
@@ -735,7 +736,7 @@ bees_main(int argc, char *argv[])
 				load_target = stod(optarg);
 				break;
 			case 'm':
-				bc->roots()->set_scan_mode(static_cast<BeesRoots::ScanMode>(stoul(optarg)));
+				root_scan_mode = static_cast<BeesRoots::ScanMode>(stoul(optarg));
 				break;
 			case 'p':
 				crucible::set_relative_path("");
@@ -806,11 +807,16 @@ bees_main(int argc, char *argv[])
 	BEESLOGNOTICE("setting worker thread pool maximum size to " << thread_count);
 	TaskMaster::set_thread_count(thread_count);
 
+	// Set root path
+	string root_path = argv[optind++];
+	BEESLOGNOTICE("setting root path to '" << root_path << "'");
+	bc->set_root_path(root_path);
+
 	// Workaround for btrfs send
 	bc->roots()->set_workaround_btrfs_send(workaround_btrfs_send);
 
-	// Create a context and start crawlers
-	bc->set_root_path(argv[optind++]);
+	// Set root scan mode
+	bc->roots()->set_scan_mode(root_scan_mode);
 
 	BeesThread status_thread("status", [&]() {
 		bc->dump_status();

src/bees.h

@@ -117,6 +117,9 @@ const size_t BEES_TRANSID_FACTOR = 10;
 // The actual limit in LOGICAL_INO seems to be 2730, but let's leave a little headroom
 const size_t BEES_MAX_EXTENT_REF_COUNT = 2560;
 
+// Wait this long for a balance to stop
+const double BEES_BALANCE_POLL_INTERVAL = 60.0;
+
 // Flags
 const int FLAGS_OPEN_COMMON   = O_NOFOLLOW | O_NONBLOCK | O_CLOEXEC | O_NOATIME | O_LARGEFILE | O_NOCTTY;
 const int FLAGS_OPEN_DIR      = FLAGS_OPEN_COMMON | O_RDONLY | O_DIRECTORY;
@@ -716,6 +719,7 @@ class BeesContext : public enable_shared_from_this<BeesContext> {
 	void set_root_fd(Fd fd);
 
 	BeesResolveAddrResult resolve_addr_uncached(BeesAddress addr);
+	void wait_for_balance();
 
 	BeesFileRange scan_one_extent(const BeesFileRange &bfr, const Extent &e);
 	void rewrite_file_range(const BeesFileRange &bfr);