Merge branch 'master' of git+ssh://sourceware.org/git/lvm2

doc/lvm-disk-reading.txt

@@ -187,3 +187,45 @@ For each VG name:
   command-specific work
 
 
+
+Filter i/o
+----------
+
+Some filters must be applied before reading a device, and other filters
+must be applied after reading a device.  In all cases, the filters must be
+applied before lvm processes the device, i.e. before it looks for an lvm
+label.
+
+1. Some filters need to be applied prior to reading any devices
+   because the purpose of the filter is to avoid submitting any
+   io on the excluded devices.  The regex filter is the primary
+   example.  Other filters benefit from being applied prior to
+   reading devices because they can tell which devices to
+   exclude without doing io to the device.  An example of this
+   is the mpath filter.
+
+2. Some filters need to be applied after reading a device because
+   they are based on data/signatures seen on the device.
+   The partitioned filter is an example of this; lvm needs to
+   read a device to see if it has a partition table before it can
+   know whether to exclude the device from further processing.
+
+We apply filters from 1 before reading devices, and we apply filters from
+2 after populating bcache, but before processing the device (i.e. before
+checking for an lvm label, which is the first step in processing.)
+
+The current implementation of this makes filters return -EAGAIN if they
+want to read the device, but bcache data is not yet available.  This will
+happen when filtering runs prior to populating bcache.  In this case the
+device is flagged.  After bcache is populated, the filters are reapplied
+to the flagged devices.  The filters which need to look at device content
+are now able to get it from bcache.  Devices that do not pass filters at
+this point are excluded just like devices which were excluded earlier.
+
+(Some filters from 2 can be skipped by consulting udev for the information
+instead of reading the device.  This is not entirely reliable, so it is
+disabled by default with the config setting external_device_info_source.
+It may be worthwhile to change the filters to use the udev info as a hint,
+or only use udev info for filtering in reporting commands where
+inaccuracies are not a big problem.)
+

lib/label/label.c

@@ -884,7 +884,7 @@ int label_scan_open(struct device *dev)
 	return 1;
 }
 
-bool dev_read_bytes(struct device *dev, off_t start, size_t len, void *data)
+bool dev_read_bytes(struct device *dev, uint64_t start, size_t len, void *data)
 {
 	int ret;
 
@@ -918,7 +918,7 @@ bool dev_read_bytes(struct device *dev, off_t start, size_t len, void *data)
 
 }
 
-bool dev_write_bytes(struct device *dev, off_t start, size_t len, void *data)
+bool dev_write_bytes(struct device *dev, uint64_t start, size_t len, void *data)
 {
 	int ret;
 
@@ -961,7 +961,7 @@ bool dev_write_bytes(struct device *dev, off_t start, size_t len, void *data)
 	return true;
 }
 
-bool dev_write_zeros(struct device *dev, off_t start, size_t len)
+bool dev_write_zeros(struct device *dev, uint64_t start, size_t len)
 {
 	int ret;
 
@@ -969,15 +969,8 @@ bool dev_write_zeros(struct device *dev, off_t start, size_t len)
 		return true;
 
 	if (!scan_bcache) {
-		if (!dev_open(dev))
+		log_error("dev_write_zeros %s bcache not set up", dev_name(dev));
-			return false;
+		return false;
-
-		ret = dev_set(dev, start, len, 0, 0);
-
-		if (!dev_close(dev))
-			stack;
-
-		return ret ? true : false;
 	}
 
 	if (dev->bcache_fd <= 0) {
@@ -1004,3 +997,39 @@ bool dev_write_zeros(struct device *dev, off_t start, size_t len)
 	return true;
 }
 
+bool dev_set_bytes(struct device *dev, uint64_t start, size_t len, uint8_t val)
+{
+	int ret;
+
+	if (test_mode())
+		return true;
+
+	if (!scan_bcache) {
+		log_error("dev_set_bytes %s bcache not set up", dev_name(dev));
+		return false;
+	}
+
+	if (dev->bcache_fd <= 0) {
+		/* This is not often needed, perhaps only with lvmetad. */
+		if (!label_scan_open(dev)) {
+			log_error("dev_set_bytes %s cannot open dev", dev_name(dev));
+			return false;
+		}
+	}
+
+	if (!bcache_set_bytes(scan_bcache, dev->bcache_fd, start, len, val)) {
+		log_error("dev_set_bytes %s at %u bcache write failed invalidate fd %d",
+			  dev_name(dev), (uint32_t)start, dev->bcache_fd);
+		label_scan_invalidate(dev);
+		return false;
+	}
+
+	if (!bcache_flush(scan_bcache)) {
+		log_error("dev_set_bytes %s at %u bcache flush failed invalidate fd %d",
+			  dev_name(dev), (uint32_t)start, dev->bcache_fd);
+		label_scan_invalidate(dev);
+		return false;
+	}
+	return true;
+}
+

lib/label/label.h

@@ -119,8 +119,9 @@ int label_scan_open(struct device *dev);
  * Wrappers around bcache equivalents.
  * (these make it easier to disable bcache and revert to direct rw if needed)
  */
-bool dev_read_bytes(struct device *dev, off_t start, size_t len, void *data);
+bool dev_read_bytes(struct device *dev, uint64_t start, size_t len, void *data);
-bool dev_write_bytes(struct device *dev, off_t start, size_t len, void *data);
+bool dev_write_bytes(struct device *dev, uint64_t start, size_t len, void *data);
-bool dev_write_zeros(struct device *dev, off_t start, size_t len);
+bool dev_write_zeros(struct device *dev, uint64_t start, size_t len);
+bool dev_set_bytes(struct device *dev, uint64_t start, size_t len, uint8_t val);
 
 #endif

lib/locking/lvmlockd.c

@@ -403,22 +403,17 @@ static int _extend_sanlock_lv(struct cmd_context *cmd, struct volume_group *vg,
 		return 0;
 	}
 
-	if (!dev_open_quiet(dev)) {
+	if (!label_scan_open(dev)) {
 		log_error("Extend sanlock LV %s cannot open device.", display_lvname(lv));
 		return 0;
 	}
 
-	if (!dev_set(dev, old_size_bytes, new_size_bytes - old_size_bytes, DEV_IO_LV, 0)) {
+	if (!dev_write_zeros(dev, old_size_bytes, new_size_bytes - old_size_bytes)) {
 		log_error("Extend sanlock LV %s cannot zero device.", display_lvname(lv));
-		dev_close_immediate(dev);
 		return 0;
 	}
 
-	dev_flush(dev);
+	label_scan_invalidate(dev);
-
-	if (!dev_close_immediate(dev))
-		stack;
-
 	return 1;
 }
 

lib/metadata/lv_manip.c

@@ -7155,8 +7155,10 @@ int wipe_lv(struct logical_volume *lv, struct wipe_params wp)
 		return 0;
 	}
 
-	if (!dev_open_quiet(dev))
+	if (!label_scan_open(dev)) {
-		return_0;
+		log_error("Failed to open %s/%s for wiping and zeroing.", lv->vg->name, lv->name);
+		goto out;
+	}
 
 	if (wp.do_wipe_signatures) {
 		log_verbose("Wiping known signatures on logical volume \"%s/%s\"",
@@ -7177,15 +7179,17 @@ int wipe_lv(struct logical_volume *lv, struct wipe_params wp)
 			    display_size(lv->vg->cmd, zero_sectors),
 			    lv->vg->name, lv->name, wp.zero_value);
 
-		if (!dev_set(dev, UINT64_C(0), (size_t) zero_sectors << SECTOR_SHIFT, DEV_IO_LV, wp.zero_value))
+		if (!wp.zero_value) {
-			stack;
+			if (!dev_write_zeros(dev, UINT64_C(0), (size_t) zero_sectors << SECTOR_SHIFT))
+				stack;
+		} else {
+			if (!dev_set_bytes(dev, UINT64_C(0), (size_t) zero_sectors << SECTOR_SHIFT, (uint8_t)wp.zero_value))
+				stack;
+		}
 	}
 
-	dev_flush(dev);
+	label_scan_invalidate(dev);
-
+out:
-	if (!dev_close_immediate(dev))
-		stack;
-
 	lv->status &= ~LV_NOSCAN;
 
 	return 1;

lib/metadata/metadata-liblvm.c

@@ -66,19 +66,11 @@ static int _pvcreate_write(struct cmd_context *cmd, struct pv_to_write *pvw)
 
 		if (pvw->pp->zero) {
 			log_verbose("Zeroing start of device %s", pv_name);
-			if (!dev_open_quiet(dev)) {
-				log_error("%s not opened: device not zeroed", pv_name);
-				return 0;
-			}
 
-			if (!dev_set(dev, UINT64_C(0), (size_t) 2048, DEV_IO_LABEL, 0)) {
+			if (!dev_write_zeros(dev, UINT64_C(0), (size_t) 2048)) {
 				log_error("%s not wiped: aborting", pv_name);
-				if (!dev_close(dev))
-					stack;
 				return 0;
 			}
-			if (!dev_close(dev))
-				stack;
 		}
 	}