dmsetup: Support remove --deferred.

This patch adds a new flag --deferred to dmsetup remove. If this flag is
specified and the device is open, it is scheduled to be deleted on
close.

struct dm_info is extended.

The existing dm_task_get_info() is converted into a wrapper around the
new version dm_task_get_info_with_deferred_remove() so existing binaries
can still use the old smaller structure.

Recompiled code will pick up the new larger structure.

From: Mikulas Patocka <mpatocka@redhat.com>

7 files changed, 87 insertions, 26 deletions

diff --git a/WHATS_NEW_DM b/WHATS_NEW_DM
index 41aa89de5..4a2c8b405 100644
--- a/WHATS_NEW_DM
+++ b/WHATS_NEW_DM
@@ -1,5 +1,6 @@
 Version 1.02.89 -
 =================================
+  Support --deferred with dmsetup remove to defer removal of open devices.
   Update dm-ioctl.h to include DM_DEFERRED_REMOVE flag.
   Add support for selection to match string list subset, recognize { } operator.
   Fix string list selection with '[value]' to not match list that's superset.
diff --git a/libdm/.exported_symbols b/libdm/.exported_symbols
index 3a6986085..5bacc26e2 100644
--- a/libdm/.exported_symbols
+++ b/libdm/.exported_symbols
@@ -1,2 +1,3 @@
 dm_log
 dm_log_with_errno
+dm_task_get_info
diff --git a/libdm/ioctl/libdm-iface.c b/libdm/ioctl/libdm-iface.c
index 345651b30..4f55ddf3a 100644
--- a/libdm/ioctl/libdm-iface.c
+++ b/libdm/ioctl/libdm-iface.c
@@ -17,6 +17,7 @@
 #include "libdm-targets.h"
 #include "libdm-common.h"
 
+#include <stddef.h>
 #include <fcntl.h>
 #include <dirent.h>
 #include <sys/ioctl.h>
@@ -654,6 +655,7 @@ int dm_task_get_info(struct dm_task *dmt, struct dm_info *info)
 	info->live_table = dmt->dmi.v4->flags & DM_ACTIVE_PRESENT_FLAG ? 1 : 0;
 	info->inactive_table = dmt->dmi.v4->flags & DM_INACTIVE_PRESENT_FLAG ?
 	    1 : 0;
+	info->deferred_remove = dmt->dmi.v4->flags & DM_DEFERRED_REMOVE;
 	info->target_count = dmt->dmi.v4->target_count;
 	info->open_count = dmt->dmi.v4->open_count;
 	info->event_nr = dmt->dmi.v4->event_nr;
@@ -862,6 +864,13 @@ int dm_task_retry_remove(struct dm_task *dmt)
 	return 1;
 }
 
+int dm_task_deferred_remove(struct dm_task *dmt)
+{
+	dmt->deferred_remove = 1;
+
+	return 1;
+}
+
 int dm_task_query_inactive_table(struct dm_task *dmt)
 {
 	dmt->query_inactive_table = 1;
@@ -1137,6 +1146,9 @@ static struct dm_ioctl *_flatten(struct dm_task *dmt, unsigned repeat_count)
 		dmi->flags |= DM_READONLY_FLAG;
 	if (dmt->skip_lockfs)
 		dmi->flags |= DM_SKIP_LOCKFS_FLAG;
+	if (dmt->deferred_remove && (dmt->type == DM_DEVICE_REMOVE || dmt->type == DM_DEVICE_REMOVE_ALL))
+		dmi->flags |= DM_DEFERRED_REMOVE;
+
 	if (dmt->secure_data) {
 		if (_dm_version_minor < 20)
 			log_verbose("Secure data flag unsupported by kernel. "
@@ -1732,7 +1744,7 @@ static struct dm_ioctl *_do_dm_ioctl(struct dm_task *dmt, unsigned command,
 	}
 
 	log_debug_activation("dm %s %s%s %s%s%s %s%.0d%s%.0d%s"
-			     "%s%c%c%s%s%s%s%s%s %.0" PRIu64 " %s [%u] (*%u)",
+			     "%s%c%c%s%s%s%s%s%s%s %.0" PRIu64 " %s [%u] (*%u)",
 			     _cmd_data_v4[dmt->type].name,
 			     dmt->new_uuid ? "UUID " : "",
 			     dmi->name, dmi->uuid, dmt->newname ? " " : "",
@@ -1748,6 +1760,7 @@ static struct dm_ioctl *_do_dm_ioctl(struct dm_task *dmt, unsigned command,
 			     dmt->read_only ? "R" : "",
 			     dmt->skip_lockfs ? "S " : "",
 			     dmt->retry_remove ? "T " : "",
+			     dmt->deferred_remove ? "D " : "",
 			     dmt->secure_data ? "W " : "",
 			     dmt->query_inactive_table ? "I " : "",
 			     dmt->enable_checks ? "C" : "",
@@ -2006,3 +2019,32 @@ void dm_lib_exit(void)
 	_version_ok = 1;
 	_version_checked = 0;
 }
+
+/*
+ * This following code is here to retain ABI compatibility after adding
+ * the field deferred_remove to struct dm_info in version 1.02.89.
+ *
+ * Binaries linked against version 1.02.88 of libdevmapper or earlier
+ * will use this function that returns dm_info without the
+ * deferred_remove field.
+ *
+ * Binaries compiled against version 1.02.89 onwards will use
+ * the new function dm_task_get_info_with_deferred_remove due to the
+ * #define.
+ *
+ * N.B. Keep this function at the end of the file to make sure that
+ * no code in this file accidentally calls it.
+ */
+#undef dm_task_get_info
+int dm_task_get_info(struct dm_task *dmt, struct dm_info *info);
+int dm_task_get_info(struct dm_task *dmt, struct dm_info *info)
+{
+	struct dm_info new_info;
+
+	if (!dm_task_get_info_with_deferred_remove(dmt, &new_info))
+		return 0;
+
+	memcpy(info, &new_info, offsetof(struct dm_info, deferred_remove));
+
+	return 1;
+}
diff --git a/libdm/ioctl/libdm-targets.h b/libdm/ioctl/libdm-targets.h
index 54223c752..100681fec 100644
--- a/libdm/ioctl/libdm-targets.h
+++ b/libdm/ioctl/libdm-targets.h
@@ -65,6 +65,7 @@ struct dm_task {
 	int new_uuid;
 	int secure_data;
 	int retry_remove;
+	int deferred_remove;
 	int enable_checks;
 	int expected_errno;
 
diff --git a/libdm/libdevmapper.h b/libdm/libdevmapper.h
index e37ee39b9..bf8a65831 100644
--- a/libdm/libdevmapper.h
+++ b/libdm/libdevmapper.h
@@ -149,6 +149,8 @@ struct dm_info {
 	int read_only;		/* 0:read-write; 1:read-only */
 
 	int32_t target_count;
+
+	int deferred_remove;
 };
 
 struct dm_deps {
@@ -172,6 +174,8 @@ struct dm_versions {
 
 int dm_get_library_version(char *version, size_t size);
 int dm_task_get_driver_version(struct dm_task *dmt, char *version, size_t size);
+
+#define dm_task_get_info dm_task_get_info_with_deferred_remove
 int dm_task_get_info(struct dm_task *dmt, struct dm_info *dmi);
 
 /*
@@ -222,6 +226,7 @@ int dm_task_query_inactive_table(struct dm_task *dmt);
 int dm_task_suppress_identical_reload(struct dm_task *dmt);
 int dm_task_secure_data(struct dm_task *dmt);
 int dm_task_retry_remove(struct dm_task *dmt);
+int dm_task_deferred_remove(struct dm_task *dmt);
 
 /*
  * Enable checks for common mistakes such as issuing ioctls in an unsafe order.
diff --git a/man/dmsetup.8.in b/man/dmsetup.8.in
index b18b0b008..477e4252b 100644
--- a/man/dmsetup.8.in
+++ b/man/dmsetup.8.in
@@ -81,10 +81,12 @@ dmsetup \(em low level logical volume management
 .B dmsetup remove
 .RB [ \-f | \-\-force ]
 .RB [ \-\-retry ]
+.RB [ \-\-deferred ]
 .I device_name
 .br
 .B dmsetup remove_all
 .RB [ \-f | \-\-force ]
+.RB [ \-\-deferred ]
 .br
 .B dmsetup rename
 .I device_name new_name
@@ -436,34 +438,35 @@ reactivating it with proper mangling mode used (see also \fB\-\-manglename\fP).
 .B remove
 .RB [ \-f | \-\-force ]
 .RB [ \-\-retry ]
+.RB [ \-\-deferred ]
 .I device_name
 .br
-Removes a device.  It will no longer be visible to dmsetup.
-Open devices cannot be removed except with older kernels
-that contain a version of device-mapper prior to 4.8.0.
-In this case the device will be deleted when its open_count
-drops to zero.  From version 4.8.0 onwards, if a device can't
-be removed because an uninterruptible process is waiting for
-I/O to return from it, adding \fB\-\-force\fP will replace the table
-with one that fails all I/O, which might allow the
-process to be killed. If an attempt to remove a device fails,
-perhaps because a process run from a quick udev rule
-temporarily opened the device, the \fB\-\-retry\fP option will cause
-the operation to be retried for a few seconds before failing.
-Do NOT combine \fB\-\-force\fP and \fB\-\-udevcookie\fP,
-as udev may start to process udev rules in the middle of error target
-replacement and result in nondeterministic result.
+Removes a device.  It will no longer be visible to dmsetup.  Open devices
+cannot be removed, but adding \fB\-\-force\fP will replace the table with one
+that fails all I/O.  \fB\-\-deferred\fP will enable deferred removal of open
+devices - the device will be removed when the last user closes it. The deferred
+removal feature is supported since version 4.27.0 of the device-mapper
+driver available in upstream kernel version 3.13.  (Use \fBdmsetup version\fP
+to check this.)  If an attempt to remove a device fails, perhaps because a process run
+from a quick udev rule temporarily opened the device, the \fB\-\-retry\fP
+option will cause the operation to be retried for a few seconds before failing.
+Do NOT combine
+\fB\-\-force\fP and \fB\-\-udevcookie\fP, as udev may start to process udev
+rules in the middle of error target replacement and result in nondeterministic
+result.
 .br
 .HP
 .B remove_all
 .RB [ \-f | \-\-force ]
-.br
-Attempts to remove all device definitions i.e. reset the driver.
-Use with care!  From version 4.8.0 onwards, if devices can't
-be removed because uninterruptible processes are waiting for
-I/O to return from them, adding \fB\-\-force\fP will replace the table
-with one that fails all I/O, which might allow the
-process to be killed.  This also runs \fBmknodes\fP afterwards.
+.RB [ \-\-deferred ]
+.br
+Attempts to remove all device definitions i.e. reset the driver.  This also runs
+\fBmknodes\fP afterwards.  Use with care!  Open devices cannot be removed, but
+adding \fB\-\-force\fP will replace the table with one that fails all I/O.
+\fB\-\-deferred\fP will enable deferred removal of open devices - the device
+will be removed when the last user closes it.  The deferred removal feature is
+supported since version 4.27.0 of the device-mapper driver available in
+upstream kernel version 3.13. 
 .br
 .HP
 .B rename
diff --git a/tools/dmsetup.c b/tools/dmsetup.c
index c679c52e5..8f64bfce6 100644
--- a/tools/dmsetup.c
+++ b/tools/dmsetup.c
@@ -120,6 +120,7 @@ enum {
 	ADD_NODE_ON_RESUME_ARG,
 	CHECKS_ARG,
 	COLS_ARG,
+	DEFERRED_ARG,
 	SELECT_ARG,
 	EXEC_ARG,
 	FORCE_ARG,
@@ -468,9 +469,10 @@ static void _display_info_long(struct dm_task *dmt, struct dm_info *info)
 
 	printf("Name:              %s\n", dm_task_get_name(dmt));
 
-	printf("State:             %s%s\n",
+	printf("State:             %s%s%s\n",
 	       info->suspended ? "SUSPENDED" : "ACTIVE",
-	       info->read_only ? " (READ-ONLY)" : "");
+	       info->read_only ? " (READ-ONLY)" : "",
+	       info->deferred_remove ? " (DEFERRED REMOVE)" : "");
 
 	/* FIXME Old value is being printed when it's being changed. */
 	if (dm_task_get_read_ahead(dmt, &read_ahead))
@@ -1321,6 +1323,9 @@ static int _simple(int task, const char *name, uint32_t event_nr, int display)
 	if (_switches[RETRY_ARG] && task == DM_DEVICE_REMOVE)
 		dm_task_retry_remove(dmt);
 
+	if (_switches[DEFERRED_ARG] && (task == DM_DEVICE_REMOVE || task == DM_DEVICE_REMOVE_ALL))
+		dm_task_deferred_remove(dmt);
+
 	r = dm_task_run(dmt);
 
       out:
@@ -3071,7 +3076,7 @@ static struct command _commands[] = {
 	  "\t                  [-u|uuid <uuid>] [{--addnodeonresume|--addnodeoncreate}]\n"
 	  "\t                  [--notable | --table <table> | <table_file>]",
 	 1, 2,0,  _create},
-	{"remove", "[-f|--force] <device>", 0, -1, 1, _remove},
+	{"remove", "[-f|--force] [--deferred] <device>", 0, -1, 1, _remove},
 	{"remove_all", "[-f|--force]", 0, 0, 0,  _remove_all},
 	{"suspend", "[--noflush] <device>", 0, -1, 1, _suspend},
 	{"resume", "<device> [{--addnodeonresume|--addnodeoncreate}]", 0, -1, 1, _resume},
@@ -3521,6 +3526,7 @@ static int _process_switches(int *argc, char ***argv, const char *dev_dir)
 		{"readonly", 0, &ind, READ_ONLY},
 		{"checks", 0, &ind, CHECKS_ARG},
 		{"columns", 0, &ind, COLS_ARG},
+		{"deferred", 0, &ind, DEFERRED_ARG},
 		{"select", 1, &ind, SELECT_ARG},
 		{"exec", 1, &ind, EXEC_ARG},
 		{"force", 0, &ind, FORCE_ARG},
@@ -3694,6 +3700,8 @@ static int _process_switches(int *argc, char ***argv, const char *dev_dir)
 			/* FIXME Accept modes as per chmod */
 			_int_args[MODE_ARG] = (int) strtol(optarg, NULL, 8);
 		}
+		if (ind == DEFERRED_ARG)
+			_switches[DEFERRED_ARG]++;
 		if (ind == EXEC_ARG) {
 			_switches[EXEC_ARG]++;
 			_command = optarg;