Add zpool_events_seek() functionality

The ZFS_IOC_EVENTS_SEEK ioctl was added to allow user space callers
to seek around the zevent file descriptor by EID.  When a specific
EID is passed and it exists the cursor will be positioned there.
If the EID is no longer cached by the kernel ENOENT is returned.
The caller may also pass ZEVENT_SEEK_START or ZEVENT_SEEK_END to seek
to those respective locations.

Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Chris Dunlap <cdunlap@llnl.gov>
Issue #2

include/libzfs.h

@@ -410,6 +410,7 @@ extern int zpool_history_unpack(char *, uint64_t, uint64_t *,
     nvlist_t ***, uint_t *);
 extern int zpool_events_next(libzfs_handle_t *, nvlist_t **, int *, int, int);
 extern int zpool_events_clear(libzfs_handle_t *, int *);
+extern int zpool_events_seek(libzfs_handle_t *, uint64_t, int);
 extern void zpool_obj_to_path(zpool_handle_t *, uint64_t, uint64_t, char *,
     size_t len);
 extern int zfs_ioctl(libzfs_handle_t *, int, struct zfs_cmd *);

include/sys/fm/util.h

@@ -71,7 +71,7 @@ typedef struct erpt_dump {
 
 #ifdef _KERNEL
 
-#define	ZEVENT_SHUTDOWN	0x1
+#define	ZEVENT_SHUTDOWN		0x1
 
 typedef void zevent_cb_t(nvlist_t *, nvlist_t *);
 
@@ -99,6 +99,7 @@ extern int zfs_zevent_fd_hold(int, minor_t *, zfs_zevent_t **);
 extern void zfs_zevent_fd_rele(int);
 extern int zfs_zevent_next(zfs_zevent_t *, nvlist_t **, uint64_t *, uint64_t *);
 extern int zfs_zevent_wait(zfs_zevent_t *);
+extern int zfs_zevent_seek(zfs_zevent_t *, uint64_t);
 extern void zfs_zevent_init(zfs_zevent_t **);
 extern void zfs_zevent_destroy(zfs_zevent_t *);
 

include/sys/fs/zfs.h

@@ -864,6 +864,7 @@ typedef enum zfs_ioc {
 	ZFS_IOC_LINUX = ('Z' << 8) + 0x80,
 	ZFS_IOC_EVENTS_NEXT,
 	ZFS_IOC_EVENTS_CLEAR,
+	ZFS_IOC_EVENTS_SEEK,
 
 	/*
 	 * FreeBSD - 1/64 numbers reserved.

include/sys/zfs_ioctl.h

@@ -262,6 +262,9 @@ typedef struct zinject_record {
 #define	ZEVENT_NONBLOCK		0x1
 #define	ZEVENT_SIZE		1024
 
+#define	ZEVENT_SEEK_START	0
+#define	ZEVENT_SEEK_END		UINT64_MAX
+
 typedef enum zinject_type {
 	ZINJECT_UNINITIALIZED,
 	ZINJECT_DATA_FAULT,

lib/libzfs/libzfs_pool.c

@@ -3870,6 +3870,42 @@ zpool_events_clear(libzfs_handle_t *hdl, int *count)
 	return (0);
 }
 
+/*
+ * Seek to a specific EID, ZEVENT_SEEK_START, or ZEVENT_SEEK_END for
+ * the passed zevent_fd file handle.  On success zero is returned,
+ * otherwise -1 is returned and hdl->libzfs_error is set to the errno.
+ */
+int
+zpool_events_seek(libzfs_handle_t *hdl, uint64_t eid, int zevent_fd)
+{
+	zfs_cmd_t zc = {"\0"};
+	int error = 0;
+
+	zc.zc_guid = eid;
+	zc.zc_cleanup_fd = zevent_fd;
+
+	if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_SEEK, &zc) != 0) {
+		switch (errno) {
+		case ENOENT:
+			error = zfs_error_fmt(hdl, EZFS_NOENT,
+			    dgettext(TEXT_DOMAIN, "cannot get event"));
+			break;
+
+		case ENOMEM:
+			error = zfs_error_fmt(hdl, EZFS_NOMEM,
+			    dgettext(TEXT_DOMAIN, "cannot get event"));
+			break;
+
+		default:
+			error = zpool_standard_error_fmt(hdl, errno,
+			    dgettext(TEXT_DOMAIN, "cannot get event"));
+			break;
+		}
+	}
+
+	return (error);
+}
+
 void
 zpool_obj_to_path(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj,
     char *pathname, size_t len)

module/zfs/fm.c

@@ -664,6 +664,67 @@ out:
 	return (error);
 }
 
+/*
+ * The caller may seek to a specific EID by passing that EID.  If the EID
+ * is still available in the posted list of events the cursor is positioned
+ * there.  Otherwise ENOENT is returned and the cursor is not moved.
+ *
+ * There are two reserved EIDs which may be passed and will never fail.
+ * ZEVENT_SEEK_START positions the cursor at the start of the list, and
+ * ZEVENT_SEEK_END positions the cursor at the end of the list.
+ */
+int
+zfs_zevent_seek(zfs_zevent_t *ze, uint64_t eid)
+{
+	zevent_t *ev;
+	int error = 0;
+
+	mutex_enter(&zevent_lock);
+
+	if (eid == ZEVENT_SEEK_START) {
+		if (ze->ze_zevent)
+			list_remove(&ze->ze_zevent->ev_ze_list, ze);
+
+		ze->ze_zevent = NULL;
+		goto out;
+	}
+
+	if (eid == ZEVENT_SEEK_END) {
+		if (ze->ze_zevent)
+			list_remove(&ze->ze_zevent->ev_ze_list, ze);
+
+		ev = list_head(&zevent_list);
+		if (ev) {
+			ze->ze_zevent = ev;
+			list_insert_head(&ev->ev_ze_list, ze);
+		} else {
+			ze->ze_zevent = NULL;
+		}
+
+		goto out;
+	}
+
+	for (ev = list_tail(&zevent_list); ev != NULL;
+	    ev = list_prev(&zevent_list, ev)) {
+		if (ev->ev_eid == eid) {
+			if (ze->ze_zevent)
+				list_remove(&ze->ze_zevent->ev_ze_list, ze);
+
+			ze->ze_zevent = ev;
+			list_insert_head(&ev->ev_ze_list, ze);
+			break;
+		}
+	}
+
+	if (ev == NULL)
+		error = ENOENT;
+
+out:
+	mutex_exit(&zevent_lock);
+
+	return (error);
+}
+
 void
 zfs_zevent_init(zfs_zevent_t **zep)
 {

module/zfs/zfs_ioctl.c

@@ -4919,6 +4919,28 @@ zfs_ioc_events_clear(zfs_cmd_t *zc)
 	return (0);
 }
 
+/*
+ * inputs:
+ * zc_guid		eid | ZEVENT_SEEK_START | ZEVENT_SEEK_END
+ * zc_cleanup		zevent file descriptor
+ */
+static int
+zfs_ioc_events_seek(zfs_cmd_t *zc)
+{
+	zfs_zevent_t *ze;
+	minor_t minor;
+	int error;
+
+	error = zfs_zevent_fd_hold(zc->zc_cleanup_fd, &minor, &ze);
+	if (error != 0)
+		return (error);
+
+	error = zfs_zevent_seek(ze, zc->zc_guid);
+	zfs_zevent_fd_rele(zc->zc_cleanup_fd);
+
+	return (error);
+}
+
 /*
  * inputs:
  * zc_name		name of new filesystem or snapshot
@@ -5393,6 +5415,8 @@ zfs_ioctl_init(void)
 	    zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_NONE);
 	zfs_ioctl_register_legacy(ZFS_IOC_EVENTS_CLEAR, zfs_ioc_events_clear,
 	    zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_NONE);
+	zfs_ioctl_register_legacy(ZFS_IOC_EVENTS_SEEK, zfs_ioc_events_seek,
+	    zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_NONE);
 }
 
 int