Device Mapper Mirror
Subsystems (2)
PATCH 1/1 Device Mapper Mirror
This patch is not yet finish and should not be pushed upstream in its current form. I provide it here so people can look at it and give me feedback.
brassow
drivers/md/dm-clog-tfr.c new file +216 -0
@@ -0,0 +1,216 @@ +/* +* Copyright (C) 2006 Red Hat, Inc. +* +* This file is released under the LGPL. +*/ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <net/sock.h> +#include <linux/workqueue.h> +#include <linux/dm-clog-tfr.h> +#include "dm.h" + +/* +* Pre-allocated nominal request area for speed +*/ +#define DM_CLOG_NOMINAL_REQUEST_SIZE 512 +static char nominal_request[DM_CLOG_NOMINAL_REQUEST_SIZE ]; + +static struct sock *nl_sk =NULL ; +static DEFINE_MUTEX (_lock); +static int users =0 ; +//static int server_pid = 0; + +static void nl_data_ready (struct sock *sk,int len) +{ +wake_up_interruptible (sk->sk_sleep); +} + +static int find_server_process (void ) +{ +int r = -ENOSYS ; +struct sk_buff *skb =NULL ; +struct nlmsghdr *nlh =NULL ; + +BUG_ON (!nl_sk); + + skb =alloc_skb (NLMSG_SPACE (0 ),GFP_KERNEL ); +if (!skb) +return -ENOMEM ; + + nlh = (struct nlmsghdr *)skb->data; + nlh->nlmsg_len =NLMSG_SPACE (0 ); + nlh->nlmsg_pid =0 ; + nlh->nlmsg_flags =0 ; + +/* Not in newer kernels +NETLINK_CB(skb).groups = 1; +NETLINK_CB(skb).dst_groups = 1; +*/ +NETLINK_CB (skb).pid =0 ; +NETLINK_CB (skb).dst_pid =0 ; +NETLINK_CB (skb).dst_group =1 ; + +netlink_broadcast (nl_sk, skb,0 ,1 ,GFP_KERNEL ); + +/* Get reply to find out server pid */ +return r; +} + +static int dm_clog_sendto_server (struct clog_tfr *tfr) +{ +BUG_ON (!nl_sk); +return -ENOSYS ; +} + +static int dm_clog_recvfrom_server (char *rdata,int *rdata_size) +{ +struct sk_buff *skb =NULL ; +struct clog_tfr *tfr =NULL ; +int r; + +BUG_ON (!nl_sk); + skb =skb_recv_datagram (nl_sk,0 ,0 , &r); +if (r) +return r; + + tfr =NLMSG_DATA ((struct nlmsghdr *)skb->data); + +if (tfr->error) { + r = tfr->error; + }else if (tfr->data_size > *rdata_size) { + r = -ENOSPC ; + }else { + *rdata_size = tfr->data_size; +memcpy (rdata, tfr->data, tfr->data_size); + } + +return r; +} + +/* +* dm_clog_consult_server +* @uuid: log's uuid (must be DM_UUID_LEN in size) +* @request_type: +* @data: data to tx to the server +* @data_size: size of data in bytes +* @rdata: place to put return data from server +* @rdata_size: value-result (amount of space given/amount of space used) +* +* Only one process at a time can communicate with the server. +* rdata_size is undefined on failure. +* +* Returns: 0 on success, -EXXX on failure +*/ +int dm_clog_consult_server (const char *uuid,int request_type, +char *data,int data_size, +char *rdata,int *rdata_size) +{ +int r =0 ; +struct clog_tfr *tfr = (struct clog_tfr *)nominal_request; + +if (data_size > (DM_CLOG_NOMINAL_REQUEST_SIZE -sizeof (*tfr))) +/* FIXME: is kmalloc sufficient if we need this much space? */ + tfr =kmalloc (data_size +sizeof (*tfr),GFP_KERNEL ); + +if (!tfr) +return -ENOMEM ; + +mutex_lock (&_lock); + +memcpy (tfr->uuid, uuid,DM_UUID_LEN ); + tfr->request_type = request_type; + tfr->data_size = data_size; + + r =dm_clog_sendto_server (tfr); +if (r <0 ) { +// DMERR("Unable to send cluster log request to server: %d", r); + goto fail; + } + +if (rdata) { + r =dm_clog_recvfrom_server (rdata, rdata_size); + +if (r <0 ) + goto fail; + }else { +/* +* FIXME: If we are using netlink, we may want an +* ack from the server to know that it got the +* request. (Ack is implicit if we are receiving +* data.) +*/ + r = -ENOSYS ; + } + +fail: +if (tfr && (tfr != (struct clog_tfr *)nominal_request)) +kfree (tfr); + +mutex_unlock (&_lock); +return r; +} + +/* +* dm_clog_tfr_get +* +* Set-up or increment reference count on the +* netlink used to communicate with the user-space +* server. +* +* Returns: 0 on success, -EXXX on failure +*/ +int dm_clog_tfr_get (void ) +{ +int r =0 ; + +mutex_lock (&_lock); +if (!users) { + +BUG_ON (nl_sk);/* no socket should be present */ + + nl_sk =netlink_kernel_create (NETLINK_DMCLOG ,1 , + nl_data_ready,THIS_MODULE ); + +if (!nl_sk) { + r = -ENOMEM ; + goto fail; + } + + r =find_server_process (); +if (r) + goto fail; + } + users++; +fail: +if (r && nl_sk) { +sock_release (nl_sk->sk_socket); + nl_sk =NULL ; + } +mutex_unlock (&_lock); + +return r; +} + +/* +* dm_clog_tfr_put +* +* Decrement reference count on the netlink used to +* communicate with the user-space server - releasing +* the socket if no references remain. +* +* Returns: 0 on success, -EXXX on failure +*/ +int dm_clog_tfr_put (void ) +{ +mutex_lock (&_lock); +if (!--users) { +sock_release (nl_sk->sk_socket); + + nl_sk =NULL ; + } +mutex_unlock (&_lock); + +return 0 ; +}
drivers/md/dm-clog.c new file +596 -0
596 lines not on this page · raw message
drivers/md/Kconfig +9 -0
@@ -230,6 +230,15 @@ config DM_MIRROR Allow volume managers to mirror logical volumes, also needed for live data migration tools such as 'pvmove'. +config DM_CLOG + tristate "Mirror cluster logging (EXPERIMENTAL)" + depends on DM_MIRROR && EXPERIMENTAL + ---help--- + Cluster logging allows mirroring to become cluster-aware. + Mirror devices can be used by multiple machines at the + same time. Note: this will not make your applications + cluster-aware. + config DM_ZERO tristate "Zero target (EXPERIMENTAL)" depends on BLK_DEV_DM && EXPERIMENTAL
drivers/md/Makefile +2 -0
@@ -7,6 +7,7 @@ dm-mod-objs := dm.o dm-table.o dm-target dm-multipath-objs := dm-hw-handler.o dm-path-selector.o dm-mpath.o dm-snapshot-objs := dm-snap.o dm-exception-store.o dm-mirror-objs := dm-log.o dm-raid1.o +dm-clulog-objs := dm-clog.o dm-clog-tfr.o md-mod-objs := md.o bitmap.o raid456-objs := raid5.o raid6algos.o raid6recov.o raid6tables.o \ raid6int1.o raid6int2.o raid6int4.o \
@@ -35,6 +36,7 @@ obj-$(CONFIG_DM_MULTIPATH) += dm-multipa obj-$(CONFIG_DM_MULTIPATH_EMC) += dm-emc.o obj-$(CONFIG_DM_SNAPSHOT) += dm-snapshot.o obj-$(CONFIG_DM_MIRROR) += dm-mirror.o +obj-$(CONFIG_DM_CLOG) += dm-clulog.o obj-$(CONFIG_DM_ZERO) += dm-zero.o quiet_cmd_unroll = UNROLL $@
include/linux/dm-clog-tfr.h new file +50 -0
50 lines not on this page · raw message