dlm: bring message parsing to softirq context
Alexander Aring
maintainerv3 cover dlm: bring message parsing to softirq context
Hi,
this is version 3 of my attempt to bring dlm message parsing to softirq context. This will "minimal" improve the dlm message parsing that is still message after message parsing. There is still a context switch involed that is the dlm callback workqueue. This patch mostly have the effect that we call "queue_work()" more often. In future we will get rid of the dlm callback workqueue and call the DLM user callback directly in the dlm message processing softirq context. However this requires that DLM users adapt the changes and signal this capability over a new lockspace flag.
- Alex
changes since v3:
- add patches that got meanwhile on the mailinglist to resend them.
- add "remove schedule in dlm receive path" to remove a schedule() in the dlm msg processing path.
- change commit message of "dlm: do dlm message processing in softirq context" to mention the
changes since v2:
I changed in v2 that we split the root_list to root_list and a per lockspace masters list. The root_list can be used as a stack list variable in ls_recover() as it's only used for recovery handling. The masters list is somehow special because it is being used for other dlm nodes to dump the nodes master rsbs. The current implementation guarantees that this happens in a very special part of the recovery handling by using a kind of distributed cluster barriers. I added more sanity checks for this handling and a more per node based recovery log mechanism. There is also a TODO that describes we should keep track of all masters rsb while lockspace locking handling and not create them while recovery handling which I think should improve the handling to maybe get rid of those barriers.
Alexander Aring (16):
dlm: fix off-by-one waiters refcount handling
dlm: put lkbs instead of force free
dlm: remove allocation parameter in msg allocation
dlm: switch to GFP_ATOMIC in dlm allocations
dlm: move root_list functionality to recover.c
dlm: move master dir dump to own list
dlm: move root_list to ls_recover() stack
dlm: implement directory dump context
dlm: drop holding waiters mutex in waiters recovery
dlm: convert ls_waiters_mutex to spinlock
dlm: convert res_lock to spinlock
dlm: make requestqueue handling non sleepable
dlm: ls_recv_active semaphore to rwlock
dlm: remove schedule in dlm receive path
dlm: convert message parsing locks to disable bh
dlm: do dlm message processing in softirq context
Kunwu Chan (2):
fs: dlm: Simplify the allocation of slab caches in
dlm_midcomms_cache_create
fs: dlm: Simplify the allocation of slab caches in
dlm_lowcomms_msg_cache_create
fs/dlm/ast.c | 28 ++--- fs/dlm/debug_fs.c | 36 +++--- fs/dlm/dir.c | 147 ++++++++++++++++++---- fs/dlm/dir.h | 3 +- fs/dlm/dlm_internal.h | 20 +-- fs/dlm/lock.c | 277 +++++++++++++++++++++++------------------- fs/dlm/lock.h | 8 +- fs/dlm/lockspace.c | 96 ++++++++------- fs/dlm/lowcomms.c | 68 +++++------ fs/dlm/lowcomms.h | 5 +- fs/dlm/member.c | 23 ++-- fs/dlm/memory.c | 14 +-- fs/dlm/memory.h | 4 +- fs/dlm/midcomms.c | 67 +++++----- fs/dlm/midcomms.h | 3 +- fs/dlm/rcom.c | 33 +++-- fs/dlm/recover.c | 126 ++++++------------- fs/dlm/recover.h | 10 +- fs/dlm/recoverd.c | 125 ++++++++++++++++--- fs/dlm/requestqueue.c | 43 ++----- fs/dlm/user.c | 34 +++--- 21 files changed, 657 insertions(+), 513 deletions(-)
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 1/18 fs: dlm: Simplify the allocation of slab caches in dlm_midcomms_cache_create
From: Kunwu Chan <[email protected]>
Use the new KMEM_CACHE() macro instead of direct kmem_cache_create to simplify the creation of SLAB caches.
--- fs/dlm/midcomms.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
fs/dlm/midcomms.c +1 -2
@@ -226,8 +226,7 @@ static DEFINE_MUTEX(close_lock);struct kmem_cache *dlm_midcomms_cache_create (void ) { -return kmem_cache_create ("dlm_mhandle" ,sizeof (struct dlm_mhandle ), -0 ,0 ,NULL ); +return KMEM_CACHE (dlm_mhandle,0 ); }static inline const char *dlm_state_str (int state)
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 2/18 fs: dlm: Simplify the allocation of slab caches in dlm_lowcomms_msg_cache_create
In reply to Alexander AringFrom: Kunwu Chan <[email protected]>
Use the new KMEM_CACHE() macro instead of direct kmem_cache_create to simplify the creation of SLAB caches.
--- fs/dlm/lowcomms.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
fs/dlm/lowcomms.c +1 -1
@@ -248,7 +248,7 @@ struct kmem_cache *dlm_lowcomms_writequeue_cache_create(void)struct kmem_cache *dlm_lowcomms_msg_cache_create (void ) { -return kmem_cache_create ("dlm_msg" ,sizeof (struct dlm_msg ),0 ,0 ,NULL ); +return KMEM_CACHE (dlm_msg,0 ); }/* need to held writequeue_lock */
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 3/18 dlm: fix off-by-one waiters refcount handling
In reply to Alexander AringThere was a wrong conversion to atomic counters in commit 75a7d60134ce ("fs: dlm: handle lkb wait count as atomic_t"), when atomic_dec_and_test() returns true it will decrement at first and then return true if it hits zero. This means we will mis a unhold_lkb() for the last iteration. This patch fixes this issue and if the last reference is taken we will remove the lkb from the waiters list as this is how it's supposed to work.
--- fs/dlm/lock.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
fs/dlm/lock.c +6 -4
@@ -5070,11 +5070,13 @@ int dlm_recover_waiters_post(struct dlm_ls *ls)/* drop all wait_count references we still * hold a reference for this iteration. */ -while (!atomic_dec_and_test (&lkb->lkb_wait_count)) -unhold_lkb (lkb); -mutex_lock (&ls->ls_waiters_mutex); -list_del_init (&lkb->lkb_wait_reply); +while (atomic_read (&lkb->lkb_wait_count)) { +if (atomic_dec_and_test (&lkb->lkb_wait_count)) +list_del_init (&lkb->lkb_wait_reply); + +unhold_lkb (lkb); + }mutex_unlock (&ls->ls_waiters_mutex);if (oc || ou) {
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 4/18 dlm: put lkbs instead of force free
In reply to Alexander AringThis patch converts a force free of the lkb idr and switch to use the lkbs put functionality. If there are still references hold due the lkb programming logic and its state it will be drop before. Instead of force freeing the lkbs of the idr using the refcounters makes sure we using the reference counters correctly. If we do that, then no rsb should be left on the lockspace keep hash bucket which is an additonally check added to this patch. All rsbs on the toss list should have a reference counter of 1.
--- fs/dlm/lock.c | 2 +- fs/dlm/lock.h | 1 + fs/dlm/lockspace.c | 31 +++++++++++++++++++++---------- 3 files changed, 23 insertions(+), 11 deletions(-)
fs/dlm/lock.c +1 -1
@@ -1368,7 +1368,7 @@ static void add_lkb(struct dlm_rsb *r, struct dlm_lkb *lkb, int status) } } -static void del_lkb (struct dlm_rsb *r,struct dlm_lkb *lkb) +void del_lkb (struct dlm_rsb *r,struct dlm_lkb *lkb) { lkb->lkb_status =0 ;list_del (&lkb->lkb_statequeue);
fs/dlm/lock.h +1 -0
@@ -60,6 +60,7 @@ int dlm_debug_add_lkb(struct dlm_ls *ls, uint32_t lkb_id, char *name, int len, int lkb_nodeid, unsigned int lkb_flags, int lkb_status);int dlm_debug_add_lkb_to_waiters (struct dlm_ls *ls,uint32_t lkb_id,int mstype,int to_nodeid); +void del_lkb (struct dlm_rsb *r,struct dlm_lkb *lkb);static inline int is_master (struct dlm_rsb *r) {
fs/dlm/lockspace.c +21 -10
@@ -737,14 +737,28 @@ static int lkb_idr_is_any(int id, void *p, void *data)return 1 ; } -static int lkb_idr_free (int id,void *p,void *data) +/* +* No locking required, lockspace usage should be synchronized +* to have any activity anymore. +*/ +static int lkb_idr_put (int id,void *p,void *data) {struct dlm_lkb *lkb = p; -if (lkb->lkb_lvbptr &&test_bit (DLM_IFL_MSTCPY_BIT , &lkb->lkb_iflags)) -dlm_free_lvb (lkb->lkb_lvbptr); +if (lkb->lkb_status) +del_lkb (lkb->lkb_resource, lkb); -dlm_free_lkb (lkb); +/* drop all wait_count references we still +* hold a reference for this iteration. +*/ +while (atomic_read (&lkb->lkb_wait_count)) { +if (atomic_dec_and_test (&lkb->lkb_wait_count)) +list_del_init (&lkb->lkb_wait_reply); + +WARN_ON_ONCE (dlm_put_lkb (lkb)); + } + +WARN_ON_ONCE (!dlm_put_lkb (lkb));return 0 ; }
@@ -826,7 +840,7 @@ static int release_lockspace(struct dlm_ls *ls, int force)* Free all lkb's in idr */ -idr_for_each (&ls->ls_lkbidr, lkb_idr_free, ls); +idr_for_each (&ls->ls_lkbidr, lkb_idr_put, ls);idr_destroy (&ls->ls_lkbidr);/*
@@ -834,15 +848,12 @@ static int release_lockspace(struct dlm_ls *ls, int force)*/ for (i =0 ; i < ls->ls_rsbtbl_size; i++) { -while ((n =rb_first (&ls->ls_rsbtbl[i].keep))) { - rsb = rb_entry(n, struct dlm_rsb, res_hashnode); -rb_erase (n, &ls->ls_rsbtbl[i].keep); -dlm_free_rsb (rsb); - } +WARN_ON_ONCE (!RB_EMPTY_ROOT (&ls->ls_rsbtbl[i].keep));while ((n =rb_first (&ls->ls_rsbtbl[i].toss))) { rsb = rb_entry(n, struct dlm_rsb, res_hashnode);rb_erase (n, &ls->ls_rsbtbl[i].toss); +WARN_ON_ONCE (kref_read (&rsb->res_ref) !=1 );dlm_free_rsb (rsb); } }
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 5/18 dlm: remove allocation parameter in msg allocation
In reply to Alexander AringThis patch removes the context parameter for message allocations and replace it by always do GFP_ATOMIC. We are preparing to process dlm message in softirq context therefore it's necessary to switch to GFP_ATOMIC allocation as we cannot sleep in this context. To simplify the code overall we just drop the allocation flag and have GFP_ATOMIC hardcoded when calling the allocation function.
--- fs/dlm/lock.c | 31 ++++++++++++------------------- fs/dlm/lowcomms.c | 16 +++++++--------- fs/dlm/lowcomms.h | 5 ++--- fs/dlm/memory.c | 8 ++++---- fs/dlm/memory.h | 4 ++-- fs/dlm/midcomms.c | 24 ++++++++++-------------- fs/dlm/midcomms.h | 3 +-- fs/dlm/rcom.c | 7 +++---- 8 files changed, 41 insertions(+), 57 deletions(-)
fs/dlm/lock.c +12 -19
@@ -3329,8 +3329,7 @@ int dlm_unlock(dlm_lockspace_t *lockspace,static int _create_message (struct dlm_ls *ls,int mb_len,int to_nodeid,int mstype,struct dlm_message **ms_ret, -struct dlm_mhandle **mh_ret, -gfp_t allocation) +struct dlm_mhandle **mh_ret) {struct dlm_message *ms;struct dlm_mhandle *mh;
@@ -3340,7 +3339,7 @@ static int _create_message(struct dlm_ls *ls, int mb_len,pass into midcomms_commit and a message buffer (mb) that we write our data into */ - mh =dlm_midcomms_get_mhandle (to_nodeid, mb_len, allocation, &mb); + mh =dlm_midcomms_get_mhandle (to_nodeid, mb_len, &mb);if (!mh)return -ENOBUFS ;
@@ -3362,8 +3361,7 @@ static int _create_message(struct dlm_ls *ls, int mb_len,static int create_message (struct dlm_rsb *r,struct dlm_lkb *lkb,int to_nodeid,int mstype,struct dlm_message **ms_ret, -struct dlm_mhandle **mh_ret, -gfp_t allocation) +struct dlm_mhandle **mh_ret) {int mb_len =sizeof (struct dlm_message );
@@ -3384,7 +3382,7 @@ static int create_message(struct dlm_rsb *r, struct dlm_lkb *lkb, }return _create_message (r->res_ls, mb_len, to_nodeid, mstype, - ms_ret, mh_ret, allocation); + ms_ret, mh_ret); }/* further lowcomms enhancements or alternate implementations may make
@@ -3453,7 +3451,7 @@ static int send_common(struct dlm_rsb *r, struct dlm_lkb *lkb, int mstype)if (error)return error; - error =create_message (r, lkb, to_nodeid, mstype, &ms, &mh,GFP_NOFS ); + error =create_message (r, lkb, to_nodeid, mstype, &ms, &mh);if (error) goto fail;
@@ -3513,8 +3511,7 @@ static int send_grant(struct dlm_rsb *r, struct dlm_lkb *lkb) to_nodeid = lkb->lkb_nodeid; - error =create_message (r, lkb, to_nodeid,DLM_MSG_GRANT , &ms, &mh, -GFP_NOFS ); + error =create_message (r, lkb, to_nodeid,DLM_MSG_GRANT , &ms, &mh);if (error) goto out;
@@ -3535,8 +3532,7 @@ static int send_bast(struct dlm_rsb *r, struct dlm_lkb *lkb, int mode) to_nodeid = lkb->lkb_nodeid; - error =create_message (r,NULL , to_nodeid,DLM_MSG_BAST , &ms, &mh, -GFP_NOFS ); + error =create_message (r,NULL , to_nodeid,DLM_MSG_BAST , &ms, &mh);if (error) goto out;
@@ -3561,8 +3557,7 @@ static int send_lookup(struct dlm_rsb *r, struct dlm_lkb *lkb)if (error)return error; - error =create_message (r,NULL , to_nodeid,DLM_MSG_LOOKUP , &ms, &mh, -GFP_NOFS ); + error =create_message (r,NULL , to_nodeid,DLM_MSG_LOOKUP , &ms, &mh);if (error) goto fail;
@@ -3586,8 +3581,7 @@ static int send_remove(struct dlm_rsb *r) to_nodeid =dlm_dir_nodeid (r); - error =create_message (r,NULL , to_nodeid,DLM_MSG_REMOVE , &ms, &mh, -GFP_ATOMIC ); + error =create_message (r,NULL , to_nodeid,DLM_MSG_REMOVE , &ms, &mh);if (error) goto out;
@@ -3608,7 +3602,7 @@ static int send_common_reply(struct dlm_rsb *r, struct dlm_lkb *lkb, to_nodeid = lkb->lkb_nodeid; - error =create_message (r, lkb, to_nodeid, mstype, &ms, &mh,GFP_NOFS ); + error =create_message (r, lkb, to_nodeid, mstype, &ms, &mh);if (error) goto out;
@@ -3650,8 +3644,7 @@ static int send_lookup_reply(struct dlm_ls *ls,struct dlm_mhandle *mh;int error, nodeid =le32_to_cpu (ms_in->m_header.h_nodeid); - error =create_message (r,NULL , nodeid,DLM_MSG_LOOKUP_REPLY , &ms, &mh, -GFP_NOFS ); + error =create_message (r,NULL , nodeid,DLM_MSG_LOOKUP_REPLY , &ms, &mh);if (error) goto out;
@@ -6065,7 +6058,7 @@ static int send_purge(struct dlm_ls *ls, int nodeid, int pid)int error; error =_create_message (ls,sizeof (struct dlm_message ), nodeid, -DLM_MSG_PURGE , &ms, &mh,GFP_NOFS ); +DLM_MSG_PURGE , &ms, &mh);if (error)return error; ms->m_nodeid =cpu_to_le32 (nodeid);
fs/dlm/lowcomms.c +7 -9
@@ -1229,14 +1229,13 @@ static struct writequeue_entry *new_wq_entry(struct connection *con, int len, };static struct dlm_msg *dlm_lowcomms_new_msg_con (struct connection *con,int len, -gfp_t allocation,char **ppc, -void (*cb)(void *data), +char **ppc,void (*cb)(void *data),void *data) {struct writequeue_entry *e;struct dlm_msg *msg; - msg =dlm_allocate_msg (allocation); + msg =dlm_allocate_msg ();if (!msg)return NULL ;
@@ -1261,9 +1260,8 @@ static struct dlm_msg *dlm_lowcomms_new_msg_con(struct connection *con, int len,* dlm_lowcomms_commit_msg which is a must call if success */ #ifndef __CHECKER__ -struct dlm_msg *dlm_lowcomms_new_msg (int nodeid,int len,gfp_t allocation, -char **ppc,void (*cb)(void *data), -void *data) +struct dlm_msg *dlm_lowcomms_new_msg (int nodeid,int len,char **ppc, +void (*cb)(void *data),void *data) {struct connection *con;struct dlm_msg *msg;
@@ -1284,7 +1282,7 @@ struct dlm_msg *dlm_lowcomms_new_msg(int nodeid, int len, gfp_t allocation,return NULL ; } - msg =dlm_lowcomms_new_msg_con (con, len, allocation, ppc, cb, data); + msg =dlm_lowcomms_new_msg_con (con, len, ppc, cb, data);if (!msg) {srcu_read_unlock (&connections_srcu, idx);return NULL ;
@@ -1348,8 +1346,8 @@ int dlm_lowcomms_resend_msg(struct dlm_msg *msg)if (msg->retransmit)return 1 ; - msg_resend =dlm_lowcomms_new_msg_con (msg->entry->con, msg->len, -GFP_ATOMIC , &ppc,NULL ,NULL ); + msg_resend =dlm_lowcomms_new_msg_con (msg->entry->con, msg->len, &ppc, +NULL ,NULL );if (!msg_resend)return -ENOMEM ;
fs/dlm/lowcomms.h +2 -3
@@ -39,9 +39,8 @@ void dlm_lowcomms_stop(void);void dlm_lowcomms_init (void );void dlm_lowcomms_exit (void );int dlm_lowcomms_close (int nodeid); -struct dlm_msg *dlm_lowcomms_new_msg (int nodeid,int len,gfp_t allocation, -char **ppc,void (*cb)(void *data), -void *data); +struct dlm_msg *dlm_lowcomms_new_msg (int nodeid,int len,char **ppc, +void (*cb)(void *data),void *data);void dlm_lowcomms_commit_msg (struct dlm_msg *msg);void dlm_lowcomms_put_msg (struct dlm_msg *msg);int dlm_lowcomms_resend_msg (struct dlm_msg *msg);
fs/dlm/memory.c +4 -4
@@ -134,9 +134,9 @@ void dlm_free_lkb(struct dlm_lkb *lkb)kmem_cache_free (lkb_cache, lkb); } -struct dlm_mhandle *dlm_allocate_mhandle (gfp_t allocation) +struct dlm_mhandle *dlm_allocate_mhandle (void ) { -return kmem_cache_alloc (mhandle_cache, allocation); +return kmem_cache_alloc (mhandle_cache,GFP_ATOMIC ); }void dlm_free_mhandle (struct dlm_mhandle *mhandle)
@@ -154,9 +154,9 @@ void dlm_free_writequeue(struct writequeue_entry *writequeue)kmem_cache_free (writequeue_cache, writequeue); } -struct dlm_msg *dlm_allocate_msg (gfp_t allocation) +struct dlm_msg *dlm_allocate_msg (void ) { -return kmem_cache_alloc (msg_cache, allocation); +return kmem_cache_alloc (msg_cache,GFP_ATOMIC ); }void dlm_free_msg (struct dlm_msg *msg)
fs/dlm/memory.h +2 -2
@@ -20,11 +20,11 @@ struct dlm_lkb *dlm_allocate_lkb(struct dlm_ls *ls);void dlm_free_lkb (struct dlm_lkb *l);char *dlm_allocate_lvb (struct dlm_ls *ls);void dlm_free_lvb (char *l); -struct dlm_mhandle *dlm_allocate_mhandle (gfp_t allocation); +struct dlm_mhandle *dlm_allocate_mhandle (void );void dlm_free_mhandle (struct dlm_mhandle *mhandle);struct writequeue_entry *dlm_allocate_writequeue (void );void dlm_free_writequeue (struct writequeue_entry *writequeue); -struct dlm_msg *dlm_allocate_msg (gfp_t allocation); +struct dlm_msg *dlm_allocate_msg (void );void dlm_free_msg (struct dlm_msg *msg);struct dlm_callback *dlm_allocate_cb (void );void dlm_free_cb (struct dlm_callback *cb);
fs/dlm/midcomms.c +10 -14
@@ -379,8 +379,7 @@ static int dlm_send_ack(int nodeid, uint32_t seq)struct dlm_msg *msg;char *ppc; - msg =dlm_lowcomms_new_msg (nodeid, mb_len,GFP_ATOMIC , &ppc, -NULL ,NULL ); + msg =dlm_lowcomms_new_msg (nodeid, mb_len, &ppc,NULL ,NULL );if (!msg)return -ENOMEM ;
@@ -428,7 +427,7 @@ static int dlm_send_fin(struct midcomms_node *node,struct dlm_mhandle *mh;char *ppc; - mh =dlm_midcomms_get_mhandle (node->nodeid, mb_len,GFP_ATOMIC , &ppc); + mh =dlm_midcomms_get_mhandle (node->nodeid, mb_len, &ppc);if (!mh)return -ENOMEM ;
@@ -976,13 +975,13 @@ static void midcomms_new_msg_cb(void *data) }static struct dlm_msg *dlm_midcomms_get_msg_3_2 (struct dlm_mhandle *mh,int nodeid, -int len,gfp_t allocation,char **ppc) +int len,char **ppc) {struct dlm_opts *opts;struct dlm_msg *msg; msg =dlm_lowcomms_new_msg (nodeid, len +DLM_MIDCOMMS_OPT_LEN , - allocation, ppc, midcomms_new_msg_cb, mh); + ppc, midcomms_new_msg_cb, mh);if (!msg)return NULL ;
@@ -1001,8 +1000,7 @@ static struct dlm_msg *dlm_midcomms_get_msg_3_2(struct dlm_mhandle *mh, int node* dlm_midcomms_commit_mhandle which is a must call if success */ #ifndef __CHECKER__ -struct dlm_mhandle *dlm_midcomms_get_mhandle (int nodeid,int len, -gfp_t allocation,char **ppc) +struct dlm_mhandle *dlm_midcomms_get_mhandle (int nodeid,int len,char **ppc) {struct midcomms_node *node;struct dlm_mhandle *mh;
@@ -1017,7 +1015,7 @@ struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len,/* this is a bug, however we going on and hope it will be resolved */ WARN_ON_ONCE (test_bit (DLM_NODE_FLAG_STOP_TX , &node->flags)); - mh =dlm_allocate_mhandle (allocation); + mh =dlm_allocate_mhandle ();if (!mh) goto err;
@@ -1028,8 +1026,7 @@ struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len,switch (node->version) {case DLM_VERSION_3_1 : - msg =dlm_lowcomms_new_msg (nodeid, len, allocation, ppc, -NULL ,NULL ); + msg =dlm_lowcomms_new_msg (nodeid, len, ppc,NULL ,NULL );if (!msg) {dlm_free_mhandle (mh); goto err;
@@ -1040,8 +1037,7 @@ struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len,/* send ack back if necessary */ dlm_send_ack_threshold (node,DLM_SEND_ACK_BACK_MSG_THRESHOLD ); - msg =dlm_midcomms_get_msg_3_2 (mh, nodeid, len, allocation, - ppc); + msg =dlm_midcomms_get_msg_3_2 (mh, nodeid, len, ppc);if (!msg) {dlm_free_mhandle (mh); goto err;
@@ -1501,8 +1497,8 @@ int dlm_midcomms_rawmsg_send(struct midcomms_node *node, void *buf, rd.node = node; rd.buf = buf; - msg =dlm_lowcomms_new_msg (node->nodeid, buflen,GFP_NOFS , - &msgbuf, midcomms_new_rawmsg_cb, &rd); + msg =dlm_lowcomms_new_msg (node->nodeid, buflen, &msgbuf, + midcomms_new_rawmsg_cb, &rd);if (!msg)return -ENOMEM ;
fs/dlm/midcomms.h +1 -2
@@ -16,8 +16,7 @@ struct midcomms_node;int dlm_validate_incoming_buffer (int nodeid,unsigned char *buf,int len);int dlm_process_incoming_buffer (int nodeid,unsigned char *buf,int buflen); -struct dlm_mhandle *dlm_midcomms_get_mhandle (int nodeid,int len, -gfp_t allocation,char **ppc); +struct dlm_mhandle *dlm_midcomms_get_mhandle (int nodeid,int len,char **ppc);void dlm_midcomms_commit_mhandle (struct dlm_mhandle *mh,const void *name,int namelen);int dlm_midcomms_addr (int nodeid,struct sockaddr_storage *addr,int len);
fs/dlm/rcom.c +3 -4
@@ -55,7 +55,7 @@ static int create_rcom(struct dlm_ls *ls, int to_nodeid, int type, int len,struct dlm_mhandle *mh;char *mb; - mh =dlm_midcomms_get_mhandle (to_nodeid, mb_len,GFP_NOFS , &mb); + mh =dlm_midcomms_get_mhandle (to_nodeid, mb_len, &mb);if (!mh) {log_print ("%s to %d type %d len %d ENOBUFS" , __func__, to_nodeid, type, len);
@@ -75,8 +75,7 @@ static int create_rcom_stateless(struct dlm_ls *ls, int to_nodeid, int type,struct dlm_msg *msg;char *mb; - msg =dlm_lowcomms_new_msg (to_nodeid, mb_len,GFP_NOFS , &mb, -NULL ,NULL ); + msg =dlm_lowcomms_new_msg (to_nodeid, mb_len, &mb,NULL ,NULL );if (!msg) {log_print ("create_rcom to %d type %d len %d ENOBUFS" , to_nodeid, type, len);
@@ -510,7 +509,7 @@ int dlm_send_ls_not_ready(int nodeid, const struct dlm_rcom *rc_in)char *mb;int mb_len =sizeof (struct dlm_rcom ) +sizeof (struct rcom_config ); - mh =dlm_midcomms_get_mhandle (nodeid, mb_len,GFP_NOFS , &mb); + mh =dlm_midcomms_get_mhandle (nodeid, mb_len, &mb);if (!mh)return -ENOBUFS ;
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 6/18 dlm: switch to GFP_ATOMIC in dlm allocations
In reply to Alexander AringThis patch prepares to process dlm messages in softirq context. To switch dlm to parse messages in softirq context some code parts either runs inside the softirq context or need to switch to run while a spinlock is held. This patch prepares to switch the allocation context to GFP_ATOMIC for those places. It's not possible anymore to preload idr allocations. However this is only a performance speedup and we might switch to xarray implementation with more lockless readers paradigms.
--- fs/dlm/lock.c | 2 -- fs/dlm/memory.c | 6 +++--- fs/dlm/recover.c | 2 -- fs/dlm/requestqueue.c | 2 +- 4 files changed, 4 insertions(+), 8 deletions(-)
fs/dlm/lock.c +0 -2
@@ -1208,13 +1208,11 @@ static int _create_lkb(struct dlm_ls *ls, struct dlm_lkb **lkb_ret,spin_lock_init (&lkb->lkb_cb_lock);INIT_WORK (&lkb->lkb_cb_work, dlm_callback_work); -idr_preload (GFP_NOFS );spin_lock (&ls->ls_lkbidr_spin); rv =idr_alloc (&ls->ls_lkbidr, lkb, start, end,GFP_NOWAIT );if (rv >=0 ) lkb->lkb_id = rv;spin_unlock (&ls->ls_lkbidr_spin); -idr_preload_end ();if (rv <0 ) {log_error (ls,"create_lkb idr error %d" , rv);
fs/dlm/memory.c +3 -3
@@ -84,7 +84,7 @@ char *dlm_allocate_lvb(struct dlm_ls *ls) {char *p; - p =kzalloc (ls->ls_lvblen,GFP_NOFS ); + p =kzalloc (ls->ls_lvblen,GFP_ATOMIC );return p; }
@@ -97,7 +97,7 @@ struct dlm_rsb *dlm_allocate_rsb(struct dlm_ls *ls) {struct dlm_rsb *r; - r =kmem_cache_zalloc (rsb_cache,GFP_NOFS ); + r =kmem_cache_zalloc (rsb_cache,GFP_ATOMIC );return r; }
@@ -112,7 +112,7 @@ struct dlm_lkb *dlm_allocate_lkb(struct dlm_ls *ls) {struct dlm_lkb *lkb; - lkb =kmem_cache_zalloc (lkb_cache,GFP_NOFS ); + lkb =kmem_cache_zalloc (lkb_cache,GFP_ATOMIC );return lkb; }
fs/dlm/recover.c +0 -2
@@ -310,7 +310,6 @@ static int recover_idr_add(struct dlm_rsb *r)struct dlm_ls *ls = r->res_ls;int rv; -idr_preload (GFP_NOFS );spin_lock (&ls->ls_recover_idr_lock);if (r->res_id) { rv =-1 ;
@@ -326,7 +325,6 @@ static int recover_idr_add(struct dlm_rsb *r) rv =0 ; out_unlock:spin_unlock (&ls->ls_recover_idr_lock); -idr_preload_end ();return rv; }
fs/dlm/requestqueue.c +1 -1
@@ -37,7 +37,7 @@ void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid,int length =le16_to_cpu (ms->m_header.h_length) -sizeof (struct dlm_message ); - e =kmalloc (sizeof (struct rq_entry ) + length,GFP_NOFS ); + e =kmalloc (sizeof (struct rq_entry ) + length,GFP_ATOMIC );if (!e) {log_print ("dlm_add_requestqueue: out of memory len %d" , length);return ;
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 7/18 dlm: move root_list functionality to recover.c
In reply to Alexander AringThis patch moves dlm_create_root_list() and dlm_release_root_list() to recover.c and declare them static because they are only used there.
--- fs/dlm/recover.c | 42 ------------------------------------------ fs/dlm/recover.h | 2 -- fs/dlm/recoverd.c | 39 +++++++++++++++++++++++++++++++++++++++ 3 files changed, 39 insertions(+), 44 deletions(-)
fs/dlm/recover.c +0 -42
@@ -889,48 +889,6 @@ void dlm_recover_rsbs(struct dlm_ls *ls)/* Create a single list of all root rsb's to be used during recovery */ -int dlm_create_root_list (struct dlm_ls *ls) -{ -struct rb_node *n; -struct dlm_rsb *r; -int i, error =0 ; - -down_write (&ls->ls_root_sem); -if (!list_empty (&ls->ls_root_list)) { -log_error (ls,"root list not empty" ); - error = -EINVAL ; - goto out; - } - -for (i =0 ; i < ls->ls_rsbtbl_size; i++) { -spin_lock (&ls->ls_rsbtbl[i].lock); -for (n =rb_first (&ls->ls_rsbtbl[i].keep); n; n =rb_next (n)) { - r = rb_entry(n, struct dlm_rsb, res_hashnode); -list_add (&r->res_root_list, &ls->ls_root_list); -dlm_hold_rsb (r); - } - -if (!RB_EMPTY_ROOT (&ls->ls_rsbtbl[i].toss)) -log_error (ls,"dlm_create_root_list toss not empty" ); -spin_unlock (&ls->ls_rsbtbl[i].lock); - } - out: -up_write (&ls->ls_root_sem); -return error; -} - -void dlm_release_root_list (struct dlm_ls *ls) -{ -struct dlm_rsb *r, *safe; - -down_write (&ls->ls_root_sem); -list_for_each_entry_safe (r, safe, &ls->ls_root_list, res_root_list) { -list_del_init (&r->res_root_list); -dlm_put_rsb (r); - } -up_write (&ls->ls_root_sem); -} -void dlm_clear_toss (struct dlm_ls *ls) {struct rb_node *n, *next;
fs/dlm/recover.h +0 -2
@@ -23,8 +23,6 @@ int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq);int dlm_recover_master_reply (struct dlm_ls *ls,const struct dlm_rcom *rc);int dlm_recover_locks (struct dlm_ls *ls,uint64_t seq);void dlm_recovered_lock (struct dlm_rsb *r); -int dlm_create_root_list (struct dlm_ls *ls); -void dlm_release_root_list (struct dlm_ls *ls);void dlm_clear_toss (struct dlm_ls *ls);void dlm_recover_rsbs (struct dlm_ls *ls);
fs/dlm/recoverd.c +39 -0
@@ -20,6 +20,45 @@#include "requestqueue.h" #include "recoverd.h" +static void dlm_create_root_list (struct dlm_ls *ls) +{ +struct rb_node *n; +struct dlm_rsb *r; +int i; + +down_write (&ls->ls_root_sem); +if (!list_empty (&ls->ls_root_list)) { +log_error (ls,"root list not empty" ); + goto out; + } + +for (i =0 ; i < ls->ls_rsbtbl_size; i++) { +spin_lock_bh (&ls->ls_rsbtbl[i].lock); +for (n =rb_first (&ls->ls_rsbtbl[i].keep); n; n =rb_next (n)) { + r = rb_entry(n, struct dlm_rsb, res_hashnode); +list_add (&r->res_root_list, &ls->ls_root_list); +dlm_hold_rsb (r); + } + +if (!RB_EMPTY_ROOT (&ls->ls_rsbtbl[i].toss)) +log_error (ls,"%s toss not empty" , __func__); +spin_unlock_bh (&ls->ls_rsbtbl[i].lock); + } + out: +up_write (&ls->ls_root_sem); +} + +static void dlm_release_root_list (struct dlm_ls *ls) +{ +struct dlm_rsb *r, *safe; + +down_write (&ls->ls_root_sem); +list_for_each_entry_safe (r, safe, &ls->ls_root_list, res_root_list) { +list_del_init (&r->res_root_list); +dlm_put_rsb (r); + } +up_write (&ls->ls_root_sem); +}/* If the start for which we're re-enabling locking (seq) has been superseded by a newer stop (ls_recover_seq), we need to leave locking disabled.
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 8/18 dlm: move master dir dump to own list
In reply to Alexander AringThis patch moves the master directory dump, means dlm_rsbs where we are the master of (res_nodeid == 0), to it's own list handling. Currently the only mutual access to ls->root_list is due the master directory dump. Put it into it's own list handling allows us to put the root_list out of the global per lockspace context and make it lockless. While on it move the rw semaphore to a rwlock as the context allows it.
Add a comment that we should keep track of our own master rsbs while locking occurs instead of recovery creates it in a snapshot like mode.
--- fs/dlm/dir.c | 22 ++++++--------- fs/dlm/dlm_internal.h | 3 ++ fs/dlm/lockspace.c | 2 ++ fs/dlm/recoverd.c | 64 +++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 77 insertions(+), 14 deletions(-)
fs/dlm/dir.c +8 -14
@@ -216,16 +216,13 @@ static struct dlm_rsb *find_rsb_root(struct dlm_ls *ls, const char *name,if (!rv)return r; -down_read (&ls->ls_root_sem); -list_for_each_entry (r, &ls->ls_root_list, res_root_list) { +list_for_each_entry (r, &ls->ls_masters_list, res_masters_list) {if (len == r->res_length && !memcmp (name, r->res_name, len)) { -up_read (&ls->ls_root_sem);log_debug (ls,"find_rsb_root revert to root_list %s" , r->res_name);return r; } } -up_read (&ls->ls_root_sem);return NULL ; }
@@ -241,7 +238,7 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,int offset =0 , dir_nodeid;__be16 be_namelen; -down_read (&ls->ls_root_sem); +read_lock (&ls->ls_masters_lock);if (inlen >1 ) { r =find_rsb_root (ls, inbuf, inlen);
@@ -250,16 +247,13 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen, nodeid, inlen, inlen, inbuf); goto out; } - list = r->res_root_list.next; + list = r->res_masters_list.next; }else { - list = ls->ls_root_list.next; + list = ls->ls_masters_list.next; } -for (offset =0 ; list != &ls->ls_root_list; list = list->next) { - r = list_entry(list, struct dlm_rsb, res_root_list); -if (r->res_nodeid) -continue ; - +for (offset =0 ; list != &ls->ls_masters_list; list = list->next) { + r = list_entry(list, struct dlm_rsb, res_masters_list); dir_nodeid =dlm_dir_nodeid (r);if (dir_nodeid != nodeid)continue ;
@@ -294,7 +288,7 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,* terminating record. */ -if ((list == &ls->ls_root_list) && +if ((list == &ls->ls_masters_list) && (offset +sizeof (uint16_t ) <= outlen)) { be_namelen =cpu_to_be16 (0xFFFF );memcpy (outbuf + offset, &be_namelen,sizeof (__be16));
@@ -302,6 +296,6 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen, ls->ls_recover_dir_sent_msg++; } out: -up_read (&ls->ls_root_sem); +read_unlock (&ls->ls_masters_lock); }
fs/dlm/dlm_internal.h +3 -0
@@ -312,6 +312,7 @@ struct dlm_rsb {struct list_head res_waitqueue;struct list_head res_root_list;/* used for recovery */ +struct list_head res_masters_list;/* used for recovery */ struct list_head res_recover_list;/* used for recovery */ int res_recover_locks_count;
@@ -645,6 +646,8 @@ struct dlm_ls {struct list_head ls_root_list;/* root resources */ struct rw_semaphore ls_root_sem;/* protect root_list */ +struct list_head ls_masters_list;/* root resources */ +rwlock_t ls_masters_lock;/* protect root_list */ const struct dlm_lockspace_ops *ls_ops;void *ls_ops_arg;
fs/dlm/lockspace.c +2 -0
@@ -582,6 +582,8 @@ static int new_lockspace(const char *name, const char *cluster,init_waitqueue_head (&ls->ls_wait_general);INIT_LIST_HEAD (&ls->ls_root_list);init_rwsem (&ls->ls_root_sem); +INIT_LIST_HEAD (&ls->ls_masters_list); +rwlock_init (&ls->ls_masters_lock);spin_lock (&lslist_lock); ls->ls_create_count =1 ;
fs/dlm/recoverd.c +64 -0
97 lines not on this page · raw message
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 9/18 dlm: move root_list to ls_recover() stack
In reply to Alexander AringThis patch moves the per lockspace ls_root_list list which is mainly used for snapshoting all dlm_rsb from a hash to a list to do recovery handling into the recovery function ls_recover() as stack variable.
Doing that shows that there is no need for locking the ls_root_list which is created at the beginning of ls_recover() and destroyed at the ending of ls_recover(). In between only functionality is called doing read only access to the root_list stack variable.
A special case is assigning the per lockspace ls_recover_dir_root_list variable to the stack variable. The ls_recover_dir_root_list get accessed by another concurrent process dlm_copy_master_names() during the time between ls_recover_dir_root_list is set and set to NULL again. This is done by a special distributed barrier functionality between dlm_recover_members_wait() and dlm_recover_directory_wait(). A comment was made to mention about this handling which might be changed to a better behaviour in future. However setting the ls_recover_dir_root_list to the stack variable and set it to NULL in this specific time will show us potential issues with the recovery handling if it's breaks.
--- fs/dlm/dir.c | 6 ++---- fs/dlm/dir.h | 3 ++- fs/dlm/dlm_internal.h | 6 ++---- fs/dlm/lock.c | 6 ++---- fs/dlm/lock.h | 2 +- fs/dlm/lockspace.c | 2 -- fs/dlm/recover.c | 30 ++++++++++-------------------- fs/dlm/recover.h | 8 +++++--- fs/dlm/recoverd.c | 35 +++++++++++++---------------------- 9 files changed, 37 insertions(+), 61 deletions(-)
fs/dlm/dir.c +2 -4
@@ -47,15 +47,13 @@ int dlm_dir_nodeid(struct dlm_rsb *r)return r->res_dir_nodeid; } -void dlm_recover_dir_nodeid (struct dlm_ls *ls) +void dlm_recover_dir_nodeid (struct dlm_ls *ls,const struct list_head *root_list) {struct dlm_rsb *r; -down_read (&ls->ls_root_sem); -list_for_each_entry (r, &ls->ls_root_list, res_root_list) { +list_for_each_entry (r, root_list, res_root_list) { r->res_dir_nodeid =dlm_hash2nodeid (ls, r->res_hash); } -up_read (&ls->ls_root_sem); }int dlm_recover_directory (struct dlm_ls *ls,uint64_t seq)
fs/dlm/dir.h +2 -1
@@ -14,7 +14,8 @@int dlm_dir_nodeid (struct dlm_rsb *rsb);int dlm_hash2nodeid (struct dlm_ls *ls,uint32_t hash); -void dlm_recover_dir_nodeid (struct dlm_ls *ls); +void dlm_recover_dir_nodeid (struct dlm_ls *ls, +const struct list_head *root_list);int dlm_recover_directory (struct dlm_ls *ls,uint64_t seq);void dlm_copy_master_names (struct dlm_ls *ls,const char *inbuf,int inlen,char *outbuf,int outlen,int nodeid);
fs/dlm/dlm_internal.h +2 -4
@@ -644,10 +644,8 @@ struct dlm_ls {wait_queue_head_t ls_recover_lock_wait;spinlock_t ls_clear_proc_locks; -struct list_head ls_root_list;/* root resources */ -struct rw_semaphore ls_root_sem;/* protect root_list */ -struct list_head ls_masters_list;/* root resources */ -rwlock_t ls_masters_lock;/* protect root_list */ +struct list_head ls_masters_list;/* root resources */ +rwlock_t ls_masters_lock;/* protect root_list */ const struct dlm_lockspace_ops *ls_ops;void *ls_ops_arg;
fs/dlm/lock.c +2 -4
@@ -5187,7 +5187,7 @@ static void purge_dead_list(struct dlm_ls *ls, struct dlm_rsb *r,/* Get rid of locks held by nodes that are gone. */ -void dlm_recover_purge (struct dlm_ls *ls) +void dlm_recover_purge (struct dlm_ls *ls,const struct list_head *root_list) {struct dlm_rsb *r;struct dlm_member *memb;
@@ -5206,8 +5206,7 @@ void dlm_recover_purge(struct dlm_ls *ls)if (!nodes_count)return ; -down_write (&ls->ls_root_sem); -list_for_each_entry (r, &ls->ls_root_list, res_root_list) { +list_for_each_entry (r, root_list, res_root_list) {hold_rsb (r);lock_rsb (r);if (is_master (r)) {
@@ -5222,7 +5221,6 @@ void dlm_recover_purge(struct dlm_ls *ls)unhold_rsb (r);cond_resched (); } -up_write (&ls->ls_root_sem);if (lkb_count) log_rinfo(ls, "dlm_recover_purge %u locks for %u nodes",
fs/dlm/lock.h +1 -1
@@ -31,7 +31,7 @@ int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,int dlm_search_rsb_tree (struct rb_root *tree,const void *name,int len,struct dlm_rsb **r_ret); -void dlm_recover_purge (struct dlm_ls *ls); +void dlm_recover_purge (struct dlm_ls *ls,const struct list_head *root_list);void dlm_purge_mstcpy_locks (struct dlm_rsb *r);void dlm_recover_grant (struct dlm_ls *ls);int dlm_recover_waiters_post (struct dlm_ls *ls);
fs/dlm/lockspace.c +0 -2
@@ -580,8 +580,6 @@ static int new_lockspace(const char *name, const char *cluster, ls->ls_recover_list_count =0 ; ls->ls_local_handle = ls;init_waitqueue_head (&ls->ls_wait_general); -INIT_LIST_HEAD (&ls->ls_root_list); -init_rwsem (&ls->ls_root_sem);INIT_LIST_HEAD (&ls->ls_masters_list);rwlock_init (&ls->ls_masters_lock);
fs/dlm/recover.c +10 -20
@@ -519,7 +519,8 @@ static int recover_master_static(struct dlm_rsb *r, unsigned int *count)* the correct dir node. */ -int dlm_recover_masters (struct dlm_ls *ls,uint64_t seq) +int dlm_recover_masters (struct dlm_ls *ls,uint64_t seq, +const struct list_head *root_list) {struct dlm_rsb *r;unsigned int total =0 ;
@@ -529,10 +530,8 @@ int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq)log_rinfo (ls,"dlm_recover_masters" ); -down_read (&ls->ls_root_sem); -list_for_each_entry (r, &ls->ls_root_list, res_root_list) { +list_for_each_entry (r, root_list, res_root_list) {if (dlm_recovery_stopped (ls)) { -up_read (&ls->ls_root_sem); error = -EINTR ; goto out; }
@@ -546,12 +545,9 @@ int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq)cond_resched (); total++; -if (error) { -up_read (&ls->ls_root_sem); +if (error) goto out; - } } -up_read (&ls->ls_root_sem);log_rinfo (ls,"dlm_recover_masters %u of %u" , count, total);
@@ -656,13 +652,13 @@ static int recover_locks(struct dlm_rsb *r, uint64_t seq)return error; } -int dlm_recover_locks (struct dlm_ls *ls,uint64_t seq) +int dlm_recover_locks (struct dlm_ls *ls,uint64_t seq, +const struct list_head *root_list) {struct dlm_rsb *r;int error, count =0 ; -down_read (&ls->ls_root_sem); -list_for_each_entry (r, &ls->ls_root_list, res_root_list) { +list_for_each_entry (r, root_list, res_root_list) {if (is_master (r)) {rsb_clear_flag (r,RSB_NEW_MASTER );continue ;
@@ -673,19 +669,15 @@ int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq)if (dlm_recovery_stopped (ls)) { error = -EINTR ; -up_read (&ls->ls_root_sem); goto out; } error =recover_locks (r, seq); -if (error) { -up_read (&ls->ls_root_sem); +if (error) goto out; - } count += r->res_recover_locks_count; } -up_read (&ls->ls_root_sem);log_rinfo (ls,"dlm_recover_locks %d out" , count);
@@ -854,13 +846,12 @@ static void recover_grant(struct dlm_rsb *r)rsb_set_flag (r,RSB_RECOVER_GRANT ); } -void dlm_recover_rsbs (struct dlm_ls *ls) +void dlm_recover_rsbs (struct dlm_ls *ls,const struct list_head *root_list) {struct dlm_rsb *r;unsigned int count =0 ; -down_read (&ls->ls_root_sem); -list_for_each_entry (r, &ls->ls_root_list, res_root_list) { +list_for_each_entry (r, root_list, res_root_list) {lock_rsb (r);if (is_master (r)) {if (rsb_flag (r,RSB_RECOVER_CONVERT ))
@@ -881,7 +872,6 @@ void dlm_recover_rsbs(struct dlm_ls *ls)rsb_clear_flag (r,RSB_NEW_MASTER2 );unlock_rsb (r); } -up_read (&ls->ls_root_sem);if (count)log_rinfo (ls,"dlm_recover_rsbs %d done" , count);
fs/dlm/recover.h +5 -3
@@ -19,12 +19,14 @@ int dlm_recover_members_wait(struct dlm_ls *ls, uint64_t seq);int dlm_recover_directory_wait (struct dlm_ls *ls,uint64_t seq);int dlm_recover_locks_wait (struct dlm_ls *ls,uint64_t seq);int dlm_recover_done_wait (struct dlm_ls *ls,uint64_t seq); -int dlm_recover_masters (struct dlm_ls *ls,uint64_t seq); +int dlm_recover_masters (struct dlm_ls *ls,uint64_t seq, +const struct list_head *root_list);int dlm_recover_master_reply (struct dlm_ls *ls,const struct dlm_rcom *rc); -int dlm_recover_locks (struct dlm_ls *ls,uint64_t seq); +int dlm_recover_locks (struct dlm_ls *ls,uint64_t seq, +const struct list_head *root_list);void dlm_recovered_lock (struct dlm_rsb *r);void dlm_clear_toss (struct dlm_ls *ls); -void dlm_recover_rsbs (struct dlm_ls *ls); +void dlm_recover_rsbs (struct dlm_ls *ls,const struct list_head *root_list);#endif /* __RECOVER_DOT_H__ */
fs/dlm/recoverd.c +13 -22
@@ -62,23 +62,17 @@ static void dlm_release_masters_list(struct dlm_ls *ls)write_unlock (&ls->ls_masters_lock); } -static void dlm_create_root_list (struct dlm_ls *ls) +static void dlm_create_root_list (struct dlm_ls *ls,struct list_head *root_list) {struct rb_node *n;struct dlm_rsb *r;int i; -down_write (&ls->ls_root_sem); -if (!list_empty (&ls->ls_root_list)) { -log_error (ls,"root list not empty" ); - goto out; - } -for (i =0 ; i < ls->ls_rsbtbl_size; i++) {spin_lock_bh (&ls->ls_rsbtbl[i].lock);for (n =rb_first (&ls->ls_rsbtbl[i].keep); n; n =rb_next (n)) { r = rb_entry(n, struct dlm_rsb, res_hashnode); -list_add (&r->res_root_list, &ls->ls_root_list); +list_add (&r->res_root_list, root_list);dlm_hold_rsb (r); }
@@ -86,20 +80,16 @@ static void dlm_create_root_list(struct dlm_ls *ls)log_error (ls,"%s toss not empty" , __func__);spin_unlock_bh (&ls->ls_rsbtbl[i].lock); } - out: -up_write (&ls->ls_root_sem); } -static void dlm_release_root_list (struct dlm_ls *ls) +static void dlm_release_root_list (struct list_head *root_list) {struct dlm_rsb *r, *safe; -down_write (&ls->ls_root_sem); -list_for_each_entry_safe (r, safe, &ls->ls_root_list, res_root_list) { +list_for_each_entry_safe (r, safe, root_list, res_root_list) {list_del_init (&r->res_root_list);dlm_put_rsb (r); } -up_write (&ls->ls_root_sem); }/* If the start for which we're re-enabling locking (seq) has been superseded
@@ -131,6 +121,7 @@ static int enable_locking(struct dlm_ls *ls, uint64_t seq)static int ls_recover (struct dlm_ls *ls,struct dlm_recover *rv) { +LIST_HEAD (root_list);unsigned long start;int error, neg =0 ;
@@ -147,7 +138,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)* routines. */ -dlm_create_root_list (ls); +dlm_create_root_list (ls, &root_list);/* * Add or remove nodes from the lockspace's ls_nodes list.
@@ -163,7 +154,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv) goto fail; } -dlm_recover_dir_nodeid (ls); +dlm_recover_dir_nodeid (ls, &root_list);/* Create a snapshot of all active rsbs were we are the master of. * During the barrier between dlm_recover_members_wait() and
@@ -241,14 +232,14 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)* Clear lkb's for departed nodes. */ -dlm_recover_purge (ls); +dlm_recover_purge (ls, &root_list);/* * Get new master nodeid's for rsb's that were mastered on * departed nodes. */ - error =dlm_recover_masters (ls, rv->seq); + error =dlm_recover_masters (ls, rv->seq, &root_list);if (error) {log_rinfo (ls,"dlm_recover_masters error %d" , error); goto fail;
@@ -258,7 +249,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)* Send our locks on remastered rsb's to the new masters. */ - error =dlm_recover_locks (ls, rv->seq); + error =dlm_recover_locks (ls, rv->seq, &root_list);if (error) {log_rinfo (ls,"dlm_recover_locks error %d" , error); goto fail;
@@ -281,7 +272,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)* settings. */ -dlm_recover_rsbs (ls); +dlm_recover_rsbs (ls, &root_list); }else {/* * Other lockspace members may be going through the "neg" steps
@@ -297,7 +288,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv) } } -dlm_release_root_list (ls); +dlm_release_root_list (&root_list);/* * Purge directory-related requests that are saved in requestqueue.
@@ -347,7 +338,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)return 0 ; fail: -dlm_release_root_list (ls); +dlm_release_root_list (&root_list);mutex_unlock (&ls->ls_recoverd_active);return error;
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 10/18 dlm: implement directory dump context
In reply to Alexander AringThis patch introduce to keep track of an directory dump in DLM. For now we only add more sanity checks if e.g. the recovery sequence number has been changed while dumping the directory. Another change is that we can keep track of a per nodeid directory dump that can be later being used to add log messages about how much entries in how many chunks was being sent to a specific nodeid.
That the whole dump depends on the recovery barrier, because the resource list is not manipulated during this time may later being improved. For now we add more sanity checks in the recovery low path to confirm there is no issue with the current behaviour e.g. it also checks if the same list entry was being returned from the last resource lookup vs last resource list entry.
--- fs/dlm/dir.c | 115 ++++++++++++++++++++++++++++++++++++++++-- fs/dlm/dlm_internal.h | 4 +- fs/dlm/lockspace.c | 2 + fs/dlm/recoverd.c | 5 -- 4 files changed, 116 insertions(+), 10 deletions(-)
fs/dlm/dir.c +112 -3
@@ -224,6 +224,80 @@ static struct dlm_rsb *find_rsb_root(struct dlm_ls *ls, const char *name,return NULL ; } +struct dlm_dir_dump { +/* init values to match if whole +* dump fits to one seq. Sanity check only. +*/ +uint64_t seq_init; +uint64_t nodeid_init; +/* compare local pointer with last lookup, +* just a sanity check. +*/ +struct list_head *last; + +unsigned int sent_res;/* for log info */ +unsigned int sent_msg;/* for log info */ + +struct list_head list; +}; + +static void drop_dir_ctx (struct dlm_ls *ls,int nodeid) +{ +struct dlm_dir_dump *dd, *safe; + +write_lock (&ls->ls_dir_dump_lock); +list_for_each_entry_safe (dd, safe, &ls->ls_dir_dump_list, list) { +if (dd->nodeid_init == nodeid) { +log_error (ls,"drop dump seq %llu" , + (unsigned long long )dd->seq_init); +list_del (&dd->list); +kfree (dd); + } + } +write_unlock (&ls->ls_dir_dump_lock); +} + +static struct dlm_dir_dump *lookup_dir_dump (struct dlm_ls *ls,int nodeid) +{ +struct dlm_dir_dump *iter, *dd =NULL ; + +read_lock (&ls->ls_dir_dump_lock); +list_for_each_entry (iter, &ls->ls_dir_dump_list, list) { +if (iter->nodeid_init == nodeid) { + dd = iter; +break ; + } + } +read_unlock (&ls->ls_dir_dump_lock); + +return dd; +} + +static struct dlm_dir_dump *init_dir_dump (struct dlm_ls *ls,int nodeid) +{ +struct dlm_dir_dump *dd; + + dd =lookup_dir_dump (ls, nodeid); +if (dd) { +log_error (ls,"found ongoing dir dump for node %d, will drop it" , + nodeid); +drop_dir_ctx (ls, nodeid); + } + + dd =kzalloc (sizeof (*dd),GFP_ATOMIC ); +if (!dd) +return NULL ; + + dd->seq_init = ls->ls_recover_seq; + dd->nodeid_init = nodeid; + +write_lock (&ls->ls_dir_dump_lock); +list_add (&dd->list, &ls->ls_dir_dump_list); +write_unlock (&ls->ls_dir_dump_lock); + +return dd; +} +/* Find the rsb where we left off (or start again), then send rsb names for rsb's we're master of and whose directory node matches the requesting node. inbuf is the rsb name last sent, inlen is the name's length */
@@ -234,11 +308,20 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,struct list_head *list;struct dlm_rsb *r;int offset =0 , dir_nodeid; +struct dlm_dir_dump *dd;__be16 be_namelen;read_lock (&ls->ls_masters_lock);if (inlen >1 ) { + dd =lookup_dir_dump (ls, nodeid); +if (!dd) { +log_error (ls,"failed to lookup dir dump context nodeid: %d" , + nodeid); + goto out; + } + +/* next chunk in dump */ r =find_rsb_root (ls, inbuf, inlen);if (!r) { log_error(ls, "copy_master_names from %d start %d %.*s",
@@ -246,8 +329,25 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen, goto out; } list = r->res_masters_list.next; + +/* sanity checks */ +if (dd->last != &r->res_masters_list || + dd->seq_init != ls->ls_recover_seq) { +log_error (ls,"failed dir dump sanity check seq_init: %llu seq: %llu" , + (unsigned long long )dd->seq_init, + (unsigned long long )ls->ls_recover_seq); + goto out; + } }else { + dd =init_dir_dump (ls, nodeid); +if (!dd) { +log_error (ls,"failed to allocate dir dump context" ); + goto out; + } + +/* start dump */ list = ls->ls_masters_list.next; + dd->last = list; }for (offset =0 ; list != &ls->ls_masters_list; list = list->next) {
@@ -269,7 +369,7 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen, be_namelen =cpu_to_be16 (0 );memcpy (outbuf + offset, &be_namelen,sizeof (__be16)); offset +=sizeof (__be16); - ls->ls_recover_dir_sent_msg++; + dd->sent_msg++; goto out; }
@@ -278,7 +378,8 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen, offset +=sizeof (__be16);memcpy (outbuf + offset, r->res_name, r->res_length); offset += r->res_length; - ls->ls_recover_dir_sent_res++; + dd->sent_res++; + dd->last = list; }/*
@@ -288,10 +389,18 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,if ((list == &ls->ls_masters_list) && (offset +sizeof (uint16_t ) <= outlen)) { +/* end dump */ be_namelen =cpu_to_be16 (0xFFFF );memcpy (outbuf + offset, &be_namelen,sizeof (__be16)); offset +=sizeof (__be16); - ls->ls_recover_dir_sent_msg++; + dd->sent_msg++; +log_rinfo (ls,"dlm_recover_directory nodeid %d sent %u res out %u messages" , + nodeid, dd->sent_res, dd->sent_msg); + +write_lock (&ls->ls_dir_dump_lock); +list_del_init (&dd->list); +write_unlock (&ls->ls_dir_dump_lock); +kfree (dd); } out:read_unlock (&ls->ls_masters_lock);
fs/dlm/dlm_internal.h +2 -2
@@ -630,8 +630,6 @@ struct dlm_ls {struct mutex ls_requestqueue_mutex;struct dlm_rcom *ls_recover_buf;int ls_recover_nodeid;/* for debugging */ -unsigned int ls_recover_dir_sent_res;/* for log info */ -unsigned int ls_recover_dir_sent_msg;/* for log info */ unsigned int ls_recover_locks_in;/* for log info */ uint64_t ls_rcom_seq;spinlock_t ls_rcom_spin;
@@ -646,6 +644,8 @@ struct dlm_ls {struct list_head ls_masters_list;/* root resources */ rwlock_t ls_masters_lock;/* protect root_list */ +struct list_head ls_dir_dump_list;/* root resources */ +rwlock_t ls_dir_dump_lock;/* protect root_list */ const struct dlm_lockspace_ops *ls_ops;void *ls_ops_arg;
fs/dlm/lockspace.c +2 -0
@@ -582,6 +582,8 @@ static int new_lockspace(const char *name, const char *cluster,init_waitqueue_head (&ls->ls_wait_general);INIT_LIST_HEAD (&ls->ls_masters_list);rwlock_init (&ls->ls_masters_lock); +INIT_LIST_HEAD (&ls->ls_dir_dump_list); +rwlock_init (&ls->ls_dir_dump_lock);spin_lock (&lslist_lock); ls->ls_create_count =1 ;
fs/dlm/recoverd.c +0 -5
@@ -173,8 +173,6 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv) goto fail; } - ls->ls_recover_dir_sent_res =0 ; - ls->ls_recover_dir_sent_msg =0 ; ls->ls_recover_locks_in =0 ;dlm_set_recover_status (ls,DLM_RS_NODES );
@@ -211,9 +209,6 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)dlm_release_masters_list (ls); -log_rinfo (ls,"dlm_recover_directory %u out %u messages" , - ls->ls_recover_dir_sent_res, ls->ls_recover_dir_sent_msg); -/* * We may have outstanding operations that are waiting for a reply from * a failed node. Mark these to be resent after recovery. Unlock and
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 11/18 dlm: drop holding waiters mutex in waiters recovery
In reply to Alexander AringThis patch drops to hold the ls_waiters_mutex in dlm_recover_waiters_pre(). The dlm_recover_waiters_pre() function is only being called when recovery handling is being done for the specific lockspace. During this time there can't be new lock requests initiated or dlm messages being processed that manipulates the lockspace waiters list.
Only debugfs can access the lockspace waiters list when dlm_recover_waiters_pre() may manipulates it. This is not possible anymore because debugfs will hold the recovery lock when it's acccessing the waiters list.
A check was introduced in remove_from_waiters_ms() for local dlm messaging to check if really the lockspace is stopped and no new lock requests can be initiated.
--- fs/dlm/debug_fs.c | 4 ++++ fs/dlm/lock.c | 17 +++++++++-------- fs/dlm/lock.h | 1 + 3 files changed, 14 insertions(+), 8 deletions(-)
fs/dlm/debug_fs.c +4 -0
@@ -823,6 +823,7 @@ static ssize_t waiters_read(struct file *file, char __user *userbuf,size_t len =DLM_DEBUG_BUF_LEN , pos =0 , ret, rv;mutex_lock (&debug_buf_lock); +dlm_lock_recovery (ls);mutex_lock (&ls->ls_waiters_mutex);memset (debug_buf,0 ,sizeof (debug_buf));
@@ -835,6 +836,7 @@ static ssize_t waiters_read(struct file *file, char __user *userbuf, pos += ret; }mutex_unlock (&ls->ls_waiters_mutex); +dlm_unlock_recovery (ls); rv =simple_read_from_buffer (userbuf, count, ppos, debug_buf, pos);mutex_unlock (&debug_buf_lock);
@@ -858,7 +860,9 @@ static ssize_t waiters_write(struct file *file, const char __user *user_buf,if (n !=3 )return -EINVAL ; +dlm_lock_recovery (ls); error =dlm_debug_add_lkb_to_waiters (ls, lkb_id, mstype, to_nodeid); +dlm_unlock_recovery (ls);if (error)return error;
fs/dlm/lock.c +9 -8
@@ -201,7 +201,7 @@ void dlm_dump_rsb(struct dlm_rsb *r)/* Threads cannot use the lockspace while it's being recovered */ -static inline void dlm_lock_recovery (struct dlm_ls *ls) +void dlm_lock_recovery (struct dlm_ls *ls) {down_read (&ls->ls_in_recovery); }
@@ -1553,7 +1553,11 @@ static int remove_from_waiters(struct dlm_lkb *lkb, int mstype) }/* Handles situations where we might be processing a "fake" or "local" reply in -which we can't try to take waiters_mutex again. */ +* the recovery context which stops any locking activity. Only debugfs might +* change the lockspace waiters but they will held the recovery lock to ensure +* remove_from_waiters_ms() in local case will be the only user manipulating the +* lockspace waiters in recovery context. +*/ static int remove_from_waiters_ms (struct dlm_lkb *lkb,const struct dlm_message *ms,bool local)
@@ -1563,6 +1567,9 @@ static int remove_from_waiters_ms(struct dlm_lkb *lkb,if (!local)mutex_lock (&ls->ls_waiters_mutex); +else +WARN_ON_ONCE (!rwsem_is_locked (&ls->ls_in_recovery) || + !dlm_locking_stopped (ls)); error =_remove_from_waiters (lkb,le32_to_cpu (ms->m_type), ms);if (!local)mutex_unlock (&ls->ls_waiters_mutex);
@@ -4395,7 +4402,6 @@ static void _receive_convert_reply(struct dlm_lkb *lkb,if (error) goto out; -/* local reply can happen with waiters_mutex held */ error =remove_from_waiters_ms (lkb, ms, local);if (error) goto out;
@@ -4434,7 +4440,6 @@ static void _receive_unlock_reply(struct dlm_lkb *lkb,if (error) goto out; -/* local reply can happen with waiters_mutex held */ error =remove_from_waiters_ms (lkb, ms, local);if (error) goto out;
@@ -4486,7 +4491,6 @@ static void _receive_cancel_reply(struct dlm_lkb *lkb,if (error) goto out; -/* local reply can happen with waiters_mutex held */ error =remove_from_waiters_ms (lkb, ms, local);if (error) goto out;
@@ -4887,8 +4891,6 @@ void dlm_recover_waiters_pre(struct dlm_ls *ls)if (!ms_local)return ; -mutex_lock (&ls->ls_waiters_mutex); -list_for_each_entry_safe (lkb, safe, &ls->ls_waiters, lkb_wait_reply) { dir_nodeid =dlm_dir_nodeid (lkb->lkb_resource);
@@ -4981,7 +4983,6 @@ void dlm_recover_waiters_pre(struct dlm_ls *ls) }schedule (); } -mutex_unlock (&ls->ls_waiters_mutex);kfree (ms_local); }
fs/dlm/lock.h +1 -0
@@ -23,6 +23,7 @@ void dlm_hold_rsb(struct dlm_rsb *r);int dlm_put_lkb (struct dlm_lkb *lkb);void dlm_scan_rsbs (struct dlm_ls *ls);int dlm_lock_recovery_try (struct dlm_ls *ls); +void dlm_lock_recovery (struct dlm_ls *ls);void dlm_unlock_recovery (struct dlm_ls *ls); int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 12/18 dlm: convert ls_waiters_mutex to spinlock
In reply to Alexander AringThis patch converts the per dlm lockspace waiters lock from a mutex to a spinlock.
--- fs/dlm/debug_fs.c | 4 ++-- fs/dlm/dlm_internal.h | 2 +- fs/dlm/lock.c | 20 ++++++++++---------- fs/dlm/lockspace.c | 2 +- 4 files changed, 14 insertions(+), 14 deletions(-)
fs/dlm/debug_fs.c +2 -2
@@ -824,7 +824,7 @@ static ssize_t waiters_read(struct file *file, char __user *userbuf,mutex_lock (&debug_buf_lock);dlm_lock_recovery (ls); -mutex_lock (&ls->ls_waiters_mutex); +spin_lock (&ls->ls_waiters_lock);memset (debug_buf,0 ,sizeof (debug_buf));list_for_each_entry (lkb, &ls->ls_waiters, lkb_wait_reply) {
@@ -835,7 +835,7 @@ static ssize_t waiters_read(struct file *file, char __user *userbuf,break ; pos += ret; } -mutex_unlock (&ls->ls_waiters_mutex); +spin_unlock (&ls->ls_waiters_lock);dlm_unlock_recovery (ls); rv =simple_read_from_buffer (userbuf, count, ppos, debug_buf, pos);
fs/dlm/dlm_internal.h +1 -1
@@ -565,7 +565,7 @@ struct dlm_ls {struct dlm_rsbtable *ls_rsbtbl;uint32_t ls_rsbtbl_size; -struct mutex ls_waiters_mutex; +spinlock_t ls_waiters_lock;struct list_head ls_waiters;/* lkbs needing a reply */ struct mutex ls_orphans_mutex;
fs/dlm/lock.c +10 -10
@@ -1407,7 +1407,7 @@ static int add_to_waiters(struct dlm_lkb *lkb, int mstype, int to_nodeid)int error =0 ;int wc; -mutex_lock (&ls->ls_waiters_mutex); +spin_lock (&ls->ls_waiters_lock);if (is_overlap_unlock (lkb) || (is_overlap_cancel (lkb) && (mstype ==DLM_MSG_CANCEL ))) {
@@ -1447,7 +1447,7 @@ static int add_to_waiters(struct dlm_lkb *lkb, int mstype, int to_nodeid)log_error (ls,"addwait error %x %d flags %x %d %d %s" , lkb->lkb_id, error,dlm_iflags_val (lkb), mstype, lkb->lkb_wait_type, lkb->lkb_resource->res_name); -mutex_unlock (&ls->ls_waiters_mutex); +spin_unlock (&ls->ls_waiters_lock);return error; }
@@ -1546,9 +1546,9 @@ static int remove_from_waiters(struct dlm_lkb *lkb, int mstype)struct dlm_ls *ls = lkb->lkb_resource->res_ls;int error; -mutex_lock (&ls->ls_waiters_mutex); +spin_lock (&ls->ls_waiters_lock); error =_remove_from_waiters (lkb, mstype,NULL ); -mutex_unlock (&ls->ls_waiters_mutex); +spin_unlock (&ls->ls_waiters_lock);return error; }
@@ -1566,13 +1566,13 @@ static int remove_from_waiters_ms(struct dlm_lkb *lkb,int error;if (!local) -mutex_lock (&ls->ls_waiters_mutex); +spin_lock (&ls->ls_waiters_lock);else WARN_ON_ONCE (!rwsem_is_locked (&ls->ls_in_recovery) || !dlm_locking_stopped (ls)); error =_remove_from_waiters (lkb,le32_to_cpu (ms->m_type), ms);if (!local) -mutex_unlock (&ls->ls_waiters_mutex); +spin_unlock (&ls->ls_waiters_lock);return error; }
@@ -4990,7 +4990,7 @@ static struct dlm_lkb *find_resend_waiter(struct dlm_ls *ls) {struct dlm_lkb *lkb =NULL , *iter; -mutex_lock (&ls->ls_waiters_mutex); +spin_lock (&ls->ls_waiters_lock);list_for_each_entry (iter, &ls->ls_waiters, lkb_wait_reply) {if (test_bit (DLM_IFL_RESEND_BIT , &iter->lkb_iflags)) {hold_lkb (iter);
@@ -4998,7 +4998,7 @@ static struct dlm_lkb *find_resend_waiter(struct dlm_ls *ls)break ; } } -mutex_unlock (&ls->ls_waiters_mutex); +spin_unlock (&ls->ls_waiters_lock);return lkb; }
@@ -5062,14 +5062,14 @@ int dlm_recover_waiters_post(struct dlm_ls *ls)/* drop all wait_count references we still * hold a reference for this iteration. */ -mutex_lock (&ls->ls_waiters_mutex); +spin_lock (&ls->ls_waiters_lock);while (atomic_read (&lkb->lkb_wait_count)) {if (atomic_dec_and_test (&lkb->lkb_wait_count))list_del_init (&lkb->lkb_wait_reply);unhold_lkb (lkb); } -mutex_unlock (&ls->ls_waiters_mutex); +spin_unlock (&ls->ls_waiters_lock);if (oc || ou) {/* do an unlock or cancel instead of resending */
fs/dlm/lockspace.c +1 -1
@@ -515,7 +515,7 @@ static int new_lockspace(const char *name, const char *cluster,spin_lock_init (&ls->ls_lkbidr_spin);INIT_LIST_HEAD (&ls->ls_waiters); -mutex_init (&ls->ls_waiters_mutex); +spin_lock_init (&ls->ls_waiters_lock);INIT_LIST_HEAD (&ls->ls_orphans);mutex_init (&ls->ls_orphans_mutex);
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 13/18 dlm: convert res_lock to spinlock
In reply to Alexander AringThis patch converts the per dlm rsb res_lock from a mutex to a spinlock.
--- fs/dlm/dlm_internal.h | 2 +- fs/dlm/lock.c | 2 +- fs/dlm/lock.h | 4 ++-- 3 files changed, 4 insertions(+), 4 deletions(-)
fs/dlm/dlm_internal.h +1 -1
@@ -290,7 +290,7 @@ struct dlm_lkb {struct dlm_rsb {struct dlm_ls *res_ls;/* the lockspace */ struct kref res_ref; -struct mutex res_mutex; +spinlock_t res_lock;unsigned long res_flags;int res_length;/* length of rsb name */ int res_nodeid;
fs/dlm/lock.c +1 -1
@@ -415,7 +415,7 @@ static int get_rsb_struct(struct dlm_ls *ls, const void *name, int len, r->res_ls = ls; r->res_length = len;memcpy (r->res_name, name, len); -mutex_init (&r->res_mutex); +spin_lock_init (&r->res_lock);INIT_LIST_HEAD (&r->res_lookup);INIT_LIST_HEAD (&r->res_grantqueue);
fs/dlm/lock.h +2 -2
@@ -70,12 +70,12 @@ static inline int is_master(struct dlm_rsb *r)static inline void lock_rsb (struct dlm_rsb *r) { -mutex_lock (&r->res_mutex); +spin_lock (&r->res_lock); }static inline void unlock_rsb (struct dlm_rsb *r) { -mutex_unlock (&r->res_mutex); +spin_unlock (&r->res_lock); }#endif
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 14/18 dlm: make requestqueue handling non sleepable
In reply to Alexander AringThis patch removes the ls_requestqueue_wait and convert the ls_requestqueue_mutex to a rw lock. Instead of calling wait_event() in dlm processing which waits until all messages are processed and allow new message processing after recovering is done, this patch is using a bitflag to signal when a message should be saved for future or not. When recovery processes all saved messages we will clear this bit again and allow new messages to processed directly.
--- fs/dlm/dlm_internal.h | 5 ++--- fs/dlm/lock.c | 16 ++++++++++++++-- fs/dlm/lockspace.c | 4 +--- fs/dlm/member.c | 5 +++++ fs/dlm/requestqueue.c | 41 ++++++++--------------------------------- 5 files changed, 30 insertions(+), 41 deletions(-)
fs/dlm/dlm_internal.h +2 -3
@@ -625,9 +625,7 @@ struct dlm_ls {struct rw_semaphore ls_in_recovery;/* block local requests */ struct rw_semaphore ls_recv_active;/* block dlm_recv */ struct list_head ls_requestqueue;/* queue remote requests */ -atomic_t ls_requestqueue_cnt; -wait_queue_head_t ls_requestqueue_wait; -struct mutex ls_requestqueue_mutex; +rwlock_t ls_requestqueue_lock;struct dlm_rcom *ls_recover_buf;int ls_recover_nodeid;/* for debugging */ unsigned int ls_recover_locks_in;/* for log info */
@@ -687,6 +685,7 @@ struct dlm_ls {#define LSFL_UEVENT_WAIT 7#define LSFL_CB_DELAY 9#define LSFL_NODIR 10 +#define LSFL_RECV_MSG_BLOCKED 11/* much of this is just saving user space pointers associated with the lock that we pass back to the user lib with an ast */
fs/dlm/lock.c +14 -2
@@ -4749,20 +4749,32 @@ static void _receive_message(struct dlm_ls *ls, const struct dlm_message *ms,static void dlm_receive_message (struct dlm_ls *ls,const struct dlm_message *ms,int nodeid) { -if (dlm_locking_stopped (ls)) { +try_again: +read_lock (&ls->ls_requestqueue_lock); +if (test_bit (LSFL_RECV_MSG_BLOCKED , &ls->ls_flags)) {/* If we were a member of this lockspace, left, and rejoined, other nodes may still be sending us messages from the lockspace generation before we left. */ if (WARN_ON_ONCE (!ls->ls_generation)) { +read_unlock (&ls->ls_requestqueue_lock);log_limit (ls,"receive %d from %d ignore old gen" ,le32_to_cpu (ms->m_type), nodeid);return ; } +read_unlock (&ls->ls_requestqueue_lock); +write_lock (&ls->ls_requestqueue_lock); +/* recheck because we hold writelock now */ +if (!test_bit (LSFL_RECV_MSG_BLOCKED , &ls->ls_flags)) { +write_unlock_bh (&ls->ls_requestqueue_lock); + goto try_again; + } +dlm_add_requestqueue (ls, nodeid, ms); +write_unlock (&ls->ls_requestqueue_lock); }else { -dlm_wait_requestqueue (ls);_receive_message (ls, ms,0 ); +read_unlock (&ls->ls_requestqueue_lock); } }
fs/dlm/lockspace.c +1 -3
@@ -554,9 +554,7 @@ static int new_lockspace(const char *name, const char *cluster,init_rwsem (&ls->ls_in_recovery);init_rwsem (&ls->ls_recv_active);INIT_LIST_HEAD (&ls->ls_requestqueue); -atomic_set (&ls->ls_requestqueue_cnt,0 ); -init_waitqueue_head (&ls->ls_requestqueue_wait); -mutex_init (&ls->ls_requestqueue_mutex); +rwlock_init (&ls->ls_requestqueue_lock);spin_lock_init (&ls->ls_clear_proc_locks);/* Due backwards compatibility with 3.1 we need to use maximum
fs/dlm/member.c +5 -0
@@ -642,6 +642,11 @@ int dlm_ls_stop(struct dlm_ls *ls)set_bit (LSFL_RECOVER_STOP , &ls->ls_flags); new =test_and_clear_bit (LSFL_RUNNING , &ls->ls_flags); ls->ls_recover_seq++; + +/* activate requestqueue and stop processing */ +write_lock (&ls->ls_requestqueue_lock); +set_bit (LSFL_RECV_MSG_BLOCKED , &ls->ls_flags); +write_unlock (&ls->ls_requestqueue_lock);spin_unlock (&ls->ls_recover_lock);/*
fs/dlm/requestqueue.c +8 -33
92 lines not on this page · raw message
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 15/18 dlm: ls_recv_active semaphore to rwlock
In reply to Alexander AringThis patch converts the ls_recv_active semaphore to a rwlock to not sleep during dlm message processing.
--- fs/dlm/dlm_internal.h | 2 +- fs/dlm/lock.c | 4 ++-- fs/dlm/lockspace.c | 2 +- fs/dlm/member.c | 4 ++-- fs/dlm/recoverd.c | 4 ++-- 5 files changed, 8 insertions(+), 8 deletions(-)
fs/dlm/dlm_internal.h +1 -1
@@ -623,7 +623,7 @@ struct dlm_ls {uint64_t ls_recover_seq;struct dlm_recover *ls_recover_args;struct rw_semaphore ls_in_recovery;/* block local requests */ -struct rw_semaphore ls_recv_active;/* block dlm_recv */ +rwlock_t ls_recv_active;/* block dlm_recv */ struct list_head ls_requestqueue;/* queue remote requests */ rwlock_t ls_requestqueue_lock;struct dlm_rcom *ls_recover_buf;
fs/dlm/lock.c +2 -2
@@ -4834,7 +4834,7 @@ void dlm_receive_buffer(const union dlm_packet *p, int nodeid)/* this rwsem allows dlm_ls_stop() to wait for all dlm_recv threads to be inactive (in this ls) before transitioning to recovery mode */ -down_read (&ls->ls_recv_active); +read_lock (&ls->ls_recv_active);if (hd->h_cmd ==DLM_MSG )dlm_receive_message (ls, &p->message, nodeid);else if (hd->h_cmd ==DLM_RCOM )
@@ -4842,7 +4842,7 @@ void dlm_receive_buffer(const union dlm_packet *p, int nodeid)else log_error (ls,"invalid h_cmd %d from %d lockspace %x" , hd->h_cmd, nodeid,le32_to_cpu (hd->u.h_lockspace)); -up_read (&ls->ls_recv_active); +read_unlock (&ls->ls_recv_active);dlm_put_lockspace (ls); }
fs/dlm/lockspace.c +1 -1
@@ -552,7 +552,7 @@ static int new_lockspace(const char *name, const char *cluster, ls->ls_recover_seq =get_random_u64 (); ls->ls_recover_args =NULL ;init_rwsem (&ls->ls_in_recovery); -init_rwsem (&ls->ls_recv_active); +rwlock_init (&ls->ls_recv_active);INIT_LIST_HEAD (&ls->ls_requestqueue);rwlock_init (&ls->ls_requestqueue_lock);spin_lock_init (&ls->ls_clear_proc_locks);
fs/dlm/member.c +2 -2
@@ -630,7 +630,7 @@ int dlm_ls_stop(struct dlm_ls *ls)* message to the requestqueue without races. */ -down_write (&ls->ls_recv_active); +write_lock (&ls->ls_recv_active);/* * Abort any recovery that's in progress (see RECOVER_STOP,
@@ -654,7 +654,7 @@ int dlm_ls_stop(struct dlm_ls *ls)* requestqueue for later. */ -up_write (&ls->ls_recv_active); +write_unlock (&ls->ls_recv_active);/* * This in_recovery lock does two things:
fs/dlm/recoverd.c +2 -2
@@ -103,7 +103,7 @@ static int enable_locking(struct dlm_ls *ls, uint64_t seq) {int error = -EINTR ; -down_write (&ls->ls_recv_active); +write_lock (&ls->ls_recv_active);spin_lock (&ls->ls_recover_lock);if (ls->ls_recover_seq == seq) {
@@ -115,7 +115,7 @@ static int enable_locking(struct dlm_ls *ls, uint64_t seq) }spin_unlock (&ls->ls_recover_lock); -up_write (&ls->ls_recv_active); +write_unlock (&ls->ls_recv_active);return error; }
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 16/18 dlm: remove schedule in dlm receive path
In reply to Alexander AringThis patch removes an explicit schedule() call in the receive path of dlm message processing. The goal in DLM is to not trigger any additional scheduling while processing DLM messages. This schedule() is directly called inside the dlm message processing path. As soon we handle the dlm messaging processing in softirq context we cannot call schedule() in this context anymore. This patch prepares for this transition by simple removing this schedule() call.
--- fs/dlm/lock.c | 1 - 1 file changed, 1 deletion(-)
fs/dlm/lock.c +0 -1
@@ -2540,7 +2540,6 @@ static void process_lookup_list(struct dlm_rsb *r)list_for_each_entry_safe (lkb, safe, &r->res_lookup, lkb_rsb_lookup) {list_del_init (&lkb->lkb_rsb_lookup);_request_lock (r, lkb); -schedule (); } }
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 17/18 dlm: convert message parsing locks to disable bh
In reply to Alexander AringThis patch converts all spinlocks involved in message parsing to it's _bh version. The reason to do that is to convert the message parsing into softirq context and we need to prevent that those locks can be interrupted by a softirq if those are held.
--- fs/dlm/ast.c | 28 +++--- fs/dlm/debug_fs.c | 32 +++---- fs/dlm/dir.c | 24 ++--- fs/dlm/lock.c | 204 ++++++++++++++++++++++++------------------ fs/dlm/lock.h | 4 +- fs/dlm/lockspace.c | 51 ++++++----- fs/dlm/lowcomms.c | 16 ++-- fs/dlm/member.c | 22 ++--- fs/dlm/midcomms.c | 40 ++++----- fs/dlm/rcom.c | 26 +++--- fs/dlm/recover.c | 52 +++++------ fs/dlm/recoverd.c | 20 ++--- fs/dlm/requestqueue.c | 12 +-- fs/dlm/user.c | 34 +++---- 14 files changed, 296 insertions(+), 269 deletions(-)
fs/dlm/ast.c +14 -14
@@ -127,19 +127,19 @@ void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,return ; } -spin_lock (&lkb->lkb_cb_lock); +spin_lock_bh (&lkb->lkb_cb_lock); rv =dlm_enqueue_lkb_callback (lkb, flags, mode, status, sbflags);switch (rv) {case DLM_ENQUEUE_CALLBACK_NEED_SCHED :kref_get (&lkb->lkb_ref); -spin_lock (&ls->ls_cb_lock); +spin_lock_bh (&ls->ls_cb_lock);if (test_bit (LSFL_CB_DELAY , &ls->ls_flags)) {list_add (&lkb->lkb_cb_list, &ls->ls_cb_delay); }else {queue_work (ls->ls_callback_wq, &lkb->lkb_cb_work); } -spin_unlock (&ls->ls_cb_lock); +spin_unlock_bh (&ls->ls_cb_lock);break ;case DLM_ENQUEUE_CALLBACK_FAILURE :WARN_ON_ONCE (1 );
@@ -150,7 +150,7 @@ void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,WARN_ON_ONCE (1 );break ; } -spin_unlock (&lkb->lkb_cb_lock); +spin_unlock_bh (&lkb->lkb_cb_lock); }void dlm_callback_work (struct work_struct *work)
@@ -162,14 +162,14 @@ void dlm_callback_work(struct work_struct *work)struct dlm_callback *cb;int rv; -spin_lock (&lkb->lkb_cb_lock); +spin_lock_bh (&lkb->lkb_cb_lock); rv =dlm_dequeue_lkb_callback (lkb, &cb);if (WARN_ON_ONCE (rv ==DLM_DEQUEUE_CALLBACK_EMPTY )) {clear_bit (DLM_IFL_CB_PENDING_BIT , &lkb->lkb_iflags); -spin_unlock (&lkb->lkb_cb_lock); +spin_unlock_bh (&lkb->lkb_cb_lock); goto out; } -spin_unlock (&lkb->lkb_cb_lock); +spin_unlock_bh (&lkb->lkb_cb_lock);for (;;) { castfn = lkb->lkb_astfn;
@@ -190,14 +190,14 @@ void dlm_callback_work(struct work_struct *work)kref_put (&cb->ref, dlm_release_callback); -spin_lock (&lkb->lkb_cb_lock); +spin_lock_bh (&lkb->lkb_cb_lock); rv =dlm_dequeue_lkb_callback (lkb, &cb);if (rv ==DLM_DEQUEUE_CALLBACK_EMPTY ) {clear_bit (DLM_IFL_CB_PENDING_BIT , &lkb->lkb_iflags); -spin_unlock (&lkb->lkb_cb_lock); +spin_unlock_bh (&lkb->lkb_cb_lock);break ; } -spin_unlock (&lkb->lkb_cb_lock); +spin_unlock_bh (&lkb->lkb_cb_lock); } out:
@@ -225,9 +225,9 @@ void dlm_callback_stop(struct dlm_ls *ls)void dlm_callback_suspend (struct dlm_ls *ls) {if (ls->ls_callback_wq) { -spin_lock (&ls->ls_cb_lock); +spin_lock_bh (&ls->ls_cb_lock);set_bit (LSFL_CB_DELAY , &ls->ls_flags); -spin_unlock (&ls->ls_cb_lock); +spin_unlock_bh (&ls->ls_cb_lock);flush_workqueue (ls->ls_callback_wq); }
@@ -245,7 +245,7 @@ void dlm_callback_resume(struct dlm_ls *ls)return ; more: -spin_lock (&ls->ls_cb_lock); +spin_lock_bh (&ls->ls_cb_lock);list_for_each_entry_safe (lkb, safe, &ls->ls_cb_delay, lkb_cb_list) {list_del_init (&lkb->lkb_cb_list);queue_work (ls->ls_callback_wq, &lkb->lkb_cb_work);
@@ -256,7 +256,7 @@ void dlm_callback_resume(struct dlm_ls *ls) empty =list_empty (&ls->ls_cb_delay);if (empty)clear_bit (LSFL_CB_DELAY , &ls->ls_flags); -spin_unlock (&ls->ls_cb_lock); +spin_unlock_bh (&ls->ls_cb_lock); sum += count;if (!empty) {
fs/dlm/debug_fs.c +16 -16
@@ -372,7 +372,7 @@ static void print_format5_lock(struct seq_file *s, struct dlm_lkb *lkb)/* lkb_id lkb_flags mode flags sb_status sb_flags */ -spin_lock (&lkb->lkb_cb_lock); +spin_lock_bh (&lkb->lkb_cb_lock);list_for_each_entry (cb, &lkb->lkb_callbacks, list) { seq_printf(s, "%x %x %d %x %d %x\n", lkb->lkb_id,
@@ -382,7 +382,7 @@ static void print_format5_lock(struct seq_file *s, struct dlm_lkb *lkb) cb->sb_status, cb->sb_flags); } -spin_unlock (&lkb->lkb_cb_lock); +spin_unlock_bh (&lkb->lkb_cb_lock); }static void print_format5 (struct dlm_rsb *r,struct seq_file *s)
@@ -508,7 +508,7 @@ static void *table_seq_start(struct seq_file *seq, loff_t *pos) tree = toss ? &ls->ls_rsbtbl[bucket].toss : &ls->ls_rsbtbl[bucket].keep; -spin_lock (&ls->ls_rsbtbl[bucket].lock); +spin_lock_bh (&ls->ls_rsbtbl[bucket].lock);if (!RB_EMPTY_ROOT (tree)) {for (node =rb_first (tree); node; node =rb_next (node)) { r = rb_entry(node, struct dlm_rsb, res_hashnode);
@@ -516,12 +516,12 @@ static void *table_seq_start(struct seq_file *seq, loff_t *pos)dlm_hold_rsb (r); ri->rsb = r; ri->bucket = bucket; -spin_unlock (&ls->ls_rsbtbl[bucket].lock); +spin_unlock_bh (&ls->ls_rsbtbl[bucket].lock);return ri; } } } -spin_unlock (&ls->ls_rsbtbl[bucket].lock); +spin_unlock_bh (&ls->ls_rsbtbl[bucket].lock);/* * move to the first rsb in the next non-empty bucket
@@ -540,18 +540,18 @@ static void *table_seq_start(struct seq_file *seq, loff_t *pos) } tree = toss ? &ls->ls_rsbtbl[bucket].toss : &ls->ls_rsbtbl[bucket].keep; -spin_lock (&ls->ls_rsbtbl[bucket].lock); +spin_lock_bh (&ls->ls_rsbtbl[bucket].lock);if (!RB_EMPTY_ROOT (tree)) { node =rb_first (tree); r = rb_entry(node, struct dlm_rsb, res_hashnode);dlm_hold_rsb (r); ri->rsb = r; ri->bucket = bucket; -spin_unlock (&ls->ls_rsbtbl[bucket].lock); +spin_unlock_bh (&ls->ls_rsbtbl[bucket].lock); *pos = n;return ri; } -spin_unlock (&ls->ls_rsbtbl[bucket].lock); +spin_unlock_bh (&ls->ls_rsbtbl[bucket].lock); } }
@@ -572,7 +572,7 @@ static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos)* move to the next rsb in the same bucket */ -spin_lock (&ls->ls_rsbtbl[bucket].lock); +spin_lock_bh (&ls->ls_rsbtbl[bucket].lock); rp = ri->rsb; next =rb_next (&rp->res_hashnode);
@@ -580,12 +580,12 @@ static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos) r = rb_entry(next, struct dlm_rsb, res_hashnode);dlm_hold_rsb (r); ri->rsb = r; -spin_unlock (&ls->ls_rsbtbl[bucket].lock); +spin_unlock_bh (&ls->ls_rsbtbl[bucket].lock);dlm_put_rsb (rp); ++*pos;return ri; } -spin_unlock (&ls->ls_rsbtbl[bucket].lock); +spin_unlock_bh (&ls->ls_rsbtbl[bucket].lock);dlm_put_rsb (rp);/*
@@ -606,18 +606,18 @@ static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos) } tree = toss ? &ls->ls_rsbtbl[bucket].toss : &ls->ls_rsbtbl[bucket].keep; -spin_lock (&ls->ls_rsbtbl[bucket].lock); +spin_lock_bh (&ls->ls_rsbtbl[bucket].lock);if (!RB_EMPTY_ROOT (tree)) { next =rb_first (tree); r = rb_entry(next, struct dlm_rsb, res_hashnode);dlm_hold_rsb (r); ri->rsb = r; ri->bucket = bucket; -spin_unlock (&ls->ls_rsbtbl[bucket].lock); +spin_unlock_bh (&ls->ls_rsbtbl[bucket].lock); *pos = n;return ri; } -spin_unlock (&ls->ls_rsbtbl[bucket].lock); +spin_unlock_bh (&ls->ls_rsbtbl[bucket].lock); } }
@@ -824,7 +824,7 @@ static ssize_t waiters_read(struct file *file, char __user *userbuf,mutex_lock (&debug_buf_lock);dlm_lock_recovery (ls); -spin_lock (&ls->ls_waiters_lock); +spin_lock_bh (&ls->ls_waiters_lock);memset (debug_buf,0 ,sizeof (debug_buf));list_for_each_entry (lkb, &ls->ls_waiters, lkb_wait_reply) {
@@ -835,7 +835,7 @@ static ssize_t waiters_read(struct file *file, char __user *userbuf,break ; pos += ret; } -spin_unlock (&ls->ls_waiters_lock); +spin_unlock_bh (&ls->ls_waiters_lock);dlm_unlock_recovery (ls); rv =simple_read_from_buffer (userbuf, count, ppos, debug_buf, pos);
fs/dlm/dir.c +12 -12
@@ -204,12 +204,12 @@ static struct dlm_rsb *find_rsb_root(struct dlm_ls *ls, const char *name, hash =jhash (name, len,0 ); bucket = hash & (ls->ls_rsbtbl_size -1 ); -spin_lock (&ls->ls_rsbtbl[bucket].lock); +spin_lock_bh (&ls->ls_rsbtbl[bucket].lock); rv =dlm_search_rsb_tree (&ls->ls_rsbtbl[bucket].keep, name, len, &r);if (rv) rv =dlm_search_rsb_tree (&ls->ls_rsbtbl[bucket].toss, name, len, &r); -spin_unlock (&ls->ls_rsbtbl[bucket].lock); +spin_unlock_bh (&ls->ls_rsbtbl[bucket].lock);if (!rv)return r;
@@ -245,7 +245,7 @@ static void drop_dir_ctx(struct dlm_ls *ls, int nodeid) {struct dlm_dir_dump *dd, *safe; -write_lock (&ls->ls_dir_dump_lock); +write_lock_bh (&ls->ls_dir_dump_lock);list_for_each_entry_safe (dd, safe, &ls->ls_dir_dump_list, list) {if (dd->nodeid_init == nodeid) { log_error(ls, "drop dump seq %llu",
@@ -254,21 +254,21 @@ static void drop_dir_ctx(struct dlm_ls *ls, int nodeid)kfree (dd); } } -write_unlock (&ls->ls_dir_dump_lock); +write_unlock_bh (&ls->ls_dir_dump_lock); }static struct dlm_dir_dump *lookup_dir_dump (struct dlm_ls *ls,int nodeid) {struct dlm_dir_dump *iter, *dd =NULL ; -read_lock (&ls->ls_dir_dump_lock); +read_lock_bh (&ls->ls_dir_dump_lock);list_for_each_entry (iter, &ls->ls_dir_dump_list, list) {if (iter->nodeid_init == nodeid) { dd = iter;break ; } } -read_unlock (&ls->ls_dir_dump_lock); +read_unlock_bh (&ls->ls_dir_dump_lock);return dd; }
@@ -291,9 +291,9 @@ static struct dlm_dir_dump *init_dir_dump(struct dlm_ls *ls, int nodeid) dd->seq_init = ls->ls_recover_seq; dd->nodeid_init = nodeid; -write_lock (&ls->ls_dir_dump_lock); +write_lock_bh (&ls->ls_dir_dump_lock);list_add (&dd->list, &ls->ls_dir_dump_list); -write_unlock (&ls->ls_dir_dump_lock); +write_unlock_bh (&ls->ls_dir_dump_lock);return dd; }
@@ -311,7 +311,7 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,struct dlm_dir_dump *dd;__be16 be_namelen; -read_lock (&ls->ls_masters_lock); +read_lock_bh (&ls->ls_masters_lock);if (inlen >1 ) { dd =lookup_dir_dump (ls, nodeid);
@@ -397,12 +397,12 @@ void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,log_rinfo (ls,"dlm_recover_directory nodeid %d sent %u res out %u messages" , nodeid, dd->sent_res, dd->sent_msg); -write_lock (&ls->ls_dir_dump_lock); +write_lock_bh (&ls->ls_dir_dump_lock);list_del_init (&dd->list); -write_unlock (&ls->ls_dir_dump_lock); +write_unlock_bh (&ls->ls_dir_dump_lock);kfree (dd); } out: -read_unlock (&ls->ls_masters_lock); +read_unlock_bh (&ls->ls_masters_lock); }
fs/dlm/lock.c +116 -88
642 lines not on this page · raw message
fs/dlm/lock.h +2 -2
@@ -70,12 +70,12 @@ static inline int is_master(struct dlm_rsb *r)static inline void lock_rsb (struct dlm_rsb *r) { -spin_lock (&r->res_lock); +spin_lock_bh (&r->res_lock); }static inline void unlock_rsb (struct dlm_rsb *r) { -spin_unlock (&r->res_lock); +spin_unlock_bh (&r->res_lock); }#endif
fs/dlm/lockspace.c +25 -26
178 lines not on this page · raw message
fs/dlm/lowcomms.c +8 -8
@@ -867,36 +867,36 @@ static void process_dlm_messages(struct work_struct *work) {struct processqueue_entry *pentry; -spin_lock (&processqueue_lock); +spin_lock_bh (&processqueue_lock); pentry =list_first_entry_or_null (&processqueue, struct processqueue_entry, list);if (WARN_ON_ONCE (!pentry)) { process_dlm_messages_pending = false; -spin_unlock (&processqueue_lock); +spin_unlock_bh (&processqueue_lock);return ; }list_del (&pentry->list);atomic_dec (&processqueue_count); -spin_unlock (&processqueue_lock); +spin_unlock_bh (&processqueue_lock);for (;;) {dlm_process_incoming_buffer (pentry->nodeid, pentry->buf, pentry->buflen);free_processqueue_entry (pentry); -spin_lock (&processqueue_lock); +spin_lock_bh (&processqueue_lock); pentry =list_first_entry_or_null (&processqueue, struct processqueue_entry, list);if (!pentry) { process_dlm_messages_pending = false; -spin_unlock (&processqueue_lock); +spin_unlock_bh (&processqueue_lock);break ; }list_del (&pentry->list);atomic_dec (&processqueue_count); -spin_unlock (&processqueue_lock); +spin_unlock_bh (&processqueue_lock); } }
@@ -966,14 +966,14 @@ static int receive_from_sock(struct connection *con, int buflen)memmove (con->rx_leftover_buf, pentry->buf + ret, con->rx_leftover); -spin_lock (&processqueue_lock); +spin_lock_bh (&processqueue_lock); ret =atomic_inc_return (&processqueue_count);list_add_tail (&pentry->list, &processqueue);if (!process_dlm_messages_pending) { process_dlm_messages_pending = true;queue_work (process_workqueue, &process_work); } -spin_unlock (&processqueue_lock); +spin_unlock_bh (&processqueue_lock);if (ret >DLM_MAX_PROCESS_BUFFERS )return DLM_IO_FLUSH ;
fs/dlm/member.c +11 -11
@@ -630,7 +630,7 @@ int dlm_ls_stop(struct dlm_ls *ls)* message to the requestqueue without races. */ -write_lock (&ls->ls_recv_active); +write_lock_bh (&ls->ls_recv_active);/* * Abort any recovery that's in progress (see RECOVER_STOP,
@@ -638,23 +638,23 @@ int dlm_ls_stop(struct dlm_ls *ls)* dlm to quit any processing (see RUNNING, dlm_locking_stopped()). */ -spin_lock (&ls->ls_recover_lock); +spin_lock_bh (&ls->ls_recover_lock);set_bit (LSFL_RECOVER_STOP , &ls->ls_flags); new =test_and_clear_bit (LSFL_RUNNING , &ls->ls_flags); ls->ls_recover_seq++;/* activate requestqueue and stop processing */ -write_lock (&ls->ls_requestqueue_lock); +write_lock_bh (&ls->ls_requestqueue_lock);set_bit (LSFL_RECV_MSG_BLOCKED , &ls->ls_flags); -write_unlock (&ls->ls_requestqueue_lock); -spin_unlock (&ls->ls_recover_lock); +write_unlock_bh (&ls->ls_requestqueue_lock); +spin_unlock_bh (&ls->ls_recover_lock);/* * Let dlm_recv run again, now any normal messages will be saved on the * requestqueue for later. */ -write_unlock (&ls->ls_recv_active); +write_unlock_bh (&ls->ls_recv_active);/* * This in_recovery lock does two things:
@@ -679,13 +679,13 @@ int dlm_ls_stop(struct dlm_ls *ls)dlm_recoverd_suspend (ls); -spin_lock (&ls->ls_recover_lock); +spin_lock_bh (&ls->ls_recover_lock);kfree (ls->ls_slots); ls->ls_slots =NULL ; ls->ls_num_slots =0 ; ls->ls_slots_size =0 ; ls->ls_recover_status =0 ; -spin_unlock (&ls->ls_recover_lock); +spin_unlock_bh (&ls->ls_recover_lock);dlm_recoverd_resume (ls);
@@ -719,12 +719,12 @@ int dlm_ls_start(struct dlm_ls *ls)if (error <0 ) goto fail_rv; -spin_lock (&ls->ls_recover_lock); +spin_lock_bh (&ls->ls_recover_lock);/* the lockspace needs to be stopped before it can be started */ if (!dlm_locking_stopped (ls)) { -spin_unlock (&ls->ls_recover_lock); +spin_unlock_bh (&ls->ls_recover_lock);log_error (ls,"start ignored: lockspace running" ); error = -EINVAL ; goto fail;
@@ -735,7 +735,7 @@ int dlm_ls_start(struct dlm_ls *ls) rv->seq = ++ls->ls_recover_seq; rv_old = ls->ls_recover_args; ls->ls_recover_args = rv; -spin_unlock (&ls->ls_recover_lock); +spin_unlock_bh (&ls->ls_recover_lock);if (rv_old) { log_error(ls, "unused recovery %llx %d",
fs/dlm/midcomms.c +20 -20
@@ -364,9 +364,9 @@ int dlm_midcomms_addr(int nodeid, struct sockaddr_storage *addr, int len) node->users =0 ;midcomms_node_reset (node); -spin_lock (&nodes_lock); +spin_lock_bh (&nodes_lock);hlist_add_head_rcu (&node->hlist, &node_hash[r]); -spin_unlock (&nodes_lock); +spin_unlock_bh (&nodes_lock); node->debugfs =dlm_create_debug_comms_file (nodeid, node);return 0 ;
@@ -477,7 +477,7 @@ static void dlm_receive_ack(struct midcomms_node *node, uint32_t seq)static void dlm_pas_fin_ack_rcv (struct midcomms_node *node) { -spin_lock (&node->state_lock); +spin_lock_bh (&node->state_lock);pr_debug ("receive passive fin ack from node %d with state %s\n" , node->nodeid,dlm_state_str (node->state));
@@ -491,13 +491,13 @@ static void dlm_pas_fin_ack_rcv(struct midcomms_node *node)wake_up (&node->shutdown_wait);break ;default : -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock);log_print ("%s: unexpected state: %d" , __func__, node->state);WARN_ON_ONCE (1 );return ; } -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock); } static void dlm_receive_buffer_3_2_trace(uint32_t seq,
@@ -534,7 +534,7 @@ static void dlm_midcomms_receive_buffer(const union dlm_packet *p,if (is_expected_seq) {switch (p->header.h_cmd) {case DLM_FIN : -spin_lock (&node->state_lock); +spin_lock_bh (&node->state_lock);pr_debug ("receive fin msg from node %d with state %s\n" , node->nodeid,dlm_state_str (node->state));
@@ -575,13 +575,13 @@ static void dlm_midcomms_receive_buffer(const union dlm_packet *p,/* probably remove_member caught it, do nothing */ break ;default : -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock);log_print ("%s: unexpected state: %d" , __func__, node->state);WARN_ON_ONCE (1 );return ; } -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock);break ;default :WARN_ON_ONCE (test_bit (DLM_NODE_FLAG_STOP_RX , &node->flags));
@@ -1182,7 +1182,7 @@ void dlm_midcomms_exit(void)static void dlm_act_fin_ack_rcv (struct midcomms_node *node) { -spin_lock (&node->state_lock); +spin_lock_bh (&node->state_lock);pr_debug ("receive active fin ack from node %d with state %s\n" , node->nodeid,dlm_state_str (node->state));
@@ -1202,13 +1202,13 @@ static void dlm_act_fin_ack_rcv(struct midcomms_node *node)wake_up (&node->shutdown_wait);break ;default : -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock);log_print ("%s: unexpected state: %d" , __func__, node->state);WARN_ON_ONCE (1 );return ; } -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock); }void dlm_midcomms_add_member (int nodeid)
@@ -1223,7 +1223,7 @@ void dlm_midcomms_add_member(int nodeid)return ; } -spin_lock (&node->state_lock); +spin_lock_bh (&node->state_lock);if (!node->users) {pr_debug ("receive add member from node %d with state %s\n" , node->nodeid,dlm_state_str (node->state));
@@ -1251,7 +1251,7 @@ void dlm_midcomms_add_member(int nodeid) node->users++;pr_debug ("node %d users inc count %d\n" , nodeid, node->users); -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock);srcu_read_unlock (&nodes_srcu, idx); }
@@ -1269,13 +1269,13 @@ void dlm_midcomms_remove_member(int nodeid)return ; } -spin_lock (&node->state_lock); +spin_lock_bh (&node->state_lock);/* case of dlm_midcomms_addr() created node but * was not added before because dlm_midcomms_close() * removed the node */ if (!node->users) { -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock);srcu_read_unlock (&nodes_srcu, idx);return ; }
@@ -1313,7 +1313,7 @@ void dlm_midcomms_remove_member(int nodeid)break ; } } -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock);srcu_read_unlock (&nodes_srcu, idx); }
@@ -1351,7 +1351,7 @@ static void midcomms_shutdown(struct midcomms_node *node)return ; } -spin_lock (&node->state_lock); +spin_lock_bh (&node->state_lock);pr_debug ("receive active shutdown for node %d with state %s\n" , node->nodeid,dlm_state_str (node->state));switch (node->state) {
@@ -1370,7 +1370,7 @@ static void midcomms_shutdown(struct midcomms_node *node)*/ break ; } -spin_unlock (&node->state_lock); +spin_unlock_bh (&node->state_lock);if (DLM_DEBUG_FENCE_TERMINATION )msleep (5000 );
@@ -1441,9 +1441,9 @@ int dlm_midcomms_close(int nodeid) ret =dlm_lowcomms_close (nodeid);dlm_delete_debug_comms_file (node->debugfs); -spin_lock (&nodes_lock); +spin_lock_bh (&nodes_lock);hlist_del_rcu (&node->hlist); -spin_unlock (&nodes_lock); +spin_unlock_bh (&nodes_lock);srcu_read_unlock (&nodes_srcu, idx);/* wait that all readers left until flush send queue */
fs/dlm/rcom.c +13 -13
@@ -143,18 +143,18 @@ static int check_rcom_config(struct dlm_ls *ls, struct dlm_rcom *rc, int nodeid)static void allow_sync_reply (struct dlm_ls *ls,__le64 *new_seq) { -spin_lock (&ls->ls_rcom_spin); +spin_lock_bh (&ls->ls_rcom_spin); *new_seq =cpu_to_le64 (++ls->ls_rcom_seq);set_bit (LSFL_RCOM_WAIT , &ls->ls_flags); -spin_unlock (&ls->ls_rcom_spin); +spin_unlock_bh (&ls->ls_rcom_spin); }static void disallow_sync_reply (struct dlm_ls *ls) { -spin_lock (&ls->ls_rcom_spin); +spin_lock_bh (&ls->ls_rcom_spin);clear_bit (LSFL_RCOM_WAIT , &ls->ls_flags);clear_bit (LSFL_RCOM_READY , &ls->ls_flags); -spin_unlock (&ls->ls_rcom_spin); +spin_unlock_bh (&ls->ls_rcom_spin); }/*
@@ -245,10 +245,10 @@ static void receive_rcom_status(struct dlm_ls *ls, goto do_create; } -spin_lock (&ls->ls_recover_lock); +spin_lock_bh (&ls->ls_recover_lock); status = ls->ls_recover_status; num_slots = ls->ls_num_slots; -spin_unlock (&ls->ls_recover_lock); +spin_unlock_bh (&ls->ls_recover_lock); len += num_slots *sizeof (struct rcom_slot ); do_create:
@@ -266,9 +266,9 @@ static void receive_rcom_status(struct dlm_ls *ls,if (!num_slots) goto do_send; -spin_lock (&ls->ls_recover_lock); +spin_lock_bh (&ls->ls_recover_lock);if (ls->ls_num_slots != num_slots) { -spin_unlock (&ls->ls_recover_lock); +spin_unlock_bh (&ls->ls_recover_lock);log_debug (ls,"receive_rcom_status num_slots %d to %d" , num_slots, ls->ls_num_slots); rc->rc_result =0 ;
@@ -277,7 +277,7 @@ static void receive_rcom_status(struct dlm_ls *ls, }dlm_slots_copy_out (ls, rc); -spin_unlock (&ls->ls_recover_lock); +spin_unlock_bh (&ls->ls_recover_lock); do_send:send_rcom_stateless (msg, rc);
@@ -285,7 +285,7 @@ static void receive_rcom_status(struct dlm_ls *ls,static void receive_sync_reply (struct dlm_ls *ls,const struct dlm_rcom *rc_in) { -spin_lock (&ls->ls_rcom_spin); +spin_lock_bh (&ls->ls_rcom_spin);if (!test_bit (LSFL_RCOM_WAIT , &ls->ls_flags) ||le64_to_cpu (rc_in->rc_id) != ls->ls_rcom_seq) { log_debug(ls, "reject reply %d from %d seq %llx expect %llx",
@@ -301,7 +301,7 @@ static void receive_sync_reply(struct dlm_ls *ls, const struct dlm_rcom *rc_in)clear_bit (LSFL_RCOM_WAIT , &ls->ls_flags);wake_up (&ls->ls_wait_general); out: -spin_unlock (&ls->ls_rcom_spin); +spin_unlock_bh (&ls->ls_rcom_spin); } int dlm_rcom_names(struct dlm_ls *ls, int nodeid, char *last_name,
@@ -613,11 +613,11 @@ void dlm_receive_rcom(struct dlm_ls *ls, const struct dlm_rcom *rc, int nodeid)break ; } -spin_lock (&ls->ls_recover_lock); +spin_lock_bh (&ls->ls_recover_lock); status = ls->ls_recover_status; stop =dlm_recovery_stopped (ls); seq = ls->ls_recover_seq; -spin_unlock (&ls->ls_recover_lock); +spin_unlock_bh (&ls->ls_recover_lock);if (stop && (rc->rc_type !=cpu_to_le32 (DLM_RCOM_STATUS ))) goto ignore;
fs/dlm/recover.c +26 -26
175 lines not on this page · raw message
fs/dlm/recoverd.c +10 -10
@@ -26,7 +26,7 @@ static int dlm_create_masters_list(struct dlm_ls *ls)struct dlm_rsb *r;int i, error =0 ; -write_lock (&ls->ls_masters_lock); +write_lock_bh (&ls->ls_masters_lock);if (!list_empty (&ls->ls_masters_list)) {log_error (ls,"root list not empty" ); error = -EINVAL ;
@@ -46,7 +46,7 @@ static int dlm_create_masters_list(struct dlm_ls *ls)spin_unlock_bh (&ls->ls_rsbtbl[i].lock); } out: -write_unlock (&ls->ls_masters_lock); +write_unlock_bh (&ls->ls_masters_lock);return error; }
@@ -54,12 +54,12 @@ static void dlm_release_masters_list(struct dlm_ls *ls) {struct dlm_rsb *r, *safe; -write_lock (&ls->ls_masters_lock); +write_lock_bh (&ls->ls_masters_lock);list_for_each_entry_safe (r, safe, &ls->ls_masters_list, res_masters_list) {list_del_init (&r->res_masters_list);dlm_put_rsb (r); } -write_unlock (&ls->ls_masters_lock); +write_unlock_bh (&ls->ls_masters_lock); }static void dlm_create_root_list (struct dlm_ls *ls,struct list_head *root_list)
@@ -103,9 +103,9 @@ static int enable_locking(struct dlm_ls *ls, uint64_t seq) {int error = -EINTR ; -write_lock (&ls->ls_recv_active); +write_lock_bh (&ls->ls_recv_active); -spin_lock (&ls->ls_recover_lock); +spin_lock_bh (&ls->ls_recover_lock);if (ls->ls_recover_seq == seq) {set_bit (LSFL_RUNNING , &ls->ls_flags);/* unblocks processes waiting to enter the dlm */
@@ -113,9 +113,9 @@ static int enable_locking(struct dlm_ls *ls, uint64_t seq)clear_bit (LSFL_RECOVER_LOCK , &ls->ls_flags); error =0 ; } -spin_unlock (&ls->ls_recover_lock); +spin_unlock_bh (&ls->ls_recover_lock); -write_unlock (&ls->ls_recv_active); +write_unlock_bh (&ls->ls_recv_active);return error; }
@@ -348,12 +348,12 @@ static void do_ls_recovery(struct dlm_ls *ls)struct dlm_recover *rv =NULL ;int error; -spin_lock (&ls->ls_recover_lock); +spin_lock_bh (&ls->ls_recover_lock); rv = ls->ls_recover_args; ls->ls_recover_args =NULL ;if (rv && ls->ls_recover_seq == rv->seq)clear_bit (LSFL_RECOVER_STOP , &ls->ls_flags); -spin_unlock (&ls->ls_recover_lock); +spin_unlock_bh (&ls->ls_recover_lock);if (rv) { error =ls_recover (ls, rv);
fs/dlm/requestqueue.c +6 -6
@@ -68,7 +68,7 @@ int dlm_process_requestqueue(struct dlm_ls *ls)struct dlm_message *ms;int error =0 ; -write_lock (&ls->ls_requestqueue_lock); +write_lock_bh (&ls->ls_requestqueue_lock);for (;;) {if (list_empty (&ls->ls_requestqueue)) {clear_bit (LSFL_RECV_MSG_BLOCKED , &ls->ls_flags);
@@ -96,11 +96,11 @@ int dlm_process_requestqueue(struct dlm_ls *ls) error = -EINTR ;break ; } -write_unlock (&ls->ls_requestqueue_lock); +write_unlock_bh (&ls->ls_requestqueue_lock);schedule (); -write_lock (&ls->ls_requestqueue_lock); +write_lock_bh (&ls->ls_requestqueue_lock); } -write_unlock (&ls->ls_requestqueue_lock); +write_unlock_bh (&ls->ls_requestqueue_lock);return error; }
@@ -135,7 +135,7 @@ void dlm_purge_requestqueue(struct dlm_ls *ls)struct dlm_message *ms;struct rq_entry *e, *safe; -write_lock (&ls->ls_requestqueue_lock); +write_lock_bh (&ls->ls_requestqueue_lock);list_for_each_entry_safe (e, safe, &ls->ls_requestqueue, list) { ms = &e->request;
@@ -144,6 +144,6 @@ void dlm_purge_requestqueue(struct dlm_ls *ls)kfree (e); } } -write_unlock (&ls->ls_requestqueue_lock); +write_unlock_bh (&ls->ls_requestqueue_lock); }
fs/dlm/user.c +17 -17
@@ -206,7 +206,7 @@ void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,return ; ls = lkb->lkb_resource->res_ls; -spin_lock (&ls->ls_clear_proc_locks); +spin_lock_bh (&ls->ls_clear_proc_locks);/* If ORPHAN/DEAD flag is set, it means the process is dead so an ast can't be delivered. For ORPHAN's, dlm_clear_proc_locks() freed
@@ -228,12 +228,12 @@ void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,if ((flags &DLM_CB_CAST ) &&lkb_is_endoflife (mode, status))set_bit (DLM_IFL_ENDOFLIFE_BIT , &lkb->lkb_iflags); -spin_lock (&proc->asts_spin); +spin_lock_bh (&proc->asts_spin); rv =dlm_enqueue_lkb_callback (lkb, flags, mode, status, sbflags);switch (rv) {case DLM_ENQUEUE_CALLBACK_FAILURE : -spin_unlock (&proc->asts_spin); +spin_unlock_bh (&proc->asts_spin);WARN_ON_ONCE (1 ); goto out;case DLM_ENQUEUE_CALLBACK_NEED_SCHED :
@@ -247,19 +247,19 @@ void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,WARN_ON_ONCE (1 );break ; } -spin_unlock (&proc->asts_spin); +spin_unlock_bh (&proc->asts_spin);if (test_bit (DLM_IFL_ENDOFLIFE_BIT , &lkb->lkb_iflags)) {/* N.B. spin_lock locks_spin, not asts_spin */ -spin_lock (&proc->locks_spin); +spin_lock_bh (&proc->locks_spin);if (!list_empty (&lkb->lkb_ownqueue)) {list_del_init (&lkb->lkb_ownqueue);dlm_put_lkb (lkb); } -spin_unlock (&proc->locks_spin); +spin_unlock_bh (&proc->locks_spin); } out: -spin_unlock (&ls->ls_clear_proc_locks); +spin_unlock_bh (&ls->ls_clear_proc_locks); } static int device_user_lock(struct dlm_user_proc *proc,
@@ -832,10 +832,10 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count,if (test_bit (DLM_PROC_FLAGS_CLOSING , &proc->flags))return -EINVAL ; -spin_lock (&proc->asts_spin); +spin_lock_bh (&proc->asts_spin);if (list_empty (&proc->asts)) {if (file->f_flags &O_NONBLOCK ) { -spin_unlock (&proc->asts_spin); +spin_unlock_bh (&proc->asts_spin);return -EAGAIN ; }
@@ -844,16 +844,16 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count, repeat:set_current_state (TASK_INTERRUPTIBLE );if (list_empty (&proc->asts) && !signal_pending (current)) { -spin_unlock (&proc->asts_spin); +spin_unlock_bh (&proc->asts_spin);schedule (); -spin_lock (&proc->asts_spin); +spin_lock_bh (&proc->asts_spin); goto repeat; }set_current_state (TASK_RUNNING );remove_wait_queue (&proc->wait, &wait);if (signal_pending (current)) { -spin_unlock (&proc->asts_spin); +spin_unlock_bh (&proc->asts_spin);return -ERESTARTSYS ; } }
@@ -875,7 +875,7 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count,*/ log_print ("dlm_rem_lkb_callback empty %x" , lkb->lkb_id);list_del_init (&lkb->lkb_cb_list); -spin_unlock (&proc->asts_spin); +spin_unlock_bh (&proc->asts_spin);/* removes ref for proc->asts, may cause lkb to be freed */ dlm_put_lkb (lkb);WARN_ON_ONCE (1 );
@@ -890,7 +890,7 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count,WARN_ON_ONCE (1 );break ; } -spin_unlock (&proc->asts_spin); +spin_unlock_bh (&proc->asts_spin);if (cb->flags &DLM_CB_BAST ) {trace_dlm_bast (lkb->lkb_resource->res_ls, lkb, cb->mode);
@@ -925,12 +925,12 @@ static __poll_t device_poll(struct file *file, poll_table *wait)poll_wait (file, &proc->wait, wait); -spin_lock (&proc->asts_spin); +spin_lock_bh (&proc->asts_spin);if (!list_empty (&proc->asts)) { -spin_unlock (&proc->asts_spin); +spin_unlock_bh (&proc->asts_spin);return EPOLLIN |EPOLLRDNORM ; } -spin_unlock (&proc->asts_spin); +spin_unlock_bh (&proc->asts_spin);return 0 ; }
Signature · 1 line
2.43.0
Alexander Aring
maintainerPATCH v3 18/18 dlm: do dlm message processing in softirq context
In reply to Alexander AringThis patch moves the dlm message processing from a ordered workqueue context to a ordered softirq context. Later we want to call the user defined ast/bast callbacks directly inside the dlm message processing context instead of doing an additional context switch to the exisiting callback workqueue. This should slightly improve the dlm message parsing behaviour. There are two main reasons why to change to this behaviour:
1.
Allow fewer scheduling possibilities for dlm message parsing context. This should deliver faster DLM user responses to ast/bast callbacks. Fewer interrupting of lock requests processing that might trigger a new lock request avoids situations that we don't finish lock requests. In future the DLM callback workqueue can be disabled by a kernel lockspace flag to signal the DLM kernel user is capable of exectuing the callbacks in softirq context. If this flag is set, the dlm processing gets rid of an additional queue_work() context switch that should take more advantage about the new softirq context because the last preemption possibility is removed from the message processing context.
2. Bringing the ast/callback callback to softirq context that the use is aware it should not block into this context. Later patches will introduce a per lockspace flag to signal that the user is capable to handling these callbacks in softirq context to solve backwards compatibility.
3. We can easily switch to a per dlm instance concurrent dlm message parsing when DLM is ready to handle it. This instance could be in e.g. per lockspace instance or more fine granulatiry instance such as per lock instance.
Futher patches will unveil more improvements to switch to a per message softirq parsing context. Especially if we getting DLM in a state that we can allow concurrent message parsing.
--- fs/dlm/lowcomms.c | 34 ++++++++++++---------------------- 1 file changed, 12 insertions(+), 22 deletions(-)
fs/dlm/lowcomms.c +12 -22
@@ -183,7 +183,6 @@ static int dlm_local_count;/* Work queues */ static struct workqueue_struct *io_workqueue; -static struct workqueue_struct *process_workqueue;static struct hlist_head connection_hash[CONN_HASH_SIZE ];static DEFINE_SPINLOCK (connections_lock);
@@ -199,9 +198,9 @@ static const struct dlm_proto_ops *dlm_proto_ops;static void process_recv_sockets (struct work_struct *work);static void process_send_sockets (struct work_struct *work); -static void process_dlm_messages (struct work_struct *work); +static void process_dlm_messages (struct tasklet_struct *tasklet); -static DECLARE_WORK(process_work, process_dlm_messages); +static DECLARE_TASKLET_DISABLED(process_tasklet, process_dlm_messages);static DEFINE_SPINLOCK (processqueue_lock);static bool process_dlm_messages_pending;static atomic_t processqueue_count;
@@ -863,7 +862,7 @@ struct dlm_processed_nodes {struct list_head list; }; -static void process_dlm_messages (struct work_struct *work) +static void process_dlm_messages (struct tasklet_struct *tasklet) {struct processqueue_entry *pentry;
@@ -971,7 +970,7 @@ static int receive_from_sock(struct connection *con, int buflen)list_add_tail (&pentry->list, &processqueue);if (!process_dlm_messages_pending) { process_dlm_messages_pending = true; -queue_work (process_workqueue, &process_work); +tasklet_schedule (&process_tasklet); }spin_unlock_bh (&processqueue_lock);
@@ -1511,7 +1510,8 @@ static void process_recv_sockets(struct work_struct *work)/* CF_RECV_PENDING cleared */ break ;case DLM_IO_FLUSH : -flush_workqueue (process_workqueue); +tasklet_disable (&process_tasklet); +tasklet_enable (&process_tasklet); fallthrough;case DLM_IO_RESCHED :cond_resched ();
@@ -1686,10 +1686,7 @@ static void work_stop(void) io_workqueue =NULL ; } -if (process_workqueue) { -destroy_workqueue (process_workqueue); - process_workqueue =NULL ; - } +tasklet_disable (&process_tasklet); }static int work_start (void )
@@ -1701,17 +1698,7 @@ static int work_start(void)return -ENOMEM ; } -/* ordered dlm message process queue, -* should be converted to a tasklet -*/ - process_workqueue =alloc_ordered_workqueue ("dlm_process" , -WQ_HIGHPRI |WQ_MEM_RECLAIM ); -if (!process_workqueue) { -log_print ("can't start dlm_process" ); -destroy_workqueue (io_workqueue); - io_workqueue =NULL ; -return -ENOMEM ; - } +tasklet_enable (&process_tasklet);return 0 ; }
@@ -1734,7 +1721,10 @@ void dlm_lowcomms_shutdown(void)hlist_for_each_entry_rcu (con, &connection_hash[i], list) {shutdown_connection (con, true);stop_connection_io (con); -flush_workqueue (process_workqueue); + +tasklet_disable (&process_tasklet); +tasklet_enable (&process_tasklet); +close_connection (con, true);clean_one_writequeue (con);
Signature · 1 line
2.43.0