/*
* Set up the argument/result storage required for the RPC call.
*/
static int nfs_write_rpcsetup(struct nfs_page *req,
struct nfs_write_data *data,
const struct rpc_call_ops *call_ops,
unsigned int count, unsigned int offset,
int how)
{
struct inode *inode = req->wb_context->path.dentry->d_inode;
int flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
int priority = flush_task_priority(how);
struct rpc_task *task;
struct rpc_message msg = {
.rpc_argp = &data->args,
.rpc_resp = &data->res,
.rpc_cred = req->wb_context->cred,
};
struct rpc_task_setup task_setup_data = {
.rpc_client = NFS_CLIENT(inode),
.task = &data->task,
.rpc_message = &msg,
.callback_ops = call_ops,
.callback_data = data,
.workqueue = nfsiod_workqueue,
.flags = flags,
.priority = priority,
};
/* Set up the RPC argument and reply structs
* NB: take care not to mess about with data->commit et al. */
data->req = req;
data->inode = inode = req->wb_context->path.dentry->d_inode;
data->cred = msg.rpc_cred;
data->args.fh = NFS_FH(inode);
data->args.offset = req_offset(req) + offset;
data->args.pgbase = req->wb_pgbase + offset;
data->args.pages = data->pagevec;
data->args.count = count;
data->args.context = get_nfs_open_context(req->wb_context);
data->args.stable = NFS_UNSTABLE;
if (how & FLUSH_STABLE) {
data->args.stable = NFS_DATA_SYNC;
if (!nfs_need_commit(NFS_I(inode)))
data->args.stable = NFS_FILE_SYNC;
}
data->res.fattr = &data->fattr;
data->res.count = count;
data->res.verf = &data->verf;
nfs_fattr_init(&data->fattr);
/* Set up the initial task struct. */
NFS_PROTO(inode)->write_setup(data, &msg);
dprintk("NFS: %5u initiated write call "
"(req %s/%lld, %u bytes @ offset %llu)\n",
data->task.tk_pid,
inode->i_sb->s_id,
(long long)NFS_FILEID(inode),
count,
(unsigned long long)data->args.offset);
task = rpc_run_task(&task_setup_data);
if (IS_ERR(task))
return PTR_ERR(task);
rpc_put_task(task);
return 0;
}
/* If a nfs_flush_* function fails, it should remove reqs from @head and
* call this on each, which will prepare them to be retried on next
* writeback using standard nfs.
*/
static void nfs_redirty_request(struct nfs_page *req)
{
nfs_mark_request_dirty(req);
nfs_end_page_writeback(req->wb_page);
nfs_clear_page_tag_locked(req);
}
/*
* Generate multiple small requests to write out a single
* contiguous dirty area on one page.
*/
static int nfs_flush_multi(struct inode *inode, struct list_head *head, unsigned int npages, size_t count, int how)
{
struct nfs_page *req = nfs_list_entry(head->next);
struct page *page = req->wb_page;
struct nfs_write_data *data;
size_t wsize = NFS_SERVER(inode)->wsize, nbytes;
unsigned int offset;
int requests = 0;
int ret = 0;
LIST_HEAD(list);
nfs_list_remove_request(req);
nbytes = count;
do {
size_t len = min(nbytes, wsize);
data = nfs_writedata_alloc(1);
if (!data)
goto out_bad;
list_add(&data->pages, &list);
requests++;
nbytes -= len;
} while (nbytes != 0);
atomic_set(&req->wb_complete, requests);
ClearPageError(page);
offset = 0;
nbytes = count;
do {
int ret2;
data = list_entry(list.next, struct nfs_write_data, pages);
list_del_init(&data->pages);
data->pagevec[0] = page;
if (nbytes < wsize)
wsize = nbytes;
ret2 = nfs_write_rpcsetup(req, data, &nfs_write_partial_ops,
wsize, offset, how);
if (ret == 0)
ret = ret2;
offset += wsize;
nbytes -= wsize;
} while (nbytes != 0);
return ret;
out_bad:
while (!list_empty(&list)) {
data = list_entry(list.next, struct nfs_write_data, pages);
list_del(&data->pages);
nfs_writedata_release(data);
}
nfs_redirty_request(req);
return -ENOMEM;
}
/*
* Create an RPC task for the given write request and kick it.
* The page must have been locked by the caller.
*
* It may happen that the page we're passed is not marked dirty.
* This is the case if nfs_updatepage detects a conflicting request
* that has been written but not committed.
*/
static int nfs_flush_one(struct inode *inode, struct list_head *head, unsigned int npages, size_t count, int how)
{
struct nfs_page *req;
struct page **pages;
struct nfs_write_data *data;
data = nfs_writedata_alloc(npages);
if (!data)
goto out_bad;
pages = data->pagevec;
while (!list_empty(head)) {
req = nfs_list_entry(head->next);
nfs_list_remove_request(req);
nfs_list_add_request(req, &data->pages);
ClearPageError(req->wb_page);
*pages++ = req->wb_page;
}
req = nfs_list_entry(data->pages.next);
/* Set up the argument struct */
return nfs_write_rpcsetup(req, data, &nfs_write_full_ops, count, 0, how);
out_bad:
while (!list_empty(head)) {
req = nfs_list_entry(head->next);
nfs_list_remove_request(req);
nfs_redirty_request(req);
}
return -ENOMEM;
}
static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
struct inode *inode, int ioflags)
{
size_t wsize = NFS_SERVER(inode)->wsize;
if (wsize < PAGE_CACHE_SIZE)
nfs_pageio_init(pgio, inode, nfs_flush_multi, wsize, ioflags);
else
nfs_pageio_init(pgio, inode, nfs_flush_one, wsize, ioflags);
}
/*
* the back merge hash support functions
*/
static inline void __deadline_del_drq_hash(struct deadline_rq *drq)
{
drq->on_hash = 0;
list_del_init(&drq->hash);
}
/**
* v9fs_unregister_trans - unregister a 9p transport
* @m: the transport to remove
*
*/
void v9fs_unregister_trans(struct p9_trans_module *m)
{
spin_lock(&v9fs_trans_lock);
list_del_init(&m->list);
spin_unlock(&v9fs_trans_lock);
}
void RxPktPendingTimeout(unsigned long data)
{
PRX_TS_RECORD pRxTs = (PRX_TS_RECORD)data;
struct rtllib_device *ieee = container_of(pRxTs, struct rtllib_device, RxTsRecord[pRxTs->num]);
PRX_REORDER_ENTRY pReorderEntry = NULL;
unsigned long flags = 0;
struct rtllib_rxb *stats_IndicateArray[REORDER_WIN_SIZE];
u8 index = 0;
bool bPktInBuf = false;
spin_lock_irqsave(&(ieee->reorder_spinlock), flags);
if(pRxTs->RxTimeoutIndicateSeq != 0xffff)
{
while(!list_empty(&pRxTs->RxPendingPktList))
{
pReorderEntry = (PRX_REORDER_ENTRY)list_entry(pRxTs->RxPendingPktList.prev,RX_REORDER_ENTRY,List);
if(index == 0)
pRxTs->RxIndicateSeq = pReorderEntry->SeqNum;
if( SN_LESS(pReorderEntry->SeqNum, pRxTs->RxIndicateSeq) ||
SN_EQUAL(pReorderEntry->SeqNum, pRxTs->RxIndicateSeq) )
{
list_del_init(&pReorderEntry->List);
if(SN_EQUAL(pReorderEntry->SeqNum, pRxTs->RxIndicateSeq))
pRxTs->RxIndicateSeq = (pRxTs->RxIndicateSeq + 1) % 4096;
RTLLIB_DEBUG(RTLLIB_DL_REORDER,"%s(): Indicate SeqNum: %d\n",__func__, pReorderEntry->SeqNum);
stats_IndicateArray[index] = pReorderEntry->prxb;
index++;
list_add_tail(&pReorderEntry->List, &ieee->RxReorder_Unused_List);
}
else
{
bPktInBuf = true;
break;
}
}
}
if(index>0){
pRxTs->RxTimeoutIndicateSeq = 0xffff;
if(index > REORDER_WIN_SIZE){
RTLLIB_DEBUG(RTLLIB_DL_ERR, "RxReorderIndicatePacket(): Rx Reorer buffer full!! \n");
spin_unlock_irqrestore(&(ieee->reorder_spinlock), flags);
return;
}
rtllib_indicate_packets(ieee, stats_IndicateArray, index);
bPktInBuf = false;
}
if(bPktInBuf && (pRxTs->RxTimeoutIndicateSeq==0xffff)){
pRxTs->RxTimeoutIndicateSeq = pRxTs->RxIndicateSeq;
mod_timer(&pRxTs->RxPktPendingTimer, jiffies + MSECS(ieee->pHTInfo->RxReorderPendingTime));
}
spin_unlock_irqrestore(&(ieee->reorder_spinlock), flags);
}
/*
* AFS Cache Manager kernel thread
*/
static int kafscmd(void *arg)
{
DECLARE_WAITQUEUE(myself, current);
struct rxrpc_call *call;
_SRXAFSCM_xxxx_t func;
int die;
printk("kAFS: Started kafscmd %d\n", current->pid);
daemonize("kafscmd");
complete(&kafscmd_alive);
/* loop around looking for things to attend to */
do {
if (list_empty(&kafscmd_attention_list)) {
set_current_state(TASK_INTERRUPTIBLE);
add_wait_queue(&kafscmd_sleepq, &myself);
for (;;) {
set_current_state(TASK_INTERRUPTIBLE);
if (!list_empty(&kafscmd_attention_list) ||
signal_pending(current) ||
kafscmd_die)
break;
schedule();
}
remove_wait_queue(&kafscmd_sleepq, &myself);
set_current_state(TASK_RUNNING);
}
die = kafscmd_die;
/* dequeue the next call requiring attention */
call = NULL;
spin_lock(&kafscmd_attention_lock);
if (!list_empty(&kafscmd_attention_list)) {
call = list_entry(kafscmd_attention_list.next,
struct rxrpc_call,
app_attn_link);
list_del_init(&call->app_attn_link);
die = 0;
}
spin_unlock(&kafscmd_attention_lock);
if (call) {
/* act upon it */
_debug("@@@ Begin Attend Call %p", call);
func = call->app_user;
if (func)
func(call);
rxrpc_put_call(call);
_debug("@@@ End Attend Call %p", call);
}
} while(!die);
static void shadow_lru_isolate(struct list_head *item,
spinlock_t *lru_lock)
{
struct address_space *mapping;
struct radix_tree_node *node;
unsigned int i;
/*
* Page cache insertions and deletions synchroneously maintain
* the shadow node LRU under the mapping->tree_lock and the
* lru_lock. Because the page cache tree is emptied before
* the inode can be destroyed, holding the lru_lock pins any
* address_space that has radix tree nodes on the LRU.
*
* We can then safely transition to the mapping->tree_lock to
* pin only the address_space of the particular node we want
* to reclaim, take the node off-LRU, and drop the lru_lock.
*/
node = container_of(item, struct radix_tree_node, private_list);
mapping = node->private_data;
/* Coming from the list, invert the lock order */
if (!spin_trylock(&mapping->tree_lock)) {
spin_unlock(lru_lock);
goto out;
}
list_del_init(item);
nr_shadow_nodes--;
spin_unlock(lru_lock);
/*
* The nodes should only contain one or more shadow entries,
* no pages, so we expect to be able to remove them all and
* delete and free the empty node afterwards.
*/
BUG_ON(!node->count);
BUG_ON(node->count & RADIX_TREE_COUNT_MASK);
for (i = 0; i < RADIX_TREE_MAP_SIZE; i++) {
if (node->slots[i]) {
BUG_ON(!radix_tree_exceptional_entry(node->slots[i]));
node->slots[i] = NULL;
BUG_ON(node->count < (1U << RADIX_TREE_COUNT_SHIFT));
node->count -= 1U << RADIX_TREE_COUNT_SHIFT;
BUG_ON(!mapping->nrshadows);
mapping->nrshadows--;
}
}
BUG_ON(node->count);
inc_zone_state(page_zone(virt_to_page(node)), WORKINGSET_NODERECLAIM);
if (!__radix_tree_delete_node(&mapping->page_tree, node))
BUG();
spin_unlock(&mapping->tree_lock);
out:
local_irq_enable();
cond_resched();
local_irq_disable();
spin_lock(lru_lock);
}
/**ltl
功能:合并两个request:把next合到rq之后
参数:
返回值:
说明；rq和next两个请求的sector相邻。
*/
static void noop_merged_requests(request_queue_t *q, struct request *rq,
struct request *next)
{
list_del_init(&next->queuelist);
}
/**
* Searches the specified queue for the specified queue for the command
* to abort.
*
* @param [in] a
* @param [in] abort_request
* @param [in] cmd
* t
* @return 0 on failure, 1 if command was not found, 2 if command was found
*/
static int esas2r_check_active_queue(struct esas2r_adapter *a,
struct esas2r_request **abort_request,
struct scsi_cmnd *cmd,
struct list_head *queue)
{
bool found = false;
struct esas2r_request *ar = *abort_request;
struct esas2r_request *rq;
struct list_head *element, *next;
list_for_each_safe(element, next, queue) {
rq = list_entry(element, struct esas2r_request, req_list);
if (rq->cmd == cmd) {
/* Found the request. See what to do with it. */
if (queue == &a->active_list) {
/*
* We are searching the active queue, which
* means that we need to send an abort request
* to the firmware.
*/
ar = esas2r_alloc_request(a);
if (ar == NULL) {
esas2r_log_dev(ESAS2R_LOG_WARN,
&(a->host->shost_gendev),
"unable to allocate an abort request for cmd %p",
cmd);
return 0; /* Failure */
}
/*
* Task management request must be formatted
* with a lock held.
*/
ar->sense_len = 0;
ar->vrq->scsi.length = 0;
ar->target_id = rq->target_id;
ar->vrq->scsi.flags |= cpu_to_le32(
(u8)le32_to_cpu(rq->vrq->scsi.flags));
memset(ar->vrq->scsi.cdb, 0,
sizeof(ar->vrq->scsi.cdb));
ar->vrq->scsi.flags |= cpu_to_le32(
FCP_CMND_TRM);
ar->vrq->scsi.u.abort_handle =
rq->vrq->scsi.handle;
} else {
/*
* The request is pending but not active on
* the firmware. Just free it now and we'll
* report the successful abort below.
*/
list_del_init(&rq->req_list);
esas2r_free_request(a, rq);
}
found = true;
break;
}
}
/* entry of listen service */
NEINT32 common_update_entry(struct listen_contex *listen_info)
{
NEINT32 ret ,sleep = 0 ;
struct cm_manager *pmanger ;
struct ne_client_map *client;
NEUINT16 session_id = 0;
#ifdef USER_UPDATE_LIST
struct list_head *pos ;
swap_list_t *swap ;
#else
// NEINT32 i, num ;
cmlist_iterator_t cm_iterator ;
#endif
pmanger = ne_listensrv_get_cmmamager(listen_info) ;
if(ne_atomic_read(&pmanger->connect_num) <= 0) {
ne_sleep(100) ;
return 0 ;
}
#ifdef USER_UPDATE_LIST
swap = &listen_info->wait_add ;
if(0==ne_mutex_trylock(&swap->lock) ) {
list_join(&swap->list, &listen_info->conn_list) ;
INIT_LIST_HEAD(&swap->list) ;
ne_mutex_unlock(&swap->lock) ;
}
pos = listen_info->conn_list.next ;
while (pos!=&listen_info->conn_list) {
client = list_entry(pos,struct ne_client_map , map_list) ;
pos = pos->next ;
session_id = client->connect_node.session_id ;
client = pmanger->trylock(pmanger, session_id) ;
if(client) {
++sleep;
ret = tryto_close_tcpsession((ne_session_handle)client, listen_info->operate_timeout ) ;
if(ret) {
if(-1==ret) list_del_init(&client->map_list) ;
pmanger->unlock(pmanger,session_id) ;
continue ;
}
ret = ne_do_netmsg(client,&listen_info->tcp) ;
if(ret > 0) {
ne_tcpnode_flush_sendbuf(&(client->connect_node)) ;
}
else if(ret ==0){
if(0==tcp_client_close(client,1) ) {
list_del_init(&client->map_list) ;
}
}
pmanger->unlock(pmanger,session_id) ;
}
}
#else
// num = ne_atomic_read(&pmanger->connect_num);
// i = 0 ;
for(client = pmanger->lock_first (pmanger,&cm_iterator) ; client; client = pmanger->lock_next (pmanger,&cm_iterator) ) {
if(tryto_close_tcpsession((ne_session_handle)client, listen_info->operate_timeout ) ) {
continue ;
}
ret = ne_do_netmsg(client,&listen_info->tcp) ;
if(ret>0) {
ne_tcpnode_flush_sendbuf(&(client->connect_node)) ;
sleep++ ;
}
else if(0==ret) {
tcp_client_close(client,1) ;
}
}
#endif
if(!sleep ){
ne_sleep(100);
}
return 0;
}
static int osio_dispatch(struct request_queue *q, int force)
{
struct osio_data *od = q->elevator->elevator_data;
const unsigned int non_empty[3] = {!list_empty(&od->fifo_head[OSIO_DIR_READ]),
!list_empty(&od->fifo_head[OSIO_DIR_SYNC_WRITE]),
!list_empty(&od->fifo_head[OSIO_DIR_ASYNC_WRITE]),};
struct request *rq = NULL;
osio_dbg("1, od->fifo_dir = %d\n", od->fifo_dir);
osio_dbg("1, non_empty[0] = %d\n", non_empty[0]);
osio_dbg("1, non_empty[1] = %d\n", non_empty[1]);
osio_dbg("1, non_empty[2] = %d\n", non_empty[2]);
/* dispatch a batch of rq */
if (od->fifo_dir != OSIO_DIR_UNDEF) {
if ((od->batching >= od->fifo_batch[od->fifo_dir]) || (!non_empty[od->fifo_dir])) {
od->fifo_dir = OSIO_DIR_UNDEF;
} else {
goto dispatch_request;
}
}
/* redecide the direction */
if (non_empty[OSIO_DIR_READ]) {
goto dir_read;
}
if (non_empty[OSIO_DIR_SYNC_WRITE]) {
goto dir_sync_write;
}
if (non_empty[OSIO_DIR_ASYNC_WRITE]) {
goto dir_async_write;
}
return 0;
dir_read:
/* find a starved write rq */
if ((od->write_starved[OSIO_SYNC] > od->write_starved_line[OSIO_SYNC]) && non_empty[OSIO_DIR_SYNC_WRITE]) {
goto dir_sync_write;
} else if ((od->write_starved[OSIO_ASYNC] > od->write_starved_line[OSIO_ASYNC]) && non_empty[OSIO_DIR_ASYNC_WRITE]) {
goto dir_async_write;
}
od->fifo_dir = OSIO_DIR_READ;
od->batching = 0;
od->write_starved[OSIO_SYNC] += non_empty[OSIO_DIR_SYNC_WRITE];
od->write_starved[OSIO_ASYNC] += non_empty[OSIO_DIR_ASYNC_WRITE];
goto dispatch_request;
dir_sync_write:
if ((od->write_starved[OSIO_ASYNC] > od->write_starved_line[OSIO_ASYNC]) && non_empty[OSIO_DIR_ASYNC_WRITE]) {
goto dir_async_write;
}
od->fifo_dir = OSIO_DIR_SYNC_WRITE;
od->batching = 0;
od->write_starved[OSIO_SYNC] = 0;
od->write_starved[OSIO_ASYNC] += non_empty[OSIO_DIR_ASYNC_WRITE];
goto dispatch_request;
dir_async_write:
od->fifo_dir = OSIO_DIR_ASYNC_WRITE;
od->batching = 0;
od->write_starved[OSIO_ASYNC] = 0;
od->write_starved[OSIO_SYNC] += non_empty[OSIO_DIR_SYNC_WRITE];
goto dispatch_request;
dispatch_request:
/* dispatch req */
osio_dbg("2, od->fifo_dir = %d\n", od->fifo_dir);
osio_dbg("2, od->batching = %d\n", od->batching);
rq = rq_entry_fifo(od->fifo_head[od->fifo_dir].next);
list_del_init(&rq->queuelist);
elv_dispatch_add_tail(q, rq);
od->batching ++;
return 1;
}
static ssize_t ksb_fs_read(struct file *fp, char __user *buf,
size_t count, loff_t *pos)
{
int ret;
unsigned long flags;
struct ks_bridge *ksb = fp->private_data;
struct data_pkt *pkt;
size_t space, copied;
read_start:
if (!test_bit(USB_DEV_CONNECTED, &ksb->flags))
return -ENODEV;
spin_lock_irqsave(&ksb->lock, flags);
if (list_empty(&ksb->to_ks_list)) {
spin_unlock_irqrestore(&ksb->lock, flags);
ret = wait_event_interruptible(ksb->ks_wait_q,
!list_empty(&ksb->to_ks_list) ||
!test_bit(USB_DEV_CONNECTED, &ksb->flags));
if (ret < 0)
return ret;
goto read_start;
}
space = count;
copied = 0;
while (!list_empty(&ksb->to_ks_list) && space) {
size_t len;
pkt = list_first_entry(&ksb->to_ks_list, struct data_pkt, list);
len = min_t(size_t, space, pkt->len);
pkt->n_read += len;
spin_unlock_irqrestore(&ksb->lock, flags);
ret = copy_to_user(buf + copied, pkt->buf, len);
if (ret) {
pr_err("copy_to_user failed err:%d\n", ret);
ksb_free_data_pkt(pkt);
return ret;
}
space -= len;
copied += len;
spin_lock_irqsave(&ksb->lock, flags);
if (pkt->n_read == pkt->len) {
/*
* re-init the packet and queue it
* for more data.
*/
list_del_init(&pkt->list);
pkt->n_read = 0;
pkt->len = MAX_DATA_PKT_SIZE;
spin_unlock_irqrestore(&ksb->lock, flags);
submit_one_urb(ksb, GFP_KERNEL, pkt);
spin_lock_irqsave(&ksb->lock, flags);
}
}
spin_unlock_irqrestore(&ksb->lock, flags);
dbg_log_event(ksb, "KS_READ", copied, 0);
pr_debug("count:%d space:%d copied:%d", count, space, copied);
return copied;
}
/******** functions ********/
static void osio_merged_requests(struct request_queue *q, struct request *rq, struct request *next)
{
list_del_init(&next->queuelist);
}
/* All actions that we need after sending hello on passive conn:
* 1) Cope with 1st easy case: conn is already linked to a peer
* 2) Cope with 2nd easy case: remove zombie conn
* 3) Resolve race:
* a) find the peer
* b) link the conn to the peer if conn[idx] is empty
* c) if the conn[idx] isn't empty and is in READY state,
* remove the conn as duplicated
* d) if the conn[idx] isn't empty and isn't in READY state,
* override conn[idx] with the conn
*/
int
usocklnd_passiveconn_hellosent(usock_conn_t *conn)
{
usock_conn_t *conn2;
usock_peer_t *peer;
struct list_head tx_list;
struct list_head zcack_list;
int idx;
int rc = 0;
/* almost nothing to do if conn is already linked to peer hash table */
if (conn->uc_peer != NULL)
goto passive_hellosent_done;
/* conn->uc_peer == NULL, so the conn isn't accessible via
* peer hash list, so nobody can touch the conn but us */
if (conn->uc_ni == NULL) /* remove zombie conn */
goto passive_hellosent_connkill;
/* all code below is race resolution, because normally
* passive conn is linked to peer just after receiving hello */
CFS_INIT_LIST_HEAD (&tx_list);
CFS_INIT_LIST_HEAD (&zcack_list);
/* conn is passive and isn't linked to any peer,
so its tx and zc_ack lists have to be empty */
LASSERT (list_empty(&conn->uc_tx_list) &&
list_empty(&conn->uc_zcack_list) &&
conn->uc_sending == 0);
rc = usocklnd_find_or_create_peer(conn->uc_ni, conn->uc_peerid, &peer);
if (rc)
return rc;
idx = usocklnd_type2idx(conn->uc_type);
/* try to link conn to peer */
pthread_mutex_lock(&peer->up_lock);
if (peer->up_conns[idx] == NULL) {
usocklnd_link_conn_to_peer(conn, peer, idx);
usocklnd_conn_addref(conn);
conn->uc_peer = peer;
usocklnd_peer_addref(peer);
} else {
conn2 = peer->up_conns[idx];
pthread_mutex_lock(&conn2->uc_lock);
if (conn2->uc_state == UC_READY) {
/* conn2 is in READY state, so conn is "duplicated" */
pthread_mutex_unlock(&conn2->uc_lock);
pthread_mutex_unlock(&peer->up_lock);
usocklnd_peer_decref(peer);
usocklnd_conn_kill(conn2);
goto passive_hellosent_connkill;
}
/* uc_state != UC_READY => switch conn and conn2 */
/* Relink txs and zc_acks from conn2 to conn.
* We're sure that nobody but us can access to conn,
* nevertheless we use mutex (if we're wrong yet,
* deadlock is easy to see that corrupted list */
list_add(&tx_list, &conn2->uc_tx_list);
list_del_init(&conn2->uc_tx_list);
list_add(&zcack_list, &conn2->uc_zcack_list);
list_del_init(&conn2->uc_zcack_list);
pthread_mutex_lock(&conn->uc_lock);
list_add_tail(&conn->uc_tx_list, &tx_list);
list_del_init(&tx_list);
list_add_tail(&conn->uc_zcack_list, &zcack_list);
list_del_init(&zcack_list);
conn->uc_peer = peer;
pthread_mutex_unlock(&conn->uc_lock);
conn2->uc_peer = NULL; /* make conn2 zombie */
pthread_mutex_unlock(&conn2->uc_lock);
usocklnd_conn_decref(conn2);
usocklnd_link_conn_to_peer(conn, peer, idx);
usocklnd_conn_addref(conn);
conn->uc_peer = peer;
}
lnet_ni_decref(conn->uc_ni);
conn->uc_ni = NULL;
pthread_mutex_unlock(&peer->up_lock);
usocklnd_peer_decref(peer);
passive_hellosent_done:
/* safely transit to UC_READY state */
/* rc == 0 */
pthread_mutex_lock(&conn->uc_lock);
if (conn->uc_state != UC_DEAD) {
usocklnd_rx_ksmhdr_state_transition(conn);
/* we're ready to recive incoming packets and maybe
already have smth. to transmit */
LASSERT (conn->uc_sending == 0);
if ( list_empty(&conn->uc_tx_list) &&
list_empty(&conn->uc_zcack_list) ) {
conn->uc_tx_flag = 0;
rc = usocklnd_add_pollrequest(conn, POLL_SET_REQUEST,
POLLIN);
} else {
conn->uc_tx_deadline =
cfs_time_shift(usock_tuns.ut_timeout);
conn->uc_tx_flag = 1;
rc = usocklnd_add_pollrequest(conn, POLL_SET_REQUEST,
POLLIN | POLLOUT);
}
if (rc == 0)
conn->uc_state = UC_READY;
}
pthread_mutex_unlock(&conn->uc_lock);
return rc;
passive_hellosent_connkill:
usocklnd_conn_kill(conn);
return 0;
}
/* All actions that we need after receiving hello on active conn:
* 1) Schedule removing if we're zombie
* 2) Restart active conn if we lost the race
* 3) Else: update RX part to receive KSM header
*/
int
usocklnd_activeconn_hellorecv(usock_conn_t *conn)
{
int rc = 0;
ksock_hello_msg_t *hello = conn->uc_rx_hello;
usock_peer_t *peer = conn->uc_peer;
/* Active conn with peer==NULL is zombie.
* Don't try to link it to peer because the conn
* has already had a chance to proceed at the beginning */
if (peer == NULL) {
LASSERT(list_empty(&conn->uc_tx_list) &&
list_empty(&conn->uc_zcack_list));
usocklnd_conn_kill(conn);
return 0;
}
peer->up_last_alive = cfs_time_current();
/* peer says that we lost the race */
if (hello->kshm_ctype == (__u32)SOCKLND_CONN_NONE) {
/* Start new active conn, relink txs and zc_acks from
* the conn to new conn, schedule removing the conn.
* Actually, we're expecting that a passive conn will
* make us zombie soon and take care of our txs and
* zc_acks */
struct list_head tx_list, zcack_list;
usock_conn_t *conn2;
int idx = usocklnd_type2idx(conn->uc_type);
CFS_INIT_LIST_HEAD (&tx_list);
CFS_INIT_LIST_HEAD (&zcack_list);
/* Block usocklnd_send() to check peer->up_conns[idx]
* and to enqueue more txs */
pthread_mutex_lock(&peer->up_lock);
pthread_mutex_lock(&conn->uc_lock);
/* usocklnd_shutdown() could kill us */
if (conn->uc_state == UC_DEAD) {
pthread_mutex_unlock(&conn->uc_lock);
pthread_mutex_unlock(&peer->up_lock);
return 0;
}
LASSERT (peer == conn->uc_peer);
LASSERT (peer->up_conns[idx] == conn);
rc = usocklnd_create_active_conn(peer, conn->uc_type, &conn2);
if (rc) {
conn->uc_errored = 1;
pthread_mutex_unlock(&conn->uc_lock);
pthread_mutex_unlock(&peer->up_lock);
return rc;
}
usocklnd_link_conn_to_peer(conn2, peer, idx);
conn2->uc_peer = peer;
/* unlink txs and zcack from the conn */
list_add(&tx_list, &conn->uc_tx_list);
list_del_init(&conn->uc_tx_list);
list_add(&zcack_list, &conn->uc_zcack_list);
list_del_init(&conn->uc_zcack_list);
/* link they to the conn2 */
list_add(&conn2->uc_tx_list, &tx_list);
list_del_init(&tx_list);
list_add(&conn2->uc_zcack_list, &zcack_list);
list_del_init(&zcack_list);
/* make conn zombie */
conn->uc_peer = NULL;
usocklnd_peer_decref(peer);
/* schedule conn2 for processing */
rc = usocklnd_add_pollrequest(conn2, POLL_ADD_REQUEST, POLLOUT);
if (rc) {
peer->up_conns[idx] = NULL;
usocklnd_conn_decref(conn2); /* should destroy conn */
} else {
usocklnd_conn_kill_locked(conn);
}
pthread_mutex_unlock(&conn->uc_lock);
pthread_mutex_unlock(&peer->up_lock);
usocklnd_conn_decref(conn);
} else { /* hello->kshm_ctype != SOCKLND_CONN_NONE */
if (conn->uc_type != usocklnd_invert_type(hello->kshm_ctype))
return -EPROTO;
pthread_mutex_lock(&peer->up_lock);
usocklnd_cleanup_stale_conns(peer, hello->kshm_src_incarnation,
conn);
pthread_mutex_unlock(&peer->up_lock);
/* safely transit to UC_READY state */
/* rc == 0 */
pthread_mutex_lock(&conn->uc_lock);
if (conn->uc_state != UC_DEAD) {
usocklnd_rx_ksmhdr_state_transition(conn);
/* POLLIN is already set because we just
* received hello, but maybe we've smth. to
* send? */
LASSERT (conn->uc_sending == 0);
if ( !list_empty(&conn->uc_tx_list) ||
!list_empty(&conn->uc_zcack_list) ) {
conn->uc_tx_deadline =
cfs_time_shift(usock_tuns.ut_timeout);
conn->uc_tx_flag = 1;
rc = usocklnd_add_pollrequest(conn,
POLL_SET_REQUEST,
POLLIN | POLLOUT);
}
if (rc == 0)
conn->uc_state = UC_READY;
}
pthread_mutex_unlock(&conn->uc_lock);
}
return rc;
}
/**
* isci_terminate_request_core() - This function will terminate the given
* request, and wait for it to complete. This function must only be called
* from a thread that can wait. Note that the request is terminated and
* completed (back to the host, if started there).
* @ihost: This SCU.
* @idev: The target.
* @isci_request: The I/O request to be terminated.
*
*/
static void isci_terminate_request_core(struct isci_host *ihost,
struct isci_remote_device *idev,
struct isci_request *isci_request)
{
enum sci_status status = SCI_SUCCESS;
bool was_terminated = false;
bool needs_cleanup_handling = false;
unsigned long flags;
unsigned long termination_completed = 1;
struct completion *io_request_completion;
dev_dbg(&ihost->pdev->dev,
"%s: device = %p; request = %p\n",
__func__, idev, isci_request);
spin_lock_irqsave(&ihost->scic_lock, flags);
io_request_completion = isci_request->io_request_completion;
/* Note that we are not going to control
* the target to abort the request.
*/
set_bit(IREQ_COMPLETE_IN_TARGET, &isci_request->flags);
/* Make sure the request wasn't just sitting around signalling
* device condition (if the request handle is NULL, then the
* request completed but needed additional handling here).
*/
if (!test_bit(IREQ_TERMINATED, &isci_request->flags)) {
was_terminated = true;
needs_cleanup_handling = true;
status = sci_controller_terminate_request(ihost,
idev,
isci_request);
}
spin_unlock_irqrestore(&ihost->scic_lock, flags);
/*
* The only time the request to terminate will
* fail is when the io request is completed and
* being aborted.
*/
if (status != SCI_SUCCESS) {
dev_dbg(&ihost->pdev->dev,
"%s: sci_controller_terminate_request"
" returned = 0x%x\n",
__func__, status);
isci_request->io_request_completion = NULL;
} else {
if (was_terminated) {
dev_dbg(&ihost->pdev->dev,
"%s: before completion wait (%p/%p)\n",
__func__, isci_request, io_request_completion);
/* Wait here for the request to complete. */
termination_completed
= wait_for_completion_timeout(
io_request_completion,
msecs_to_jiffies(ISCI_TERMINATION_TIMEOUT_MSEC));
if (!termination_completed) {
/* The request to terminate has timed out. */
spin_lock_irqsave(&ihost->scic_lock, flags);
/* Check for state changes. */
if (!test_bit(IREQ_TERMINATED,
&isci_request->flags)) {
/* The best we can do is to have the
* request die a silent death if it
* ever really completes.
*/
isci_request_mark_zombie(ihost,
isci_request);
needs_cleanup_handling = true;
} else
termination_completed = 1;
spin_unlock_irqrestore(&ihost->scic_lock,
flags);
if (!termination_completed) {
dev_dbg(&ihost->pdev->dev,
"%s: *** Timeout waiting for "
"termination(%p/%p)\n",
__func__, io_request_completion,
isci_request);
/* The request can no longer be referenced
* safely since it may go away if the
* termination every really does complete.
*/
isci_request = NULL;
}
}
if (termination_completed)
dev_dbg(&ihost->pdev->dev,
"%s: after completion wait (%p/%p)\n",
__func__, isci_request, io_request_completion);
}
if (termination_completed) {
isci_request->io_request_completion = NULL;
/* Peek at the status of the request. This will tell
* us if there was special handling on the request such that it
* needs to be detached and freed here.
*/
spin_lock_irqsave(&isci_request->state_lock, flags);
needs_cleanup_handling
= isci_request_is_dealloc_managed(
isci_request->status);
spin_unlock_irqrestore(&isci_request->state_lock, flags);
}
if (needs_cleanup_handling) {
dev_dbg(&ihost->pdev->dev,
"%s: cleanup isci_device=%p, request=%p\n",
__func__, idev, isci_request);
if (isci_request != NULL) {
spin_lock_irqsave(&ihost->scic_lock, flags);
isci_free_tag(ihost, isci_request->io_tag);
isci_request_change_state(isci_request, unallocated);
list_del_init(&isci_request->dev_node);
spin_unlock_irqrestore(&ihost->scic_lock, flags);
}
}
}
}
static int vmw_gmr_build_descriptors(struct list_head *desc_pages,
struct page *pages[],
unsigned long num_pages)
{
struct page *page, *next;
struct svga_guest_mem_descriptor *page_virtual = NULL;
struct svga_guest_mem_descriptor *desc_virtual = NULL;
unsigned int desc_per_page;
unsigned long prev_pfn;
unsigned long pfn;
int ret;
desc_per_page = PAGE_SIZE /
sizeof(struct svga_guest_mem_descriptor) - 1;
while (likely(num_pages != 0)) {
page = alloc_page(__GFP_HIGHMEM);
if (unlikely(page == NULL)) {
ret = -ENOMEM;
goto out_err;
}
list_add_tail(&page->lru, desc_pages);
/*
* Point previous page terminating descriptor to this
* page before unmapping it.
*/
if (likely(page_virtual != NULL)) {
desc_virtual->ppn = page_to_pfn(page);
kunmap_atomic(page_virtual, KM_USER0);
}
page_virtual = kmap_atomic(page, KM_USER0);
desc_virtual = page_virtual - 1;
prev_pfn = ~(0UL);
while (likely(num_pages != 0)) {
pfn = page_to_pfn(*pages);
if (pfn != prev_pfn + 1) {
if (desc_virtual - page_virtual ==
desc_per_page - 1)
break;
(++desc_virtual)->ppn = cpu_to_le32(pfn);
desc_virtual->num_pages = cpu_to_le32(1);
} else {
uint32_t tmp =
le32_to_cpu(desc_virtual->num_pages);
desc_virtual->num_pages = cpu_to_le32(tmp + 1);
}
prev_pfn = pfn;
--num_pages;
++pages;
}
(++desc_virtual)->ppn = cpu_to_le32(0);
desc_virtual->num_pages = cpu_to_le32(0);
}
if (likely(page_virtual != NULL))
kunmap_atomic(page_virtual, KM_USER0);
return 0;
out_err:
list_for_each_entry_safe(page, next, desc_pages, lru) {
list_del_init(&page->lru);
__free_page(page);
}
static void hub_events(void)
{
unsigned long flags;
struct list_head *tmp;
struct usb_device *dev;
struct usb_hub *hub;
u16 hubstatus;
u16 hubchange;
u16 portstatus;
u16 portchange;
int i, ret;
int m=0;
/*
* We restart the list every time to avoid a deadlock with
* deleting hubs downstream from this one. This should be
* safe since we delete the hub from the event list.
* Not the most efficient, but avoids deadlocks.
*/
while (m<5) {
m++;
spin_lock_irqsave(&hub_event_lock, flags);
if (list_empty(&hub_event_list))
break;
/* Grab the next entry from the beginning of the list */
tmp = hub_event_list.next;
hub = list_entry(tmp, struct usb_hub, event_list);
dev = interface_to_usbdev(hub->intf);
list_del_init(tmp);
if (unlikely(down_trylock(&hub->khubd_sem)))
BUG(); /* never blocks, we were on list */
spin_unlock_irqrestore(&hub_event_lock, flags);
if (hub->error) {
dev_dbg (&hub->intf->dev, "resetting for error %d\n",
hub->error);
if (hub_reset(hub)) {
dev_dbg (&hub->intf->dev,
"can't reset; disconnecting\n");
up(&hub->khubd_sem);
hub_start_disconnect(dev);
continue;
}
hub->nerrors = 0;
hub->error = 0;
}
for (i = 0; i < hub->descriptor->bNbrPorts; i++) {
ret = hub_port_status(dev, i, &portstatus, &portchange);
if (ret < 0) {
continue;
}
if (portchange & USB_PORT_STAT_C_CONNECTION) {
hub_port_connect_change(hub, i, portstatus, portchange);
} else if (portchange & USB_PORT_STAT_C_ENABLE) {
dev_dbg (hubdev (dev),
"port %d enable change, status %x\n",
i + 1, portstatus);
clear_port_feature(dev,
i + 1, USB_PORT_FEAT_C_ENABLE);
/*
* EM interference sometimes causes badly
* shielded USB devices to be shutdown by
* the hub, this hack enables them again.
* Works at least with mouse driver.
*/
if (!(portstatus & USB_PORT_STAT_ENABLE)
&& (portstatus & USB_PORT_STAT_CONNECTION)
&& (dev->children[i])) {
dev_err (&hub->intf->dev,
"port %i "
"disabled by hub (EMI?), "
"re-enabling...",
i + 1);
hub_port_connect_change(hub,
i, portstatus, portchange);
}
}
if (portchange & USB_PORT_STAT_C_SUSPEND) {
dev_dbg (&hub->intf->dev,
"suspend change on port %d\n",
i + 1);
clear_port_feature(dev,
i + 1, USB_PORT_FEAT_C_SUSPEND);
}
if (portchange & USB_PORT_STAT_C_OVERCURRENT) {
dev_err (&hub->intf->dev,
"over-current change on port %d\n",
i + 1);
clear_port_feature(dev,
i + 1, USB_PORT_FEAT_C_OVER_CURRENT);
hub_power_on(hub);
}
if (portchange & USB_PORT_STAT_C_RESET) {
dev_dbg (&hub->intf->dev,
"reset change on port %d\n",
i + 1);
clear_port_feature(dev,
i + 1, USB_PORT_FEAT_C_RESET);
}
} /* end for i */
/* deal with hub status changes */
if (hub_hub_status(hub, &hubstatus, &hubchange) < 0)
dev_err (&hub->intf->dev, "get_hub_status failed\n");
else {
if (hubchange & HUB_CHANGE_LOCAL_POWER) {
dev_dbg (&hub->intf->dev, "power change\n");
clear_hub_feature(dev, C_HUB_LOCAL_POWER);
}
if (hubchange & HUB_CHANGE_OVERCURRENT) {
dev_dbg (&hub->intf->dev, "overcurrent change\n");
wait_ms(500); /* Cool down */
clear_hub_feature(dev, C_HUB_OVER_CURRENT);
hub_power_on(hub);
}
}
up(&hub->khubd_sem);
} /* end while (1) */
spin_unlock_irqrestore(&hub_event_lock, flags);
}
/**
* kthread_create - create a kthread.
* @threadfn: the function to run until signal_pending(current).
* @data: data ptr for @threadfn.
* @namefmt: printf-style name for the thread.
*
* Description: This helper function creates and names a kernel
* thread. The thread will be stopped: use wake_up_process() to start
* it. See also kthread_run(), kthread_create_on_cpu().
*
* When woken, the thread will run @threadfn() with @data as its
* argument. @threadfn() can either call do_exit() directly if it is a
* standalone thread for which noone will call kthread_stop(), or
* return when 'kthread_should_stop()' is true (which means
* kthread_stop() has been called). The return value should be zero
* or a negative error number; it will be passed to kthread_stop().
*
* Returns a task_struct or ERR_PTR(-ENOMEM).
*/
struct task_struct *kthread_create(int (*threadfn)(void *data),
void *data,
const char namefmt[],
...)
{
struct kthread_create_info create;
create.threadfn = threadfn;
create.data = data;
init_completion(&create.done);
spin_lock(&kthread_create_lock);
list_add_tail(&create.list, &kthread_create_list);
spin_unlock(&kthread_create_lock);
wake_up_process(kthreadd_task);
wait_for_completion(&create.done);
if (!IS_ERR(create.result)) {
struct sched_param param = { .sched_priority = 0 };
va_list args;
va_start(args, namefmt);
vsnprintf(create.result->comm, sizeof(create.result->comm),
namefmt, args);
va_end(args);
/*
* root may have changed our (kthreadd's) priority or CPU mask.
* The kernel thread should not inherit these properties.
*/
sched_setscheduler_nocheck(create.result, SCHED_NORMAL, ¶m);
#ifdef CONFIG_SYS_HAS_CONTROL_CPU
set_cpus_allowed_ptr(create.result, cpumask_of(0));
#else
set_cpus_allowed_ptr(create.result, cpu_all_mask);
#endif
}
return create.result;
}
EXPORT_SYMBOL(kthread_create);
/**
* kthread_stop - stop a thread created by kthread_create().
* @k: thread created by kthread_create().
*
* Sets kthread_should_stop() for @k to return true, wakes it, and
* waits for it to exit. This can also be called after kthread_create()
* instead of calling wake_up_process(): the thread will exit without
* calling threadfn().
*
* If threadfn() may call do_exit() itself, the caller must ensure
* task_struct can't go away.
*
* Returns the result of threadfn(), or %-EINTR if wake_up_process()
* was never called.
*/
int kthread_stop(struct task_struct *k)
{
struct kthread *kthread;
int ret;
trace_sched_kthread_stop(k);
get_task_struct(k);
kthread = to_kthread(k);
barrier(); /* it might have exited */
if (k->vfork_done != NULL) {
kthread->should_stop = 1;
wake_up_process(k);
wait_for_completion(&kthread->exited);
}
ret = k->exit_code;
put_task_struct(k);
trace_sched_kthread_stop_ret(ret);
return ret;
}
EXPORT_SYMBOL(kthread_stop);
int kthreadd(void *unused)
{
struct task_struct *tsk = current;
/* Setup a clean context for our children to inherit. */
set_task_comm(tsk, "kthreadd");
ignore_signals(tsk);
#ifdef CONFIG_SYS_HAS_CONTROL_CPU
set_cpus_allowed_ptr(tsk, cpu_control_mask);
#else
set_cpus_allowed_ptr(tsk, cpu_all_mask);
#endif
set_mems_allowed(node_possible_map);
current->flags |= PF_NOFREEZE | PF_FREEZER_NOSIG;
for (;;) {
set_current_state(TASK_INTERRUPTIBLE);
if (list_empty(&kthread_create_list))
schedule();
__set_current_state(TASK_RUNNING);
spin_lock(&kthread_create_lock);
while (!list_empty(&kthread_create_list)) {
struct kthread_create_info *create;
create = list_entry(kthread_create_list.next,
struct kthread_create_info, list);
list_del_init(&create->list);
spin_unlock(&kthread_create_lock);
create_kthread(create);
spin_lock(&kthread_create_lock);
}
spin_unlock(&kthread_create_lock);
}
return 0;
}
static inline void _dequeue_task_dummy(struct task_struct *p)
{
struct sched_dummy_entity *dummy_se = &p->dummy_se;
list_del_init(&dummy_se->run_list);
}
static ssize_t
printer_read(struct file *fd, char __user *buf, size_t len, loff_t *ptr)
{
struct printer_dev *dev = fd->private_data;
unsigned long flags;
size_t size;
size_t bytes_copied;
struct usb_request *req;
/* This is a pointer to the current USB rx request. */
struct usb_request *current_rx_req;
/* This is the number of bytes in the current rx buffer. */
size_t current_rx_bytes;
/* This is a pointer to the current rx buffer. */
u8 *current_rx_buf;
if (len == 0)
return -EINVAL;
DBG(dev, "printer_read trying to read %d bytes\n", (int)len);
mutex_lock(&dev->lock_printer_io);
spin_lock_irqsave(&dev->lock, flags);
/* We will use this flag later to check if a printer reset happened
* after we turn interrupts back on.
*/
dev->reset_printer = 0;
setup_rx_reqs(dev);
bytes_copied = 0;
current_rx_req = dev->current_rx_req;
current_rx_bytes = dev->current_rx_bytes;
current_rx_buf = dev->current_rx_buf;
dev->current_rx_req = NULL;
dev->current_rx_bytes = 0;
dev->current_rx_buf = NULL;
/* Check if there is any data in the read buffers. Please note that
* current_rx_bytes is the number of bytes in the current rx buffer.
* If it is zero then check if there are any other rx_buffers that
* are on the completed list. We are only out of data if all rx
* buffers are empty.
*/
if ((current_rx_bytes == 0) &&
(likely(list_empty(&dev->rx_buffers)))) {
/* Turn interrupts back on before sleeping. */
spin_unlock_irqrestore(&dev->lock, flags);
/*
* If no data is available check if this is a NON-Blocking
* call or not.
*/
if (fd->f_flags & (O_NONBLOCK|O_NDELAY)) {
mutex_unlock(&dev->lock_printer_io);
return -EAGAIN;
}
/* Sleep until data is available */
wait_event_interruptible(dev->rx_wait,
(likely(!list_empty(&dev->rx_buffers))));
spin_lock_irqsave(&dev->lock, flags);
}
/* We have data to return then copy it to the caller's buffer.*/
while ((current_rx_bytes || likely(!list_empty(&dev->rx_buffers)))
&& len) {
if (current_rx_bytes == 0) {
req = container_of(dev->rx_buffers.next,
struct usb_request, list);
list_del_init(&req->list);
if (req->actual && req->buf) {
current_rx_req = req;
current_rx_bytes = req->actual;
current_rx_buf = req->buf;
} else {
list_add(&req->list, &dev->rx_reqs);
continue;
}
}
/* Don't leave irqs off while doing memory copies */
spin_unlock_irqrestore(&dev->lock, flags);
if (len > current_rx_bytes)
size = current_rx_bytes;
else
size = len;
size -= copy_to_user(buf, current_rx_buf, size);
bytes_copied += size;
len -= size;
buf += size;
spin_lock_irqsave(&dev->lock, flags);
/* We've disconnected or reset so return. */
if (dev->reset_printer) {
list_add(¤t_rx_req->list, &dev->rx_reqs);
spin_unlock_irqrestore(&dev->lock, flags);
mutex_unlock(&dev->lock_printer_io);
return -EAGAIN;
}
/* If we not returning all the data left in this RX request
* buffer then adjust the amount of data left in the buffer.
* Othewise if we are done with this RX request buffer then
* requeue it to get any incoming data from the USB host.
*/
if (size < current_rx_bytes) {
current_rx_bytes -= size;
current_rx_buf += size;
} else {
list_add(¤t_rx_req->list, &dev->rx_reqs);
current_rx_bytes = 0;
current_rx_buf = NULL;
current_rx_req = NULL;
}
}
/**
* kthread_create_on_node - create a kthread.
* @threadfn: the function to run until signal_pending(current).
* @data: data ptr for @threadfn.
* @node: memory node number.
* @namefmt: printf-style name for the thread.
*
* Description: This helper function creates and names a kernel
* thread. The thread will be stopped: use wake_up_process() to start
* it. See also kthread_run().
*
* If thread is going to be bound on a particular cpu, give its node
* in @node, to get NUMA affinity for kthread stack, or else give -1.
* When woken, the thread will run @threadfn() with @data as its
* argument. @threadfn() can either call do_exit() directly if it is a
* standalone thread for which no one will call kthread_stop(), or
* return when 'kthread_should_stop()' is true (which means
* kthread_stop() has been called). The return value should be zero
* or a negative error number; it will be passed to kthread_stop().
*
* Returns a task_struct or ERR_PTR(-ENOMEM).
*/
struct task_struct *kthread_create_on_node(int (*threadfn)(void *data),
void *data,
int node,
const char namefmt[],
...)
{
struct kthread_create_info create;
create.threadfn = threadfn;
create.data = data;
create.node = node;
init_completion(&create.done);
spin_lock(&kthread_create_lock);
list_add_tail(&create.list, &kthread_create_list);
spin_unlock(&kthread_create_lock);
wake_up_process(kthreadd_task);
wait_for_completion(&create.done);
if (!IS_ERR(create.result)) {
static const struct sched_param param = { .sched_priority = 0 };
va_list args;
va_start(args, namefmt);
vsnprintf(create.result->comm, sizeof(create.result->comm),
namefmt, args);
va_end(args);
/*
* root may have changed our (kthreadd's) priority or CPU mask.
* The kernel thread should not inherit these properties.
*/
sched_setscheduler_nocheck(create.result, SCHED_NORMAL, ¶m);
set_cpus_allowed_ptr(create.result, cpu_all_mask);
}
return create.result;
}
EXPORT_SYMBOL(kthread_create_on_node);
/**
* kthread_bind - bind a just-created kthread to a cpu.
* @p: thread created by kthread_create().
* @cpu: cpu (might not be online, must be possible) for @k to run on.
*
* Description: This function is equivalent to set_cpus_allowed(),
* except that @cpu doesn't need to be online, and the thread must be
* stopped (i.e., just returned from kthread_create()).
*/
void kthread_bind(struct task_struct *p, unsigned int cpu)
{
/* Must have done schedule() in kthread() before we set_task_cpu */
if (!wait_task_inactive(p, TASK_UNINTERRUPTIBLE)) {
WARN_ON(1);
return;
}
/* It's safe because the task is inactive. */
do_set_cpus_allowed(p, cpumask_of(cpu));
p->flags |= PF_THREAD_BOUND;
}
EXPORT_SYMBOL(kthread_bind);
/**
* kthread_stop - stop a thread created by kthread_create().
* @k: thread created by kthread_create().
*
* Sets kthread_should_stop() for @k to return true, wakes it, and
* waits for it to exit. This can also be called after kthread_create()
* instead of calling wake_up_process(): the thread will exit without
* calling threadfn().
*
* If threadfn() may call do_exit() itself, the caller must ensure
* task_struct can't go away.
*
* Returns the result of threadfn(), or %-EINTR if wake_up_process()
* was never called.
*/
int kthread_stop(struct task_struct *k)
{
struct kthread *kthread;
int ret;
trace_sched_kthread_stop(k);
get_task_struct(k);
kthread = to_kthread(k);
barrier(); /* it might have exited */
if (k->vfork_done != NULL) {
kthread->should_stop = 1;
wake_up_process(k);
wait_for_completion(&kthread->exited);
}
ret = k->exit_code;
put_task_struct(k);
trace_sched_kthread_stop_ret(ret);
return ret;
}
EXPORT_SYMBOL(kthread_stop);
int kthreadd(void *unused)
{
struct task_struct *tsk = current;
/* Setup a clean context for our children to inherit. */
set_task_comm(tsk, "kthreadd");
ignore_signals(tsk);
set_cpus_allowed_ptr(tsk, cpu_all_mask);
set_mems_allowed(node_states[N_HIGH_MEMORY]);
current->flags |= PF_NOFREEZE | PF_FREEZER_NOSIG;
for (;;) {
set_current_state(TASK_INTERRUPTIBLE);
if (list_empty(&kthread_create_list))
schedule();
__set_current_state(TASK_RUNNING);
spin_lock(&kthread_create_lock);
while (!list_empty(&kthread_create_list)) {
struct kthread_create_info *create;
create = list_entry(kthread_create_list.next,
struct kthread_create_info, list);
list_del_init(&create->list);
spin_unlock(&kthread_create_lock);
create_kthread(create);
spin_lock(&kthread_create_lock);
}
spin_unlock(&kthread_create_lock);
}
return 0;
}
void __init_kthread_worker(struct kthread_worker *worker,
const char *name,
struct lock_class_key *key)
{
spin_lock_init(&worker->lock);
lockdep_set_class_and_name(&worker->lock, key, name);
INIT_LIST_HEAD(&worker->work_list);
worker->task = NULL;
}
s32 rtl8188eu_xmitframe_complete(struct adapter *adapt, struct xmit_priv *pxmitpriv, struct xmit_buf *pxmitbuf)
{
struct hal_data_8188e *haldata = GET_HAL_DATA(adapt);
struct xmit_frame *pxmitframe = NULL;
struct xmit_frame *pfirstframe = NULL;
/* aggregate variable */
struct hw_xmit *phwxmit;
struct sta_info *psta = NULL;
struct tx_servq *ptxservq = NULL;
struct list_head *xmitframe_plist = NULL, *xmitframe_phead = NULL;
u32 pbuf; /* next pkt address */
u32 pbuf_tail; /* last pkt tail */
u32 len; /* packet length, except TXDESC_SIZE and PKT_OFFSET */
u32 bulksize = haldata->UsbBulkOutSize;
u8 desc_cnt;
u32 bulkptr;
/* dump frame variable */
u32 ff_hwaddr;
RT_TRACE(_module_rtl8192c_xmit_c_, _drv_info_, ("+xmitframe_complete\n"));
/* check xmitbuffer is ok */
if (pxmitbuf == NULL) {
pxmitbuf = rtw_alloc_xmitbuf(pxmitpriv);
if (pxmitbuf == NULL)
return false;
}
/* 3 1. pick up first frame */
rtw_free_xmitframe(pxmitpriv, pxmitframe);
pxmitframe = rtw_dequeue_xframe(pxmitpriv, pxmitpriv->hwxmits, pxmitpriv->hwxmit_entry);
if (pxmitframe == NULL) {
/* no more xmit frame, release xmit buffer */
rtw_free_xmitbuf(pxmitpriv, pxmitbuf);
return false;
}
pxmitframe->pxmitbuf = pxmitbuf;
pxmitframe->buf_addr = pxmitbuf->pbuf;
pxmitbuf->priv_data = pxmitframe;
pxmitframe->agg_num = 1; /* alloc xmitframe should assign to 1. */
pxmitframe->pkt_offset = 1; /* first frame of aggregation, reserve offset */
rtw_xmitframe_coalesce(adapt, pxmitframe->pkt, pxmitframe);
/* always return ndis_packet after rtw_xmitframe_coalesce */
rtw_os_xmit_complete(adapt, pxmitframe);
/* 3 2. aggregate same priority and same DA(AP or STA) frames */
pfirstframe = pxmitframe;
len = xmitframe_need_length(pfirstframe) + TXDESC_SIZE + (pfirstframe->pkt_offset*PACKET_OFFSET_SZ);
pbuf_tail = len;
pbuf = round_up(pbuf_tail, 8);
/* check pkt amount in one bulk */
desc_cnt = 0;
bulkptr = bulksize;
if (pbuf < bulkptr) {
desc_cnt++;
} else {
desc_cnt = 0;
bulkptr = ((pbuf / bulksize) + 1) * bulksize; /* round to next bulksize */
}
/* dequeue same priority packet from station tx queue */
psta = pfirstframe->attrib.psta;
switch (pfirstframe->attrib.priority) {
case 1:
case 2:
ptxservq = &(psta->sta_xmitpriv.bk_q);
phwxmit = pxmitpriv->hwxmits + 3;
break;
case 4:
case 5:
ptxservq = &(psta->sta_xmitpriv.vi_q);
phwxmit = pxmitpriv->hwxmits + 1;
break;
case 6:
case 7:
ptxservq = &(psta->sta_xmitpriv.vo_q);
phwxmit = pxmitpriv->hwxmits;
break;
case 0:
case 3:
default:
ptxservq = &(psta->sta_xmitpriv.be_q);
phwxmit = pxmitpriv->hwxmits + 2;
break;
}
spin_lock_bh(&pxmitpriv->lock);
xmitframe_phead = get_list_head(&ptxservq->sta_pending);
xmitframe_plist = xmitframe_phead->next;
while (xmitframe_phead != xmitframe_plist) {
pxmitframe = container_of(xmitframe_plist, struct xmit_frame, list);
xmitframe_plist = xmitframe_plist->next;
pxmitframe->agg_num = 0; /* not first frame of aggregation */
pxmitframe->pkt_offset = 0; /* not first frame of aggregation, no need to reserve offset */
len = xmitframe_need_length(pxmitframe) + TXDESC_SIZE + (pxmitframe->pkt_offset*PACKET_OFFSET_SZ);
if (round_up(pbuf + len, 8) > MAX_XMITBUF_SZ) {
pxmitframe->agg_num = 1;
pxmitframe->pkt_offset = 1;
break;
}
list_del_init(&pxmitframe->list);
ptxservq->qcnt--;
phwxmit->accnt--;
pxmitframe->buf_addr = pxmitbuf->pbuf + pbuf;
rtw_xmitframe_coalesce(adapt, pxmitframe->pkt, pxmitframe);
/* always return ndis_packet after rtw_xmitframe_coalesce */
rtw_os_xmit_complete(adapt, pxmitframe);
/* (len - TXDESC_SIZE) == pxmitframe->attrib.last_txcmdsz */
update_txdesc(pxmitframe, pxmitframe->buf_addr, pxmitframe->attrib.last_txcmdsz, true);
/* don't need xmitframe any more */
rtw_free_xmitframe(pxmitpriv, pxmitframe);
/* handle pointer and stop condition */
pbuf_tail = pbuf + len;
pbuf = round_up(pbuf_tail, 8);
pfirstframe->agg_num++;
if (MAX_TX_AGG_PACKET_NUMBER == pfirstframe->agg_num)
break;
if (pbuf < bulkptr) {
desc_cnt++;
if (desc_cnt == haldata->UsbTxAggDescNum)
break;
} else {
desc_cnt = 0;
bulkptr = ((pbuf / bulksize) + 1) * bulksize;
}
} /* end while (aggregate same priority and same DA(AP or STA) frames) */
if (list_empty(&ptxservq->sta_pending.queue))
list_del_init(&ptxservq->tx_pending);
spin_unlock_bh(&pxmitpriv->lock);
if ((pfirstframe->attrib.ether_type != 0x0806) &&
(pfirstframe->attrib.ether_type != 0x888e) &&
(pfirstframe->attrib.ether_type != 0x88b4) &&
(pfirstframe->attrib.dhcp_pkt != 1))
rtw_issue_addbareq_cmd(adapt, pfirstframe);
/* 3 3. update first frame txdesc */
if ((pbuf_tail % bulksize) == 0) {
/* remove pkt_offset */
pbuf_tail -= PACKET_OFFSET_SZ;
pfirstframe->buf_addr += PACKET_OFFSET_SZ;
pfirstframe->pkt_offset--;
}
update_txdesc(pfirstframe, pfirstframe->buf_addr, pfirstframe->attrib.last_txcmdsz, true);
/* 3 4. write xmit buffer to USB FIFO */
ff_hwaddr = rtw_get_ff_hwaddr(pfirstframe);
usb_write_port(adapt, ff_hwaddr, pbuf_tail, (u8 *)pxmitbuf);
/* 3 5. update statisitc */
pbuf_tail -= (pfirstframe->agg_num * TXDESC_SIZE);
pbuf_tail -= (pfirstframe->pkt_offset * PACKET_OFFSET_SZ);
rtw_count_tx_stats(adapt, pfirstframe, pbuf_tail);
rtw_free_xmitframe(pxmitpriv, pfirstframe);
return true;
}
void
schedule(void) {
bool intr_flag;
struct proc_struct *next;
#ifndef MT_SUPPORT
list_entry_t head;
int lapic_id = pls_read(lapic_id);
#endif
local_intr_save(intr_flag);
int lcpu_count = pls_read(lcpu_count);
{
current->need_resched = 0;
#ifndef MT_SUPPORT
if (current->mm)
{
assert(current->mm->lapic == lapic_id);
current->mm->lapic = -1;
}
#endif
if (current->state == PROC_RUNNABLE && current->pid >= lcpu_count) {
sched_class_enqueue(current);
}
#ifndef MT_SUPPORT
list_init(&head);
while (1)
{
next = sched_class_pick_next();
if (next != NULL) sched_class_dequeue(next);
if (next && next->mm && next->mm->lapic != -1)
{
list_add(&head, &(next->run_link));
}
else
{
list_entry_t *cur;
while ((cur = list_next(&head)) != &head)
{
list_del_init(cur);
sched_class_enqueue(le2proc(cur, run_link));
}
break;
}
}
#else
next = sched_class_pick_next();
if (next != NULL)
sched_class_dequeue(next);
#endif /* !MT_SUPPORT */
if (next == NULL) {
next = idleproc;
}
next->runs ++;
/* Collect information here*/
if (sched_collect_info) {
int lcpu_count = pls_read(lcpu_count);
int lcpu_idx = pls_read(lcpu_idx);
int loc = sched_info_head[lcpu_idx];
int prev = sched_info_pid[loc*lcpu_count + lcpu_idx];
if (next->pid == prev)
sched_info_times[loc*lcpu_count + lcpu_idx] ++;
else {
sched_info_head[lcpu_idx] ++;
if (sched_info_head[lcpu_idx] >= PGSIZE / sizeof(uint16_t) / lcpu_count)
sched_info_head[lcpu_idx] = 0;
loc = sched_info_head[lcpu_idx];
uint16_t prev_pid = sched_info_pid[loc*lcpu_count + lcpu_idx];
uint16_t prev_times = sched_info_times[loc*lcpu_count + lcpu_idx];
if (prev_times > 0 && prev_pid >= lcpu_count + 2)
sched_slices[lcpu_idx][prev_pid % SLICEPOOL_SIZE] += prev_times;
sched_info_pid[loc*lcpu_count + lcpu_idx] = next->pid;
sched_info_times[loc*lcpu_count + lcpu_idx] = 1;
}
}
#ifndef MT_SUPPORT
assert(!next->mm || next->mm->lapic == -1);
if (next->mm)
next->mm->lapic = lapic_id;
#endif
if (next != current) {
#if 0
kprintf("N %d to %d\n", current->pid, next->pid);
#endif
proc_run(next);
}
}
local_intr_restore(intr_flag);
}
/**
* process_cursors - do action on each cursor attached to inode
* @inode:
* @act: action to do
*
* Finds all cursors of @inode in reiser4's super block radix tree of cursors
* and performs action specified by @act on each of cursors.
*/
static void process_cursors(struct inode *inode, enum cursor_action act)
{
oid_t oid;
dir_cursor *start;
struct list_head *head;
reiser4_context *ctx;
struct d_cursor_info *info;
/* this can be called by
*
* kswapd->...->prune_icache->..reiser4_destroy_inode
*
* without reiser4_context
*/
ctx = reiser4_init_context(inode->i_sb);
if (IS_ERR(ctx)) {
warning("vs-23", "failed to init context");
return;
}
assert("nikita-3558", inode != NULL);
info = d_info(inode);
oid = get_inode_oid(inode);
spin_lock_inode(inode);
head = get_readdir_list(inode);
spin_lock(&d_lock);
/* find any cursor for this oid: reference to it is hanging of radix
* tree */
start = lookup(info, (unsigned long)oid);
if (start != NULL) {
dir_cursor *scan;
reiser4_file_fsdata *fsdata;
/* process circular list of cursors for this oid */
scan = start;
do {
dir_cursor *next;
next = list_entry(scan->list.next, dir_cursor, list);
fsdata = scan->fsdata;
assert("nikita-3557", fsdata != NULL);
if (scan->key.oid == oid) {
switch (act) {
case CURSOR_DISPOSE:
list_del_init(&fsdata->dir.linkage);
break;
case CURSOR_LOAD:
list_add(&fsdata->dir.linkage, head);
break;
case CURSOR_KILL:
kill_cursor(scan);
break;
}
}
if (scan == next)
/* last cursor was just killed */
break;
scan = next;
} while (scan != start);
}
spin_unlock(&d_lock);
/* check that we killed 'em all */
assert("nikita-3568",
ergo(act == CURSOR_KILL,
list_empty_careful(get_readdir_list(inode))));
assert("nikita-3569",
ergo(act == CURSOR_KILL, lookup(info, oid) == NULL));
spin_unlock_inode(inode);
reiser4_exit_context(ctx);
}
void drbd_req_destroy(struct kref *kref)
{
struct drbd_request *req = container_of(kref, struct drbd_request, kref);
struct drbd_conf *mdev = req->w.mdev;
const unsigned s = req->rq_state;
if ((req->master_bio && !(s & RQ_POSTPONED)) ||
atomic_read(&req->completion_ref) ||
(s & RQ_LOCAL_PENDING) ||
((s & RQ_NET_MASK) && !(s & RQ_NET_DONE))) {
dev_err(DEV, "drbd_req_destroy: Logic BUG rq_state = 0x%x, completion_ref = %d\n",
s, atomic_read(&req->completion_ref));
return;
}
/* remove it from the transfer log.
* well, only if it had been there in the first
* place... if it had not (local only or conflicting
* and never sent), it should still be "empty" as
* initialized in drbd_req_new(), so we can list_del() it
* here unconditionally */
list_del_init(&req->tl_requests);
/* if it was a write, we may have to set the corresponding
* bit(s) out-of-sync first. If it had a local part, we need to
* release the reference to the activity log. */
if (s & RQ_WRITE) {
/* Set out-of-sync unless both OK flags are set
* (local only or remote failed).
* Other places where we set out-of-sync:
* READ with local io-error */
/* There is a special case:
* we may notice late that IO was suspended,
* and postpone, or schedule for retry, a write,
* before it even was submitted or sent.
* In that case we do not want to touch the bitmap at all.
*/
if ((s & (RQ_POSTPONED|RQ_LOCAL_MASK|RQ_NET_MASK)) != RQ_POSTPONED) {
if (!(s & RQ_NET_OK) || !(s & RQ_LOCAL_OK))
drbd_set_out_of_sync(mdev, req->i.sector, req->i.size);
if ((s & RQ_NET_OK) && (s & RQ_LOCAL_OK) && (s & RQ_NET_SIS))
drbd_set_in_sync(mdev, req->i.sector, req->i.size);
}
/* one might be tempted to move the drbd_al_complete_io
* to the local io completion callback drbd_request_endio.
* but, if this was a mirror write, we may only
* drbd_al_complete_io after this is RQ_NET_DONE,
* otherwise the extent could be dropped from the al
* before it has actually been written on the peer.
* if we crash before our peer knows about the request,
* but after the extent has been dropped from the al,
* we would forget to resync the corresponding extent.
*/
if (s & RQ_IN_ACT_LOG) {
if (get_ldev_if_state(mdev, D_FAILED)) {
drbd_al_complete_io(mdev, &req->i);
put_ldev(mdev);
} else if (__ratelimit(&drbd_ratelimit_state)) {
dev_warn(DEV, "Should have called drbd_al_complete_io(, %llu, %u), "
"but my Disk seems to have failed :(\n",
(unsigned long long) req->i.sector, req->i.size);
}
}
}
mempool_free(req, drbd_request_mempool);
}
void
wait_queue_del(wait_queue_t *queue, wait_t *wait) {
assert(!list_empty(&(wait->wait_link)) && wait->wait_queue == queue);
list_del_init(&(wait->wait_link));
}
/**
* kthread_create - create a kthread.
* @threadfn: the function to run until signal_pending(current).
* @data: data ptr for @threadfn.
* @namefmt: printf-style name for the thread.
*
* Description: This helper function creates and names a kernel
* thread. The thread will be stopped: use wake_up_process() to start
* it. See also kthread_run(), kthread_create_on_cpu().
*
* When woken, the thread will run @threadfn() with @data as its
* argument. @threadfn() can either call do_exit() directly if it is a
* standalone thread for which noone will call kthread_stop(), or
* return when 'kthread_should_stop()' is true (which means
* kthread_stop() has been called). The return value should be zero
* or a negative error number; it will be passed to kthread_stop().
*
* Returns a task_struct or ERR_PTR(-ENOMEM).
*/
struct task_struct *kthread_create(int (*threadfn)(void *data),
void *data,
const char namefmt[],
...)
{
struct kthread_create_info create;
create.threadfn = threadfn;
create.data = data;
init_completion(&create.done);
spin_lock(&kthread_create_lock);
list_add_tail(&create.list, &kthread_create_list);
spin_unlock(&kthread_create_lock);
wake_up_process(kthreadd_task);
wait_for_completion(&create.done);
if (!IS_ERR(create.result)) {
struct sched_param param = { .sched_priority = 0 };
va_list args;
va_start(args, namefmt);
vsnprintf(create.result->comm, sizeof(create.result->comm),
namefmt, args);
va_end(args);
/*
* root may have changed our (kthreadd's) priority or CPU mask.
* The kernel thread should not inherit these properties.
*/
sched_setscheduler_nocheck(create.result, SCHED_NORMAL, ¶m);
set_cpus_allowed_ptr(create.result, cpu_all_mask);
}
return create.result;
}
EXPORT_SYMBOL(kthread_create);
/**
* kthread_stop - stop a thread created by kthread_create().
* @k: thread created by kthread_create().
*
* Sets kthread_should_stop() for @k to return true, wakes it, and
* waits for it to exit. This can also be called after kthread_create()
* instead of calling wake_up_process(): the thread will exit without
* calling threadfn().
*
* If threadfn() may call do_exit() itself, the caller must ensure
* task_struct can't go away.
*
* Returns the result of threadfn(), or %-EINTR if wake_up_process()
* was never called.
*/
int kthread_stop(struct task_struct *k)
{
struct kthread *kthread;
int ret;
trace_sched_kthread_stop(k);
get_task_struct(k);
kthread = to_kthread(k);
barrier(); /* it might have exited */
if (k->vfork_done != NULL) {
kthread->should_stop = 1;
wake_up_process(k);
wait_for_completion(&kthread->exited);
}
ret = k->exit_code;
put_task_struct(k);
trace_sched_kthread_stop_ret(ret);
return ret;
}
EXPORT_SYMBOL(kthread_stop);
int kthreadd(void *unused)
{
struct task_struct *tsk = current;
/* Setup a clean context for our children to inherit. */
set_task_comm(tsk, "kthreadd");
ignore_signals(tsk);
set_cpus_allowed_ptr(tsk, cpu_all_mask);
set_mems_allowed(node_states[N_HIGH_MEMORY]);
current->flags |= PF_NOFREEZE | PF_FREEZER_NOSIG;
for (;;) {
set_current_state(TASK_INTERRUPTIBLE);
if (list_empty(&kthread_create_list))
schedule();
__set_current_state(TASK_RUNNING);
spin_lock(&kthread_create_lock);
while (!list_empty(&kthread_create_list)) {
struct kthread_create_info *create;
create = list_entry(kthread_create_list.next,
struct kthread_create_info, list);
list_del_init(&create->list);
spin_unlock(&kthread_create_lock);
create_kthread(create);
spin_lock(&kthread_create_lock);
}
spin_unlock(&kthread_create_lock);
}
return 0;
}
/**
* kthread_worker_fn - kthread function to process kthread_worker
* @worker_ptr: pointer to initialized kthread_worker
*
* This function can be used as @threadfn to kthread_create() or
* kthread_run() with @worker_ptr argument pointing to an initialized
* kthread_worker. The started kthread will process work_list until
* the it is stopped with kthread_stop(). A kthread can also call
* this function directly after extra initialization.
*
* Different kthreads can be used for the same kthread_worker as long
* as there's only one kthread attached to it at any given time. A
* kthread_worker without an attached kthread simply collects queued
* kthread_works.
*/
int kthread_worker_fn(void *worker_ptr)
{
struct kthread_worker *worker = worker_ptr;
struct kthread_work *work;
WARN_ON(worker->task);
worker->task = current;
repeat:
set_current_state(TASK_INTERRUPTIBLE); /* mb paired w/ kthread_stop */
if (kthread_should_stop()) {
__set_current_state(TASK_RUNNING);
spin_lock_irq(&worker->lock);
worker->task = NULL;
spin_unlock_irq(&worker->lock);
return 0;
}
work = NULL;
spin_lock_irq(&worker->lock);
if (!list_empty(&worker->work_list)) {
work = list_first_entry(&worker->work_list,
struct kthread_work, node);
list_del_init(&work->node);
}
worker->current_work = work;
spin_unlock_irq(&worker->lock);
if (work) {
__set_current_state(TASK_RUNNING);
work->func(work);
} else if (!freezing(current))
schedule();
try_to_freeze();
goto repeat;
}
/*
* Any time a mark is getting freed we end up here.
* The caller had better be holding a reference to this mark so we don't actually
* do the final put under the mark->lock
*/
void fsnotify_destroy_mark_locked(struct fsnotify_mark *mark,
struct fsnotify_group *group)
{
struct inode *inode = NULL;
BUG_ON(!mutex_is_locked(&group->mark_mutex));
spin_lock(&mark->lock);
/* something else already called this function on this mark */
if (!(mark->flags & FSNOTIFY_MARK_FLAG_ALIVE)) {
spin_unlock(&mark->lock);
return;
}
mark->flags &= ~FSNOTIFY_MARK_FLAG_ALIVE;
if (mark->flags & FSNOTIFY_MARK_FLAG_INODE) {
inode = mark->inode;
fsnotify_destroy_inode_mark(mark);
} else if (mark->flags & FSNOTIFY_MARK_FLAG_VFSMOUNT)
fsnotify_destroy_vfsmount_mark(mark);
else if (mark->flags & FSNOTIFY_MARK_FLAG_TASK) {
fsnotify_destroy_task_mark(mark);
}
else
BUG();
list_del_init(&mark->g_list);
spin_unlock(&mark->lock);
if (inode && (mark->flags & FSNOTIFY_MARK_FLAG_OBJECT_PINNED))
iput(inode);
/* release lock temporarily */
mutex_unlock(&group->mark_mutex);
spin_lock(&destroy_lock);
list_add(&mark->g_list, &destroy_list);
spin_unlock(&destroy_lock);
wake_up(&destroy_waitq);
/*
* We don't necessarily have a ref on mark from caller so the above destroy
* may have actually freed it, unless this group provides a 'freeing_mark'
* function which must be holding a reference.
*/
/*
* Some groups like to know that marks are being freed. This is a
* callback to the group function to let it know that this mark
* is being freed.
*/
if (group->ops->freeing_mark)
group->ops->freeing_mark(mark, group);
/*
* __fsnotify_update_child_dentry_flags(inode);
*
* I really want to call that, but we can't, we have no idea if the inode
* still exists the second we drop the mark->lock.
*
* The next time an event arrive to this inode from one of it's children
* __fsnotify_parent will see that the inode doesn't care about it's
* children and will update all of these flags then. So really this
* is just a lazy update (and could be a perf win...)
*/
atomic_dec(&group->num_marks);
mutex_lock_nested(&group->mark_mutex, SINGLE_DEPTH_NESTING);
}
/*
* rrpc_move_valid_pages -- migrate live data off the block
* @rrpc: the 'rrpc' structure
* @block: the block from which to migrate live pages
*
* Description:
* GC algorithms may call this function to migrate remaining live
* pages off the block prior to erasing it. This function blocks
* further execution until the operation is complete.
*/
static int rrpc_move_valid_pages(struct rrpc *rrpc, struct rrpc_block *rblk)
{
struct nvm_tgt_dev *dev = rrpc->dev;
struct request_queue *q = dev->q;
struct rrpc_rev_addr *rev;
struct nvm_rq *rqd;
struct bio *bio;
struct page *page;
int slot;
int nr_sec_per_blk = dev->geo.sec_per_blk;
u64 phys_addr;
DECLARE_COMPLETION_ONSTACK(wait);
if (bitmap_full(rblk->invalid_pages, nr_sec_per_blk))
return 0;
bio = bio_alloc(GFP_NOIO, 1);
if (!bio) {
pr_err("nvm: could not alloc bio to gc\n");
return -ENOMEM;
}
page = mempool_alloc(rrpc->page_pool, GFP_NOIO);
if (!page) {
bio_put(bio);
return -ENOMEM;
}
while ((slot = find_first_zero_bit(rblk->invalid_pages,
nr_sec_per_blk)) < nr_sec_per_blk) {
/* Lock laddr */
phys_addr = rrpc_blk_to_ppa(rrpc, rblk) + slot;
try:
spin_lock(&rrpc->rev_lock);
/* Get logical address from physical to logical table */
rev = &rrpc->rev_trans_map[phys_addr];
/* already updated by previous regular write */
if (rev->addr == ADDR_EMPTY) {
spin_unlock(&rrpc->rev_lock);
continue;
}
rqd = rrpc_inflight_laddr_acquire(rrpc, rev->addr, 1);
if (IS_ERR_OR_NULL(rqd)) {
spin_unlock(&rrpc->rev_lock);
schedule();
goto try;
}
spin_unlock(&rrpc->rev_lock);
/* Perform read to do GC */
bio->bi_iter.bi_sector = rrpc_get_sector(rev->addr);
bio_set_op_attrs(bio, REQ_OP_READ, 0);
bio->bi_private = &wait;
bio->bi_end_io = rrpc_end_sync_bio;
/* TODO: may fail when EXP_PG_SIZE > PAGE_SIZE */
bio_add_pc_page(q, bio, page, RRPC_EXPOSED_PAGE_SIZE, 0);
if (rrpc_submit_io(rrpc, bio, rqd, NVM_IOTYPE_GC)) {
pr_err("rrpc: gc read failed.\n");
rrpc_inflight_laddr_release(rrpc, rqd);
goto finished;
}
wait_for_completion_io(&wait);
if (bio->bi_error) {
rrpc_inflight_laddr_release(rrpc, rqd);
goto finished;
}
bio_reset(bio);
reinit_completion(&wait);
bio->bi_iter.bi_sector = rrpc_get_sector(rev->addr);
bio_set_op_attrs(bio, REQ_OP_WRITE, 0);
bio->bi_private = &wait;
bio->bi_end_io = rrpc_end_sync_bio;
bio_add_pc_page(q, bio, page, RRPC_EXPOSED_PAGE_SIZE, 0);
/* turn the command around and write the data back to a new
* address
*/
if (rrpc_submit_io(rrpc, bio, rqd, NVM_IOTYPE_GC)) {
pr_err("rrpc: gc write failed.\n");
rrpc_inflight_laddr_release(rrpc, rqd);
goto finished;
}
wait_for_completion_io(&wait);
rrpc_inflight_laddr_release(rrpc, rqd);
if (bio->bi_error)
goto finished;
bio_reset(bio);
}
finished:
mempool_free(page, rrpc->page_pool);
bio_put(bio);
if (!bitmap_full(rblk->invalid_pages, nr_sec_per_blk)) {
pr_err("nvm: failed to garbage collect block\n");
return -EIO;
}
return 0;
}
static void rrpc_block_gc(struct work_struct *work)
{
struct rrpc_block_gc *gcb = container_of(work, struct rrpc_block_gc,
ws_gc);
struct rrpc *rrpc = gcb->rrpc;
struct rrpc_block *rblk = gcb->rblk;
struct rrpc_lun *rlun = rblk->rlun;
struct nvm_tgt_dev *dev = rrpc->dev;
struct ppa_addr ppa;
mempool_free(gcb, rrpc->gcb_pool);
pr_debug("nvm: block 'ch:%d,lun:%d,blk:%d' being reclaimed\n",
rlun->bppa.g.ch, rlun->bppa.g.lun,
rblk->id);
if (rrpc_move_valid_pages(rrpc, rblk))
goto put_back;
ppa.ppa = 0;
ppa.g.ch = rlun->bppa.g.ch;
ppa.g.lun = rlun->bppa.g.lun;
ppa.g.blk = rblk->id;
if (nvm_erase_blk(dev, &ppa, 0))
goto put_back;
rrpc_put_blk(rrpc, rblk);
return;
put_back:
spin_lock(&rlun->lock);
list_add_tail(&rblk->prio, &rlun->prio_list);
spin_unlock(&rlun->lock);
}
/* the block with highest number of invalid pages, will be in the beginning
* of the list
*/
static struct rrpc_block *rblk_max_invalid(struct rrpc_block *ra,
struct rrpc_block *rb)
{
if (ra->nr_invalid_pages == rb->nr_invalid_pages)
return ra;
return (ra->nr_invalid_pages < rb->nr_invalid_pages) ? rb : ra;
}
/* linearly find the block with highest number of invalid pages
* requires lun->lock
*/
static struct rrpc_block *block_prio_find_max(struct rrpc_lun *rlun)
{
struct list_head *prio_list = &rlun->prio_list;
struct rrpc_block *rblk, *max;
BUG_ON(list_empty(prio_list));
max = list_first_entry(prio_list, struct rrpc_block, prio);
list_for_each_entry(rblk, prio_list, prio)
max = rblk_max_invalid(max, rblk);
return max;
}
static void rrpc_lun_gc(struct work_struct *work)
{
struct rrpc_lun *rlun = container_of(work, struct rrpc_lun, ws_gc);
struct rrpc *rrpc = rlun->rrpc;
struct nvm_tgt_dev *dev = rrpc->dev;
struct rrpc_block_gc *gcb;
unsigned int nr_blocks_need;
nr_blocks_need = dev->geo.blks_per_lun / GC_LIMIT_INVERSE;
if (nr_blocks_need < rrpc->nr_luns)
nr_blocks_need = rrpc->nr_luns;
spin_lock(&rlun->lock);
while (nr_blocks_need > rlun->nr_free_blocks &&
!list_empty(&rlun->prio_list)) {
struct rrpc_block *rblk = block_prio_find_max(rlun);
if (!rblk->nr_invalid_pages)
break;
gcb = mempool_alloc(rrpc->gcb_pool, GFP_ATOMIC);
if (!gcb)
break;
list_del_init(&rblk->prio);
WARN_ON(!block_is_full(rrpc, rblk));
pr_debug("rrpc: selected block 'ch:%d,lun:%d,blk:%d' for GC\n",
rlun->bppa.g.ch, rlun->bppa.g.lun,
rblk->id);
gcb->rrpc = rrpc;
gcb->rblk = rblk;
INIT_WORK(&gcb->ws_gc, rrpc_block_gc);
queue_work(rrpc->kgc_wq, &gcb->ws_gc);
nr_blocks_need--;
}
spin_unlock(&rlun->lock);
/* TODO: Hint that request queue can be started again */
}
static void rrpc_gc_queue(struct work_struct *work)
{
struct rrpc_block_gc *gcb = container_of(work, struct rrpc_block_gc,
ws_gc);
struct rrpc *rrpc = gcb->rrpc;
struct rrpc_block *rblk = gcb->rblk;
struct rrpc_lun *rlun = rblk->rlun;
spin_lock(&rlun->lock);
list_add_tail(&rblk->prio, &rlun->prio_list);
spin_unlock(&rlun->lock);
mempool_free(gcb, rrpc->gcb_pool);
pr_debug("nvm: block 'ch:%d,lun:%d,blk:%d' full, allow GC (sched)\n",
rlun->bppa.g.ch, rlun->bppa.g.lun,
rblk->id);
}
static const struct block_device_operations rrpc_fops = {
.owner = THIS_MODULE,
};
static struct rrpc_lun *rrpc_get_lun_rr(struct rrpc *rrpc, int is_gc)
{
unsigned int i;
struct rrpc_lun *rlun, *max_free;
if (!is_gc)
return get_next_lun(rrpc);
/* during GC, we don't care about RR, instead we want to make
* sure that we maintain evenness between the block luns.
*/
max_free = &rrpc->luns[0];
/* prevent GC-ing lun from devouring pages of a lun with
* little free blocks. We don't take the lock as we only need an
* estimate.
*/
rrpc_for_each_lun(rrpc, rlun, i) {
if (rlun->nr_free_blocks > max_free->nr_free_blocks)
max_free = rlun;
}
return max_free;
}
static void sas_ata_task_done(struct sas_task *task)
{
struct ata_queued_cmd *qc = task->uldd_task;
struct domain_device *dev = task->dev;
struct task_status_struct *stat = &task->task_status;
struct ata_task_resp *resp = (struct ata_task_resp *)stat->buf;
struct sas_ha_struct *sas_ha = dev->port->ha;
enum ata_completion_errors ac;
unsigned long flags;
struct ata_link *link;
struct ata_port *ap;
spin_lock_irqsave(&dev->done_lock, flags);
if (test_bit(SAS_HA_FROZEN, &sas_ha->state))
task = NULL;
else if (qc && qc->scsicmd)
ASSIGN_SAS_TASK(qc->scsicmd, NULL);
spin_unlock_irqrestore(&dev->done_lock, flags);
/* check if libsas-eh got to the task before us */
if (unlikely(!task))
return;
if (!qc)
goto qc_already_gone;
ap = qc->ap;
link = &ap->link;
spin_lock_irqsave(ap->lock, flags);
/* check if we lost the race with libata/sas_ata_post_internal() */
if (unlikely(ap->pflags & ATA_PFLAG_FROZEN)) {
spin_unlock_irqrestore(ap->lock, flags);
if (qc->scsicmd)
goto qc_already_gone;
else {
/* if eh is not involved and the port is frozen then the
* ata internal abort process has taken responsibility
* for this sas_task
*/
return;
}
}
if (stat->stat == SAS_PROTO_RESPONSE || stat->stat == SAM_STAT_GOOD ||
((stat->stat == SAM_STAT_CHECK_CONDITION &&
dev->sata_dev.command_set == ATAPI_COMMAND_SET))) {
ata_tf_from_fis(resp->ending_fis, &dev->sata_dev.tf);
if (!link->sactive) {
qc->err_mask |= ac_err_mask(dev->sata_dev.tf.command);
} else {
link->eh_info.err_mask |= ac_err_mask(dev->sata_dev.tf.command);
if (unlikely(link->eh_info.err_mask))
qc->flags |= ATA_QCFLAG_FAILED;
}
} else {
ac = sas_to_ata_err(stat);
if (ac) {
SAS_DPRINTK("%s: SAS error %x\n", __func__,
stat->stat);
/* We saw a SAS error. Send a vague error. */
if (!link->sactive) {
qc->err_mask = ac;
} else {
link->eh_info.err_mask |= AC_ERR_DEV;
qc->flags |= ATA_QCFLAG_FAILED;
}
dev->sata_dev.tf.feature = 0x04; /* status err */
dev->sata_dev.tf.command = ATA_ERR;
}
}
qc->lldd_task = NULL;
ata_qc_complete(qc);
spin_unlock_irqrestore(ap->lock, flags);
qc_already_gone:
list_del_init(&task->list);
sas_free_task(task);
}
