#include <rdma/ib_verbs.h>
#include <rdma/rdma_cm.h>
+#include <linux/pci.h>
+#include <linux/slab.h>
#include "rds.h"
#include "rdma_transport.h"
#define RDS_FMR_SIZE 256
-#define RDS_FMR_POOL_SIZE 4096
+#define RDS_FMR_POOL_SIZE 8192
#define RDS_IB_MAX_SGE 8
#define RDS_IB_RECV_SGE 2
#define RDS_IB_SUPPORTED_PROTOCOLS 0x00000003 /* minor versions supported */
+#define RDS_IB_RECYCLE_BATCH_COUNT 32
+
+extern struct rw_semaphore rds_ib_devices_lock;
extern struct list_head rds_ib_devices;
/*
* try and minimize the amount of memory tied up both the device and
* socket receive queues.
*/
-/* page offset of the final full frag that fits in the page */
-#define RDS_PAGE_LAST_OFF (((PAGE_SIZE / RDS_FRAG_SIZE) - 1) * RDS_FRAG_SIZE)
struct rds_page_frag {
struct list_head f_item;
- struct page *f_page;
- unsigned long f_offset;
- dma_addr_t f_mapped;
+ struct list_head f_cache_entry;
+ struct scatterlist f_sg;
};
struct rds_ib_incoming {
struct list_head ii_frags;
+ struct list_head ii_cache_entry;
struct rds_incoming ii_inc;
};
+struct rds_ib_cache_head {
+ struct list_head *first;
+ unsigned long count;
+};
+
+struct rds_ib_refill_cache {
+ struct rds_ib_cache_head *percpu;
+ struct list_head *xfer;
+ struct list_head *ready;
+};
+
struct rds_ib_connect_private {
/* Add new fields at the end, and don't permute existing fields. */
__be32 dp_saddr;
};
struct rds_ib_send_work {
- struct rds_message *s_rm;
- struct rds_rdma_op *s_op;
+ void *s_op;
struct ib_send_wr s_wr;
struct ib_sge s_sge[RDS_IB_MAX_SGE];
unsigned long s_queued;
/* tx */
struct rds_ib_work_ring i_send_ring;
- struct rds_message *i_rm;
+ struct rm_data_op *i_data_op;
struct rds_header *i_send_hdrs;
u64 i_send_hdrs_dma;
struct rds_ib_send_work *i_sends;
+ atomic_t i_signaled_sends;
/* rx */
struct tasklet_struct i_recv_tasklet;
struct rds_header *i_recv_hdrs;
u64 i_recv_hdrs_dma;
struct rds_ib_recv_work *i_recvs;
- struct rds_page_frag i_frag;
u64 i_ack_recv; /* last ACK received */
+ struct rds_ib_refill_cache i_cache_incs;
+ struct rds_ib_refill_cache i_cache_frags;
/* sending acks */
unsigned long i_ack_flags;
/* Batched completions */
unsigned int i_unsignaled_wrs;
- long i_unsignaled_bytes;
};
/* This assumes that atomic_t is at least 32 bits */
unsigned int max_initiator_depth;
unsigned int max_responder_resources;
spinlock_t spinlock; /* protect the above */
+ atomic_t refcount;
+ struct work_struct free_work;
};
+#define pcidev_to_node(pcidev) pcibus_to_node(pcidev->bus)
+#define ibdev_to_node(ibdev) pcidev_to_node(to_pci_dev(ibdev->dma_device))
+#define rdsibdev_to_node(rdsibdev) ibdev_to_node(rdsibdev->dev)
+
/* bits for i_ack_flags */
#define IB_ACK_IN_FLIGHT 0
#define IB_ACK_REQUESTED 1
uint64_t s_ib_rdma_mr_pool_flush;
uint64_t s_ib_rdma_mr_pool_wait;
uint64_t s_ib_rdma_mr_pool_depleted;
+ uint64_t s_ib_atomic_cswp;
+ uint64_t s_ib_atomic_fadd;
};
extern struct workqueue_struct *rds_ib_wq;
extern struct rds_transport rds_ib_transport;
extern void rds_ib_add_one(struct ib_device *device);
extern void rds_ib_remove_one(struct ib_device *device);
+struct rds_ib_device *rds_ib_get_client_data(struct ib_device *device);
+void rds_ib_dev_put(struct rds_ib_device *rds_ibdev);
extern struct ib_client rds_ib_client;
extern unsigned int fmr_pool_size;
int rds_ib_conn_connect(struct rds_connection *conn);
void rds_ib_conn_shutdown(struct rds_connection *conn);
void rds_ib_state_change(struct sock *sk);
-int __init rds_ib_listen_init(void);
+int rds_ib_listen_init(void);
void rds_ib_listen_stop(void);
void __rds_ib_conn_error(struct rds_connection *conn, const char *, ...);
int rds_ib_cm_handle_connect(struct rdma_cm_id *cm_id,
int rds_ib_update_ipaddr(struct rds_ib_device *rds_ibdev, __be32 ipaddr);
void rds_ib_add_conn(struct rds_ib_device *rds_ibdev, struct rds_connection *conn);
void rds_ib_remove_conn(struct rds_ib_device *rds_ibdev, struct rds_connection *conn);
-void __rds_ib_destroy_conns(struct list_head *list, spinlock_t *list_lock);
-static inline void rds_ib_destroy_nodev_conns(void)
-{
- __rds_ib_destroy_conns(&ib_nodev_conns, &ib_nodev_conns_lock);
-}
-static inline void rds_ib_destroy_conns(struct rds_ib_device *rds_ibdev)
-{
- __rds_ib_destroy_conns(&rds_ibdev->conn_list, &rds_ibdev->spinlock);
-}
+void rds_ib_destroy_nodev_conns(void);
struct rds_ib_mr_pool *rds_ib_create_mr_pool(struct rds_ib_device *);
void rds_ib_get_mr_info(struct rds_ib_device *rds_ibdev, struct rds_info_rdma_connection *iinfo);
void rds_ib_destroy_mr_pool(struct rds_ib_mr_pool *);
void rds_ib_sync_mr(void *trans_private, int dir);
void rds_ib_free_mr(void *trans_private, int invalidate);
void rds_ib_flush_mrs(void);
+int rds_ib_fmr_init(void);
+void rds_ib_fmr_exit(void);
/* ib_recv.c */
-int __init rds_ib_recv_init(void);
+int rds_ib_recv_init(void);
void rds_ib_recv_exit(void);
int rds_ib_recv(struct rds_connection *conn);
-int rds_ib_recv_refill(struct rds_connection *conn, gfp_t kptr_gfp,
- gfp_t page_gfp, int prefill);
+int rds_ib_recv_alloc_caches(struct rds_ib_connection *ic);
+void rds_ib_recv_free_caches(struct rds_ib_connection *ic);
+void rds_ib_recv_refill(struct rds_connection *conn, int prefill);
void rds_ib_inc_free(struct rds_incoming *inc);
int rds_ib_inc_copy_to_user(struct rds_incoming *inc, struct iovec *iov,
size_t size);
void rds_ib_send_cq_comp_handler(struct ib_cq *cq, void *context);
void rds_ib_send_init_ring(struct rds_ib_connection *ic);
void rds_ib_send_clear_ring(struct rds_ib_connection *ic);
-int rds_ib_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op);
+int rds_ib_xmit_rdma(struct rds_connection *conn, struct rm_rdma_op *op);
void rds_ib_send_add_credits(struct rds_connection *conn, unsigned int credits);
void rds_ib_advertise_credits(struct rds_connection *conn, unsigned int posted);
int rds_ib_send_grab_credits(struct rds_ib_connection *ic, u32 wanted,
unsigned int avail);
/* ib_sysctl.c */
-int __init rds_ib_sysctl_init(void);
+int rds_ib_sysctl_init(void);
void rds_ib_sysctl_exit(void);
extern unsigned long rds_ib_sysctl_max_send_wr;
extern unsigned long rds_ib_sysctl_max_recv_wr;