[sr-dev] git:andrei/tcp_tls_changes: tls: added a minimum overhead shm buffer queue
Andrei Pelinescu-Onciul
andrei at iptel.org
Thu May 20 17:14:22 CEST 2010
Module: sip-router
Branch: andrei/tcp_tls_changes
Commit: c914809c9f70fe15d1a2459fa0668d722d83726c
URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=c914809c9f70fe15d1a2459fa0668d722d83726c
Author: Andrei Pelinescu-Onciul <andrei at iptel.org>
Committer: Andrei Pelinescu-Onciul <andrei at iptel.org>
Date: Thu May 20 16:22:16 2010 +0200
tls: added a minimum overhead shm buffer queue
Minimum overhead buffer queue in shm memory, based on
tcp_wbuffer_queue (tcp_conn.h).
---
modules/tls/sbufq.h | 292 +++++++++++++++++++++++++++++++++++++++++++++++++++
1 files changed, 292 insertions(+), 0 deletions(-)
diff --git a/modules/tls/sbufq.h b/modules/tls/sbufq.h
new file mode 100644
index 0000000..6b59751
--- /dev/null
+++ b/modules/tls/sbufq.h
@@ -0,0 +1,292 @@
+/*
+ * $Id$
+ *
+ * Copyright (C) 2010 iptelorg GmbH
+ *
+ * Permission to use, copy, modify, and distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+/** minimal overhead buffer queue in shm memory.
+ * @file modules/tls/sbufq.h
+ * @ingroup: tls
+ * Module: @ref tls
+ */
+/*
+ * History:
+ * --------
+ * 2010-03-31 initial version, based on tcp_conn.h tcp_wbuffer_queue (andrei)
+*/
+
+#ifndef __sbufq_h
+#define __sbufq_h
+
+#include "../../compiler_opt.h"
+#include "../../ut.h"
+#include "../../mem/shm_mem.h"
+#include "../../timer_ticks.h"
+#include "../../timer.h"
+#include "../../dprint.h"
+#include <string.h>
+
+
+struct sbuf_elem {
+ struct sbuf_elem* next;
+ unsigned int b_size; /**< buf size */
+ char buf[1]; /**< variable size buffer */
+};
+
+struct sbuffer_queue {
+ struct sbuf_elem* first;
+ struct sbuf_elem* last;
+ ticks_t last_chg; /**< last change (creation time or partial flush)*/
+ unsigned int queued; /**< total size */
+ unsigned int offset; /**< offset in the first buffer where unflushed data
+ starts */
+ unsigned int last_used; /**< how much of the last buffer is used */
+};
+
+
+/* sbufq_flush() output flags */
+#define F_BUFQ_EMPTY 1
+#define F_BUFQ_ERROR_FLUSH 2
+
+
+#define sbufq_empty(bq) ((bq)->first==0)
+#define sbufq_non_empty(bq) ((bq)->first!=0)
+
+
+
+/** adds/appends data to a buffer queue.
+ * WARNING: it does no attempt to synchronize access/lock. If needed it should
+ * be called under lock.
+ * @param q - buffer queue
+ * @param data
+ * @param size
+ * @param min_buf_size - min size to allocate for new buffer elements
+ * @return 0 on success, -1 on error (mem. allocation)
+ */
+inline static int sbufq_add(struct sbuffer_queue* q, const void* data,
+ unsigned int size, unsigned int min_buf_size)
+{
+ struct sbuf_elem* b;
+ unsigned int last_free;
+ unsigned int b_size;
+ unsigned int crt_size;
+ ticks_t t;
+
+ t=get_ticks_raw();
+
+ if (likely(q->last==0)) {
+ b_size=MAX_unsigned(min_buf_size, size);
+ b=shm_malloc(sizeof(*b)+b_size-sizeof(b->buf));
+ if (unlikely(b==0))
+ goto error;
+ b->b_size=b_size;
+ b->next=0;
+ q->last=b;
+ q->first=b;
+ q->last_used=0;
+ q->offset=0;
+ q->last_chg=get_ticks_raw();
+ last_free=b_size;
+ crt_size=size;
+ goto data_cpy;
+ }else{
+ b=q->last;
+ }
+
+ while(size){
+ last_free=b->b_size-q->last_used;
+ if (last_free==0){
+ b_size=MAX_unsigned(min_buf_size, size);
+ b=shm_malloc(sizeof(*b)+b_size-sizeof(b->buf));
+ if (unlikely(b==0))
+ goto error;
+ b->b_size=b_size;
+ b->next=0;
+ q->last->next=b;
+ q->last=b;
+ q->last_used=0;
+ last_free=b->b_size;
+ }
+ crt_size=MIN_unsigned(last_free, size);
+data_cpy:
+ memcpy(b->buf+q->last_used, data, crt_size);
+ q->last_used+=crt_size;
+ size-=crt_size;
+ data+=crt_size;
+ q->queued+=crt_size;
+ }
+ return 0;
+error:
+ return -1;
+}
+
+
+
+/** inserts data (at the beginning) in a buffer queue.
+ * Note: should never be called after sbufq_run().
+ * WARNING: it does no attempt to synchronize access/lock. If needed it should
+ * be called under lock.
+ * @param q - buffer queue
+ * @param data
+ * @param size
+ * @param min_buf_size - min size to allocate for new buffer elements
+ * @return 0 on success, -1 on error (mem. allocation)
+ */
+inline static int sbufq_insert(struct sbuffer_queue* q, const void* data,
+ unsigned int size, unsigned int min_buf_size)
+{
+ struct sbuf_elem* b;
+
+ if (likely(q->first==0)) /* if empty, use sbufq_add */
+ return sbufq_add(q, data, size, min_buf_size);
+
+ if (unlikely(q->offset)){
+ LOG(L_CRIT, "BUG: non-null offset %d (bad call, should"
+ "never be called after sbufq_run())\n", q->offset);
+ goto error;
+ }
+ if ((q->first==q->last) && ((q->last->b_size-q->last_used)>=size)){
+ /* one block with enough space in it for size bytes */
+ memmove(q->first->buf+size, q->first->buf, size);
+ memcpy(q->first->buf, data, size);
+ q->last_used+=size;
+ }else{
+ /* create a size bytes block directly */
+ b=shm_malloc(sizeof(*b)+size-sizeof(b->buf));
+ if (unlikely(b==0))
+ goto error;
+ b->b_size=size;
+ /* insert it */
+ b->next=q->first;
+ q->first=b;
+ memcpy(b->buf, data, size);
+ }
+
+ q->queued+=size;
+ return 0;
+error:
+ return -1;
+}
+
+
+/** destroy a buffer queue.
+ * Only the content is destroyed (shm_free()'d), the queue head is
+ * re-intialized.
+ * WARNING: it does no attempt to synchronize access/lock. If needed it should
+ * be called under lock.
+ * @param q - buffer queue
+ * @return - number of bytes that used to be queued (>=0).
+ */
+inline static unsigned int sbufq_destroy(struct sbuffer_queue* q)
+{
+ struct sbuf_elem* b;
+ struct sbuf_elem* next_b;
+ int unqueued;
+
+ unqueued=0;
+ if (likely(q->first)){
+ b=q->first;
+ do{
+ next_b=b->next;
+ unqueued+=(b==q->last)?q->last_used:b->b_size;
+ if (b==q->first)
+ unqueued-=q->offset;
+ shm_free(b);
+ b=next_b;
+ }while(b);
+ }
+ memset(q, 0, sizeof(*q));
+ return unqueued;
+}
+
+
+
+/** tries to flush the queue.
+ * Tries to flush as much as possible from the given queue, using the
+ * given callback.
+ * WARNING: it does no attempt to synchronize access/lock. If needed it should
+ * be called under lock.
+ * @param q - buffer queue
+ * @param *flags - set to:
+ * F_BUFQ_EMPTY if the queued is completely flushed
+ * F_BUFQ_ERROR_FLUSH if the flush_f callback returned error.
+ * @param flush_f - flush function (callback). modeled after write():
+ * flush_f(param1, param2, const void* buf, unsigned size).
+ * It should return the number of bytes "flushed" on
+ * success, or <0 on error. If the number of bytes
+ * "flushed" is smaller then the requested size, it
+ * would be assumed that no more bytes can be flushed
+ * and sbufq_flush will exit.
+ * @param flush_p1 - parameter for the flush function callback.
+ * @param flush_p2 - parameter for the flush function callback.
+ * @return -1 on internal error, or the number of bytes flushed on
+ * success (>=0). Note that the flags param is
+ * always set and it should be used to check for errors, since
+ * a flush_f() failure will not result in a negative return.
+ */
+inline static int sbufq_flush(struct sbuffer_queue* q, int* flags,
+ int (*flush_f)(void* p1, void* p2,
+ const void* buf,
+ unsigned size),
+ void* flush_p1, void* flush_p2)
+{
+ struct sbuf_elem *b;
+ int n;
+ int ret;
+ int block_size;
+ char* buf;
+
+ *flags=0;
+ ret=0;
+ while(q->first){
+ block_size=((q->first==q->last)?q->last_used:q->first->b_size)-
+ q->offset;
+ buf=q->first->buf+q->offset;
+ n=flush_f(flush_p1, flush_p2, buf, block_size);
+ if (likely(n>0)){
+ ret+=n;
+ if (likely(n==block_size)){
+ b=q->first;
+ q->first=q->first->next;
+ shm_free(b);
+ q->offset=0;
+ q->queued-=block_size;
+ ret+=block_size;
+ }else{
+ q->offset+=n;
+ q->queued-=n;
+ ret+=n;
+ break;
+ }
+ }else{
+ if (unlikely(n<0))
+ *flags|=F_BUFQ_ERROR_FLUSH;
+ break;
+ }
+ }
+ if (likely(q->first==0)){
+ q->last=0;
+ q->last_used=0;
+ q->offset=0;
+ *flags|=F_BUFQ_EMPTY;
+ }
+ return ret;
+}
+
+
+
+
+#endif /*__sbufq_h*/
+
+/* vi: set ts=4 sw=4 tw=79:ai:cindent: */
More information about the sr-dev
mailing list