MEDIUM: stream-int/conn-stream: Handle I/O subscriptions in the conn-stream

wait_event structure is moved in the conn-stream. The tasklet is only
created if the conn-stream is attached to a mux and released when the mux is
detached. This implies a subtle change. In stream_int_chk_rcv() function,
the wakeup of the tasklet was removed because there is no longer tasklet at
this stage (stream_int_chk_rcv() is a callback function of si_embedded_ops).
This commit is contained in:
Christopher Faulet 2022-03-31 11:09:28 +02:00
parent 070b91bc11
commit 2f35e7b6ab
6 changed files with 55 additions and 34 deletions

View File

@ -157,6 +157,7 @@ struct conn_stream {
unsigned int flags; /* CS_FL_* */
unsigned int hcto; /* half-closed timeout (0 = unset) */
struct wait_event wait_event; /* We're in a wait list */
struct cs_endpoint *endp; /* points to the end point (MUX stream or appctx) */
enum obj_type *app; /* points to the applicative point (stream or check) */
struct stream_interface *si;

View File

@ -63,8 +63,6 @@ struct stream_interface {
unsigned int flags; /* SI_FL_* */
struct conn_stream *cs; /* points to the conn-streams that owns the endpoint (connection or applet) */
struct si_ops *ops; /* general operations at the stream interface layer */
struct wait_event wait_event; /* We're in a wait list */
};
/* operations available on a stream-interface */

View File

@ -108,12 +108,6 @@ static inline int si_init(struct stream_interface *si)
si->flags &= SI_FL_ISBACK;
si->cs = NULL;
si->ops = &si_embedded_ops;
si->wait_event.tasklet = tasklet_new();
if (!si->wait_event.tasklet)
return -1;
si->wait_event.tasklet->process = si_cs_io_cb;
si->wait_event.tasklet->context = si;
si->wait_event.events = 0;
return 0;
}

View File

@ -64,6 +64,9 @@ struct conn_stream *cs_new(struct cs_endpoint *endp)
cs->data_cb = NULL;
cs->src = NULL;
cs->dst = NULL;
cs->wait_event.tasklet = NULL;
cs->wait_event.events = 0;
if (!endp) {
endp = cs_endpoint_new();
if (unlikely(!endp))
@ -156,6 +159,8 @@ void cs_free(struct conn_stream *cs)
BUG_ON(!(cs->endp->flags & CS_EP_DETACHED));
cs_endpoint_free(cs->endp);
}
if (cs->wait_event.tasklet)
tasklet_free(cs->wait_event.tasklet);
pool_free(pool_head_connstream, cs);
}
@ -172,6 +177,15 @@ int cs_attach_mux(struct conn_stream *cs, void *target, void *ctx)
if (!conn->ctx)
conn->ctx = cs;
if (cs_strm(cs)) {
if (!cs->wait_event.tasklet) {
cs->wait_event.tasklet = tasklet_new();
if (!cs->wait_event.tasklet)
return -1;
cs->wait_event.tasklet->process = si_cs_io_cb;
cs->wait_event.tasklet->context = cs->si;
cs->wait_event.events = 0;
}
cs->si->ops = &si_conn_ops;
cs->data_cb = &si_conn_cb;
}
@ -204,8 +218,19 @@ int cs_attach_strm(struct conn_stream *cs, struct stream *strm)
cs->si = si_new(cs);
if (unlikely(!cs->si))
return -1;
cs->endp->flags &= ~CS_EP_ORPHAN;
if (cs->endp->flags & CS_EP_T_MUX) {
cs->wait_event.tasklet = tasklet_new();
if (!cs->wait_event.tasklet) {
si_free(cs->si);
cs->si = NULL;
return -1;
}
cs->wait_event.tasklet->process = si_cs_io_cb;
cs->wait_event.tasklet->context = cs->si;
cs->wait_event.events = 0;
cs->si->ops = &si_conn_ops;
cs->data_cb = &si_conn_cb;
}
@ -237,8 +262,8 @@ void cs_detach_endp(struct conn_stream *cs)
if (conn->mux) {
/* TODO: handle unsubscribe for healthchecks too */
cs->endp->flags |= CS_EP_ORPHAN;
if (cs->si && cs->si->wait_event.events != 0)
conn->mux->unsubscribe(cs, cs->si->wait_event.events, &cs->si->wait_event);
if (cs->wait_event.events != 0)
conn->mux->unsubscribe(cs, cs->wait_event.events, &cs->wait_event);
conn->mux->detach(cs);
cs->endp = NULL;
}
@ -290,6 +315,12 @@ void cs_detach_app(struct conn_stream *cs)
cs->data_cb = NULL;
sockaddr_free(&cs->src);
sockaddr_free(&cs->dst);
if (cs->wait_event.tasklet)
tasklet_free(cs->wait_event.tasklet);
cs->wait_event.tasklet = NULL;
cs->wait_event.events = 0;
if (!cs->endp || (cs->endp->flags & CS_EP_DETACHED))
cs_free(cs);
}

View File

@ -1502,9 +1502,8 @@ int stream_set_http_mode(struct stream *s, const struct mux_proto_list *mux_prot
* mux will probably want to subscribe to
* the underlying XPRT
*/
if (cs_si(s->csf)->wait_event.events)
conn->mux->unsubscribe(cs, cs_si(s->csf)->wait_event.events,
&(cs_si(s->csf)->wait_event));
if (s->csf->wait_event.events)
conn->mux->unsubscribe(cs, s->csf->wait_event.events, &(s->csf->wait_event));
if (conn->mux->flags & MX_FL_NO_UPG)
return 0;
@ -3278,22 +3277,22 @@ static int stats_dump_full_strm_to_buffer(struct conn_stream *cs, struct stream
strm->txn->req.flags, strm->txn->rsp.flags);
chunk_appendf(&trash,
" si[0]=%p (flags=0x%02x endp0=%s:%p sub=%d)\n",
" si[0]=%p (flags=0x%02x endp0=%s:%p)\n",
strm->csf->si,
strm->csf->si->flags,
(strm->csf->endp->flags & CS_EP_T_MUX ? "CONN" : "APPCTX"),
__cs_endp_target(strm->csf), strm->csf->si->wait_event.events);
__cs_endp_target(strm->csf));
chunk_appendf(&trash,
" si[1]=%p (flags=0x%02x endp1=%s:%p sub=%d)\n",
" si[1]=%p (flags=0x%02x endp1=%s:%p)\n",
strm->csb->si,
strm->csb->si->flags,
(strm->csb->endp->flags & CS_EP_T_MUX ? "CONN" : "APPCTX"),
__cs_endp_target(strm->csb), strm->csb->si->wait_event.events);
__cs_endp_target(strm->csb));
csf = strm->csf;
chunk_appendf(&trash, " cs=%p csf=0x%08x state=%s endp=%p,0x%08x\n", csf, csf->flags,
cs_state_str(csf->state), csf->endp->target, csf->endp->flags);
chunk_appendf(&trash, " cs=%p csf=0x%08x state=%s endp=%p,0x%08x sub=%d\n", csf, csf->flags,
cs_state_str(csf->state), csf->endp->target, csf->endp->flags, csf->wait_event.events);
if ((conn = cs_conn(csf)) != NULL) {
chunk_appendf(&trash,
@ -3329,8 +3328,8 @@ static int stats_dump_full_strm_to_buffer(struct conn_stream *cs, struct stream
}
csb = strm->csb;
chunk_appendf(&trash, " cs=%p csb=0x%08x state=%s endp=%p,0x%08x\n", csb, csb->flags,
cs_state_str(csb->state), csb->endp->target, csb->endp->flags);
chunk_appendf(&trash, " cs=%p csb=0x%08x state=%s endp=%p,0x%08x sub=%d\n", csb, csb->flags,
cs_state_str(csb->state), csb->endp->target, csb->endp->flags, csb->wait_event.events);
if ((conn = cs_conn(csb)) != NULL) {
chunk_appendf(&trash,
" co1=%p ctrl=%s xprt=%s mux=%s data=%s target=%s:%p\n",

View File

@ -124,7 +124,6 @@ void si_free(struct stream_interface *si)
if (!si)
return;
tasklet_free(si->wait_event.tasklet);
pool_free(pool_head_streaminterface, si);
}
@ -235,7 +234,6 @@ static void stream_int_chk_rcv(struct stream_interface *si)
}
else {
/* (re)start reading */
tasklet_wakeup(si->wait_event.tasklet);
if (!(si->cs->flags & CS_FL_DONT_WAKE))
task_wakeup(si_task(si), TASK_WOKEN_IO);
}
@ -555,7 +553,7 @@ static int si_cs_process(struct conn_stream *cs)
BUG_ON(!conn);
/* If we have data to send, try it now */
if (!channel_is_empty(oc) && !(si->wait_event.events & SUB_RETRY_SEND))
if (!channel_is_empty(oc) && !(si->cs->wait_event.events & SUB_RETRY_SEND))
si_cs_send(cs);
/* First step, report to the conn-stream what was detected at the
@ -658,7 +656,7 @@ static int si_cs_send(struct conn_stream *cs)
}
/* We're already waiting to be able to send, give up */
if (si->wait_event.events & SUB_RETRY_SEND)
if (si->cs->wait_event.events & SUB_RETRY_SEND)
return 0;
/* we might have been called just after an asynchronous shutw */
@ -773,11 +771,11 @@ static int si_cs_send(struct conn_stream *cs)
/* We couldn't send all of our data, let the mux know we'd like to send more */
if (!channel_is_empty(oc))
conn->mux->subscribe(cs, SUB_RETRY_SEND, &si->wait_event);
conn->mux->subscribe(cs, SUB_RETRY_SEND, &si->cs->wait_event);
return did_send;
}
/* This is the ->process() function for any stream-interface's wait_event task.
/* This is the ->process() function for any conn-stream's wait_event task.
* It's assigned during the stream-interface's initialization, for any type of
* stream interface. Thus it is always safe to perform a tasklet_wakeup() on a
* stream interface, as the presence of the CS is checked there.
@ -791,9 +789,9 @@ struct task *si_cs_io_cb(struct task *t, void *ctx, unsigned int state)
if (!cs_conn(cs))
return t;
if (!(si->wait_event.events & SUB_RETRY_SEND) && !channel_is_empty(si_oc(si)))
if (!(cs->wait_event.events & SUB_RETRY_SEND) && !channel_is_empty(si_oc(si)))
ret = si_cs_send(cs);
if (!(si->wait_event.events & SUB_RETRY_RECV))
if (!(cs->wait_event.events & SUB_RETRY_RECV))
ret |= si_cs_recv(cs);
if (ret != 0)
si_cs_process(cs);
@ -909,7 +907,7 @@ int si_sync_recv(struct stream_interface *si)
if (!cs_conn_mux(si->cs))
return 0; // only conn_streams are supported
if (si->wait_event.events & SUB_RETRY_RECV)
if (si->cs->wait_event.events & SUB_RETRY_RECV)
return 0; // already subscribed
if (!si_rx_endp_ready(si) || si_rx_blocked(si))
@ -1111,7 +1109,7 @@ static void stream_int_chk_rcv_conn(struct stream_interface *si)
{
/* (re)start reading */
if (cs_state_in(si->cs->state, CS_SB_CON|CS_SB_RDY|CS_SB_EST))
tasklet_wakeup(si->wait_event.tasklet);
tasklet_wakeup(si->cs->wait_event.tasklet);
}
@ -1138,7 +1136,7 @@ static void stream_int_chk_snd_conn(struct stream_interface *si)
!(si->flags & SI_FL_WAIT_DATA)) /* not waiting for data */
return;
if (!(si->wait_event.events & SUB_RETRY_SEND) && !channel_is_empty(si_oc(si)))
if (!(si->cs->wait_event.events & SUB_RETRY_SEND) && !channel_is_empty(si_oc(si)))
si_cs_send(cs);
if (cs->endp->flags & (CS_EP_ERROR|CS_EP_ERR_PENDING) || si_is_conn_error(si)) {
@ -1232,7 +1230,7 @@ static int si_cs_recv(struct conn_stream *cs)
/* If another call to si_cs_recv() failed, and we subscribed to
* recv events already, give up now.
*/
if (si->wait_event.events & SUB_RETRY_RECV)
if (si->cs->wait_event.events & SUB_RETRY_RECV)
return 0;
/* maybe we were called immediately after an asynchronous shutr */
@ -1532,7 +1530,7 @@ static int si_cs_recv(struct conn_stream *cs)
}
else if (!si_rx_blocked(si)) {
/* Subscribe to receive events if we're blocking on I/O */
conn->mux->subscribe(cs, SUB_RETRY_RECV, &si->wait_event);
conn->mux->subscribe(cs, SUB_RETRY_RECV, &si->cs->wait_event);
si_rx_endp_done(si);
} else {
si_rx_endp_more(si);