diff options
author | wpaul <wpaul@FreeBSD.org> | 1996-11-30 22:38:44 +0000 |
---|---|---|
committer | wpaul <wpaul@FreeBSD.org> | 1996-11-30 22:38:44 +0000 |
commit | 7c59df49d939e5f564cc78e9a17c7d2393e11420 (patch) | |
tree | 11a2a1fc5ef8902ed9155bf6cb194e478f34da19 /usr.sbin/ypserv/yp_server.c | |
parent | 697c32345ed353ef71779fcebe2ad32f2747a771 (diff) | |
download | FreeBSD-src-7c59df49d939e5f564cc78e9a17c7d2393e11420.zip FreeBSD-src-7c59df49d939e5f564cc78e9a17c7d2393e11420.tar.gz |
This commit changes the YPPROC_ALL procecdure so that it handles requests
_without_ using fork().
The problem with YPPROC_ALL is that it transmits an entire map through
a TCP pipe as the result of a single RPC call. First of all, this requires
certain hackery in the XDR filter. Second, if the map being sent is
large, the server can end up spending lots of time in the XDR filter
sending to just the one client, while requests for other clients will
go unanswered.
My original solution for this was to fork() the request into a child
process which terminates after the map has been transmitted (or the
transfer is interrupted due to an error). This leaves the parent free
to handle other requests. But this solution is kind of lame: fork()
is relatively expensive, and we have to keep a cap on the number of
child processes to keep from swamping the system.
What we do now is grab control of the service transport handle and XDR
handle from the RPC library and send the records one at a time ourselves
instead of letting the RPC library do it. We send a record, then go
back to the svc_run() loop and select() on the socket. If select() says
we can still write data, we send the next record. Then we call
svc_getreqset() and handle other RPCs and loop around again. This way,
we can handle other RPCs between records.
We manage multiple YPPROC_ALL requests using a circular queue. When a
request is done, we dequeue it and destroy the handle. We also tag
each request with a ttl which is decremented whevever we run the queue
and a handle isn't serviced. This lets us nuke requests that have sat
idle for too long (if we didn't do this, we might run out of socket
descriptors.)
Now all I have to do is come up with an async resolver, and ypserv
won't need to fork() at all. :)
Note: these changes should not go into 2.2 unless they get a very
throrough shakedown before the final cutoff date.
Diffstat (limited to 'usr.sbin/ypserv/yp_server.c')
-rw-r--r-- | usr.sbin/ypserv/yp_server.c | 66 |
1 files changed, 16 insertions, 50 deletions
diff --git a/usr.sbin/ypserv/yp_server.c b/usr.sbin/ypserv/yp_server.c index 9c9fe30..837c1cc 100644 --- a/usr.sbin/ypserv/yp_server.c +++ b/usr.sbin/ypserv/yp_server.c @@ -43,6 +43,7 @@ #include <netinet/in.h> #include <arpa/inet.h> #include <rpc/rpc.h> +#include <setjmp.h> #ifndef lint static const char rcsid[] = "$Id: yp_server.c,v 1.10 1996/05/31 16:01:51 wpaul Exp $"; @@ -51,10 +52,12 @@ static const char rcsid[] = "$Id: yp_server.c,v 1.10 1996/05/31 16:01:51 wpaul E int forked = 0; int children = 0; static DB *spec_dbp = NULL; /* Special global DB handle for ypproc_all. */ +static SVCXPRT *xprt; /* Special SVCXPRT handle for ypproc_all. */ static char *master_string = "YP_MASTER_NAME"; static char *order_string = "YP_LAST_MODIFIED"; static int master_sz = sizeof("YP_MASTER_NAME") - 1; static int order_sz = sizeof("YP_LAST_MODIFIED") - 1; +static jmp_buf env; /* * NIS v2 support. This is where most of the action happens. @@ -473,34 +476,19 @@ ypproc_clear_2_svc(void *argp, struct svc_req *rqstp) */ /* - * Custom XDR routine for serialzing results of ypproc_all: keep - * reading from the database and spew until we run out of records - * or encounter an error. + * Custom XDR routine for serialzing results of ypproc_all: grab control + * of the transport and xdr handle from the RPC library and this request + * to the async queue. It will multiplex the record transmission in such + * a way that we can service other requests between transmissions and + * avoid blocking. (It will also close the DB handle for us when the + * request is done.) */ static bool_t xdr_my_ypresp_all(register XDR *xdrs, ypresp_all *objp) { - DBT key = { NULL, 0 } , data = { NULL, 0 }; - - while (1) { - /* Get a record. */ - if ((objp->ypresp_all_u.val.stat = - yp_next_record(spec_dbp,&key,&data,1,0)) == YP_TRUE) { - objp->ypresp_all_u.val.val.valdat_len = data.size; - objp->ypresp_all_u.val.val.valdat_val = data.data; - objp->ypresp_all_u.val.key.keydat_len = key.size; - objp->ypresp_all_u.val.key.keydat_val = key.data; - objp->more = TRUE; - } else { - objp->more = FALSE; - } - - /* Serialize. */ - if (!xdr_ypresp_all(xdrs, objp)) - return(FALSE); - if (objp->more == FALSE) - return(TRUE); - } + if (yp_add_async(xdrs, xprt, spec_dbp) == FALSE) + return(FALSE); + longjmp(env, 1); /* XXX EVIL!! */ } ypresp_all * @@ -531,38 +519,16 @@ ypproc_all_2_svc(ypreq_nokey *argp, struct svc_req *rqstp) return (&result); } - /* - * The ypproc_all procedure can take a while to complete. - * Best to handle it in a subprocess so the parent doesn't - * block. (Is there a better way to do this? Maybe with - * async socket I/O?) - */ - if (!debug && children < MAX_CHILDREN && fork()) { - children++; - forked = 0; - return (NULL); - } else { - forked++; - } - -#ifndef DB_CACHE if ((spec_dbp = yp_open_db(argp->domain, argp->map)) == NULL) { result.ypresp_all_u.val.stat = yp_errno; return(&result); } -#else - if ((spec_dbp = yp_open_db_cache(argp->domain, argp->map, NULL, 0)) == NULL) { - result.ypresp_all_u.val.stat = yp_errno; - return(&result); - } -#endif /* Kick off the actual data transfer. */ - svc_sendreply(rqstp->rq_xprt, xdr_my_ypresp_all, (char *)&result); - -#ifndef DB_CACHE - (void)(spec_dbp->close)(spec_dbp); -#endif + xprt = rqstp->rq_xprt; + if (!setjmp(env)) /* XXX EVIL!!! */ + svc_sendreply(rqstp->rq_xprt, xdr_my_ypresp_all, + (char *)&result); /* * Returning NULL prevents the dispatcher from calling * svc_sendreply() since we already did it. |