Date: Wed, 20 Aug 2014 09:34:41 +0200 From: Hans Petter Selasky <hps@selasky.org> To: freebsd-net@freebsd.org, FreeBSD Current <freebsd-current@freebsd.org> Subject: [RFC] Add support for hardware transmit rate limiting queues [WAS: Add support for changing the flow ID of TCP connections] Message-ID: <53F44F91.2060006@selasky.org> In-Reply-To: <20140709163146.GA21731@ox> References: <53BC2E73.6090700@selasky.org> <53BC43AE.3040409@FreeBSD.org> <53BD5385.4090208@selasky.org> <20140709163146.GA21731@ox>
next in thread | previous in thread | raw e-mail | index | archive | help
This is a multi-part message in MIME format.
--------------080406030702000505070708
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Hi,
A month has passed since the last e-mail on this topic, and in the
meanwhile some new patches have been created and tested:
Basically the approach has been changed a little bit:
- The creation of hardware transmit rings has been made independent of
the TCP stack. This allows firewall applications to forward traffic into
hardware transmit rings aswell, and not only native TCP applications.
This should be one more reason to get the feature into the kernel.
- A hardware transmit ring basically can have two modes: FIXED-RATE or
AUTOMATIC-RATE. In the fixed rate mode all traffic is sent at a fixed
bytes per second rate. In the automatic mode you can configure a time
after which the TX queue must be empty. The hardware driver uses this to
configure the actual rate. In automatic mode you can also set an upper
and lower transmit rate limit.
- The MBUF has got a new field in the packet header: "txringid"
- IOCTLs for TCP v4 and v6 sockets has been updated to allow setting of
the "txringid" field in the mbuf.
The current patch [see attachment] should be much simpler and less
intrusive than the previous one.
Any comments ?
--HPS
--------------080406030702000505070708
Content-Type: text/x-diff;
name="net_ratectl.diff"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename="net_ratectl.diff"
=== sys/net/if.h
==================================================================
--- sys/net/if.h (revision 270138)
+++ sys/net/if.h (local)
@@ -239,6 +239,7 @@
#define IFCAP_RXCSUM_IPV6 0x200000 /* can offload checksum on IPv6 RX */
#define IFCAP_TXCSUM_IPV6 0x400000 /* can offload checksum on IPv6 TX */
#define IFCAP_HWSTATS 0x800000 /* manages counters internally */
+#define IFCAP_HWTXRINGS 0x1000000 /* hardware supports TX rings */
#define IFCAP_HWCSUM_IPV6 (IFCAP_RXCSUM_IPV6 | IFCAP_TXCSUM_IPV6)
=== sys/netinet/in.c
==================================================================
--- sys/netinet/in.c (revision 270138)
+++ sys/netinet/in.c (local)
@@ -42,6 +42,7 @@
#include <sys/malloc.h>
#include <sys/priv.h>
#include <sys/socket.h>
+#include <sys/socketvar.h>
#include <sys/jail.h>
#include <sys/kernel.h>
#include <sys/proc.h>
@@ -201,9 +202,23 @@
struct in_ifaddr *ia;
int error;
- if (ifp == NULL)
- return (EADDRNOTAVAIL);
+ if (ifp == NULL) {
+ struct inpcb *inp;
+ switch (cmd) {
+ case SIOCSTXRINGID:
+ inp = sotoinpcb(so);
+ if (inp == NULL)
+ return (EINVAL);
+ INP_WLOCK(inp);
+ inp->inp_txringid = *(unsigned *)data;
+ INP_WUNLOCK(inp);
+ return (0);
+ default:
+ return (EADDRNOTAVAIL);
+ }
+ }
+
/*
* Filter out 4 ioctls we implement directly. Forward the rest
* to specific functions and ifp->if_ioctl().
=== sys/netinet/in_pcb.h
==================================================================
--- sys/netinet/in_pcb.h (revision 270138)
+++ sys/netinet/in_pcb.h (local)
@@ -46,6 +46,7 @@
#ifdef _KERNEL
#include <sys/lock.h>
#include <sys/rwlock.h>
+#include <sys/mbuf.h>
#include <net/vnet.h>
#include <vm/uma.h>
#endif
@@ -177,7 +178,8 @@
u_char inp_ip_ttl; /* (i) time to live proto */
u_char inp_ip_p; /* (c) protocol proto */
u_char inp_ip_minttl; /* (i) minimum TTL or drop */
- uint32_t inp_flowid; /* (x) flow id / queue id */
+ m_flowid_t inp_flowid; /* (x) flow ID */
+ m_txringid_t inp_txringid; /* (x) transmit ring ID */
u_int inp_refcount; /* (i) refcount */
void *inp_pspare[5]; /* (x) route caching / general use */
uint32_t inp_flowtype; /* (x) M_HASHTYPE value */
=== sys/netinet/in_var.h
==================================================================
--- sys/netinet/in_var.h (revision 270138)
+++ sys/netinet/in_var.h (local)
@@ -33,6 +33,7 @@
#ifndef _NETINET_IN_VAR_H_
#define _NETINET_IN_VAR_H_
+#include <sys/mbuf.h>
#include <sys/queue.h>
#include <sys/fnv_hash.h>
#include <sys/tree.h>
@@ -81,6 +82,18 @@
struct sockaddr_in ifra_mask;
int ifra_vhid;
};
+
+struct in_ratectlreq {
+ char ifreq_name[IFNAMSIZ];
+ m_txringid_t tx_ring_id;
+ uint32_t min_bytes_per_interval;
+ uint32_t max_bytes_per_interval;
+ uint32_t micros_per_interval;
+ uint32_t mode;
+#define IN_RATECTLREQ_MODE_FIXED 0 /* min rate = max rate */
+#define IN_RATECTLREQ_MODE_AUTOMATIC 1 /* bounded by min/max */
+};
+
/*
* Given a pointer to an in_ifaddr (ifaddr),
* return a pointer to the addr as a sockaddr_in.
=== sys/netinet/ip_output.c
==================================================================
--- sys/netinet/ip_output.c (revision 270138)
+++ sys/netinet/ip_output.c (local)
@@ -145,6 +145,7 @@
if (inp != NULL) {
INP_LOCK_ASSERT(inp);
M_SETFIB(m, inp->inp_inc.inc_fibnum);
+ m->m_pkthdr.txringid = inp->inp_txringid;
if (inp->inp_flags & (INP_HW_FLOWID|INP_SW_FLOWID)) {
m->m_pkthdr.flowid = inp->inp_flowid;
M_HASHTYPE_SET(m, inp->inp_flowtype);
=== sys/netinet6/in6.c
==================================================================
--- sys/netinet6/in6.c (revision 270138)
+++ sys/netinet6/in6.c (local)
@@ -235,6 +235,23 @@
int error;
u_long ocmd = cmd;
+ if (ifp == NULL) {
+ struct inpcb *inp;
+
+ switch (cmd) {
+ case SIOCSTXRINGID:
+ inp = sotoinpcb(so);
+ if (inp == NULL)
+ return (EINVAL);
+ INP_WLOCK(inp);
+ inp->inp_txringid = *(unsigned *)data;
+ INP_WUNLOCK(inp);
+ return (0);
+ default:
+ break;
+ }
+ }
+
/*
* Compat to make pre-10.x ifconfig(8) operable.
*/
=== sys/sys/mbuf.h
==================================================================
--- sys/sys/mbuf.h (revision 270138)
+++ sys/sys/mbuf.h (local)
@@ -114,6 +114,10 @@
void (*m_tag_free)(struct m_tag *);
};
+typedef uint32_t m_flowid_t;
+typedef uint32_t m_txringid_t;
+#define M_TXRINGID_UNDEFINED 0
+
/*
* Record/packet header in first mbuf of chain; valid only if M_PKTHDR is set.
* Size ILP32: 48
@@ -125,7 +129,8 @@
int32_t len; /* total packet length */
/* Layer crossing persistent information. */
- uint32_t flowid; /* packet's 4-tuple system */
+ m_flowid_t flowid; /* packet's 4-tuple system */
+ m_txringid_t txringid; /* transmit ring ID */
uint64_t csum_flags; /* checksum and offload features */
uint16_t fibnum; /* this packet should use this fib */
uint8_t cosqos; /* class/quality of service */
=== sys/sys/sockio.h
==================================================================
--- sys/sys/sockio.h (revision 270138)
+++ sys/sys/sockio.h (local)
@@ -43,6 +43,7 @@
#define SIOCATMARK _IOR('s', 7, int) /* at oob mark? */
#define SIOCSPGRP _IOW('s', 8, int) /* set process group */
#define SIOCGPGRP _IOR('s', 9, int) /* get process group */
+#define SIOCSTXRINGID _IOW('s', 10, unsigned) /* set transmit ring ID */
/* SIOCADDRT _IOW('r', 10, struct ortentry) 4.3BSD */
/* SIOCDELRT _IOW('r', 11, struct ortentry) 4.3BSD */
@@ -128,4 +129,9 @@
#define SIOCDIFGROUP _IOW('i', 137, struct ifgroupreq) /* delete ifgroup */
#define SIOCGIFGMEMB _IOWR('i', 138, struct ifgroupreq) /* get members */
+#define SIOCARATECTL _IOWR('i', 139, struct in_ratectlreq) /* add new new rate control HW ring */
+#define SIOCSRATECTL _IOWR('i', 140, struct in_ratectlreq) /* set parameters for existing HW ring */
+#define SIOCGRATECTL _IOWR('i', 141, struct in_ratectlreq) /* get parameters for existing HW ring */
+#define SIOCDRATECTL _IOW('i', 142, struct in_ratectlreq) /* delete existing HW ring */
+
#endif /* !_SYS_SOCKIO_H_ */
--------------080406030702000505070708--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?53F44F91.2060006>