Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S265691AbTGDCVt (ORCPT ); Thu, 3 Jul 2003 22:21:49 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S265591AbTGDCTO (ORCPT ); Thu, 3 Jul 2003 22:19:14 -0400 Received: from mta2.srv.hcvlny.cv.net ([167.206.5.5]:40103 "EHLO mta2.srv.hcvlny.cv.net") by vger.kernel.org with ESMTP id S265665AbTGDCRR (ORCPT ); Thu, 3 Jul 2003 22:17:17 -0400 Date: Thu, 03 Jul 2003 22:31:27 -0400 From: Jeff Sipek Subject: [PATCH - RFC] [1/5] 64-bit network statistics - generic net To: Kernel Mailing List Cc: Andrew Morton , Dave Jones , Linus Torvalds , Jeff Garzik Message-id: <200307032231.39842.jeffpc@optonline.net> MIME-version: 1.0 Content-type: multipart/mixed; boundary="Boundary_(ID_5+ztBL0+KOSUBUfAD6Oe6Q)" User-Agent: KMail/1.5.2 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 14688 Lines: 388 --Boundary_(ID_5+ztBL0+KOSUBUfAD6Oe6Q) Content-type: Text/Plain; charset=us-ascii Content-transfer-encoding: 7BIT Content-disposition: inline Content-description: clearsigned data -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The variables for network statistics (in struct net_device_stats) are unsigned longs. On 32-bit architectures, this makes them overflow every 4GB or 2^32 packets. The following series of patches [against 2.5.74] makes the statistics variable type configurable. The default is to leave everything the way it was (unsigned long). However, when NETSTATS64 is set in the config, the statistics use 64-bit variables (u_int64_t) - this works only on 32-bit architectures. These patches are *not* ready to be included due to the fact that they add an API layer that controls the statistics and all the drivers have to be checked. For easier coding, the patch renames the fields in struct net_device_stat - this breaks everything (so all the previous code is easily visible.) (The names are prefixed with _.) This is a RFC, please let me know what you think should be improved. *NOTE* I am aware that locking is suboptimal - I am looking for other solutions. The patches include: - Documentation/networking/64bitstats.txt with description of the API, and current status of this project (I'm not sure about some of the arch dependent suff) - updated drivers to use new API - 8139too - 8390 - dummy - eepro100 - loopback - ne2k-pci - pcnet32 - update for procfs and sysfs interfaces Thanks, Jeff. - -- You measure democracy by the freedom it gives its dissidents, not the freedom it gives its assimilated conformists. - Abbie Hoffman -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2 (GNU/Linux) iD8DBQE/BOcHwFP0+seVj/4RAs8dAJ9UqAyKB3ADE11o6Wo5kP2MYMUrVACeMWKw OiQPdQD85s9u+G4F/t0GraA= =aZf/ -----END PGP SIGNATURE----- --Boundary_(ID_5+ztBL0+KOSUBUfAD6Oe6Q) Content-type: text/x-diff; charset=us-ascii; name=net.diff Content-transfer-encoding: 7BIT Content-disposition: attachment; filename=net.diff diff -X dontdiff -Naur linux-2.5.74-vanilla/net/Kconfig linux-2.5.74-nx/net/Kconfig --- linux-2.5.74-vanilla/net/Kconfig 2003-07-02 16:40:31.000000000 -0400 +++ linux-2.5.74-nx/net/Kconfig 2003-07-03 15:11:00.000000000 -0400 @@ -48,6 +48,20 @@ mechanism that results in faster communication. If unsure, say N. + +config NETSTATS64 + bool "64-bit statistics" + depends on NET && (X86 || ARM || PARISC || SPARC32) && !(X86_64) + help + If you say Y here, the network statistics will use 64-bit values + no matter what architecture you are on. If you are on a 64-bit + architecture, say N here, the overhead required by this code is + not justifiable. If, however you are on a 32-bit architecture, + you have a choice. You can say: + Y and statistics will be 64-bit + N and statistics will remain 32-bit + + If unsure, say N. config NETLINK_DEV tristate "Netlink device emulation" diff -X dontdiff -Naur linux-2.5.74-vanilla/net/core/dev.c linux-2.5.74-nx/net/core/dev.c --- linux-2.5.74-vanilla/net/core/dev.c 2003-07-02 16:48:08.000000000 -0400 +++ linux-2.5.74-nx/net/core/dev.c 2003-07-03 15:11:00.000000000 -0400 @@ -18,6 +18,7 @@ * Alexey Kuznetsov * Adam Sulmicki * Pekka Riikonen + * Josef "Jeff" Sipek * * Changes: * D.J. Barrow : Fixed bug where dev->refcnt gets set @@ -70,6 +71,7 @@ * indefinitely on dev->refcnt * J Hadi Salim : - Backlog queue sampling * - netif_rx() feedback + * Josef "Jeff" Sipek : - using new API for network statistics */ #include @@ -1869,23 +1871,23 @@ struct net_device_stats *stats = dev->get_stats ? dev->get_stats(dev) : NULL; if (stats) - seq_printf(seq, "%6s:%8lu %7lu %4lu %4lu %4lu %5lu %10lu %9lu " - "%8lu %7lu %4lu %4lu %4lu %5lu %7lu %10lu\n", - dev->name, stats->rx_bytes, stats->rx_packets, - stats->rx_errors, - stats->rx_dropped + stats->rx_missed_errors, - stats->rx_fifo_errors, - stats->rx_length_errors + stats->rx_over_errors + - stats->rx_crc_errors + stats->rx_frame_errors, - stats->rx_compressed, stats->multicast, - stats->tx_bytes, stats->tx_packets, - stats->tx_errors, stats->tx_dropped, - stats->tx_fifo_errors, stats->collisions, - stats->tx_carrier_errors + - stats->tx_aborted_errors + - stats->tx_window_errors + - stats->tx_heartbeat_errors, - stats->tx_compressed); + seq_printf(seq, "%6s:%8llu %7llu %4llu %4llu %4llu %5llu %10llu %9llu " + "%8llu %7llu %4llu %4llu %4llu %5llu %7llu %10llu\n", + dev->name, net_stats_get(rx_bytes,stats), net_stats_get(rx_packets,stats), + net_stats_get(rx_errors,stats), + net_stats_get(rx_dropped,stats) + net_stats_get(rx_missed_errors,stats), + net_stats_get(rx_fifo_errors,stats), + net_stats_get(rx_length_errors,stats) + net_stats_get(rx_over_errors,stats) + + net_stats_get(rx_crc_errors,stats) + net_stats_get(rx_frame_errors,stats), + net_stats_get(rx_compressed,stats), net_stats_get(multicast,stats), + net_stats_get(tx_bytes,stats), net_stats_get(tx_packets,stats), + net_stats_get(tx_errors,stats), net_stats_get(tx_dropped,stats), + net_stats_get(tx_fifo_errors,stats), net_stats_get(collisions,stats), + net_stats_get(tx_carrier_errors,stats) + + net_stats_get(tx_aborted_errors,stats) + + net_stats_get(tx_window_errors,stats) + + net_stats_get(tx_heartbeat_errors,stats), + net_stats_get(tx_compressed,stats)); else seq_printf(seq, "%6s: No statistics available.\n", dev->name); } diff -X dontdiff -Naur linux-2.5.74-vanilla/net/core/net-sysfs.c linux-2.5.74-nx/net/core/net-sysfs.c --- linux-2.5.74-vanilla/net/core/net-sysfs.c 2003-07-02 16:55:46.000000000 -0400 +++ linux-2.5.74-nx/net/core/net-sysfs.c 2003-07-03 15:11:00.000000000 -0400 @@ -2,6 +2,9 @@ * net-sysfs.c - network device class and attributes * * Copyright (c) 2003 Stephen Hemminber + * + * Copyright (c) 2003 Josef "Jeff" Sipek + * - Now using the "New" Networks statistics API * */ @@ -180,21 +183,21 @@ struct netstat_fs_entry { struct attribute attr; - ssize_t (*show)(const struct net_device_stats *, char *); + ssize_t (*show)(struct net_device_stats *, char *); ssize_t (*store)(struct net_device_stats *, const char *, size_t); }; -static ssize_t net_device_stat_show(unsigned long var, char *buf) +static ssize_t net_device_stat_show(u_int64_t var, char *buf) { - return sprintf(buf, "%ld\n", var); + return sprintf(buf, "%llu\n", var); } /* generate a read-only statistics attribute */ #define NETDEVICE_STAT(_NAME) \ -static ssize_t show_stat_##_NAME(const struct net_device_stats *stats, \ +static ssize_t show_stat_##_NAME(struct net_device_stats *stats, \ char *buf) \ { \ - return net_device_stat_show(stats->_NAME, buf); \ + return net_device_stat_show(net_stats_get(_NAME,stats), buf); \ } \ static struct netstat_fs_entry net_stat_##_NAME = { \ .attr = {.name = __stringify(_NAME), .mode = S_IRUGO }, \ diff -X dontdiff -Naur linux-2.5.74-vanilla/include/linux/if_vlan.h linux-2.5.74-nx/include/linux/if_vlan.h --- linux-2.5.74-vanilla/include/linux/if_vlan.h 2003-07-02 16:42:17.000000000 -0400 +++ linux-2.5.74-nx/include/linux/if_vlan.h 2003-07-03 15:11:00.000000000 -0400 @@ -2,6 +2,7 @@ * VLAN An implementation of 802.1Q VLAN tagging. * * Authors: Ben Greear + * Josef "Jeff" Sipek * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License @@ -162,8 +163,8 @@ skb->dev->last_rx = jiffies; stats = vlan_dev_get_stats(skb->dev); - stats->rx_packets++; - stats->rx_bytes += skb->len; + net_stats_inc(rx_packets,stats); + net_stats_add(rx_bytes,stats,skb->len); skb->priority = vlan_get_ingress_priority(skb->dev, vlan_tag); switch (skb->pkt_type) { @@ -171,7 +172,7 @@ break; case PACKET_MULTICAST: - stats->multicast++; + net_stats_inc(multicast,stats); break; case PACKET_OTHERHOST: diff -X dontdiff -Naur linux-2.5.74-vanilla/include/linux/netdevice.h linux-2.5.74-nx/include/linux/netdevice.h --- linux-2.5.74-vanilla/include/linux/netdevice.h 2003-07-02 16:56:07.000000000 -0400 +++ linux-2.5.74-nx/include/linux/netdevice.h 2003-07-03 15:11:00.000000000 -0400 @@ -14,6 +14,7 @@ * Alan Cox, * Bjorn Ekwall. * Pekka Riikonen + * Josef "Jeff" Sipek * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License @@ -95,39 +96,75 @@ * with byte counters. */ +#ifdef CONFIG_NETSTATS64 +#define NETSTAT_TYPE u_int64_t +struct net_device_stats_locks +{ + spinlock_t rx_packets; + spinlock_t tx_packets; + spinlock_t rx_bytes; + spinlock_t tx_bytes; + spinlock_t rx_errors; + spinlock_t tx_errors; + spinlock_t rx_dropped; + spinlock_t tx_dropped; + spinlock_t multicast; + spinlock_t collisions; + spinlock_t rx_length_errors; + spinlock_t rx_over_errors; + spinlock_t rx_crc_errors; + spinlock_t rx_frame_errors; + spinlock_t rx_fifo_errors; + spinlock_t rx_missed_errors; + spinlock_t tx_aborted_errors; + spinlock_t tx_carrier_errors; + spinlock_t tx_fifo_errors; + spinlock_t tx_heartbeat_errors; + spinlock_t tx_window_errors; + spinlock_t rx_compressed; + spinlock_t tx_compressed; +}; +#else +#define NETSTAT_TYPE unsigned long +#endif + struct net_device_stats { - unsigned long rx_packets; /* total packets received */ - unsigned long tx_packets; /* total packets transmitted */ - unsigned long rx_bytes; /* total bytes received */ - unsigned long tx_bytes; /* total bytes transmitted */ - unsigned long rx_errors; /* bad packets received */ - unsigned long tx_errors; /* packet transmit problems */ - unsigned long rx_dropped; /* no space in linux buffers */ - unsigned long tx_dropped; /* no space available in linux */ - unsigned long multicast; /* multicast packets received */ - unsigned long collisions; + NETSTAT_TYPE _rx_packets; /* total packets received */ + NETSTAT_TYPE _tx_packets; /* total packets transmitted */ + NETSTAT_TYPE _rx_bytes; /* total bytes received */ + NETSTAT_TYPE _tx_bytes; /* total bytes transmitted */ + NETSTAT_TYPE _rx_errors; /* bad packets received */ + NETSTAT_TYPE _tx_errors; /* packet transmit problems */ + NETSTAT_TYPE _rx_dropped; /* no space in linux buffers */ + NETSTAT_TYPE _tx_dropped; /* no space available in linux */ + NETSTAT_TYPE _multicast; /* multicast packets received */ + NETSTAT_TYPE _collisions; /* detailed rx_errors: */ - unsigned long rx_length_errors; - unsigned long rx_over_errors; /* receiver ring buff overflow */ - unsigned long rx_crc_errors; /* recved pkt with crc error */ - unsigned long rx_frame_errors; /* recv'd frame alignment error */ - unsigned long rx_fifo_errors; /* recv'r fifo overrun */ - unsigned long rx_missed_errors; /* receiver missed packet */ + NETSTAT_TYPE _rx_length_errors; + NETSTAT_TYPE _rx_over_errors; /* receiver ring buff overflow */ + NETSTAT_TYPE _rx_crc_errors; /* recved pkt with crc error */ + NETSTAT_TYPE _rx_frame_errors; /* recv'd frame alignment error */ + NETSTAT_TYPE _rx_fifo_errors; /* recv'r fifo overrun */ + NETSTAT_TYPE _rx_missed_errors; /* receiver missed packet */ /* detailed tx_errors */ - unsigned long tx_aborted_errors; - unsigned long tx_carrier_errors; - unsigned long tx_fifo_errors; - unsigned long tx_heartbeat_errors; - unsigned long tx_window_errors; - + NETSTAT_TYPE _tx_aborted_errors; + NETSTAT_TYPE _tx_carrier_errors; + NETSTAT_TYPE _tx_fifo_errors; + NETSTAT_TYPE _tx_heartbeat_errors; + NETSTAT_TYPE _tx_window_errors; + /* for cslip etc */ - unsigned long rx_compressed; - unsigned long tx_compressed; -}; + NETSTAT_TYPE _rx_compressed; + NETSTAT_TYPE _tx_compressed; +#ifdef CONFIG_NETSTATS64 + /* locks */ + struct net_device_stats_locks locks; +#endif +}; /* Media selection options. */ enum { @@ -865,6 +902,46 @@ extern void dev_clear_fastroute(struct net_device *dev); #endif +static inline void net_stats_init(struct net_device_stats* stats) +{ +#ifdef CONFIG_NETSTATS64 + spin_lock_init(&stats->locks.rx_packets); + spin_lock_init(&stats->locks.tx_packets); + spin_lock_init(&stats->locks.rx_bytes); + spin_lock_init(&stats->locks.tx_bytes); + spin_lock_init(&stats->locks.rx_errors); + spin_lock_init(&stats->locks.tx_errors); + spin_lock_init(&stats->locks.rx_dropped); + spin_lock_init(&stats->locks.tx_dropped); + spin_lock_init(&stats->locks.multicast); + spin_lock_init(&stats->locks.collisions); + spin_lock_init(&stats->locks.rx_length_errors); + spin_lock_init(&stats->locks.rx_over_errors); + spin_lock_init(&stats->locks.rx_crc_errors); + spin_lock_init(&stats->locks.rx_frame_errors); + spin_lock_init(&stats->locks.rx_fifo_errors); + spin_lock_init(&stats->locks.rx_missed_errors); + spin_lock_init(&stats->locks.tx_aborted_errors); + spin_lock_init(&stats->locks.tx_carrier_errors); + spin_lock_init(&stats->locks.tx_fifo_errors); + spin_lock_init(&stats->locks.tx_heartbeat_errors); + spin_lock_init(&stats->locks.tx_window_errors); + spin_lock_init(&stats->locks.rx_compressed); + spin_lock_init(&stats->locks.tx_compressed); +#endif +} + +#ifdef CONFIG_NETSTATS64 +#define net_stats_add(what,stats,num) locked_add64(&(stats)->_##what,(u_int64_t) num,&(stats)->locks.what) +#define net_stats_inc(what,stats) locked_add64(&(stats)->_##what,1,&(stats)->locks.what) +#define net_stats_set(what,stats,num) locked_set64(&(stats)->_##what,(u_int64_t) num,&(stats)->locks.what) +#define net_stats_get(what,stats) locked_get64(&(stats)->_##what,&(stats)->locks.what) +#else +#define net_stats_add(what,stats,num) (stats)->_##what += (NETSTAT_TYPE) num +#define net_stats_inc(what,stats) ((stats)->_##what++) +#define net_stats_set(what,stats,num) (stats)->_##what = (NETSTAT_TYPE) num +#define net_stats_get(what,stats) ((u_int64_t) (stats)->_##what) +#endif #endif /* __KERNEL__ */ --Boundary_(ID_5+ztBL0+KOSUBUfAD6Oe6Q)-- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/