iftop-users

iftop-users@lists.beasts.org

195 discussions

iftop is overloaded when the bandwidth is 5Gbit/s or more
by Nguyễn Anh Dũng 17 Apr '24

17 Apr '24

Hi Paul Warren and everyone - I'm having a problem when using iftop to analyze bandwidth greater than 5Gbit/s, it will slow down and give non-standard bandwidth. The total output bandwidth at this time is only about 1Gbit/s. - As far as I know, iftop runs 2 threads, 1 thread to analyze bandwidth and 1 thread to display, so if the CPU has many CPU cores, iftop will not use all of those CPU cores. - I also thought about filtering by subnet (-F) to divide the bandwidth into smaller amounts, it is more effective. But still not optimized. Because I have a number of IPs with bandwidth up to 5Gbit/s, I don't know how to filter or divide that bandwidth. - Now I'm thinking about optimizing by making iftop multi-threaded on the CPU, thereby expanding performance through more powerful hardware. I tried editing the source code iftop but it didn't work, and got many errors. Hope to get everyone's support Thanks all [image: Biểu tượng Cộng đồng đã xác minh] I also thought about filtering by subnet (-F) to divide the bandwidth into smaller amounts, it is more effective. But still not optimized. Because I have a number of IPs with bandwidth up to 5Gbit/s, I don't know how to filter or divide that bandwidth.

1 0

Improve C99 compatibility of the configure script
by Florian Weimer 22 Dec '22

22 Dec '22

This patch is needed to fix a C99 compatibility issue in the configure script. Future compilers are likely to reject implicit function declarations by default, causing these configure checks to fail always. Related to: <https://fedoraproject.org/wiki/Changes/PortingToModernC> <https://fedoraproject.org/wiki/Toolchain/PortingToModernC> diff --git a/configure b/configure index a0bb579d62bd6658..4ea776e4318a39e7 100755 --- a/configure +++ b/configure @@ -5079,9 +5079,9 @@ else int main() { $type dummy; FILE *f=fopen("conftestval", "w"); - if (!f) exit(1); + if (!f) return 1; fprintf(f, "%d\n", sizeof($1)); - exit(0); + return 0; } _ACEOF diff --git a/configure.ac b/configure.ac index 886a23a353c01856..e9ef7f57d002c54e 100644 --- a/configure.ac +++ b/configure.ac @@ -111,9 +111,9 @@ for type in u_int8_t u_int16_t u_int32_t ; do int main() { $type dummy; FILE *f=fopen("conftestval", "w"); - if (!f) exit(1); + if (!f) return 1; fprintf(f, "%d\n", sizeof($1)); - exit(0); + return 0; } ])], [ x=`cat conftestval` Thanks, Florian

1 0

iftop resolving IPs as local hostname instead of regular public PTR
by Hoggins! 25 May '22

25 May '22

Hi there, First time posting here, I'm having questions/issues regarding DNS PTR resolution when running iftop. I'm asking this because "it used to work" and now it's not outputting what I expect, and I'd like to understand why and how to circumvent the problem. I'm aware that my issue might very well be system-related, but I'm sure discussing this with you will enable me finding the culprit. Here's the thing: I have dynamic (public) IPs attached to a system (using Keepalived, everything works fine), and I monitor the traffic going on this public interface with iftop. These public IPs all have correct PTR records. eg.with dig : ;; ANSWER SECTION: 50.52.196.5.in-addr.arpa. 86369 IN PTR www.radiom.fr. But when I use iftop to monitor this public interface, all I get as resolved hostname is my local hostname. It *used* to work, and I'm pretty sure it's due to a system upgrade, maybe a library that doesn't behave as I expect it to. I checked my /etc/hosts file, but there are no entries linking public IPs and the local hostname. So my question is: where should I look to get the correct info? Is iftop using something like gethostbyaddr() to get the hostname? Thanks! Hoggins!

2 4

[PATCHES] Connection tracking map & multi-interface monitoring
by Luke Dashjr 27 Dec '21

27 Dec '21

First patch looks up connections in Linux's connection tracking (NAT) table. Handy to iftop the WAN interface on my router and still see LAN IPs. Second series (4 patches) enables combining multiple network interfaces. I'm using this to monitor my real WAN plus inside-a-VPN traffic (and a filter rule to exclude the VPN traffic on the real WAN). (It would have been nicer to take multiple -i options rather than use a comma separator, but the options code was too messy and I didn't care to spend additional time figuring out how to refactor it.) Luke

1 0

[PATCH] Reverse order of pthread_cond_signal()/pthread_mutex_unlock()
by Bill Nickless 08 Feb '21

08 Feb '21

Valgrand was not happy with the order in which pthread_cond_signal() and pthread_mutex_unlock() were being called. Also see https://docs.oracle.com/cd/E19455-01/806-5257/6je9h032r/index.html which states "Call pthread_cond_signal() under the protection of the same mutex used with the condition variable being signaled. Otherwise, the condition variable could be signaled between the test of the associated condition and blocking in pthread_cond_wait(), which can cause an infinite wait." Patch also available at: https://github.com/wknickless/iftop/commit/4c2fc858dc90b497819f0580fa4723a2… Signed-off-by: Bill Nickless <bill(a)nonick.org> --- resolver.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/resolver.c b/resolver.c index 6c4c21d..70394c2 100644 --- a/resolver.c +++ b/resolver.c @@ -508,12 +508,13 @@ void resolve(int af, void* addr, char* result, int buflen) { added = 1; } } - pthread_mutex_unlock(&resolver_queue_mutex); if(added == 1) { pthread_cond_signal(&resolver_queue_cond); } + pthread_mutex_unlock(&resolver_queue_mutex); + if(result != NULL && buflen > 1) { strncpy(result, hostname, buflen - 1); result[buflen - 1] = '\0'; -- 2.25.1

1 0

[PATCH] sethostent() and setprotoent() are not thread safe
by Bill Nickless 08 Feb '21

08 Feb '21

Remove calls to sethostent() and setprotoent() because they are only optimizations and are not thread safe. Signed-off-by: Bill Nickless <bill(a)nonick.org> --- resolver.c | 1 - serv_hash.c | 1 - 2 files changed, 2 deletions(-) diff --git a/resolver.c b/resolver.c index 8557125..6c4c21d 100644 --- a/resolver.c +++ b/resolver.c @@ -407,7 +407,6 @@ char *do_resolve(struct addr_storage *addr) { void resolver_worker(void* ignored) { pthread_mutex_lock(&resolver_queue_mutex); - sethostent(1); while(1) { /* Wait until we are told that an address has been added to the * queue. */ diff --git a/serv_hash.c b/serv_hash.c index 279032e..10af520 100644 --- a/serv_hash.c +++ b/serv_hash.c @@ -52,7 +52,6 @@ void serv_hash_initialise(hash_type* sh) { struct servent* ent; struct protoent* pent; ip_service* service; - setprotoent(1); while((ent = getservent()) != NULL) { pent = getprotobyname(ent->s_proto); if(pent != NULL) { -- 2.25.1

1 0

[PATCH] Fix -Wformat-truncation warnings
by Bill Nickless 08 Feb '21

08 Feb '21

Fix -Wformat-truncation warnings and get rid of the magic constant 9 in tui.c. Also available at https://github.com/wknickless/iftop/commit/6f5c565a5fdaf127cfcdc236c22fdc96… Signed-off-by: Bill Nickless <bill(a)nonick.org> --- tui.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/tui.c b/tui.c index af10a2f..59221a5 100644 --- a/tui.c +++ b/tui.c @@ -34,6 +34,7 @@ /* Width of the host column in the output */ #define PRINT_WIDTH 40 +#define PRINT_WIDTH_LONG 49 /* @@ -55,7 +56,7 @@ void tui_print() { if (!labellong) { xfree(labellong); - labellong = (char *)calloc(PRINT_WIDTH + 1 + 9, 1); + labellong = (char *)calloc(PRINT_WIDTH_LONG + 1, 1); } if (options.paused ) { @@ -63,7 +64,7 @@ void tui_print() { } /* Headings */ - snprintf(label, PRINT_WIDTH, "%-*s", PRINT_WIDTH, "Host name (port/service if enabled)"); + snprintf(label, PRINT_WIDTH + 1, "%-*s", PRINT_WIDTH, "Host name (port/service if enabled)"); printf("%s %s %10s %10s %10s %10s\n", " #", label, "last 2s", "last 10s", "last 40s", "cumulative"); /* Divider line */ @@ -112,21 +113,21 @@ void tui_print() { printf("\n"); /* Rate totals */ - snprintf(labellong, PRINT_WIDTH + 9, "%-*s", PRINT_WIDTH + 9, "Total send rate:"); + snprintf(labellong, PRINT_WIDTH_LONG + 1, "%-*s", PRINT_WIDTH_LONG, "Total send rate:"); printf("%s ", labellong); for(j = 0; j < HISTORY_DIVISIONS; j++) { readable_size(((host_pair_line *)&totals)->sent[j], buf0_10, 10, 1024, options.bandwidth_unit); printf("%10s%c", buf0_10, j == HISTORY_DIVISIONS - 1 ? '\n' : ' '); } - snprintf(labellong, PRINT_WIDTH + 9, "%-*s", PRINT_WIDTH + 9, "Total receive rate:"); + snprintf(labellong, PRINT_WIDTH_LONG + 1, "%-*s", PRINT_WIDTH_LONG, "Total receive rate:"); printf("%s ", labellong); for(j = 0; j < HISTORY_DIVISIONS; j++) { readable_size(((host_pair_line *)&totals)->recv[j], buf0_10, 10, 1024, options.bandwidth_unit); printf("%10s%c", buf0_10, j == HISTORY_DIVISIONS - 1 ? '\n' : ' '); } - snprintf(labellong, PRINT_WIDTH + 9, "%-*s", PRINT_WIDTH + 9, "Total send and receive rate:"); + snprintf(labellong, PRINT_WIDTH_LONG + 1, "%-*s", PRINT_WIDTH_LONG, "Total send and receive rate:"); printf("%s ", labellong); for(j = 0; j < HISTORY_DIVISIONS; j++) { readable_size(((host_pair_line *)&totals)->sent[j] + ((host_pair_line *)&totals)->recv[j], buf0_10, 10, 1024, options.bandwidth_unit); @@ -140,14 +141,14 @@ void tui_print() { printf("\n"); /* Peak traffic */ - snprintf(labellong, PRINT_WIDTH + 9, "%-*s", PRINT_WIDTH + 9, "Peak rate (sent/received/total):"); + snprintf(labellong, PRINT_WIDTH_LONG + 1, "%-*s", PRINT_WIDTH_LONG, "Peak rate (sent/received/total):"); readable_size(peaksent / RESOLUTION, buf0_10, 10, 1024, options.bandwidth_unit); readable_size(peakrecv / RESOLUTION, buf1_10, 10, 1024, options.bandwidth_unit); readable_size(peaktotal / RESOLUTION, buf2_10, 10, 1024, options.bandwidth_unit); printf("%s %10s %10s %10s\n", labellong, buf0_10, buf1_10, buf2_10); /* Cumulative totals */ - snprintf(labellong, PRINT_WIDTH + 9, "%-*s", PRINT_WIDTH + 9, "Cumulative (sent/received/total):"); + snprintf(labellong, PRINT_WIDTH_LONG + 1, "%-*s", PRINT_WIDTH_LONG, "Cumulative (sent/received/total):"); readable_size(history_totals.total_sent, buf0_10, 10, 1024, 1); readable_size(history_totals.total_recv, buf1_10, 10, 1024, 1); readable_size(history_totals.total_recv + history_totals.total_sent, buf2_10, 10, 1024, 1); -- 2.25.1

1 0

[PATCH] Support passively sniffing IPv6
by Bill Nickless 07 Feb '21

07 Feb '21

Support passively sniffing and displaying IPv6 traffic, previously silently ignored. Also submitted: https://github.com/gdm85/iftop/pull/1 Signed-off-by: Bill Nickless <bill(a)nonick.org> --- iftop.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/iftop.c b/iftop.c index 311401e..5e829f9 100644 --- a/iftop.c +++ b/iftop.c @@ -352,6 +352,13 @@ static void handle_ip_packet(struct ip* iptr, int hw_dir, int pld_len) assign_addr_pair(&ap, iptr, 0); direction = 0; } + else if(IP_V(iptr) == 6 && memcmp(&ip6tr->ip6_src, + &ip6tr->ip6_dst, + sizeof(struct in6_addr)) < 0) { + assign_addr_pair(&ap, iptr, 1); + } else if(IP_V(iptr) == 6) { + assign_addr_pair(&ap, iptr, 0); + } /* Drop other uncertain packages. */ else return; -- 2.25.1

1 0

Feature request: pppd decapsulation (I think)
by James Lay 25 Sep '20

25 Sep '20

So, I have an interface, eth1, that is used for pppd (connects direct to the modem) that is used to create the ppp0 virtual interface. When using iftop on ppp0, it works great. Sadly sometimes that interface bounces, and iftop stops reporting data after it comes back up. When trying to use iftop on eth1, no data is reported even when ppp0 is up. Can we get iftop to work on ppp0? I hope I explained it well...thank you! James

3 5

[PATCH v2] Fix minor ui annoyance when ligatures are suported.
by Ariel Popper 25 Sep '20

25 Sep '20

When `iftop` is run under a terminal/font combination that supports ligatures, the "<=" characters form a single, stacked "less than or equal to" glyph rather than an arrow pointing left. This patch replaces "<=" with "<-" along with some others in order to remain visually consistant. --- I'm not sure if anyone cares about this but me, but thought I might share it with whoever wants it. Anyhow, this is a more complete patch than the last one. iftop.8 | 6 +++--- tui.c | 4 ++-- ui.c | 10 +++++----- 3 files changed, 10 insertions(+), 10 deletions(-) diff --git a/iftop.8 b/iftop.8 index 6bc38a7..298d636 100644 --- a/iftop.8 +++ b/iftop.8 @@ -116,12 +116,12 @@ visual indication of traffic. The main part of the display lists, for each pair of hosts, the rate at which data has been sent and received over the preceding 2, 10 and 40 second -intervals. The direction of data flow is indicated by arrows, <= and =>. For +intervals. The direction of data flow is indicated by arrows, <- and ->. For instance, .nf -foo.example.com => bar.example.com 1Kb 500b 100b - <= 2Mb 2Mb 2Mb +foo.example.com -> bar.example.com 1Kb 500b 100b + <- 2Mb 2Mb 2Mb .fi shows, on the first line, traffic from \fBfoo.example.com\fP to diff --git a/tui.c b/tui.c index af10a2f..d9b26b6 100644 --- a/tui.c +++ b/tui.c @@ -85,7 +85,7 @@ void tui_print() { sprint_host(host2, screen_line->ap.af, &(screen_line->ap.dst6), screen_line->ap.dst_port, screen_line->ap.protocol, PRINT_WIDTH, options.aggregate_dest); /* Send rate per connection */ - printf("%4d %s%s", l, host1, " =>"); + printf("%4d %s%s", l, host1, " ->"); for(j = 0; j < HISTORY_DIVISIONS; j++) { readable_size(screen_line->sent[j], buf0_10, 10, 1024, options.bandwidth_unit); printf(" %10s", buf0_10); @@ -95,7 +95,7 @@ void tui_print() { printf(" %10s\n", buf0_10); /* Receive rate per connection */ - printf(" %s%s", host2, " <="); + printf(" %s%s", host2, " <-"); for(j = 0; j < HISTORY_DIVISIONS; j++) { readable_size(screen_line->recv[j], buf0_10, 10, 1024, options.bandwidth_unit); printf(" %10s", buf0_10); diff --git a/ui.c b/ui.c index abaadd2..242bd6e 100644 --- a/ui.c +++ b/ui.c @@ -330,17 +330,17 @@ void ui_print() { switch(options.linedisplay) { case OPTION_LINEDISPLAY_TWO_LINE: - mvaddstr(y, x, " => "); - mvaddstr(y+1, x, " <= "); + mvaddstr(y, x, " -> "); + mvaddstr(y+1, x, " <- "); break; case OPTION_LINEDISPLAY_ONE_LINE_BOTH: - mvaddstr(y, x, "<=> "); + mvaddstr(y, x, "<-> "); break; case OPTION_LINEDISPLAY_ONE_LINE_SENT: - mvaddstr(y, x, " => "); + mvaddstr(y, x, " -> "); break; case OPTION_LINEDISPLAY_ONE_LINE_RECV: - mvaddstr(y, x, " <= "); + mvaddstr(y, x, " <- "); break; } -- 2.28.0

1 0

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

iftop-users