Skip to contentSkip to conversation

Abstract columns, properly space and wrap fields

PATCH In review Stefano Brivio 5 participants 19 messages Dec 12, 2017 – Feb 13, 2019

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

v2 cover Abstract columns, properly space and wrap fields

Currently, 'ss' simply subdivides the whole available screen width between available columns, starting from a set of hardcoded amount of spacing and growing column widths.

This makes the output unreadable in several cases, as it doesn't take into account the actual content width.

Fix this by introducing a simple abstraction for columns, buffering the output, measuring the width of the fields, grouping fields into lines as they fit, equally distributing any remaining whitespace, and finally rendering the result. Some examples are reported below [1].

This implementation doesn't seem to cause any significant performance issues, as reported in 3/4.

Patch 1/4 replaces all relevant printf() calls by the out() helper, which simply consists of the usual printf() implementation.

Patch 2/4 implements column abstraction, with configurable column width and delimiters, and 3/4 splits buffering and rendering phases, employing a simple buffering mechanism with chunked allocation and introducing a rendering function.

Up to this point, the output is still unchanged.

Finally, 4/4 introduces field width calculation based on content length measured while buffering, in order to split fields onto multiple lines and equally space them within the single lines.

Now that column behaviour is well-defined and more easily configurable, it should be easier to further improve the output by splitting logically separable information (e.g. TCP details) into additional columns. However, this patchset keeps the full "extended" information into a single column, for the moment being.

v2: rebase after conflict with 00ac78d39c29 ("ss: print tcpi_rcv_ssthresh")

[1]

- 80 columns terminal, ss -Z -f netlink
* before:
Recv-Q Send-Q Local Address:Port Peer Address:Port

0 0 rtnl:evolution-calen/2075 * pr
oc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:abrt-applet/32700 * pr
oc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:firefox/21619 * pr
oc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:evolution-calen/32639 * p
roc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 [...]

* after:
Recv-Q Send-Q Local Address:Port Peer Address:Port
0 0 rtnl:evolution-calen/2075 *
proc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:abrt-applet/32700 *
proc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:firefox/21619 *
proc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:evolution-calen/32639 *
proc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
[...]

- 80 columns terminal, ss -tunpl
* before:

Netid  State      Recv-Q Send-Q Local Address:Port               Peer Address:Port
udp    UNCONN     0      0         *:37732                 *:*
udp    UNCONN     0      0         *:5353                  *:*
udp    UNCONN     0      0      192.168.122.1:53                    *:*
udp    UNCONN     0      0      *%virbr0:67                    *:*

[...]

* after:

Netid   State    Recv-Q   Send-Q     Local Address:Port      Peer Address:Port
udp     UNCONN   0        0                      *:37732                *:*
udp     UNCONN   0        0                      *:5353                 *:*
udp     UNCONN   0        0          192.168.122.1:53                   *:*
udp     UNCONN   0        0               *%virbr0:67                   *:*

[...]

- 66 columns terminal, ss -tunpl
* before:
Netid State Recv-Q Send-Q Local Address:Port P
eer Address:Port
udp UNCONN 0 0 *:37732 *:*

udp UNCONN 0 0 *:5353 *:*

udp UNCONN 0 0 192.168.122.1:53
*:*
udp UNCONN 0 0 *%virbr0:67 *:*
[...]

* after:
Netid State Recv-Q Send-Q Local Address:Port Peer Address:Port

udp   UNCONN 0      0                  *:37732             *:*
udp   UNCONN 0      0                  *:5353              *:*
udp   UNCONN 0      0      192.168.122.1:53                *:*
udp   UNCONN 0      0           *%virbr0:67                *:*

[...]

Stefano Brivio (4):
ss: Replace printf() calls for "main" output by calls to helper
ss: Introduce columns lightweight abstraction
ss: Buffer raw fields first, then render them as a table
ss: Implement automatic column width calculation

 misc/ss.c | 895 +++++++++++++++++++++++++++++++++++++++++++-------------------
 1 file changed, 621 insertions(+), 274 deletions(-)
Signature · 1 line
2.9.4
#1 lore raw

PATCH v2 1/4 ss: Replace printf() calls for "main" output by calls to helper

This is preparation work for output buffering, which will allow us to use optimal spacing and alignment of logical "columns".

The new out() function is just a re-implementation of a typical libc's printf(), except that the return value of vfprintf() is ignored as no callers use it. This implementation will be replaced in the next patches to provide column width adjustment and adequate spacing.

All printf() calls that output parts of the socket list are now replaced by calls to out(). Output of summary and version is excluded from this.

No functional differences here, output not affected.

---
v2: rebase after conflict with 00ac78d39c29 ("ss: print tcpi_rcv_ssthresh")

 misc/ss.c | 399 ++++++++++++++++++++++++++++++++------------------------------
 1 file changed, 205 insertions(+), 194 deletions(-)
misc/ss.c +205 -194
index da52d5edeb7e..a7d3b89e1478 100644
@@ -26,6 +26,7 @@
 #include <getopt.h>
 #include <stdbool.h>
 #include <limits.h>
+#include <stdarg.h>
 
 #include "utils.h"
 #include "rt_names.h"
@@ -823,6 +824,15 @@ static const char *vsock_netid_name(int type)
 	}
 }
 
+static void out(const char *fmt, ...)
+{
+	va_list args;
+
+	va_start(args, fmt);
+	vfprintf(stdout, fmt, args);
+	va_end(args);
+}
+
 static void sock_state_print(struct sockstat *s)
 {
 	const char *sock_name;
@@ -863,39 +873,39 @@ static void sock_state_print(struct sockstat *s)
 	}
 
 	if (netid_width)
-		printf("%-*s ", netid_width,
-		       is_sctp_assoc(s, sock_name) ? "" : sock_name);
+		out("%-*s ", netid_width,
+		    is_sctp_assoc(s, sock_name) ? "" : sock_name);
 	if (state_width) {
 		if (is_sctp_assoc(s, sock_name))
-			printf("`- %-*s ", state_width - 3,
-			       sctp_sstate_name[s->state]);
+			out("`- %-*s ", state_width - 3,
+			    sctp_sstate_name[s->state]);
 		else
-			printf("%-*s ", state_width, sstate_name[s->state]);
+			out("%-*s ", state_width, sstate_name[s->state]);
 	}
 
-	printf("%-6d %-6d %s", s->rq, s->wq, odd_width_pad);
+	out("%-6d %-6d %s", s->rq, s->wq, odd_width_pad);
 }
 
 static void sock_details_print(struct sockstat *s)
 {
 	if (s->uid)
-		printf(" uid:%u", s->uid);
+		out(" uid:%u", s->uid);
 
-	printf(" ino:%u", s->ino);
-	printf(" sk:%llx", s->sk);
+	out(" ino:%u", s->ino);
+	out(" sk:%llx", s->sk);
 
 	if (s->mark)
-		printf(" fwmark:0x%x", s->mark);
+		out(" fwmark:0x%x", s->mark);
 }
 
 static void sock_addr_print_width(int addr_len, const char *addr, char *delim,
 		int port_len, const char *port, const char *ifname)
 {
 	if (ifname) {
-		printf("%*s%%%s%s%-*s ", addr_len, addr, ifname, delim,
-				port_len, port);
+		out("%*s%%%s%s%-*s ", addr_len, addr, ifname, delim,
+		    port_len, port);
 	} else {
-		printf("%*s%s%-*s ", addr_len, addr, delim, port_len, port);
+		out("%*s%s%-*s ", addr_len, addr, delim, port_len, port);
 	}
 }
 
@@ -1793,12 +1803,12 @@ static void proc_ctx_print(struct sockstat *s)
 		if (find_entry(s->ino, &buf,
 				(show_proc_ctx & show_sock_ctx) ?
 				PROC_SOCK_CTX : PROC_CTX) > 0) {
-			printf(" users:(%s)", buf);
+			out(" users:(%s)", buf);
 			free(buf);
 		}
 	} else if (show_users) {
 		if (find_entry(s->ino, &buf, USERS) > 0) {
-			printf(" users:(%s)", buf);
+			out(" users:(%s)", buf);
 			free(buf);
 		}
 	}
@@ -1878,51 +1888,51 @@ static char *sprint_bw(char *buf, double bw)
 static void sctp_stats_print(struct sctp_info *s)
 {
 	if (s->sctpi_tag)
-		printf(" tag:%x", s->sctpi_tag);
+		out(" tag:%x", s->sctpi_tag);
 	if (s->sctpi_state)
-		printf(" state:%s", sctp_sstate_name[s->sctpi_state]);
+		out(" state:%s", sctp_sstate_name[s->sctpi_state]);
 	if (s->sctpi_rwnd)
-		printf(" rwnd:%d", s->sctpi_rwnd);
+		out(" rwnd:%d", s->sctpi_rwnd);
 	if (s->sctpi_unackdata)
-		printf(" unackdata:%d", s->sctpi_unackdata);
+		out(" unackdata:%d", s->sctpi_unackdata);
 	if (s->sctpi_penddata)
-		printf(" penddata:%d", s->sctpi_penddata);
+		out(" penddata:%d", s->sctpi_penddata);
 	if (s->sctpi_instrms)
-		printf(" instrms:%d", s->sctpi_instrms);
+		out(" instrms:%d", s->sctpi_instrms);
 	if (s->sctpi_outstrms)
-		printf(" outstrms:%d", s->sctpi_outstrms);
+		out(" outstrms:%d", s->sctpi_outstrms);
 	if (s->sctpi_inqueue)
-		printf(" inqueue:%d", s->sctpi_inqueue);
+		out(" inqueue:%d", s->sctpi_inqueue);
 	if (s->sctpi_outqueue)
-		printf(" outqueue:%d", s->sctpi_outqueue);
+		out(" outqueue:%d", s->sctpi_outqueue);
 	if (s->sctpi_overall_error)
-		printf(" overerr:%d", s->sctpi_overall_error);
+		out(" overerr:%d", s->sctpi_overall_error);
 	if (s->sctpi_max_burst)
-		printf(" maxburst:%d", s->sctpi_max_burst);
+		out(" maxburst:%d", s->sctpi_max_burst);
 	if (s->sctpi_maxseg)
-		printf(" maxseg:%d", s->sctpi_maxseg);
+		out(" maxseg:%d", s->sctpi_maxseg);
 	if (s->sctpi_peer_rwnd)
-		printf(" prwnd:%d", s->sctpi_peer_rwnd);
+		out(" prwnd:%d", s->sctpi_peer_rwnd);
 	if (s->sctpi_peer_tag)
-		printf(" ptag:%x", s->sctpi_peer_tag);
+		out(" ptag:%x", s->sctpi_peer_tag);
 	if (s->sctpi_peer_capable)
-		printf(" pcapable:%d", s->sctpi_peer_capable);
+		out(" pcapable:%d", s->sctpi_peer_capable);
 	if (s->sctpi_peer_sack)
-		printf(" psack:%d", s->sctpi_peer_sack);
+		out(" psack:%d", s->sctpi_peer_sack);
 	if (s->sctpi_s_autoclose)
-		printf(" autoclose:%d", s->sctpi_s_autoclose);
+		out(" autoclose:%d", s->sctpi_s_autoclose);
 	if (s->sctpi_s_adaptation_ind)
-		printf(" adapind:%d", s->sctpi_s_adaptation_ind);
+		out(" adapind:%d", s->sctpi_s_adaptation_ind);
 	if (s->sctpi_s_pd_point)
-		printf(" pdpoint:%d", s->sctpi_s_pd_point);
+		out(" pdpoint:%d", s->sctpi_s_pd_point);
 	if (s->sctpi_s_nodelay)
-		printf(" nodealy:%d", s->sctpi_s_nodelay);
+		out(" nodealy:%d", s->sctpi_s_nodelay);
 	if (s->sctpi_s_disable_fragments)
-		printf(" nofrag:%d", s->sctpi_s_disable_fragments);
+		out(" nofrag:%d", s->sctpi_s_disable_fragments);
 	if (s->sctpi_s_v4mapped)
-		printf(" v4mapped:%d", s->sctpi_s_v4mapped);
+		out(" v4mapped:%d", s->sctpi_s_v4mapped);
 	if (s->sctpi_s_frag_interleave)
-		printf(" fraginl:%d", s->sctpi_s_frag_interleave);
+		out(" fraginl:%d", s->sctpi_s_frag_interleave);
 }
 
 static void tcp_stats_print(struct tcpstat *s)
@@ -1930,65 +1940,65 @@ static void tcp_stats_print(struct tcpstat *s)
 	char b1[64];
 
 	if (s->has_ts_opt)
-		printf(" ts");
+		out(" ts");
 	if (s->has_sack_opt)
-		printf(" sack");
+		out(" sack");
 	if (s->has_ecn_opt)
-		printf(" ecn");
+		out(" ecn");
 	if (s->has_ecnseen_opt)
-		printf(" ecnseen");
+		out(" ecnseen");
 	if (s->has_fastopen_opt)
-		printf(" fastopen");
+		out(" fastopen");
 	if (s->cong_alg[0])
-		printf(" %s", s->cong_alg);
+		out(" %s", s->cong_alg);
 	if (s->has_wscale_opt)
-		printf(" wscale:%d,%d", s->snd_wscale, s->rcv_wscale);
+		out(" wscale:%d,%d", s->snd_wscale, s->rcv_wscale);
 	if (s->rto)
-		printf(" rto:%g", s->rto);
+		out(" rto:%g", s->rto);
 	if (s->backoff)
-		printf(" backoff:%u", s->backoff);
+		out(" backoff:%u", s->backoff);
 	if (s->rtt)
-		printf(" rtt:%g/%g", s->rtt, s->rttvar);
+		out(" rtt:%g/%g", s->rtt, s->rttvar);
 	if (s->ato)
-		printf(" ato:%g", s->ato);
+		out(" ato:%g", s->ato);
 
 	if (s->qack)
-		printf(" qack:%d", s->qack);
+		out(" qack:%d", s->qack);
 	if (s->qack & 1)
-		printf(" bidir");
+		out(" bidir");
 
 	if (s->mss)
-		printf(" mss:%d", s->mss);
+		out(" mss:%d", s->mss);
 	if (s->rcv_mss)
-		printf(" rcvmss:%d", s->rcv_mss);
+		out(" rcvmss:%d", s->rcv_mss);
 	if (s->advmss)
-		printf(" advmss:%d", s->advmss);
+		out(" advmss:%d", s->advmss);
 	if (s->cwnd)
-		printf(" cwnd:%u", s->cwnd);
+		out(" cwnd:%u", s->cwnd);
 	if (s->ssthresh)
-		printf(" ssthresh:%d", s->ssthresh);
+		out(" ssthresh:%d", s->ssthresh);
 
 	if (s->bytes_acked)
-		printf(" bytes_acked:%llu", s->bytes_acked);
+		out(" bytes_acked:%llu", s->bytes_acked);
 	if (s->bytes_received)
-		printf(" bytes_received:%llu", s->bytes_received);
+		out(" bytes_received:%llu", s->bytes_received);
 	if (s->segs_out)
-		printf(" segs_out:%u", s->segs_out);
+		out(" segs_out:%u", s->segs_out);
 	if (s->segs_in)
-		printf(" segs_in:%u", s->segs_in);
+		out(" segs_in:%u", s->segs_in);
 	if (s->data_segs_out)
-		printf(" data_segs_out:%u", s->data_segs_out);
+		out(" data_segs_out:%u", s->data_segs_out);
 	if (s->data_segs_in)
-		printf(" data_segs_in:%u", s->data_segs_in);
+		out(" data_segs_in:%u", s->data_segs_in);
 
 	if (s->dctcp && s->dctcp->enabled) {
 		struct dctcpstat *dctcp = s->dctcp;
 
-		printf(" dctcp:(ce_state:%u,alpha:%u,ab_ecn:%u,ab_tot:%u)",
-				dctcp->ce_state, dctcp->alpha, dctcp->ab_ecn,
-				dctcp->ab_tot);
+		out(" dctcp:(ce_state:%u,alpha:%u,ab_ecn:%u,ab_tot:%u)",
+			     dctcp->ce_state, dctcp->alpha, dctcp->ab_ecn,
+			     dctcp->ab_tot);
 	} else if (s->dctcp) {
-		printf(" dctcp:fallback_mode");
+		out(" dctcp:fallback_mode");
 	}
 
 	if (s->bbr_info) {
@@ -1998,73 +2008,72 @@ static void tcp_stats_print(struct tcpstat *s)
 		bw <<= 32;
 		bw |= s->bbr_info->bbr_bw_lo;
 
-		printf(" bbr:(bw:%sbps,mrtt:%g",
-		       sprint_bw(b1, bw * 8.0),
-		       (double)s->bbr_info->bbr_min_rtt / 1000.0);
+		out(" bbr:(bw:%sbps,mrtt:%g",
+		    sprint_bw(b1, bw * 8.0),
+		    (double)s->bbr_info->bbr_min_rtt / 1000.0);
 		if (s->bbr_info->bbr_pacing_gain)
-			printf(",pacing_gain:%g",
-			       (double)s->bbr_info->bbr_pacing_gain / 256.0);
+			out(",pacing_gain:%g",
+			    (double)s->bbr_info->bbr_pacing_gain / 256.0);
 		if (s->bbr_info->bbr_cwnd_gain)
-			printf(",cwnd_gain:%g",
-			       (double)s->bbr_info->bbr_cwnd_gain / 256.0);
-		printf(")");
+			out(",cwnd_gain:%g",
+			    (double)s->bbr_info->bbr_cwnd_gain / 256.0);
+		out(")");
 	}
 
 	if (s->send_bps)
-		printf(" send %sbps", sprint_bw(b1, s->send_bps));
+		out(" send %sbps", sprint_bw(b1, s->send_bps));
 	if (s->lastsnd)
-		printf(" lastsnd:%u", s->lastsnd);
+		out(" lastsnd:%u", s->lastsnd);
 	if (s->lastrcv)
-		printf(" lastrcv:%u", s->lastrcv);
+		out(" lastrcv:%u", s->lastrcv);
 	if (s->lastack)
-		printf(" lastack:%u", s->lastack);
+		out(" lastack:%u", s->lastack);
 
 	if (s->pacing_rate) {
-		printf(" pacing_rate %sbps", sprint_bw(b1, s->pacing_rate));
+		out(" pacing_rate %sbps", sprint_bw(b1, s->pacing_rate));
 		if (s->pacing_rate_max)
-				printf("/%sbps", sprint_bw(b1,
-							s->pacing_rate_max));
+			out("/%sbps", sprint_bw(b1, s->pacing_rate_max));
 	}
 
 	if (s->delivery_rate)
-		printf(" delivery_rate %sbps", sprint_bw(b1, s->delivery_rate));
+		out(" delivery_rate %sbps", sprint_bw(b1, s->delivery_rate));
 	if (s->app_limited)
-		printf(" app_limited");
+		out(" app_limited");
 
 	if (s->busy_time) {
-		printf(" busy:%llums", s->busy_time / 1000);
+		out(" busy:%llums", s->busy_time / 1000);
 		if (s->rwnd_limited)
-			printf(" rwnd_limited:%llums(%.1f%%)",
-			       s->rwnd_limited / 1000,
-			       100.0 * s->rwnd_limited / s->busy_time);
+			out(" rwnd_limited:%llums(%.1f%%)",
+			    s->rwnd_limited / 1000,
+			    100.0 * s->rwnd_limited / s->busy_time);
 		if (s->sndbuf_limited)
-			printf(" sndbuf_limited:%llums(%.1f%%)",
-			       s->sndbuf_limited / 1000,
-			       100.0 * s->sndbuf_limited / s->busy_time);
+			out(" sndbuf_limited:%llums(%.1f%%)",
+			    s->sndbuf_limited / 1000,
+			    100.0 * s->sndbuf_limited / s->busy_time);
 	}
 
 	if (s->unacked)
-		printf(" unacked:%u", s->unacked);
+		out(" unacked:%u", s->unacked);
 	if (s->retrans || s->retrans_total)
-		printf(" retrans:%u/%u", s->retrans, s->retrans_total);
+		out(" retrans:%u/%u", s->retrans, s->retrans_total);
 	if (s->lost)
-		printf(" lost:%u", s->lost);
+		out(" lost:%u", s->lost);
 	if (s->sacked && s->ss.state != SS_LISTEN)
-		printf(" sacked:%u", s->sacked);
+		out(" sacked:%u", s->sacked);
 	if (s->fackets)
-		printf(" fackets:%u", s->fackets);
+		out(" fackets:%u", s->fackets);
 	if (s->reordering != 3)
-		printf(" reordering:%d", s->reordering);
+		out(" reordering:%d", s->reordering);
 	if (s->rcv_rtt)
-		printf(" rcv_rtt:%g", s->rcv_rtt);
+		out(" rcv_rtt:%g", s->rcv_rtt);
 	if (s->rcv_space)
-		printf(" rcv_space:%d", s->rcv_space);
+		out(" rcv_space:%d", s->rcv_space);
 	if (s->rcv_ssthresh)
-		printf(" rcv_ssthresh:%u", s->rcv_ssthresh);
+		out(" rcv_ssthresh:%u", s->rcv_ssthresh);
 	if (s->not_sent)
-		printf(" notsent:%u", s->not_sent);
+		out(" notsent:%u", s->not_sent);
 	if (s->min_rtt)
-		printf(" minrtt:%g", s->min_rtt);
+		out(" minrtt:%g", s->min_rtt);
 }
 
 static void tcp_timer_print(struct tcpstat *s)
@@ -2081,18 +2090,18 @@ static void tcp_timer_print(struct tcpstat *s)
 	if (s->timer) {
 		if (s->timer > 4)
 			s->timer = 5;
-		printf(" timer:(%s,%s,%d)",
-				tmr_name[s->timer],
-				print_ms_timer(s->timeout),
-				s->retrans);
+		out(" timer:(%s,%s,%d)",
+			     tmr_name[s->timer],
+			     print_ms_timer(s->timeout),
+			     s->retrans);
 	}
 }
 
 static void sctp_timer_print(struct tcpstat *s)
 {
 	if (s->timer)
-		printf(" timer:(T3_RTX,%s,%d)",
-		       print_ms_timer(s->timeout), s->retrans);
+		out(" timer:(T3_RTX,%s,%d)",
+		    print_ms_timer(s->timeout), s->retrans);
 }
 
 static int tcp_show_line(char *line, const struct filter *f, int family)
@@ -2151,13 +2160,13 @@ static int tcp_show_line(char *line, const struct filter *f, int family)
 	if (show_details) {
 		sock_details_print(&s.ss);
 		if (opt[0])
-			printf(" opt:\"%s\"", opt);
+			out(" opt:\"%s\"", opt);
 	}
 
 	if (show_tcpinfo)
 		tcp_stats_print(&s);
 
-	printf("\n");
+	out("\n");
 	return 0;
 }
 
@@ -2200,44 +2209,44 @@ static void print_skmeminfo(struct rtattr *tb[], int attrtype)
 			const struct inet_diag_meminfo *minfo =
 				RTA_DATA(tb[INET_DIAG_MEMINFO]);
 
-			printf(" mem:(r%u,w%u,f%u,t%u)",
-					minfo->idiag_rmem,
-					minfo->idiag_wmem,
-					minfo->idiag_fmem,
-					minfo->idiag_tmem);
+			out(" mem:(r%u,w%u,f%u,t%u)",
+				   minfo->idiag_rmem,
+				   minfo->idiag_wmem,
+				   minfo->idiag_fmem,
+				   minfo->idiag_tmem);
 		}
 		return;
 	}
 
 	skmeminfo = RTA_DATA(tb[attrtype]);
 
-	printf(" skmem:(r%u,rb%u,t%u,tb%u,f%u,w%u,o%u",
-	       skmeminfo[SK_MEMINFO_RMEM_ALLOC],
-	       skmeminfo[SK_MEMINFO_RCVBUF],
-	       skmeminfo[SK_MEMINFO_WMEM_ALLOC],
-	       skmeminfo[SK_MEMINFO_SNDBUF],
-	       skmeminfo[SK_MEMINFO_FWD_ALLOC],
-	       skmeminfo[SK_MEMINFO_WMEM_QUEUED],
-	       skmeminfo[SK_MEMINFO_OPTMEM]);
+	out(" skmem:(r%u,rb%u,t%u,tb%u,f%u,w%u,o%u",
+		     skmeminfo[SK_MEMINFO_RMEM_ALLOC],
+		     skmeminfo[SK_MEMINFO_RCVBUF],
+		     skmeminfo[SK_MEMINFO_WMEM_ALLOC],
+		     skmeminfo[SK_MEMINFO_SNDBUF],
+		     skmeminfo[SK_MEMINFO_FWD_ALLOC],
+		     skmeminfo[SK_MEMINFO_WMEM_QUEUED],
+		     skmeminfo[SK_MEMINFO_OPTMEM]);
 
 	if (RTA_PAYLOAD(tb[attrtype]) >=
 		(SK_MEMINFO_BACKLOG + 1) * sizeof(__u32))
-		printf(",bl%u", skmeminfo[SK_MEMINFO_BACKLOG]);
+		out(",bl%u", skmeminfo[SK_MEMINFO_BACKLOG]);
 
 	if (RTA_PAYLOAD(tb[attrtype]) >=
 		(SK_MEMINFO_DROPS + 1) * sizeof(__u32))
-		printf(",d%u", skmeminfo[SK_MEMINFO_DROPS]);
+		out(",d%u", skmeminfo[SK_MEMINFO_DROPS]);
 
-	printf(")");
+	out(")");
 }
 
 static void print_md5sig(struct tcp_diag_md5sig *sig)
 {
-	printf("%s/%d=",
-	       format_host(sig->tcpm_family,
-			   sig->tcpm_family == AF_INET6 ? 16 : 4,
-			   &sig->tcpm_addr),
-	       sig->tcpm_prefixlen);
+	out("%s/%d=",
+	    format_host(sig->tcpm_family,
+			sig->tcpm_family == AF_INET6 ? 16 : 4,
+			&sig->tcpm_addr),
+	    sig->tcpm_prefixlen);
 	print_escape_buf(sig->tcpm_key, sig->tcpm_keylen, " ,");
 }
 
@@ -2382,10 +2391,10 @@ static void tcp_show_info(const struct nlmsghdr *nlh, struct inet_diag_msg *r,
 		struct tcp_diag_md5sig *sig = RTA_DATA(tb[INET_DIAG_MD5SIG]);
 		int len = RTA_PAYLOAD(tb[INET_DIAG_MD5SIG]);
 
-		printf(" md5keys:");
+		out(" md5keys:");
 		print_md5sig(sig++);
 		for (len -= sizeof(*sig); len > 0; len -= sizeof(*sig)) {
-			printf(",");
+			out(",");
 			print_md5sig(sig++);
 		}
 	}
@@ -2420,18 +2429,18 @@ static void sctp_show_info(const struct nlmsghdr *nlh, struct inet_diag_msg *r,
 		len = RTA_PAYLOAD(tb[INET_DIAG_LOCALS]);
 		sa = RTA_DATA(tb[INET_DIAG_LOCALS]);
 
-		printf("locals:%s", format_host_sa(sa));
+		out("locals:%s", format_host_sa(sa));
 		for (sa++, len -= sizeof(*sa); len > 0; sa++, len -= sizeof(*sa))
-			printf(",%s", format_host_sa(sa));
+			out(",%s", format_host_sa(sa));
 
 	}
 	if (tb[INET_DIAG_PEERS]) {
 		len = RTA_PAYLOAD(tb[INET_DIAG_PEERS]);
 		sa = RTA_DATA(tb[INET_DIAG_PEERS]);
 
-		printf(" peers:%s", format_host_sa(sa));
+		out(" peers:%s", format_host_sa(sa));
 		for (sa++, len -= sizeof(*sa); len > 0; sa++, len -= sizeof(*sa))
-			printf(",%s", format_host_sa(sa));
+			out(",%s", format_host_sa(sa));
 	}
 	if (tb[INET_DIAG_INFO]) {
 		struct sctp_info *info;
@@ -2518,18 +2527,19 @@ static int inet_show_sock(struct nlmsghdr *nlh,
 	if (show_details) {
 		sock_details_print(s);
 		if (s->local.family == AF_INET6 && tb[INET_DIAG_SKV6ONLY])
-			printf(" v6only:%u", v6only);
+			out(" v6only:%u", v6only);
 
 		if (tb[INET_DIAG_SHUTDOWN]) {
 			unsigned char mask;
 
 			mask = rta_getattr_u8(tb[INET_DIAG_SHUTDOWN]);
-			printf(" %c-%c", mask & 1 ? '-' : '<', mask & 2 ? '-' : '>');
+			out(" %c-%c",
+			    mask & 1 ? '-' : '<', mask & 2 ? '-' : '>');
 		}
 	}
 
 	if (show_mem || (show_tcpinfo && s->type != IPPROTO_UDP)) {
-		printf("\n\t");
+		out("\n\t");
 		if (s->type == IPPROTO_SCTP)
 			sctp_show_info(nlh, r, tb);
 		else
@@ -2537,7 +2547,7 @@ static int inet_show_sock(struct nlmsghdr *nlh,
 	}
 	sctp_ino = s->ino;
 
-	printf("\n");
+	out("\n");
 	return 0;
 }
 
@@ -2993,9 +3003,9 @@ static int dgram_show_line(char *line, const struct filter *f, int family)
 	inet_stats_print(&s, false);
 
 	if (show_details && opt[0])
-		printf(" opt:\"%s\"", opt);
+		out(" opt:\"%s\"", opt);
 
-	printf("\n");
+	out("\n");
 	return 0;
 }
 
@@ -3170,10 +3180,11 @@ static int unix_show_sock(const struct sockaddr_nl *addr, struct nlmsghdr *nlh,
 			unsigned char mask;
 
 			mask = rta_getattr_u8(tb[UNIX_DIAG_SHUTDOWN]);
-			printf(" %c-%c", mask & 1 ? '-' : '<', mask & 2 ? '-' : '>');
+			out(" %c-%c",
+			    mask & 1 ? '-' : '<', mask & 2 ? '-' : '>');
 		}
 	}
-	printf("\n");
+	out("\n");
 
 	return 0;
 }
@@ -3330,7 +3341,7 @@ static int unix_show(struct filter *f)
 		if (++cnt > MAX_UNIX_REMEMBER) {
 			while (list) {
 				unix_stats_print(list, f);
-				printf("\n");
+				out("\n");
 
 				unix_list_drop_first(&list);
 			}
@@ -3340,7 +3351,7 @@ static int unix_show(struct filter *f)
 	fclose(fp);
 	while (list) {
 		unix_stats_print(list, f);
-		printf("\n");
+		out("\n");
 
 		unix_list_drop_first(&list);
 	}
@@ -3386,12 +3397,12 @@ static int packet_stats_print(struct sockstat *s, const struct filter *f)
 
 static void packet_show_ring(struct packet_diag_ring *ring)
 {
-	printf("blk_size:%d", ring->pdr_block_size);
-	printf(",blk_nr:%d", ring->pdr_block_nr);
-	printf(",frm_size:%d", ring->pdr_frame_size);
-	printf(",frm_nr:%d", ring->pdr_frame_nr);
-	printf(",tmo:%d", ring->pdr_retire_tmo);
-	printf(",features:0x%x", ring->pdr_features);
+	out("blk_size:%d", ring->pdr_block_size);
+	out(",blk_nr:%d", ring->pdr_block_nr);
+	out(",frm_size:%d", ring->pdr_frame_size);
+	out(",frm_nr:%d", ring->pdr_frame_nr);
+	out(",tmo:%d", ring->pdr_retire_tmo);
+	out(",features:0x%x", ring->pdr_features);
 }
 
 static int packet_show_sock(const struct sockaddr_nl *addr,
@@ -3449,56 +3460,56 @@ static int packet_show_sock(const struct sockaddr_nl *addr,
 
 	if (show_details) {
 		if (pinfo) {
-			printf("\n\tver:%d", pinfo->pdi_version);
-			printf(" cpy_thresh:%d", pinfo->pdi_copy_thresh);
-			printf(" flags( ");
+			out("\n\tver:%d", pinfo->pdi_version);
+			out(" cpy_thresh:%d", pinfo->pdi_copy_thresh);
+			out(" flags( ");
 			if (pinfo->pdi_flags & PDI_RUNNING)
-				printf("running");
+				out("running");
 			if (pinfo->pdi_flags & PDI_AUXDATA)
-				printf(" auxdata");
+				out(" auxdata");
 			if (pinfo->pdi_flags & PDI_ORIGDEV)
-				printf(" origdev");
+				out(" origdev");
 			if (pinfo->pdi_flags & PDI_VNETHDR)
-				printf(" vnethdr");
+				out(" vnethdr");
 			if (pinfo->pdi_flags & PDI_LOSS)
-				printf(" loss");
+				out(" loss");
 			if (!pinfo->pdi_flags)
-				printf("0");
-			printf(" )");
+				out("0");
+			out(" )");
 		}
 		if (ring_rx) {
-			printf("\n\tring_rx(");
+			out("\n\tring_rx(");
 			packet_show_ring(ring_rx);
-			printf(")");
+			out(")");
 		}
 		if (ring_tx) {
-			printf("\n\tring_tx(");
+			out("\n\tring_tx(");
 			packet_show_ring(ring_tx);
-			printf(")");
+			out(")");
 		}
 		if (has_fanout) {
 			uint16_t type = (fanout >> 16) & 0xffff;
 
-			printf("\n\tfanout(");
-			printf("id:%d,", fanout & 0xffff);
-			printf("type:");
+			out("\n\tfanout(");
+			out("id:%d,", fanout & 0xffff);
+			out("type:");
 
 			if (type == 0)
-				printf("hash");
+				out("hash");
 			else if (type == 1)
-				printf("lb");
+				out("lb");
 			else if (type == 2)
-				printf("cpu");
+				out("cpu");
 			else if (type == 3)
-				printf("roll");
+				out("roll");
 			else if (type == 4)
-				printf("random");
+				out("random");
 			else if (type == 5)
-				printf("qm");
+				out("qm");
 			else
-				printf("0x%x", type);
+				out("0x%x", type);
 
-			printf(")");
+			out(")");
 		}
 	}
 
@@ -3508,15 +3519,15 @@ static int packet_show_sock(const struct sockaddr_nl *addr,
 		int num = RTA_PAYLOAD(tb[PACKET_DIAG_FILTER]) /
 			  sizeof(struct sock_filter);
 
-		printf("\n\tbpf filter (%d): ", num);
+		out("\n\tbpf filter (%d): ", num);
 		while (num) {
-			printf(" 0x%02x %u %u %u,",
-			      fil->code, fil->jt, fil->jf, fil->k);
+			out(" 0x%02x %u %u %u,",
+			    fil->code, fil->jt, fil->jf, fil->k);
 			num--;
 			fil++;
 		}
 	}
-	printf("\n");
+	out("\n");
 	return 0;
 }
 
@@ -3559,7 +3570,7 @@ static int packet_show_line(char *buf, const struct filter *f, int fam)
 	if (packet_stats_print(&stat, f))
 		return 0;
 
-	printf("\n");
+	out("\n");
 	return 0;
 }
 
@@ -3672,14 +3683,14 @@ static int netlink_show_one(struct filter *f,
 		else if (pid > 0)
 			getpidcon(pid, &pid_context);
 
-		printf(" proc_ctx=%s", pid_context ? : "unavailable");
+		out(" proc_ctx=%s", pid_context ? : "unavailable");
 		free(pid_context);
 	}
 
 	if (show_details) {
-		printf(" sk=%llx cb=%llx groups=0x%08x", sk, cb, groups);
+		out(" sk=%llx cb=%llx groups=0x%08x", sk, cb, groups);
 	}
-	printf("\n");
+	out("\n");
 
 	return 0;
 }
@@ -3715,9 +3726,9 @@ static int netlink_show_sock(const struct sockaddr_nl *addr,
 	}
 
 	if (show_mem) {
-		printf("\t");
+		out("\t");
 		print_skmeminfo(tb, NETLINK_DIAG_MEMINFO);
-		printf("\n");
+		out("\n");
 	}
 
 	return 0;
@@ -3808,7 +3819,7 @@ static void vsock_stats_print(struct sockstat *s, struct filter *f)
 
 	proc_ctx_print(s);
 
-	printf("\n");
+	out("\n");
 }
 
 static int vsock_show_sock(const struct sockaddr_nl *addr,
@@ -4573,10 +4584,10 @@ int main(int argc, char *argv[])
 
 	if (show_header) {
 		if (netid_width)
-			printf("%-*s ", netid_width, "Netid");
+			out("%-*s ", netid_width, "Netid");
 		if (state_width)
-			printf("%-*s ", state_width, "State");
-		printf("%-6s %-6s %s", "Recv-Q", "Send-Q", odd_width_pad);
+			out("%-*s ", state_width, "State");
+		out("%-6s %-6s %s", "Recv-Q", "Send-Q", odd_width_pad);
 	}
 
 	/* Make enough space for the local/remote port field */
@@ -4584,9 +4595,9 @@ int main(int argc, char *argv[])
 	serv_width += 13;
 
 	if (show_header) {
-		printf("%*s:%-*s %*s:%-*s\n",
-		       addr_width, "Local Address", serv_width, "Port",
-		       addr_width, "Peer Address", serv_width, "Port");
+		out("%*s:%-*s %*s:%-*s\n",
+		    addr_width, "Local Address", serv_width, "Port",
+		    addr_width, "Peer Address", serv_width, "Port");
 	}
 
 	fflush(stdout);
Signature · 1 line
2.9.4
#2 lore raw

PATCH v2 2/4 ss: Introduce columns lightweight abstraction

In reply to Stefano Brivio

Instead of embedding spacing directly while printing contents, logically declare columns and functions to buffer their content, to print left and right spacing around fields, to flush them to screen, and to print headers.

This makes it a bit easier to handle layout changes and prepares for full output buffering, needed for optimal spacing in field output layout.

Columns are currently set up to retain exactly the same output as before. This needs some slight adjustments of the values previously calculated in main(), as the width value introduced here already includes the width of left delimiters and spacing is not explicitly printed anymore whenever a field is printed. These calculations will go away altogether once automatic width calculation is implemented.

We can also remove explicit printing of newlines after the final content for a given line is printed, flushing the last field on a line will cause field_flush() to print newlines where appropriate.

No changes in output expected here.

---
v2: rebase after conflict with 00ac78d39c29 ("ss: print tcpi_rcv_ssthresh")

 misc/ss.c | 291 ++++++++++++++++++++++++++++++++++++++++++--------------------
 1 file changed, 198 insertions(+), 93 deletions(-)
misc/ss.c +198 -93
index a7d3b89e1478..42310ba4120d 100644
@@ -103,11 +103,48 @@ int show_header = 1;
 int follow_events;
 int sctp_ino;
 
-int netid_width;
-int state_width;
-int addr_width;
-int serv_width;
-char *odd_width_pad = "";
+enum col_id {
+	COL_NETID,
+	COL_STATE,
+	COL_RECVQ,
+	COL_SENDQ,
+	COL_ADDR,
+	COL_SERV,
+	COL_RADDR,
+	COL_RSERV,
+	COL_EXT,
+	COL_MAX
+};
+
+enum col_align {
+	ALIGN_LEFT,
+	ALIGN_CENTER,
+	ALIGN_RIGHT
+};
+
+struct column {
+	const enum col_align align;
+	const char *header;
+	const char *ldelim;
+	int width;	/* Including delimiter. -1: fit to content, 0: hide */
+	int stored;	/* Characters buffered */
+	int printed;	/* Characters printed so far */
+};
+
+static struct column columns[] = {
+	{ ALIGN_LEFT,	"Netid",		"",	0,	0,	0 },
+	{ ALIGN_LEFT,	"State",		" ",	0,	0,	0 },
+	{ ALIGN_LEFT,	"Recv-Q",		" ",	7,	0,	0 },
+	{ ALIGN_LEFT,	"Send-Q",		" ",	7,	0,	0 },
+	{ ALIGN_RIGHT,	"Local Address:",	" ",	0,	0,	0 },
+	{ ALIGN_LEFT,	"Port",			"",	0,	0,	0 },
+	{ ALIGN_RIGHT,	"Peer Address:",	" ",	0,	0,	0 },
+	{ ALIGN_LEFT,	"Port",			"",	0,	0,	0 },
+	{ ALIGN_LEFT,	"",			"",	-1,	0,	0 },
+};
+
+static struct column *current_field = columns;
+static char field_buf[BUFSIZ];
 
 static const char *TCP_PROTO = "tcp";
 static const char *SCTP_PROTO = "sctp";
@@ -826,13 +863,113 @@ static const char *vsock_netid_name(int type)
 
 static void out(const char *fmt, ...)
 {
+	struct column *f = current_field;
 	va_list args;
 
 	va_start(args, fmt);
-	vfprintf(stdout, fmt, args);
+	f->stored += vsnprintf(field_buf + f->stored, BUFSIZ - f->stored,
+			       fmt, args);
 	va_end(args);
 }
 
+static int print_left_spacing(struct column *f)
+{
+	int s;
+
+	if (f->width < 0 || f->align == ALIGN_LEFT)
+		return 0;
+
+	s = f->width - f->stored - f->printed;
+	if (f->align == ALIGN_CENTER)
+		/* If count of total spacing is odd, shift right by one */
+		s = (s + 1) / 2;
+
+	if (s > 0)
+		return printf("%*c", s, ' ');
+
+	return 0;
+}
+
+static void print_right_spacing(struct column *f)
+{
+	int s;
+
+	if (f->width < 0 || f->align == ALIGN_RIGHT)
+		return;
+
+	s = f->width - f->printed;
+	if (f->align == ALIGN_CENTER)
+		s /= 2;
+
+	if (s > 0)
+		printf("%*c", s, ' ');
+}
+
+static int field_needs_delimiter(struct column *f)
+{
+	if (!f->stored)
+		return 0;
+
+	/* Was another field already printed on this line? */
+	for (f--; f >= columns; f--)
+		if (f->width)
+			return 1;
+
+	return 0;
+}
+
+/* Flush given field to screen together with delimiter and spacing */
+static void field_flush(struct column *f)
+{
+	if (!f->width)
+		return;
+
+	if (field_needs_delimiter(f))
+		f->printed = printf("%s", f->ldelim);
+
+	f->printed += print_left_spacing(f);
+	f->printed += printf("%s", field_buf);
+	print_right_spacing(f);
+
+	*field_buf = 0;
+	f->printed = 0;
+	f->stored = 0;
+}
+
+static int field_is_last(struct column *f)
+{
+	return f - columns == COL_MAX - 1;
+}
+
+static void field_next(void)
+{
+	field_flush(current_field);
+
+	if (field_is_last(current_field)) {
+		printf("\n");
+		current_field = columns;
+	} else {
+		current_field++;
+	}
+}
+
+/* Walk through fields and flush them until we reach the desired one */
+static void field_set(enum col_id id)
+{
+	while (id != current_field - columns)
+		field_next();
+}
+
+/* Print header for all non-empty columns */
+static void print_header(void)
+{
+	while (!field_is_last(current_field)) {
+		if (current_field->width)
+			out(current_field->header);
+		field_next();
+	}
+}
+
 static void sock_state_print(struct sockstat *s)
 {
 	const char *sock_name;
@@ -872,18 +1009,21 @@ static void sock_state_print(struct sockstat *s)
 		sock_name = "unknown";
 	}
 
-	if (netid_width)
-		out("%-*s ", netid_width,
-		    is_sctp_assoc(s, sock_name) ? "" : sock_name);
-	if (state_width) {
-		if (is_sctp_assoc(s, sock_name))
-			out("`- %-*s ", state_width - 3,
-			    sctp_sstate_name[s->state]);
-		else
-			out("%-*s ", state_width, sstate_name[s->state]);
+	if (is_sctp_assoc(s, sock_name)) {
+		field_set(COL_STATE);		/* Empty Netid field */
+		out("`- %s", sctp_sstate_name[s->state]);
+	} else {
+		field_set(COL_NETID);
+		out("%s", sock_name);
+		field_set(COL_STATE);
+		out("%s", sstate_name[s->state]);
 	}
 
-	out("%-6d %-6d %s", s->rq, s->wq, odd_width_pad);
+	field_set(COL_RECVQ);
+	out("%-6d", s->rq);
+	field_set(COL_SENDQ);
+	out("%-6d", s->wq);
+	field_set(COL_ADDR);
 }
 
 static void sock_details_print(struct sockstat *s)
@@ -898,21 +1038,17 @@ static void sock_details_print(struct sockstat *s)
 		out(" fwmark:0x%x", s->mark);
 }
 
-static void sock_addr_print_width(int addr_len, const char *addr, char *delim,
-		int port_len, const char *port, const char *ifname)
-{
-	if (ifname) {
-		out("%*s%%%s%s%-*s ", addr_len, addr, ifname, delim,
-		    port_len, port);
-	} else {
-		out("%*s%s%-*s ", addr_len, addr, delim, port_len, port);
-	}
-}
-
 static void sock_addr_print(const char *addr, char *delim, const char *port,
 		const char *ifname)
 {
-	sock_addr_print_width(addr_width, addr, delim, serv_width, port, ifname);
+	if (ifname)
+		out("%s" "%%" "%s%s", addr, ifname, delim);
+	else
+		out("%s%s", addr, delim);
+
+	field_next();
+	out("%s", port);
+	field_next();
 }
 
 static const char *print_ms_timer(unsigned int timeout)
@@ -1093,7 +1229,6 @@ static void inet_addr_print(const inet_prefix *a, int port,
 {
 	char buf[1024];
 	const char *ap = buf;
-	int est_len = addr_width;
 	const char *ifname = NULL;
 
 	if (a->family == AF_INET) {
@@ -1112,24 +1247,13 @@ static void inet_addr_print(const inet_prefix *a, int port,
 					 "[%s]", ap);
 				ap = buf;
 			}
-
-			est_len = strlen(ap);
-			if (est_len <= addr_width)
-				est_len = addr_width;
-			else
-				est_len = addr_width + ((est_len-addr_width+3)/4)*4;
 		}
 	}
 
-	if (ifindex) {
-		ifname   = ll_index_to_name(ifindex);
-		est_len -= strlen(ifname) + 1;  /* +1 for percent char */
-		if (est_len < 0)
-			est_len = 0;
-	}
+	if (ifindex)
+		ifname = ll_index_to_name(ifindex);
 
-	sock_addr_print_width(est_len, ap, ":", serv_width, resolve_service(port),
-			ifname);
+	sock_addr_print(ap, ":", resolve_service(port), ifname);
 }
 
 struct aafilter {
@@ -2166,7 +2290,6 @@ static int tcp_show_line(char *line, const struct filter *f, int family)
 	if (show_tcpinfo)
 		tcp_stats_print(&s);
 
-	out("\n");
 	return 0;
 }
 
@@ -2547,7 +2670,6 @@ static int inet_show_sock(struct nlmsghdr *nlh,
 	}
 	sctp_ino = s->ino;
 
-	out("\n");
 	return 0;
 }
 
@@ -3005,7 +3127,6 @@ static int dgram_show_line(char *line, const struct filter *f, int family)
 	if (show_details && opt[0])
 		out(" opt:\"%s\"", opt);
 
-	out("\n");
 	return 0;
 }
 
@@ -3184,7 +3305,6 @@ static int unix_show_sock(const struct sockaddr_nl *addr, struct nlmsghdr *nlh,
 			    mask & 1 ? '-' : '<', mask & 2 ? '-' : '>');
 		}
 	}
-	out("\n");
 
 	return 0;
 }
@@ -3341,8 +3461,6 @@ static int unix_show(struct filter *f)
 		if (++cnt > MAX_UNIX_REMEMBER) {
 			while (list) {
 				unix_stats_print(list, f);
-				out("\n");
-
 				unix_list_drop_first(&list);
 			}
 			cnt = 0;
@@ -3351,8 +3469,6 @@ static int unix_show(struct filter *f)
 	fclose(fp);
 	while (list) {
 		unix_stats_print(list, f);
-		out("\n");
-
 		unix_list_drop_first(&list);
 	}
 
@@ -3527,7 +3643,6 @@ static int packet_show_sock(const struct sockaddr_nl *addr,
 			fil++;
 		}
 	}
-	out("\n");
 	return 0;
 }
 
@@ -3570,7 +3685,6 @@ static int packet_show_line(char *buf, const struct filter *f, int fam)
 	if (packet_stats_print(&stat, f))
 		return 0;
 
-	out("\n");
 	return 0;
 }
 
@@ -3690,7 +3804,6 @@ static int netlink_show_one(struct filter *f,
 	if (show_details) {
 		out(" sk=%llx cb=%llx groups=0x%08x", sk, cb, groups);
 	}
-	out("\n");
 
 	return 0;
 }
@@ -3728,7 +3841,6 @@ static int netlink_show_sock(const struct sockaddr_nl *addr,
 	if (show_mem) {
 		out("\t");
 		print_skmeminfo(tb, NETLINK_DIAG_MEMINFO);
-		out("\n");
 	}
 
 	return 0;
@@ -3818,8 +3930,6 @@ static void vsock_stats_print(struct sockstat *s, struct filter *f)
 	vsock_addr_print(&s->remote, s->rport);
 
 	proc_ctx_print(s);
-
-	out("\n");
 }
 
 static int vsock_show_sock(const struct sockaddr_nl *addr,
@@ -4539,13 +4649,17 @@ int main(int argc, char *argv[])
 	if (ssfilter_parse(&current_filter.f, argc, argv, filter_fp))
 		usage();
 
-	netid_width = 0;
 	if (current_filter.dbs&(current_filter.dbs-1))
-		netid_width = 5;
+		columns[COL_NETID].width = 6;
 
-	state_width = 0;
 	if (current_filter.states&(current_filter.states-1))
-		state_width = 10;
+		columns[COL_STATE].width = 10;
+
+	/* If Netid or State are hidden, no delimiter before next column */
+	if (!columns[COL_NETID].width)
+		columns[COL_STATE].width--;
+	else if (!columns[COL_STATE].width)
+		columns[COL_RECVQ].width--;
 
 	if (isatty(STDOUT_FILENO)) {
 		struct winsize w;
@@ -4556,49 +4670,38 @@ int main(int argc, char *argv[])
 		}
 	}
 
-	addrp_width = screen_width;
-	if (netid_width)
-		addrp_width -= netid_width + 1;
-	if (state_width)
-		addrp_width -= state_width + 1;
-	addrp_width -= 14;
+	addrp_width = screen_width -
+		      columns[COL_NETID].width -
+		      columns[COL_STATE].width -
+		      columns[COL_RECVQ].width -
+		      columns[COL_SENDQ].width;
 
 	if (addrp_width&1) {
-		if (netid_width)
-			netid_width++;
-		else if (state_width)
-			state_width++;
+		if (columns[COL_NETID].width)
+			columns[COL_NETID].width++;
+		else if (columns[COL_STATE].width)
+			columns[COL_STATE].width++;
 		else
-			odd_width_pad = " ";
+			columns[COL_SENDQ].width++;
 	}
 
 	addrp_width /= 2;
-	addrp_width--;
 
-	serv_width = resolve_services ? 7 : 5;
+	columns[COL_SERV].width = resolve_services ? 8 : 6;
+	if (addrp_width < 15 + columns[COL_SERV].width)
+		addrp_width = 15 + columns[COL_SERV].width;
 
-	if (addrp_width < 15+serv_width+1)
-		addrp_width = 15+serv_width+1;
-
-	addr_width = addrp_width - serv_width - 1;
-
-	if (show_header) {
-		if (netid_width)
-			out("%-*s ", netid_width, "Netid");
-		if (state_width)
-			out("%-*s ", state_width, "State");
-		out("%-6s %-6s %s", "Recv-Q", "Send-Q", odd_width_pad);
-	}
+	columns[COL_ADDR].width = addrp_width - columns[COL_SERV].width;
 
 	/* Make enough space for the local/remote port field */
-	addr_width -= 13;
-	serv_width += 13;
+	columns[COL_ADDR].width -= 13;
+	columns[COL_SERV].width += 13;
 
-	if (show_header) {
-		out("%*s:%-*s %*s:%-*s\n",
-		    addr_width, "Local Address", serv_width, "Port",
-		    addr_width, "Peer Address", serv_width, "Port");
-	}
+	columns[COL_RADDR].width = columns[COL_ADDR].width;
+	columns[COL_RSERV].width = columns[COL_SERV].width;
+
+	if (show_header)
+		print_header();
 
 	fflush(stdout);
 
@@ -4627,5 +4730,7 @@ int main(int argc, char *argv[])
 	if (show_users || show_proc_ctx || show_sock_ctx)
 		user_ent_destroy();
 
+	field_next();
+
 	return 0;
 }
Signature · 1 line
2.9.4
#3 lore raw

PATCH v2 3/4 ss: Buffer raw fields first, then render them as a table

In reply to Stefano Brivio

This allows us to measure the maximum field length for each column before printing fields and will permit us to apply optimal field spacing and distribution. Structure of the output buffer with chunked allocation is described in comments.

Output is still unchanged, original spacing is used.

Running over one million sockets with -tul options by simply modifying main() to loop 50,000 times over the *_show() functions, buffering the whole output and rendering it at the end, with 10 UDP sockets, 10 TCP sockets, while throwing output away, doesn't show significant changes in execution time on my laptop with an Intel i7-6600U CPU:

- before this patch:

$ time ./ss -tul > /dev/null
real	0m29.899s
user	0m2.017s
sys	0m27.801s

- after this patch:

$ time ./ss -tul > /dev/null
real	0m29.827s
user	0m1.942s
sys	0m27.812s

---
v2: rebase after conflict with 00ac78d39c29 ("ss: print tcpi_rcv_ssthresh")

 misc/ss.c | 271 +++++++++++++++++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 225 insertions(+), 46 deletions(-)
misc/ss.c +225 -46
index 42310ba4120d..166267974c36 100644
@@ -47,6 +47,8 @@
 #include <linux/vm_sockets_diag.h>
 
 #define MAGIC_SEQ 123456
+#define BUF_CHUNK (1024 * 1024)
+#define LEN_ALIGN(x) (((x) + 1) & ~1)
 
 #define DIAG_REQUEST(_req, _r)						    \
 	struct {							    \
@@ -127,24 +129,45 @@ struct column {
 	const char *header;
 	const char *ldelim;
 	int width;	/* Including delimiter. -1: fit to content, 0: hide */
-	int stored;	/* Characters buffered */
-	int printed;	/* Characters printed so far */
 };
 
 static struct column columns[] = {
-	{ ALIGN_LEFT,	"Netid",		"",	0,	0,	0 },
-	{ ALIGN_LEFT,	"State",		" ",	0,	0,	0 },
-	{ ALIGN_LEFT,	"Recv-Q",		" ",	7,	0,	0 },
-	{ ALIGN_LEFT,	"Send-Q",		" ",	7,	0,	0 },
-	{ ALIGN_RIGHT,	"Local Address:",	" ",	0,	0,	0 },
-	{ ALIGN_LEFT,	"Port",			"",	0,	0,	0 },
-	{ ALIGN_RIGHT,	"Peer Address:",	" ",	0,	0,	0 },
-	{ ALIGN_LEFT,	"Port",			"",	0,	0,	0 },
-	{ ALIGN_LEFT,	"",			"",	-1,	0,	0 },
+	{ ALIGN_LEFT,	"Netid",		"",	0 },
+	{ ALIGN_LEFT,	"State",		" ",	0 },
+	{ ALIGN_LEFT,	"Recv-Q",		" ",	7 },
+	{ ALIGN_LEFT,	"Send-Q",		" ",	7 },
+	{ ALIGN_RIGHT,	"Local Address:",	" ",	0 },
+	{ ALIGN_LEFT,	"Port",			"",	0 },
+	{ ALIGN_RIGHT,	"Peer Address:",	" ",	0 },
+	{ ALIGN_LEFT,	"Port",			"",	0 },
+	{ ALIGN_LEFT,	"",			"",	-1 },
 };
 
 static struct column *current_field = columns;
-static char field_buf[BUFSIZ];
+
+/* Output buffer: chained chunks of BUF_CHUNK bytes. Each field is written to
+ * the buffer as a variable size token. A token consists of a 16 bits length
+ * field, followed by a string which is not NULL-terminated.
+ *
+ * A new chunk is allocated and linked when the current chunk doesn't have
+ * enough room to store the current token as a whole.
+ */
+struct buf_chunk {
+	struct buf_chunk *next;	/* Next chained chunk */
+	char *end;		/* Current end of content */
+	char data[0];
+};
+
+struct buf_token {
+	uint16_t len;		/* Data length, excluding length descriptor */
+	char data[0];
+};
+
+static struct {
+	struct buf_token *cur;	/* Position of current token in chunk */
+	struct buf_chunk *head;	/* First chunk */
+	struct buf_chunk *tail;	/* Current chunk */
+} buffer;
 
 static const char *TCP_PROTO = "tcp";
 static const char *SCTP_PROTO = "sctp";
@@ -861,25 +884,109 @@ static const char *vsock_netid_name(int type)
 	}
 }
 
+/* Allocate and initialize a new buffer chunk */
+static struct buf_chunk *buf_chunk_new(void)
+{
+	struct buf_chunk *new = malloc(BUF_CHUNK);
+
+	if (!new)
+		abort();
+
+	new->next = NULL;
+
+	/* This is also the last block */
+	buffer.tail = new;
+
+	/* Next token will be stored at the beginning of chunk data area, and
+	 * its initial length is zero.
+	 */
+	buffer.cur = (struct buf_token *)new->data;
+	buffer.cur->len = 0;
+
+	new->end = buffer.cur->data;
+
+	return new;
+}
+
+/* Return available tail room in given chunk */
+static int buf_chunk_avail(struct buf_chunk *chunk)
+{
+	return BUF_CHUNK - offsetof(struct buf_chunk, data) -
+	       (chunk->end - chunk->data);
+}
+
+/* Update end pointer and token length, link new chunk if we hit the end of the
+ * current one. Return -EAGAIN if we got a new chunk, caller has to print again.
+ */
+static int buf_update(int len)
+{
+	struct buf_chunk *chunk = buffer.tail;
+	struct buf_token *t = buffer.cur;
+
+	/* Claim success if new content fits in the current chunk, and anyway
+	 * if this is the first token in the chunk: in the latter case,
+	 * allocating a new chunk won't help, so we'll just cut the output.
+	 */
+	if ((len < buf_chunk_avail(chunk) && len != -1 /* glibc < 2.0.6 */) ||
+	    t == (struct buf_token *)chunk->data) {
+		len = min(len, buf_chunk_avail(chunk));
+
+		/* Total field length can't exceed 2^16 bytes, cut as needed */
+		len = min(len, USHRT_MAX - t->len);
+
+		chunk->end += len;
+		t->len += len;
+		return 0;
+	}
+
+	/* Content truncated, time to allocate more */
+	chunk->next = buf_chunk_new();
+
+	/* Copy current token over to new chunk, including length descriptor */
+	memcpy(chunk->next->data, t, sizeof(t->len) + t->len);
+	chunk->next->end += t->len;
+
+	/* Discard partially written field in old chunk */
+	chunk->end -= t->len + sizeof(t->len);
+
+	return -EAGAIN;
+}
+
+/* Append content to buffer as part of the current field */
 static void out(const char *fmt, ...)
 {
 	struct column *f = current_field;
 	va_list args;
+	char *pos;
+	int len;
+
+	if (!f->width)
+		return;
+
+	if (!buffer.head)
+		buffer.head = buf_chunk_new();
+
+again:	/* Append to buffer: if we have a new chunk, print again */
 
+	pos = buffer.cur->data + buffer.cur->len;
 	va_start(args, fmt);
-	f->stored += vsnprintf(field_buf + f->stored, BUFSIZ - f->stored,
-			       fmt, args);
+
+	/* Limit to tail room. If we hit the limit, buf_update() will tell us */
+	len = vsnprintf(pos, buf_chunk_avail(buffer.tail), fmt, args);
 	va_end(args);
+
+	if (buf_update(len))
+		goto again;
 }
 
-static int print_left_spacing(struct column *f)
+static int print_left_spacing(struct column *f, int stored, int printed)
 {
 	int s;
 
 	if (f->width < 0 || f->align == ALIGN_LEFT)
 		return 0;
 
-	s = f->width - f->stored - f->printed;
+	s = f->width - stored - printed;
 	if (f->align == ALIGN_CENTER)
 		/* If count of total spacing is odd, shift right by one */
 		s = (s + 1) / 2;
@@ -890,14 +997,14 @@ static int print_left_spacing(struct column *f)
 	return 0;
 }
 
-static void print_right_spacing(struct column *f)
+static void print_right_spacing(struct column *f, int printed)
 {
 	int s;
 
 	if (f->width < 0 || f->align == ALIGN_RIGHT)
 		return;
 
-	s = f->width - f->printed;
+	s = f->width - printed;
 	if (f->align == ALIGN_CENTER)
 		s /= 2;
 
@@ -905,35 +1012,29 @@ static void print_right_spacing(struct column *f)
 		printf("%*c", s, ' ');
 }
 
-static int field_needs_delimiter(struct column *f)
-{
-	if (!f->stored)
-		return 0;
-
-	/* Was another field already printed on this line? */
-	for (f--; f >= columns; f--)
-		if (f->width)
-			return 1;
-
-	return 0;
-}
-
-/* Flush given field to screen together with delimiter and spacing */
+/* Done with field: update buffer pointer, start new token after current one */
 static void field_flush(struct column *f)
 {
+	struct buf_chunk *chunk = buffer.tail;
+	unsigned int pad = buffer.cur->len % 2;
+
 	if (!f->width)
 		return;
 
-	if (field_needs_delimiter(f))
-		f->printed = printf("%s", f->ldelim);
-
-	f->printed += print_left_spacing(f);
-	f->printed += printf("%s", field_buf);
-	print_right_spacing(f);
+	/* We need a new chunk if we can't store the next length descriptor.
+	 * Mind the gap between end of previous token and next aligned position
+	 * for length descriptor.
+	 */
+	if (buf_chunk_avail(chunk) - pad < sizeof(buffer.cur->len)) {
+		chunk->end += pad;
+		chunk->next = buf_chunk_new();
+		return;
+	}
 
-	*field_buf = 0;
-	f->printed = 0;
-	f->stored = 0;
+	buffer.cur = (struct buf_token *)(buffer.cur->data +
+					  LEN_ALIGN(buffer.cur->len));
+	buffer.cur->len = 0;
+	buffer.tail->end = buffer.cur->data;
 }
 
 static int field_is_last(struct column *f)
@@ -945,12 +1046,10 @@ static void field_next(void)
 {
 	field_flush(current_field);
 
-	if (field_is_last(current_field)) {
-		printf("\n");
+	if (field_is_last(current_field))
 		current_field = columns;
-	} else {
+	else
 		current_field++;
-	}
 }
 
 /* Walk through fields and flush them until we reach the desired one */
@@ -970,6 +1069,86 @@ static void print_header(void)
 	}
 }
 
+/* Get the next available token in the buffer starting from the current token */
+static struct buf_token *buf_token_next(struct buf_token *cur)
+{
+	struct buf_chunk *chunk = buffer.tail;
+
+	/* If we reached the end of chunk contents, get token from next chunk */
+	if (cur->data + LEN_ALIGN(cur->len) == chunk->end) {
+		buffer.tail = chunk = chunk->next;
+		return chunk ? (struct buf_token *)chunk->data : NULL;
+	}
+
+	return (struct buf_token *)(cur->data + LEN_ALIGN(cur->len));
+}
+
+/* Free up all allocated buffer chunks */
+static void buf_free_all(void)
+{
+	struct buf_chunk *tmp;
+
+	for (buffer.tail = buffer.head; buffer.tail; ) {
+		tmp = buffer.tail;
+		buffer.tail = buffer.tail->next;
+		free(tmp);
+	}
+	buffer.head = NULL;
+}
+
+/* Render buffered output with spacing and delimiters, then free up buffers */
+static void render(void)
+{
+	struct buf_token *token = (struct buf_token *)buffer.head->data;
+	int printed, line_started = 0, need_newline = 0;
+	struct column *f;
+
+	/* Ensure end alignment of last token, it wasn't necessarily flushed */
+	buffer.tail->end += buffer.cur->len % 2;
+
+	/* Rewind and replay */
+	buffer.tail = buffer.head;
+
+	f = columns;
+	while (!f->width)
+		f++;
+
+	while (token) {
+		/* Print left delimiter only if we already started a line */
+		if (line_started++)
+			printed = printf("%s", current_field->ldelim);
+		else
+			printed = 0;
+
+		/* Print field content from token data with spacing */
+		printed += print_left_spacing(f, token->len, printed);
+		printed += fwrite(token->data, 1, token->len, stdout);
+		print_right_spacing(f, printed);
+
+		/* Variable field size or overflow, won't align to screen */
+		if (printed > f->width)
+			need_newline = 1;
+
+		/* Go to next non-empty field, deal with end-of-line */
+		do {
+			if (field_is_last(f)) {
+				if (need_newline) {
+					printf("\n");
+					need_newline = 0;
+				}
+				f = columns;
+				line_started = 0;
+			} else {
+				f++;
+			}
+		} while (!f->width);
+
+		token = buf_token_next(token);
+	}
+
+	buf_free_all();
+}
+
 static void sock_state_print(struct sockstat *s)
 {
 	const char *sock_name;
@@ -4730,7 +4909,7 @@ int main(int argc, char *argv[])
 	if (show_users || show_proc_ctx || show_sock_ctx)
 		user_ent_destroy();
 
-	field_next();
+	render();
 
 	return 0;
 }
Signature · 1 line
2.9.4
#4 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 20 lines · Stefano Brivio

This allows us to measure the maximum field length for each column before printing fields and will permit us to apply optimal field spacing and distribution. Structure of the output buffer with chunked allocation is described in comments.

Output is still unchanged, original spacing is used.

Running over one million sockets with -tul options by simply modifying main() to loop 50,000 times over the *_show() functions, buffering the whole output and rendering it at the end, with 10 UDP sockets, 10 TCP sockets, while throwing output away, doesn't show significant changes in execution time on my laptop with an Intel i7-6600U CPU:

- before this patch:

$ time ./ss -tul > /dev/null
real	0m29.899s
user	0m2.017s
sys	0m27.801s
Quoted 6 lines

- after this patch:

$ time ./ss -tul > /dev/null
real	0m29.827s
user	0m1.942s
sys	0m27.812s

I do not get it.

"ss -emoi " uses almost 1KB per socket.

10,000,000 sockets -> we need about 10GB of memory ???

This is a serious regression.

#5 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 7 lines · On Tue, 12 Feb 2019 16:42:04 -0800 Eric Dumazet <[email protected]> wrote:

I do not get it.

"ss -emoi " uses almost 1KB per socket.

10,000,000 sockets -> we need about 10GB of memory ???

This is a serious regression.

I guess this is rather subjective: the worst case I considered back then was the output of 'ss -tei0' (less than 500 bytes) for one million sockets, which gives 500M of memory, which should in turn be fine on a machine handling one million sockets.

Now, if 'ss -emoi' on 10 million sockets is an actual use case (out of
curiosity: how are you going to process that output? Would JSON help?), I see two easy options to solve this:

1. flush the output every time we reach a given buffer size (1M perhaps). This might make the resulting blocks slightly unaligned, with occasional loss of readability on lines occurring every 1k to 10k sockets approximately, even though after 1k sockets column sizes won't change much (it looks anyway better than the original), and I don't expect anybody to actually scroll that output

2. add a switch for unbuffered output, but then you need to remember to pass it manually, and the whole output would be as bad as the original in case you need the switch.

I'd rather go with 1., it's easy to implement (we already have partial flushing with '--events') and it looks like a good compromise on usability. Thoughts?

Signature · 1 line
Stefano
#6 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 27 lines · On Wed, 13 Feb 2019 09:37:11 +0100 Stefano Brivio <[email protected]> wrote:
Quoted 7 lines · On Tue, 12 Feb 2019 16:42:04 -0800 Eric Dumazet <[email protected]> wrote:

I do not get it.

"ss -emoi " uses almost 1KB per socket.

10,000,000 sockets -> we need about 10GB of memory ???

This is a serious regression.

I guess this is rather subjective: the worst case I considered back then was the output of 'ss -tei0' (less than 500 bytes) for one million sockets, which gives 500M of memory, which should in turn be fine on a machine handling one million sockets.

Now, if 'ss -emoi' on 10 million sockets is an actual use case (out of
curiosity: how are you going to process that output? Would JSON help?), I see two easy options to solve this:

1. flush the output every time we reach a given buffer size (1M perhaps). This might make the resulting blocks slightly unaligned, with occasional loss of readability on lines occurring every 1k to 10k sockets approximately, even though after 1k sockets column sizes won't change much (it looks anyway better than the original), and I don't expect anybody to actually scroll that output

Quoted 8 lines

2. add a switch for unbuffered output, but then you need to remember to pass it manually, and the whole output would be as bad as the original in case you need the switch.

I'd rather go with 1., it's easy to implement (we already have partial flushing with '--events') and it looks like a good compromise on usability. Thoughts?

I agree with eric. The benefits of buffering are not worth it.
Let's just choose a reasonable field width, if something is too big, columns won't line up which i snot a big deal.

Unless you come up with a better solution, I am going to revert this.

#7 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 34 lines · On Wed, 13 Feb 2019 08:51:01 -0800 Stephen Hemminger…
Quoted 31 lines · On Wed, 13 Feb 2019 09:37:11 +0100 Stefano Brivio <[email protected]> wrote:
Quoted 7 lines · On Tue, 12 Feb 2019 16:42:04 -0800 Eric Dumazet <[email protected]> wrote:

I do not get it.

"ss -emoi " uses almost 1KB per socket.

10,000,000 sockets -> we need about 10GB of memory ???

This is a serious regression.

I guess this is rather subjective: the worst case I considered back then was the output of 'ss -tei0' (less than 500 bytes) for one million sockets, which gives 500M of memory, which should in turn be fine on a machine handling one million sockets.

Now, if 'ss -emoi' on 10 million sockets is an actual use case (out of
curiosity: how are you going to process that output? Would JSON help?), I see two easy options to solve this:

1. flush the output every time we reach a given buffer size (1M perhaps). This might make the resulting blocks slightly unaligned, with occasional loss of readability on lines occurring every 1k to 10k sockets approximately, even though after 1k sockets column sizes won't change much (it looks anyway better than the original), and I don't expect anybody to actually scroll that output

2. add a switch for unbuffered output, but then you need to remember to pass it manually, and the whole output would be as bad as the original in case you need the switch.

Quoted 7 lines
Quoted 4 lines

I'd rather go with 1., it's easy to implement (we already have partial flushing with '--events') and it looks like a good compromise on usability. Thoughts?

I agree with eric. The benefits of buffering are not worth it.
Let's just choose a reasonable field width, if something is too big, columns won't line up which i snot a big deal.

That's how it was before, and we couldn't even get fields aligned with TCP and UDP sockets in a 80 columns wide terminal. See examples at:
https://patchwork.ozlabs.org/cover/847301/.

I tried, but I think it's impossible to find a "reasonable" field width, especially when you mix a number of socket types.

Quoted 1 line

Unless you come up with a better solution, I am going to revert this.

That's why I asked for feedback about my proposals 1. and 2. above. I'll go for 1. then.

Signature · 1 line
Stefano
#8 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 3 lines · Stefano Brivio

That's why I asked for feedback about my proposals 1. and 2. above. I'll go for 1. then.

If stdout is a regular file or a pipe, just do not try to align things ?

#9 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

In reply to Stefano Brivio
Quoted 11 lines · Stefano Brivio
Quoted 7 lines · On Tue, 12 Feb 2019 16:42:04 -0800 Eric Dumazet <[email protected]> wrote:

I do not get it.

"ss -emoi " uses almost 1KB per socket.

10,000,000 sockets -> we need about 10GB of memory ???

This is a serious regression.

Quoted 8 lines

I guess this is rather subjective: the worst case I considered back then was the output of 'ss -tei0' (less than 500 bytes) for one million sockets, which gives 500M of memory, which should in turn be fine on a machine handling one million sockets.

Now, if 'ss -emoi' on 10 million sockets is an actual use case (out of
curiosity: how are you going to process that output? Would JSON help?), I see two easy options to solve this:

ss -temoi | parser (written in shell or awk or whatever...)

This is a use case, I just got bitten because using ss command actually OOM my container, while trying to debug a busy GFE.

The host itself can have 10,000,000 TCP sockets, but usually sysadmin shells run in a container with no more than 500 MB available.

Otherwise, it would be too easy for a buggy program to OOM the whole machine and have angry customers.

Quoted 8 lines

1. flush the output every time we reach a given buffer size (1M perhaps). This might make the resulting blocks slightly unaligned, with occasional loss of readability on lines occurring every 1k to 10k sockets approximately, even though after 1k sockets column sizes won't change much (it looks anyway better than the original), and I don't expect anybody to actually scroll that output

Quoted 8 lines

2. add a switch for unbuffered output, but then you need to remember to pass it manually, and the whole output would be as bad as the original in case you need the switch.

I'd rather go with 1., it's easy to implement (we already have partial flushing with '--events') and it looks like a good compromise on usability. Thoughts?

1 seems fine, but a switch for 'please do not try to format' would be fine.

I wonder why we try to 'format' when stdout is a pipe or a regular file .

#10 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 24 lines · On Wed, 13 Feb 2019 09:31:03 -0800 Eric Dumazet <[email protected]> wrote:
Quoted 11 lines · Stefano Brivio
Quoted 7 lines · On Tue, 12 Feb 2019 16:42:04 -0800 Eric Dumazet <[email protected]> wrote:

I do not get it.

"ss -emoi " uses almost 1KB per socket.

10,000,000 sockets -> we need about 10GB of memory ???

This is a serious regression.

Quoted 8 lines

I guess this is rather subjective: the worst case I considered back then was the output of 'ss -tei0' (less than 500 bytes) for one million sockets, which gives 500M of memory, which should in turn be fine on a machine handling one million sockets.

Now, if 'ss -emoi' on 10 million sockets is an actual use case (out of
curiosity: how are you going to process that output? Would JSON help?), I see two easy options to solve this:

ss -temoi | parser (written in shell or awk or whatever...)

Quoted 8 lines

This is a use case, I just got bitten because using ss command actually OOM my container, while trying to debug a busy GFE.

The host itself can have 10,000,000 TCP sockets, but usually sysadmin shells run in a container with no more than 500 MB available.

Otherwise, it would be too easy for a buggy program to OOM the whole machine and have angry customers.

Ouch, I see.

Quoted 12 lines
Quoted 12 lines

1. flush the output every time we reach a given buffer size (1M perhaps). This might make the resulting blocks slightly unaligned, with occasional loss of readability on lines occurring every 1k to 10k sockets approximately, even though after 1k sockets column sizes won't change much (it looks anyway better than the original), and I don't expect anybody to actually scroll that output

2. add a switch for unbuffered output, but then you need to remember to pass it manually, and the whole output would be as bad as the original in case you need the switch.

Quoted 6 lines
Quoted 4 lines

I'd rather go with 1., it's easy to implement (we already have partial flushing with '--events') and it looks like a good compromise on usability. Thoughts?

1 seems fine, but a switch for 'please do not try to format' would be fine.

Let me try with 1. first then -- we already have a huge number of switches.

Quoted 1 line

I wonder why we try to 'format' when stdout is a pipe or a regular file .

As stupid as it might sound, I just didn't think of fixing that :)

What would you suggest in that case, single whitespaces? Tabs?

Signature · 1 line
Stefano
#11 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 3 lines · Stefano Brivio

What would you suggest in that case, single whitespaces? Tabs?

We have spaces at the moment, I would leave them.

#12 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

In reply to Eric Dumazet
Quoted 45 lines · On Wed, 13 Feb 2019 09:31:03 -0800 Eric Dumazet <[email protected]> wrote:
Quoted 11 lines · Stefano Brivio
Quoted 7 lines · On Tue, 12 Feb 2019 16:42:04 -0800 Eric Dumazet <[email protected]> wrote:

I do not get it.

"ss -emoi " uses almost 1KB per socket.

10,000,000 sockets -> we need about 10GB of memory ???

This is a serious regression.

Quoted 8 lines

I guess this is rather subjective: the worst case I considered back then was the output of 'ss -tei0' (less than 500 bytes) for one million sockets, which gives 500M of memory, which should in turn be fine on a machine handling one million sockets.

Now, if 'ss -emoi' on 10 million sockets is an actual use case (out of
curiosity: how are you going to process that output? Would JSON help?), I see two easy options to solve this:

ss -temoi | parser (written in shell or awk or whatever...)

This is a use case, I just got bitten because using ss command actually OOM my container, while trying to debug a busy GFE.

The host itself can have 10,000,000 TCP sockets, but usually sysadmin shells run in a container with no more than 500 MB available.

Otherwise, it would be too easy for a buggy program to OOM the whole machine and have angry customers.

Quoted 12 lines

1. flush the output every time we reach a given buffer size (1M perhaps). This might make the resulting blocks slightly unaligned, with occasional loss of readability on lines occurring every 1k to 10k sockets approximately, even though after 1k sockets column sizes won't change much (it looks anyway better than the original), and I don't expect anybody to actually scroll that output

2. add a switch for unbuffered output, but then you need to remember to pass it manually, and the whole output would be as bad as the original in case you need the switch.

Quoted 8 lines
Quoted 4 lines

I'd rather go with 1., it's easy to implement (we already have partial flushing with '--events') and it looks like a good compromise on usability. Thoughts?

1 seems fine, but a switch for 'please do not try to format' would be fine.

I wonder why we try to 'format' when stdout is a pipe or a regular file .

On a second thought: what about | less, or | grep [ports], or > readable.log? I guess those might also be rather common use cases, what do you think?

I'm tempted to skip this for the moment and just go with option 1.

Signature · 1 line
Stefano
#13 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 55 lines · On Wed, 13 Feb 2019 22:17:16 +0100 Stefano Brivio <[email protected]> wrote:
Quoted 52 lines · On Wed, 13 Feb 2019 09:31:03 -0800 Eric Dumazet <[email protected]> wrote:
Quoted 11 lines · Stefano Brivio
Quoted 7 lines · On Tue, 12 Feb 2019 16:42:04 -0800 Eric Dumazet <[email protected]> wrote:

I do not get it.

"ss -emoi " uses almost 1KB per socket.

10,000,000 sockets -> we need about 10GB of memory ???

This is a serious regression.

Quoted 8 lines

I guess this is rather subjective: the worst case I considered back then was the output of 'ss -tei0' (less than 500 bytes) for one million sockets, which gives 500M of memory, which should in turn be fine on a machine handling one million sockets.

Now, if 'ss -emoi' on 10 million sockets is an actual use case (out of
curiosity: how are you going to process that output? Would JSON help?), I see two easy options to solve this:

ss -temoi | parser (written in shell or awk or whatever...)

This is a use case, I just got bitten because using ss command actually OOM my container, while trying to debug a busy GFE.

The host itself can have 10,000,000 TCP sockets, but usually sysadmin shells run in a container with no more than 500 MB available.

Otherwise, it would be too easy for a buggy program to OOM the whole machine and have angry customers.

Quoted 8 lines

1. flush the output every time we reach a given buffer size (1M perhaps). This might make the resulting blocks slightly unaligned, with occasional loss of readability on lines occurring every 1k to 10k sockets approximately, even though after 1k sockets column sizes won't change much (it looks anyway better than the original), and I don't expect anybody to actually scroll that output

Quoted 8 lines

2. add a switch for unbuffered output, but then you need to remember to pass it manually, and the whole output would be as bad as the original in case you need the switch.

I'd rather go with 1., it's easy to implement (we already have partial flushing with '--events') and it looks like a good compromise on usability. Thoughts?

1 seems fine, but a switch for 'please do not try to format' would be fine.

Quoted 8 lines
Quoted 1 line

I wonder why we try to 'format' when stdout is a pipe or a regular file .

On a second thought: what about | less, or | grep [ports], or > readable.log? I guess those might also be rather common use cases, what do you think?

I'm tempted to skip this for the moment and just go with option 1.

What I would favor:
* use big enough columns that for the common case everything lines up fine
* if column is to wide just print that element wider (which is what print %Ns does)
and
* add json output for programs that want to parse
* use print_uint etc for that

The buffering patch (in iproute2-next) can/will be reverted.

#14 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 3 lines · On Wed, 13 Feb 2019 13:55:34 -0800 Stephen Hemminger…

What I would favor:
* use big enough columns that for the common case everything lines up fine
* if column is to wide just print that element wider (which is what print %Ns does)

This is very close to what was done before, but as soon as you mix, say, UNIX sockets with TCP sockets, "big enough" columns typically make output for TCP sockets unreadable.

With buffering, instead, I can decide that a line split is needed, and keep fields aligned no matter what.

Quoted 3 lines

and
* add json output for programs that want to parse
* use print_uint etc for that

Sure, I think we all agree with this, but it's not going to be quick to implement (even though it's perhaps a bit easier with abstracted columns and buffering). Eric reported a problem and I'm trying to fix it quickly.

Quoted 1 line

The buffering patch (in iproute2-next) can/will be reverted.

I think it received generally good feedback (also from users, later on) and this is the first report of a serious issue -- it's also an issue which looks easy to fix (I'm half way through that by now).

By the way, this patch was merged in iproute2 more than one year ago (December 2017, by you).

Signature · 1 line
Stefano
#15 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

In reply to Stephen Hemminger

Hi Stephen,

Quoted 64 lines · Stephen Hemminger
Quoted 61 lines · On Wed, 13 Feb 2019 22:17:16 +0100 Stefano Brivio <[email protected]> wrote:
Quoted 45 lines · On Wed, 13 Feb 2019 09:31:03 -0800 Eric Dumazet <[email protected]> wrote:
Quoted 11 lines · Stefano Brivio
Quoted 7 lines · On Tue, 12 Feb 2019 16:42:04 -0800 Eric Dumazet <[email protected]> wrote:

I do not get it.

"ss -emoi " uses almost 1KB per socket.

10,000,000 sockets -> we need about 10GB of memory ???

This is a serious regression.

Quoted 8 lines

I guess this is rather subjective: the worst case I considered back then was the output of 'ss -tei0' (less than 500 bytes) for one million sockets, which gives 500M of memory, which should in turn be fine on a machine handling one million sockets.

Now, if 'ss -emoi' on 10 million sockets is an actual use case (out of
curiosity: how are you going to process that output? Would JSON help?), I see two easy options to solve this:

ss -temoi | parser (written in shell or awk or whatever...)

This is a use case, I just got bitten because using ss command actually OOM my container, while trying to debug a busy GFE.

The host itself can have 10,000,000 TCP sockets, but usually sysadmin shells run in a container with no more than 500 MB available.

Otherwise, it would be too easy for a buggy program to OOM the whole machine and have angry customers.

Quoted 12 lines

1. flush the output every time we reach a given buffer size (1M perhaps). This might make the resulting blocks slightly unaligned, with occasional loss of readability on lines occurring every 1k to 10k sockets approximately, even though after 1k sockets column sizes won't change much (it looks anyway better than the original), and I don't expect anybody to actually scroll that output

2. add a switch for unbuffered output, but then you need to remember to pass it manually, and the whole output would be as bad as the original in case you need the switch.

Quoted 8 lines
Quoted 4 lines

I'd rather go with 1., it's easy to implement (we already have partial flushing with '--events') and it looks like a good compromise on usability. Thoughts?

1 seems fine, but a switch for 'please do not try to format' would be fine.

I wonder why we try to 'format' when stdout is a pipe or a regular file .

On a second thought: what about | less, or | grep [ports], or > readable.log? I guess those might also be rather common use cases, what do you think?

Quoted 6 lines
Quoted 2 lines

I'm tempted to skip this for the moment and just go with option 1.

What I would favor:
* use big enough columns that for the common case everything lines up fine
* if column is to wide just print that element wider (which is what print %Ns does)

This is pretty much the situation Stefano attempted to improve, minus scaling the columns to max terminal width. ss output formatting being quirky and unreadable with either small or large terminals was the number one reason I heard so far why people prefer netstat.

Quoted 3 lines

and
* add json output for programs that want to parse
* use print_uint etc for that

For Eric's use-case, skipping any buffering and tabular output if stdout is not a TTY suffices. In fact, iproute2 does this already for colored output (see check_enable_color() for reference).

Adding JSON output support everywhere is a nice feature when it comes to scripting, but it won't help console users. Unless you expect CLI frontends to come turning that JSON into human-readable output.

IMHO, JSON output wouldn't even help in this case - unless Eric indeed prefers to write/use a JSON parser for his analysis instead of something along 'ss | grep'.

Quoted 1 line

The buffering patch (in iproute2-next) can/will be reverted.

It's not fair to claim that despite Stefano's commitment to fix the reported issues. His ss output rewrite is there since v4.15.0 and according to git history it needed only two fixes so far. I've had one-liners which required more follow-ups than that! Also, we're still discovering issues introduced by all the jsonify patches. Allowing for people to get things right not the first time but after a few tries is important. If you want to revert something, start with features which have a fundamental design issue in the exact situation they tried to improve, like the MSG_PEEK | MSG_TRUNC thing Hangbin and me wrote.

Thanks, Phil

#16 lore raw

Re: [PATCH iproute2 net-next v2 3/4] ss: Buffer raw fields first, then render them as a table

Quoted 8 lines · Phil Sutter
Quoted 3 lines

What I would favor:
* use big enough columns that for the common case everything lines up fine
* if column is to wide just print that element wider (which is what print %Ns does)

This is pretty much the situation Stefano attempted to improve, minus scaling the columns to max terminal width. ss output formatting being quirky and unreadable with either small or large terminals was the number one reason I heard so far why people prefer netstat.

+1.

prior to Stefano's change ss was a PITA trying to read in an xterm. I for one would run the command and then have to adjust the terminal to get it to display an actual readable format.

Quoted 9 lines
Quoted 4 lines

and
* add json output for programs that want to parse
* use print_uint etc for that

For Eric's use-case, skipping any buffering and tabular output if stdout is not a TTY suffices. In fact, iproute2 does this already for colored output (see check_enable_color() for reference).

Quoted 7 lines

Adding JSON output support everywhere is a nice feature when it comes to scripting, but it won't help console users. Unless you expect CLI frontends to come turning that JSON into human-readable output.

IMHO, JSON output wouldn't even help in this case - unless Eric indeed prefers to write/use a JSON parser for his analysis instead of something along 'ss | grep'.

I agree. json has its uses, console/xterm for humans is not one and piping into something like jq to selectively pick columns is not a user friendly solution.

Quoted 12 lines
Quoted 2 lines

The buffering patch (in iproute2-next) can/will be reverted.

It's not fair to claim that despite Stefano's commitment to fix the reported issues. His ss output rewrite is there since v4.15.0 and according to git history it needed only two fixes so far. I've had one-liners which required more follow-ups than that! Also, we're still discovering issues introduced by all the jsonify patches. Allowing for people to get things right not the first time but after a few tries is important. If you want to revert something, start with features which have a fundamental design issue in the exact situation they tried to improve, like the MSG_PEEK | MSG_TRUNC thing Hangbin and me wrote.

I was just looking at the overhead of that. While it is deceiving to twice as many recvmsg calls as you expect, the overhead of the peek in reading 700k+ routes is on the order of 3% with the 32k min buffer size. The true overhead of the dump functions for ip is the device index to name mapping (just like the overhead of a batch is the name to index mapping). I will send a v2 of my patches soon.

#17 lore raw

PATCH v2 4/4 ss: Implement automatic column width calculation

In reply to Stefano Brivio

Group fitting fields into lines and space them equally using the remaining screen width for each line. If columns don't fit on one line, break them into the least possible amount of lines and keep them aligned across lines.

This is done by:
- recording the length of the longest item in each column during formatting and buffering (which was added in the previous patch)
- fitting as many fields as possible on each line of output
- distributing the remaining padding space equally between the columns

---
v2: rebase after conflict with 00ac78d39c29 ("ss: print tcpi_rcv_ssthresh")

 misc/ss.c | 188 +++++++++++++++++++++++++++++++++++++++-----------------------
 1 file changed, 120 insertions(+), 68 deletions(-)
misc/ss.c +120 -68
index 166267974c36..9d21ed7a0705 100644
@@ -128,19 +128,21 @@ struct column {
 	const enum col_align align;
 	const char *header;
 	const char *ldelim;
-	int width;	/* Including delimiter. -1: fit to content, 0: hide */
+	int disabled;
+	int width;	/* Calculated, including additional layout spacing */
+	int max_len;	/* Measured maximum field length in this column */
 };
 
 static struct column columns[] = {
-	{ ALIGN_LEFT,	"Netid",		"",	0 },
-	{ ALIGN_LEFT,	"State",		" ",	0 },
-	{ ALIGN_LEFT,	"Recv-Q",		" ",	7 },
-	{ ALIGN_LEFT,	"Send-Q",		" ",	7 },
-	{ ALIGN_RIGHT,	"Local Address:",	" ",	0 },
-	{ ALIGN_LEFT,	"Port",			"",	0 },
-	{ ALIGN_RIGHT,	"Peer Address:",	" ",	0 },
-	{ ALIGN_LEFT,	"Port",			"",	0 },
-	{ ALIGN_LEFT,	"",			"",	-1 },
+	{ ALIGN_LEFT,	"Netid",		"",	0, 0, 0 },
+	{ ALIGN_LEFT,	"State",		" ",	0, 0, 0 },
+	{ ALIGN_LEFT,	"Recv-Q",		" ",	0, 0, 0 },
+	{ ALIGN_LEFT,	"Send-Q",		" ",	0, 0, 0 },
+	{ ALIGN_RIGHT,	"Local Address:",	" ",	0, 0, 0 },
+	{ ALIGN_LEFT,	"Port",			"",	0, 0, 0 },
+	{ ALIGN_RIGHT,	"Peer Address:",	" ",	0, 0, 0 },
+	{ ALIGN_LEFT,	"Port",			"",	0, 0, 0 },
+	{ ALIGN_LEFT,	"",			"",	0, 0, 0 },
 };
 
 static struct column *current_field = columns;
@@ -960,7 +962,7 @@ static void out(const char *fmt, ...)
 	char *pos;
 	int len;
 
-	if (!f->width)
+	if (f->disabled)
 		return;
 
 	if (!buffer.head)
@@ -983,7 +985,7 @@ static int print_left_spacing(struct column *f, int stored, int printed)
 {
 	int s;
 
-	if (f->width < 0 || f->align == ALIGN_LEFT)
+	if (!f->width || f->align == ALIGN_LEFT)
 		return 0;
 
 	s = f->width - stored - printed;
@@ -1001,7 +1003,7 @@ static void print_right_spacing(struct column *f, int printed)
 {
 	int s;
 
-	if (f->width < 0 || f->align == ALIGN_RIGHT)
+	if (!f->width || f->align == ALIGN_RIGHT)
 		return;
 
 	s = f->width - printed;
@@ -1018,9 +1020,12 @@ static void field_flush(struct column *f)
 	struct buf_chunk *chunk = buffer.tail;
 	unsigned int pad = buffer.cur->len % 2;
 
-	if (!f->width)
+	if (f->disabled)
 		return;
 
+	if (buffer.cur->len > f->max_len)
+		f->max_len = buffer.cur->len;
+
 	/* We need a new chunk if we can't store the next length descriptor.
 	 * Mind the gap between end of previous token and next aligned position
 	 * for length descriptor.
@@ -1063,7 +1068,7 @@ static void field_set(enum col_id id)
 static void print_header(void)
 {
 	while (!field_is_last(current_field)) {
-		if (current_field->width)
+		if (!current_field->disabled)
 			out(current_field->header);
 		field_next();
 	}
@@ -1096,16 +1101,106 @@ static void buf_free_all(void)
 	buffer.head = NULL;
 }
 
+/* Calculate column width from contents length. If columns don't fit on one
+ * line, break them into the least possible amount of lines and keep them
+ * aligned across lines. Available screen space is equally spread between fields
+ * as additional spacing.
+ */
+static void render_calc_width(int screen_width)
+{
+	int first, len = 0, linecols = 0;
+	struct column *c, *eol = columns - 1;
+
+	/* First pass: set width for each column to measured content length */
+	for (first = 1, c = columns; c - columns < COL_MAX; c++) {
+		if (c->disabled)
+			continue;
+
+		if (!first && c->max_len)
+			c->width = c->max_len + strlen(c->ldelim);
+		else
+			c->width = c->max_len;
+
+		/* But don't exceed screen size. If we exceed the screen size
+		 * for even a single field, it will just start on a line of its
+		 * own and then naturally wrap.
+		 */
+		c->width = min(c->width, screen_width);
+
+		if (c->width)
+			first = 0;
+	}
+
+	/* Second pass: find out newlines and distribute available spacing */
+	for (c = columns; c - columns < COL_MAX; c++) {
+		int pad, spacing, rem, last;
+		struct column *tmp;
+
+		if (!c->width)
+			continue;
+
+		linecols++;
+		len += c->width;
+
+		for (last = 1, tmp = c + 1; tmp - columns < COL_MAX; tmp++) {
+			if (tmp->width) {
+				last = 0;
+				break;
+			}
+		}
+
+		if (!last && len < screen_width) {
+			/* Columns fit on screen so far, nothing to do yet */
+			continue;
+		}
+
+		if (len == screen_width) {
+			/* Exact fit, just start with new line */
+			goto newline;
+		}
+
+		if (len > screen_width) {
+			/* Screen width exceeded: go back one column */
+			len -= c->width;
+			c--;
+			linecols--;
+		}
+
+		/* Distribute remaining space to columns on this line */
+		pad = screen_width - len;
+		spacing = pad / linecols;
+		rem = pad % linecols;
+		for (tmp = c; tmp > eol; tmp--) {
+			if (!tmp->width)
+				continue;
+
+			tmp->width += spacing;
+			if (rem) {
+				tmp->width++;
+				rem--;
+			}
+		}
+
+newline:
+		/* Line break: reset line counters, mark end-of-line */
+		eol = c;
+		len = 0;
+		linecols = 0;
+	}
+}
+
 /* Render buffered output with spacing and delimiters, then free up buffers */
-static void render(void)
+static void render(int screen_width)
 {
 	struct buf_token *token = (struct buf_token *)buffer.head->data;
-	int printed, line_started = 0, need_newline = 0;
+	int printed, line_started = 0;
 	struct column *f;
 
 	/* Ensure end alignment of last token, it wasn't necessarily flushed */
 	buffer.tail->end += buffer.cur->len % 2;
 
+	render_calc_width(screen_width);
+
 	/* Rewind and replay */
 	buffer.tail = buffer.head;
 
@@ -1125,23 +1220,16 @@ static void render(void)
 		printed += fwrite(token->data, 1, token->len, stdout);
 		print_right_spacing(f, printed);
 
-		/* Variable field size or overflow, won't align to screen */
-		if (printed > f->width)
-			need_newline = 1;
-
 		/* Go to next non-empty field, deal with end-of-line */
 		do {
 			if (field_is_last(f)) {
-				if (need_newline) {
-					printf("\n");
-					need_newline = 0;
-				}
+				printf("\n");
 				f = columns;
 				line_started = 0;
 			} else {
 				f++;
 			}
-		} while (!f->width);
+		} while (f->disabled);
 
 		token = buf_token_next(token);
 	}
@@ -4532,7 +4620,7 @@ int main(int argc, char *argv[])
 	FILE *filter_fp = NULL;
 	int ch;
 	int state_filter = 0;
-	int addrp_width, screen_width = 80;
+	int screen_width = 80;
 
 	while ((ch = getopt_long(argc, argv,
 				 "dhaletuwxnro460spbEf:miA:D:F:vVzZN:KHS",
@@ -4828,17 +4916,11 @@ int main(int argc, char *argv[])
 	if (ssfilter_parse(&current_filter.f, argc, argv, filter_fp))
 		usage();
 
-	if (current_filter.dbs&(current_filter.dbs-1))
-		columns[COL_NETID].width = 6;
-
-	if (current_filter.states&(current_filter.states-1))
-		columns[COL_STATE].width = 10;
+	if (!(current_filter.dbs & (current_filter.dbs - 1)))
+		columns[COL_NETID].disabled = 1;
 
-	/* If Netid or State are hidden, no delimiter before next column */
-	if (!columns[COL_NETID].width)
-		columns[COL_STATE].width--;
-	else if (!columns[COL_STATE].width)
-		columns[COL_RECVQ].width--;
+	if (!(current_filter.states & (current_filter.states - 1)))
+		columns[COL_STATE].disabled = 1;
 
 	if (isatty(STDOUT_FILENO)) {
 		struct winsize w;
@@ -4849,36 +4931,6 @@ int main(int argc, char *argv[])
 		}
 	}
 
-	addrp_width = screen_width -
-		      columns[COL_NETID].width -
-		      columns[COL_STATE].width -
-		      columns[COL_RECVQ].width -
-		      columns[COL_SENDQ].width;
-
-	if (addrp_width&1) {
-		if (columns[COL_NETID].width)
-			columns[COL_NETID].width++;
-		else if (columns[COL_STATE].width)
-			columns[COL_STATE].width++;
-		else
-			columns[COL_SENDQ].width++;
-	}
-
-	addrp_width /= 2;
-
-	columns[COL_SERV].width = resolve_services ? 8 : 6;
-	if (addrp_width < 15 + columns[COL_SERV].width)
-		addrp_width = 15 + columns[COL_SERV].width;
-
-	columns[COL_ADDR].width = addrp_width - columns[COL_SERV].width;
-
-	/* Make enough space for the local/remote port field */
-	columns[COL_ADDR].width -= 13;
-	columns[COL_SERV].width += 13;
-
-	columns[COL_RADDR].width = columns[COL_ADDR].width;
-	columns[COL_RSERV].width = columns[COL_SERV].width;
-
 	if (show_header)
 		print_header();
 
@@ -4909,7 +4961,7 @@ int main(int argc, char *argv[])
 	if (show_users || show_proc_ctx || show_sock_ctx)
 		user_ent_destroy();
 
-	render();
+	render(screen_width);
 
 	return 0;
 }
Signature · 1 line
2.9.4
#18 lore raw
In reply to Stefano Brivio
Quoted 103 lines · On Tue, 12 Dec 2017 01:46:29 +0100 Stefano Brivio <[email protected]> wrote:

Currently, 'ss' simply subdivides the whole available screen width between available columns, starting from a set of hardcoded amount of spacing and growing column widths.

This makes the output unreadable in several cases, as it doesn't take into account the actual content width.

Fix this by introducing a simple abstraction for columns, buffering the output, measuring the width of the fields, grouping fields into lines as they fit, equally distributing any remaining whitespace, and finally rendering the result. Some examples are reported below [1].

This implementation doesn't seem to cause any significant performance issues, as reported in 3/4.

Patch 1/4 replaces all relevant printf() calls by the out() helper, which simply consists of the usual printf() implementation.

Patch 2/4 implements column abstraction, with configurable column width and delimiters, and 3/4 splits buffering and rendering phases, employing a simple buffering mechanism with chunked allocation and introducing a rendering function.

Up to this point, the output is still unchanged.

Finally, 4/4 introduces field width calculation based on content length measured while buffering, in order to split fields onto multiple lines and equally space them within the single lines.

Now that column behaviour is well-defined and more easily configurable, it should be easier to further improve the output by splitting logically separable information (e.g. TCP details) into additional columns. However, this patchset keeps the full "extended" information into a single column, for the moment being.

v2: rebase after conflict with 00ac78d39c29 ("ss: print tcpi_rcv_ssthresh")

[1]

- 80 columns terminal, ss -Z -f netlink
* before:
Recv-Q Send-Q Local Address:Port Peer Address:Port

0 0 rtnl:evolution-calen/2075 * pr
oc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:abrt-applet/32700 * pr
oc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:firefox/21619 * pr
oc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:evolution-calen/32639 * p
roc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 [...]

* after:
Recv-Q Send-Q Local Address:Port Peer Address:Port
0 0 rtnl:evolution-calen/2075 *
proc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:abrt-applet/32700 *
proc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:firefox/21619 *
proc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
0 0 rtnl:evolution-calen/32639 *
proc_ctx=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
[...]

- 80 columns terminal, ss -tunpl
* before:

Netid  State      Recv-Q Send-Q Local Address:Port               Peer Address:Port
udp    UNCONN     0      0         *:37732                 *:*
udp    UNCONN     0      0         *:5353                  *:*
udp    UNCONN     0      0      192.168.122.1:53                    *:*
udp    UNCONN     0      0      *%virbr0:67                    *:*

[...]

* after:

Netid   State    Recv-Q   Send-Q     Local Address:Port      Peer Address:Port
udp     UNCONN   0        0                      *:37732                *:*
udp     UNCONN   0        0                      *:5353                 *:*
udp     UNCONN   0        0          192.168.122.1:53                   *:*
udp     UNCONN   0        0               *%virbr0:67                   *:*

[...]

- 66 columns terminal, ss -tunpl
* before:
Netid State Recv-Q Send-Q Local Address:Port P
eer Address:Port
udp UNCONN 0 0 *:37732 *:*

udp UNCONN 0 0 *:5353 *:*

udp UNCONN 0 0 192.168.122.1:53
*:*
udp UNCONN 0 0 *%virbr0:67 *:*
[...]

* after:
Netid State Recv-Q Send-Q Local Address:Port Peer Address:Port

udp   UNCONN 0      0                  *:37732             *:*
udp   UNCONN 0      0                  *:5353              *:*
udp   UNCONN 0      0      192.168.122.1:53                *:*
udp   UNCONN 0      0           *%virbr0:67                *:*
Quoted 8 lines

[...]

Stefano Brivio (4):
ss: Replace printf() calls for "main" output by calls to helper
ss: Introduce columns lightweight abstraction
ss: Buffer raw fields first, then render them as a table
ss: Implement automatic column width calculation

This looks so good, I went ahead and applied to current release (not net-next) Thanks Stefano