Skip to contentSkip to conversation

add json support on tc u32

PATCH Superseded Wen Liang 5 participants 14 messages Jan 6, 2022 – Jan 31, 2022

Subsystems (2)

v3 of 5 Superseded Latest is v5 →

lore
Share & tools
lore

Wen Liang

author
#0 lore raw

v3 cover add json support on tc u32

This adds support for json output on tc u32.
- The first patch is replacing with proper json functions in u32_print_opt()
- The second patch is fixing the json support in u32 print_raw(), print_ipv4() and print_ipv6()

Wen Liang (2):
tc: u32: add support for json output
tc: u32: add json support in print_raw, print_ipv4, print_ipv6

 tc/f_u32.c | 173 +++++++++++++++++++++++++++++------------------------
 1 file changed, 96 insertions(+), 77 deletions(-)
Signature · 1 line
2.26.3
author
#1 lore raw

PATCH v3 1/2 tc: u32: add support for json output

Currently u32 filter output does not support json. This commit uses proper json functions to add support for it.

---
 tc/f_u32.c | 72 ++++++++++++++++++++++++++++--------------------------
 1 file changed, 37 insertions(+), 35 deletions(-)
tc/f_u32.c +37 -35
index a5747f67..be16ba1e 100644
@@ -1213,11 +1213,11 @@ static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt,
 
 	if (handle) {
 		SPRINT_BUF(b1);
-		fprintf(f, "fh %s ", sprint_u32_handle(handle, b1));
+		print_string(PRINT_ANY, "fh", "fh %s ", sprint_u32_handle(handle, b1));
 	}
 
 	if (TC_U32_NODE(handle))
-		fprintf(f, "order %d ", TC_U32_NODE(handle));
+		print_int(PRINT_ANY, "order", "order %d ", TC_U32_NODE(handle));
 
 	if (tb[TCA_U32_SEL]) {
 		if (RTA_PAYLOAD(tb[TCA_U32_SEL])  < sizeof(*sel))
@@ -1227,15 +1227,13 @@ static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt,
 	}
 
 	if (tb[TCA_U32_DIVISOR]) {
-		fprintf(f, "ht divisor %d ",
-			rta_getattr_u32(tb[TCA_U32_DIVISOR]));
+		print_int(PRINT_ANY, "ht_divisor", "ht divisor %d ", rta_getattr_u32(tb[TCA_U32_DIVISOR]));
 	} else if (tb[TCA_U32_HASH]) {
 		__u32 htid = rta_getattr_u32(tb[TCA_U32_HASH]);
-
-		fprintf(f, "key ht %x bkt %x ", TC_U32_USERHTID(htid),
-			TC_U32_HASH(htid));
+		print_hex(PRINT_ANY, "key_ht", "key ht %x ", TC_U32_USERHTID(htid));
+		print_hex(PRINT_ANY, "bkt", "bkt %x ", TC_U32_HASH(htid));
 	} else {
-		fprintf(f, "??? ");
+		fprintf(stderr, "divisor and hash missing ");
 	}
 	if (tb[TCA_U32_CLASSID]) {
 		SPRINT_BUF(b1);
@@ -1244,12 +1242,11 @@ static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt,
 			sprint_tc_classid(rta_getattr_u32(tb[TCA_U32_CLASSID]),
 					  b1));
 	} else if (sel && sel->flags & TC_U32_TERMINAL) {
-		fprintf(f, "terminal flowid ??? ");
+		print_bool(PRINT_ANY, "terminal_flowid", "terminal flowid ??? ", true);
 	}
 	if (tb[TCA_U32_LINK]) {
 		SPRINT_BUF(b1);
-		fprintf(f, "link %s ",
-			sprint_u32_handle(rta_getattr_u32(tb[TCA_U32_LINK]),
+		print_string(PRINT_ANY, "link", "link %s ", sprint_u32_handle(rta_getattr_u32(tb[TCA_U32_LINK]),
 					  b1));
 	}
 
@@ -1257,14 +1254,14 @@ static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt,
 		__u32 flags = rta_getattr_u32(tb[TCA_U32_FLAGS]);
 
 		if (flags & TCA_CLS_FLAGS_SKIP_HW)
-			fprintf(f, "skip_hw ");
+			print_bool(PRINT_ANY, "skip_hw", "skip_hw ", true);
 		if (flags & TCA_CLS_FLAGS_SKIP_SW)
-			fprintf(f, "skip_sw ");
+			print_bool(PRINT_ANY, "skip_sw", "skip_sw ", true);
 
 		if (flags & TCA_CLS_FLAGS_IN_HW)
-			fprintf(f, "in_hw ");
+			print_bool(PRINT_ANY, "in_hw", "in_hw ", true);
 		else if (flags & TCA_CLS_FLAGS_NOT_IN_HW)
-			fprintf(f, "not_in_hw ");
+			print_bool(PRINT_ANY, "not_in_hw", "not_in_hw ", true);
 	}
 
 	if (tb[TCA_U32_PCNT]) {
@@ -1275,10 +1272,10 @@ static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt,
 		pf = RTA_DATA(tb[TCA_U32_PCNT]);
 	}
 
-	if (sel && show_stats && NULL != pf)
-		fprintf(f, " (rule hit %llu success %llu)",
-			(unsigned long long) pf->rcnt,
-			(unsigned long long) pf->rhit);
+	if (sel && show_stats && NULL != pf) {
+		print_u64(PRINT_ANY, "rule_hit", "(rule hit %llu ", pf->rcnt);
+		print_u64(PRINT_ANY, "success", "success %llu)", pf->rhit);
+	}
 
 	if (tb[TCA_U32_MARK]) {
 		struct tc_u32_mark *mark = RTA_DATA(tb[TCA_U32_MARK]);
@@ -1286,8 +1283,10 @@ static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt,
 		if (RTA_PAYLOAD(tb[TCA_U32_MARK]) < sizeof(*mark)) {
 			fprintf(f, "\n  Invalid mark (kernel&iproute2 mismatch)\n");
 		} else {
-			fprintf(f, "\n  mark 0x%04x 0x%04x (success %d)",
-				mark->val, mark->mask, mark->success);
+			print_nl();
+			print_0xhex(PRINT_ANY, "fwmark_value", "  mark 0x%04x ", mark->val);
+			print_0xhex(PRINT_ANY, "fwmark_mask", "0x%04x ", mark->mask);
+			print_int(PRINT_ANY, "fwmark_success", "(success %d)", mark->success);
 		}
 	}
 
@@ -1298,38 +1297,41 @@ static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt,
 			for (i = 0; i < sel->nkeys; i++) {
 				show_keys(f, sel->keys + i);
 				if (show_stats && NULL != pf)
-					fprintf(f, " (success %llu ) ",
-						(unsigned long long) pf->kcnts[i]);
+					print_u64(PRINT_ANY, "success", " (success %llu ) ", pf->kcnts[i]);
 			}
 		}
 
 		if (sel->flags & (TC_U32_VAROFFSET | TC_U32_OFFSET)) {
-			fprintf(f, "\n    offset ");
-			if (sel->flags & TC_U32_VAROFFSET)
-				fprintf(f, "%04x>>%d at %d ",
-					ntohs(sel->offmask),
-					sel->offshift,  sel->offoff);
+			print_nl();
+			print_string(PRINT_ANY, NULL, "%s", "    offset ");
+			if (sel->flags & TC_U32_VAROFFSET) {
+				print_hex(PRINT_ANY, "offset_mask", "%04x", ntohs(sel->offmask));
+				print_int(PRINT_ANY, "offset_shift", ">>%d ", sel->offshift);
+				print_int(PRINT_ANY, "offset_off", "at %d ", sel->offoff);
+			}
 			if (sel->off)
-				fprintf(f, "plus %d ", sel->off);
+				print_int(PRINT_ANY, "plus", "plus %d ", sel->off);
 		}
 		if (sel->flags & TC_U32_EAT)
-			fprintf(f, " eat ");
+			print_string(PRINT_ANY, NULL, "%s", " eat ");
 
 		if (sel->hmask) {
-			fprintf(f, "\n    hash mask %08x at %d ",
-				(unsigned int)htonl(sel->hmask), sel->hoff);
+			print_nl();
+			print_hex(PRINT_ANY, "hash_mask", "    hash mask %08x ", (unsigned int)htonl(sel->hmask));
+			print_int(PRINT_ANY, "hash_off", "at %d ", sel->hoff);
 		}
 	}
 
 	if (tb[TCA_U32_POLICE]) {
-		fprintf(f, "\n");
+		print_nl();
 		tc_print_police(f, tb[TCA_U32_POLICE]);
 	}
 
 	if (tb[TCA_U32_INDEV]) {
 		struct rtattr *idev = tb[TCA_U32_INDEV];
-
-		fprintf(f, "\n  input dev %s\n", rta_getattr_str(idev));
+		print_nl();
+		print_string(PRINT_ANY, "input_dev", "  input dev %s", rta_getattr_str(idev));
+		print_nl();
 	}
 
 	if (tb[TCA_U32_ACT])
Signature · 1 line
2.26.3
author
#13 lore raw

PATCH v3 2/2 tc: u32: add json support in `print_raw`, `print_ipv4`, `print_ipv6`

In reply to Wen Liang
---
 tc/f_u32.c | 101 +++++++++++++++++++++++++++++++----------------------
 1 file changed, 59 insertions(+), 42 deletions(-)
tc/f_u32.c +59 -42
index be16ba1e..1dd9bfcc 100644
@@ -824,23 +824,24 @@ static void print_ipv4(FILE *f, const struct tc_u32_key *key)
 {
 	char abuf[256];
 
+	open_json_object("match");
 	switch (key->off) {
 	case 0:
 		switch (ntohl(key->mask)) {
 		case 0x0f000000:
-			fprintf(f, "\n  match IP ihl %u",
-				ntohl(key->val) >> 24);
+			print_nl();
+			print_uint(PRINT_ANY, "ip_ihl", "  match IP ihl %u", ntohl(key->val) >> 24);
 			return;
 		case 0x00ff0000:
-			fprintf(f, "\n  match IP dsfield %#x",
-				ntohl(key->val) >> 16);
+			print_nl();
+			print_0xhex(PRINT_ANY, "ip_dsfield", "  match IP dsfield %#x", ntohl(key->val) >> 16);
 			return;
 		}
 		break;
 	case 8:
 		if (ntohl(key->mask) == 0x00ff0000) {
-			fprintf(f, "\n  match IP protocol %d",
-				ntohl(key->val) >> 16);
+			print_nl();
+			print_int(PRINT_ANY, "ip_protocol", "  match IP protocol %d", ntohl(key->val) >> 16);
 			return;
 		}
 		break;
@@ -849,11 +850,18 @@ static void print_ipv4(FILE *f, const struct tc_u32_key *key)
 			int bits = mask2bits(key->mask);
 
 			if (bits >= 0) {
-				fprintf(f, "\n  %s %s/%d",
-					key->off == 12 ? "match IP src" : "match IP dst",
-					inet_ntop(AF_INET, &key->val,
-						  abuf, sizeof(abuf)),
-					bits);
+				 if (key->off == 12) {
+				       print_nl();
+				       print_null(PRINT_FP, NULL, "  match IP src ", NULL);
+				       open_json_object("src");
+				 } else {
+				       print_nl();
+				       print_null(PRINT_FP, NULL, "  match IP dst ", NULL);
+				       open_json_object("dst");
+				 }
+				 print_string(PRINT_ANY, "address", "%s", inet_ntop(AF_INET, &key->val, abuf, sizeof(abuf)));
+				 print_int(PRINT_ANY, "prefixlen", "/%d", bits);
+				 close_json_object();
 				return;
 			}
 		}
@@ -862,45 +870,45 @@ static void print_ipv4(FILE *f, const struct tc_u32_key *key)
 	case 20:
 		switch (ntohl(key->mask)) {
 		case 0x0000ffff:
-			fprintf(f, "\n  match dport %u",
-				ntohl(key->val) & 0xffff);
+			print_uint(PRINT_ANY, "dport", "match dport %u", ntohl(key->val) & 0xffff);
 			return;
 		case 0xffff0000:
-			fprintf(f, "\n  match sport %u",
-				ntohl(key->val) >> 16);
+			print_nl();
+			print_uint(PRINT_ANY, "sport", "  match sport %u", ntohl(key->val) >> 16);
 			return;
 		case 0xffffffff:
-			fprintf(f, "\n  match dport %u, match sport %u",
-				ntohl(key->val) & 0xffff,
-				ntohl(key->val) >> 16);
-
+			print_nl();
+			print_uint(PRINT_ANY, "dport", "  match dport %u, ", ntohl(key->val) & 0xffff);
+			print_uint(PRINT_ANY, "sport", "match sport %u", ntohl(key->val) >> 16);
 			return;
 		}
 		/* XXX: Default print_raw */
 	}
+	close_json_object();
 }
 
 static void print_ipv6(FILE *f, const struct tc_u32_key *key)
 {
 	char abuf[256];
 
+	open_json_object("match");
 	switch (key->off) {
 	case 0:
 		switch (ntohl(key->mask)) {
 		case 0x0f000000:
-			fprintf(f, "\n  match IP ihl %u",
-				ntohl(key->val) >> 24);
+			print_nl();
+			print_uint(PRINT_ANY, "ip_ihl", "  match IP ihl %u", ntohl(key->val) >> 24);
 			return;
 		case 0x00ff0000:
-			fprintf(f, "\n  match IP dsfield %#x",
-				ntohl(key->val) >> 16);
+			print_nl();
+			print_0xhex(PRINT_ANY, "ip_dsfield", "  match IP dsfield %#x", ntohl(key->val) >> 16);
 			return;
 		}
 		break;
 	case 8:
 		if (ntohl(key->mask) == 0x00ff0000) {
-			fprintf(f, "\n  match IP protocol %d",
-				ntohl(key->val) >> 16);
+			print_nl();
+			print_int(PRINT_ANY, "ip_protocol", "  match IP protocol %d", ntohl(key->val) >> 16);
 			return;
 		}
 		break;
@@ -909,11 +917,18 @@ static void print_ipv6(FILE *f, const struct tc_u32_key *key)
 			int bits = mask2bits(key->mask);
 
 			if (bits >= 0) {
-				fprintf(f, "\n  %s %s/%d",
-					key->off == 12 ? "match IP src" : "match IP dst",
-					inet_ntop(AF_INET, &key->val,
-						  abuf, sizeof(abuf)),
-					bits);
+				if (key->off == 12) {
+			              print_nl();
+				      print_null(PRINT_FP, NULL, "  match IP src ", NULL);
+				      open_json_object("src");
+				} else {
+			              print_nl();
+				      print_null(PRINT_FP, NULL, "  match IP dst ", NULL);
+				      open_json_object("dst");
+				}
+				print_string(PRINT_ANY, "address", "%s", inet_ntop(AF_INET, &key->val, abuf, sizeof(abuf)));
+				print_int(PRINT_ANY, "prefixlen", "/%d", bits);
+				close_json_object();
 				return;
 			}
 		}
@@ -922,31 +937,33 @@ static void print_ipv6(FILE *f, const struct tc_u32_key *key)
 	case 20:
 		switch (ntohl(key->mask)) {
 		case 0x0000ffff:
-			fprintf(f, "\n  match sport %u",
-				ntohl(key->val) & 0xffff);
+			print_nl();
+			print_uint(PRINT_ANY, "sport", "  match sport %u", ntohl(key->val) & 0xffff);
 			return;
 		case 0xffff0000:
-			fprintf(f, "\n  match dport %u",
-				ntohl(key->val) >> 16);
+			print_uint(PRINT_ANY, "dport", "match dport %u", ntohl(key->val) >> 16);
 			return;
 		case 0xffffffff:
-			fprintf(f, "\n  match sport %u, match dport %u",
-				ntohl(key->val) & 0xffff,
-				ntohl(key->val) >> 16);
+			print_nl();
+			print_uint(PRINT_ANY, "sport", "  match sport %u, ", ntohl(key->val) & 0xffff);
+			print_uint(PRINT_ANY, "dport", "match dport %u", ntohl(key->val) >> 16);
 
 			return;
 		}
 		/* XXX: Default print_raw */
 	}
+	close_json_object();
 }
 
 static void print_raw(FILE *f, const struct tc_u32_key *key)
 {
-	fprintf(f, "\n  match %08x/%08x at %s%d",
-		(unsigned int)ntohl(key->val),
-		(unsigned int)ntohl(key->mask),
-		key->offmask ? "nexthdr+" : "",
-		key->off);
+	open_json_object("match");
+	print_nl();
+	print_hex(PRINT_ANY, "value", "  match %08x", (unsigned int)ntohl(key->val));
+	print_hex(PRINT_ANY, "mask", "/%08x ", (unsigned int)ntohl(key->mask));
+	print_string(PRINT_ANY, "offmask", "at %s", key->offmask ? "nexthdr+" : "");
+	print_int(PRINT_ANY, "off", "%d", key->off);
+	close_json_object();
 }
 
 static const struct {
Signature · 1 line
2.26.3
#2 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

In reply to Wen Liang

tc/f_u32.c static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt, · line 1245

Quoted 3 lines · On Thu, 6 Jan 2022 13:45:51 -0500 Wen Liang <[email protected]> wrote:
tc/f_u32.c +1 -1
 	} else if (sel && sel->flags & TC_U32_TERMINAL) {
-		fprintf(f, "terminal flowid ??? ");
+		print_bool(PRINT_ANY, "terminal_flowid", "terminal flowid ??? ", true);

This looks like another error (ie to stderr) like the earlier case

tc/f_u32.c static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt, · line 1250

Quoted 6 lines
tc/f_u32.c +1 -2
 	if (tb[TCA_U32_LINK]) {
 		SPRINT_BUF(b1);
-		fprintf(f, "link %s ",
-			sprint_u32_handle(rta_getattr_u32(tb[TCA_U32_LINK]),
+		print_string(PRINT_ANY, "link", "link %s ", sprint_u32_handle(rta_getattr_u32(tb[TCA_U32_LINK]),
 					  b1));

Break that long line up. Would look better with a temporary variable.

FYI - good test is to run the json output into python's json parser to make sure it is valid.

#3 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

Quoted 8 lines · Stephen Hemminger

tc/f_u32.c static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt, · line 1250

Quoted 6 lines
tc/f_u32.c +1 -2
 	if (tb[TCA_U32_LINK]) {
 		SPRINT_BUF(b1);
-		fprintf(f, "link %s ",
-			sprint_u32_handle(rta_getattr_u32(tb[TCA_U32_LINK]),
+		print_string(PRINT_ANY, "link", "link %s ", sprint_u32_handle(rta_getattr_u32(tb[TCA_U32_LINK]),
 					  b1));

Break that long line up. Would look better with a temporary variable.

+1 and that comment applies to all lines in both patches.

#4 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

In reply to Stephen Hemminger
Quoted 9 lines · Stephen Hemminger

tc/f_u32.c static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt, · line 1245

Quoted 3 lines · On Thu, 6 Jan 2022 13:45:51 -0500 Wen Liang <[email protected]> wrote:
tc/f_u32.c +1 -1
 	} else if (sel && sel->flags & TC_U32_TERMINAL) {
-		fprintf(f, "terminal flowid ??? ");
+		print_bool(PRINT_ANY, "terminal_flowid", "terminal flowid ??? ", true);

This looks like another error (ie to stderr) like the earlier case

Hi Stephen,
Sorry for coming to this so late, but this doesn't look like an error to me.

As far as I can see, TC_U32_TERMINAL is set in this file together with CLASSID or when "action" or "policy" are used. The latter case should be the one that this else branch should catch.

Now, "terminal flowid ???" looks to me like a message printed when we don't actually have a flowid to show, and indeed that is specified when this flag is set (see the comment at line 1169). As such this is probably more a useless log message, than an error one.

If this is the case, we can probably maintain this message on the PRINT_FP output (only to not break script parsing this bit of info out there), and disregard this bit of info on the JSON output.

What do you think?

Regards,
Andrea

#5 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

Quoted 23 lines · On Mon, 24 Jan 2022 19:25:06 +0100 Andrea Claudi <[email protected]> wrote:
Quoted 9 lines · Stephen Hemminger

tc/f_u32.c static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt, · line 1245

Quoted 3 lines · On Thu, 6 Jan 2022 13:45:51 -0500 Wen Liang <[email protected]> wrote:
tc/f_u32.c +1 -1
 	} else if (sel && sel->flags & TC_U32_TERMINAL) {
-		fprintf(f, "terminal flowid ??? ");
+		print_bool(PRINT_ANY, "terminal_flowid", "terminal flowid ??? ", true);  

This looks like another error (ie to stderr) like the earlier case

Hi Stephen,
Sorry for coming to this so late, but this doesn't look like an error to me.

As far as I can see, TC_U32_TERMINAL is set in this file together with CLASSID or when "action" or "policy" are used. The latter case should be the one that this else branch should catch.

Now, "terminal flowid ???" looks to me like a message printed when we don't actually have a flowid to show, and indeed that is specified when this flag is set (see the comment at line 1169). As such this is probably more a useless log message, than an error one.

Quoted 9 lines

If this is the case, we can probably maintain this message on the PRINT_FP output (only to not break script parsing this bit of info out there), and disregard this bit of info on the JSON output.

What do you think?

Regards,
Andrea

Just always put the same original message on stderr.

#6 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

Quoted 30 lines · Stephen Hemminger
Quoted 27 lines · On Mon, 24 Jan 2022 19:25:06 +0100 Andrea Claudi <[email protected]> wrote:
Quoted 9 lines · Stephen Hemminger

tc/f_u32.c static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt, · line 1245

Quoted 3 lines · On Thu, 6 Jan 2022 13:45:51 -0500 Wen Liang <[email protected]> wrote:
tc/f_u32.c +1 -1
 	} else if (sel && sel->flags & TC_U32_TERMINAL) {
-		fprintf(f, "terminal flowid ??? ");
+		print_bool(PRINT_ANY, "terminal_flowid", "terminal flowid ??? ", true);  

This looks like another error (ie to stderr) like the earlier case

Hi Stephen,
Sorry for coming to this so late, but this doesn't look like an error to me.

As far as I can see, TC_U32_TERMINAL is set in this file together with CLASSID or when "action" or "policy" are used. The latter case should be the one that this else branch should catch.

Now, "terminal flowid ???" looks to me like a message printed when we don't actually have a flowid to show, and indeed that is specified when this flag is set (see the comment at line 1169). As such this is probably more a useless log message, than an error one.

If this is the case, we can probably maintain this message on the PRINT_FP output (only to not break script parsing this bit of info out there), and disregard this bit of info on the JSON output.

Quoted 8 lines
Quoted 5 lines

What do you think?

Regards,
Andrea

Just always put the same original message on stderr.

Let me phrase my case better: I think the "terminal flowid" message should not be on stderr, as I don't think this is an error message.

Indeed, "terminal flowid ???" is printed every time we use "action" or "policy" (see my previous email for details), even when no error is present and cls_u32 is working ok. In these cases, not having a flowid is legitimate and not an error.

As this is the case, I think the proper course of action is to have this message printed out only in non-json output to preserve the same output of older iproute versions. It would be even better if we decide to remove this message altogether, as it is not adding any valuable info to the user.

#7 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

Quoted 48 lines · On Mon, 24 Jan 2022 22:30:21 +0100 Andrea Claudi <[email protected]> wrote:
Quoted 30 lines · Stephen Hemminger
Quoted 27 lines · On Mon, 24 Jan 2022 19:25:06 +0100 Andrea Claudi <[email protected]> wrote:
Quoted 9 lines · Stephen Hemminger

tc/f_u32.c static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt, · line 1245

Quoted 3 lines · On Thu, 6 Jan 2022 13:45:51 -0500 Wen Liang <[email protected]> wrote:
tc/f_u32.c +1 -1
 	} else if (sel && sel->flags & TC_U32_TERMINAL) {
-		fprintf(f, "terminal flowid ??? ");
+		print_bool(PRINT_ANY, "terminal_flowid", "terminal flowid ??? ", true);    

This looks like another error (ie to stderr) like the earlier case

Hi Stephen,
Sorry for coming to this so late, but this doesn't look like an error to me.

As far as I can see, TC_U32_TERMINAL is set in this file together with CLASSID or when "action" or "policy" are used. The latter case should be the one that this else branch should catch.

Now, "terminal flowid ???" looks to me like a message printed when we don't actually have a flowid to show, and indeed that is specified when this flag is set (see the comment at line 1169). As such this is probably more a useless log message, than an error one.

If this is the case, we can probably maintain this message on the PRINT_FP output (only to not break script parsing this bit of info out there), and disregard this bit of info on the JSON output.

Quoted 8 lines
Quoted 5 lines

What do you think?

Regards,
Andrea

Just always put the same original message on stderr.

Let me phrase my case better: I think the "terminal flowid" message should not be on stderr, as I don't think this is an error message.

Indeed, "terminal flowid ???" is printed every time we use "action" or "policy" (see my previous email for details), even when no error is present and cls_u32 is working ok. In these cases, not having a flowid is legitimate and not an error.

Quoted 5 lines

As this is the case, I think the proper course of action is to have this message printed out only in non-json output to preserve the same output of older iproute versions. It would be even better if we decide to remove this message altogether, as it is not adding any valuable info to the user.

Agree, I have never used this obscure corner of u32 so will defer to others. But the existing message is unhelpful and looks like a bug.
The output should be clear and correct for both json and non-json cases; and any ??? kind of output should be reserved for cases where some bogus result is being returned by the kernel. Some version skew, or partial result of previous operation maybe.

#8 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

Quoted 31 lines · Stephen Hemminger
Quoted 28 lines · On Mon, 24 Jan 2022 22:30:21 +0100 Andrea Claudi <[email protected]> wrote:
Quoted 16 lines · Stephen Hemminger
Quoted 11 lines · On Mon, 24 Jan 2022 19:25:06 +0100 Andrea Claudi <[email protected]> wrote:
Quoted 9 lines · Stephen Hemminger

tc/f_u32.c static int u32_print_opt(struct filter_util *qu, FILE *f, struct rtattr *opt, · line 1245

Quoted 3 lines · On Thu, 6 Jan 2022 13:45:51 -0500 Wen Liang <[email protected]> wrote:
  	} else if (sel && sel->flags & TC_U32_TERMINAL) {
tc/f_u32.c +1 -1
-		fprintf(f, "terminal flowid ??? ");
+		print_bool(PRINT_ANY, "terminal_flowid", "terminal flowid ??? ", true);

This looks like another error (ie to stderr) like the earlier case

Quoted 2 lines

Just always put the same original message on stderr.

Let me phrase my case better: I think the "terminal flowid" message should not be on stderr, as I don't think this is an error message.

Indeed, "terminal flowid ???" is printed every time we use "action" or "policy" (see my previous email for details), even when no error is present and cls_u32 is working ok. In these cases, not having a flowid is legitimate and not an error.

Quoted 7 lines
Quoted 5 lines

As this is the case, I think the proper course of action is to have this message printed out only in non-json output to preserve the same output of older iproute versions. It would be even better if we decide to remove this message altogether, as it is not adding any valuable info to the user.

Agree, I have never used this obscure corner of u32 so will defer to others.

Andrea is correct: it is not an error and not deserving to be in stderr. And it is _not_ an obscure case just for u32. The classid/flowid is a classifier concept - in most cases is used to select a queue upon a filter match but could also be used to uniquely identify a flow. Consider it a mini-action which identifies the flow. It is omni-present. In this case u32 is only reporting it wasnt set. It doesnt affect the datapath functionality. So u32 is doing the right thing. Flower is a big culprit in that when you visit the googles hardly any example shows this field being set and the iproute2 side of flower code just ignores it.

Quoted 5 lines

But the existing message is unhelpful and looks like a bug. The output should be clear and correct for both json and non-json cases; and any ??? kind of output should be reserved for cases where some bogus result is being returned by the kernel. Some version skew, or partial result of previous operation maybe.

Makes sense in particular if we have formal output format like json. If this breaks tdc it would be worth to fix tdc (and not be backward compatible)

So: Since none of the tc maintainers was Cced in this thread, can we please impose a rule where any changes to tc subdir needs to have tdc tests run (and hopefully whoever is making the change will be gracious to contribute an additional testcase)?
Do you need a patch for that in some documentation?

cheers,
jamal

#9 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

Quoted 9 lines · Jamal Hadi Salim

Makes sense in particular if we have formal output format like json. If this breaks tdc it would be worth to fix tdc (and not be backward compatible)

So: Since none of the tc maintainers was Cced in this thread, can we please impose a rule where any changes to tc subdir needs to have tdc tests run (and hopefully whoever is making the change will be gracious to contribute an additional testcase)?

I can try to remember to run tdc tests for tc patches. I looked into it a few days ago and seems straightforward to run tdc.sh. The output of those tests could be simplified - when all is good you get the one line summary of the test name with PASS/FAIL with an option to run in verbose mode to get the details of failures. As it is, the person running the tests has to wade through a lot of output.

Quoted 2 lines

Do you need a patch for that in some documentation?

How about adding some comments to README.devel?

#10 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

Quoted 6 lines · David Ahern
Quoted 5 lines · Jamal Hadi Salim

Makes sense in particular if we have formal output format like json. If this breaks tdc it would be worth to fix tdc (and not be backward compatible)

Quoted 7 lines
Quoted 4 lines

So: Since none of the tc maintainers was Cced in this thread, can we please impose a rule where any changes to tc subdir needs to have tdc tests run (and hopefully whoever is making the change will be gracious to contribute an additional testcase)?

I can try to remember to run tdc tests for tc patches. I looked into it a few days ago and seems straightforward to run tdc.sh.

Note tdc.sh is meant for the bot. It skips a lot things per Davide's comment that he was worried the robot will end up spending too many cycles. Good source at the moment is the README.

Quoted 6 lines

The output of
those tests could be simplified - when all is good you get the one line summary of the test name with PASS/FAIL with an option to run in verbose mode to get the details of failures. As it is, the person running the tests has to wade through a lot of output.

We are going to put some cycles improving things. Your input is useful.

Quoted 4 lines
Quoted 2 lines

Do you need a patch for that in some documentation?

How about adding some comments to README.devel?

Sure - but it wont be sufficient IMO.
Best course of action is for the maintainers to remind people to run tests.

BTW: We found out that Stephen's patches still break the latest -next.

cheers,
jamal

#11 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

Quoted 9 lines · Jamal Hadi Salim
Quoted 4 lines
Quoted 2 lines

Do you need a patch for that in some documentation?

How about adding some comments to README.devel?

Sure - but it wont be sufficient IMO.
Best course of action is for the maintainers to remind people to run tests.

Above, you said the tests were meant for bots.

Quoted 3 lines

BTW: We found out that Stephen's patches still break the latest -next.

ugh. I committed them after running tdc.sh and not seeing a change in output. We'll need fixup patches then.

Clearly some work is needed on getting the test suite usable by a wider audience. I am new to running those tests as well and probably had some pilot errors running them.

I do wait for ACKs from tc folks, but can't wait forever. Right now there is the 'skip_hw' and 'skip_sw' patch that Victor sent a Tested-by. When I apply that v2 patch, I see errors in the tdc.sh output so my mileage varies from Victor's. There is also the v5 of this set which I have no applied yet; it could use some acks and tdc testing as well.

#12 lore raw

Re: [PATCH iproute2 v3 1/2] tc: u32: add support for json output

Quoted 4 lines · David Ahern
Quoted 3 lines · Jamal Hadi Salim
Quoted 3 lines
Quoted 2 lines

Do you need a patch for that in some documentation?

Quoted 9 lines
Quoted 6 lines
Quoted 1 line

How about adding some comments to README.devel?

Sure - but it wont be sufficient IMO.
Best course of action is for the maintainers to remind people to run tests.

Above, you said the tests were meant for bots.

The invocation that tdc.sh makes is targeted for the bots. It only tests actions and qdiscs.
If you run ./tdc.py -h you'll see more options.

Quoted 11 lines
Quoted 3 lines

BTW: We found out that Stephen's patches still break the latest -next.

ugh. I committed them after running tdc.sh and not seeing a change in output. We'll need fixup patches then.

Clearly some work is needed on getting the test suite usable by a wider audience. I am new to running those tests as well and probably had some pilot errors running them.

One of the issues is sometimes some of the tests take a while to complete, so some tweaking of the timer may be needed depending on your setup etc. Probably that is what you may be running into.

Quoted 5 lines

I do wait for ACKs from tc folks, but can't wait forever. Right now there is the 'skip_hw' and 'skip_sw' patch that Victor sent a Tested-by. When I apply that v2 patch, I see errors in the tdc.sh output so my mileage varies from Victor's. There is also the v5 of this set which I have no applied yet; it could use some acks and tdc testing as well.

ok. Punting to Victor for testing the v5 - was it posted on the list? I will Ack Bowen's patch.

cheers,
jamal