Skip to contentSkip to conversation

examples/ipsec-secgw: set default

PATCH Superseded Bernard Iremonger 6 participants 34 messages Aug 29, 2019 – Nov 18, 2019

lore
Share & tools
lore
#0 lore raw

cover examples/ipsec-secgw: set default

This patch set, sets the default code path in the ipsec-secgw application to use the librte_ipsec.
The _old test scripts have been modified to use the legacy code patch

Bernard Iremonger (2):
examples/ipsec-secgw: set default to IPsec library mode examples/ipsec-secgw: add -l 0 parameter to old scripts

 doc/guides/rel_notes/release_19_11.rst             |  8 ++++
 doc/guides/sample_app_ug/ipsec_secgw.rst           |  6 ++-
 examples/ipsec-secgw/ipsec-secgw.c                 | 44 ++++++++++++++--------
 .../ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/trs_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/tun_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh   |  2 +-
 13 files changed, 50 insertions(+), 28 deletions(-)
Signature · 1 line
2.7.4
#1 lore raw

PATCH 1/2 examples/ipsec-secgw: set default to IPsec library mode

Set the default code path to librte_ipsec mode.
Add parameter 0 | 1 to -l option
Check for conflicting options, -w -a -e and reassembly options are not supported in legacy mode.
Show fragment table size.
Update print_usage.
Update the ipsec-secgw guide.
Update the release notes.

---
 doc/guides/rel_notes/release_19_11.rst   |  8 ++++++
 doc/guides/sample_app_ug/ipsec_secgw.rst |  6 +++--
 examples/ipsec-secgw/ipsec-secgw.c       | 44 ++++++++++++++++++++------------
 3 files changed, 40 insertions(+), 18 deletions(-)
doc/guides/rel_notes/release_19_11.rst +8 -0
index 8490d89..70143c5 100644
@@ -56,6 +56,14 @@ New Features
      Also, make sure to start the actual text at the margin.
      =========================================================
 
+* **Updated the IPsec Security Gateway application.**
+
+  The ``librte_ipsec`` code path is now the default code path in
+  ``ipsec-secgw``
+
+  * The ``-l`` command line option has been extended to take a 0 | 1 argument.
+    0 enables the code path that uses legacy code.
+    1 enables the code path that uses ``librte_ipsec``.
 
 Removed Items
 -------------
doc/guides/sample_app_ug/ipsec_secgw.rst +4 -2
index ad2d79e..17b00c0 100644
@@ -92,7 +92,7 @@ The application has a number of command line options::
 
    ./build/ipsec-secgw [EAL options] --
                         -p PORTMASK -P -u PORTMASK -j FRAMESIZE
-                        -l -w REPLAY_WINOW_SIZE -e -a
+                        -l 0 -w REPLAY_WINOW_SIZE -e -a
                         --config (port,queue,lcore)[,(port,queue,lcore]
                         --single-sa SAIDX
                         --rxoffload MASK
@@ -120,7 +120,9 @@ Where:
     Minimum value: RTE_MBUF_DEFAULT_BUF_SIZE (2176)
     Maximum value: UINT16_MAX (65535).
 
-*   ``-l``: enables code-path that uses librte_ipsec.
+*   ``-l 0 | 1``: Default is ``librte_ipsec`` code path.
+    0 enables the code path that uses legacy code.
+    1 enables the code path that uses ``librte_ipsec``.
 
 *   ``-w REPLAY_WINOW_SIZE``: specifies the IPsec sequence number replay window
     size for each Security Association (available only with librte_ipsec
examples/ipsec-secgw/ipsec-secgw.c +28 -16

94 lines not on this page · raw message

Signature · 1 line
2.7.4
Acked-by Tested-by on v1 1/2
#2 lore raw

Re: [dpdk-dev] [PATCH 1/2] examples/ipsec-secgw: set default to IPsec library mode

Quoted 112 lines

Set the default code path to librte_ipsec mode.
Add parameter 0 | 1 to -l option
Check for conflicting options, -w -a -e and reassembly options are not supported in legacy mode.
Show fragment table size.
Update print_usage.
Update the ipsec-secgw guide.
Update the release notes.

---
 doc/guides/rel_notes/release_19_11.rst   |  8 ++++++
 doc/guides/sample_app_ug/ipsec_secgw.rst |  6 +++--
 examples/ipsec-secgw/ipsec-secgw.c       | 44 ++++++++++++++++++++------------
 3 files changed, 40 insertions(+), 18 deletions(-)
doc/guides/rel_notes/release_19_11.rst +8 -0
index 8490d89..70143c5 100644
@@ -56,6 +56,14 @@ New Features
      Also, make sure to start the actual text at the margin.
      =========================================================
 
+* **Updated the IPsec Security Gateway application.**
+
+  The ``librte_ipsec`` code path is now the default code path in
+  ``ipsec-secgw``
+
+  * The ``-l`` command line option has been extended to take a 0 | 1 argument.
+    0 enables the code path that uses legacy code.
+    1 enables the code path that uses ``librte_ipsec``.
 
 Removed Items
 -------------
doc/guides/sample_app_ug/ipsec_secgw.rst +4 -2
index ad2d79e..17b00c0 100644
@@ -92,7 +92,7 @@ The application has a number of command line options::
 
    ./build/ipsec-secgw [EAL options] --
                         -p PORTMASK -P -u PORTMASK -j FRAMESIZE
-                        -l -w REPLAY_WINOW_SIZE -e -a
+                        -l 0 -w REPLAY_WINOW_SIZE -e -a
                         --config (port,queue,lcore)[,(port,queue,lcore]
                         --single-sa SAIDX
                         --rxoffload MASK
@@ -120,7 +120,9 @@ Where:
     Minimum value: RTE_MBUF_DEFAULT_BUF_SIZE (2176)
     Maximum value: UINT16_MAX (65535).
 
-*   ``-l``: enables code-path that uses librte_ipsec.
+*   ``-l 0 | 1``: Default is ``librte_ipsec`` code path.
+    0 enables the code path that uses legacy code.
+    1 enables the code path that uses ``librte_ipsec``.
 
 *   ``-w REPLAY_WINOW_SIZE``: specifies the IPsec sequence number replay window
     size for each Security Association (available only with librte_ipsec
examples/ipsec-secgw/ipsec-secgw.c +6 -4
index 0d1fd6a..3e4b9e0 100644
@@ -188,7 +188,7 @@ static uint32_t frame_buf_size = RTE_MBUF_DEFAULT_BUF_SIZE;
 static uint32_t mtu_size = RTE_ETHER_MTU;
 
 /* application wide librte_ipsec/SA parameters */
-struct app_sa_prm app_sa_prm = {.enable = 0};
+struct app_sa_prm app_sa_prm = {.enable = 1};
 
 struct lcore_rx_queue {
 	uint16_t port_id;
@@ -1259,7 +1259,7 @@ print_usage(const char *prgname)
 		" [-P]"
 		" [-u PORTMASK]"
 		" [-j FRAMESIZE]"
-		" [-l]"
+		" [-l 0 | 1]"
 		" [-w REPLAY_WINDOW_SIZE]"
 		" [-e]"
 		" [-a]"
@@ -1277,7 +1277,8 @@ print_usage(const char *prgname)
 		"  -u PORTMASK: Hexadecimal bitmask of unprotected ports\n"
 		"  -j FRAMESIZE: Data buffer size, minimum (and default)\n"
 		"     value: RTE_MBUF_DEFAULT_BUF_SIZE\n"
-		"  -l enables code-path that uses librte_ipsec\n"
+		"  -l 0 enables code-path that uses the legacy code\n"
+		"  -l 1 enables code-path that uses librte_ipsec\n"
 		"  -w REPLAY_WINDOW_SIZE specifies IPsec SQN replay window\n"
 		"     size for each SA\n"
 		"  -e enables ESN\n"
@@ -1418,6 +1419,7 @@ print_app_sa_prm(const struct app_sa_prm *prm)
 	printf("replay window size: %u\n", prm->window_size);
 	printf("ESN: %s\n", (prm->enable_esn == 0) ? "disabled" : "enabled");
 	printf("SA flags: %#" PRIx64 "\n", prm->flags);
+	printf("Fragment Table size %u\n", frag_tbl_sz);
 }
 
 static int32_t
@@ -1431,7 +1433,7 @@ parse_args(int32_t argc, char **argv)
 
 	argvopt = argv;
 
-	while ((opt = getopt_long(argc, argvopt, "aelp:Pu:f:j:w:",
+	while ((opt = getopt_long(argc, argvopt, "aep:Pu:f:j:w:l:",
 				lgopts, &option_index)) != EOF) {
 
 		switch (opt) {

examples/ipsec-secgw/ipsec-secgw.c parse_args(int32_t argc, char **argv) · line 1490

Quoted 7 lines
examples/ipsec-secgw/ipsec-secgw.c +3 -1
@@ -1483,18 +1485,28 @@ parse_args(int32_t argc, char **argv)
 			printf("Custom frame buffer size %u\n", frame_buf_size);
 			break;
 		case 'l':
-			app_sa_prm.enable = 1;
+			ret = parse_decimal(optarg);
+			if (ret == -1) {
+				printf("Invalid argument l %s\n", optarg);

Just as a nit, probably a bit better:
printf("Invalid argument: \'%s\' for: \'%c\' option\n", optarg, opt); Apart from that:

Quoted 50 lines
examples/ipsec-secgw/ipsec-secgw.c +19 -12
+				print_usage(prgname);
+				return -1;
+			} else if (ret == 0)
+				app_sa_prm.enable = 0;
+			else if (ret == 1)
+				app_sa_prm.enable = 1;
+			else {
+				printf("Invalid argument l %d\n", ret);
+				print_usage(prgname);
+				return -1;
+			}
 			break;
 		case 'w':
-			app_sa_prm.enable = 1;
 			app_sa_prm.window_size = parse_decimal(optarg);
 			break;
 		case 'e':
-			app_sa_prm.enable = 1;
 			app_sa_prm.enable_esn = 1;
 			break;
 		case 'a':
-			app_sa_prm.enable = 1;
 			app_sa_prm.flags |= RTE_IPSEC_SAFLAG_SQN_ATOM;
 			break;
 		case CMD_LINE_OPT_CONFIG_NUM:
@@ -1579,14 +1591,14 @@ parse_args(int32_t argc, char **argv)
 		return -1;
 	}
 
-	/* check do we need to enable multi-seg support */
-	if (multi_seg_required()) {
-		/* legacy mode doesn't support multi-seg */
-		app_sa_prm.enable = 1;
-		printf("frame buf size: %u, mtu: %u, "
-			"number of reassemble entries: %u\n"
-			"multi-segment support is required\n",
-			frame_buf_size, mtu_size, frag_tbl_sz);
+	if (app_sa_prm.enable == 0 &&
+			(app_sa_prm.window_size > 0 ||
+					app_sa_prm.enable_esn ||
+					app_sa_prm.flags != 0 ||
+					multi_seg_required())) {
+		printf("-w -e -a and reassembly options are not "
+				"supported in legacy mode\n");
+		return -1;
 	}
 
 	print_app_sa_prm(&app_sa_prm);
--

2.7.4

#3 lore raw

Re: [dpdk-dev] [PATCH 1/2] examples/ipsec-secgw: set default to IPsec library mode

Hi Konstantin,

<snip>

Quoted 133 lines
Quoted 9 lines

Set the default code path to librte_ipsec mode.
Add parameter 0 | 1 to -l option
Check for conflicting options, -w -a -e and reassembly options are not supported in legacy mode.
Show fragment table size.
Update print_usage.
Update the ipsec-secgw guide.
Update the release notes.

Quoted 5 lines
---
 doc/guides/rel_notes/release_19_11.rst   |  8 ++++++
 doc/guides/sample_app_ug/ipsec_secgw.rst |  6 +++--
 examples/ipsec-secgw/ipsec-secgw.c       | 44 ++++++++++++++++++++---

---------

doc/guides/rel_notes/release_19_11.rst New Features · line 57

Quoted 10 lines
 3 files changed, 40 insertions(+), 18 deletions(-)
doc/guides/rel_notes/release_19_11.rst +0 -0
index 8490d89..70143c5 100644
@@ -56,6 +56,14 @@ New Features
      Also, make sure to start the actual text at the margin.
 

=========================================================

doc/guides/rel_notes/release_19_11.rst New Features

Quoted 7 lines
doc/guides/rel_notes/release_19_11.rst +6 -0
+* **Updated the IPsec Security Gateway application.**
+
+  The ``librte_ipsec`` code path is now the default code path in
+ ``ipsec-secgw``
+
+  * The ``-l`` command line option has been extended to take a 0 | 1

argument.

doc/guides/sample_app_ug/ipsec_secgw.rst The application has a number of command line

Quoted 11 lines
doc/guides/rel_notes/release_19_11.rst +2 -0
+    0 enables the code path that uses legacy code.
+    1 enables the code path that uses ``librte_ipsec``.
 
 Removed Items
 -------------
doc/guides/sample_app_ug/ipsec_secgw.rst +0 -0
index ad2d79e..17b00c0 100644
@@ -92,7 +92,7 @@ The application has a number of command line

options::

Quoted 11 lines
doc/guides/sample_app_ug/ipsec_secgw.rst +1 -1
    ./build/ipsec-secgw [EAL options] --
                         -p PORTMASK -P -u PORTMASK -j FRAMESIZE
-                        -l -w REPLAY_WINOW_SIZE -e -a
+                        -l 0 -w REPLAY_WINOW_SIZE -e -a
                         --config (port,queue,lcore)[,(port,queue,lcore]
                         --single-sa SAIDX
                         --rxoffload MASK @@ -120,7 +120,9 @@ Where:
     Minimum value: RTE_MBUF_DEFAULT_BUF_SIZE (2176)
     Maximum value: UINT16_MAX (65535).

doc/guides/sample_app_ug/ipsec_secgw.rst Where:

Quoted 6 lines
doc/guides/sample_app_ug/ipsec_secgw.rst +3 -1
-*   ``-l``: enables code-path that uses librte_ipsec.
+*   ``-l 0 | 1``: Default is ``librte_ipsec`` code path.
+    0 enables the code path that uses legacy code.
+    1 enables the code path that uses ``librte_ipsec``.
 
 *   ``-w REPLAY_WINOW_SIZE``: specifies the IPsec sequence number

replay window

Quoted 8 lines

size for each Security Association (available only with
librte_ipsec diff --git a/examples/ipsec-secgw/ipsec-secgw.c b/examples/ipsec-secgw/ipsec-secgw.c
index 0d1fd6a..3e4b9e0 100644

examples/ipsec-secgw/ipsec-secgw.c +0 -0
@@ -188,7 +188,7 @@ static uint32_t frame_buf_size =

RTE_MBUF_DEFAULT_BUF_SIZE; static uint32_t mtu_size =

RTE_ETHER_MTU;

Quoted 17 lines

/* application wide librte_ipsec/SA parameters */ -struct app_sa_prm
app_sa_prm = {.enable = 0};

examples/ipsec-secgw/ipsec-secgw.c +2 -1
+struct app_sa_prm app_sa_prm = {.enable = 1};
 
 struct lcore_rx_queue {
 	uint16_t port_id;
@@ -1259,7 +1259,7 @@ print_usage(const char *prgname)
 		" [-P]"
 		" [-u PORTMASK]"
 		" [-j FRAMESIZE]"
-		" [-l]"
+		" [-l 0 | 1]"
 		" [-w REPLAY_WINDOW_SIZE]"
 		" [-e]"
 		" [-a]"

examples/ipsec-secgw/ipsec-secgw.c print_usage(const char *prgname) · line 1277

Quoted 1 line
examples/ipsec-secgw/ipsec-secgw.c +0 -0
@@ -1277,7 +1277,8 @@ print_usage(const char *prgname)
 		"  -u PORTMASK: Hexadecimal bitmask of unprotected

ports\n"

examples/ipsec-secgw/ipsec-secgw.c print_usage(const char *prgname)

Quoted 6 lines
examples/ipsec-secgw/ipsec-secgw.c +2 -1
 		"  -j FRAMESIZE: Data buffer size, minimum (and default)\n"
 		"     value: RTE_MBUF_DEFAULT_BUF_SIZE\n"
-		"  -l enables code-path that uses librte_ipsec\n"
+		"  -l 0 enables code-path that uses the legacy code\n"
+		"  -l 1 enables code-path that uses librte_ipsec\n"
 		"  -w REPLAY_WINDOW_SIZE specifies IPsec SQN replay

window\n"

examples/ipsec-secgw/ipsec-secgw.c print_app_sa_prm(const struct app_sa_prm

Quoted 3 lines
examples/ipsec-secgw/ipsec-secgw.c +0 -0
 		"     size for each SA\n"
 		"  -e enables ESN\n"
@@ -1418,6 +1419,7 @@ print_app_sa_prm(const struct app_sa_prm

*prm)

examples/ipsec-secgw/ipsec-secgw.c print_app_sa_prm(const struct app_sa_prm · line 1420

Quoted 2 lines
examples/ipsec-secgw/ipsec-secgw.c +0 -0
 	printf("replay window size: %u\n", prm->window_size);
 	printf("ESN: %s\n", (prm->enable_esn == 0) ? "disabled" :

"enabled");

Quoted 15 lines
examples/ipsec-secgw/ipsec-secgw.c +2 -1
 	printf("SA flags: %#" PRIx64 "\n", prm->flags);
+	printf("Fragment Table size %u\n", frag_tbl_sz);
 }
 
 static int32_t
@@ -1431,7 +1433,7 @@ parse_args(int32_t argc, char **argv)
 
 	argvopt = argv;
 
-	while ((opt = getopt_long(argc, argvopt, "aelp:Pu:f:j:w:",
+	while ((opt = getopt_long(argc, argvopt, "aep:Pu:f:j:w:l:",
 				lgopts, &option_index)) != EOF) {
 
 		switch (opt) {

examples/ipsec-secgw/ipsec-secgw.c parse_args(int32_t argc, char **argv) · line 1485

Quoted 1 line
examples/ipsec-secgw/ipsec-secgw.c +0 -0
@@ -1483,18 +1485,28 @@ parse_args(int32_t argc, char **argv)
 			printf("Custom frame buffer size %u\n",

frame_buf_size);

examples/ipsec-secgw/ipsec-secgw.c parse_args(int32_t argc, char **argv) · line 1490

Quoted 6 lines
examples/ipsec-secgw/ipsec-secgw.c +3 -1
 			break;
 		case 'l':
-			app_sa_prm.enable = 1;
+			ret = parse_decimal(optarg);
+			if (ret == -1) {
+				printf("Invalid argument l %s\n", optarg);

Just as a nit, probably a bit better:
printf("Invalid argument: \'%s\' for: \'%c\' option\n", optarg, opt); Apart from that:

Quoted 2 lines

I will send a v2.

Regards,

Bernard

<snip>

#4 lore raw

PATCH 2/2 examples/ipsec-secgw: add -l 0 parameter to old scripts

In reply to Bernard Iremonger

enable legacy mode in *_old scripts

---
 examples/ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh         | 2 +-
 examples/ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh | 2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh               | 2 +-
 examples/ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh         | 2 +-
 examples/ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh | 2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh               | 2 +-
 10 files changed, 10 insertions(+), 10 deletions(-)
examples/ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh +1 -1
index ffd945b..1f3a9f4 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_3descbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh +1 -1
index a3abb61..093b0c9 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aescbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh +1 -1
index 3aa0712..c6e8851 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesctr_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh +1 -1
index 0523049..e9fc4d8 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesgcm_inline_crypto_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh +1 -1
index 951e6b6..1653d29 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesgcm_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh +1 -1
index eaf248a..8fec47a 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_3descbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh +1 -1
index 3c0d8d1..9ff08f1 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aescbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh +1 -1
index 26f0d02..b23f0db 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesctr_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh +1 -1
index de65961..7574618 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesgcm_inline_crypto_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh +1 -1
index e0a015e..fabfe63 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesgcm_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
Signature · 1 line
2.7.4
#5 lore raw

Re: [dpdk-dev] [PATCH 2/2] examples/ipsec-secgw: add -l 0 parameter to old scripts

Hi Bernard,

Quoted 4 lines

enable legacy mode in *_old scripts

Quoted 12 lines
---
 examples/ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh         | 2 +-
 examples/ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh | 2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh               | 2 +-
 examples/ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh         | 2 +-
 examples/ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh | 2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh               | 2 +-
 10 files changed, 10 insertions(+), 10 deletions(-)

examples/ipsec-secgw/test/run_test.sh -p fails.
I think you need to update examples/ipsec-secgw/test/pkttest.sh too. Konstantin

Quoted 105 lines
examples/ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh +1 -1
index ffd945b..1f3a9f4 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_3descbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh +1 -1
index a3abb61..093b0c9 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aescbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh +1 -1
index 3aa0712..c6e8851 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesctr_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec- renamed from examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh +0 -0

secgw/test/trs_aesgcm_inline_crypto_old_defs.sh
index 0523049..e9fc4d8 100644

examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh +1 -1
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesgcm_inline_crypto_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh +1 -1
index 951e6b6..1653d29 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesgcm_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh +1 -1
index eaf248a..8fec47a 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_3descbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh +1 -1
index 3c0d8d1..9ff08f1 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aescbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh +1 -1
index 26f0d02..b23f0db 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesctr_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec- renamed from examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh +0 -0

secgw/test/tun_aesgcm_inline_crypto_old_defs.sh
index de65961..7574618 100644

examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh +1 -1
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesgcm_inline_crypto_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh +1 -2
index e0a015e..fabfe63 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesgcm_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
--

2.7.4

#6 lore raw

Re: [dpdk-dev] [PATCH 2/2] examples/ipsec-secgw: add -l 0 parameter to old scripts

Hi Konstantin,

Quoted 28 lines

-----Original Message-----
From: Ananyev, Konstantin
Sent: Thursday, September 26, 2019 2:43 PM
To: Iremonger, Bernard <[email protected]>; [email protected]; [email protected]
Subject: RE: [PATCH 2/2] examples/ipsec-secgw: add -l 0 parameter to old scripts

Hi Bernard,

Quoted 4 lines

enable legacy mode in *_old scripts

Quoted 12 lines
---
 examples/ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh         | 2 +-
 examples/ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh | 2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh               | 2 +-
 examples/ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh         | 2 +-
 examples/ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh | 2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh               | 2 +-
 10 files changed, 10 insertions(+), 10 deletions(-)
Quoted 3 lines

examples/ipsec-secgw/test/run_test.sh -p fails.
I think you need to update examples/ipsec-secgw/test/pkttest.sh too. Konstantin

I will check it.

Quoted 108 lines
Quoted 108 lines
examples/ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh +1 -1
index ffd945b..1f3a9f4 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_3descbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh +1 -1
index a3abb61..093b0c9 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aescbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh +1 -1
index 3aa0712..c6e8851 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesctr_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'

diff --git a/examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh b/examples/ipsec- secgw/test/trs_aesgcm_inline_crypto_old_defs.sh index 0523049..e9fc4d8 100644

examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh +1 -1
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesgcm_inline_crypto_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh +1 -1
index 951e6b6..1653d29 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesgcm_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh +1 -1
index eaf248a..8fec47a 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_3descbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh +1 -1
index 3c0d8d1..9ff08f1 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aescbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh +1 -1
index 26f0d02..b23f0db 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesctr_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'

diff --git a/examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh b/examples/ipsec- secgw/test/tun_aesgcm_inline_crypto_old_defs.sh index de65961..7574618 100644

examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh +1 -1
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesgcm_inline_crypto_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh +0 -0
index e0a015e..fabfe63 100644
@@ -2,4 +2,4 @@
 
Quoted 6 lines
Quoted 6 lines
examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh +1 -2
 . ${DIR}/tun_aesgcm_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
--

2.7.4

Regards,

Bernard

#7 lore raw

v2 cover examples/ipsec-secgw: set default

In reply to Bernard Iremonger

This patch set, sets the default code path in the ipsec-secgw application to use the librte_ipsec.
The *_old test scripts have been modified to use the legacy code path.

Changes in v2:
-------------
The error messages for the -l option have been updated. The pktest.sh script has been updated to drop the -l option.

Bernard Iremonger (3):
examples/ipsec-secgw: set default to IPsec library mode examples/ipsec-secgw: add -l 0 parameter to old scripts examples/ipsec-secgw: update pktest.sh script

 doc/guides/rel_notes/release_19_11.rst             |  8 ++++
 doc/guides/sample_app_ug/ipsec_secgw.rst           |  6 ++-
 examples/ipsec-secgw/ipsec-secgw.c                 | 46 ++++++++++++++--------
 examples/ipsec-secgw/test/pkttest.sh               |  1 -
 .../ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/trs_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/tun_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh   |  2 +-
 14 files changed, 52 insertions(+), 29 deletions(-)
Signature · 1 line
2.7.4
#8 lore raw

Hi Bernard,

Quoted 10 lines

This patch set, sets the default code path in the ipsec-secgw application to use the librte_ipsec.
The *_old test scripts have been modified to use the legacy code path.

Changes in v2:
-------------
The error messages for the -l option have been updated. The pktest.sh script has been updated to drop the -l option.

The patches looks good to me.
Just one thing - shouldn't we also update the docs (AG for ipsec-secgw)? Konstantin

Quoted 24 lines

Bernard Iremonger (3):
examples/ipsec-secgw: set default to IPsec library mode examples/ipsec-secgw: add -l 0 parameter to old scripts examples/ipsec-secgw: update pktest.sh script

 doc/guides/rel_notes/release_19_11.rst             |  8 ++++
 doc/guides/sample_app_ug/ipsec_secgw.rst           |  6 ++-
 examples/ipsec-secgw/ipsec-secgw.c                 | 46 ++++++++++++++--------
 examples/ipsec-secgw/test/pkttest.sh               |  1 -
 .../ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/trs_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/tun_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh   |  2 +-
 14 files changed, 52 insertions(+), 29 deletions(-)

--
2.7.4

#9 lore raw

Hi Konstantin,

Quoted 15 lines

-----Original Message-----
From: Ananyev, Konstantin
Sent: Tuesday, October 1, 2019 6:31 PM
To: Iremonger, Bernard <[email protected]>; [email protected]; [email protected]
Subject: RE: [PATCH v2 0/3] examples/ipsec-secgw: set default

Hi Bernard,

Quoted 5 lines

This patch set, sets the default code path in the ipsec-secgw application to use the librte_ipsec.
The *_old test scripts have been modified to use the legacy code path.

Quoted 8 lines
Quoted 4 lines

Changes in v2:
-------------
The error messages for the -l option have been updated. The pktest.sh script has been updated to drop the -l option.

The patches looks good to me.
Just one thing - shouldn't we also update the docs (AG for ipsec-secgw)? Konstantin

The 19.11 release notes and the sample application guide for ipsec-secgw have been updated. The updates are in the first patch (0001) with ipsec-secgw.c updates.

Quoted 5 lines
Quoted 5 lines

Bernard Iremonger (3):
examples/ipsec-secgw: set default to IPsec library mode examples/ipsec-secgw: add -l 0 parameter to old scripts examples/ipsec-secgw: update pktest.sh script

Quoted 19 lines
Quoted 3 lines
 doc/guides/rel_notes/release_19_11.rst             |  8 ++++
 doc/guides/sample_app_ug/ipsec_secgw.rst           |  6 ++-
 examples/ipsec-secgw/ipsec-secgw.c                 | 46 ++++++++++++++-------

-

Quoted 15 lines
 examples/ipsec-secgw/test/pkttest.sh               |  1 -
 .../ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/trs_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/tun_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh   |  2 +-
 14 files changed, 52 insertions(+), 29 deletions(-)

--
2.7.4

Regards,

Bernard.

#10 lore raw
Quoted 30 lines

-----Original Message-----
From: Iremonger, Bernard
Sent: Wednesday, October 2, 2019 10:00 AM
To: Ananyev, Konstantin <[email protected]>; [email protected]; [email protected]
Subject: RE: [PATCH v2 0/3] examples/ipsec-secgw: set default

Hi Konstantin,

Quoted 20 lines

-----Original Message-----
From: Ananyev, Konstantin
Sent: Tuesday, October 1, 2019 6:31 PM
To: Iremonger, Bernard <[email protected]>; [email protected]; [email protected]
Subject: RE: [PATCH v2 0/3] examples/ipsec-secgw: set default

Hi Bernard,

Quoted 9 lines

This patch set, sets the default code path in the ipsec-secgw application to use the librte_ipsec.
The *_old test scripts have been modified to use the legacy code path.

Changes in v2:
-------------
The error messages for the -l option have been updated. The pktest.sh script has been updated to drop the -l option.

Quoted 6 lines
Quoted 3 lines

The patches looks good to me.
Just one thing - shouldn't we also update the docs (AG for ipsec-secgw)? Konstantin

The 19.11 release notes and the sample application guide for ipsec-secgw have been updated. The updates are in the first patch (0001) with ipsec-secgw.c updates.

Ah, ok sorry for the noise then :)
Konstantin

Quoted 30 lines
Quoted 7 lines
Quoted 7 lines

Bernard Iremonger (3):
examples/ipsec-secgw: set default to IPsec library mode examples/ipsec-secgw: add -l 0 parameter to old scripts examples/ipsec-secgw: update pktest.sh script

Quoted 19 lines
Quoted 3 lines
 doc/guides/rel_notes/release_19_11.rst             |  8 ++++
 doc/guides/sample_app_ug/ipsec_secgw.rst           |  6 ++-
 examples/ipsec-secgw/ipsec-secgw.c                 | 46 ++++++++++++++-------

-

Quoted 15 lines
 examples/ipsec-secgw/test/pkttest.sh               |  1 -
 .../ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/trs_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/tun_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh   |  2 +-
 14 files changed, 52 insertions(+), 29 deletions(-)

--
2.7.4

Regards,

Bernard.

#11 lore raw
In reply to Bernard Iremonger

Hi All,

This patchset would need ack from more vendors as it will impact user experience on a key example application which is normally demonstrated to customers.

IPSec library is still evolving and there are new functionality added every release. Atleast from NXP side we are not OK with this change.

I would hold this patch till RC2 atleast.

Regards,
Akhil

Quoted 41 lines

-----Original Message-----
From: Bernard Iremonger <[email protected]>
Sent: Tuesday, October 1, 2019 8:48 PM
To: [email protected]; [email protected]; Akhil Goyal <[email protected]>
Cc: Bernard Iremonger <[email protected]>
Subject: [PATCH v2 0/3] examples/ipsec-secgw: set default

This patch set, sets the default code path in the ipsec-secgw application to use the librte_ipsec.
The *_old test scripts have been modified to use the legacy code path.

Changes in v2:
-------------
The error messages for the -l option have been updated. The pktest.sh script has been updated to drop the -l option.

Bernard Iremonger (3):
examples/ipsec-secgw: set default to IPsec library mode examples/ipsec-secgw: add -l 0 parameter to old scripts examples/ipsec-secgw: update pktest.sh script

 doc/guides/rel_notes/release_19_11.rst             |  8 ++++
 doc/guides/sample_app_ug/ipsec_secgw.rst           |  6 ++-
 examples/ipsec-secgw/ipsec-secgw.c                 | 46 ++++++++++++++--------
 examples/ipsec-secgw/test/pkttest.sh               |  1 -
 .../ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/trs_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/tun_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh   |  2 +-
 14 files changed, 52 insertions(+), 29 deletions(-)

--
2.7.4

#12 lore raw

Hi Akhil,

With this patch applied the legacy code path in the ipsec-secgw application is still available. The default code path is now to use librte_ipsec. The "-l 0" option at startup allows the legacy code path to be used.
Both code paths are still available.

Regards,

Bernard

Quoted 65 lines

-----Original Message-----
From: Akhil Goyal [mailto:[email protected]]
Sent: Friday, October 11, 2019 1:40 PM
To: Iremonger, Bernard <[email protected]>; [email protected]; Ananyev, Konstantin <[email protected]>; [email protected]; [email protected]; 'Thomas Monjalon' <[email protected]>
Subject: RE: [PATCH v2 0/3] examples/ipsec-secgw: set default

Hi All,

This patchset would need ack from more vendors as it will impact user experience on a key example application which is normally demonstrated to customers.

IPSec library is still evolving and there are new functionality added every release.
Atleast from NXP side we are not OK with this change.

I would hold this patch till RC2 atleast.

Regards,
Akhil

Quoted 17 lines

-----Original Message-----
From: Bernard Iremonger <[email protected]>
Sent: Tuesday, October 1, 2019 8:48 PM
To: [email protected]; [email protected]; Akhil Goyal <[email protected]>
Cc: Bernard Iremonger <[email protected]>
Subject: [PATCH v2 0/3] examples/ipsec-secgw: set default

This patch set, sets the default code path in the ipsec-secgw application to use the librte_ipsec.
The *_old test scripts have been modified to use the legacy code path.

Changes in v2:
-------------
The error messages for the -l option have been updated. The pktest.sh script has been updated to drop the -l option.

Quoted 8 lines

Bernard Iremonger (3):
examples/ipsec-secgw: set default to IPsec library mode examples/ipsec-secgw: add -l 0 parameter to old scripts examples/ipsec-secgw: update pktest.sh script

 doc/guides/rel_notes/release_19_11.rst             |  8 ++++
 doc/guides/sample_app_ug/ipsec_secgw.rst           |  6 ++-
 examples/ipsec-secgw/ipsec-secgw.c                 | 46 ++++++++++++++-------

-

Quoted 15 lines
 examples/ipsec-secgw/test/pkttest.sh               |  1 -
 .../ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/trs_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/tun_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh   |  2 +-
 14 files changed, 52 insertions(+), 29 deletions(-)

--
2.7.4

#13 lore raw
In reply to Akhil Goyal

11/10/2019 14:40, Akhil Goyal:

Quoted 7 lines

Hi All,

This patchset would need ack from more vendors as it will impact user experience on a key example application which is normally demonstrated to customers.

IPSec library is still evolving and there are new functionality added every release. Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

#14 lore raw
Quoted 7 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 3 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 9 lines
Quoted 3 lines

on a key example application which is normally demonstrated to customers.

IPSec library is still evolving and there are new functionality added every

release.

Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance numbers.

#15 lore raw

Hi Akhil,

Quoted 23 lines

-----Original Message-----
From: Akhil Goyal [mailto:[email protected]]
Sent: Tuesday, October 15, 2019 7:43 AM
To: Thomas Monjalon <[email protected]>
Cc: Iremonger, Bernard <[email protected]>; [email protected]; Ananyev, Konstantin <[email protected]>; [email protected]; [email protected]
Subject: RE: [PATCH v2 0/3] examples/ipsec-secgw: set default

Quoted 8 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 4 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 1 line

on a key example application which is normally demonstrated to

customers.

Quoted 4 lines
Quoted 3 lines

IPSec library is still evolving and there are new functionality added every

release.

Quoted 8 lines
Quoted 5 lines
Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance numbers.

Could you give some details on the performance issues you are seeing.

Regards,

Bernard.

#16 lore raw
Quoted 17 lines
Quoted 17 lines
Quoted 10 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 4 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 1 line

on a key example application which is normally demonstrated to

customers.

Quoted 6 lines
Quoted 3 lines

IPSec library is still evolving and there are new functionality added every

release.

Quoted 1 line

Atleast from NXP side we are not OK with this change.

Quoted 9 lines
Quoted 6 lines
Quoted 3 lines

What can be changed in the library to make it acceptable as a default in this example?

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance numbers.

Could you give some details on the performance issues you are seeing.

We were observing about 4-5% drop when using the ipsec-lib instead of the Legacy code path. We would again measure it on RC1. That is why I say, I will Hold this patch till RC2, unless some other vendor also confirms that.

#17 lore raw
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 7 lines
Quoted 5 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 1 line

on a key example application which is normally demonstrated to

customers.

Quoted 9 lines
Quoted 3 lines

IPSec library is still evolving and there are new functionality added every

release.

Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance numbers.

Could you give some details on the performance issues you are seeing.

We were observing about 4-5% drop when using the ipsec-lib instead of the Legacy code path. We would again measure it on RC1.

Yeh, I remember you already mentioned that few months ago... Though we don't have access to NXP HW, so obviously we do need some help from you guys here to address it:
configuration/test details so we can try to reproduce it probably some insight from you what in particular causing that slowdown.. patches would be even better :)
Konstantin

Quoted 2 lines

That is why I say, I will
Hold this patch till RC2, unless some other vendor also confirms that.

#18 lore raw
In reply to Akhil Goyal

Hi Akhil,

Quoted 32 lines

-----Original Message-----
From: Akhil Goyal <[email protected]>
Sent: Tuesday, October 15, 2019 4:05 PM
To: Iremonger, Bernard <[email protected]>; Thomas Monjalon <[email protected]>
Cc: [email protected]; Ananyev, Konstantin <[email protected]>; [email protected]; [email protected]
Subject: RE: [PATCH v2 0/3] examples/ipsec-secgw: set default

Quoted 13 lines
Quoted 13 lines
Quoted 8 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 4 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 1 line

on a key example application which is normally demonstrated to

customers.

Quoted 4 lines
Quoted 3 lines

IPSec library is still evolving and there are new functionality added every

release.

Quoted 8 lines
Quoted 8 lines
Quoted 5 lines
Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance

Quoted 8 lines

numbers.

Quoted 3 lines

Could you give some details on the performance issues you are seeing.

We were observing about 4-5% drop when using the ipsec-lib instead of the Legacy code path. We would again measure it on RC1. That is why I say, I will Hold this patch till RC2, unless some other vendor also confirms that.

Is there any update on performance measurements on 19.11-rc1 ?

Regards,

Bernard.

#19 lore raw

Hi Bernard,

Quoted 28 lines

Hi Akhil,

Quoted 25 lines
Quoted 15 lines
Quoted 15 lines
Quoted 10 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 4 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 1 line

on a key example application which is normally demonstrated to

customers.

Quoted 4 lines
Quoted 3 lines

IPSec library is still evolving and there are new functionality added every

release.

Quoted 8 lines
Quoted 8 lines
Quoted 5 lines
Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance

numbers.

Quoted 1 line
Quoted 9 lines
Quoted 6 lines
Quoted 2 lines

Could you give some details on the performance issues you are seeing.

We were observing about 4-5% drop when using the ipsec-lib instead of the Legacy code path. We would again measure it on RC1. That is why I say, I will Hold this patch till RC2, unless some other vendor also confirms that.

Is there any update on performance measurements on 19.11-rc1 ?

The performance impact of this patch is huge ~10% w.r.t. 19.11-rc1 base on NXP hardware.

We cannot merge this. Anoob also reported performance issues on Marvell hardware.

Quoted 3 lines

Regards,

Bernard.

#20 lore raw

Hi Akhil,

Quoted 30 lines
Quoted 30 lines
Quoted 29 lines
Quoted 13 lines
Quoted 13 lines
Quoted 8 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 4 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 1 line

on a key example application which is normally demonstrated to

customers.

Quoted 4 lines
Quoted 3 lines

IPSec library is still evolving and there are new functionality added every

release.

Quoted 8 lines
Quoted 8 lines
Quoted 5 lines
Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance

numbers.

Quoted 3 lines

Could you give some details on the performance issues you are seeing.

We were observing about 4-5% drop when using the ipsec-lib instead of the Legacy code path. We would again measure it on RC1. That is why I say, I will Hold this patch till RC2, unless some other vendor also confirms that.

Quoted 5 lines
Quoted 2 lines

Is there any update on performance measurements on 19.11-rc1 ?

The performance impact of this patch is huge ~10% w.r.t. 19.11-rc1 base on NXP hardware.

We cannot merge this. Anoob also reported performance issues on Marvell hardware.

Sure, 10% is a lot, so more than understandable.
Though, I think we do need to decide our future goals for it.
I see two main options here:
1. Make lib code-path on par with legacy one in terms of performance,
deprecate and then remove legacy code-path.
Till that happen (deprecation/removal) to minimize code divergence,
forbid to add new features to legacy code path only.
New features should be added to both paths, or library code path.
Obviously that one looks like a preferred option to me,
but it requires some effort from all interested parties (Intel, NXP, Marvell, ...). If everyone is ok with it, then I think it would be good to have some draft timeline here. If you guys are not interested in this effort, then the only other approach I can think about:
2. split ipsec-secgw app into 2 (one using librte_ipsec, second using raw devices (legacy one)). We probably can still try to keep some code shared by 2 apps:
(configuration/initialization/session management (SAD, SPD)),
but actual packet processing path will be different.
I really don't like that option, but I think we need to come-up with clear decision, one way or another.

Thanks
Konstantin

#21 lore raw

Hi Konstantin,

Quoted 56 lines

Hi Akhil,

Quoted 19 lines
Quoted 19 lines
Quoted 19 lines
Quoted 19 lines
Quoted 19 lines
Quoted 8 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 4 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 1 line

on a key example application which is normally demonstrated to

customers.

Quoted 9 lines
Quoted 3 lines

IPSec library is still evolving and there are new functionality added every

release.

Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 2 lines
Quoted 2 lines

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance

numbers.

Quoted 3 lines

Could you give some details on the performance issues you are seeing.

We were observing about 4-5% drop when using the ipsec-lib instead of

the

Quoted 1 line
Quoted 1 line
Quoted 1 line

Legacy code path. We would again measure it on RC1. That is why I say, I

will

Quoted 5 lines
Quoted 4 lines
Quoted 1 line

Hold this patch till RC2, unless some other vendor also confirms that.

Is there any update on performance measurements on 19.11-rc1 ?

The performance impact of this patch is huge ~10% w.r.t. 19.11-rc1 base on

NXP hardware.

Quoted 2 lines

We cannot merge this. Anoob also reported performance issues on Marvell

hardware.

Sure, 10% is a lot, so more than understandable.
Though, I think we do need to decide our future goals for it.
I see two main options here:
1. Make lib code-path on par with legacy one in terms of performance, deprecate and then remove legacy code-path.
Till that happen (deprecation/removal) to minimize code divergence,
forbid to add new features to legacy code path only.
New features should be added to both paths, or library code path.
Obviously that one looks like a preferred option to me,
but it requires some effort from all interested parties (Intel, NXP, Marvell, ...). If everyone is ok with it, then I think it would be good to have some draft timeline here.
If you guys are not interested in this effort, then the only other approach I can

Quoted 9 lines

think about:
2. split ipsec-secgw app into 2 (one using librte_ipsec, second using raw devices (legacy one)).
We probably can still try to keep some code shared by 2 apps:
(configuration/initialization/session management (SAD, SPD)),
but actual packet processing path will be different.
I really don't like that option, but I think we need to come-up with clear decision, one way or another.

IMO, Option 1 is the only way forward. From NXP side, we can start our work on this post 19.11 release and should target in 20.02 release.

Regards,
Akhil

#22 lore raw

Hi Akhil,

Quoted 64 lines
Quoted 54 lines
Quoted 20 lines
Quoted 20 lines
Quoted 20 lines
Quoted 20 lines
Quoted 20 lines
Quoted 9 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 4 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 1 line

on a key example application which is normally demonstrated to

customers.

Quoted 9 lines
Quoted 3 lines

IPSec library is still evolving and there are new functionality added every

release.

Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 2 lines
Quoted 2 lines

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance

numbers.

Quoted 3 lines

Could you give some details on the performance issues you are seeing.

We were observing about 4-5% drop when using the ipsec-lib instead of

the

Quoted 1 line
Quoted 1 line
Quoted 1 line

Legacy code path. We would again measure it on RC1. That is why I say, I

will

Quoted 5 lines
Quoted 4 lines
Quoted 1 line

Hold this patch till RC2, unless some other vendor also confirms that.

Is there any update on performance measurements on 19.11-rc1 ?

The performance impact of this patch is huge ~10% w.r.t. 19.11-rc1 base on

NXP hardware.

Quoted 2 lines

We cannot merge this. Anoob also reported performance issues on Marvell

hardware.

Sure, 10% is a lot, so more than understandable.
Though, I think we do need to decide our future goals for it.
I see two main options here:
1. Make lib code-path on par with legacy one in terms of performance, deprecate and then remove legacy code-path.
Till that happen (deprecation/removal) to minimize code divergence,
forbid to add new features to legacy code path only.
New features should be added to both paths, or library code path.
Obviously that one looks like a preferred option to me,
but it requires some effort from all interested parties (Intel, NXP, Marvell, ...). If everyone is ok with it, then I think it would be good to have some draft timeline here.
If you guys are not interested in this effort, then the only other approach I can

Quoted 9 lines

think about:
2. split ipsec-secgw app into 2 (one using librte_ipsec, second using raw devices (legacy one)).
We probably can still try to keep some code shared by 2 apps:
(configuration/initialization/session management (SAD, SPD)),
but actual packet processing path will be different.
I really don't like that option, but I think we need to come-up with clear decision, one way or another.

Quoted 1 line

IMO, Option 1 is the only way forward. From NXP side, we can start our work on this post 19.11 release and should target in 20.02 release.

Great to hear.
Thanks for clarification.
Konstantin

#23 lore raw
In reply to Ananyev, Konstantin

Hi Akhil, Konstantin,

Please see inline.

Thanks,
Anoob

Quoted 55 lines

-----Original Message-----
From: Ananyev, Konstantin <[email protected]>
Sent: Monday, November 4, 2019 8:55 PM
To: Akhil Goyal <[email protected]>; Iremonger, Bernard <[email protected]>; Thomas Monjalon <[email protected]>
Cc: [email protected]; Anoob Joseph <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>; dpdk-techboard <[email protected]>
Subject: [EXT] RE: [PATCH v2 0/3] examples/ipsec-secgw: set default

External Email

----------------------------------------------------------------------
Hi Akhil,

Quoted 23 lines
Quoted 23 lines
Quoted 23 lines
Quoted 23 lines
Quoted 23 lines
Quoted 9 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 4 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 2 lines

on a key example application which is normally
demonstrated to

customers.

Quoted 9 lines
Quoted 3 lines

IPSec library is still evolving and there are new
functionality added every

release.

Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

numbers.

Quoted 3 lines

Could you give some details on the performance issues you are seeing.

We were observing about 4-5% drop when using the ipsec-lib instead of the Legacy code path. We would again measure it on RC1. That is why I say, I will Hold this patch till RC2, unless some other vendor also

confirms that.

Quoted 4 lines
Quoted 3 lines

Is there any update on performance measurements on 19.11-rc1 ?

The performance impact of this patch is huge ~10% w.r.t. 19.11-rc1 base on

NXP hardware.

Quoted 2 lines

We cannot merge this. Anoob also reported performance issues on Marvell

hardware.

Quoted 14 lines

Sure, 10% is a lot, so more than understandable.
Though, I think we do need to decide our future goals for it.
I see two main options here:
1. Make lib code-path on par with legacy one in terms of performance, deprecate and then remove legacy code-path.
Till that happen (deprecation/removal) to minimize code divergence,
forbid to add new features to legacy code path only.
New features should be added to both paths, or library code path.
Obviously that one looks like a preferred option to me, but it requires some effort from all interested parties (Intel, NXP, Marvell, ...).
If everyone is ok with it, then I think it would be good to have some draft timeline here.
If you guys are not interested in this effort, then the only other approach I can think about:

[Anoob] I would say this is the right option. But then, there are features getting added only to lib ipsec mode. There are features like "fallback session" or anti replay which is only added for lib ipsec mode and not for non lib ipsec mode. Such features are good to have but would cost extra cycles in the datapath. Since it is only added for lib ipsec mode, the perf divergence between lib ipsec mode and non lib ipsec mode is fairly obvious.

So what is the solution? Both the modes need to be at par if our end strategy is going to deprecate one. If we need to reach a state where we can do apples to apples comparison, new features should be added to both modes and there should be a consensus on the feature implementations.

Also, looking at the "fallback session" feature, the feature was "pushed" without properly addressing the issues. The solution was hardly suitable for inline ipsec and the comments were ignored. Marvell is not ok with adding checks in datapath for an incomplete feature. Marvell withdrew the objections when it was conveyed that the review comments were deemed invaluable. Also because it was added to lib ipsec path which is not being used currently for our benchmarks. Marvell will be submitting an RFC for supporting fallback session with few changes in the rte_security library. When we attempt that, we can take care of only non lib ipsec mode as lib ipsec mode already has one and Marvell cannot work on two different fronts, when the other contributors themselves don't care about other established modes.

Also considering that lib ipsec is under constant change, (and is still experimental), I would not be too excited about making it the default immediately. I see that there are usages supported by the non lib ipsec mode, which is not supported by lib ipsec mode.

For example, ipv4 & ipv6 using same SPI is valid for non lib ipsec, but not for lib ipsec. Because of this, the default conf doesn't even run when launched with lib ipsec mode. I'm not sure whether we should rush with making it the default when the whole idea is still "experimental".

Quoted 15 lines

2. split ipsec-secgw app into 2 (one using librte_ipsec, second using raw devices (legacy one)).
We probably can still try to keep some code shared by 2 apps:
(configuration/initialization/session management (SAD, SPD)), but actual packet processing path will be different.
I really don't like that option, but I think we need to come-up with clear decision, one way or another.

Thanks
Konstantin

#24 lore raw
Quoted 58 lines
Quoted 43 lines
Quoted 25 lines
Quoted 25 lines
Quoted 25 lines
Quoted 25 lines
Quoted 25 lines
Quoted 11 lines

11/10/2019 14:40, Akhil Goyal:

Quoted 4 lines

Hi All,

This patchset would need ack from more vendors as it will impact user

experience

Quoted 2 lines

on a key example application which is normally
demonstrated to

customers.

Quoted 9 lines
Quoted 3 lines

IPSec library is still evolving and there are new
functionality added every

release.

Quoted 1 line

Atleast from NXP side we are not OK with this change.

What can be changed in the library to make it acceptable as a default in this example?

We are observing performance issues with ipsec library. So would request other Vendors to confirm if they are OK with the performance

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

numbers.

Quoted 3 lines

Could you give some details on the performance issues you are seeing.

We were observing about 4-5% drop when using the ipsec-lib instead of the Legacy code path. We would again measure it on RC1. That is why I say, I will Hold this patch till RC2, unless some other vendor also

confirms that.

Quoted 4 lines
Quoted 3 lines

Is there any update on performance measurements on 19.11-rc1 ?

The performance impact of this patch is huge ~10% w.r.t. 19.11-rc1 base on

NXP hardware.

Quoted 2 lines

We cannot merge this. Anoob also reported performance issues on Marvell

hardware.

Quoted 14 lines

Sure, 10% is a lot, so more than understandable.
Though, I think we do need to decide our future goals for it.
I see two main options here:
1. Make lib code-path on par with legacy one in terms of performance, deprecate and then remove legacy code-path.
Till that happen (deprecation/removal) to minimize code divergence,
forbid to add new features to legacy code path only.
New features should be added to both paths, or library code path.
Obviously that one looks like a preferred option to me, but it requires some effort from all interested parties (Intel, NXP, Marvell, ...).
If everyone is ok with it, then I think it would be good to have some draft timeline here.
If you guys are not interested in this effort, then the only other approach I can think about:

Quoted 4 lines

[Anoob] I would say this is the right option. But then, there are features getting added only to lib ipsec mode. There are features like "fallback session" or anti replay which is only added for lib ipsec mode and not for non lib ipsec mode. Such features are good to have but would cost extra cycles in the datapath. Since it is only added for lib ipsec mode, the perf divergence between lib ipsec mode and non lib ipsec mode is fairly obvious.

AFAIK, all these features are optional and can be disabled.
For performance comparison, I'd expect we run lib mode with extra features disabled.

Quoted 35 lines

So what is the solution? Both the modes need to be at par if our end strategy is going to deprecate one. If we need to reach a state where we can do apples to apples comparison, new features should be added to both modes and there should be a consensus on the feature implementations.

Also, looking at the "fallback session" feature, the feature was "pushed" without properly addressing the issues. The solution was hardly suitable for inline ipsec and the comments were ignored. Marvell is not ok with adding checks in datapath for an incomplete feature. Marvell withdrew the objections when it was conveyed that the review comments were deemed invaluable. Also because it was added to lib ipsec path which is not being used currently for our benchmarks. Marvell will be submitting an RFC for supporting fallback session with few changes in the rte_security library. When we attempt that, we can take care of only non lib ipsec mode as lib ipsec mode already has one and Marvell cannot work on two different fronts, when the other contributors themselves don't care about other established modes.

Also considering that lib ipsec is under constant change, (and is still experimental), I would not be too excited about making it the default immediately. I see that there are usages supported by the non lib ipsec mode, which is not supported by lib ipsec mode.

For example, ipv4 & ipv6 using same SPI is valid for non lib ipsec, but not for lib ipsec. Because of this, the default conf doesn't even run when launched with lib ipsec mode. I'm not sure whether we should rush with making it the default when the whole idea is still "experimental".

Quoted 15 lines

2. split ipsec-secgw app into 2 (one using librte_ipsec, second using raw devices (legacy one)).
We probably can still try to keep some code shared by 2 apps:
(configuration/initialization/session management (SAD, SPD)), but actual packet processing path will be different.
I really don't like that option, but I think we need to come-up with clear decision, one way or another.

Thanks
Konstantin

#25 lore raw
In reply to Akhil Goyal

Hi Akhil,

Marvell is NOT OK with this change in this cycle. The library is still emerging and the boundary between ipsec-secgw and lib ipsec still hasn't emerged clearly. Moreover, lib_ipsec mode isn't even working with the default conf (I believe the fix for this still in discussion. The issue was reported by a Marvell engineer some time back). There are questionable features getting pushed, for which Marvell's concerns were subdued.

I would expect this taken up after lib ipsec is found stable for atleast one release cycle.

As for the performance, we found 2-3% drop with the lib ipsec in the last release (lookaside crypto mode). Yet to try out with latest codebase. With some of the recent submissions, I expect an even higher performance hit. Also, eventmode ipsec-secgw (already submitted RFC) is using non lib ipsec mode. We don't have plans to move that to lib ipsec mode until lib ipsec is made stable.

Thanks,
Anoob

Quoted 68 lines

-----Original Message-----
From: Akhil Goyal <[email protected]>
Sent: Friday, October 11, 2019 6:10 PM
To: Bernard Iremonger <[email protected]>; [email protected]; [email protected]; Anoob Joseph <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>; 'Thomas Monjalon' <[email protected]>
Subject: [EXT] RE: [PATCH v2 0/3] examples/ipsec-secgw: set default

External Email

----------------------------------------------------------------------
Hi All,

This patchset would need ack from more vendors as it will impact user experience on a key example application which is normally demonstrated to customers.

IPSec library is still evolving and there are new functionality added every release.
Atleast from NXP side we are not OK with this change.

I would hold this patch till RC2 atleast.

Regards,
Akhil

Quoted 17 lines

-----Original Message-----
From: Bernard Iremonger <[email protected]>
Sent: Tuesday, October 1, 2019 8:48 PM
To: [email protected]; [email protected]; Akhil Goyal <[email protected]>
Cc: Bernard Iremonger <[email protected]>
Subject: [PATCH v2 0/3] examples/ipsec-secgw: set default

This patch set, sets the default code path in the ipsec-secgw application to use the librte_ipsec.
The *_old test scripts have been modified to use the legacy code path.

Changes in v2:
-------------
The error messages for the -l option have been updated. The pktest.sh script has been updated to drop the -l option.

Quoted 8 lines

Bernard Iremonger (3):
examples/ipsec-secgw: set default to IPsec library mode examples/ipsec-secgw: add -l 0 parameter to old scripts examples/ipsec-secgw: update pktest.sh script

 doc/guides/rel_notes/release_19_11.rst             |  8 ++++
 doc/guides/sample_app_ug/ipsec_secgw.rst           |  6 ++-
 examples/ipsec-secgw/ipsec-secgw.c                 | 46 ++++++++++++++-------

-

Quoted 15 lines
 examples/ipsec-secgw/test/pkttest.sh               |  1 -
 .../ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/trs_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh  |  2 +-
 .../ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh   |  2 +-
 .../ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh   |  2 +-
 .../test/tun_aesgcm_inline_crypto_old_defs.sh      |  2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh   |  2 +-
 14 files changed, 52 insertions(+), 29 deletions(-)

--
2.7.4

#26 lore raw

PATCH v2 1/3 examples/ipsec-secgw: set default to IPsec library mode

In reply to Bernard Iremonger

Set the default code path to librte_ipsec mode.
Add parameter 0 | 1 to -l option and update error message. Check for conflicting options, -w -a -e and reassembly options are not supported in legacy mode.
Show fragment table size.
Update print_usage.
Update the ipsec-secgw guide.
Update the release notes.

---
 doc/guides/rel_notes/release_19_11.rst   |  8 ++++++
 doc/guides/sample_app_ug/ipsec_secgw.rst |  6 +++--
 examples/ipsec-secgw/ipsec-secgw.c       | 46 +++++++++++++++++++++-----------
 3 files changed, 42 insertions(+), 18 deletions(-)
doc/guides/rel_notes/release_19_11.rst +8 -0
index 27cfbd9..b07e3bf 100644
@@ -56,6 +56,14 @@ New Features
      Also, make sure to start the actual text at the margin.
      =========================================================
 
+* **Updated the IPsec Security Gateway application.**
+
+  The ``librte_ipsec`` code path is now the default code path in
+  ``ipsec-secgw``
+
+  * The ``-l`` command line option has been extended to take a 0 | 1 argument.
+    0 enables the code path that uses legacy code.
+    1 enables the code path that uses ``librte_ipsec``.
 
 Removed Items
 -------------
doc/guides/sample_app_ug/ipsec_secgw.rst +4 -2
index ad2d79e..17b00c0 100644
@@ -92,7 +92,7 @@ The application has a number of command line options::
 
    ./build/ipsec-secgw [EAL options] --
                         -p PORTMASK -P -u PORTMASK -j FRAMESIZE
-                        -l -w REPLAY_WINOW_SIZE -e -a
+                        -l 0 -w REPLAY_WINOW_SIZE -e -a
                         --config (port,queue,lcore)[,(port,queue,lcore]
                         --single-sa SAIDX
                         --rxoffload MASK
@@ -120,7 +120,9 @@ Where:
     Minimum value: RTE_MBUF_DEFAULT_BUF_SIZE (2176)
     Maximum value: UINT16_MAX (65535).
 
-*   ``-l``: enables code-path that uses librte_ipsec.
+*   ``-l 0 | 1``: Default is ``librte_ipsec`` code path.
+    0 enables the code path that uses legacy code.
+    1 enables the code path that uses ``librte_ipsec``.
 
 *   ``-w REPLAY_WINOW_SIZE``: specifies the IPsec sequence number replay window
     size for each Security Association (available only with librte_ipsec
examples/ipsec-secgw/ipsec-secgw.c +30 -16

96 lines not on this page · raw message

Signature · 1 line
2.7.4
Acked-by Radu Nicolau on v2 1/3
#27 lore raw

Re: [dpdk-dev] [PATCH v2 1/3] examples/ipsec-secgw: set default to IPsec library mode

Quoted 9 lines · Bernard Iremonger

Set the default code path to librte_ipsec mode.
Add parameter 0 | 1 to -l option and update error message. Check for conflicting options, -w -a -e and reassembly options are not supported in legacy mode.
Show fragment table size.
Update print_usage.
Update the ipsec-secgw guide.
Update the release notes.

Quoted 4 lines

---

Acked-by Radu Nicolau <[email protected] <mailto:[email protected]>>
#28 lore raw

PATCH v2 2/3 examples/ipsec-secgw: add -l 0 parameter to old scripts

In reply to Bernard Iremonger

enable legacy mode in *_old scripts

---
 examples/ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh         | 2 +-
 examples/ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh | 2 +-
 examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh               | 2 +-
 examples/ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh         | 2 +-
 examples/ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh          | 2 +-
 examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh | 2 +-
 examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh               | 2 +-
 10 files changed, 10 insertions(+), 10 deletions(-)
examples/ipsec-secgw/test/trs_3descbc_sha1_old_defs.sh +1 -1
index ffd945b..1f3a9f4 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_3descbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aescbc_sha1_old_defs.sh +1 -1
index a3abb61..093b0c9 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aescbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesctr_sha1_old_defs.sh +1 -1
index 3aa0712..c6e8851 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesctr_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesgcm_inline_crypto_old_defs.sh +1 -1
index 0523049..e9fc4d8 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesgcm_inline_crypto_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/trs_aesgcm_old_defs.sh +1 -1
index 951e6b6..1653d29 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/trs_aesgcm_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_3descbc_sha1_old_defs.sh +1 -1
index eaf248a..8fec47a 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_3descbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aescbc_sha1_old_defs.sh +1 -1
index 3c0d8d1..9ff08f1 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aescbc_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesctr_sha1_old_defs.sh +1 -1
index 26f0d02..b23f0db 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesctr_sha1_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesgcm_inline_crypto_old_defs.sh +1 -1
index de65961..7574618 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesgcm_inline_crypto_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
examples/ipsec-secgw/test/tun_aesgcm_old_defs.sh +1 -1
index e0a015e..fabfe63 100644
@@ -2,4 +2,4 @@
 
 . ${DIR}/tun_aesgcm_defs.sh
 
-SGW_CMD_XPRM=
+SGW_CMD_XPRM='-l 0'
Signature · 1 line
2.7.4
Acked-by Tested-by on v2 2/3
#29 lore raw

Re: [dpdk-dev] [PATCH v2 2/3] examples/ipsec-secgw: add -l 0 parameter to old scripts

Quoted 9 lines

-----Original Message-----
From: Iremonger, Bernard
Sent: Tuesday, October 1, 2019 4:18 PM
To: [email protected]; Ananyev, Konstantin <[email protected]>; [email protected]
Cc: Iremonger, Bernard <[email protected]>
Subject: [PATCH v2 2/3] examples/ipsec-secgw: add -l 0 parameter to old scripts

Quoted 4 lines

enable legacy mode in *_old scripts

---

Quoted 1 line

2.7.4

#31 lore raw

PATCH v2 3/3 examples/ipsec-secgw: update pktest.sh script

In reply to Bernard Iremonger

Drop SGW_CMD_XPRM=l, as default is IPsec library now.

---
 examples/ipsec-secgw/test/pkttest.sh | 1 -
 1 file changed, 1 deletion(-)
examples/ipsec-secgw/test/pkttest.sh +0 -1
index 04cd96b..7c78de7 100755
@@ -18,7 +18,6 @@ fi
 CRYPTO_DEV="--vdev=crypto_null0"
 SGW_PORT_CFG="--vdev=net_tap0,mac=fixed --vdev=net_tap1,mac=fixed"
 SGW_EAL_XPRM="--no-pci"
-SGW_CMD_XPRM=-l
 SGW_WAIT_DEV="dtap0"
 . ${DIR}/common_defs_secgw.sh
 
Signature · 1 line
2.7.4
Acked-by Tested-by on v2 3/3
#32 lore raw

Re: [dpdk-dev] [PATCH v2 3/3] examples/ipsec-secgw: update pktest.sh script

Quoted 21 lines

-----Original Message-----
From: Iremonger, Bernard
Sent: Tuesday, October 1, 2019 4:18 PM
To: [email protected]; Ananyev, Konstantin <[email protected]>; [email protected]
Cc: Iremonger, Bernard <[email protected]>
Subject: [PATCH v2 3/3] examples/ipsec-secgw: update pktest.sh script

Drop SGW_CMD_XPRM=l, as default is IPsec library now.

---
 examples/ipsec-secgw/test/pkttest.sh | 1 -
 1 file changed, 1 deletion(-)
examples/ipsec-secgw/test/pkttest.sh +0 -0
index 04cd96b..7c78de7 100755

examples/ipsec-secgw/test/pkttest.sh fi

Quoted 8 lines
examples/ipsec-secgw/test/pkttest.sh +0 -2
@@ -18,7 +18,6 @@ fi
 CRYPTO_DEV="--vdev=crypto_null0"
 SGW_PORT_CFG="--vdev=net_tap0,mac=fixed --vdev=net_tap1,mac=fixed"
 SGW_EAL_XPRM="--no-pci"
-SGW_CMD_XPRM=-l
 SGW_WAIT_DEV="dtap0"
 . ${DIR}/common_defs_secgw.sh
 
--
Quoted 1 line

2.7.4