Skip to contentSkip to conversation

20.05-rc1 fixes for OVS

PATCH In review David Marchand 7 participants 20 messages Apr 27, 2020 – May 6, 2020

lore
Share & tools
lore
#0 lore raw

cover 20.05-rc1 fixes for OVS

Couple of fixes caught while checking OVS integration.

Extra bonus points for reviewers that will review the OVS series :-)
https://patchwork.ozlabs.org/project/openvswitch/list/?series=172917

Signature · 14 lines
David Marchand

David Marchand (3):
  ring: fix build with -Wswitch-enum
  eal: fix typo in endian conversion macros
  ethdev: fix build warning on 64-bit value

 lib/librte_eal/include/generic/rte_byteorder.h | 6 +++---
 lib/librte_ethdev/rte_flow.h                   | 2 +-
 lib/librte_ring/rte_ring_peek.h                | 8 ++++++++
 3 files changed, 12 insertions(+), 4 deletions(-)

-- 
2.23.0
#1 lore raw

PATCH 1/3 ring: fix build with -Wswitch-enum

Some popular vswitch implementation might use a gcc option that complains about missing enums in switch statements. Fix this by listing all possible values.

Fixes 664ff4b1729b ("ring: introduce peek style API")
---
 lib/librte_ring/rte_ring_peek.h | 8 ++++++++
 1 file changed, 8 insertions(+)
lib/librte_ring/rte_ring_peek.h +8 -0
index 9e7f4db323..d5e6ea1cf3 100644
@@ -68,6 +68,8 @@ __rte_ring_do_enqueue_start(struct rte_ring *r, uint32_t n,
 		n =  __rte_ring_hts_move_prod_head(r, n, behavior,
 			&head, &free);
 		break;
+	case RTE_RING_SYNC_MT:
+	case RTE_RING_SYNC_MT_RTS:
 	default:
 		/* unsupported mode, shouldn't be here */
 		RTE_ASSERT(0);
@@ -217,6 +219,8 @@ rte_ring_enqueue_elem_finish(struct rte_ring *r, const void *obj_table,
 			__rte_ring_enqueue_elems(r, tail, obj_table, esize, n);
 		__rte_ring_hts_set_head_tail(&r->hts_prod, tail, n, 1);
 		break;
+	case RTE_RING_SYNC_MT:
+	case RTE_RING_SYNC_MT_RTS:
 	default:
 		/* unsupported mode, shouldn't be here */
 		RTE_ASSERT(0);
@@ -263,6 +267,8 @@ __rte_ring_do_dequeue_start(struct rte_ring *r, void *obj_table,
 		n =  __rte_ring_hts_move_cons_head(r, n, behavior,
 			&head, &avail);
 		break;
+	case RTE_RING_SYNC_MT:
+	case RTE_RING_SYNC_MT_RTS:
 	default:
 		/* unsupported mode, shouldn't be here */
 		RTE_ASSERT(0);
@@ -414,6 +420,8 @@ rte_ring_dequeue_elem_finish(struct rte_ring *r, unsigned int n)
 		n = __rte_ring_hts_get_tail(&r->hts_cons, &tail, n);
 		__rte_ring_hts_set_head_tail(&r->hts_cons, tail, n, 0);
 		break;
+	case RTE_RING_SYNC_MT:
+	case RTE_RING_SYNC_MT_RTS:
 	default:
 		/* unsupported mode, shouldn't be here */
 		RTE_ASSERT(0);
Signature · 1 line
2.23.0
Acked-by on v1 1/3
#2 lore raw

Re: [dpdk-dev] [PATCH 1/3] ring: fix build with -Wswitch-enum

Quoted 44 lines

Some popular vswitch implementation might use a gcc option that complains about missing enums in switch statements. Fix this by listing all possible values.

Fixes 664ff4b1729b ("ring: introduce peek style API")
---
 lib/librte_ring/rte_ring_peek.h | 8 ++++++++
 1 file changed, 8 insertions(+)
lib/librte_ring/rte_ring_peek.h +6 -0
index 9e7f4db323..d5e6ea1cf3 100644
@@ -68,6 +68,8 @@ __rte_ring_do_enqueue_start(struct rte_ring *r, uint32_t n,
 		n =  __rte_ring_hts_move_prod_head(r, n, behavior,
 			&head, &free);
 		break;
+	case RTE_RING_SYNC_MT:
+	case RTE_RING_SYNC_MT_RTS:
 	default:
 		/* unsupported mode, shouldn't be here */
 		RTE_ASSERT(0);
@@ -217,6 +219,8 @@ rte_ring_enqueue_elem_finish(struct rte_ring *r, const void *obj_table,
 			__rte_ring_enqueue_elems(r, tail, obj_table, esize, n);
 		__rte_ring_hts_set_head_tail(&r->hts_prod, tail, n, 1);
 		break;
+	case RTE_RING_SYNC_MT:
+	case RTE_RING_SYNC_MT_RTS:
 	default:
 		/* unsupported mode, shouldn't be here */
 		RTE_ASSERT(0);
@@ -263,6 +267,8 @@ __rte_ring_do_dequeue_start(struct rte_ring *r, void *obj_table,
 		n =  __rte_ring_hts_move_cons_head(r, n, behavior,
 			&head, &avail);
 		break;
+	case RTE_RING_SYNC_MT:
+	case RTE_RING_SYNC_MT_RTS:
 	default:
 		/* unsupported mode, shouldn't be here */
 		RTE_ASSERT(0);

lib/librte_ring/rte_ring_peek.h rte_ring_dequeue_elem_finish(struct rte_ring *r, unsigned int n)

Quoted 9 lines
lib/librte_ring/rte_ring_peek.h +2 -1
@@ -414,6 +420,8 @@ rte_ring_dequeue_elem_finish(struct rte_ring *r, unsigned int n)
 		n = __rte_ring_hts_get_tail(&r->hts_cons, &tail, n);
 		__rte_ring_hts_set_head_tail(&r->hts_cons, tail, n, 0);
 		break;
+	case RTE_RING_SYNC_MT:
+	case RTE_RING_SYNC_MT_RTS:
 	default:
 		/* unsupported mode, shouldn't be here */
 		RTE_ASSERT(0);
--
Quoted 1 line

2.23.0

#3 lore raw

PATCH 2/3 eal: fix typo in endian conversion macros

In reply to David Marchand

Caught by code inspection, for little endian, RTE_LEXX macros should provide rte_leXX_t type values.

Fixes b75667ef9f7e ("eal: add static endianness conversion macros")
---
 lib/librte_eal/include/generic/rte_byteorder.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
lib/librte_eal/include/generic/rte_byteorder.h +3 -3
index 38e8cfd32b..9ca960932f 100644
@@ -93,9 +93,9 @@
 #define RTE_BE16(v) (rte_be16_t)(RTE_STATIC_BSWAP16(v))
 #define RTE_BE32(v) (rte_be32_t)(RTE_STATIC_BSWAP32(v))
 #define RTE_BE64(v) (rte_be64_t)(RTE_STATIC_BSWAP64(v))
-#define RTE_LE16(v) (rte_be16_t)(v)
-#define RTE_LE32(v) (rte_be32_t)(v)
-#define RTE_LE64(v) (rte_be64_t)(v)
+#define RTE_LE16(v) (rte_le16_t)(v)
+#define RTE_LE32(v) (rte_le32_t)(v)
+#define RTE_LE64(v) (rte_le64_t)(v)
 #else
 #error Unsupported endianness.
 #endif
Signature · 1 line
2.23.0
#5 lore raw

PATCH 3/3 ethdev: fix build warning on 64-bit value

In reply to David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/librte_ethdev/rte_flow.h +1 -1
index 132b44edc6..1fb94f35e8 100644
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
 	.s_field = 0x01,
-	.seid = RTE_BE64(0xffffffffffffffff),
+	.seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),
 };
 #endif
 
Signature · 1 line
2.23.0
Reviewed-by on v1 3/3
#6 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 17 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/librte_ethdev/rte_flow.h +0 -0
index 132b44edc6..1fb94f35e8 100644

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1540

Quoted 8 lines
lib/librte_ethdev/rte_flow.h +1 -1
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
 	.s_field = 0x01,
-	.seid = RTE_BE64(0xffffffffffffffff),
+	.seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),
 };
 #endif
 
#7 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

In reply to David Marchand
Quoted 16 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/librte_ethdev/rte_flow.h +0 -0
index 132b44edc6..1fb94f35e8 100644

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 5 lines
lib/librte_ethdev/rte_flow.h +1 -1
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
 	.s_field = 0x01,
-	.seid = RTE_BE64(0xffffffffffffffff),
+	.seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end. If we are going to cast, why not just put "-1" in to save some digits.

/Bruce

#8 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 13 lines · Bruce Richardson
Quoted 11 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 13 lines

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 10 lines
lib/librte_ethdev/rte_flow.h +1 -1
index 132b44edc6..1fb94f35e8 100644
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
      .s_field = 0x01,
-     .seid = RTE_BE64(0xffffffffffffffff),
+     .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end. If we are going to cast, why not just put "-1" in to save some digits.

I preferred this form in the hope future developers who want 0x0fffffffffffffff will copy/paste this.

Signature · 1 line
David Marchand
#9 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 39 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of David Marchand
Sent: Monday, April 27, 2020 2:37 PM
To: Richardson, Bruce <[email protected]>
Cc: dev <[email protected]>; Ori Kam <[email protected]>; Thomas Monjalon <[email protected]>; Yigit, Ferruh <[email protected]>; Andrew Rybchenko <[email protected]>; Zhang, Xiao <[email protected]>
Subject: Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 13 lines · Bruce Richardson
Quoted 11 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 13 lines

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 10 lines
lib/librte_ethdev/rte_flow.h +1 -1
index 132b44edc6..1fb94f35e8 100644
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
      .s_field = 0x01,
-     .seid = RTE_BE64(0xffffffffffffffff),
+     .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end. If we are going to cast, why not just put "-1" in to save some digits.

Quoted 3 lines

I preferred this form in the hope future developers who want 0x0fffffffffffffff will copy/paste this.

As I remember there should be UINT64_MAX in stdint.h.

#10 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 40 lines · Konstantin
Quoted 40 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of David Marchand
Sent: Monday, April 27, 2020 2:37 PM
To: Richardson, Bruce <[email protected]>
Cc: dev <[email protected]>; Ori Kam <[email protected]>; Thomas Monjalon <[email protected]>; Yigit, Ferruh <[email protected]>; Andrew Rybchenko <[email protected]>; Zhang, Xiao <[email protected]>
Subject: Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 26 lines · Bruce Richardson
Quoted 16 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/librte_ethdev/rte_flow.h +0 -0
index 132b44edc6..1fb94f35e8 100644

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 5 lines
lib/librte_ethdev/rte_flow.h +1 -1
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
      .s_field = 0x01,
-     .seid = RTE_BE64(0xffffffffffffffff),
+     .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end. If we are going to cast, why not just put "-1" in to save some digits.

Quoted 5 lines
Quoted 3 lines

I preferred this form in the hope future developers who want 0x0fffffffffffffff will copy/paste this.

As I remember there should be UINT64_MAX in stdint.h.

Yes, we could go with:
+ .seid = RTE_BE64(UINT64_MAX),

And then next time, for any value like 0x0fff ffff ffff ffff (had to group the digits of what I had written), pretty sure we will miss this and I will catch it only when building ovs.

--
David Marchand

#11 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 42 lines · David Marchand
Quoted 40 lines · Konstantin
Quoted 40 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of David Marchand
Sent: Monday, April 27, 2020 2:37 PM
To: Richardson, Bruce <[email protected]>
Cc: dev <[email protected]>; Ori Kam <[email protected]>; Thomas Monjalon <[email protected]>; Yigit, Ferruh <[email protected]>; Andrew Rybchenko <[email protected]>; Zhang, Xiao <[email protected]>
Subject: Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 26 lines · Bruce Richardson
Quoted 16 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/librte_ethdev/rte_flow.h +0 -0
index 132b44edc6..1fb94f35e8 100644

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 5 lines
lib/librte_ethdev/rte_flow.h +1 -1
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
      .s_field = 0x01,
-     .seid = RTE_BE64(0xffffffffffffffff),
+     .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end. If we are going to cast, why not just put "-1" in to save some digits.

Quoted 8 lines
Quoted 5 lines
Quoted 3 lines

I preferred this form in the hope future developers who want 0x0fffffffffffffff will copy/paste this.

As I remember there should be UINT64_MAX in stdint.h.

Yes, we could go with:
+ .seid = RTE_BE64(UINT64_MAX),

This is something else but if the value is 'UINT64_MAX', do we need 'RTE_BE64' macro?

Quoted 9 lines

And then next time, for any value like 0x0fff ffff ffff ffff (had to group the digits of what I had written), pretty sure we will miss this and I will catch it only when building ovs.

--
David Marchand

#12 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 48 lines · Ferruh Yigit
Quoted 46 lines · David Marchand
Quoted 44 lines · Konstantin
Quoted 43 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of David Marchand
Sent: Monday, April 27, 2020 2:37 PM
To: Richardson, Bruce <[email protected]>
Cc: dev <[email protected]>; Ori Kam <[email protected]>; Thomas Monjalon <[email protected]>; Yigit, Ferruh <[email protected]>; Andrew Rybchenko <[email protected]>; Zhang, Xiao <[email protected]>
Subject: Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 13 lines · Bruce Richardson
Quoted 11 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 13 lines

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 10 lines
lib/librte_ethdev/rte_flow.h +1 -1
index 132b44edc6..1fb94f35e8 100644
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
      .s_field = 0x01,
-     .seid = RTE_BE64(0xffffffffffffffff),
+     .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end. If we are going to cast, why not just put "-1" in to save some digits.

I preferred this form in the hope future developers who want 0x0fffffffffffffff will copy/paste this.

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

As I remember there should be UINT64_MAX in stdint.h.

Yes, we could go with:
+ .seid = RTE_BE64(UINT64_MAX),

This is something else but if the value is 'UINT64_MAX', do we need 'RTE_BE64' macro?

In OVS case, sparse validates that a rte_be64_t value (mapped to ovs_be64_t) is passed to .seid.

.../build/install/usr/local/include/rte_flow.h:1537:17: error:
incorrect type in initializer (different base types)
.../build/install/usr/local/include/rte_flow.h:1537:17: expected
restricted ovs_be64 [usertype] seid
.../build/install/usr/local/include/rte_flow.h:1537:17: got unsigned long

Signature · 1 line
David Marchand
#13 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 57 lines · David Marchand
Quoted 48 lines · Ferruh Yigit
Quoted 46 lines · David Marchand
Quoted 44 lines · Konstantin
Quoted 43 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of David Marchand
Sent: Monday, April 27, 2020 2:37 PM
To: Richardson, Bruce <[email protected]>
Cc: dev <[email protected]>; Ori Kam <[email protected]>; Thomas Monjalon <[email protected]>; Yigit, Ferruh <[email protected]>; Andrew Rybchenko <[email protected]>; Zhang, Xiao <[email protected]>
Subject: Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 13 lines · Bruce Richardson
Quoted 11 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 13 lines

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 10 lines
lib/librte_ethdev/rte_flow.h +1 -1
index 132b44edc6..1fb94f35e8 100644
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
      .s_field = 0x01,
-     .seid = RTE_BE64(0xffffffffffffffff),
+     .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end. If we are going to cast, why not just put "-1" in to save some digits.

I preferred this form in the hope future developers who want 0x0fffffffffffffff will copy/paste this.

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

As I remember there should be UINT64_MAX in stdint.h.

Yes, we could go with:
+ .seid = RTE_BE64(UINT64_MAX),

This is something else but if the value is 'UINT64_MAX', do we need 'RTE_BE64' macro?

Quoted 10 lines

In OVS case, sparse validates that a rte_be64_t value (mapped to ovs_be64_t) is passed to .seid.

.../build/install/usr/local/include/rte_flow.h:1537:17: error:
incorrect type in initializer (different base types)
.../build/install/usr/local/include/rte_flow.h:1537:17: expected
restricted ovs_be64 [usertype] seid
.../build/install/usr/local/include/rte_flow.h:1537:17: got unsigned long

Got it, thanks for clarification.

#14 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

In reply to David Marchand

27/04/2020 16:00, David Marchand:

Quoted 27 lines · Konstantin
Quoted 26 lines

From: dev <[email protected]> On Behalf Of David Marchand

Quoted 24 lines
Quoted 7 lines · Bruce Richardson
Quoted 6 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
Quoted 11 lines

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 8 lines

---

lib/librte_ethdev/rte_flow.h +1 -1
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
      .s_field = 0x01,
-     .seid = RTE_BE64(0xffffffffffffffff),
+     .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end. If we are going to cast, why not just put "-1" in to save some digits.

I preferred this form in the hope future developers who want 0x0fffffffffffffff will copy/paste this.

Quoted 8 lines
Quoted 1 line

As I remember there should be UINT64_MAX in stdint.h.

Yes, we could go with:
+ .seid = RTE_BE64(UINT64_MAX),

And then next time, for any value like 0x0fff ffff ffff ffff (had to group the digits of what I had written), pretty sure we will miss this and I will catch it only when building ovs.

I agree with David (in general and especially here).

RTE_BE64 is required for static analyzers and is an explicit info.

UINT64_C is better than ULL suffix because it
- is generic
- gives size explicitly

UINT64_C(0xffffffffffffffff) is better than UINT64_MAX because
- rte_flow.h has a lot of bitmasks
- it is copy/paste safe

#15 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 45 lines

From: dev [mailto:[email protected]] On Behalf Of Thomas Monjalon
Sent: Tuesday, April 28, 2020 11:28 AM

27/04/2020 16:00, David Marchand:

Quoted 6 lines · Konstantin
Quoted 5 lines

From: dev <[email protected]> On Behalf Of David Marchand

Quoted 4 lines
Quoted 2 lines · Bruce Richardson
Quoted 1 line · David Marchand

Building OVS with dpdk, sparse complains about 64-bit

constant being

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned

long

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

lib/librte_ethdev/rte_flow.h

Quoted 8 lines
Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")

---

lib/librte_ethdev/rte_flow.h +0 -0
 static const struct rte_flow_item_pfcp

rte_flow_item_pfcp_mask = {

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 3 lines
      .s_field = 0x01,
lib/librte_ethdev/rte_flow.h +1 -1
-     .seid = RTE_BE64(0xffffffffffffffff),
+     .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end. If we are going

to cast,

Quoted 6 lines
Quoted 6 lines
Quoted 5 lines
Quoted 1 line

why not just put "-1" in to save some digits.

I preferred this form in the hope future developers who want 0x0fffffffffffffff will copy/paste this.

Quoted 7 lines
Quoted 1 line

As I remember there should be UINT64_MAX in stdint.h.

Yes, we could go with:
+ .seid = RTE_BE64(UINT64_MAX),

And then next time, for any value like 0x0fff ffff ffff ffff (had to group the digits of what I had written), pretty sure we will miss

this

Quoted 1 line

and I will catch it only when building ovs.

Quoted 7 lines

I agree with David (in general and especially here).

RTE_BE64 is required for static analyzers and is an explicit info.

UINT64_C is better than ULL suffix because it
- is generic
- gives size explicitly

Certainly. Explicit is preferred in code for embedded systems. "unsigned long long" means 64 bit or more, which also applies to the "ULL" postfix. "uint64_t" and "UINT64_C" means exactly 64 bit.

Quoted 7 lines

UINT64_C(0xffffffffffffffff) is better than UINT64_MAX because
- rte_flow.h has a lot of bitmasks
- it is copy/paste safe

And shouldn't the struct rte_flow_item_pfcp be packed? I would expect the compiler to add 32 bit padding before seid to ensure its 64 bit alignment.

Med venlig hilsen / kind regards
- Morten Brørup

#16 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

In reply to Bruce Richardson
Quoted 12 lines · Bruce Richardson
Quoted 11 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 11 lines

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 10 lines
lib/librte_ethdev/rte_flow.h +1 -1
index 132b44edc6..1fb94f35e8 100644
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
 	.s_field = 0x01,
-	.seid = RTE_BE64(0xffffffffffffffff),
+	.seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end.

It is not a cast as far as I can see, it is exactly ULL (or UL):
/usr/include/stdint.h

# if __WORDSIZE == 64
#  define UINT64_C(c)   c ## UL
# else
#  define UINT64_C(c)   c ## ULL
# endif
Quoted 3 lines

If we are going to cast, why not just put "-1" in to save some digits.

/Bruce

#17 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 26 lines · Andrew Rybchenko
Quoted 12 lines · Bruce Richardson
Quoted 11 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 11 lines

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 10 lines
lib/librte_ethdev/rte_flow.h +1 -1
index 132b44edc6..1fb94f35e8 100644
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
     .s_field = 0x01,
-    .seid = RTE_BE64(0xffffffffffffffff),
+    .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end.

Quoted 7 lines

It is not a cast as far as I can see, it is exactly ULL (or UL):
/usr/include/stdint.h

# if __WORDSIZE == 64
#  define UINT64_C(c)   c ## UL
# else
#  define UINT64_C(c)   c ## ULL
# endif

Yes.

Signature · 1 line
David Marchand
#18 lore raw

Re: [dpdk-dev] [PATCH 3/3] ethdev: fix build warning on 64-bit value

Quoted 36 lines · David Marchand
Quoted 26 lines · Andrew Rybchenko
Quoted 12 lines · Bruce Richardson
Quoted 11 lines · David Marchand

Building OVS with dpdk, sparse complains about 64-bit constant being passed as a normal integer that can't fit it:
error: constant 0xffffffffffffffff is so big it is unsigned long

Fixes ecbc8570131d ("ethdev: add PFCP header to flow API")
---
 lib/librte_ethdev/rte_flow.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 11 lines

lib/librte_ethdev/rte_flow.h struct rte_flow_item_pfcp { · line 1537

Quoted 10 lines
lib/librte_ethdev/rte_flow.h +1 -1
index 132b44edc6..1fb94f35e8 100644
@@ -1534,7 +1534,7 @@ struct rte_flow_item_pfcp {
 #ifndef __cplusplus
 static const struct rte_flow_item_pfcp rte_flow_item_pfcp_mask = {
     .s_field = 0x01,
-    .seid = RTE_BE64(0xffffffffffffffff),
+    .seid = RTE_BE64(UINT64_C(0xffffffffffffffff)),

Rather than cast, why not put "ULL" at the end.

Quoted 7 lines

It is not a cast as far as I can see, it is exactly ULL (or UL):
/usr/include/stdint.h

# if __WORDSIZE == 64
#  define UINT64_C(c)   c ## UL
# else
#  define UINT64_C(c)   c ## ULL
# endif
Quoted 2 lines

Yes.

Thanks. I'd prefer just sticking on the "ULL" myself as it's shorter, but this is ok.

#19 lore raw
In reply to David Marchand
Quoted 8 lines · David Marchand

Couple of fixes caught while checking OVS integration.

David Marchand (3):
ring: fix build with -Wswitch-enum
eal: fix typo in endian conversion macros
ethdev: fix build warning on 64-bit value

Quoted 5 lines
 lib/librte_eal/include/generic/rte_byteorder.h | 6 +++---
 lib/librte_ethdev/rte_flow.h                   | 2 +-
 lib/librte_ring/rte_ring_peek.h                | 8 ++++++++
 3 files changed, 12 insertions(+), 4 deletions(-)

Series applied.

Signature · 1 line
David Marchand