* [dpdk-stable] [PATCH] pipeline: fix endianness conversions
@ 2021-04-03 0:17 Cristian Dumitrescu
2021-04-20 0:34 ` [dpdk-stable] [dpdk-dev] " Thomas Monjalon
0 siblings, 1 reply; 2+ messages in thread
From: Cristian Dumitrescu @ 2021-04-03 0:17 UTC (permalink / raw)
To: dev; +Cc: stable
The SWX pipeline instructions work with operands of different types:
header fields (h.header.field), packet meta-data (m.field), extern
object mailbox field (e.obj.field), extern function (f.field), action
data read from table entries (t.field), or immediate values; hence the
HMEFTI acronym. The H operands are stored in network byte order (NBO),
while the MEFT operands are stored in host byte order (HBO), hence the
need to operate endianness conversions.
Some of the endianness conversion macros were not working correctly
for some cases such as operands of different sizes, and they are fixed
now. Affected instructions: mov, and, or, xor, jmpeq, jmpneq.
Fixes: 7210349d5baa ("pipeline: add SWX move instruction")
Fixes: 650195cf965a ("pipeline: introduce SWX and instruction")
Fixes: 8f796198dcda ("pipeline: introduce SWX or instruction")
Fixes: b4e607f9fd5e ("pipeline: introduce SWX XOR instruction")
Fixes: b3947e25bed4 ("pipeline: introduce SWX jump and return instructions")
Cc: stable@dpdk.org
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
---
lib/librte_pipeline/rte_swx_pipeline.c | 470 ++++++++++++++++++++-----
1 file changed, 382 insertions(+), 88 deletions(-)
diff --git a/lib/librte_pipeline/rte_swx_pipeline.c b/lib/librte_pipeline/rte_swx_pipeline.c
index f46f90c34..dc2a155ed 100644
--- a/lib/librte_pipeline/rte_swx_pipeline.c
+++ b/lib/librte_pipeline/rte_swx_pipeline.c
@@ -318,9 +318,11 @@ enum instruction_type {
* dst = src
* dst = HMEF, src = HMEFTI
*/
- INSTR_MOV, /* dst = MEF, src = MEFT */
- INSTR_MOV_S, /* (dst, src) = (MEF, H) or (dst, src) = (H, MEFT) */
- INSTR_MOV_I, /* dst = HMEF, src = I */
+ INSTR_MOV, /* dst = MEF, src = MEFT */
+ INSTR_MOV_MH, /* dst = MEF, src = H */
+ INSTR_MOV_HM, /* dst = H, src = MEFT */
+ INSTR_MOV_HH, /* dst = H, src = H */
+ INSTR_MOV_I, /* dst = HMEF, src = I */
/* dma h.header t.field
* memcpy(h.header, t.field, sizeof(h.header))
@@ -374,25 +376,31 @@ enum instruction_type {
* dst &= src
* dst = HMEF, src = HMEFTI
*/
- INSTR_ALU_AND, /* dst = MEF, src = MEFT */
- INSTR_ALU_AND_S, /* (dst, src) = (MEF, H) or (dst, src) = (H, MEFT) */
- INSTR_ALU_AND_I, /* dst = HMEF, src = I */
+ INSTR_ALU_AND, /* dst = MEF, src = MEFT */
+ INSTR_ALU_AND_MH, /* dst = MEF, src = H */
+ INSTR_ALU_AND_HM, /* dst = H, src = MEFT */
+ INSTR_ALU_AND_HH, /* dst = H, src = H */
+ INSTR_ALU_AND_I, /* dst = HMEF, src = I */
/* or dst src
* dst |= src
* dst = HMEF, src = HMEFTI
*/
- INSTR_ALU_OR, /* dst = MEF, src = MEFT */
- INSTR_ALU_OR_S, /* (dst, src) = (MEF, H) or (dst, src) = (H, MEFT) */
- INSTR_ALU_OR_I, /* dst = HMEF, src = I */
+ INSTR_ALU_OR, /* dst = MEF, src = MEFT */
+ INSTR_ALU_OR_MH, /* dst = MEF, src = H */
+ INSTR_ALU_OR_HM, /* dst = H, src = MEFT */
+ INSTR_ALU_OR_HH, /* dst = H, src = H */
+ INSTR_ALU_OR_I, /* dst = HMEF, src = I */
/* xor dst src
* dst ^= src
* dst = HMEF, src = HMEFTI
*/
- INSTR_ALU_XOR, /* dst = MEF, src = MEFT */
- INSTR_ALU_XOR_S, /* (dst, src) = (MEF, H) or (dst, src) = (H, MEFT) */
- INSTR_ALU_XOR_I, /* dst = HMEF, src = I */
+ INSTR_ALU_XOR, /* dst = MEF, src = MEFT */
+ INSTR_ALU_XOR_MH, /* dst = MEF, src = H */
+ INSTR_ALU_XOR_HM, /* dst = H, src = MEFT */
+ INSTR_ALU_XOR_HH, /* dst = H, src = H */
+ INSTR_ALU_XOR_I, /* dst = HMEF, src = I */
/* shl dst src
* dst <<= src
@@ -533,41 +541,45 @@ enum instruction_type {
INSTR_JMP_ACTION_MISS,
/* jmpeq LABEL a b
- * Jump is a is equal to b
+ * Jump if a is equal to b
* a = HMEFT, b = HMEFTI
*/
- INSTR_JMP_EQ, /* (a, b) = (MEFT, MEFT) or (a, b) = (H, H) */
- INSTR_JMP_EQ_S, /* (a, b) = (MEFT, H) or (a, b) = (H, MEFT) */
- INSTR_JMP_EQ_I, /* (a, b) = (MEFT, I) or (a, b) = (H, I) */
+ INSTR_JMP_EQ, /* a = MEFT, b = MEFT */
+ INSTR_JMP_EQ_MH, /* a = MEFT, b = H */
+ INSTR_JMP_EQ_HM, /* a = H, b = MEFT */
+ INSTR_JMP_EQ_HH, /* a = H, b = H */
+ INSTR_JMP_EQ_I, /* (a, b) = (MEFT, I) or (a, b) = (H, I) */
/* jmpneq LABEL a b
- * Jump is a is not equal to b
+ * Jump if a is not equal to b
* a = HMEFT, b = HMEFTI
*/
- INSTR_JMP_NEQ, /* (a, b) = (MEFT, MEFT) or (a, b) = (H, H) */
- INSTR_JMP_NEQ_S, /* (a, b) = (MEFT, H) or (a, b) = (H, MEFT) */
- INSTR_JMP_NEQ_I, /* (a, b) = (MEFT, I) or (a, b) = (H, I) */
+ INSTR_JMP_NEQ, /* a = MEFT, b = MEFT */
+ INSTR_JMP_NEQ_MH, /* a = MEFT, b = H */
+ INSTR_JMP_NEQ_HM, /* a = H, b = MEFT */
+ INSTR_JMP_NEQ_HH, /* a = H, b = H */
+ INSTR_JMP_NEQ_I, /* (a, b) = (MEFT, I) or (a, b) = (H, I) */
/* jmplt LABEL a b
* Jump if a is less than b
* a = HMEFT, b = HMEFTI
*/
- INSTR_JMP_LT, /* a = MEF, b = MEF */
- INSTR_JMP_LT_MH, /* a = MEF, b = H */
- INSTR_JMP_LT_HM, /* a = H, b = MEF */
+ INSTR_JMP_LT, /* a = MEFT, b = MEFT */
+ INSTR_JMP_LT_MH, /* a = MEFT, b = H */
+ INSTR_JMP_LT_HM, /* a = H, b = MEFT */
INSTR_JMP_LT_HH, /* a = H, b = H */
- INSTR_JMP_LT_MI, /* a = MEF, b = I */
+ INSTR_JMP_LT_MI, /* a = MEFT, b = I */
INSTR_JMP_LT_HI, /* a = H, b = I */
/* jmpgt LABEL a b
* Jump if a is greater than b
* a = HMEFT, b = HMEFTI
*/
- INSTR_JMP_GT, /* a = MEF, b = MEF */
- INSTR_JMP_GT_MH, /* a = MEF, b = H */
- INSTR_JMP_GT_HM, /* a = H, b = MEF */
+ INSTR_JMP_GT, /* a = MEFT, b = MEFT */
+ INSTR_JMP_GT_MH, /* a = MEFT, b = H */
+ INSTR_JMP_GT_HM, /* a = H, b = MEFT */
INSTR_JMP_GT_HH, /* a = H, b = H */
- INSTR_JMP_GT_MI, /* a = MEF, b = I */
+ INSTR_JMP_GT_MI, /* a = MEFT, b = I */
INSTR_JMP_GT_HI, /* a = H, b = I */
/* return
@@ -897,7 +909,7 @@ struct thread {
#if RTE_BYTE_ORDER == RTE_LITTLE_ENDIAN
-#define ALU_S(thread, ip, operator) \
+#define ALU_MH(thread, ip, operator) \
{ \
uint8_t *dst_struct = (thread)->structs[(ip)->alu.dst.struct_id]; \
uint64_t *dst64_ptr = (uint64_t *)&dst_struct[(ip)->alu.dst.offset]; \
@@ -915,8 +927,6 @@ struct thread {
*dst64_ptr = (dst64 & ~dst64_mask) | (result & dst64_mask); \
}
-#define ALU_MH ALU_S
-
#define ALU_HM(thread, ip, operator) \
{ \
uint8_t *dst_struct = (thread)->structs[(ip)->alu.dst.struct_id]; \
@@ -937,6 +947,25 @@ struct thread {
*dst64_ptr = (dst64 & ~dst64_mask) | result; \
}
+#define ALU_HM_FAST(thread, ip, operator) \
+{ \
+ uint8_t *dst_struct = (thread)->structs[(ip)->alu.dst.struct_id]; \
+ uint64_t *dst64_ptr = (uint64_t *)&dst_struct[(ip)->alu.dst.offset]; \
+ uint64_t dst64 = *dst64_ptr; \
+ uint64_t dst64_mask = UINT64_MAX >> (64 - (ip)->alu.dst.n_bits); \
+ uint64_t dst = dst64 & dst64_mask; \
+ \
+ uint8_t *src_struct = (thread)->structs[(ip)->alu.src.struct_id]; \
+ uint64_t *src64_ptr = (uint64_t *)&src_struct[(ip)->alu.src.offset]; \
+ uint64_t src64 = *src64_ptr; \
+ uint64_t src64_mask = UINT64_MAX >> (64 - (ip)->alu.src.n_bits); \
+ uint64_t src = hton64(src64 & src64_mask) >> (64 - (ip)->alu.dst.n_bits); \
+ \
+ uint64_t result = dst operator src; \
+ \
+ *dst64_ptr = (dst64 & ~dst64_mask) | result; \
+}
+
#define ALU_HH(thread, ip, operator) \
{ \
uint8_t *dst_struct = (thread)->structs[(ip)->alu.dst.struct_id]; \
@@ -956,12 +985,31 @@ struct thread {
*dst64_ptr = (dst64 & ~dst64_mask) | result; \
}
+#define ALU_HH_FAST(thread, ip, operator) \
+{ \
+ uint8_t *dst_struct = (thread)->structs[(ip)->alu.dst.struct_id]; \
+ uint64_t *dst64_ptr = (uint64_t *)&dst_struct[(ip)->alu.dst.offset]; \
+ uint64_t dst64 = *dst64_ptr; \
+ uint64_t dst64_mask = UINT64_MAX >> (64 - (ip)->alu.dst.n_bits); \
+ uint64_t dst = dst64 & dst64_mask; \
+ \
+ uint8_t *src_struct = (thread)->structs[(ip)->alu.src.struct_id]; \
+ uint64_t *src64_ptr = (uint64_t *)&src_struct[(ip)->alu.src.offset]; \
+ uint64_t src64 = *src64_ptr; \
+ uint64_t src = (src64 << (64 - (ip)->alu.src.n_bits)) >> (64 - (ip)->alu.dst.n_bits); \
+ \
+ uint64_t result = dst operator src; \
+ \
+ *dst64_ptr = (dst64 & ~dst64_mask) | result; \
+}
+
#else
-#define ALU_S ALU
#define ALU_MH ALU
#define ALU_HM ALU
+#define ALU_HM_FAST ALU
#define ALU_HH ALU
+#define ALU_HH_FAST ALU
#endif
@@ -1024,7 +1072,7 @@ struct thread {
#if RTE_BYTE_ORDER == RTE_LITTLE_ENDIAN
-#define MOV_S(thread, ip) \
+#define MOV_MH(thread, ip) \
{ \
uint8_t *dst_struct = (thread)->structs[(ip)->mov.dst.struct_id]; \
uint64_t *dst64_ptr = (uint64_t *)&dst_struct[(ip)->mov.dst.offset]; \
@@ -1039,9 +1087,44 @@ struct thread {
*dst64_ptr = (dst64 & ~dst64_mask) | (src & dst64_mask); \
}
+#define MOV_HM(thread, ip) \
+{ \
+ uint8_t *dst_struct = (thread)->structs[(ip)->mov.dst.struct_id]; \
+ uint64_t *dst64_ptr = (uint64_t *)&dst_struct[(ip)->mov.dst.offset]; \
+ uint64_t dst64 = *dst64_ptr; \
+ uint64_t dst64_mask = UINT64_MAX >> (64 - (ip)->mov.dst.n_bits); \
+ \
+ uint8_t *src_struct = (thread)->structs[(ip)->mov.src.struct_id]; \
+ uint64_t *src64_ptr = (uint64_t *)&src_struct[(ip)->mov.src.offset]; \
+ uint64_t src64 = *src64_ptr; \
+ uint64_t src64_mask = UINT64_MAX >> (64 - (ip)->mov.src.n_bits); \
+ uint64_t src = src64 & src64_mask; \
+ \
+ src = hton64(src) >> (64 - (ip)->mov.dst.n_bits); \
+ *dst64_ptr = (dst64 & ~dst64_mask) | src; \
+}
+
+#define MOV_HH(thread, ip) \
+{ \
+ uint8_t *dst_struct = (thread)->structs[(ip)->mov.dst.struct_id]; \
+ uint64_t *dst64_ptr = (uint64_t *)&dst_struct[(ip)->mov.dst.offset]; \
+ uint64_t dst64 = *dst64_ptr; \
+ uint64_t dst64_mask = UINT64_MAX >> (64 - (ip)->mov.dst.n_bits); \
+ \
+ uint8_t *src_struct = (thread)->structs[(ip)->mov.src.struct_id]; \
+ uint64_t *src64_ptr = (uint64_t *)&src_struct[(ip)->mov.src.offset]; \
+ uint64_t src64 = *src64_ptr; \
+ \
+ uint64_t src = src64 << (64 - (ip)->mov.src.n_bits); \
+ src = src >> (64 - (ip)->mov.dst.n_bits); \
+ *dst64_ptr = (dst64 & ~dst64_mask) | src; \
+}
+
#else
-#define MOV_S MOV
+#define MOV_MH MOV
+#define MOV_HM MOV
+#define MOV_HH MOV
#endif
@@ -1076,7 +1159,7 @@ struct thread {
#if RTE_BYTE_ORDER == RTE_LITTLE_ENDIAN
-#define JMP_CMP_S(thread, ip, operator) \
+#define JMP_CMP_MH(thread, ip, operator) \
{ \
uint8_t *a_struct = (thread)->structs[(ip)->jmp.a.struct_id]; \
uint64_t *a64_ptr = (uint64_t *)&a_struct[(ip)->jmp.a.offset]; \
@@ -1092,8 +1175,6 @@ struct thread {
(thread)->ip = (a operator b) ? (ip)->jmp.ip : ((thread)->ip + 1); \
}
-#define JMP_CMP_MH JMP_CMP_S
-
#define JMP_CMP_HM(thread, ip, operator) \
{ \
uint8_t *a_struct = (thread)->structs[(ip)->jmp.a.struct_id]; \
@@ -1125,12 +1206,27 @@ struct thread {
(thread)->ip = (a operator b) ? (ip)->jmp.ip : ((thread)->ip + 1); \
}
+#define JMP_CMP_HH_FAST(thread, ip, operator) \
+{ \
+ uint8_t *a_struct = (thread)->structs[(ip)->jmp.a.struct_id]; \
+ uint64_t *a64_ptr = (uint64_t *)&a_struct[(ip)->jmp.a.offset]; \
+ uint64_t a64 = *a64_ptr; \
+ uint64_t a = a64 << (64 - (ip)->jmp.a.n_bits); \
+ \
+ uint8_t *b_struct = (thread)->structs[(ip)->jmp.b.struct_id]; \
+ uint64_t *b64_ptr = (uint64_t *)&b_struct[(ip)->jmp.b.offset]; \
+ uint64_t b64 = *b64_ptr; \
+ uint64_t b = b64 << (64 - (ip)->jmp.b.n_bits); \
+ \
+ (thread)->ip = (a operator b) ? (ip)->jmp.ip : ((thread)->ip + 1); \
+}
+
#else
-#define JMP_CMP_S JMP_CMP
#define JMP_CMP_MH JMP_CMP
#define JMP_CMP_HM JMP_CMP
#define JMP_CMP_HH JMP_CMP
+#define JMP_CMP_HH_FAST JMP_CMP
#endif
@@ -2524,10 +2620,14 @@ instruction_is_jmp(struct instruction *instr)
case INSTR_JMP_ACTION_HIT:
case INSTR_JMP_ACTION_MISS:
case INSTR_JMP_EQ:
- case INSTR_JMP_EQ_S:
+ case INSTR_JMP_EQ_MH:
+ case INSTR_JMP_EQ_HM:
+ case INSTR_JMP_EQ_HH:
case INSTR_JMP_EQ_I:
case INSTR_JMP_NEQ:
- case INSTR_JMP_NEQ_S:
+ case INSTR_JMP_NEQ_MH:
+ case INSTR_JMP_NEQ_HM:
+ case INSTR_JMP_NEQ_HH:
case INSTR_JMP_NEQ_I:
case INSTR_JMP_LT:
case INSTR_JMP_LT_MH:
@@ -3508,13 +3608,16 @@ instr_mov_translate(struct rte_swx_pipeline *p,
fdst = struct_field_parse(p, NULL, dst, &dst_struct_id);
CHECK(fdst, EINVAL);
- /* MOV or MOV_S. */
+ /* MOV, MOV_MH, MOV_HM or MOV_HH. */
fsrc = struct_field_parse(p, action, src, &src_struct_id);
if (fsrc) {
instr->type = INSTR_MOV;
- if ((dst[0] == 'h' && src[0] != 'h') ||
- (dst[0] != 'h' && src[0] == 'h'))
- instr->type = INSTR_MOV_S;
+ if (dst[0] != 'h' && src[0] == 'h')
+ instr->type = INSTR_MOV_MH;
+ if (dst[0] == 'h' && src[0] != 'h')
+ instr->type = INSTR_MOV_HM;
+ if (dst[0] == 'h' && src[0] == 'h')
+ instr->type = INSTR_MOV_HH;
instr->mov.dst.struct_id = (uint8_t)dst_struct_id;
instr->mov.dst.n_bits = fdst->n_bits;
@@ -3556,15 +3659,45 @@ instr_mov_exec(struct rte_swx_pipeline *p)
}
static inline void
-instr_mov_s_exec(struct rte_swx_pipeline *p)
+instr_mov_mh_exec(struct rte_swx_pipeline *p)
{
struct thread *t = &p->threads[p->thread_id];
struct instruction *ip = t->ip;
- TRACE("[Thread %2u] mov (s)\n",
+ TRACE("[Thread %2u] mov (mh)\n",
p->thread_id);
- MOV_S(t, ip);
+ MOV_MH(t, ip);
+
+ /* Thread. */
+ thread_ip_inc(p);
+}
+
+static inline void
+instr_mov_hm_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] mov (hm)\n",
+ p->thread_id);
+
+ MOV_HM(t, ip);
+
+ /* Thread. */
+ thread_ip_inc(p);
+}
+
+static inline void
+instr_mov_hh_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] mov (hh)\n",
+ p->thread_id);
+
+ MOV_HH(t, ip);
/* Thread. */
thread_ip_inc(p);
@@ -4060,13 +4193,16 @@ instr_alu_and_translate(struct rte_swx_pipeline *p,
fdst = struct_field_parse(p, NULL, dst, &dst_struct_id);
CHECK(fdst, EINVAL);
- /* AND or AND_S. */
+ /* AND, AND_MH, AND_HM, AND_HH. */
fsrc = struct_field_parse(p, action, src, &src_struct_id);
if (fsrc) {
instr->type = INSTR_ALU_AND;
- if ((dst[0] == 'h' && src[0] != 'h') ||
- (dst[0] != 'h' && src[0] == 'h'))
- instr->type = INSTR_ALU_AND_S;
+ if (dst[0] != 'h' && src[0] == 'h')
+ instr->type = INSTR_ALU_AND_MH;
+ if (dst[0] == 'h' && src[0] != 'h')
+ instr->type = INSTR_ALU_AND_HM;
+ if (dst[0] == 'h' && src[0] == 'h')
+ instr->type = INSTR_ALU_AND_HH;
instr->alu.dst.struct_id = (uint8_t)dst_struct_id;
instr->alu.dst.n_bits = fdst->n_bits;
@@ -4110,13 +4246,16 @@ instr_alu_or_translate(struct rte_swx_pipeline *p,
fdst = struct_field_parse(p, NULL, dst, &dst_struct_id);
CHECK(fdst, EINVAL);
- /* OR or OR_S. */
+ /* OR, OR_MH, OR_HM, OR_HH. */
fsrc = struct_field_parse(p, action, src, &src_struct_id);
if (fsrc) {
instr->type = INSTR_ALU_OR;
- if ((dst[0] == 'h' && src[0] != 'h') ||
- (dst[0] != 'h' && src[0] == 'h'))
- instr->type = INSTR_ALU_OR_S;
+ if (dst[0] != 'h' && src[0] == 'h')
+ instr->type = INSTR_ALU_OR_MH;
+ if (dst[0] == 'h' && src[0] != 'h')
+ instr->type = INSTR_ALU_OR_HM;
+ if (dst[0] == 'h' && src[0] == 'h')
+ instr->type = INSTR_ALU_OR_HH;
instr->alu.dst.struct_id = (uint8_t)dst_struct_id;
instr->alu.dst.n_bits = fdst->n_bits;
@@ -4160,13 +4299,16 @@ instr_alu_xor_translate(struct rte_swx_pipeline *p,
fdst = struct_field_parse(p, NULL, dst, &dst_struct_id);
CHECK(fdst, EINVAL);
- /* XOR or XOR_S. */
+ /* XOR, XOR_MH, XOR_HM, XOR_HH. */
fsrc = struct_field_parse(p, action, src, &src_struct_id);
if (fsrc) {
instr->type = INSTR_ALU_XOR;
- if ((dst[0] == 'h' && src[0] != 'h') ||
- (dst[0] != 'h' && src[0] == 'h'))
- instr->type = INSTR_ALU_XOR_S;
+ if (dst[0] != 'h' && src[0] == 'h')
+ instr->type = INSTR_ALU_XOR_MH;
+ if (dst[0] == 'h' && src[0] != 'h')
+ instr->type = INSTR_ALU_XOR_HM;
+ if (dst[0] == 'h' && src[0] == 'h')
+ instr->type = INSTR_ALU_XOR_HH;
instr->alu.dst.struct_id = (uint8_t)dst_struct_id;
instr->alu.dst.n_bits = fdst->n_bits;
@@ -4568,15 +4710,45 @@ instr_alu_and_exec(struct rte_swx_pipeline *p)
}
static inline void
-instr_alu_and_s_exec(struct rte_swx_pipeline *p)
+instr_alu_and_mh_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] and (mh)\n", p->thread_id);
+
+ /* Structs. */
+ ALU_MH(t, ip, &);
+
+ /* Thread. */
+ thread_ip_inc(p);
+}
+
+static inline void
+instr_alu_and_hm_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] and (hm)\n", p->thread_id);
+
+ /* Structs. */
+ ALU_HM_FAST(t, ip, &);
+
+ /* Thread. */
+ thread_ip_inc(p);
+}
+
+static inline void
+instr_alu_and_hh_exec(struct rte_swx_pipeline *p)
{
struct thread *t = &p->threads[p->thread_id];
struct instruction *ip = t->ip;
- TRACE("[Thread %2u] and (s)\n", p->thread_id);
+ TRACE("[Thread %2u] and (hh)\n", p->thread_id);
/* Structs. */
- ALU_S(t, ip, &);
+ ALU_HH_FAST(t, ip, &);
/* Thread. */
thread_ip_inc(p);
@@ -4613,15 +4785,45 @@ instr_alu_or_exec(struct rte_swx_pipeline *p)
}
static inline void
-instr_alu_or_s_exec(struct rte_swx_pipeline *p)
+instr_alu_or_mh_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] or (mh)\n", p->thread_id);
+
+ /* Structs. */
+ ALU_MH(t, ip, |);
+
+ /* Thread. */
+ thread_ip_inc(p);
+}
+
+static inline void
+instr_alu_or_hm_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] or (hm)\n", p->thread_id);
+
+ /* Structs. */
+ ALU_HM_FAST(t, ip, |);
+
+ /* Thread. */
+ thread_ip_inc(p);
+}
+
+static inline void
+instr_alu_or_hh_exec(struct rte_swx_pipeline *p)
{
struct thread *t = &p->threads[p->thread_id];
struct instruction *ip = t->ip;
- TRACE("[Thread %2u] or (s)\n", p->thread_id);
+ TRACE("[Thread %2u] or (hh)\n", p->thread_id);
/* Structs. */
- ALU_S(t, ip, |);
+ ALU_HH_FAST(t, ip, |);
/* Thread. */
thread_ip_inc(p);
@@ -4658,15 +4860,45 @@ instr_alu_xor_exec(struct rte_swx_pipeline *p)
}
static inline void
-instr_alu_xor_s_exec(struct rte_swx_pipeline *p)
+instr_alu_xor_mh_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] xor (mh)\n", p->thread_id);
+
+ /* Structs. */
+ ALU_MH(t, ip, ^);
+
+ /* Thread. */
+ thread_ip_inc(p);
+}
+
+static inline void
+instr_alu_xor_hm_exec(struct rte_swx_pipeline *p)
{
struct thread *t = &p->threads[p->thread_id];
struct instruction *ip = t->ip;
- TRACE("[Thread %2u] xor (s)\n", p->thread_id);
+ TRACE("[Thread %2u] xor (hm)\n", p->thread_id);
/* Structs. */
- ALU_S(t, ip, ^);
+ ALU_HM_FAST(t, ip, ^);
+
+ /* Thread. */
+ thread_ip_inc(p);
+}
+
+static inline void
+instr_alu_xor_hh_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] xor (hh)\n", p->thread_id);
+
+ /* Structs. */
+ ALU_HH_FAST(t, ip, ^);
/* Thread. */
thread_ip_inc(p);
@@ -6819,13 +7051,16 @@ instr_jmp_eq_translate(struct rte_swx_pipeline *p,
fa = struct_field_parse(p, action, a, &a_struct_id);
CHECK(fa, EINVAL);
- /* JMP_EQ or JMP_EQ_S. */
+ /* JMP_EQ, JMP_EQ_MH, JMP_EQ_HM, JMP_EQ_HH. */
fb = struct_field_parse(p, action, b, &b_struct_id);
if (fb) {
instr->type = INSTR_JMP_EQ;
- if ((a[0] == 'h' && b[0] != 'h') ||
- (a[0] != 'h' && b[0] == 'h'))
- instr->type = INSTR_JMP_EQ_S;
+ if (a[0] != 'h' && b[0] == 'h')
+ instr->type = INSTR_JMP_EQ_MH;
+ if (a[0] == 'h' && b[0] != 'h')
+ instr->type = INSTR_JMP_EQ_HM;
+ if (a[0] == 'h' && b[0] == 'h')
+ instr->type = INSTR_JMP_EQ_HH;
instr->jmp.ip = NULL; /* Resolved later. */
instr->jmp.a.struct_id = (uint8_t)a_struct_id;
@@ -6873,13 +7108,16 @@ instr_jmp_neq_translate(struct rte_swx_pipeline *p,
fa = struct_field_parse(p, action, a, &a_struct_id);
CHECK(fa, EINVAL);
- /* JMP_NEQ or JMP_NEQ_S. */
+ /* JMP_NEQ, JMP_NEQ_MH, JMP_NEQ_HM, JMP_NEQ_HH. */
fb = struct_field_parse(p, action, b, &b_struct_id);
if (fb) {
instr->type = INSTR_JMP_NEQ;
- if ((a[0] == 'h' && b[0] != 'h') ||
- (a[0] != 'h' && b[0] == 'h'))
- instr->type = INSTR_JMP_NEQ_S;
+ if (a[0] != 'h' && b[0] == 'h')
+ instr->type = INSTR_JMP_NEQ_MH;
+ if (a[0] == 'h' && b[0] != 'h')
+ instr->type = INSTR_JMP_NEQ_HM;
+ if (a[0] == 'h' && b[0] == 'h')
+ instr->type = INSTR_JMP_NEQ_HH;
instr->jmp.ip = NULL; /* Resolved later. */
instr->jmp.a.struct_id = (uint8_t)a_struct_id;
@@ -7114,14 +7352,36 @@ instr_jmp_eq_exec(struct rte_swx_pipeline *p)
}
static inline void
-instr_jmp_eq_s_exec(struct rte_swx_pipeline *p)
+instr_jmp_eq_mh_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] jmpeq (mh)\n", p->thread_id);
+
+ JMP_CMP_MH(t, ip, ==);
+}
+
+static inline void
+instr_jmp_eq_hm_exec(struct rte_swx_pipeline *p)
{
struct thread *t = &p->threads[p->thread_id];
struct instruction *ip = t->ip;
- TRACE("[Thread %2u] jmpeq (s)\n", p->thread_id);
+ TRACE("[Thread %2u] jmpeq (hm)\n", p->thread_id);
- JMP_CMP_S(t, ip, ==);
+ JMP_CMP_HM(t, ip, ==);
+}
+
+static inline void
+instr_jmp_eq_hh_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] jmpeq (hh)\n", p->thread_id);
+
+ JMP_CMP_HH_FAST(t, ip, ==);
}
static inline void
@@ -7147,14 +7407,36 @@ instr_jmp_neq_exec(struct rte_swx_pipeline *p)
}
static inline void
-instr_jmp_neq_s_exec(struct rte_swx_pipeline *p)
+instr_jmp_neq_mh_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] jmpneq (mh)\n", p->thread_id);
+
+ JMP_CMP_MH(t, ip, !=);
+}
+
+static inline void
+instr_jmp_neq_hm_exec(struct rte_swx_pipeline *p)
+{
+ struct thread *t = &p->threads[p->thread_id];
+ struct instruction *ip = t->ip;
+
+ TRACE("[Thread %2u] jmpneq (hm)\n", p->thread_id);
+
+ JMP_CMP_HM(t, ip, !=);
+}
+
+static inline void
+instr_jmp_neq_hh_exec(struct rte_swx_pipeline *p)
{
struct thread *t = &p->threads[p->thread_id];
struct instruction *ip = t->ip;
- TRACE("[Thread %2u] jmpneq (s)\n", p->thread_id);
+ TRACE("[Thread %2u] jmpneq (hh)\n", p->thread_id);
- JMP_CMP_S(t, ip, !=);
+ JMP_CMP_HH_FAST(t, ip, !=);
}
static inline void
@@ -8136,7 +8418,9 @@ static instr_exec_t instruction_table[] = {
[INSTR_HDR_INVALIDATE] = instr_hdr_invalidate_exec,
[INSTR_MOV] = instr_mov_exec,
- [INSTR_MOV_S] = instr_mov_s_exec,
+ [INSTR_MOV_MH] = instr_mov_mh_exec,
+ [INSTR_MOV_HM] = instr_mov_hm_exec,
+ [INSTR_MOV_HH] = instr_mov_hh_exec,
[INSTR_MOV_I] = instr_mov_i_exec,
[INSTR_DMA_HT] = instr_dma_ht_exec,
@@ -8168,15 +8452,21 @@ static instr_exec_t instruction_table[] = {
[INSTR_ALU_CKSUB_FIELD] = instr_alu_cksub_field_exec,
[INSTR_ALU_AND] = instr_alu_and_exec,
- [INSTR_ALU_AND_S] = instr_alu_and_s_exec,
+ [INSTR_ALU_AND_MH] = instr_alu_and_mh_exec,
+ [INSTR_ALU_AND_HM] = instr_alu_and_hm_exec,
+ [INSTR_ALU_AND_HH] = instr_alu_and_hh_exec,
[INSTR_ALU_AND_I] = instr_alu_and_i_exec,
[INSTR_ALU_OR] = instr_alu_or_exec,
- [INSTR_ALU_OR_S] = instr_alu_or_s_exec,
+ [INSTR_ALU_OR_MH] = instr_alu_or_mh_exec,
+ [INSTR_ALU_OR_HM] = instr_alu_or_hm_exec,
+ [INSTR_ALU_OR_HH] = instr_alu_or_hh_exec,
[INSTR_ALU_OR_I] = instr_alu_or_i_exec,
[INSTR_ALU_XOR] = instr_alu_xor_exec,
- [INSTR_ALU_XOR_S] = instr_alu_xor_s_exec,
+ [INSTR_ALU_XOR_MH] = instr_alu_xor_mh_exec,
+ [INSTR_ALU_XOR_HM] = instr_alu_xor_hm_exec,
+ [INSTR_ALU_XOR_HH] = instr_alu_xor_hh_exec,
[INSTR_ALU_XOR_I] = instr_alu_xor_i_exec,
[INSTR_ALU_SHL] = instr_alu_shl_exec,
@@ -8254,11 +8544,15 @@ static instr_exec_t instruction_table[] = {
[INSTR_JMP_ACTION_MISS] = instr_jmp_action_miss_exec,
[INSTR_JMP_EQ] = instr_jmp_eq_exec,
- [INSTR_JMP_EQ_S] = instr_jmp_eq_s_exec,
+ [INSTR_JMP_EQ_MH] = instr_jmp_eq_mh_exec,
+ [INSTR_JMP_EQ_HM] = instr_jmp_eq_hm_exec,
+ [INSTR_JMP_EQ_HH] = instr_jmp_eq_hh_exec,
[INSTR_JMP_EQ_I] = instr_jmp_eq_i_exec,
[INSTR_JMP_NEQ] = instr_jmp_neq_exec,
- [INSTR_JMP_NEQ_S] = instr_jmp_neq_s_exec,
+ [INSTR_JMP_NEQ_MH] = instr_jmp_neq_mh_exec,
+ [INSTR_JMP_NEQ_HM] = instr_jmp_neq_hm_exec,
+ [INSTR_JMP_NEQ_HH] = instr_jmp_neq_hh_exec,
[INSTR_JMP_NEQ_I] = instr_jmp_neq_i_exec,
[INSTR_JMP_LT] = instr_jmp_lt_exec,
--
2.17.1
^ permalink raw reply [flat|nested] 2+ messages in thread* Re: [dpdk-stable] [dpdk-dev] [PATCH] pipeline: fix endianness conversions
2021-04-03 0:17 [dpdk-stable] [PATCH] pipeline: fix endianness conversions Cristian Dumitrescu
@ 2021-04-20 0:34 ` Thomas Monjalon
0 siblings, 0 replies; 2+ messages in thread
From: Thomas Monjalon @ 2021-04-20 0:34 UTC (permalink / raw)
To: Cristian Dumitrescu; +Cc: dev, stable
03/04/2021 02:17, Cristian Dumitrescu:
> The SWX pipeline instructions work with operands of different types:
> header fields (h.header.field), packet meta-data (m.field), extern
> object mailbox field (e.obj.field), extern function (f.field), action
> data read from table entries (t.field), or immediate values; hence the
> HMEFTI acronym. The H operands are stored in network byte order (NBO),
> while the MEFT operands are stored in host byte order (HBO), hence the
> need to operate endianness conversions.
>
> Some of the endianness conversion macros were not working correctly
> for some cases such as operands of different sizes, and they are fixed
> now. Affected instructions: mov, and, or, xor, jmpeq, jmpneq.
>
> Fixes: 7210349d5baa ("pipeline: add SWX move instruction")
> Fixes: 650195cf965a ("pipeline: introduce SWX and instruction")
> Fixes: 8f796198dcda ("pipeline: introduce SWX or instruction")
> Fixes: b4e607f9fd5e ("pipeline: introduce SWX XOR instruction")
> Fixes: b3947e25bed4 ("pipeline: introduce SWX jump and return instructions")
> Cc: stable@dpdk.org
>
> Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Applied, thanks
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2021-04-20 0:34 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-04-03 0:17 [dpdk-stable] [PATCH] pipeline: fix endianness conversions Cristian Dumitrescu
2021-04-20 0:34 ` [dpdk-stable] [dpdk-dev] " Thomas Monjalon
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).