perf intel-pt: Fix ip compression

PATCH In review Adrian Hunter 3 participants 4 messages Jul 20, 2016 – Aug 16, 2016 perf events

Share & tools

Adrian Hunter

author reviewer

2016-07-20 09:04 UTC #0 lore raw

PATCH 1/1 perf intel-pt: Fix ip compression

June 2015 Intel SDM introduced IP Compression types 4 and 6. Refer section 36.4.2.2 Target IP (TIP) Packet - IP Compression.

Existing Intel PT packet decoder did not support type 4, and got type 6 wrong. Because type 3 and type 4 have the same number of bytes, the packet 'count' has been changed from being the number of ip bytes to being the type code. That allows the Intel PT decoder to correctly decide whether to sign-extend or use the last ip. However that also meant the code had to be adjusted in a number of places.

Currently hardware is not using the new compression types, so this fix has no effect on existing hardware.

Signed-off-by Adrian Hunter <[email protected]>

---
 .../perf/util/intel-pt-decoder/intel-pt-decoder.c  | 44 +++++++++++-----------
 .../util/intel-pt-decoder/intel-pt-pkt-decoder.c   | 24 ++++++++----
 2 files changed, 40 insertions(+), 28 deletions(-)

tools/perf/util/intel-pt-decoder/intel-pt-decoder.c +23 ~~-21~~

index 9c8f15da86ce..8ff6c6a61291 100644
@@ -123,8 +123,6 @@ struct intel_pt_decoder {
 	bool have_calc_cyc_to_tsc;
 	int exec_mode;
 	unsigned int insn_bytes;
-	uint64_t sign_bit;
-	uint64_t sign_bits;
 	uint64_t period;
 	enum intel_pt_period_type period_type;
 	uint64_t tot_insn_cnt;

@@ -191,9 +189,6 @@ struct intel_pt_decoder *intel_pt_decoder_new(struct intel_pt_params *params)
 	decoder->data               = params->data;
 	decoder->return_compression = params->return_compression;
 
-	decoder->sign_bit           = (uint64_t)1 << 47;
-	decoder->sign_bits          = ~(((uint64_t)1 << 48) - 1);
-
 	decoder->period             = params->period;
 	decoder->period_type        = params->period_type;

@@ -362,21 +357,30 @@ int intel_pt__strerror(int code, char *buf, size_t buflen)
 	return 0;
 }
 
-static uint64_t intel_pt_calc_ip(struct intel_pt_decoder *decoder,
-				 const struct intel_pt_pkt *packet,
+static uint64_t intel_pt_calc_ip(const struct intel_pt_pkt *packet,
 				 uint64_t last_ip)
 {
 	uint64_t ip;
 
 	switch (packet->count) {
-	case 2:
+	case 1:
 		ip = (last_ip & (uint64_t)0xffffffffffff0000ULL) |
 		     packet->payload;
 		break;
-	case 4:
+	case 2:
 		ip = (last_ip & (uint64_t)0xffffffff00000000ULL) |
 		     packet->payload;
 		break;
+	case 3:
+		ip = packet->payload;
+		/* Sign-extend 6-byte ip */
+		if (ip & (uint64_t)0x800000000000ULL)
+			ip |= (uint64_t)0xffff000000000000ULL;
+		break;
+	case 4:
+		ip = (last_ip & (uint64_t)0xffff000000000000ULL) |
+		     packet->payload;
+		break;
 	case 6:
 		ip = packet->payload;
 		break;

@@ -384,16 +388,12 @@ static uint64_t intel_pt_calc_ip(struct intel_pt_decoder *decoder,
 		return 0;
 	}
 
-	if (ip & decoder->sign_bit)
-		return ip | decoder->sign_bits;
-
 	return ip;
 }
 
 static inline void intel_pt_set_last_ip(struct intel_pt_decoder *decoder)
 {
-	decoder->last_ip = intel_pt_calc_ip(decoder, &decoder->packet,
-					    decoder->last_ip);
+	decoder->last_ip = intel_pt_calc_ip(&decoder->packet, decoder->last_ip);
 }
 
 static inline void intel_pt_set_ip(struct intel_pt_decoder *decoder)

@@ -1657,6 +1657,12 @@ next:
 	}
 }
 
+static inline bool intel_pt_have_ip(struct intel_pt_decoder *decoder)
+{
+	return decoder->last_ip || decoder->packet.count == 0 ||
+	       decoder->packet.count == 3 || decoder->packet.count == 6;
+}
+
 /* Walk PSB+ packets to get in sync. */
 static int intel_pt_walk_psb(struct intel_pt_decoder *decoder)
 {

@@ -1677,8 +1683,7 @@ static int intel_pt_walk_psb(struct intel_pt_decoder *decoder)
 
 		case INTEL_PT_FUP:
 			decoder->pge = true;
-			if (decoder->last_ip || decoder->packet.count == 6 ||
-			    decoder->packet.count == 0) {
+			if (intel_pt_have_ip(decoder)) {
 				uint64_t current_ip = decoder->ip;
 
 				intel_pt_set_ip(decoder);

@@ -1767,8 +1772,7 @@ static int intel_pt_walk_to_ip(struct intel_pt_decoder *decoder)
 		case INTEL_PT_TIP_PGE:
 		case INTEL_PT_TIP:
 			decoder->pge = decoder->packet.type != INTEL_PT_TIP_PGD;
-			if (decoder->last_ip || decoder->packet.count == 6 ||
-			    decoder->packet.count == 0)
+			if (intel_pt_have_ip(decoder))
 				intel_pt_set_ip(decoder);
 			if (decoder->ip)
 				return 0;

@@ -1776,9 +1780,7 @@ static int intel_pt_walk_to_ip(struct intel_pt_decoder *decoder)
 
 		case INTEL_PT_FUP:
 			if (decoder->overflow) {
-				if (decoder->last_ip ||
-				    decoder->packet.count == 6 ||
-				    decoder->packet.count == 0)
+				if (intel_pt_have_ip(decoder))
 					intel_pt_set_ip(decoder);
 				if (decoder->ip)
 					return 0;

tools/perf/util/intel-pt-decoder/intel-pt-pkt-decoder.c +17 -7

index b1257c816310..4f7b32020487 100644
@@ -292,36 +292,46 @@ static int intel_pt_get_ip(enum intel_pt_pkt_type type, unsigned int byte,
 			   const unsigned char *buf, size_t len,
 			   struct intel_pt_pkt *packet)
 {
-	switch (byte >> 5) {
+	int ip_len;
+
+	packet->count = byte >> 5;
+
+	switch (packet->count) {
 	case 0:
-		packet->count = 0;
+		ip_len = 0;
 		break;
 	case 1:
 		if (len < 3)
 			return INTEL_PT_NEED_MORE_BYTES;
-		packet->count = 2;
+		ip_len = 2;
 		packet->payload = le16_to_cpu(*(uint16_t *)(buf + 1));
 		break;
 	case 2:
 		if (len < 5)
 			return INTEL_PT_NEED_MORE_BYTES;
-		packet->count = 4;
+		ip_len = 4;
 		packet->payload = le32_to_cpu(*(uint32_t *)(buf + 1));
 		break;
 	case 3:
-	case 6:
+	case 4:
 		if (len < 7)
 			return INTEL_PT_NEED_MORE_BYTES;
-		packet->count = 6;
+		ip_len = 6;
 		memcpy_le64(&packet->payload, buf + 1, 6);
 		break;
+	case 6:
+		if (len < 9)
+			return INTEL_PT_NEED_MORE_BYTES;
+		ip_len = 8;
+		packet->payload = le64_to_cpu(*(uint64_t *)(buf + 1));
+		break;
 	default:
 		return INTEL_PT_BAD_PACKET;
 	}
 
 	packet->type = type;
 
-	return packet->count + 1;
+	return ip_len + 1;
 }
 
 static int intel_pt_get_mode(const unsigned char *buf, size_t len,

Signature · 1 line

1.9.1

Adrian Hunter

author reviewer

2016-08-10 18:03 UTC #1 lore raw

Quoted 10 lines · Adrian Hunter

June 2015 Intel SDM introduced IP Compression types 4 and 6. Refer section 36.4.2.2 Target IP (TIP) Packet - IP Compression.
Existing Intel PT packet decoder did not support type 4, and got type 6 wrong. Because type 3 and type 4 have the same number of bytes, the packet 'count' has been changed from being the number of ip bytes to being the type code. That allows the Intel PT decoder to correctly decide whether to sign-extend or use the last ip. However that also meant the code had to be adjusted in a number of places.

Quoted 8 lines

Currently hardware is not using the new compression types, so this fix has no effect on existing hardware.
Signed-off-by Adrian Hunter <[email protected]>
---
 .../perf/util/intel-pt-decoder/intel-pt-decoder.c  | 44 +++++++++++-----------
 .../util/intel-pt-decoder/intel-pt-pkt-decoder.c   | 24 ++++++++----
 2 files changed, 40 insertions(+), 28 deletions(-)

This is outstanding but still applies.

Regards
Adrian

Arnaldo Carvalho de Melo

maintainer

2016-08-10 19:01 UTC #2 lore raw

Em Wed, Aug 10, 2016 at 11:32:21AM +0300, Adrian Hunter escreveu:

Quoted 20 lines · Adrian Hunter

Quoted 10 lines
June 2015 Intel SDM introduced IP Compression types 4 and 6. Refer section 36.4.2.2 Target IP (TIP) Packet - IP Compression.
Existing Intel PT packet decoder did not support type 4, and got type 6 wrong. Because type 3 and type 4 have the same number of bytes, the packet 'count' has been changed from being the number of ip bytes to being the type code. That allows the Intel PT decoder to correctly decide whether to sign-extend or use the last ip. However that also meant the code had to be adjusted in a number of places.
Quoted 8 lines
Currently hardware is not using the new compression types, so this fix has no effect on existing hardware.
Signed-off-by Adrian Hunter <[email protected]>
---
 .../perf/util/intel-pt-decoder/intel-pt-decoder.c  | 44 +++++++++++-----------
 .../util/intel-pt-decoder/intel-pt-pkt-decoder.c   | 24 ++++++++----
 2 files changed, 40 insertions(+), 28 deletions(-)

Quoted 3 lines

Hi
This is outstanding but still applies.

Oops, will get that merged, sorry for the delay.

- Arnaldo

tip-bot for Adrian Hunter

merge bot

2016-08-16 18:15 UTC #3 lore raw

In reply to Adrian Hunter

perf intel-pt: Fix ip compression

Commit-ID:  e1717e0485af4f47fc4da1e979ac817f9ad61b0f
Gitweb:     http://git.kernel.org/tip/e1717e0485af4f47fc4da1e979ac817f9ad61b0f
Author:     Adrian Hunter <[email protected]>
AuthorDate: Wed, 20 Jul 2016 12:00:06 +0300
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Fri, 12 Aug 2016 14:39:48 -0300

perf intel-pt: Fix ip compression

The June 2015 Intel SDM introduced IP Compression types 4 and 6. Refer to section 36.4.2.2 Target IP (TIP) Packet - IP Compression.

Currently hardware is not using the new compression types, so this fix has no effect on existing hardware.

Signed-off-by Adrian Hunter <[email protected]>

Cc Jiri Olsa <[email protected]>

Link http://lkml.kernel.org/r/[email protected] open here

Signed-off-by Arnaldo Carvalho de Melo <[email protected]>

---
 .../perf/util/intel-pt-decoder/intel-pt-decoder.c  | 44 +++++++++++-----------
 .../util/intel-pt-decoder/intel-pt-pkt-decoder.c   | 24 ++++++++----
 2 files changed, 40 insertions(+), 28 deletions(-)

tools/perf/util/intel-pt-decoder/intel-pt-decoder.c +23 ~~-21~~

index 9c8f15d..8ff6c6a 100644
@@ -123,8 +123,6 @@ struct intel_pt_decoder {
 	bool have_calc_cyc_to_tsc;
 	int exec_mode;
 	unsigned int insn_bytes;
-	uint64_t sign_bit;
-	uint64_t sign_bits;
 	uint64_t period;
 	enum intel_pt_period_type period_type;
 	uint64_t tot_insn_cnt;

@@ -191,9 +189,6 @@ struct intel_pt_decoder *intel_pt_decoder_new(struct intel_pt_params *params)
 	decoder->data               = params->data;
 	decoder->return_compression = params->return_compression;
 
-	decoder->sign_bit           = (uint64_t)1 << 47;
-	decoder->sign_bits          = ~(((uint64_t)1 << 48) - 1);
-
 	decoder->period             = params->period;
 	decoder->period_type        = params->period_type;

@@ -362,21 +357,30 @@ int intel_pt__strerror(int code, char *buf, size_t buflen)
 	return 0;
 }
 
-static uint64_t intel_pt_calc_ip(struct intel_pt_decoder *decoder,
-				 const struct intel_pt_pkt *packet,
+static uint64_t intel_pt_calc_ip(const struct intel_pt_pkt *packet,
 				 uint64_t last_ip)
 {
 	uint64_t ip;
 
 	switch (packet->count) {
-	case 2:
+	case 1:
 		ip = (last_ip & (uint64_t)0xffffffffffff0000ULL) |
 		     packet->payload;
 		break;
-	case 4:
+	case 2:
 		ip = (last_ip & (uint64_t)0xffffffff00000000ULL) |
 		     packet->payload;
 		break;
+	case 3:
+		ip = packet->payload;
+		/* Sign-extend 6-byte ip */
+		if (ip & (uint64_t)0x800000000000ULL)
+			ip |= (uint64_t)0xffff000000000000ULL;
+		break;
+	case 4:
+		ip = (last_ip & (uint64_t)0xffff000000000000ULL) |
+		     packet->payload;
+		break;
 	case 6:
 		ip = packet->payload;
 		break;

@@ -384,16 +388,12 @@ static uint64_t intel_pt_calc_ip(struct intel_pt_decoder *decoder,
 		return 0;
 	}
 
-	if (ip & decoder->sign_bit)
-		return ip | decoder->sign_bits;
-
 	return ip;
 }
 
 static inline void intel_pt_set_last_ip(struct intel_pt_decoder *decoder)
 {
-	decoder->last_ip = intel_pt_calc_ip(decoder, &decoder->packet,
-					    decoder->last_ip);
+	decoder->last_ip = intel_pt_calc_ip(&decoder->packet, decoder->last_ip);
 }
 
 static inline void intel_pt_set_ip(struct intel_pt_decoder *decoder)

@@ -1657,6 +1657,12 @@ next:
 	}
 }
 
+static inline bool intel_pt_have_ip(struct intel_pt_decoder *decoder)
+{
+	return decoder->last_ip || decoder->packet.count == 0 ||
+	       decoder->packet.count == 3 || decoder->packet.count == 6;
+}
+
 /* Walk PSB+ packets to get in sync. */
 static int intel_pt_walk_psb(struct intel_pt_decoder *decoder)
 {

@@ -1677,8 +1683,7 @@ static int intel_pt_walk_psb(struct intel_pt_decoder *decoder)
 
 		case INTEL_PT_FUP:
 			decoder->pge = true;
-			if (decoder->last_ip || decoder->packet.count == 6 ||
-			    decoder->packet.count == 0) {
+			if (intel_pt_have_ip(decoder)) {
 				uint64_t current_ip = decoder->ip;
 
 				intel_pt_set_ip(decoder);

@@ -1767,8 +1772,7 @@ static int intel_pt_walk_to_ip(struct intel_pt_decoder *decoder)
 		case INTEL_PT_TIP_PGE:
 		case INTEL_PT_TIP:
 			decoder->pge = decoder->packet.type != INTEL_PT_TIP_PGD;
-			if (decoder->last_ip || decoder->packet.count == 6 ||
-			    decoder->packet.count == 0)
+			if (intel_pt_have_ip(decoder))
 				intel_pt_set_ip(decoder);
 			if (decoder->ip)
 				return 0;

@@ -1776,9 +1780,7 @@ static int intel_pt_walk_to_ip(struct intel_pt_decoder *decoder)
 
 		case INTEL_PT_FUP:
 			if (decoder->overflow) {
-				if (decoder->last_ip ||
-				    decoder->packet.count == 6 ||
-				    decoder->packet.count == 0)
+				if (intel_pt_have_ip(decoder))
 					intel_pt_set_ip(decoder);
 				if (decoder->ip)
 					return 0;

tools/perf/util/intel-pt-decoder/intel-pt-pkt-decoder.c +17 -7

index b1257c8..4f7b320 100644
@@ -292,36 +292,46 @@ static int intel_pt_get_ip(enum intel_pt_pkt_type type, unsigned int byte,
 			   const unsigned char *buf, size_t len,
 			   struct intel_pt_pkt *packet)
 {
-	switch (byte >> 5) {
+	int ip_len;
+
+	packet->count = byte >> 5;
+
+	switch (packet->count) {
 	case 0:
-		packet->count = 0;
+		ip_len = 0;
 		break;
 	case 1:
 		if (len < 3)
 			return INTEL_PT_NEED_MORE_BYTES;
-		packet->count = 2;
+		ip_len = 2;
 		packet->payload = le16_to_cpu(*(uint16_t *)(buf + 1));
 		break;
 	case 2:
 		if (len < 5)
 			return INTEL_PT_NEED_MORE_BYTES;
-		packet->count = 4;
+		ip_len = 4;
 		packet->payload = le32_to_cpu(*(uint32_t *)(buf + 1));
 		break;
 	case 3:
-	case 6:
+	case 4:
 		if (len < 7)
 			return INTEL_PT_NEED_MORE_BYTES;
-		packet->count = 6;
+		ip_len = 6;
 		memcpy_le64(&packet->payload, buf + 1, 6);
 		break;
+	case 6:
+		if (len < 9)
+			return INTEL_PT_NEED_MORE_BYTES;
+		ip_len = 8;
+		packet->payload = le64_to_cpu(*(uint64_t *)(buf + 1));
+		break;
 	default:
 		return INTEL_PT_BAD_PACKET;
 	}
 
 	packet->type = type;
 
-	return packet->count + 1;
+	return ip_len + 1;
 }
 
 static int intel_pt_get_mode(const unsigned char *buf, size_t len,