Merge tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-tracegrafted

Pull tracing fixes from Steven Rostedt:

 - Don't destroy user event fields when removal fails

   User event fields are destroyed before the event is removed from
   visibility. But that can fail leaving the still visible event with no
   fields. Move the destroying of the fields to after the event is
   successfully removed from visibility.

 - Initialize function graph state is fork before calling
   copy_exec_state()

   For non-CLONE_VM forks, copy_exec_state() allocates a new
   task_exec_state. If that allocation fails, ftrace_graph_exit_task()
   will free the tasks ret_stack pointer. Since that pointer is still
   using the parent's ret_stack, it mistakenly frees the parent's
   pointer too.

   Call ftrace_graph_init() on the task first which will NULL out the
   new tasks's ret_stack and if the copy fails, it will not free
   anything.

 - Remove FGRAPH_MAX_INDEX

   The macro FGRAPH_MAX_INDEX was added but never used. Remove it.

 - Save ent_size in function graph printing of nested functions

   The function graph tracer needs to look at the next event to see if
   the next event is the return of the current function entry. If it is,
   it prints a single line:

	ktime_get();

   Otherwise it prints it like a nested function:

	tick_nohz_irq_exit() {
	    ktime_get();
	    kcpustat_irq_exit();
	}

   In order to look at the next event, it must save the current event so
   that it has the information to print from it. It saves the event in
   the iterator descriptor called "ent". What it doesn't save is the
   ent_size of the event which is now used to know if the function graph
   arguments are to be printed. The peek doesn't save the size so the
   size used happens to be that of the size of the last event that was
   seen.

   Save the entry event size in the iterator descriptor so that the
   correct size is used.

 - Fix several errors with freeing data in the histogram code

   The histogram code had a lot of leaked or or incorrect accounting
   when failures happen. Correct them.

 - Fix histogram regression of .percent and .graph modifiers

   Up until 6.3 histogram values could have "percent" or "graph"
   modifiers that changed how they were printed. But a change that added
   restricting histograms values from being strings, stack traces and
   other modifiers inadvertently prevented them from using the percent
   and graph modifiers, which were legal use cases for values.

   Put back the percent and graph modifiers.

 - Fix various typos in the comments

 - Set the trace_clock before initializing a histogram with clock
   argument

   The histogram API allows the user to specific which trace clock to
   use via a "clock=" string. The histogram is set up first before the
   clock is checked. If the passed in clock is not valid, it exits
   without fully fixing up the histogram leaving it on the list and a
   use-after-free can trigger.

   Update the clock argument first and if it fails then exit gracefully
   before the histogram trigger is placed on any lists.

 - Restore :mod: trailer after parsing in ftrace_set_clr_event

   The function ftrace_set_clr_event() modifies the parse string and
   needs to put it back to what was passed in. It searches for ":mod:"
   via a strsep() but fails to put back the first ':' in the string.

   Add back the ':' in the passed in string.

 - Take trace_array reference when opening a tracer options file

   The options files are dynamically created and some tracers add their
   own options. When a tracer adds their own list of options, the
   trace_array holding them has an array to hold the list of options for
   each tracer. This array increases in size via a krealloc(), and the
   new entry gets a newly allocated array to hold the options of the new
   tracer being added.

   The element in each entry of the tracer's option array holds a
   pointer back to the trace_array, a pointer to the tracer it is
   associated to, a pointer to the flags of the option.

   The issue is that these arrays are freed when the trace_array is
   freed when its instance it represents is removed from the instances
   directory. There's a race that an open of one of these options files
   can happen when the instance is being removed.

   Add a new helper function to be called by the open function of the
   options file to iterate all existing trace_arrays under a lock and
   find the one that has the given option element in one of it's tracer
   arrays. If found, then update the associated trace_array's reference
   counter to keep it from being freed. If not found, have the open call
   return -ENODEV.

 - Disable interrupts when acquiring the lock in rb_wake_up_waiters()

   The function rb_wake_up_waiters() assumes it will be called in
   interrupt context and does not disable irqs when taking
   cpu_buffer->reader_lock, which can be called in hard interrupt
   context. The issue is in PREEMPT_RT, this function is called in
   thread context leaving this lock open to a deadlock.

   Take the lock with interrupts disabled.

 - Use rcu_assign_pointer() for tmp_ops filter hash

   The tmp_ops used in update_ftrace_direct_mod() assigns its
   filter_hash field directly, but that field is annotated as __rcu and
   sparse complains. Assign it with rcu_assign_pointer()

 - Fix use-after-free in enable_trigger_private_data_free()

   The trace_event_call is accessed through the event_trigger_data's
   trace_event_file pointer to put the trace_event_call on freeing. The
   issue is that the trace_event_file data may have been freed already
   causing a use-after-free. Add a field to the event_trigger_data that
   points directly to the trace_event_call so that it can decrement its
   reference directly without needing to go through the
   trace_event_file.

 - Fix accounting of buffer data remote headers

   trace_buffer_desc_size() and trace_remote_alloc_buffer() undercount
   the number of pages is needed for the asked for size as it doesn't
   take into account the meta data on each page. Add a helper function
   to do the calculation properly and use that in these functions.

 - Catch nr_page_va overflow in ring_buffer_desc sizing

   The number of pages per remote ring buffer is capped by
   ring_buffer_desc::nr_page_va (32 bits). A buffer_size large enough to
   overflow that field would silently allocate a descriptor smaller than
   what was asked for.

 - Do not resize the subbuf order if any per_cpu buffer is disabled

   The mmapping of ring buffers disables resizing the subbuffers, but it
   is done per-cpu whereas the subbuf size change is done for all the
   per_cpu buffers under the buffer->mutex. It could change the size of
   some while the mapping is happening on others. Have the resize of the
   subbuf order check all the per_cpu buffers under the lock to see if
   any of them is disabled before starting and causing an inconsistency
   between buffers that are being mapped.

* tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (25 commits)
  ring-buffer: Check resize_disabled before publishing the new subbuf order
  tracing/remotes: Catch nr_page_va overflow in ring_buffer_desc sizing
  tracing/remotes: Account for ring buffer page header in size calculation
  tracing: Don't dereference trace_event_file in deferred trigger free
  ftrace: Use rcu_assign_pointer() for tmp_ops filter hash
  ring-buffer: Acquire the lock with irqsave in rb_wake_up_waiters()
  tracing: Take trace_array reference when opening a tracer options file
  tracing: Fix ring_buffer_read_page_size() kernel-doc
  tracing: Restore :mod: trailer after parsing in ftrace_set_clr_event()
  tracing: Fix memory corruption from a "STACKTRACE" histogram key
  tracing: Fix memory corruption from the histogram stacktrace modifier
  tracing: Undo the registration when enabling the histogram trigger fails
  tracing: Take the reference before publishing the named histogram trigger
  tracing: Set the trace clock before registering the histogram trigger
  tracing: Fix typo "preceeded" in comment
  tracing: Fix typo "availabe" in comment
  tracing: Let histogram values keep the percent and graph modifiers
  tracing: Keep the entry count when the histogram stats allocation fails
  tracing: Free histogram the field rejected for a bad modifier
  tracing: Free histogram the var ref when its initialization fails
  ...

5 files changed, 1493 insertions, 0 deletions

diff --git a/tools/firewire/Makefile b/tools/firewire/Makefile
new file mode 100644
index 000000000..67b6e9fca
--- /dev/null
+++ b/tools/firewire/Makefile
@@ -0,0 +1,20 @@
+# SPDX-License-Identifier: GPL-2.0
+prefix = /usr
+nosy-dump-version = 0.4
+
+CC = gcc
+
+all : nosy-dump
+
+nosy-dump : CFLAGS = -Wall -O2 -g
+nosy-dump : CPPFLAGS = -DVERSION=\"$(nosy-dump-version)\" -I../../drivers/firewire
+nosy-dump : LDFLAGS = -g
+nosy-dump : LDLIBS = -lpopt
+
+nosy-dump : nosy-dump.o decode-fcp.o
+
+clean :
+	rm -rf *.o nosy-dump
+
+install :
+	install nosy-dump $(prefix)/bin/nosy-dump
diff --git a/tools/firewire/decode-fcp.c b/tools/firewire/decode-fcp.c
new file mode 100644
index 000000000..f115a3be8
--- /dev/null
+++ b/tools/firewire/decode-fcp.c
@@ -0,0 +1,214 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/firewire-constants.h>
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "list.h"
+#include "nosy-dump.h"
+
+#define CSR_FCP_COMMAND			0xfffff0000b00ull
+#define CSR_FCP_RESPONSE		0xfffff0000d00ull
+
+static const char * const ctype_names[] = {
+	[0x0] = "control",		[0x8] = "not implemented",
+	[0x1] = "status",		[0x9] = "accepted",
+	[0x2] = "specific inquiry",	[0xa] = "rejected",
+	[0x3] = "notify",		[0xb] = "in transition",
+	[0x4] = "general inquiry",	[0xc] = "stable",
+	[0x5] = "(reserved 0x05)",	[0xd] = "changed",
+	[0x6] = "(reserved 0x06)",	[0xe] = "(reserved 0x0e)",
+	[0x7] = "(reserved 0x07)",	[0xf] = "interim",
+};
+
+static const char * const subunit_type_names[] = {
+	[0x00] = "monitor",		[0x10] = "(reserved 0x10)",
+	[0x01] = "audio",		[0x11] = "(reserved 0x11)",
+	[0x02] = "printer",		[0x12] = "(reserved 0x12)",
+	[0x03] = "disc",		[0x13] = "(reserved 0x13)",
+	[0x04] = "tape recorder/player",[0x14] = "(reserved 0x14)",
+	[0x05] = "tuner",		[0x15] = "(reserved 0x15)",
+	[0x06] = "ca",			[0x16] = "(reserved 0x16)",
+	[0x07] = "camera",		[0x17] = "(reserved 0x17)",
+	[0x08] = "(reserved 0x08)",	[0x18] = "(reserved 0x18)",
+	[0x09] = "panel",		[0x19] = "(reserved 0x19)",
+	[0x0a] = "bulletin board",	[0x1a] = "(reserved 0x1a)",
+	[0x0b] = "camera storage",	[0x1b] = "(reserved 0x1b)",
+	[0x0c] = "(reserved 0x0c)",	[0x1c] = "vendor unique",
+	[0x0d] = "(reserved 0x0d)",	[0x1d] = "all subunit types",
+	[0x0e] = "(reserved 0x0e)",	[0x1e] = "subunit_type extended to next byte",
+	[0x0f] = "(reserved 0x0f)",	[0x1f] = "unit",
+};
+
+struct avc_enum {
+	int value;
+	const char *name;
+};
+
+struct avc_field {
+	const char *name;	/* Short name for field. */
+	int offset;		/* Location of field, specified in bits; */
+				/* negative means from end of packet.    */
+	int width;		/* Width of field, 0 means use data_length. */
+	struct avc_enum *names;
+};
+
+struct avc_opcode_info {
+	const char *name;
+	struct avc_field fields[8];
+};
+
+struct avc_enum power_field_names[] = {
+	{ 0x70, "on" },
+	{ 0x60, "off" },
+	{ }
+};
+
+static const struct avc_opcode_info opcode_info[256] = {
+
+	/* TA Document 1999026 */
+	/* AV/C Digital Interface Command Set General Specification 4.0 */
+	[0xb2] = { "power", {
+			{ "state", 0, 8, power_field_names }
+		}
+	},
+	[0x30] = { "unit info", {
+			{ "foo", 0, 8 },
+			{ "unit_type", 8, 5 },
+			{ "unit", 13, 3 },
+			{ "company id", 16, 24 },
+		}
+	},
+	[0x31] = { "subunit info" },
+	[0x01] = { "reserve" },
+	[0xb0] = { "version" },
+	[0x00] = { "vendor dependent" },
+	[0x02] = { "plug info" },
+	[0x12] = { "channel usage" },
+	[0x24] = { "connect" },
+	[0x20] = { "connect av" },
+	[0x22] = { "connections" },
+	[0x11] = { "digital input" },
+	[0x10] = { "digital output" },
+	[0x25] = { "disconnect" },
+	[0x21] = { "disconnect av" },
+	[0x19] = { "input plug signal format" },
+	[0x18] = { "output plug signal format" },
+	[0x1f] = { "general bus setup" },
+
+	/* TA Document 1999025 */
+	/* AV/C Descriptor Mechanism Specification Version 1.0 */
+	[0x0c] = { "create descriptor" },
+	[0x08] = { "open descriptor" },
+	[0x09] = { "read descriptor" },
+	[0x0a] = { "write descriptor" },
+	[0x05] = { "open info block" },
+	[0x06] = { "read info block" },
+	[0x07] = { "write info block" },
+	[0x0b] = { "search descriptor" },
+	[0x0d] = { "object number select" },
+
+	/* TA Document 1999015 */
+	/* AV/C Command Set for Rate Control of Isochronous Data Flow 1.0 */
+	[0xb3] = { "rate", {
+			{ "subfunction", 0, 8 },
+			{ "result", 8, 8 },
+			{ "plug_type", 16, 8 },
+			{ "plug_id", 16, 8 },
+		}
+	},
+
+	/* TA Document 1999008 */
+	/* AV/C Audio Subunit Specification 1.0 */
+	[0xb8] = { "function block" },
+
+	/* TA Document 2001001 */
+	/* AV/C Panel Subunit Specification 1.1 */
+	[0x7d] = { "gui update" },
+	[0x7e] = { "push gui data" },
+	[0x7f] = { "user action" },
+	[0x7c] = { "pass through" },
+
+	/* */
+	[0x26] = { "asynchronous connection" },
+};
+
+struct avc_frame {
+	uint32_t operand0:8;
+	uint32_t opcode:8;
+	uint32_t subunit_id:3;
+	uint32_t subunit_type:5;
+	uint32_t ctype:4;
+	uint32_t cts:4;
+};
+
+static void
+decode_avc(struct link_transaction *t)
+{
+	struct avc_frame *frame =
+	    (struct avc_frame *) t->request->packet.write_block.data;
+	const struct avc_opcode_info *info;
+	const char *name;
+	char buffer[32];
+	int i;
+
+	info = &opcode_info[frame->opcode];
+	if (info->name == NULL) {
+		snprintf(buffer, sizeof(buffer),
+			 "(unknown opcode 0x%02x)", frame->opcode);
+		name = buffer;
+	} else {
+		name = info->name;
+	}
+
+	printf("av/c %s, subunit_type=%s, subunit_id=%u, opcode=%s",
+	    ctype_names[frame->ctype], subunit_type_names[frame->subunit_type],
+	    frame->subunit_id, name);
+
+	for (i = 0; info->fields[i].name != NULL; i++)
+		printf(", %s", info->fields[i].name);
+
+	printf("\n");
+}
+
+int
+decode_fcp(struct link_transaction *t)
+{
+	struct avc_frame *frame =
+	    (struct avc_frame *) t->request->packet.write_block.data;
+	unsigned long long offset =
+	    ((unsigned long long) t->request->packet.common.offset_high << 32) |
+	    t->request->packet.common.offset_low;
+
+	if (t->request->packet.common.tcode != TCODE_WRITE_BLOCK_REQUEST)
+		return 0;
+
+	if (offset == CSR_FCP_COMMAND || offset == CSR_FCP_RESPONSE) {
+		switch (frame->cts) {
+		case 0x00:
+			decode_avc(t);
+			break;
+		case 0x01:
+			printf("cal fcp frame (cts=0x01)\n");
+			break;
+		case 0x02:
+			printf("ehs fcp frame (cts=0x02)\n");
+			break;
+		case 0x03:
+			printf("havi fcp frame (cts=0x03)\n");
+			break;
+		case 0x0e:
+			printf("vendor specific fcp frame (cts=0x0e)\n");
+			break;
+		case 0x0f:
+			printf("extended cts\n");
+			break;
+		default:
+			printf("reserved fcp frame (ctx=0x%02x)\n", frame->cts);
+			break;
+		}
+		return 1;
+	}
+
+	return 0;
+}
+
diff --git a/tools/firewire/list.h b/tools/firewire/list.h
new file mode 100644
index 000000000..6278d08e9
--- /dev/null
+++ b/tools/firewire/list.h
@@ -0,0 +1,63 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+struct list {
+	struct list *next, *prev;
+};
+
+static inline void
+list_init(struct list *list)
+{
+	list->next = list;
+	list->prev = list;
+}
+
+static inline int
+list_empty(struct list *list)
+{
+	return list->next == list;
+}
+
+static inline void
+list_insert(struct list *link, struct list *new_link)
+{
+	new_link->prev		= link->prev;
+	new_link->next		= link;
+	new_link->prev->next	= new_link;
+	new_link->next->prev	= new_link;
+}
+
+static inline void
+list_append(struct list *list, struct list *new_link)
+{
+	list_insert((struct list *)list, new_link);
+}
+
+static inline void
+list_prepend(struct list *list, struct list *new_link)
+{
+	list_insert(list->next, new_link);
+}
+
+static inline void
+list_remove(struct list *link)
+{
+	link->prev->next = link->next;
+	link->next->prev = link->prev;
+}
+
+#define list_entry(link, type, member) \
+	((type *)((char *)(link)-(unsigned long)(&((type *)0)->member)))
+
+#define list_head(list, type, member)		\
+	list_entry((list)->next, type, member)
+
+#define list_tail(list, type, member)		\
+	list_entry((list)->prev, type, member)
+
+#define list_next(elm, member)					\
+	list_entry((elm)->member.next, typeof(*elm), member)
+
+#define list_for_each_entry(pos, list, member)			\
+	for (pos = list_head(list, typeof(*pos), member);	\
+	     &pos->member != (list);				\
+	     pos = list_next(pos, member))
+
diff --git a/tools/firewire/nosy-dump.c b/tools/firewire/nosy-dump.c
new file mode 100644
index 000000000..9a906de3a
--- /dev/null
+++ b/tools/firewire/nosy-dump.c
@@ -0,0 +1,1022 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * nosy-dump - Interface to snoop mode driver for TI PCILynx 1394 controllers
+ * Copyright (C) 2002-2006 Kristian Høgsberg
+ */
+
+#include <byteswap.h>
+#include <endian.h>
+#include <fcntl.h>
+#include <linux/firewire-constants.h>
+#include <poll.h>
+#include <popt.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/ioctl.h>
+#include <sys/time.h>
+#include <termios.h>
+#include <unistd.h>
+
+#include "list.h"
+#include "nosy-dump.h"
+#include "nosy-user.h"
+
+enum {
+	PACKET_FIELD_DETAIL		= 0x01,
+	PACKET_FIELD_DATA_LENGTH	= 0x02,
+	/* Marks the fields we print in transaction view. */
+	PACKET_FIELD_TRANSACTION	= 0x04,
+};
+
+static void print_packet(uint32_t *data, size_t length);
+static void decode_link_packet(struct link_packet *packet, size_t length,
+			       int include_flags, int exclude_flags);
+static int run = 1;
+sig_t sys_sigint_handler;
+
+static char *option_nosy_device = "/dev/nosy";
+static char *option_view = "packet";
+static char *option_output;
+static char *option_input;
+static int option_hex;
+static int option_iso;
+static int option_cycle_start;
+static int option_version;
+static int option_verbose;
+
+enum {
+	VIEW_TRANSACTION,
+	VIEW_PACKET,
+	VIEW_STATS,
+};
+
+static const struct poptOption options[] = {
+	{
+		.longName	= "device",
+		.shortName	= 'd',
+		.argInfo	= POPT_ARG_STRING,
+		.arg		= &option_nosy_device,
+		.descrip	= "Path to nosy device.",
+		.argDescrip	= "DEVICE"
+	},
+	{
+		.longName	= "view",
+		.argInfo	= POPT_ARG_STRING,
+		.arg		= &option_view,
+		.descrip	= "Specify view of bus traffic: packet, transaction or stats.",
+		.argDescrip	= "VIEW"
+	},
+	{
+		.longName	= "hex",
+		.shortName	= 'x',
+		.argInfo	= POPT_ARG_NONE,
+		.arg		= &option_hex,
+		.descrip	= "Print each packet in hex.",
+	},
+	{
+		.longName	= "iso",
+		.argInfo	= POPT_ARG_NONE,
+		.arg		= &option_iso,
+		.descrip	= "Print iso packets.",
+	},
+	{
+		.longName	= "cycle-start",
+		.argInfo	= POPT_ARG_NONE,
+		.arg		= &option_cycle_start,
+		.descrip	= "Print cycle start packets.",
+	},
+	{
+		.longName	= "verbose",
+		.shortName	= 'v',
+		.argInfo	= POPT_ARG_NONE,
+		.arg		= &option_verbose,
+		.descrip	= "Verbose packet view.",
+	},
+	{
+		.longName	= "output",
+		.shortName	= 'o',
+		.argInfo	= POPT_ARG_STRING,
+		.arg		= &option_output,
+		.descrip	= "Log to output file.",
+		.argDescrip	= "FILENAME"
+	},
+	{
+		.longName	= "input",
+		.shortName	= 'i',
+		.argInfo	= POPT_ARG_STRING,
+		.arg		= &option_input,
+		.descrip	= "Decode log from file.",
+		.argDescrip	= "FILENAME"
+	},
+	{
+		.longName	= "version",
+		.argInfo	= POPT_ARG_NONE,
+		.arg		= &option_version,
+		.descrip	= "Specify print version info.",
+	},
+	POPT_AUTOHELP
+	POPT_TABLEEND
+};
+
+/* Allow all ^C except the first to interrupt the program in the usual way. */
+static void
+sigint_handler(int signal_num)
+{
+	if (run == 1) {
+		run = 0;
+		signal(SIGINT, SIG_DFL);
+	}
+}
+
+static struct subaction *
+subaction_create(uint32_t *data, size_t length)
+{
+	struct subaction *sa;
+
+	/* we put the ack in the subaction struct for easy access. */
+	sa = malloc(sizeof *sa - sizeof sa->packet + length);
+	if (!sa)
+		exit(EXIT_FAILURE);
+	sa->ack = data[length / 4 - 1];
+	sa->length = length;
+	memcpy(&sa->packet, data, length);
+
+	return sa;
+}
+
+static void
+subaction_destroy(struct subaction *sa)
+{
+	free(sa);
+}
+
+static struct list pending_transaction_list = {
+	&pending_transaction_list, &pending_transaction_list
+};
+
+static struct link_transaction *
+link_transaction_lookup(int request_node, int response_node, int tlabel)
+{
+	struct link_transaction *t;
+
+	list_for_each_entry(t, &pending_transaction_list, link) {
+		if (t->request_node == request_node &&
+		    t->response_node == response_node &&
+		    t->tlabel == tlabel)
+			return t;
+	}
+
+	t = malloc(sizeof *t);
+	if (!t)
+		exit(EXIT_FAILURE);
+	t->request_node = request_node;
+	t->response_node = response_node;
+	t->tlabel = tlabel;
+	list_init(&t->request_list);
+	list_init(&t->response_list);
+
+	list_append(&pending_transaction_list, &t->link);
+
+	return t;
+}
+
+static void
+link_transaction_destroy(struct link_transaction *t)
+{
+	struct subaction *sa;
+
+	while (!list_empty(&t->request_list)) {
+		sa = list_head(&t->request_list, struct subaction, link);
+		list_remove(&sa->link);
+		subaction_destroy(sa);
+	}
+	while (!list_empty(&t->response_list)) {
+		sa = list_head(&t->response_list, struct subaction, link);
+		list_remove(&sa->link);
+		subaction_destroy(sa);
+	}
+	free(t);
+}
+
+struct protocol_decoder {
+	const char *name;
+	int (*decode)(struct link_transaction *t);
+};
+
+static const struct protocol_decoder protocol_decoders[] = {
+	{ "FCP", decode_fcp }
+};
+
+static void
+handle_transaction(struct link_transaction *t)
+{
+	struct subaction *sa;
+	int i;
+
+	if (!t->request) {
+		printf("BUG in handle_transaction\n");
+		return;
+	}
+
+	for (i = 0; i < array_length(protocol_decoders); i++)
+		if (protocol_decoders[i].decode(t))
+			break;
+
+	/* HACK: decode only fcp right now. */
+	return;
+
+	decode_link_packet(&t->request->packet, t->request->length,
+			   PACKET_FIELD_TRANSACTION, 0);
+	if (t->response)
+		decode_link_packet(&t->response->packet, t->request->length,
+				   PACKET_FIELD_TRANSACTION, 0);
+	else
+		printf("[no response]");
+
+	if (option_verbose) {
+		list_for_each_entry(sa, &t->request_list, link)
+			print_packet((uint32_t *) &sa->packet, sa->length);
+		list_for_each_entry(sa, &t->response_list, link)
+			print_packet((uint32_t *) &sa->packet, sa->length);
+	}
+	printf("\r\n");
+
+	link_transaction_destroy(t);
+}
+
+static void
+clear_pending_transaction_list(void)
+{
+	struct link_transaction *t;
+
+	while (!list_empty(&pending_transaction_list)) {
+		t = list_head(&pending_transaction_list,
+			      struct link_transaction, link);
+		list_remove(&t->link);
+		link_transaction_destroy(t);
+		/* print unfinished transactions */
+	}
+}
+
+static const char * const tcode_names[] = {
+	[0x0] = "write_quadlet_request",	[0x6] = "read_quadlet_response",
+	[0x1] = "write_block_request",		[0x7] = "read_block_response",
+	[0x2] = "write_response",		[0x8] = "cycle_start",
+	[0x3] = "reserved",			[0x9] = "lock_request",
+	[0x4] = "read_quadlet_request",		[0xa] = "iso_data",
+	[0x5] = "read_block_request",		[0xb] = "lock_response",
+};
+
+static const char * const ack_names[] = {
+	[0x0] = "no ack",			[0x8] = "reserved (0x08)",
+	[0x1] = "ack_complete",			[0x9] = "reserved (0x09)",
+	[0x2] = "ack_pending",			[0xa] = "reserved (0x0a)",
+	[0x3] = "reserved (0x03)",		[0xb] = "reserved (0x0b)",
+	[0x4] = "ack_busy_x",			[0xc] = "reserved (0x0c)",
+	[0x5] = "ack_busy_a",			[0xd] = "ack_data_error",
+	[0x6] = "ack_busy_b",			[0xe] = "ack_type_error",
+	[0x7] = "reserved (0x07)",		[0xf] = "reserved (0x0f)",
+};
+
+static const char * const rcode_names[] = {
+	[0x0] = "complete",			[0x4] = "conflict_error",
+	[0x1] = "reserved (0x01)",		[0x5] = "data_error",
+	[0x2] = "reserved (0x02)",		[0x6] = "type_error",
+	[0x3] = "reserved (0x03)",		[0x7] = "address_error",
+};
+
+static const char * const retry_names[] = {
+	[0x0] = "retry_1",
+	[0x1] = "retry_x",
+	[0x2] = "retry_a",
+	[0x3] = "retry_b",
+};
+
+enum {
+	PACKET_RESERVED,
+	PACKET_REQUEST,
+	PACKET_RESPONSE,
+	PACKET_OTHER,
+};
+
+struct packet_info {
+	const char *name;
+	int type;
+	int response_tcode;
+	const struct packet_field *fields;
+	int field_count;
+};
+
+struct packet_field {
+	const char *name; /* Short name for field. */
+	int offset;	/* Location of field, specified in bits; */
+			/* negative means from end of packet.    */
+	int width;	/* Width of field, 0 means use data_length. */
+	int flags;	/* Show options. */
+	const char * const *value_names;
+};
+
+#define COMMON_REQUEST_FIELDS						\
+	{ "dest", 0, 16, PACKET_FIELD_TRANSACTION },			\
+	{ "tl", 16, 6 },						\
+	{ "rt", 22, 2, PACKET_FIELD_DETAIL, retry_names },		\
+	{ "tcode", 24, 4, PACKET_FIELD_TRANSACTION, tcode_names },	\
+	{ "pri", 28, 4, PACKET_FIELD_DETAIL },				\
+	{ "src", 32, 16, PACKET_FIELD_TRANSACTION },			\
+	{ "offs", 48, 48, PACKET_FIELD_TRANSACTION }
+
+#define COMMON_RESPONSE_FIELDS						\
+	{ "dest", 0, 16 },						\
+	{ "tl", 16, 6 },						\
+	{ "rt", 22, 2, PACKET_FIELD_DETAIL, retry_names },		\
+	{ "tcode", 24, 4, 0, tcode_names },				\
+	{ "pri", 28, 4, PACKET_FIELD_DETAIL },				\
+	{ "src", 32, 16 },						\
+	{ "rcode", 48, 4, PACKET_FIELD_TRANSACTION, rcode_names }
+
+static const struct packet_field read_quadlet_request_fields[] = {
+	COMMON_REQUEST_FIELDS,
+	{ "crc", 96, 32, PACKET_FIELD_DETAIL },
+	{ "ack", 156, 4, 0, ack_names },
+};
+
+static const struct packet_field read_quadlet_response_fields[] = {
+	COMMON_RESPONSE_FIELDS,
+	{ "data", 96, 32, PACKET_FIELD_TRANSACTION },
+	{ "crc", 128, 32, PACKET_FIELD_DETAIL },
+	{ "ack", 188, 4, 0, ack_names },
+};
+
+static const struct packet_field read_block_request_fields[] = {
+	COMMON_REQUEST_FIELDS,
+	{ "data_length", 96, 16, PACKET_FIELD_TRANSACTION },
+	{ "extended_tcode", 112, 16 },
+	{ "crc", 128, 32, PACKET_FIELD_DETAIL },
+	{ "ack", 188, 4, 0, ack_names },
+};
+
+static const struct packet_field block_response_fields[] = {
+	COMMON_RESPONSE_FIELDS,
+	{ "data_length", 96, 16, PACKET_FIELD_DATA_LENGTH },
+	{ "extended_tcode", 112, 16 },
+	{ "crc", 128, 32, PACKET_FIELD_DETAIL },
+	{ "data", 160, 0, PACKET_FIELD_TRANSACTION },
+	{ "crc", -64, 32, PACKET_FIELD_DETAIL },
+	{ "ack", -4, 4, 0, ack_names },
+};
+
+static const struct packet_field write_quadlet_request_fields[] = {
+	COMMON_REQUEST_FIELDS,
+	{ "data", 96, 32, PACKET_FIELD_TRANSACTION },
+	{ "ack", -4, 4, 0, ack_names },
+};
+
+static const struct packet_field block_request_fields[] = {
+	COMMON_REQUEST_FIELDS,
+	{ "data_length", 96, 16, PACKET_FIELD_DATA_LENGTH | PACKET_FIELD_TRANSACTION },
+	{ "extended_tcode", 112, 16, PACKET_FIELD_TRANSACTION },
+	{ "crc", 128, 32, PACKET_FIELD_DETAIL },
+	{ "data", 160, 0, PACKET_FIELD_TRANSACTION },
+	{ "crc", -64, 32, PACKET_FIELD_DETAIL },
+	{ "ack", -4, 4, 0, ack_names },
+};
+
+static const struct packet_field write_response_fields[] = {
+	COMMON_RESPONSE_FIELDS,
+	{ "reserved", 64, 32, PACKET_FIELD_DETAIL },
+	{ "ack", -4, 4, 0, ack_names },
+};
+
+static const struct packet_field iso_data_fields[] = {
+	{ "data_length", 0, 16, PACKET_FIELD_DATA_LENGTH },
+	{ "tag", 16, 2 },
+	{ "channel", 18, 6 },
+	{ "tcode", 24, 4, 0, tcode_names },
+	{ "sy", 28, 4 },
+	{ "crc", 32, 32, PACKET_FIELD_DETAIL },
+	{ "data", 64, 0 },
+	{ "crc", -64, 32, PACKET_FIELD_DETAIL },
+	{ "ack", -4, 4, 0, ack_names },
+};
+
+static const struct packet_info packet_info[] = {
+	{
+		.name		= "write_quadlet_request",
+		.type		= PACKET_REQUEST,
+		.response_tcode	= TCODE_WRITE_RESPONSE,
+		.fields		= write_quadlet_request_fields,
+		.field_count	= array_length(write_quadlet_request_fields)
+	},
+	{
+		.name		= "write_block_request",
+		.type		= PACKET_REQUEST,
+		.response_tcode	= TCODE_WRITE_RESPONSE,
+		.fields		= block_request_fields,
+		.field_count	= array_length(block_request_fields)
+	},
+	{
+		.name		= "write_response",
+		.type		= PACKET_RESPONSE,
+		.fields		= write_response_fields,
+		.field_count	= array_length(write_response_fields)
+	},
+	{
+		.name		= "reserved",
+		.type		= PACKET_RESERVED,
+	},
+	{
+		.name		= "read_quadlet_request",
+		.type		= PACKET_REQUEST,
+		.response_tcode	= TCODE_READ_QUADLET_RESPONSE,
+		.fields		= read_quadlet_request_fields,
+		.field_count	= array_length(read_quadlet_request_fields)
+	},
+	{
+		.name		= "read_block_request",
+		.type		= PACKET_REQUEST,
+		.response_tcode	= TCODE_READ_BLOCK_RESPONSE,
+		.fields		= read_block_request_fields,
+		.field_count	= array_length(read_block_request_fields)
+	},
+	{
+		.name		= "read_quadlet_response",
+		.type		= PACKET_RESPONSE,
+		.fields		= read_quadlet_response_fields,
+		.field_count	= array_length(read_quadlet_response_fields)
+	},
+	{
+		.name		= "read_block_response",
+		.type		= PACKET_RESPONSE,
+		.fields		= block_response_fields,
+		.field_count	= array_length(block_response_fields)
+	},
+	{
+		.name		= "cycle_start",
+		.type		= PACKET_OTHER,
+		.fields		= write_quadlet_request_fields,
+		.field_count	= array_length(write_quadlet_request_fields)
+	},
+	{
+		.name		= "lock_request",
+		.type		= PACKET_REQUEST,
+		.fields		= block_request_fields,
+		.field_count	= array_length(block_request_fields)
+	},
+	{
+		.name		= "iso_data",
+		.type		= PACKET_OTHER,
+		.fields		= iso_data_fields,
+		.field_count	= array_length(iso_data_fields)
+	},
+	{
+		.name		= "lock_response",
+		.type		= PACKET_RESPONSE,
+		.fields		= block_response_fields,
+		.field_count	= array_length(block_response_fields)
+	},
+};
+
+static int
+handle_request_packet(uint32_t *data, size_t length)
+{
+	struct link_packet *p = (struct link_packet *) data;
+	struct subaction *sa, *prev;
+	struct link_transaction *t;
+
+	t = link_transaction_lookup(p->common.source, p->common.destination,
+			p->common.tlabel);
+	sa = subaction_create(data, length);
+	t->request = sa;
+
+	if (!list_empty(&t->request_list)) {
+		prev = list_tail(&t->request_list,
+				 struct subaction, link);
+
+		if (!ACK_BUSY(prev->ack)) {
+			/*
+			 * error, we should only see ack_busy_* before the
+			 * ack_pending/ack_complete -- this is an ack_pending
+			 * instead (ack_complete would have finished the
+			 * transaction).
+			 */
+		}
+
+		if (prev->packet.common.tcode != sa->packet.common.tcode ||
+		    prev->packet.common.tlabel != sa->packet.common.tlabel) {
+			/* memcmp() ? */
+			/* error, these should match for retries. */
+		}
+	}
+
+	list_append(&t->request_list, &sa->link);
+
+	switch (sa->ack) {
+	case ACK_COMPLETE:
+		if (p->common.tcode != TCODE_WRITE_QUADLET_REQUEST &&
+		    p->common.tcode != TCODE_WRITE_BLOCK_REQUEST)
+			/* error, unified transactions only allowed for write */;
+		list_remove(&t->link);
+		handle_transaction(t);
+		break;
+
+	case ACK_NO_ACK:
+	case ACK_DATA_ERROR:
+	case ACK_TYPE_ERROR:
+		list_remove(&t->link);
+		handle_transaction(t);
+		break;
+
+	case ACK_PENDING:
+		/* request subaction phase over, wait for response. */
+		break;
+
+	case ACK_BUSY_X:
+	case ACK_BUSY_A:
+	case ACK_BUSY_B:
+		/* ok, wait for retry. */
+		/* check that retry protocol is respected. */
+		break;
+	}
+
+	return 1;
+}
+
+static int
+handle_response_packet(uint32_t *data, size_t length)
+{
+	struct link_packet *p = (struct link_packet *) data;
+	struct subaction *sa, *prev;
+	struct link_transaction *t;
+
+	t = link_transaction_lookup(p->common.destination, p->common.source,
+			p->common.tlabel);
+	if (list_empty(&t->request_list)) {
+		/* unsolicited response */
+	}
+
+	sa = subaction_create(data, length);
+	t->response = sa;
+
+	if (!list_empty(&t->response_list)) {
+		prev = list_tail(&t->response_list, struct subaction, link);
+
+		if (!ACK_BUSY(prev->ack)) {
+			/*
+			 * error, we should only see ack_busy_* before the
+			 * ack_pending/ack_complete
+			 */
+		}
+
+		if (prev->packet.common.tcode != sa->packet.common.tcode ||
+		    prev->packet.common.tlabel != sa->packet.common.tlabel) {
+			/* use memcmp() instead? */
+			/* error, these should match for retries. */
+		}
+	} else {
+		prev = list_tail(&t->request_list, struct subaction, link);
+		if (prev->ack != ACK_PENDING) {
+			/*
+			 * error, should not get response unless last request got
+			 * ack_pending.
+			 */
+		}
+
+		if (packet_info[prev->packet.common.tcode].response_tcode !=
+		    sa->packet.common.tcode) {
+			/* error, tcode mismatch */
+		}
+	}
+
+	list_append(&t->response_list, &sa->link);
+
+	switch (sa->ack) {
+	case ACK_COMPLETE:
+	case ACK_NO_ACK:
+	case ACK_DATA_ERROR:
+	case ACK_TYPE_ERROR:
+		list_remove(&t->link);
+		handle_transaction(t);
+		/* transaction complete, remove t from pending list. */
+		break;
+
+	case ACK_PENDING:
+		/* error for responses. */
+		break;
+
+	case ACK_BUSY_X:
+	case ACK_BUSY_A:
+	case ACK_BUSY_B:
+		/* no problem, wait for next retry */
+		break;
+	}
+
+	return 1;
+}
+
+static int
+handle_packet(uint32_t *data, size_t length)
+{
+	if (length == 0) {
+		printf("bus reset\r\n");
+		clear_pending_transaction_list();
+	} else if (length > sizeof(struct phy_packet)) {
+		struct link_packet *p = (struct link_packet *) data;
+
+		switch (packet_info[p->common.tcode].type) {
+		case PACKET_REQUEST:
+			return handle_request_packet(data, length);
+
+		case PACKET_RESPONSE:
+			return handle_response_packet(data, length);
+
+		case PACKET_OTHER:
+		case PACKET_RESERVED:
+			return 0;
+		}
+	}
+
+	return 1;
+}
+
+static unsigned int
+get_bits(struct link_packet *packet, int offset, int width)
+{
+	uint32_t *data = (uint32_t *) packet;
+	uint32_t index, shift, mask;
+
+	index = offset / 32 + 1;
+	shift = 32 - (offset & 31) - width;
+	mask = width == 32 ? ~0 : (1 << width) - 1;
+
+	return (data[index] >> shift) & mask;
+}
+
+#if __BYTE_ORDER == __LITTLE_ENDIAN
+#define byte_index(i) ((i) ^ 3)
+#elif __BYTE_ORDER == __BIG_ENDIAN
+#define byte_index(i) (i)
+#else
+#error unsupported byte order.
+#endif
+
+static void
+dump_data(unsigned char *data, int length)
+{
+	int i, print_length;
+
+	if (length > 128)
+		print_length = 128;
+	else
+		print_length = length;
+
+	for (i = 0; i < print_length; i++)
+		printf("%s%02hhx",
+		       (i % 4 == 0 && i != 0) ? " " : "",
+		       data[byte_index(i)]);
+
+	if (print_length < length)
+		printf(" (%d more bytes)", length - print_length);
+}
+
+static void
+decode_link_packet(struct link_packet *packet, size_t length,
+		   int include_flags, int exclude_flags)
+{
+	const struct packet_info *pi;
+	int data_length = 0;
+	int i;
+
+	pi = &packet_info[packet->common.tcode];
+
+	for (i = 0; i < pi->field_count; i++) {
+		const struct packet_field *f = &pi->fields[i];
+		int offset;
+
+		if (f->flags & exclude_flags)
+			continue;
+		if (include_flags && !(f->flags & include_flags))
+			continue;
+
+		if (f->offset < 0)
+			offset = length * 8 + f->offset - 32;
+		else
+			offset = f->offset;
+
+		if (f->value_names != NULL) {
+			uint32_t bits;
+
+			bits = get_bits(packet, offset, f->width);
+			printf("%s", f->value_names[bits]);
+		} else if (f->width == 0) {
+			printf("%s=[", f->name);
+			dump_data((unsigned char *) packet + (offset / 8 + 4), data_length);
+			printf("]");
+		} else {
+			unsigned long long bits;
+			int high_width, low_width;
+
+			if ((offset & ~31) != ((offset + f->width - 1) & ~31)) {
+				/* Bit field spans quadlet boundary. */
+				high_width = ((offset + 31) & ~31) - offset;
+				low_width = f->width - high_width;
+
+				bits = get_bits(packet, offset, high_width);
+				bits = (bits << low_width) |
+					get_bits(packet, offset + high_width, low_width);
+			} else {
+				bits = get_bits(packet, offset, f->width);
+			}
+
+			printf("%s=0x%0*llx", f->name, (f->width + 3) / 4, bits);
+
+			if (f->flags & PACKET_FIELD_DATA_LENGTH)
+				data_length = bits;
+		}
+
+		if (i < pi->field_count - 1)
+			printf(", ");
+	}
+}
+
+static void
+print_packet(uint32_t *data, size_t length)
+{
+	int i;
+
+	printf("%6u  ", data[0]);
+
+	if (length == 4) {
+		printf("bus reset");
+	} else if (length < sizeof(struct phy_packet)) {
+		printf("short packet: ");
+		for (i = 1; i < length / 4; i++)
+			printf("%s%08x", i == 0 ? "[" : " ", data[i]);
+		printf("]");
+
+	} else if (length == sizeof(struct phy_packet) && data[1] == ~data[2]) {
+		struct phy_packet *pp = (struct phy_packet *) data;
+
+		/* phy packet are 3 quadlets: the 1 quadlet payload,
+		 * the bitwise inverse of the payload and the snoop
+		 * mode ack */
+
+		switch (pp->common.identifier) {
+		case PHY_PACKET_CONFIGURATION:
+			if (!pp->phy_config.set_root && !pp->phy_config.set_gap_count) {
+				printf("ext phy config: phy_id=%02x", pp->phy_config.root_id);
+			} else {
+				printf("phy config:");
+				if (pp->phy_config.set_root)
+					printf(" set_root_id=%02x", pp->phy_config.root_id);
+				if (pp->phy_config.set_gap_count)
+					printf(" set_gap_count=%u", pp->phy_config.gap_count);
+			}
+			break;
+
+		case PHY_PACKET_LINK_ON:
+			printf("link-on packet, phy_id=%02x", pp->link_on.phy_id);
+			break;
+
+		case PHY_PACKET_SELF_ID:
+			if (pp->self_id.extended) {
+				printf("extended self id: phy_id=%02x, seq=%u",
+				       pp->ext_self_id.phy_id, pp->ext_self_id.sequence);
+			} else {
+				static const char * const speed_names[] = {
+					"S100", "S200", "S400", "BETA"
+				};
+				printf("self id: phy_id=%02x, link %s, gap_count=%u speed=%s%s%s",
+				       pp->self_id.phy_id,
+				       (pp->self_id.link_active ? "active" : "not active"),
+				       pp->self_id.gap_count,
+				       speed_names[pp->self_id.phy_speed],
+				       (pp->self_id.contender ? ", irm contender" : ""),
+				       (pp->self_id.initiated_reset ? ", initiator" : ""));
+			}
+			break;
+		default:
+			printf("unknown phy packet: ");
+			for (i = 1; i < length / 4; i++)
+				printf("%s%08x", i == 0 ? "[" : " ", data[i]);
+			printf("]");
+			break;
+		}
+	} else {
+		struct link_packet *packet = (struct link_packet *) data;
+
+		decode_link_packet(packet, length, 0,
+				   option_verbose ? 0 : PACKET_FIELD_DETAIL);
+	}
+
+	if (option_hex) {
+		printf("  [");
+		dump_data((unsigned char *) data + 4, length - 4);
+		printf("]");
+	}
+
+	printf("\r\n");
+}
+
+#define HIDE_CURSOR	"\033[?25l"
+#define SHOW_CURSOR	"\033[?25h"
+#define CLEAR		"\033[H\033[2J"
+
+static void
+print_stats(uint32_t *data, size_t length)
+{
+	static int bus_reset_count, short_packet_count, phy_packet_count;
+	static int tcode_count[16];
+	static struct timeval last_update;
+	struct timeval now;
+	int i;
+
+	if (length == 0)
+		bus_reset_count++;
+	else if (length < sizeof(struct phy_packet))
+		short_packet_count++;
+	else if (length == sizeof(struct phy_packet) && data[1] == ~data[2])
+		phy_packet_count++;
+	else {
+		struct link_packet *packet = (struct link_packet *) data;
+		tcode_count[packet->common.tcode]++;
+	}
+
+	gettimeofday(&now, NULL);
+	if (now.tv_sec <= last_update.tv_sec &&
+	    now.tv_usec < last_update.tv_usec + 500000)
+		return;
+
+	last_update = now;
+	printf(CLEAR HIDE_CURSOR
+	       "  bus resets              : %8d\n"
+	       "  short packets           : %8d\n"
+	       "  phy packets             : %8d\n",
+	       bus_reset_count, short_packet_count, phy_packet_count);
+
+	for (i = 0; i < array_length(packet_info); i++)
+		if (packet_info[i].type != PACKET_RESERVED)
+			printf("  %-24s: %8d\n", packet_info[i].name, tcode_count[i]);
+	printf(SHOW_CURSOR "\n");
+}
+
+static struct termios saved_attributes;
+
+static void
+reset_input_mode(void)
+{
+	tcsetattr(STDIN_FILENO, TCSANOW, &saved_attributes);
+}
+
+static void
+set_input_mode(void)
+{
+	struct termios tattr;
+
+	/* Make sure stdin is a terminal. */
+	if (!isatty(STDIN_FILENO)) {
+		fprintf(stderr, "Not a terminal.\n");
+		exit(EXIT_FAILURE);
+	}
+
+	/* Save the terminal attributes so we can restore them later. */
+	tcgetattr(STDIN_FILENO, &saved_attributes);
+	atexit(reset_input_mode);
+
+	/* Set the funny terminal modes. */
+	tcgetattr(STDIN_FILENO, &tattr);
+	tattr.c_lflag &= ~(ICANON|ECHO); /* Clear ICANON and ECHO. */
+	tattr.c_cc[VMIN] = 1;
+	tattr.c_cc[VTIME] = 0;
+	tcsetattr(STDIN_FILENO, TCSAFLUSH, &tattr);
+}
+
+int main(int argc, const char *argv[])
+{
+	uint32_t buf[128 * 1024];
+	uint32_t filter;
+	int length, retval, view;
+	int fd = -1;
+	FILE *output = NULL, *input = NULL;
+	poptContext con;
+	char c;
+	struct pollfd pollfds[2];
+
+	sys_sigint_handler = signal(SIGINT, sigint_handler);
+
+	con = poptGetContext(NULL, argc, argv, options, 0);
+	retval = poptGetNextOpt(con);
+	if (retval < -1) {
+		poptPrintUsage(con, stdout, 0);
+		return -1;
+	}
+
+	if (option_version) {
+		printf("dump tool for nosy sniffer, version %s\n", VERSION);
+		return 0;
+	}
+
+	if (__BYTE_ORDER != __LITTLE_ENDIAN)
+		fprintf(stderr, "warning: nosy has only been tested on little "
+			"endian machines\n");
+
+	if (option_input != NULL) {
+		input = fopen(option_input, "r");
+		if (input == NULL) {
+			fprintf(stderr, "Could not open %s, %m\n", option_input);
+			return -1;
+		}
+	} else {
+		fd = open(option_nosy_device, O_RDWR);
+		if (fd < 0) {
+			fprintf(stderr, "Could not open %s, %m\n", option_nosy_device);
+			return -1;
+		}
+		set_input_mode();
+	}
+
+	if (strcmp(option_view, "transaction") == 0)
+		view = VIEW_TRANSACTION;
+	else if (strcmp(option_view, "stats") == 0)
+		view = VIEW_STATS;
+	else
+		view = VIEW_PACKET;
+
+	if (option_output) {
+		output = fopen(option_output, "w");
+		if (output == NULL) {
+			fprintf(stderr, "Could not open %s, %m\n", option_output);
+			return -1;
+		}
+	}
+
+	setvbuf(stdout, NULL, _IOLBF, BUFSIZ);
+
+	filter = ~0;
+	if (!option_iso)
+		filter &= ~(1 << TCODE_STREAM_DATA);
+	if (!option_cycle_start)
+		filter &= ~(1 << TCODE_CYCLE_START);
+	if (view == VIEW_STATS)
+		filter = ~(1 << TCODE_CYCLE_START);
+
+	ioctl(fd, NOSY_IOC_FILTER, filter);
+
+	ioctl(fd, NOSY_IOC_START);
+
+	pollfds[0].fd = fd;
+	pollfds[0].events = POLLIN;
+	pollfds[1].fd = STDIN_FILENO;
+	pollfds[1].events = POLLIN;
+
+	while (run) {
+		if (input != NULL) {
+			if (fread(&length, sizeof length, 1, input) != 1)
+				return 0;
+			fread(buf, 1, length, input);
+		} else {
+			poll(pollfds, 2, -1);
+			if (pollfds[1].revents) {
+				read(STDIN_FILENO, &c, sizeof c);
+				switch (c) {
+				case 'q':
+					if (output != NULL)
+						fclose(output);
+					return 0;
+				}
+			}
+
+			if (pollfds[0].revents)
+				length = read(fd, buf, sizeof buf);
+			else
+				continue;
+		}
+
+		if (output != NULL) {
+			fwrite(&length, sizeof length, 1, output);
+			fwrite(buf, 1, length, output);
+		}
+
+		switch (view) {
+		case VIEW_TRANSACTION:
+			handle_packet(buf, length);
+			break;
+		case VIEW_PACKET:
+			print_packet(buf, length);
+			break;
+		case VIEW_STATS:
+			print_stats(buf, length);
+			break;
+		}
+	}
+
+	if (output != NULL)
+		fclose(output);
+
+	close(fd);
+
+	poptFreeContext(con);
+
+	return 0;
+}
diff --git a/tools/firewire/nosy-dump.h b/tools/firewire/nosy-dump.h
new file mode 100644
index 000000000..69e5e594f
--- /dev/null
+++ b/tools/firewire/nosy-dump.h
@@ -0,0 +1,174 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __nosy_dump_h__
+#define __nosy_dump_h__
+
+#define array_length(array) (sizeof(array) / sizeof(array[0]))
+
+#define ACK_NO_ACK   0x0
+#define ACK_DONE(a)  ((a >> 2) == 0)
+#define ACK_BUSY(a)  ((a >> 2) == 1)
+#define ACK_ERROR(a) ((a >> 2) == 3)
+
+#include <stdint.h>
+
+struct phy_packet {
+	uint32_t timestamp;
+	union {
+		struct {
+			uint32_t zero:24;
+			uint32_t phy_id:6;
+			uint32_t identifier:2;
+		} common, link_on;
+
+		struct {
+			uint32_t zero:16;
+			uint32_t gap_count:6;
+			uint32_t set_gap_count:1;
+			uint32_t set_root:1;
+			uint32_t root_id:6;
+			uint32_t identifier:2;
+		} phy_config;
+
+		struct {
+			uint32_t more_packets:1;
+			uint32_t initiated_reset:1;
+			uint32_t port2:2;
+			uint32_t port1:2;
+			uint32_t port0:2;
+			uint32_t power_class:3;
+			uint32_t contender:1;
+			uint32_t phy_delay:2;
+			uint32_t phy_speed:2;
+			uint32_t gap_count:6;
+			uint32_t link_active:1;
+			uint32_t extended:1;
+			uint32_t phy_id:6;
+			uint32_t identifier:2;
+		} self_id;
+
+		struct {
+			uint32_t more_packets:1;
+			uint32_t reserved1:1;
+			uint32_t porth:2;
+			uint32_t portg:2;
+			uint32_t portf:2;
+			uint32_t porte:2;
+			uint32_t portd:2;
+			uint32_t portc:2;
+			uint32_t portb:2;
+			uint32_t porta:2;
+			uint32_t reserved0:2;
+			uint32_t sequence:3;
+			uint32_t extended:1;
+			uint32_t phy_id:6;
+			uint32_t identifier:2;
+		} ext_self_id;
+	};
+	uint32_t inverted;
+	uint32_t ack;
+};
+
+#define TCODE_PHY_PACKET 0x10
+
+#define PHY_PACKET_CONFIGURATION 0x00
+#define PHY_PACKET_LINK_ON 0x01
+#define PHY_PACKET_SELF_ID 0x02
+
+struct link_packet {
+	uint32_t timestamp;
+	union {
+		struct {
+			uint32_t priority:4;
+			uint32_t tcode:4;
+			uint32_t rt:2;
+			uint32_t tlabel:6;
+			uint32_t destination:16;
+
+			uint32_t offset_high:16;
+			uint32_t source:16;
+
+			uint32_t offset_low;
+		} common;
+
+		struct {
+			uint32_t common[3];
+			uint32_t crc;
+		} read_quadlet;
+
+		struct {
+			uint32_t common[3];
+			uint32_t data;
+			uint32_t crc;
+		} read_quadlet_response;
+
+		struct {
+			uint32_t common[3];
+			uint32_t extended_tcode:16;
+			uint32_t data_length:16;
+			uint32_t crc;
+		} read_block;
+
+		struct {
+			uint32_t common[3];
+			uint32_t extended_tcode:16;
+			uint32_t data_length:16;
+			uint32_t crc;
+			uint32_t data[0];
+			/* crc and ack follows. */
+		} read_block_response;
+
+		struct {
+			uint32_t common[3];
+			uint32_t data;
+			uint32_t crc;
+		} write_quadlet;
+
+		struct {
+			uint32_t common[3];
+			uint32_t extended_tcode:16;
+			uint32_t data_length:16;
+			uint32_t crc;
+			uint32_t data[0];
+			/* crc and ack follows. */
+		} write_block;
+
+		struct {
+			uint32_t common[3];
+			uint32_t crc;
+		} write_response;
+
+		struct {
+			uint32_t common[3];
+			uint32_t data;
+			uint32_t crc;
+		} cycle_start;
+
+		struct {
+			uint32_t sy:4;
+			uint32_t tcode:4;
+			uint32_t channel:6;
+			uint32_t tag:2;
+			uint32_t data_length:16;
+
+			uint32_t crc;
+		} iso_data;
+	};
+};
+
+struct subaction {
+	uint32_t ack;
+	size_t length;
+	struct list link;
+	struct link_packet packet;
+};
+
+struct link_transaction {
+	int request_node, response_node, tlabel;
+	struct subaction *request, *response;
+	struct list request_list, response_list;
+	struct list link;
+};
+
+int decode_fcp(struct link_transaction *t);
+
+#endif /* __nosy_dump_h__ */