1 In-kernel memory-mapped I/O tracing
4 Home page and links to optional user space tools:
6 http://nouveau.freedesktop.org/wiki/MmioTrace
8 MMIO tracing was originally developed by Intel around 2003 for their Fault
9 Injection Test Harness. In Dec 2006 - Jan 2007, using the code from Intel,
10 Jeff Muizelaar created a tool for tracing MMIO accesses with the Nouveau
11 project in mind. Since then many people have contributed.
13 Mmiotrace was built for reverse engineering any memory-mapped IO device with
14 the Nouveau project as the first real user. Only x86 and x86_64 architectures
17 Out-of-tree mmiotrace was originally modified for mainline inclusion and
18 ftrace framework by Pekka Paalanen <pq@iki.fi>.
24 Mmiotrace feature is compiled in by the CONFIG_MMIOTRACE option. Tracing is
25 disabled by default, so it is safe to have this set to yes. SMP systems are
26 supported, but tracing is unreliable and may miss events if more than one CPU
27 is on-line, therefore mmiotrace takes all but one CPU off-line during run-time
28 activation. You can re-enable CPUs by hand, but you have been warned, there
29 is no way to automatically detect if you are losing events due to CPUs racing.
35 $ mount -t debugfs debugfs /debug
36 $ echo mmiotrace > /debug/tracing/current_tracer
37 $ cat /debug/tracing/trace_pipe > mydump.txt &
39 $ echo "X is up" > /debug/tracing/trace_marker
40 $ echo nop > /debug/tracing/current_tracer
41 Check for lost events.
47 Make sure debugfs is mounted to /debug. If not, (requires root privileges)
48 $ mount -t debugfs debugfs /debug
50 Check that the driver you are about to trace is not loaded.
52 Activate mmiotrace (requires root privileges):
53 $ echo mmiotrace > /debug/tracing/current_tracer
55 Start storing the trace:
56 $ cat /debug/tracing/trace_pipe > mydump.txt &
57 The 'cat' process should stay running (sleeping) in the background.
59 Load the driver you want to trace and use it. Mmiotrace will only catch MMIO
60 accesses to areas that are ioremapped while mmiotrace is active.
62 During tracing you can place comments (markers) into the trace by
63 $ echo "X is up" > /debug/tracing/trace_marker
64 This makes it easier to see which part of the (huge) trace corresponds to
65 which action. It is recommended to place descriptive markers about what you
68 Shut down mmiotrace (requires root privileges):
69 $ echo nop > /debug/tracing/current_tracer
70 The 'cat' process exits. If it does not, kill it by issuing 'fg' command and
73 Check that mmiotrace did not lose events due to a buffer filling up. Either
74 $ grep -i lost mydump.txt
75 which tells you exactly how many events were lost, or use
77 to view your kernel log and look for "mmiotrace has lost events" warning. If
78 events were lost, the trace is incomplete. You should enlarge the buffers and
79 try again. Buffers are enlarged by first seeing how large the current buffers
81 $ cat /debug/tracing/trace_entries
82 gives you a number. Approximately double this number and write it back, for
84 $ echo 0 > /debug/tracing/tracing_enabled
85 $ echo 128000 > /debug/tracing/trace_entries
86 $ echo 1 > /debug/tracing/tracing_enabled
87 Then start again from the top.
89 If you are doing a trace for a driver project, e.g. Nouveau, you should also
90 do the following before sending your results:
91 $ lspci -vvv > lspci.txt
93 $ tar zcf pciid-nick-mmiotrace.tar.gz mydump.txt lspci.txt dmesg.txt
94 and then send the .tar.gz file. The trace compresses considerably. Replace
95 "pciid" and "nick" with the PCI ID or model name of your piece of hardware
96 under investigation and your nick name.
102 Access to hardware IO-memory is gained by mapping addresses from PCI bus by
103 calling one of the ioremap_*() functions. Mmiotrace is hooked into the
104 __ioremap() function and gets called whenever a mapping is created. Mapping is
105 an event that is recorded into the trace log. Note, that ISA range mappings
106 are not caught, since the mapping always exists and is returned directly.
108 MMIO accesses are recorded via page faults. Just before __ioremap() returns,
109 the mapped pages are marked as not present. Any access to the pages causes a
110 fault. The page fault handler calls mmiotrace to handle the fault. Mmiotrace
111 marks the page present, sets TF flag to achieve single stepping and exits the
112 fault handler. The instruction that faulted is executed and debug trap is
113 entered. Here mmiotrace again marks the page as not present. The instruction
114 is decoded to get the type of operation (read/write), data width and the value
115 read or written. These are stored to the trace log.
117 Setting the page present in the page fault handler has a race condition on SMP
118 machines. During the single stepping other CPUs may run freely on that page
119 and events can be missed without a notice. Re-enabling other CPUs during
120 tracing is discouraged.
126 The raw log is text and easily filtered with e.g. grep and awk. One record is
127 one line in the log. A record starts with a keyword, followed by keyword
128 dependant arguments. Arguments are separated by a space, or continue until the
129 end of line. The format for version 20070824 is as follows:
131 Explanation Keyword Space separated arguments
132 ---------------------------------------------------------------------------
134 read event R width, timestamp, map id, physical, value, PC, PID
135 write event W width, timestamp, map id, physical, value, PC, PID
136 ioremap event MAP timestamp, map id, physical, virtual, length, PC, PID
137 iounmap event UNMAP timestamp, map id, PC, PID
138 marker MARK timestamp, text
139 version VERSION the string "20070824"
140 info for reader LSPCI one line from lspci -v
141 PCI address map PCIDEV space separated /proc/bus/pci/devices data
142 unk. opcode UNKNOWN timestamp, map id, physical, data, PC, PID
144 Timestamp is in seconds with decimals. Physical is a PCI bus address, virtual
145 is a kernel virtual address. Width is the data width in bytes and value is the
146 data value. Map id is an arbitrary id number identifying the mapping that was
147 used in an operation. PC is the program counter and PID is process id. PC is
148 zero if it is not recorded. PID is always zero as tracing MMIO accesses
149 originating in user space memory is not yet supported.
151 For instance, the following awk filter will pass all 32-bit writes that target
152 physical addresses in the range [0xfb73ce40, 0xfb800000[
154 $ awk '/W 4 / { adr=strtonum($5); if (adr >= 0xfb73ce40 &&
155 adr < 0xfb800000) print; }'
161 The user space tools include utilities for:
162 - replacing numeric addresses and values with hardware register names
163 - replaying MMIO logs, i.e., re-executing the recorded writes