1 jelson 1.1 emlog -- the EMbedded-system LOG-device
2 Jeremy Elson - 6 June 2000
|
3 jelson 1.3 jelson@circlemud.org
|
4 jelson 1.1
|
5 jelson 1.5 Emlog web page:
6 http://www.circlemud.org/~jelson/software/emlog
7
|
8 jelson 1.1 --------------------------------------------------------------------------
9
10
11 What is emlog?
12 ==============
13
|
14 jelson 1.4 emlog is a Linux kernel module that makes it easy to access the most
15 recent (and *only* the most recent) output from a process. It works
16 just like "tail -f" on a log file, except that the storage required
17 never grows. This can be useful in embedded systems where there isn't
18 enough memory or disk space for keeping complete log files, but the
19 most recent debugging messages are sometimes needed (e.g., after an
20 error is observed).
|
21 jelson 1.1
|
22 jelson 1.6 The emlog kernel module implements simple character device driver.
23 The driver acts like a named pipe that has a finite, circular buffer.
24 The size of the buffer is easily configurable. As more data is
25 written into the buffer, the oldest data is discarded. A process that
26 reads from an emlog device will first read the existing buffer, then
27 see new text as it's written, similar to monitoring a log file using
28 "tail -f".
|
29 jelson 1.1
30
31 How is emlog used?
32 ==================
33
34 1: Configure, compile, and install emlog
35
36 First, decide which major number you would like to use for emlog.
37 This is configured in emlog.h using the constant
38 EMLOG_MAJOR_NUMBER. The default is 241, which is in the
39 "local/experimental use" range according to the kernel
40 documentation (similar to the 10/8 or 192.168/16 IP networks).
41 Setting the major number to 0 will cause the kernel to dynamically
42 assign a major number to emlog.
43
44 Next, compile using the Makefile provided. Typing 'make' should
45 generate a single object file, 'emlog.o'. Insert the module into
46 the kernel using the 'insmod' command; e.g. 'insmod emlog.o'.
47
48 2: Create device files for emlog
49
50 jelson 1.1 Next, you must use 'mknod' to create device files that your
51 processes can write to. The major number of the device files
52 should be whatever number you selected in Step 1 (e.g., 241). The
53 minor number is used to indicate the *size* of the ring buffer for
54 that device file, specified as the the number of kilobytes (e.g.,
55 1024 bytes). For example, to create an 8K buffer called 'testlog':
56
57 % mknod /tmp/testlog c 241 8
58
59 You can create as many devices as you like. Internally, emlog uses
60 the file's inode number to identify which buffer it refers to.
61
62 3: Write to and read from your new device file
63
64 Once the device file has been created, simply write to your device
65 file as you would any normal named pipe, e.g.
66
67 % echo hello > /tmp/testlog
68
69 Writes will never block because the buffer never runs out of space;
70 old data is simply overwritten by new data.
71 jelson 1.1
72 You can read from the log in the normal way, e.g. using cat. Note
73 that reads block, just like "tail -f", waiting for new log data.
74 For example:
75
76 % cat /tmp/testlog
77 hello [we immediately see the hello that we wrote in the previous step]
78 _ [... and here's the cursor. the 'cat' process is now
79 blocked, waiting for new input. New data will be displayed
80 as it is written to the device by other processes.]
81 ^C [use control-c, for example, to stop reading.]
82
83
84 4: Remove emlog when you're done
85
86 Type 'rmmod emlog' will remove the emlog kernel module and free all
|
87 jelson 1.3 associated buffers. This won't work until all emlog device files
88 are closed.
89
|
90 jelson 1.1
91
92 Other Usage Notes
93 =================
94
95 emlog will allocate a fixed-size buffer on behalf of a device file if
96 one of the following two conditions is true:
97
98 1- A process has the file open for reading or writing
99 2- A process has written text to the pipe that has not been read
100
101 In other words, buffers are persistent, even after a process closes
102 the pipe. If another process later reads the pipe, the text will
103 still be there. Note that it is possible (naturally) to fill virtual
104 memory by creating many such pipes, writing to all of them, and never
105 reading the data out of them. All buffers will be freed when the
106 emlog kernel module is removed.
107
108
109 Troubleshooting
110 ===============
111 jelson 1.1
112 Q: When I try insert the module using 'insmod', I get 'I/O error".
113
114 A: That probably means the major device number being registered by
115 emlog is already in use by another device driver. Try changing the
116 major device number in emlog.h (or, change it to 0 in order to get a
117 dynamically assigned major number).
118
119
120 Q: I'm seeing "I/O error" at a time *other* then when the module is
121 inserted.
122
123 A: Oops - you've found a bug in emlog. Please report it.
124
125
126 Q: When I try to access an emlog device file for reading or writing,
127 I get the error "no such device".
128
129 A: This probably means either that the emlog kernel module is not
130 loaded; or, that the major number of the device file does not match
131 the major number that emlog registered. To see which major number is
132 jelson 1.1 being used by emlog, type 'cat /proc/devices | grep emlog'.
133
134
135 Q: When I try to access an emlog device file for reading or writing,
136 I get the error "invalid argument".
137
138 A: The *minor* number of the emlog device file must be a number
139 between 1 and 128, representing the number of kilobytes (1,024 bytes)
140 that should be used for emlog's ring buffer. Make sure you're
141 specifying a valid minor number in your 'mknod' statement.
142
143
144 Q: I see "no memory" errors when I try opening new emlog files.
145
|
146 jelson 1.3 A: Looks like you're out of virtual memory, sport.
147
148
149 Q: When I try to remove the emlog driver ("rmmod emlog"), I get the
150 error "Device or resource busy".
151
152 A: That means a process is currently using an emlog device. You have
153 to wait until all processes close all emlog device files until the
154 driver can be removed. Try using "lsof" to see which files are in use
155 by which processes.
156
157
158 Q: You've made my computer crash.
159
160 A: Sorry. If you can reproduce the problem I'll try to fix it.
|
161 jelson 1.1
162
163 Known Bugs
164 ==========
165
|
166 jelson 1.6 emlog identifies buffers based solely on the inode number of the
|
167 jelson 1.1 device file being accessed. If two device files on two different
|
168 jelson 1.3 filesystems happen to have the same inode number, they will share the
|
169 jelson 1.1 same buffer, as if they were the same device file.
170
|
171 jelson 1.3 Currently, the poll() function for emlog files is unimplemented.
172 Therefore, using the select() function with an emlog file will not
173 work. (However, non-blocking reads DO work, e.g. by setting
174 O_NONBLOCK using ioctl()).
175
176 Bug reports, patches, complaints, praise, and submissions of Central
177 Services Form 27B/6, are welcomed by the author (Jeremy Elson,
178 <jelson@circlemud.org>.
|
179 jelson 1.1
180
181 Who wrote emlog, and why?
182 =========================
183
184 Emlog was written by Jeremy Elson <jelson@circlemud.org> at the
185 University of Southern California's Information Sciences Institute as
186 part of the SCADDS project <http://www.isi.edu/scadds>. SCADDS is an
187 embedded systems research project. We use small PC/104-bus-based
188 single-board-PCs using Linux. We wanted to save the debugging output
189 from certain processes, but since these things have 16MB of disk space
190 and 32MB of RAM, keeping complete log files was not an option. These
191 tiny nodes do have serial ports running PPP, though, so it's possible
192 to walk over to a node with a laptop, plug in a serial cable, and then
193 telnet into the box. Using emlog, we can always keep the most recent
194 debug messages from our processes; in case of an error, we can plug in
195 a debug console and see what went wrong.
196
|
197 jelson 1.3 This work was supported by DARPA under grant No. DABT63-99-1-0011 as
198 part of the SCADDS project, and was also made possible in part due to
199 support from Cisco Systems.
200
|