1The seq_file interface 2 3 Copyright 2003 Jonathan Corbet <corbet@lwn.net> 4 This file is originally from the LWN.net Driver Porting series at 5 http://lwn.net/Articles/driver-porting/ 6 7 8There are numerous ways for a device driver (or other kernel component) to 9provide information to the user or system administrator. One useful 10technique is the creation of virtual files, in debugfs, /proc or elsewhere. 11Virtual files can provide human-readable output that is easy to get at 12without any special utility programs; they can also make life easier for 13script writers. It is not surprising that the use of virtual files has 14grown over the years. 15 16Creating those files correctly has always been a bit of a challenge, 17however. It is not that hard to make a virtual file which returns a 18string. But life gets trickier if the output is long - anything greater 19than an application is likely to read in a single operation. Handling 20multiple reads (and seeks) requires careful attention to the reader's 21position within the virtual file - that position is, likely as not, in the 22middle of a line of output. The kernel has traditionally had a number of 23implementations that got this wrong. 24 25The 2.6 kernel contains a set of functions (implemented by Alexander Viro) 26which are designed to make it easy for virtual file creators to get it 27right. 28 29The seq_file interface is available via <linux/seq_file.h>. There are 30three aspects to seq_file: 31 32 * An iterator interface which lets a virtual file implementation 33 step through the objects it is presenting. 34 35 * Some utility functions for formatting objects for output without 36 needing to worry about things like output buffers. 37 38 * A set of canned file_operations which implement most operations on 39 the virtual file. 40 41We'll look at the seq_file interface via an extremely simple example: a 42loadable module which creates a file called /proc/sequence. The file, when 43read, simply produces a set of increasing integer values, one per line. The 44sequence will continue until the user loses patience and finds something 45better to do. The file is seekable, in that one can do something like the 46following: 47 48 dd if=/proc/sequence of=out1 count=1 49 dd if=/proc/sequence skip=1 of=out2 count=1 50 51Then concatenate the output files out1 and out2 and get the right 52result. Yes, it is a thoroughly useless module, but the point is to show 53how the mechanism works without getting lost in other details. (Those 54wanting to see the full source for this module can find it at 55http://lwn.net/Articles/22359/). 56 57 58The iterator interface 59 60Modules implementing a virtual file with seq_file must implement a simple 61iterator object that allows stepping through the data of interest. 62Iterators must be able to move to a specific position - like the file they 63implement - but the interpretation of that position is up to the iterator 64itself. A seq_file implementation that is formatting firewall rules, for 65example, could interpret position N as the Nth rule in the chain. 66Positioning can thus be done in whatever way makes the most sense for the 67generator of the data, which need not be aware of how a position translates 68to an offset in the virtual file. The one obvious exception is that a 69position of zero should indicate the beginning of the file. 70 71The /proc/sequence iterator just uses the count of the next number it 72will output as its position. 73 74Four functions must be implemented to make the iterator work. The first, 75called start() takes a position as an argument and returns an iterator 76which will start reading at that position. For our simple sequence example, 77the start() function looks like: 78 79 static void *ct_seq_start(struct seq_file *s, loff_t *pos) 80 { 81 loff_t *spos = kmalloc(sizeof(loff_t), GFP_KERNEL); 82 if (! spos) 83 return NULL; 84 *spos = *pos; 85 return spos; 86 } 87 88The entire data structure for this iterator is a single loff_t value 89holding the current position. There is no upper bound for the sequence 90iterator, but that will not be the case for most other seq_file 91implementations; in most cases the start() function should check for a 92"past end of file" condition and return NULL if need be. 93 94For more complicated applications, the private field of the seq_file 95structure can be used. There is also a special value which can be returned 96by the start() function called SEQ_START_TOKEN; it can be used if you wish 97to instruct your show() function (described below) to print a header at the 98top of the output. SEQ_START_TOKEN should only be used if the offset is 99zero, however. 100 101The next function to implement is called, amazingly, next(); its job is to 102move the iterator forward to the next position in the sequence. The 103example module can simply increment the position by one; more useful 104modules will do what is needed to step through some data structure. The 105next() function returns a new iterator, or NULL if the sequence is 106complete. Here's the example version: 107 108 static void *ct_seq_next(struct seq_file *s, void *v, loff_t *pos) 109 { 110 loff_t *spos = v; 111 *pos = ++*spos; 112 return spos; 113 } 114 115The stop() function is called when iteration is complete; its job, of 116course, is to clean up. If dynamic memory is allocated for the iterator, 117stop() is the place to free it. 118 119 static void ct_seq_stop(struct seq_file *s, void *v) 120 { 121 kfree(v); 122 } 123 124Finally, the show() function should format the object currently pointed to 125by the iterator for output. The example module's show() function is: 126 127 static int ct_seq_show(struct seq_file *s, void *v) 128 { 129 loff_t *spos = v; 130 seq_printf(s, "%lld\n", (long long)*spos); 131 return 0; 132 } 133 134If all is well, the show() function should return zero. A negative error 135code in the usual manner indicates that something went wrong; it will be 136passed back to user space. This function can also return SEQ_SKIP, which 137causes the current item to be skipped; if the show() function has already 138generated output before returning SEQ_SKIP, that output will be dropped. 139 140We will look at seq_printf() in a moment. But first, the definition of the 141seq_file iterator is finished by creating a seq_operations structure with 142the four functions we have just defined: 143 144 static const struct seq_operations ct_seq_ops = { 145 .start = ct_seq_start, 146 .next = ct_seq_next, 147 .stop = ct_seq_stop, 148 .show = ct_seq_show 149 }; 150 151This structure will be needed to tie our iterator to the /proc file in 152a little bit. 153 154It's worth noting that the iterator value returned by start() and 155manipulated by the other functions is considered to be completely opaque by 156the seq_file code. It can thus be anything that is useful in stepping 157through the data to be output. Counters can be useful, but it could also be 158a direct pointer into an array or linked list. Anything goes, as long as 159the programmer is aware that things can happen between calls to the 160iterator function. However, the seq_file code (by design) will not sleep 161between the calls to start() and stop(), so holding a lock during that time 162is a reasonable thing to do. The seq_file code will also avoid taking any 163other locks while the iterator is active. 164 165 166Formatted output 167 168The seq_file code manages positioning within the output created by the 169iterator and getting it into the user's buffer. But, for that to work, that 170output must be passed to the seq_file code. Some utility functions have 171been defined which make this task easy. 172 173Most code will simply use seq_printf(), which works pretty much like 174printk(), but which requires the seq_file pointer as an argument. It is 175common to ignore the return value from seq_printf(), but a function 176producing complicated output may want to check that value and quit if 177something non-zero is returned; an error return means that the seq_file 178buffer has been filled and further output will be discarded. 179 180For straight character output, the following functions may be used: 181 182 int seq_putc(struct seq_file *m, char c); 183 int seq_puts(struct seq_file *m, const char *s); 184 int seq_escape(struct seq_file *m, const char *s, const char *esc); 185 186The first two output a single character and a string, just like one would 187expect. seq_escape() is like seq_puts(), except that any character in s 188which is in the string esc will be represented in octal form in the output. 189 190There is also a pair of functions for printing filenames: 191 192 int seq_path(struct seq_file *m, struct path *path, char *esc); 193 int seq_path_root(struct seq_file *m, struct path *path, 194 struct path *root, char *esc) 195 196Here, path indicates the file of interest, and esc is a set of characters 197which should be escaped in the output. A call to seq_path() will output 198the path relative to the current process's filesystem root. If a different 199root is desired, it can be used with seq_path_root(). Note that, if it 200turns out that path cannot be reached from root, the value of root will be 201changed in seq_file_root() to a root which *does* work. 202 203 204Making it all work 205 206So far, we have a nice set of functions which can produce output within the 207seq_file system, but we have not yet turned them into a file that a user 208can see. Creating a file within the kernel requires, of course, the 209creation of a set of file_operations which implement the operations on that 210file. The seq_file interface provides a set of canned operations which do 211most of the work. The virtual file author still must implement the open() 212method, however, to hook everything up. The open function is often a single 213line, as in the example module: 214 215 static int ct_open(struct inode *inode, struct file *file) 216 { 217 return seq_open(file, &ct_seq_ops); 218 } 219 220Here, the call to seq_open() takes the seq_operations structure we created 221before, and gets set up to iterate through the virtual file. 222 223On a successful open, seq_open() stores the struct seq_file pointer in 224file->private_data. If you have an application where the same iterator can 225be used for more than one file, you can store an arbitrary pointer in the 226private field of the seq_file structure; that value can then be retrieved 227by the iterator functions. 228 229The other operations of interest - read(), llseek(), and release() - are 230all implemented by the seq_file code itself. So a virtual file's 231file_operations structure will look like: 232 233 static const struct file_operations ct_file_ops = { 234 .owner = THIS_MODULE, 235 .open = ct_open, 236 .read = seq_read, 237 .llseek = seq_lseek, 238 .release = seq_release 239 }; 240 241There is also a seq_release_private() which passes the contents of the 242seq_file private field to kfree() before releasing the structure. 243 244The final step is the creation of the /proc file itself. In the example 245code, that is done in the initialization code in the usual way: 246 247 static int ct_init(void) 248 { 249 struct proc_dir_entry *entry; 250 251 proc_create("sequence", 0, NULL, &ct_file_ops); 252 return 0; 253 } 254 255 module_init(ct_init); 256 257And that is pretty much it. 258 259 260seq_list 261 262If your file will be iterating through a linked list, you may find these 263routines useful: 264 265 struct list_head *seq_list_start(struct list_head *head, 266 loff_t pos); 267 struct list_head *seq_list_start_head(struct list_head *head, 268 loff_t pos); 269 struct list_head *seq_list_next(void *v, struct list_head *head, 270 loff_t *ppos); 271 272These helpers will interpret pos as a position within the list and iterate 273accordingly. Your start() and next() functions need only invoke the 274seq_list_* helpers with a pointer to the appropriate list_head structure. 275 276 277The extra-simple version 278 279For extremely simple virtual files, there is an even easier interface. A 280module can define only the show() function, which should create all the 281output that the virtual file will contain. The file's open() method then 282calls: 283 284 int single_open(struct file *file, 285 int (*show)(struct seq_file *m, void *p), 286 void *data); 287 288When output time comes, the show() function will be called once. The data 289value given to single_open() can be found in the private field of the 290seq_file structure. When using single_open(), the programmer should use 291single_release() instead of seq_release() in the file_operations structure 292to avoid a memory leak. 293