1lockfile API 2============ 3 4The lockfile API serves two purposes: 5 6* Mutual exclusion and atomic file updates. When we want to change a 7 file, we create a lockfile `<filename>.lock`, write the new file 8 contents into it, and then rename the lockfile to its final 9 destination `<filename>`. We create the `<filename>.lock` file with 10 `O_CREAT|O_EXCL` so that we can notice and fail if somebody else has 11 already locked the file, then atomically rename the lockfile to its 12 final destination to commit the changes and unlock the file. 13 14* Automatic cruft removal. If the program exits after we lock a file 15 but before the changes have been committed, we want to make sure 16 that we remove the lockfile. This is done by remembering the 17 lockfiles we have created in a linked list and setting up an 18 `atexit(3)` handler and a signal handler that clean up the 19 lockfiles. This mechanism ensures that outstanding lockfiles are 20 cleaned up if the program exits (including when `die()` is called) 21 or if the program dies on a signal. 22 23Please note that lockfiles only block other writers. Readers do not 24block, but they are guaranteed to see either the old contents of the 25file or the new contents of the file (assuming that the filesystem 26implements `rename(2)` atomically). 27 28 29Calling sequence 30---------------- 31 32The caller: 33 34* Allocates a `struct lock_file` either as a static variable or on the 35 heap, initialized to zeros. Once you use the structure to call the 36 `hold_lock_file_*` family of functions, it belongs to the lockfile 37 subsystem and its storage must remain valid throughout the life of 38 the program (i.e. you cannot use an on-stack variable to hold this 39 structure). 40 41* Attempts to create a lockfile by passing that variable and the path 42 of the final destination (e.g. `$GIT_DIR/index`) to 43 `hold_lock_file_for_update` or `hold_lock_file_for_append`. 44 45* Writes new content for the destination file by writing to the file 46 descriptor returned by those functions (also available via 47 `lock->fd`). 48 49When finished writing, the caller can: 50 51* Close the file descriptor and rename the lockfile to its final 52 destination by calling `commit_lock_file`. 53 54* Close the file descriptor and remove the lockfile by calling 55 `rollback_lock_file`. 56 57* Close the file descriptor without removing or renaming the lockfile 58 by calling `close_lock_file`, and later call `commit_lock_file`, 59 `rollback_lock_file`, or `reopen_lock_file`. 60 61Even after the lockfile is committed or rolled back, the `lock_file` 62object must not be freed or altered by the caller. However, it may be 63reused; just pass it to another call of `hold_lock_file_for_update` or 64`hold_lock_file_for_append`. 65 66If the program exits before you have called one of `commit_lock_file`, 67`rollback_lock_file`, or `close_lock_file`, an `atexit(3)` handler 68will close and remove the lockfile, rolling back any uncommitted 69changes. 70 71If you need to close the file descriptor you obtained from a 72`hold_lock_file_*` function yourself, do so by calling 73`close_lock_file`. You should never call `close(2)` yourself! 74Otherwise the `struct lock_file` structure would still think that the 75file descriptor needs to be closed, and a later call to 76`commit_lock_file` or `rollback_lock_file` or program exit would 77result in duplicate calls to `close(2)`. Worse yet, if you `close(2)` 78and then later open another file descriptor for a completely different 79purpose, then a call to `commit_lock_file` or `rollback_lock_file` 80might close that unrelated file descriptor. 81 82 83Error handling 84-------------- 85 86The `hold_lock_file_*` functions return a file descriptor on success 87or -1 on failure (unless `LOCK_DIE_ON_ERROR` is used; see below). On 88errors, `errno` describes the reason for failure. Errors can be 89reported by passing `errno` to one of the following helper functions: 90 91unable_to_lock_message:: 92 93 Append an appropriate error message to a `strbuf`. 94 95unable_to_lock_error:: 96 97 Emit an appropriate error message using `error()`. 98 99unable_to_lock_die:: 100 101 Emit an appropriate error message and `die()`. 102 103Similarly, `commit_lock_file` and `close_lock_file` return 0 on 104success. On failure they set `errno` appropriately, do their best to 105roll back the lockfile, and return -1. 106 107 108Flags 109----- 110 111The following flags can be passed to `hold_lock_file_for_update` or 112`hold_lock_file_for_append`: 113 114LOCK_NODEREF:: 115 116 Usually symbolic links in the destination path are resolved 117 and the lockfile is created by adding ".lock" to the resolved 118 path. If `LOCK_NODEREF` is set, then the lockfile is created 119 by adding ".lock" to the path argument itself. This option is 120 used, for example, when locking a symbolic reference, which 121 for backwards-compatibility reasons can be a symbolic link 122 containing the name of the referred-to-reference. 123 124LOCK_DIE_ON_ERROR:: 125 126 If a lock is already taken for the file, `die()` with an error 127 message. If this option is not specified, trying to lock a 128 file that is already locked returns -1 to the caller. 129 130 131The functions 132------------- 133 134hold_lock_file_for_update:: 135 136 Take a pointer to `struct lock_file`, the path of the file to 137 be locked (e.g. `$GIT_DIR/index`) and a flags argument (see 138 above). Attempt to create a lockfile for the destination and 139 return the file descriptor for writing to the file. 140 141hold_lock_file_for_append:: 142 143 Like `hold_lock_file_for_update`, but before returning copy 144 the existing contents of the file (if any) to the lockfile and 145 position its write pointer at the end of the file. 146 147commit_lock_file:: 148 149 Take a pointer to the `struct lock_file` initialized with an 150 earlier call to `hold_lock_file_for_update` or 151 `hold_lock_file_for_append`, close the file descriptor, and 152 rename the lockfile to its final destination. Return 0 upon 153 success. On failure, roll back the lock file and return -1, 154 with `errno` set to the value from the failing call to 155 `close(2)` or `rename(2)`. It is a bug to call 156 `commit_lock_file` for a `lock_file` object that is not 157 currently locked. 158 159rollback_lock_file:: 160 161 Take a pointer to the `struct lock_file` initialized with an 162 earlier call to `hold_lock_file_for_update` or 163 `hold_lock_file_for_append`, close the file descriptor and 164 remove the lockfile. It is a NOOP to call 165 `rollback_lock_file()` for a `lock_file` object that has 166 already been committed or rolled back. 167 168close_lock_file:: 169 170 Take a pointer to the `struct lock_file` initialized with an 171 earlier call to `hold_lock_file_for_update` or 172 `hold_lock_file_for_append`, and close the file descriptor. 173 Return 0 upon success. On failure to `close(2)`, return a 174 negative value and roll back the lock file. Usually 175 `commit_lock_file` or `rollback_lock_file` should eventually 176 be called if `close_lock_file` succeeds. 177 178reopen_lock_file:: 179 180 Re-open a lockfile that has been closed (using 181 `close_lock_file`) but not yet committed or rolled back. This 182 can be used to implement a sequence of operations like the 183 following: 184 185 * Lock file. 186 187 * Write new contents to lockfile, then `close_lock_file` to 188 cause the contents to be written to disk. 189 190 * Pass the name of the lockfile to another program to allow it 191 (and nobody else) to inspect the contents you wrote, while 192 still holding the lock yourself. 193 194 * `reopen_lock_file` to reopen the lockfile. Make further 195 updates to the contents. 196 197 * `commit_lock_file` to make the final version permanent.