2 * Copyright (c) 2011, Google Inc.
4 #include "git-compat-util.h"
6 #include "bulk-checkin.h"
9 #include "repository.h"
10 #include "csum-file.h"
13 #include "string-list.h"
14 #include "tmp-objdir.h"
16 #include "object-store.h"
18 static int odb_transaction_nesting
;
20 static struct tmp_objdir
*bulk_fsync_objdir
;
22 static struct bulk_checkin_packfile
{
26 struct pack_idx_option pack_idx_opts
;
28 struct pack_idx_entry
**written
;
29 uint32_t alloc_written
;
31 } bulk_checkin_packfile
;
33 static void finish_tmp_packfile(struct strbuf
*basename
,
34 const char *pack_tmp_name
,
35 struct pack_idx_entry
**written_list
,
37 struct pack_idx_option
*pack_idx_opts
,
40 char *idx_tmp_name
= NULL
;
42 stage_tmp_packfiles(basename
, pack_tmp_name
, written_list
, nr_written
,
43 NULL
, pack_idx_opts
, hash
, &idx_tmp_name
);
44 rename_tmp_packfile_idx(basename
, &idx_tmp_name
);
49 static void flush_bulk_checkin_packfile(struct bulk_checkin_packfile
*state
)
51 unsigned char hash
[GIT_MAX_RAWSZ
];
52 struct strbuf packname
= STRBUF_INIT
;
58 if (state
->nr_written
== 0) {
60 unlink(state
->pack_tmp_name
);
62 } else if (state
->nr_written
== 1) {
63 finalize_hashfile(state
->f
, hash
, FSYNC_COMPONENT_PACK
,
64 CSUM_HASH_IN_STREAM
| CSUM_FSYNC
| CSUM_CLOSE
);
66 int fd
= finalize_hashfile(state
->f
, hash
, FSYNC_COMPONENT_PACK
, 0);
67 fixup_pack_header_footer(fd
, hash
, state
->pack_tmp_name
,
68 state
->nr_written
, hash
,
73 strbuf_addf(&packname
, "%s/pack/pack-%s.", get_object_directory(),
75 finish_tmp_packfile(&packname
, state
->pack_tmp_name
,
76 state
->written
, state
->nr_written
,
77 &state
->pack_idx_opts
, hash
);
78 for (i
= 0; i
< state
->nr_written
; i
++)
79 free(state
->written
[i
]);
83 memset(state
, 0, sizeof(*state
));
85 strbuf_release(&packname
);
86 /* Make objects we just wrote available to ourselves */
87 reprepare_packed_git(the_repository
);
91 * Cleanup after batch-mode fsync_object_files.
93 static void flush_batch_fsync(void)
95 struct strbuf temp_path
= STRBUF_INIT
;
96 struct tempfile
*temp
;
98 if (!bulk_fsync_objdir
)
102 * Issue a full hardware flush against a temporary file to ensure
103 * that all objects are durable before any renames occur. The code in
104 * fsync_loose_object_bulk_checkin has already issued a writeout
105 * request, but it has not flushed any writeback cache in the storage
106 * hardware or any filesystem logs. This fsync call acts as a barrier
107 * to ensure that the data in each new object file is durable before
108 * the final name is visible.
110 strbuf_addf(&temp_path
, "%s/bulk_fsync_XXXXXX", get_object_directory());
111 temp
= xmks_tempfile(temp_path
.buf
);
112 fsync_or_die(get_tempfile_fd(temp
), get_tempfile_path(temp
));
113 delete_tempfile(&temp
);
114 strbuf_release(&temp_path
);
117 * Make the object files visible in the primary ODB after their data is
120 tmp_objdir_migrate(bulk_fsync_objdir
);
121 bulk_fsync_objdir
= NULL
;
124 static int already_written(struct bulk_checkin_packfile
*state
, struct object_id
*oid
)
128 /* The object may already exist in the repository */
129 if (has_object_file(oid
))
132 /* Might want to keep the list sorted */
133 for (i
= 0; i
< state
->nr_written
; i
++)
134 if (oideq(&state
->written
[i
]->oid
, oid
))
137 /* This is a new object we need to keep */
142 * Read the contents from fd for size bytes, streaming it to the
143 * packfile in state while updating the hash in ctx. Signal a failure
144 * by returning a negative value when the resulting pack would exceed
145 * the pack size limit and this is not the first object in the pack,
146 * so that the caller can discard what we wrote from the current pack
147 * by truncating it and opening a new one. The caller will then call
148 * us again after rewinding the input fd.
150 * The already_hashed_to pointer is kept untouched by the caller to
151 * make sure we do not hash the same byte when we are called
152 * again. This way, the caller does not have to checkpoint its hash
153 * status before calling us just in case we ask it to call us again
156 static int stream_to_pack(struct bulk_checkin_packfile
*state
,
157 git_hash_ctx
*ctx
, off_t
*already_hashed_to
,
158 int fd
, size_t size
, enum object_type type
,
159 const char *path
, unsigned flags
)
162 unsigned char ibuf
[16384];
163 unsigned char obuf
[16384];
166 int write_object
= (flags
& HASH_WRITE_OBJECT
);
169 git_deflate_init(&s
, pack_compression_level
);
171 hdrlen
= encode_in_pack_object_header(obuf
, sizeof(obuf
), type
, size
);
172 s
.next_out
= obuf
+ hdrlen
;
173 s
.avail_out
= sizeof(obuf
) - hdrlen
;
175 while (status
!= Z_STREAM_END
) {
176 if (size
&& !s
.avail_in
) {
177 ssize_t rsize
= size
< sizeof(ibuf
) ? size
: sizeof(ibuf
);
178 ssize_t read_result
= read_in_full(fd
, ibuf
, rsize
);
180 die_errno("failed to read from '%s'", path
);
181 if (read_result
!= rsize
)
182 die("failed to read %d bytes from '%s'",
185 if (*already_hashed_to
< offset
) {
186 size_t hsize
= offset
- *already_hashed_to
;
190 the_hash_algo
->update_fn(ctx
, ibuf
, hsize
);
191 *already_hashed_to
= offset
;
198 status
= git_deflate(&s
, size
? 0 : Z_FINISH
);
200 if (!s
.avail_out
|| status
== Z_STREAM_END
) {
202 size_t written
= s
.next_out
- obuf
;
204 /* would we bust the size limit? */
205 if (state
->nr_written
&&
206 pack_size_limit_cfg
&&
207 pack_size_limit_cfg
< state
->offset
+ written
) {
208 git_deflate_abort(&s
);
212 hashwrite(state
->f
, obuf
, written
);
213 state
->offset
+= written
;
216 s
.avail_out
= sizeof(obuf
);
225 die("unexpected deflate failure: %d", status
);
232 /* Lazily create backing packfile for the state */
233 static void prepare_to_stream(struct bulk_checkin_packfile
*state
,
236 if (!(flags
& HASH_WRITE_OBJECT
) || state
->f
)
239 state
->f
= create_tmp_packfile(&state
->pack_tmp_name
);
240 reset_pack_idx_option(&state
->pack_idx_opts
);
242 /* Pretend we are going to write only one object */
243 state
->offset
= write_pack_header(state
->f
, 1);
245 die_errno("unable to write pack header");
248 static int deflate_to_pack(struct bulk_checkin_packfile
*state
,
249 struct object_id
*result_oid
,
251 enum object_type type
, const char *path
,
254 off_t seekback
, already_hashed_to
;
256 unsigned char obuf
[16384];
258 struct hashfile_checkpoint checkpoint
= {0};
259 struct pack_idx_entry
*idx
= NULL
;
261 seekback
= lseek(fd
, 0, SEEK_CUR
);
262 if (seekback
== (off_t
) -1)
263 return error("cannot find the current offset");
265 header_len
= format_object_header((char *)obuf
, sizeof(obuf
),
267 the_hash_algo
->init_fn(&ctx
);
268 the_hash_algo
->update_fn(&ctx
, obuf
, header_len
);
270 /* Note: idx is non-NULL when we are writing */
271 if ((flags
& HASH_WRITE_OBJECT
) != 0)
272 CALLOC_ARRAY(idx
, 1);
274 already_hashed_to
= 0;
277 prepare_to_stream(state
, flags
);
279 hashfile_checkpoint(state
->f
, &checkpoint
);
280 idx
->offset
= state
->offset
;
281 crc32_begin(state
->f
);
283 if (!stream_to_pack(state
, &ctx
, &already_hashed_to
,
284 fd
, size
, type
, path
, flags
))
287 * Writing this object to the current pack will make
288 * it too big; we need to truncate it, start a new
289 * pack, and write into it.
292 BUG("should not happen");
293 hashfile_truncate(state
->f
, &checkpoint
);
294 state
->offset
= checkpoint
.offset
;
295 flush_bulk_checkin_packfile(state
);
296 if (lseek(fd
, seekback
, SEEK_SET
) == (off_t
) -1)
297 return error("cannot seek back");
299 the_hash_algo
->final_oid_fn(result_oid
, &ctx
);
303 idx
->crc32
= crc32_end(state
->f
);
304 if (already_written(state
, result_oid
)) {
305 hashfile_truncate(state
->f
, &checkpoint
);
306 state
->offset
= checkpoint
.offset
;
309 oidcpy(&idx
->oid
, result_oid
);
310 ALLOC_GROW(state
->written
,
311 state
->nr_written
+ 1,
312 state
->alloc_written
);
313 state
->written
[state
->nr_written
++] = idx
;
318 void prepare_loose_object_bulk_checkin(void)
321 * We lazily create the temporary object directory
322 * the first time an object might be added, since
323 * callers may not know whether any objects will be
324 * added at the time they call begin_odb_transaction.
326 if (!odb_transaction_nesting
|| bulk_fsync_objdir
)
329 bulk_fsync_objdir
= tmp_objdir_create("bulk-fsync");
330 if (bulk_fsync_objdir
)
331 tmp_objdir_replace_primary_odb(bulk_fsync_objdir
, 0);
334 void fsync_loose_object_bulk_checkin(int fd
, const char *filename
)
337 * If we have an active ODB transaction, we issue a call that
338 * cleans the filesystem page cache but avoids a hardware flush
339 * command. Later on we will issue a single hardware flush
340 * before renaming the objects to their final names as part of
343 if (!bulk_fsync_objdir
||
344 git_fsync(fd
, FSYNC_WRITEOUT_ONLY
) < 0) {
346 warning(_("core.fsyncMethod = batch is unsupported on this platform"));
347 fsync_or_die(fd
, filename
);
351 int index_bulk_checkin(struct object_id
*oid
,
352 int fd
, size_t size
, enum object_type type
,
353 const char *path
, unsigned flags
)
355 int status
= deflate_to_pack(&bulk_checkin_packfile
, oid
, fd
, size
, type
,
357 if (!odb_transaction_nesting
)
358 flush_bulk_checkin_packfile(&bulk_checkin_packfile
);
362 void begin_odb_transaction(void)
364 odb_transaction_nesting
+= 1;
367 void flush_odb_transaction(void)
370 flush_bulk_checkin_packfile(&bulk_checkin_packfile
);
373 void end_odb_transaction(void)
375 odb_transaction_nesting
-= 1;
376 if (odb_transaction_nesting
< 0)
377 BUG("Unbalanced ODB transaction nesting");
379 if (odb_transaction_nesting
)
382 flush_odb_transaction();