mirror of
https://github.com/git/git.git
synced 2024-11-30 21:44:02 +08:00
94e99012fc
Using the a Linux-kernel-derived doubly-linked list implementation from the Userspace RCU library allows us to enqueue and delete items from the object request queue in constant time. This change reduces enqueue times in the prefetch() function where object request queue could grow to several thousand objects. I left out the list_for_each_entry* family macros from list.h which relied on the __typeof__ operator as we support platforms without it. Thus, list_entry (aka "container_of") needs to be called explicitly inside macro-wrapped for loops. The downside is this costs us an additional pointer per object request, but this is offset by reduced overhead on queue operations leading to improved performance and shorter queue depths. Signed-off-by: Eric Wong <e@80x24.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
585 lines
14 KiB
C
585 lines
14 KiB
C
#include "cache.h"
|
|
#include "commit.h"
|
|
#include "walker.h"
|
|
#include "http.h"
|
|
#include "list.h"
|
|
|
|
struct alt_base {
|
|
char *base;
|
|
int got_indices;
|
|
struct packed_git *packs;
|
|
struct alt_base *next;
|
|
};
|
|
|
|
enum object_request_state {
|
|
WAITING,
|
|
ABORTED,
|
|
ACTIVE,
|
|
COMPLETE
|
|
};
|
|
|
|
struct object_request {
|
|
struct walker *walker;
|
|
unsigned char sha1[20];
|
|
struct alt_base *repo;
|
|
enum object_request_state state;
|
|
struct http_object_request *req;
|
|
struct list_head node;
|
|
};
|
|
|
|
struct alternates_request {
|
|
struct walker *walker;
|
|
const char *base;
|
|
struct strbuf *url;
|
|
struct strbuf *buffer;
|
|
struct active_request_slot *slot;
|
|
int http_specific;
|
|
};
|
|
|
|
struct walker_data {
|
|
const char *url;
|
|
int got_alternates;
|
|
struct alt_base *alt;
|
|
};
|
|
|
|
static LIST_HEAD(object_queue_head);
|
|
|
|
static void fetch_alternates(struct walker *walker, const char *base);
|
|
|
|
static void process_object_response(void *callback_data);
|
|
|
|
static void start_object_request(struct walker *walker,
|
|
struct object_request *obj_req)
|
|
{
|
|
struct active_request_slot *slot;
|
|
struct http_object_request *req;
|
|
|
|
req = new_http_object_request(obj_req->repo->base, obj_req->sha1);
|
|
if (req == NULL) {
|
|
obj_req->state = ABORTED;
|
|
return;
|
|
}
|
|
obj_req->req = req;
|
|
|
|
slot = req->slot;
|
|
slot->callback_func = process_object_response;
|
|
slot->callback_data = obj_req;
|
|
|
|
/* Try to get the request started, abort the request on error */
|
|
obj_req->state = ACTIVE;
|
|
if (!start_active_slot(slot)) {
|
|
obj_req->state = ABORTED;
|
|
release_http_object_request(req);
|
|
return;
|
|
}
|
|
}
|
|
|
|
static void finish_object_request(struct object_request *obj_req)
|
|
{
|
|
if (finish_http_object_request(obj_req->req))
|
|
return;
|
|
|
|
if (obj_req->req->rename == 0)
|
|
walker_say(obj_req->walker, "got %s\n", sha1_to_hex(obj_req->sha1));
|
|
}
|
|
|
|
static void process_object_response(void *callback_data)
|
|
{
|
|
struct object_request *obj_req =
|
|
(struct object_request *)callback_data;
|
|
struct walker *walker = obj_req->walker;
|
|
struct walker_data *data = walker->data;
|
|
struct alt_base *alt = data->alt;
|
|
|
|
process_http_object_request(obj_req->req);
|
|
obj_req->state = COMPLETE;
|
|
|
|
/* Use alternates if necessary */
|
|
if (missing_target(obj_req->req)) {
|
|
fetch_alternates(walker, alt->base);
|
|
if (obj_req->repo->next != NULL) {
|
|
obj_req->repo =
|
|
obj_req->repo->next;
|
|
release_http_object_request(obj_req->req);
|
|
start_object_request(walker, obj_req);
|
|
return;
|
|
}
|
|
}
|
|
|
|
finish_object_request(obj_req);
|
|
}
|
|
|
|
static void release_object_request(struct object_request *obj_req)
|
|
{
|
|
if (obj_req->req !=NULL && obj_req->req->localfile != -1)
|
|
error("fd leakage in release: %d", obj_req->req->localfile);
|
|
|
|
list_del(&obj_req->node);
|
|
free(obj_req);
|
|
}
|
|
|
|
#ifdef USE_CURL_MULTI
|
|
static int fill_active_slot(struct walker *walker)
|
|
{
|
|
struct object_request *obj_req;
|
|
struct list_head *pos, *tmp, *head = &object_queue_head;
|
|
|
|
list_for_each_safe(pos, tmp, head) {
|
|
obj_req = list_entry(pos, struct object_request, node);
|
|
if (obj_req->state == WAITING) {
|
|
if (has_sha1_file(obj_req->sha1))
|
|
obj_req->state = COMPLETE;
|
|
else {
|
|
start_object_request(walker, obj_req);
|
|
return 1;
|
|
}
|
|
}
|
|
}
|
|
return 0;
|
|
}
|
|
#endif
|
|
|
|
static void prefetch(struct walker *walker, unsigned char *sha1)
|
|
{
|
|
struct object_request *newreq;
|
|
struct walker_data *data = walker->data;
|
|
|
|
newreq = xmalloc(sizeof(*newreq));
|
|
newreq->walker = walker;
|
|
hashcpy(newreq->sha1, sha1);
|
|
newreq->repo = data->alt;
|
|
newreq->state = WAITING;
|
|
newreq->req = NULL;
|
|
|
|
http_is_verbose = walker->get_verbosely;
|
|
list_add_tail(&newreq->node, &object_queue_head);
|
|
|
|
#ifdef USE_CURL_MULTI
|
|
fill_active_slots();
|
|
step_active_slots();
|
|
#endif
|
|
}
|
|
|
|
static void process_alternates_response(void *callback_data)
|
|
{
|
|
struct alternates_request *alt_req =
|
|
(struct alternates_request *)callback_data;
|
|
struct walker *walker = alt_req->walker;
|
|
struct walker_data *cdata = walker->data;
|
|
struct active_request_slot *slot = alt_req->slot;
|
|
struct alt_base *tail = cdata->alt;
|
|
const char *base = alt_req->base;
|
|
const char null_byte = '\0';
|
|
char *data;
|
|
int i = 0;
|
|
|
|
if (alt_req->http_specific) {
|
|
if (slot->curl_result != CURLE_OK ||
|
|
!alt_req->buffer->len) {
|
|
|
|
/* Try reusing the slot to get non-http alternates */
|
|
alt_req->http_specific = 0;
|
|
strbuf_reset(alt_req->url);
|
|
strbuf_addf(alt_req->url, "%s/objects/info/alternates",
|
|
base);
|
|
curl_easy_setopt(slot->curl, CURLOPT_URL,
|
|
alt_req->url->buf);
|
|
active_requests++;
|
|
slot->in_use = 1;
|
|
if (slot->finished != NULL)
|
|
(*slot->finished) = 0;
|
|
if (!start_active_slot(slot)) {
|
|
cdata->got_alternates = -1;
|
|
slot->in_use = 0;
|
|
if (slot->finished != NULL)
|
|
(*slot->finished) = 1;
|
|
}
|
|
return;
|
|
}
|
|
} else if (slot->curl_result != CURLE_OK) {
|
|
if (!missing_target(slot)) {
|
|
cdata->got_alternates = -1;
|
|
return;
|
|
}
|
|
}
|
|
|
|
fwrite_buffer((char *)&null_byte, 1, 1, alt_req->buffer);
|
|
alt_req->buffer->len--;
|
|
data = alt_req->buffer->buf;
|
|
|
|
while (i < alt_req->buffer->len) {
|
|
int posn = i;
|
|
while (posn < alt_req->buffer->len && data[posn] != '\n')
|
|
posn++;
|
|
if (data[posn] == '\n') {
|
|
int okay = 0;
|
|
int serverlen = 0;
|
|
struct alt_base *newalt;
|
|
if (data[i] == '/') {
|
|
/*
|
|
* This counts
|
|
* http://git.host/pub/scm/linux.git/
|
|
* -----------here^
|
|
* so memcpy(dst, base, serverlen) will
|
|
* copy up to "...git.host".
|
|
*/
|
|
const char *colon_ss = strstr(base,"://");
|
|
if (colon_ss) {
|
|
serverlen = (strchr(colon_ss + 3, '/')
|
|
- base);
|
|
okay = 1;
|
|
}
|
|
} else if (!memcmp(data + i, "../", 3)) {
|
|
/*
|
|
* Relative URL; chop the corresponding
|
|
* number of subpath from base (and ../
|
|
* from data), and concatenate the result.
|
|
*
|
|
* The code first drops ../ from data, and
|
|
* then drops one ../ from data and one path
|
|
* from base. IOW, one extra ../ is dropped
|
|
* from data than path is dropped from base.
|
|
*
|
|
* This is not wrong. The alternate in
|
|
* http://git.host/pub/scm/linux.git/
|
|
* to borrow from
|
|
* http://git.host/pub/scm/linus.git/
|
|
* is ../../linus.git/objects/. You need
|
|
* two ../../ to borrow from your direct
|
|
* neighbour.
|
|
*/
|
|
i += 3;
|
|
serverlen = strlen(base);
|
|
while (i + 2 < posn &&
|
|
!memcmp(data + i, "../", 3)) {
|
|
do {
|
|
serverlen--;
|
|
} while (serverlen &&
|
|
base[serverlen - 1] != '/');
|
|
i += 3;
|
|
}
|
|
/* If the server got removed, give up. */
|
|
okay = strchr(base, ':') - base + 3 <
|
|
serverlen;
|
|
} else if (alt_req->http_specific) {
|
|
char *colon = strchr(data + i, ':');
|
|
char *slash = strchr(data + i, '/');
|
|
if (colon && slash && colon < data + posn &&
|
|
slash < data + posn && colon < slash) {
|
|
okay = 1;
|
|
}
|
|
}
|
|
/* skip "objects\n" at end */
|
|
if (okay) {
|
|
struct strbuf target = STRBUF_INIT;
|
|
strbuf_add(&target, base, serverlen);
|
|
strbuf_add(&target, data + i, posn - i - 7);
|
|
if (walker->get_verbosely)
|
|
fprintf(stderr, "Also look at %s\n",
|
|
target.buf);
|
|
newalt = xmalloc(sizeof(*newalt));
|
|
newalt->next = NULL;
|
|
newalt->base = strbuf_detach(&target, NULL);
|
|
newalt->got_indices = 0;
|
|
newalt->packs = NULL;
|
|
|
|
while (tail->next != NULL)
|
|
tail = tail->next;
|
|
tail->next = newalt;
|
|
}
|
|
}
|
|
i = posn + 1;
|
|
}
|
|
|
|
cdata->got_alternates = 1;
|
|
}
|
|
|
|
static void fetch_alternates(struct walker *walker, const char *base)
|
|
{
|
|
struct strbuf buffer = STRBUF_INIT;
|
|
struct strbuf url = STRBUF_INIT;
|
|
struct active_request_slot *slot;
|
|
struct alternates_request alt_req;
|
|
struct walker_data *cdata = walker->data;
|
|
|
|
/*
|
|
* If another request has already started fetching alternates,
|
|
* wait for them to arrive and return to processing this request's
|
|
* curl message
|
|
*/
|
|
#ifdef USE_CURL_MULTI
|
|
while (cdata->got_alternates == 0) {
|
|
step_active_slots();
|
|
}
|
|
#endif
|
|
|
|
/* Nothing to do if they've already been fetched */
|
|
if (cdata->got_alternates == 1)
|
|
return;
|
|
|
|
/* Start the fetch */
|
|
cdata->got_alternates = 0;
|
|
|
|
if (walker->get_verbosely)
|
|
fprintf(stderr, "Getting alternates list for %s\n", base);
|
|
|
|
strbuf_addf(&url, "%s/objects/info/http-alternates", base);
|
|
|
|
/*
|
|
* Use a callback to process the result, since another request
|
|
* may fail and need to have alternates loaded before continuing
|
|
*/
|
|
slot = get_active_slot();
|
|
slot->callback_func = process_alternates_response;
|
|
alt_req.walker = walker;
|
|
slot->callback_data = &alt_req;
|
|
|
|
curl_easy_setopt(slot->curl, CURLOPT_FILE, &buffer);
|
|
curl_easy_setopt(slot->curl, CURLOPT_WRITEFUNCTION, fwrite_buffer);
|
|
curl_easy_setopt(slot->curl, CURLOPT_URL, url.buf);
|
|
|
|
alt_req.base = base;
|
|
alt_req.url = &url;
|
|
alt_req.buffer = &buffer;
|
|
alt_req.http_specific = 1;
|
|
alt_req.slot = slot;
|
|
|
|
if (start_active_slot(slot))
|
|
run_active_slot(slot);
|
|
else
|
|
cdata->got_alternates = -1;
|
|
|
|
strbuf_release(&buffer);
|
|
strbuf_release(&url);
|
|
}
|
|
|
|
static int fetch_indices(struct walker *walker, struct alt_base *repo)
|
|
{
|
|
int ret;
|
|
|
|
if (repo->got_indices)
|
|
return 0;
|
|
|
|
if (walker->get_verbosely)
|
|
fprintf(stderr, "Getting pack list for %s\n", repo->base);
|
|
|
|
switch (http_get_info_packs(repo->base, &repo->packs)) {
|
|
case HTTP_OK:
|
|
case HTTP_MISSING_TARGET:
|
|
repo->got_indices = 1;
|
|
ret = 0;
|
|
break;
|
|
default:
|
|
repo->got_indices = 0;
|
|
ret = -1;
|
|
}
|
|
|
|
return ret;
|
|
}
|
|
|
|
static int http_fetch_pack(struct walker *walker, struct alt_base *repo, unsigned char *sha1)
|
|
{
|
|
struct packed_git *target;
|
|
int ret;
|
|
struct slot_results results;
|
|
struct http_pack_request *preq;
|
|
|
|
if (fetch_indices(walker, repo))
|
|
return -1;
|
|
target = find_sha1_pack(sha1, repo->packs);
|
|
if (!target)
|
|
return -1;
|
|
|
|
if (walker->get_verbosely) {
|
|
fprintf(stderr, "Getting pack %s\n",
|
|
sha1_to_hex(target->sha1));
|
|
fprintf(stderr, " which contains %s\n",
|
|
sha1_to_hex(sha1));
|
|
}
|
|
|
|
preq = new_http_pack_request(target, repo->base);
|
|
if (preq == NULL)
|
|
goto abort;
|
|
preq->lst = &repo->packs;
|
|
preq->slot->results = &results;
|
|
|
|
if (start_active_slot(preq->slot)) {
|
|
run_active_slot(preq->slot);
|
|
if (results.curl_result != CURLE_OK) {
|
|
error("Unable to get pack file %s\n%s", preq->url,
|
|
curl_errorstr);
|
|
goto abort;
|
|
}
|
|
} else {
|
|
error("Unable to start request");
|
|
goto abort;
|
|
}
|
|
|
|
ret = finish_http_pack_request(preq);
|
|
release_http_pack_request(preq);
|
|
if (ret)
|
|
return ret;
|
|
|
|
return 0;
|
|
|
|
abort:
|
|
return -1;
|
|
}
|
|
|
|
static void abort_object_request(struct object_request *obj_req)
|
|
{
|
|
release_object_request(obj_req);
|
|
}
|
|
|
|
static int fetch_object(struct walker *walker, unsigned char *sha1)
|
|
{
|
|
char *hex = sha1_to_hex(sha1);
|
|
int ret = 0;
|
|
struct object_request *obj_req = NULL;
|
|
struct http_object_request *req;
|
|
struct list_head *pos, *head = &object_queue_head;
|
|
|
|
list_for_each(pos, head) {
|
|
obj_req = list_entry(pos, struct object_request, node);
|
|
if (!hashcmp(obj_req->sha1, sha1))
|
|
break;
|
|
}
|
|
if (obj_req == NULL)
|
|
return error("Couldn't find request for %s in the queue", hex);
|
|
|
|
if (has_sha1_file(obj_req->sha1)) {
|
|
if (obj_req->req != NULL)
|
|
abort_http_object_request(obj_req->req);
|
|
abort_object_request(obj_req);
|
|
return 0;
|
|
}
|
|
|
|
#ifdef USE_CURL_MULTI
|
|
while (obj_req->state == WAITING)
|
|
step_active_slots();
|
|
#else
|
|
start_object_request(walker, obj_req);
|
|
#endif
|
|
|
|
/*
|
|
* obj_req->req might change when fetching alternates in the callback
|
|
* process_object_response; therefore, the "shortcut" variable, req,
|
|
* is used only after we're done with slots.
|
|
*/
|
|
while (obj_req->state == ACTIVE)
|
|
run_active_slot(obj_req->req->slot);
|
|
|
|
req = obj_req->req;
|
|
|
|
if (req->localfile != -1) {
|
|
close(req->localfile);
|
|
req->localfile = -1;
|
|
}
|
|
|
|
/*
|
|
* we turned off CURLOPT_FAILONERROR to avoid losing a
|
|
* persistent connection and got CURLE_OK.
|
|
*/
|
|
if (req->http_code == 404 && req->curl_result == CURLE_OK &&
|
|
(starts_with(req->url, "http://") ||
|
|
starts_with(req->url, "https://")))
|
|
req->curl_result = CURLE_HTTP_RETURNED_ERROR;
|
|
|
|
if (obj_req->state == ABORTED) {
|
|
ret = error("Request for %s aborted", hex);
|
|
} else if (req->curl_result != CURLE_OK &&
|
|
req->http_code != 416) {
|
|
if (missing_target(req))
|
|
ret = -1; /* Be silent, it is probably in a pack. */
|
|
else
|
|
ret = error("%s (curl_result = %d, http_code = %ld, sha1 = %s)",
|
|
req->errorstr, req->curl_result,
|
|
req->http_code, hex);
|
|
} else if (req->zret != Z_STREAM_END) {
|
|
walker->corrupt_object_found++;
|
|
ret = error("File %s (%s) corrupt", hex, req->url);
|
|
} else if (hashcmp(obj_req->sha1, req->real_sha1)) {
|
|
ret = error("File %s has bad hash", hex);
|
|
} else if (req->rename < 0) {
|
|
ret = error("unable to write sha1 filename %s",
|
|
sha1_file_name(req->sha1));
|
|
}
|
|
|
|
release_http_object_request(req);
|
|
release_object_request(obj_req);
|
|
return ret;
|
|
}
|
|
|
|
static int fetch(struct walker *walker, unsigned char *sha1)
|
|
{
|
|
struct walker_data *data = walker->data;
|
|
struct alt_base *altbase = data->alt;
|
|
|
|
if (!fetch_object(walker, sha1))
|
|
return 0;
|
|
while (altbase) {
|
|
if (!http_fetch_pack(walker, altbase, sha1))
|
|
return 0;
|
|
fetch_alternates(walker, data->alt->base);
|
|
altbase = altbase->next;
|
|
}
|
|
return error("Unable to find %s under %s", sha1_to_hex(sha1),
|
|
data->alt->base);
|
|
}
|
|
|
|
static int fetch_ref(struct walker *walker, struct ref *ref)
|
|
{
|
|
struct walker_data *data = walker->data;
|
|
return http_fetch_ref(data->alt->base, ref);
|
|
}
|
|
|
|
static void cleanup(struct walker *walker)
|
|
{
|
|
struct walker_data *data = walker->data;
|
|
struct alt_base *alt, *alt_next;
|
|
|
|
if (data) {
|
|
alt = data->alt;
|
|
while (alt) {
|
|
alt_next = alt->next;
|
|
|
|
free(alt->base);
|
|
free(alt);
|
|
|
|
alt = alt_next;
|
|
}
|
|
free(data);
|
|
walker->data = NULL;
|
|
}
|
|
}
|
|
|
|
struct walker *get_http_walker(const char *url)
|
|
{
|
|
char *s;
|
|
struct walker_data *data = xmalloc(sizeof(struct walker_data));
|
|
struct walker *walker = xmalloc(sizeof(struct walker));
|
|
|
|
data->alt = xmalloc(sizeof(*data->alt));
|
|
data->alt->base = xstrdup(url);
|
|
for (s = data->alt->base + strlen(data->alt->base) - 1; *s == '/'; --s)
|
|
*s = 0;
|
|
|
|
data->alt->got_indices = 0;
|
|
data->alt->packs = NULL;
|
|
data->alt->next = NULL;
|
|
data->got_alternates = -1;
|
|
|
|
walker->corrupt_object_found = 0;
|
|
walker->fetch = fetch;
|
|
walker->fetch_ref = fetch_ref;
|
|
walker->prefetch = prefetch;
|
|
walker->cleanup = cleanup;
|
|
walker->data = data;
|
|
|
|
#ifdef USE_CURL_MULTI
|
|
add_fill_function(walker, (int (*)(void *)) fill_active_slot);
|
|
#endif
|
|
|
|
return walker;
|
|
}
|