git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH] http API: fix dangling pointer issue noted by GCC 12.0

From
Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Date
Jan 26, 2022, 21:30 UTC
Message-ID
<patch-1.1-1cec367e805-20220126T212921Z-avarab@gmail.com>

The pre-release GCC 12.0 development branch has a new warning about dangling pointers in -Wall:

    http.c: In function ‘run_active_slot’:
    http.c:1332:24: error: storing the address of local variable ‘finished’ in ‘*slot.finished’ [-Werror=dangling-pointer=]
     1332 |         slot->finished = &finished;
          |         ~~~~~~~~~~~~~~~^~~~~~~~~~~
    http.c:1330:13: note: ‘finished’ declared here
     1330 |         int finished = 0;
          |             ^~~~~~~~

This is on a locally built "gcc (GCC) 12.0.1 20220120 (experimental)", built from gcc.git's 8bc700f4c3f (Enhance vec_pack_trunc for integral mode mask., 2022-01-17).

To fix this I first simply made the member "int finished", i.e. removing the pointer indirection. It turns out that nothing cared about the state of it being a NULL pointer v.s. "*ptr == 0".

But we can instead amend the code added in baa7b67d091 (HTTP slot reuse fixes, 2006-03-10) to get rid of "int *finished" entirely. I instrumented the code to add this after every use of slot->finished or slot->in_use:

    if (slot->finished && slot->in_use == *slot->finished) BUG("in-use = %d and finished = %d disconnect", slot->in_use, *slot->finished);
    if (!slot->finished && !slot->in_use) BUG("have !in-use and no finished pointer");
Which never fires, but we would get occurrences of:
    if (!slot->finished && slot->in_use) BUG("have in-use and no finished pointer");

I.e. we can simply drop the field and rely on "slot->in_use" in cases where we used "finished" before. The two fields were mirror images of each other, and the tri-state nature of "finished" wasn't something we relied upon.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
---

A look at GCC's release history suggests that GCC 12.0 will be released sometime in the next 3-6 months: https://gcc.gnu.org/releases.html

By addressing this new -Wall warning early we can get ahead of the annoying compilation error we'll run into with DEVELOPER=1 once it's released.

 http-walker.c | 4 ----
 http.c        | 8 +-------
 http.h        | 1 -
 3 files changed, 1 insertion(+), 12 deletions(-)
diff --git a/http-walker.c b/http-walker.c
index 910fae539b8..5cc369dea85 100644
--- a/http-walker.c
+++ b/http-walker.c
@@ -225,13 +225,9 @@ static void process_alternates_response(void *callback_data)
 					 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;
 		}
diff --git a/http.c b/http.c
index 229da4d1488..4507c9ac9c7 100644
--- a/http.c
+++ b/http.c
@@ -197,9 +197,6 @@ static void finish_active_slot(struct active_request_slot *slot)
 	closedown_active_slot(slot);
 	curl_easy_getinfo(slot->curl, CURLINFO_HTTP_CODE, &slot->http_code);
 
-	if (slot->finished != NULL)
-		(*slot->finished) = 1;
-
 	/* Store slot results so they can be read after the slot is reused */
 	if (slot->results != NULL) {
 		slot->results->curl_result = slot->curl_result;
@@ -1204,7 +1201,6 @@ struct active_request_slot *get_active_slot(void)
 	active_requests++;
 	slot->in_use = 1;
 	slot->results = NULL;
-	slot->finished = NULL;
 	slot->callback_data = NULL;
 	slot->callback_func = NULL;
 	curl_easy_setopt(slot->curl, CURLOPT_COOKIEFILE, curl_cookie_file);
@@ -1327,10 +1323,8 @@ void run_active_slot(struct active_request_slot *slot)
 	fd_set excfds;
 	int max_fd;
 	struct timeval select_timeout;
-	int finished = 0;
 
-	slot->finished = &finished;
-	while (!finished) {
+	while (slot->in_use) {
 		step_active_slots();
 
 		if (slot->in_use) {
diff --git a/http.h b/http.h
index df1590e53a4..81418d5fd8b 100644
--- a/http.h
+++ b/http.h
@@ -24,7 +24,6 @@ struct active_request_slot {
 	int in_use;
 	CURLcode curl_result;
 	long http_code;
-	int *finished;
 	struct slot_results *results;
 	void *callback_data;
 	void (*callback_func)(void *data);
-- 
2.35.0.890.gd7e422415d9
Next: Taylor Blau
Message 1 of 15 in “http API: fix dangling pointer issue noted by GCC 12.0”
  1. http API: fix dangling pointer issue noted by GCC 12.0Ævar Arnfjörð Bjarmason, Jan 26, 2022
  2. Taylor BlauJan 26, 2022
  3. Junio C HamanoJan 27, 2022
  4. Junio C HamanoJan 27, 2022
  5. Ævar Arnfjörð BjarmasonJan 27, 2022
  6. Junio C HamanoJan 27, 2022
  7. http API: fix dangling pointer issue noted by GCC 12.0Ævar Arnfjörð Bjarmason, Feb 25, 2022
  8. Junio C HamanoFeb 25, 2022
  9. Taylor BlauFeb 26, 2022
  10. http API: fix dangling pointer issue noted by GCC 12.0Ævar Arnfjörð Bjarmason, Mar 25, 2022
  11. Taylor BlauMar 25, 2022
  12. Junio C HamanoMar 26, 2022
  13. Ævar Arnfjörð BjarmasonApr 14, 2022
  14. Junio C HamanoApr 14, 2022
  15. Ævar Arnfjörð BjarmasonApr 15, 2022

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.