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

Re: git fails: segfault at 0 ip 00000000004076d5 sp 00007fff7806ebc0

From
Erik Faye-Lund <kusmabite@gmail.com>
Date
Oct 12, 2012, 16:29 UTC
Message-ID
<CABPQNSYyh-gk8aD9SKLzo0MDBkGECthdGZiT1mB1pA0ajdB8Cw@mail.gmail.com>
In-Reply-To
<CAJa+X0MefHXTu_D7-SN8itQVO4sMyc4O7=bK=-sUP+W4GHOu9g@mail.gmail.com>
On Fri, Oct 12, 2012 at 6:58 AM, Brad Hein <linuxbrad@gmail.com> wrote:
Show 25 quoted lines
> In Fedora 17
> With git-1.7.11.7-1.fc17.x86_64 (rpm)
>
> I try to clone a particular repository but git just returns, having
> not cloned the repo. Seems like a bug. Details follow:
>   $ git clone http://gnuradio.org/git/gnuradio.git
>
> While the command fails a message is logged to syslog. Repeated
> attempts to clone the repo yield the same result:
>   Oct 11 21:38:25 localhost kernel: [662703.442645]
> git-remote-http[25796]: segfault at 0 ip 00000000004076d5 sp
> 00007fff7806ebc0 error 4 in git-remote-http[400000+96000]
>   Oct 11 21:39:00 localhost kernel: [662737.899829]
> git-remote-http[25837]: segfault at 0 ip 00000000004076d5 sp
> 00007fff37c5ef20 error 4 in git-remote-http[400000+96000]
>   Oct 11 21:39:25 localhost kernel: [662763.341248]
> git-remote-http[25873]: segfault at 0 ip 00000000004076d5 sp
> 00007fff6310d470 error 4 in git-remote-http[400000+96000]
>
> A tcpdump reveals that the last thing the client does is requests a
> file that doesn't exist on the server (404). Details are in my post on
> FedoraForums: http://forums.fedoraforum.org/showthread.php?p=1607891&posted=1#post1607891
>
> Problem mitigated by downgrade to "git-1.7.10.1-1.fc17.x86_64" or
> "git-1.7.11.4-3.fc17.x86_64" or try to clone a different repository.

Thanks for reporting. I gave it a quick go, and the issue seems to also be present in the current 'master'.

The problem is a NULL-pointer dereferencing introduced in 8809703 ("http: factor out http error code handling"), where the code assume that slot->results still points to http_request::results. This assumption seems to be wrong.

This seems to step around the issue, but I don't know if http_request::results should be set to NULL in the first place. Jeff?

diff --git a/http.c b/http.c
index 345c171..ac3ab16 100644
--- a/http.c
+++ b/http.c
@@ -745,10 +745,8 @@ char *get_remote_object_url(const char *url,
const char *hex,
 	return strbuf_detach(&buf, NULL);
 }

-int handle_curl_result(struct active_request_slot *slot)
+int handle_curl_result(struct active_request_slot *slot, struct
slot_results *results)
 {
-	struct slot_results *results = slot->results;
-
 	if (results->curl_result == CURLE_OK) {
 		credential_approve(&http_auth);
 		return HTTP_OK;
@@ -822,7 +820,7 @@ static int http_request(const char *url, void
*result, int target, int options)

 	if (start_active_slot(slot)) {
 		run_active_slot(slot);
-		ret = handle_curl_result(slot);
+		ret = handle_curl_result(slot, &results);
 	} else {
 		error("Unable to start HTTP request for %s", url);
 		ret = HTTP_START_FAILED;
diff --git a/http.h b/http.h
index 12de255..12c27fa 100644
--- a/http.h
+++ b/http.h
@@ -78,7 +78,7 @@ extern int start_active_slot(struct
active_request_slot *slot);
 extern void run_active_slot(struct active_request_slot *slot);
 extern void finish_active_slot(struct active_request_slot *slot);
 extern void finish_all_active_slots(void);
-extern int handle_curl_result(struct active_request_slot *slot);
+extern int handle_curl_result(struct active_request_slot *slot,
struct slot_results *results);

 #ifdef USE_CURL_MULTI
 extern void fill_active_slots(void);
diff --git a/remote-curl.c b/remote-curl.c
index 10fa8f1..42716c5 100644
--- a/remote-curl.c
+++ b/remote-curl.c
@@ -356,7 +356,7 @@ static int run_slot(struct active_request_slot *slot)
 	slot->curl_result = curl_easy_perform(slot->curl);
 	finish_active_slot(slot);

-	err = handle_curl_result(slot);
+	err = handle_curl_result(slot, &results);
 	if (err != HTTP_OK && err != HTTP_REAUTH) {
 		error("RPC failed; result=%d, HTTP code = %ld",
 		      results.curl_result, results.http_code);
Previous: Junio C HamanoNext: Erik Faye-Lund
Message 8 of 10 in “git fails: segfault at 0 ip 00000000004076d5 sp 00007fff7806ebc0”
  1. Brad HeinOct 12, 2012
  2. http: fix segfault in handle_curl_resultJeff King, Oct 12, 2012
  3. Jeff KingOct 12, 2012
  4. 1/2 remote-curl: do not call run_slot repeatedlyJeff King, Oct 12, 2012
  5. 2/2 http: do not set up curl auth after a 401Jeff King, Oct 12, 2012
  6. Brad HeinOct 12, 2012
  7. Junio C HamanoOct 12, 2012
  8. Erik Faye-LundOct 12, 2012
  9. Erik Faye-LundOct 12, 2012
  10. Jeff KingOct 12, 2012

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.