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

[PATCH 1/2] Have set_try_to_free_routine return the previous routine

From
Johannes Sixt <j6t@kdbg.org>
Date
May 8, 2010, 15:13 UTC
Message-ID
<201005081713.49937.j6t@kdbg.org>
In-Reply-To
<4BDDDA21.4090304@kdbg.org>

This effectively requires from the callers of set_try_to_free_routine to treat the try-to-free-routines as a stack.

We will need this for the next patch where the only current caller cannot depend on that the previously set routine was the default routine.

Signed-off-by: Johannes Sixt <j6t@kdbg.org>
---
On Sonntag, 2. Mai 2010, Johannes Sixt wrote:
Show 12 quoted lines
> Am 02.05.2010 18:12, schrieb Junio C Hamano:
> > * np/malloc-threading (2010-04-08) 2 commits
> >
> > * js/async-thread (2010-03-09) 7 commits
>
> Note that np/malloc-threading does not help the potential xmalloc race
> condition in js/async-thread topic at all. :-( np/malloc-threading is only
> about a race that is internal in pack-objects.
>
> Now that you force me to think about the problem, I suggest something along
> these lines: Disable try_to_free_routine when GIT_TRACE is set, and require
> callers of set_try_to_free_routine to restore the previous routine.

Here is a 2-patch series that implements this idea. I wrote it on top of np/malloc-threading, because it implements the infrastructure that is necessary for this.

 builtin-pack-objects.c |    6 ++++--
 git-compat-util.h      |    3 ++-
 wrapper.c              |    6 ++++--
 3 files changed, 10 insertions(+), 5 deletions(-)
diff --git a/builtin-pack-objects.c b/builtin-pack-objects.c
index 26fc7cd..5279cd9 100644
--- a/builtin-pack-objects.c
+++ b/builtin-pack-objects.c
@@ -1556,6 +1556,8 @@ static void try_to_free_from_threads(size_t size)
 	read_unlock();
 }
 
+try_to_free_t old_try_to_free_routine;
+
 /*
  * The main thread waits on the condition that (at least) one of the workers
  * has stopped working (which is indicated in the .working member of
@@ -1590,12 +1592,12 @@ static void init_threaded_search(void)
 	pthread_mutex_init(&cache_mutex, NULL);
 	pthread_mutex_init(&progress_mutex, NULL);
 	pthread_cond_init(&progress_cond, NULL);
-	set_try_to_free_routine(try_to_free_from_threads);
+	old_try_to_free_routine = set_try_to_free_routine(try_to_free_from_threads);
 }
 
 static void cleanup_threaded_search(void)
 {
-	set_try_to_free_routine(NULL);
+	set_try_to_free_routine(old_try_to_free_routine);
 	pthread_cond_destroy(&progress_cond);
 	pthread_mutex_destroy(&read_mutex);
 	pthread_mutex_destroy(&cache_mutex);
diff --git a/git-compat-util.h b/git-compat-util.h
index 1c171db..828aada 100644
--- a/git-compat-util.h
+++ b/git-compat-util.h
@@ -346,7 +346,8 @@ static inline char *gitstrchrnul(const char *s, int c)
 
 extern void release_pack_memory(size_t, int);
 
-extern void set_try_to_free_routine(void (*routine)(size_t));
+typedef void (*try_to_free_t)(size_t);
+extern try_to_free_t set_try_to_free_routine(try_to_free_t);
 
 extern char *xstrdup(const char *str);
 extern void *xmalloc(size_t size);
diff --git a/wrapper.c b/wrapper.c
index 62edb57..8aa9df9 100644
--- a/wrapper.c
+++ b/wrapper.c
@@ -10,9 +10,11 @@ static void try_to_free_builtin(size_t size)
 
 static void (*try_to_free_routine)(size_t size) = try_to_free_builtin;
 
-void set_try_to_free_routine(void (*routine)(size_t))
+try_to_free_t set_try_to_free_routine(try_to_free_t routine)
 {
-	try_to_free_routine = (routine) ? routine : try_to_free_builtin;
+	try_to_free_t old = try_to_free_routine;
+	try_to_free_routine = routine;
+	return old;
 }
 
 char *xstrdup(const char *str)
-- 
1.7.1.64.ga1799.dirty
Previous: Johannes SixtNext: Johannes Sixt
Message 3 of 16 in “What's cooking in git.git (May 2010, #01; Sun, 2)”
  1. Junio C HamanoMay 2, 2010
  2. Johannes SixtMay 2, 2010
  3. 1/2 Have set_try_to_free_routine return the previous routineJohannes Sixt, May 8, 2010
  4. 2/2 Do not call release_pack_memory in malloc wrappers when GIT_TRACE is usedJohannes Sixt, May 8, 2010
  5. Jakub NarebskiMay 3, 2010
  6. Will PalmerMay 3, 2010
  7. jn/shortlog (Re: What's cooking in git.git (May 2010, #01; Sun, 2))Jonathan Nieder, May 4, 2010
  8. Jonathan NiederMay 4, 2010
  9. 0/5 jn/shortlogJonathan Nieder, May 4, 2010
  10. 1/5 Documentation/shortlog: scripted users should not rely on implicit HEADJonathan Nieder, May 4, 2010
  11. 2/5 t4201 (shortlog): guard setup with test_expect_successJonathan Nieder, May 4, 2010
  12. 3/5 t4201 (shortlog): Test output format with multiple authorsJonathan Nieder, May 4, 2010
  13. 4/5 shortlog: Document and test --format optionJonathan Nieder, May 4, 2010
  14. 5/5 pretty: Respect --abbrev optionJonathan Nieder, May 4, 2010
  15. Jonathan NiederMay 4, 2010
  16. Clemens BuchacherMay 8, 2010

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.