Skip to contentSkip to conversation

Fix racy fsmonitor

PATCH New Johannes Schindelin via GitGitGadget 1 participant 3 messages May 24, 2019

lore
Share & tools
lore

Johannes Schindelin via GitGitGadget

author
#0 lore raw

cover Fix racy fsmonitor

The t7519-status-fsmonitor.sh tests became a lot more flaky with the recent fsmonitor fix (js/fsmonitor-refresh-after-discarding-index). That fix, however, did not introduce the flakiness, but it just made it much more likely to be hit. And it seemed to be hit only on Windows.

The reason, though, is that the fsmonitor feature failed to mark the in-memory index as changed, i.e. in need of writing, and it was the has_racy_timestamp() test that hid this bug in most cases (although a lot less on Windows, where the files' mtimes are actually a lot more accurate than on Linux).

This fixes https://github.com/gitgitgadget/git/issues/197

Johannes Schindelin (2):
fill_stat_cache_info(): prepare for an fsmonitor fix mark_fsmonitor_valid(): mark the index as changed if needed

 apply.c                |  2 +-
 builtin/update-index.c |  2 +-
 cache.h                |  2 +-
 diff-lib.c             |  2 +-
 entry.c                |  2 +-
 fsmonitor.h            |  5 +++--
 preload-index.c        |  2 +-
 read-cache.c           | 10 +++++-----
 8 files changed, 14 insertions(+), 13 deletions(-)

base-commit: aa25c82427ae70aebf3b8f970f2afd54e9a2a8c6
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-223%2Fdscho%2Ffix-racy-fsmonitor-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-223/dscho/fix-racy-fsmonitor-v1
Pull-Request: https://github.com/gitgitgadget/git/pull/223

Signature · 1 line
gitgitgadget
author
#1 lore raw

PATCH 1/2 fill_stat_cache_info(): prepare for an fsmonitor fix

From: Johannes Schindelin <[email protected]>

We will need to pass down the struct index_state to mark_fsmonitor_valid() for an upcoming bug fix, and this here function calls that there function, so we need to extend the signature of fill_stat_cache_info() first.

Signed-off-by Johannes Schindelin <[email protected]>
---
 apply.c                | 2 +-
 builtin/update-index.c | 2 +-
 cache.h                | 2 +-
 entry.c                | 2 +-
 read-cache.c           | 6 +++---
 5 files changed, 7 insertions(+), 7 deletions(-)
apply.c +1 -1
index f15afa9f6a..4992eca416 100644
@@ -4310,7 +4310,7 @@ static int add_index_file(struct apply_state *state,
 						     "created file '%s'"),
 						   path);
 			}
-			fill_stat_cache_info(ce, &st);
+			fill_stat_cache_info(state->repo->index, ce, &st);
 		}
 		if (write_object_file(buf, size, blob_type, &ce->oid) < 0) {
 			discard_cache_entry(ce);
builtin/update-index.c +1 -1
index 27db0928bf..3f8cc6ccb4 100644
@@ -280,7 +280,7 @@ static int add_one_path(const struct cache_entry *old, const char *path, int len
 	memcpy(ce->name, path, len);
 	ce->ce_flags = create_ce_flags(0);
 	ce->ce_namelen = len;
-	fill_stat_cache_info(ce, st);
+	fill_stat_cache_info(&the_index, ce, st);
 	ce->ce_mode = ce_mode_from_stat(old, st->st_mode);
 
 	if (index_path(&the_index, &ce->oid, path, st,
cache.h +1 -1
index b4bb2e2c11..bf20337ef4 100644
@@ -826,7 +826,7 @@ int match_stat_data(const struct stat_data *sd, struct stat *st);
 int match_stat_data_racy(const struct index_state *istate,
 			 const struct stat_data *sd, struct stat *st);
 
-void fill_stat_cache_info(struct cache_entry *ce, struct stat *st);
+void fill_stat_cache_info(struct index_state *istate, struct cache_entry *ce, struct stat *st);
 
 #define REFRESH_REALLY		0x0001	/* ignore_valid */
 #define REFRESH_UNMERGED	0x0002	/* allow unmerged */
entry.c +1 -1
index 0e4f2f2910..53380bb614 100644
@@ -373,7 +373,7 @@ static int write_entry(struct cache_entry *ce,
 			if (lstat(ce->name, &st) < 0)
 				return error_errno("unable to stat just-written file %s",
 						   ce->name);
-		fill_stat_cache_info(ce, &st);
+		fill_stat_cache_info(state->istate, ce, &st);
 		ce->ce_flags |= CE_UPDATE_IN_BASE;
 		mark_fsmonitor_invalid(state->istate, ce);
 		state->istate->cache_changed |= CE_ENTRY_CHANGED;
read-cache.c +3 -3
index 22e7b9944e..d3b43ae3ba 100644
@@ -195,7 +195,7 @@ int match_stat_data(const struct stat_data *sd, struct stat *st)
  * cache, ie the parts that aren't tracked by GIT, and only used
  * to validate the cache.
  */
-void fill_stat_cache_info(struct cache_entry *ce, struct stat *st)
+void fill_stat_cache_info(struct index_state *istate, struct cache_entry *ce, struct stat *st)
 {
 	fill_stat_data(&ce->ce_stat_data, st);
 

@@ -728,7 +728,7 @@ int add_to_index(struct index_state *istate, const char *path, struct stat *st,
 	memcpy(ce->name, path, namelen);
 	ce->ce_namelen = namelen;
 	if (!intent_only)
-		fill_stat_cache_info(ce, st);
+		fill_stat_cache_info(istate, ce, st);
 	else
 		ce->ce_flags |= CE_INTENT_TO_ADD;
 

@@ -1447,7 +1447,7 @@ static struct cache_entry *refresh_cache_ent(struct index_state *istate,
 	updated = make_empty_cache_entry(istate, ce_namelen(ce));
 	copy_cache_entry(updated, ce);
 	memcpy(updated->name, ce->name, ce->ce_namelen + 1);
-	fill_stat_cache_info(updated, &st);
+	fill_stat_cache_info(istate, updated, &st);
 	/*
 	 * If ignore_valid is not set, we should leave CE_VALID bit
 	 * alone.  Otherwise, paths marked with --no-assume-unchanged
Signature · 1 line
gitgitgadget
author
#2 lore raw

PATCH 2/2 mark_fsmonitor_valid(): mark the index as changed if needed

In reply to Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

Without this bug fix, t7519's four "status doesn't detect unreported modifications" test cases would fail occasionally (and, oddly enough, *a lot* more frequently on Windows).

The reason is that these test cases intentionally use the side effect of git status to re-write the index if any updates were detected: they first clean the worktree, run git status to update the index as well as show the output to the casual reader, then make the worktree dirty again and expect no changes to reported if running with a mocked fsmonitor hook.

The problem with this strategy was that the index was written during said git status on the clean worktree for the *wrong* reason: not because the index was marked as changed (it wasn't), but because the recorded mtimes were racy with the index' own mtime.

As the mtime granularity on Windows is 100 nanoseconds (see e.g.
https://docs.microsoft.com/en-us/windows/desktop/SysInfo/file-times),
the mtimes of the files are often enough *not* racy with the index', so that that git status call currently does not always update the index (including the fsmonitor extension), causing the test case to fail.

The obvious fix: if we change *any* index entry's CE_FSMONITOR_VALID flag, we should also mark the index as changed. That will cause the index to be written upon git status, *including* an updated fsmonitor extension.

Side note: Even though the reader might think that the t7519 issue should be *much* more prevalent on Linux, given that the ext4 filesystem (that seems to be used by every Linux distribution) stores mtimes in nanosecond precision. However, ext4 uses current_kernel_time() (see
https://unix.stackexchange.com/questions/11599#comment762968_11599; it is *amazingly* hard to find any proper source of information about such ext4 questions) whose accuracy seems to depend on many factors but is safely worse than the 100-nanosecond granularity of NTFS (again, it is *horribly* hard to find anything remotely authoritative about this question). So it seems that the racy index condition that hid the bug fixed by this patch simply is a lot more likely on Linux than on Windows. But not impossible ;-)

Signed-off-by Johannes Schindelin <[email protected]>
---
 diff-lib.c      | 2 +-
 fsmonitor.h     | 5 +++--
 preload-index.c | 2 +-
 read-cache.c    | 4 ++--
 4 files changed, 7 insertions(+), 6 deletions(-)
diff-lib.c +1 -1
index a838c219ec..61812f48c2 100644
@@ -232,7 +232,7 @@ int run_diff_files(struct rev_info *revs, unsigned int option)
 
 		if (!changed && !dirty_submodule) {
 			ce_mark_uptodate(ce);
-			mark_fsmonitor_valid(ce);
+			mark_fsmonitor_valid(istate, ce);
 			if (!revs->diffopt.flags.find_copies_harder)
 				continue;
 		}
fsmonitor.h +3 -2
index 8489fa3244..739318ab6d 100644
@@ -49,9 +49,10 @@ void refresh_fsmonitor(struct index_state *istate);
  * called any time the cache entry has been updated to reflect the
  * current state of the file on disk.
  */
-static inline void mark_fsmonitor_valid(struct cache_entry *ce)
+static inline void mark_fsmonitor_valid(struct index_state *istate, struct cache_entry *ce)
 {
-	if (core_fsmonitor) {
+	if (core_fsmonitor && !(ce->ce_flags & CE_FSMONITOR_VALID)) {
+		istate->cache_changed = 1;
 		ce->ce_flags |= CE_FSMONITOR_VALID;
 		trace_printf_key(&trace_fsmonitor, "mark_fsmonitor_clean '%s'", ce->name);
 	}
preload-index.c +1 -1
index e73600ee78..ed6eaa4738 100644
@@ -78,7 +78,7 @@ static void *preload_thread(void *_data)
 		if (ie_match_stat(index, ce, &st, CE_MATCH_RACY_IS_DIRTY|CE_MATCH_IGNORE_FSMONITOR))
 			continue;
 		ce_mark_uptodate(ce);
-		mark_fsmonitor_valid(ce);
+		mark_fsmonitor_valid(index, ce);
 	} while (--nr > 0);
 	if (p->progress) {
 		struct progress_data *pd = p->progress;
read-cache.c +2 -2
index d3b43ae3ba..32816763bd 100644
@@ -204,7 +204,7 @@ void fill_stat_cache_info(struct index_state *istate, struct cache_entry *ce, st
 
 	if (S_ISREG(st->st_mode)) {
 		ce_mark_uptodate(ce);
-		mark_fsmonitor_valid(ce);
+		mark_fsmonitor_valid(istate, ce);
 	}
 }
 

@@ -1432,7 +1432,7 @@ static struct cache_entry *refresh_cache_ent(struct index_state *istate,
 			 */
 			if (!S_ISGITLINK(ce->ce_mode)) {
 				ce_mark_uptodate(ce);
-				mark_fsmonitor_valid(ce);
+				mark_fsmonitor_valid(istate, ce);
 			}
 			return ce;
 		}
Signature · 1 line
gitgitgadget