git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
* [PATCH] perl Git.pm: don't ignore signalled failure in _cmd_close()
@ 2022-02-01 20:52 Ævar Arnfjörð Bjarmason
  2022-02-01 21:58 ` Junio C Hamano
  0 siblings, 1 reply; 2+ messages in thread
From: Ævar Arnfjörð Bjarmason @ 2022-02-01 20:52 UTC (permalink / raw)
  To: git; +Cc: Junio C Hamano, Eric Wong, Ævar Arnfjörð Bjarmason

Fix misbehavior in Git.pm that dates back to the very first version of
the library in git.git added in b1edc53d062 (Introduce Git.pm (v4),
2006-06-24). When we fail to execute a command we shouldn't ignore all
signals, those can happen e.g. if abort() is called, or if the command
segfaults.

Because of this we'd consider e.g. a command that died due to LSAN
exiting with abort() successful, as is the case with the tests listed
as running successfully with SANITIZE=leak in 9081a421a6d (checkout:
fix "branch info" memory leaks, 2021-11-16). We did run them
successfully, but only because we ignored these errors.

This was then made worse by the use of "abort_on_error=1" for LSAN
added in 85b81b35ff9 (test-lib: set LSAN_OPTIONS to abort by default,
2017-09-05). Doing that makes sense, but without providing that option
we'd have a "$? >> 8" of "23" on failure, with abort_on_error=1 we'll
get "0".

All of our tests pass even without the SIGPIPE exception being added
here, but as the code appears to have been trying to ignore it let's
keep ignoring it for now.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
---
 perl/Git.pm                                 | 21 +++++++++++++++++++--
 t/t9102-git-svn-deep-rmdir.sh               |  1 -
 t/t9123-git-svn-rebuild-with-rewriteroot.sh |  1 -
 t/t9128-git-svn-cmd-branch.sh               |  1 -
 t/t9167-git-svn-cmd-branch-subproject.sh    |  1 -
 5 files changed, 19 insertions(+), 6 deletions(-)

diff --git a/perl/Git.pm b/perl/Git.pm
index 090a7df63fc..080cdc2a21d 100644
--- a/perl/Git.pm
+++ b/perl/Git.pm
@@ -1686,6 +1686,16 @@ sub _setup_git_cmd_env {
 # by searching for it at proper places.
 sub _execv_git_cmd { exec('git', @_); }
 
+sub _is_sig {
+	my ($v, $n) = @_;
+
+	# We are avoiding a "use POSIX qw(SIGPIPE SIGABRT)" in the hot
+	# Git.pm codepath.
+	require POSIX;
+	no strict 'refs';
+	$v == *{"POSIX::$n"}->();
+}
+
 # Close pipe to a subprocess.
 sub _cmd_close {
 	my $ctx = shift @_;
@@ -1698,9 +1708,16 @@ sub _cmd_close {
 		} elsif ($? >> 8) {
 			# The caller should pepper this.
 			throw Git::Error::Command($ctx, $? >> 8);
+		} elsif ($? & 127 && _is_sig($? & 127, "SIGPIPE")) {
+			# we might e.g. closed a live stream; the command
+			# dying of SIGPIPE would drive us here.
+		} elsif ($? & 127 && _is_sig($? & 127, "SIGABRT")) {
+			die sprintf('BUG?: got SIGABRT ($? = %d, $? & 127 = %d) when closing pipe',
+				    $?, $? & 127);
+		} elsif ($? & 127) {
+			die sprintf('got signal ($? = %d, $? & 127 = %d) when closing pipe',
+				    $?, $? & 127);
 		}
-		# else we might e.g. closed a live stream; the command
-		# dying of SIGPIPE would drive us here.
 	}
 }
 
diff --git a/t/t9102-git-svn-deep-rmdir.sh b/t/t9102-git-svn-deep-rmdir.sh
index 7b2049caa0c..946ef85eb98 100755
--- a/t/t9102-git-svn-deep-rmdir.sh
+++ b/t/t9102-git-svn-deep-rmdir.sh
@@ -1,7 +1,6 @@
 #!/bin/sh
 test_description='git svn rmdir'
 
-TEST_PASSES_SANITIZE_LEAK=true
 . ./lib-git-svn.sh
 
 test_expect_success 'initialize repo' '
diff --git a/t/t9123-git-svn-rebuild-with-rewriteroot.sh b/t/t9123-git-svn-rebuild-with-rewriteroot.sh
index 3320b1f39cf..ead404589eb 100755
--- a/t/t9123-git-svn-rebuild-with-rewriteroot.sh
+++ b/t/t9123-git-svn-rebuild-with-rewriteroot.sh
@@ -5,7 +5,6 @@
 
 test_description='git svn respects rewriteRoot during rebuild'
 
-TEST_PASSES_SANITIZE_LEAK=true
 . ./lib-git-svn.sh
 
 mkdir import
diff --git a/t/t9128-git-svn-cmd-branch.sh b/t/t9128-git-svn-cmd-branch.sh
index 9871f5abc93..783e3ba0c53 100755
--- a/t/t9128-git-svn-cmd-branch.sh
+++ b/t/t9128-git-svn-cmd-branch.sh
@@ -5,7 +5,6 @@
 
 test_description='git svn partial-rebuild tests'
 
-TEST_PASSES_SANITIZE_LEAK=true
 . ./lib-git-svn.sh
 
 test_expect_success 'initialize svnrepo' '
diff --git a/t/t9167-git-svn-cmd-branch-subproject.sh b/t/t9167-git-svn-cmd-branch-subproject.sh
index d9fd111c105..d8128430a8d 100755
--- a/t/t9167-git-svn-cmd-branch-subproject.sh
+++ b/t/t9167-git-svn-cmd-branch-subproject.sh
@@ -5,7 +5,6 @@
 
 test_description='git svn branch for subproject clones'
 
-TEST_PASSES_SANITIZE_LEAK=true
 . ./lib-git-svn.sh
 
 test_expect_success 'initialize svnrepo' '
-- 
2.35.0.913.g12b4baa2536


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH] perl Git.pm: don't ignore signalled failure in _cmd_close()
  2022-02-01 20:52 [PATCH] perl Git.pm: don't ignore signalled failure in _cmd_close() Ævar Arnfjörð Bjarmason
@ 2022-02-01 21:58 ` Junio C Hamano
  0 siblings, 0 replies; 2+ messages in thread
From: Junio C Hamano @ 2022-02-01 21:58 UTC (permalink / raw)
  To: Ævar Arnfjörð Bjarmason; +Cc: git, Eric Wong

Ævar Arnfjörð Bjarmason  <avarab@gmail.com> writes:

> Fix misbehavior in Git.pm that dates back to the very first version of
> the library in git.git added in b1edc53d062 (Introduce Git.pm (v4),
> 2006-06-24). When we fail to execute a command we shouldn't ignore all
> signals, those can happen e.g. if abort() is called, or if the command
> segfaults.

Good find.

> All of our tests pass even without the SIGPIPE exception being added
> here, but as the code appears to have been trying to ignore it let's
> keep ignoring it for now.
>

Sounds OK-ish to me.

> +sub _is_sig {
> +	my ($v, $n) = @_;
> +
> +	# We are avoiding a "use POSIX qw(SIGPIPE SIGABRT)" in the hot
> +	# Git.pm codepath.
> +	require POSIX;
> +	no strict 'refs';
> +	$v == *{"POSIX::$n"}->();
> +}




^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2022-02-01 21:58 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-02-01 20:52 [PATCH] perl Git.pm: don't ignore signalled failure in _cmd_close() Ævar Arnfjörð Bjarmason
2022-02-01 21:58 ` Junio C Hamano

Code repositories for project(s) associated with this public inbox

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).