File: closures-test-Avoid-timeout-on-ARM64-CPUs.patch

package info (click to toggle)
glib2.0 2.58.3-1
  • links: PTS, VCS
  • area: main
  • in suites: bullseye, sid
  • size: 48,444 kB
  • sloc: ansic: 452,195; xml: 16,781; python: 6,149; makefile: 3,776; sh: 1,499; perl: 1,140; cpp: 9
file content (73 lines) | stat: -rw-r--r-- 3,127 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
From: Simon McVittie <smcv@debian.org>
Date: Mon, 18 Dec 2017 10:47:29 +0000
Subject: closures test: Avoid timeout on ARM64 CPUs

Closures use a 16-bit atomic reference count, which is really slow
on certain ARM64 CPUs such as the Cortex-A57 (glib#1316). This is
non-trivial to solve, since the public struct field cannot be enlarged
to 32-bit while preserving ABI, and 16-bit atomic operations would be new
(and rather niche) API.

Until this can be solved properly (hopefully in GLib 2.59.x), cut down
the number of signal emission cycles and bump up the timeout in the
Meson build system, so that builds won't time out. We can't just take
another zero off the number of signal emission cycles, as was done in the
original version of this patch in Debian, because if we do that it can
result in test failures when the main thread starves the other threads.

ARM64 CPUs are backwards-compatible with 32-bit ARM, and the same
slowdown can be seen when building and testing 32-bit code on these
CPUs, so check for both 32- and 64-bit ARM.

Bug-Debian: https://bugs.debian.org/880883
Forwarded: https://gitlab.gnome.org/GNOME/glib/merge_requests/451
Applied-upstream: 2.59.0, commit:54549b7e4
Co-authored-by: Iain Lane <laney@debian.org>
Signed-off-by: Simon McVittie <smcv@debian.org>
---
 tests/refcount/closures.c  | 12 ++++++++++++
 tests/refcount/meson.build |  7 +++++++
 2 files changed, 19 insertions(+)

diff --git a/tests/refcount/closures.c b/tests/refcount/closures.c
index 25b0a27..1884067 100644
--- a/tests/refcount/closures.c
+++ b/tests/refcount/closures.c
@@ -259,7 +259,19 @@ main (int    argc,
   thread1 = g_thread_create (thread1_main, closure, TRUE, NULL);
   thread2 = g_thread_create (thread2_main, closure, TRUE, NULL);
 
+  /* The 16-bit compare-and-swap operations currently used for closure
+   * refcounts are really slow on some ARM CPUs, notably Cortex-A57.
+   * Reduce the number of iterations so that the test completes in a
+   * finite time, but don't reduce it so much that the main thread
+   * starves the other threads and causes a test failure.
+   *
+   * https://gitlab.gnome.org/GNOME/glib/issues/1316
+   * aka https://bugs.debian.org/880883 */
+#if defined(__aarch64__) || defined(__arm__)
+  for (i = 0; i < 100000; i++)
+#else
   for (i = 0; i < 1000000; i++)
+#endif
     {
       static guint count = 0;
       test_emissions (object);
diff --git a/tests/refcount/meson.build b/tests/refcount/meson.build
index 68c23cc..22b9655 100644
--- a/tests/refcount/meson.build
+++ b/tests/refcount/meson.build
@@ -54,6 +54,13 @@ foreach test_name, extra_args : refcount_tests
 
   suite = ['refcount'] + extra_args.get('suite', [])
   timeout = suite.contains('slow') ? test_timeout_slow : test_timeout
+
+  # https://gitlab.gnome.org/GNOME/glib/issues/1316
+  # aka https://bugs.debian.org/880883
+  if test_name == 'closures' and ['arm', 'aarch64'].contains(host_cpu_family)
+    timeout = timeout * 10
+  endif
+
   # FIXME? TESTS_ENVIRONMENT = LIBCHARSET_ALIAS_DIR=$(top_builddir)/glib/libcharset
   test(test_name, exe, env : test_env, timeout : timeout, suite : suite)
 endforeach