[PATCH 1/3] python3: Avoid hanging test

* [PATCH 1/3] python3: Avoid hanging test
@ 2019-04-08 11:59 Richard Purdie
  2019-04-08 11:59 ` [PATCH 2/3] openssh/util-linux/python*: Ensure ptest output is unbuffered Richard Purdie
  2019-04-08 11:59 ` [PATCH 3/3] ptest-runner: Add several logging fixes Richard Purdie
  0 siblings, 2 replies; 4+ messages in thread
From: Richard Purdie @ 2019-04-08 11:59 UTC (permalink / raw)
  To: openembedded-core

There is a python test which hangs with recent kernels, 5.0 onwards. This causes
ptest to timeout for python3. Disable the problematic test until we better understand
the real cause and fix of the issue (discussions are happening with upstream).

See the patch for details/links.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
---
 .../python/python3/ptesthack.patch            | 49 +++++++++++++++++++
 meta/recipes-devtools/python/python3_3.7.2.bb |  1 +
 2 files changed, 50 insertions(+)
 create mode 100644 meta/recipes-devtools/python/python3/ptesthack.patch

diff --git a/meta/recipes-devtools/python/python3/ptesthack.patch b/meta/recipes-devtools/python/python3/ptesthack.patch
new file mode 100644
index 00000000000..3f5589645ed
--- /dev/null
+++ b/meta/recipes-devtools/python/python3/ptesthack.patch
@@ -0,0 +1,49 @@
+This test hangs under 5.0 kernels onwards. It appears to be caused by the commit in the kernel:
+
+commit 4f693b55c3d2d2239b8a0094b518a1e533cf75d5 (HEAD, refs/bisect/bad)
+Author: Eric Dumazet <edumazet@google.com>
+Date:   Tue Nov 27 14:42:03 2018 -0800
+
+    tcp: implement coalescing on backlog queue
+    
+    In case GRO is not as efficient as it should be or disabled,
+    we might have a user thread trapped in __release_sock() while
+    softirq handler flood packets up to the point we have to drop.
+    
+    This patch balances work done from user thread and softirq,
+    to give more chances to __release_sock() to complete its work
+    before new packets are added the the backlog.
+    
+    This also helps if we receive many ACK packets, since GRO
+    does not aggregate them.
+    
+    This patch brings ~60% throughput increase on a receiver
+    without GRO, but the spectacular gain is really on
+    1000x release_sock() latency reduction I have measured.
+    
+    Signed-off-by: Eric Dumazet <edumazet@google.com>
+    Cc: Neal Cardwell <ncardwell@google.com>
+    Cc: Yuchung Cheng <ycheng@google.com>
+    Acked-by: Neal Cardwell <ncardwell@google.com>
+    Signed-off-by: David S. Miller <davem@davemloft.net>
+
+
+Reported to upstream kernel for advice: https://lore.kernel.org/netdev/85aabf9d4f41b6c57629e736993233f80a037e59.camel@linuxfoundation.org/T/#u
+
+Disable the test for now to stop ptests hanging
+
+Upstream-Status: Inappropriate [real cause of issue still TBD]
+
+Index: Python-3.7.2/Lib/test/test_httplib.py
+===================================================================
+--- Python-3.7.2.orig/Lib/test/test_httplib.py
++++ Python-3.7.2/Lib/test/test_httplib.py
+@@ -1114,7 +1114,7 @@ class BasicTest(TestCase):
+         self.assertEqual(sock.file.read(), extradata) #we read to the end
+         resp.close()
+ 
+-    def test_response_fileno(self):
++    def atest_response_fileno(self):
+         # Make sure fd returned by fileno is valid.
+         serv = socket.socket(
+             socket.AF_INET, socket.SOCK_STREAM, socket.IPPROTO_TCP)
diff --git a/meta/recipes-devtools/python/python3_3.7.2.bb b/meta/recipes-devtools/python/python3_3.7.2.bb
index 5c64bc8aa2a..3d44bdca911 100644
--- a/meta/recipes-devtools/python/python3_3.7.2.bb
+++ b/meta/recipes-devtools/python/python3_3.7.2.bb
@@ -21,6 +21,7 @@ SRC_URI = "http://www.python.org/ftp/python/${PV}/Python-${PV}.tar.xz \
            file://0001-python3-use-cc_basename-to-replace-CC-for-checking-c.patch \
            file://0002-Don-t-do-runtime-test-to-get-float-byte-order.patch \
            file://0003-setup.py-pass-missing-libraries-to-Extension-for-mul.patch \
+           file://ptesthack.patch \
            "
 
 SRC_URI_append_class-native = " \
-- 
2.20.1



^ permalink raw reply related	[flat|nested] 4+ messages in thread