[Qemu-devel] [PATCH 00/18] Acceptance Tests: target architecture support

[Qemu-devel] [PATCH 00/18] Acceptance Tests: target architecture support

[Cleber Rosa]

The current acceptance tests don't provide any type of architecture
information that can be used to influence the selection of the QEMU
binary used on them[1].  If one is running tests on a x86_64 host, the
default QEMU binary will be "x86_64-softmmu/qemu-system-x86_64".

Given the nature of QEMU, some tests will be architecture agnostic,
while others will be architecture dependent.  The "check-qtest" and
"check-qtest-TARGET" make targets exemplify that pattern.

For the acceptance tests, the same requirement exists.  Tests should
be allowed to influence the binary used, and when they don't, a
default selection mechanism should kick in[2].  The proposed solution
here requires only that an Avocado tag is set, such as:

   class My(Test):
       def test_nx_cpu_flag(self):
           """
           :avocado: tags=arch:x86_64
           """
           test_code()

The value of the "arch" key, in this case, "x86_64" will be used when
selecting the QEMU binary to use in the test.  At the same time, if
"x86_64-softmmu" is not a built target, the test will be filtered out
by "make check-acceptance"[3].

Besides the convention explained above, where the binary will be
selected from the "arch" tag, it's also possible to set an "arch"
*parameter* that will also influence the QEMU binary selection:

  $ avocado run -p arch=aarch64 works-on-many-arches.py

Finally, it's also posible to set the "qemu_bin" parameter, which will
define (instead of just influencing) the QEMU binary to be used:

 $ avocado run -p qemu_bin=qemu-bin-aarch64 test.py

As examples for the idea proposed here, a number of "boot linux
console" tests have been added, for a number of different target
architectures.  When the build environment includes them (as it has
been added to Travis CI jobs) the architecture specific tests will be
automatically executed.

As mentioned previously, this patch series include ideas present in
other patch series, and from different authors.  I tried by best
to include the information about authorship, but if I missed any,
please accept my apologies and let me know.

---

[1] - The "boot_linux_console.py" contains a "x86_64" test tag, but
      that is informational only, in the sense that it's not consumed
      by the test itself, or used by "make check-acceptance" to filter
      out tests.

[2] - This patch series doesn't attempt to change the default selection
      mechanism.  Possible changes in this area may include looking for
      any one built binary first, no matter the host architecture.

[3] - On a previous proposed version, the test class would look at the
      "arch" parameter given, and would cancel the test if there wasn't
      a match.

---

Git Info:
  - URI: https://github.com/clebergnu/qemu/tree/sent/target_arch
  - Remote: https://github.com/clebergnu/qemu
  - Branch: sent/target_arch

Cleber Rosa (17):
  scripts/qemu.py: log QEMU launch command line
  Acceptance tests: show avocado test execution by default
  Acceptance tests: improve docstring on pick_default_qemu_bin()
  Acceptance tests: fix doc reference to avocado_qemu directory
  Acceptance tests: introduce arch parameter and attribute
  Acceptance tests: use "arch:" tag to filter target specific tests
  Acceptance tests: look for target architecture in test tags first
  Boot Linux Console Test: rename the x86_64 after the arch and machine
  Boot Linux Console Test: update the x86_64 kernel
  Boot Linux Console Test: refactor the console watcher into utility
    method
  scripts/qemu.py: support adding a console with the default serial
    device
  Boot Linux Console Test: add a test for mips64el + malta
  Boot Linux Console Test: add a test for ppc64 + pseries
  Boot Linux Console Test: add a test for aarch64 + virt
  Boot Linux Console Test: add a test for arm + virt
  Boot Linux Console Test: add a test for s390x + s390-ccw-virtio
  Boot Linux Console Test: add a test for alpha + clipper

Philippe Mathieu-Daudé (1):
  Boot Linux Console Test: add a test for mips + malta

 .travis.yml                               |   4 +-
 docs/devel/testing.rst                    |  23 ++-
 scripts/qemu.py                           |  32 ++--
 tests/Makefile.include                    |   5 +-
 tests/acceptance/avocado_qemu/__init__.py |  22 ++-
 tests/acceptance/boot_linux_console.py    | 208 ++++++++++++++++++++--
 tests/requirements.txt                    |   2 +-
 7 files changed, 254 insertions(+), 42 deletions(-)

-- 
2.20.1

[Qemu-devel] [PATCH 01/18] scripts/qemu.py: log QEMU launch command line

[Cleber Rosa]

Even when the launch of QEMU succeeds, it's useful to have the command
line recorded.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 scripts/qemu.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/scripts/qemu.py b/scripts/qemu.py
index 6e3b0e6771..ec3567d4e2 100644
--- a/scripts/qemu.py
+++ b/scripts/qemu.py
@@ -312,6 +312,7 @@ class QEMUMachine(object):
         self._pre_launch()
         self._qemu_full_args = (self._wrapper + [self._binary] +
                                 self._base_args() + self._args)
+        LOG.debug('VM launch command: %r', ' '.join(self._qemu_full_args))
         self._popen = subprocess.Popen(self._qemu_full_args,
                                        stdin=devnull,
                                        stdout=self._qemu_log_file,
-- 
2.20.1

[Qemu-devel] [PATCH 02/18] Acceptance tests: show avocado test execution by default

[Cleber Rosa]

The current version of the "check-acceptance" target will only show
one line for execution of all tests.  That's probably OK if the tests
to be run are quick enough and they're always the same.

But, there's already one test alone that takes on average ~5 seconds
to run, we intend to adapt the list of tests to match the user's build
environment (among other choices).

Because of that, let's present the default Avocado UI by default.
Users can always choose a different output by setting the AVOCADO_SHOW
variable.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 .travis.yml            | 2 +-
 tests/Makefile.include | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/.travis.yml b/.travis.yml
index 93fd0164a0..844d514afa 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -188,7 +188,7 @@ matrix:
     # Acceptance (Functional) tests
     - env:
         - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu"
-        - TEST_CMD="make AVOCADO_SHOW=app check-acceptance"
+        - TEST_CMD="make check-acceptance"
       addons:
         apt:
           packages:
diff --git a/tests/Makefile.include b/tests/Makefile.include
index f403a6571d..c73298740d 100644
--- a/tests/Makefile.include
+++ b/tests/Makefile.include
@@ -956,7 +956,7 @@ TESTS_RESULTS_DIR=$(BUILD_DIR)/tests/results
 # Controls the output generated by Avocado when running tests.
 # Any number of command separated loggers are accepted.  For more
 # information please refer to "avocado --help".
-AVOCADO_SHOW=none
+AVOCADO_SHOW=app
 
 PYTHON3 = $(shell $(PYTHON) -c 'import sys; print(1 if sys.version_info >= (3, 0) else 0)')
 ifeq ($(PYTHON3), 1)
-- 
2.20.1

[Qemu-devel] [PATCH 03/18] Acceptance tests: improve docstring on pick_default_qemu_bin()

[Cleber Rosa]

Making it clear what is returned by this utility function.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 tests/acceptance/avocado_qemu/__init__.py | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/tests/acceptance/avocado_qemu/__init__.py b/tests/acceptance/avocado_qemu/__init__.py
index 1e54fd5932..d8d5b48dac 100644
--- a/tests/acceptance/avocado_qemu/__init__.py
+++ b/tests/acceptance/avocado_qemu/__init__.py
@@ -27,6 +27,10 @@ def pick_default_qemu_bin():
     """
     Picks the path of a QEMU binary, starting either in the current working
     directory or in the source tree root directory.
+
+    :returns: the path to the default QEMU binary or None if one could not
+              be found
+    :rtype: str or None
     """
     arch = os.uname()[4]
     qemu_bin_relative_path = os.path.join("%s-softmmu" % arch,
-- 
2.20.1

[Qemu-devel] [PATCH 04/18] Acceptance tests: fix doc reference to avocado_qemu directory

[Cleber Rosa]

The "this directory" reference is misleading and confusing, it's a
leftover from when this text was proposed in a README file inside
the "tests/acceptance/avocado_qemu" directory.

When that text was moved to the top level docs directory, the
reference was not updated.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 docs/devel/testing.rst | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
index 18e2c0868a..44c9b3ae74 100644
--- a/docs/devel/testing.rst
+++ b/docs/devel/testing.rst
@@ -590,8 +590,9 @@ Alternatively, follow the instructions on this link:
 Overview
 --------
 
-This directory provides the ``avocado_qemu`` Python module, containing
-the ``avocado_qemu.Test`` class.  Here's a simple usage example:
+The ``tests/acceptance/avocado_qemu`` directory provides the
+``avocado_qemu`` Python module, containing the ``avocado_qemu.Test``
+class.  Here's a simple usage example:
 
 .. code::
 
-- 
2.20.1

[Qemu-devel] [PATCH 05/18] Acceptance tests: introduce arch parameter and attribute

[Cleber Rosa]

It's useful to define the architecture that should be used in
situations such as:
 * the intended target of the QEMU binary to be used on tests
 * the architecture of code to be run within the QEMU binary, such
   as a kernel image or a full blown guest OS image

This commit introduces both a test parameter and a test instance
attribute, that will contain such a value.

Now, when the "arch" test parameter is given, it will influence the
selection of the default QEMU binary, if one is not given explicitly
by means of the "qemu_img" parameter.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 docs/devel/testing.rst                    | 18 ++++++++++++++++++
 tests/acceptance/avocado_qemu/__init__.py | 13 ++++++++++---
 2 files changed, 28 insertions(+), 3 deletions(-)

diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
index 44c9b3ae74..60a16c5834 100644
--- a/docs/devel/testing.rst
+++ b/docs/devel/testing.rst
@@ -689,6 +689,17 @@ vm
 A QEMUMachine instance, initially configured according to the given
 ``qemu_bin`` parameter.
 
+arch
+~~~~
+
+The architecture that will be used on a number of different
+scenarios.  For instance, when a QEMU binary is not explicitly given,
+the one selected will depend on this attribute.
+
+The ``arch`` attribute will be set to the test parameter of the same
+name, and if one is not given explicitly, it will default to the
+system architecture (as given by ``uname``).
+
 qemu_bin
 ~~~~~~~~
 
@@ -711,6 +722,13 @@ like the following:
 
   PARAMS (key=qemu_bin, path=*, default=x86_64-softmmu/qemu-system-x86_64) => 'x86_64-softmmu/qemu-system-x86_64
 
+arch
+~~~~
+
+The architecture that will be used on a number of different scenarios.
+This parameter has a direct relation with the ``arch`` attribute.  If
+not given, it will default to None.
+
 qemu_bin
 ~~~~~~~~
 
diff --git a/tests/acceptance/avocado_qemu/__init__.py b/tests/acceptance/avocado_qemu/__init__.py
index d8d5b48dac..7a38851753 100644
--- a/tests/acceptance/avocado_qemu/__init__.py
+++ b/tests/acceptance/avocado_qemu/__init__.py
@@ -23,16 +23,22 @@ def is_readable_executable_file(path):
     return os.path.isfile(path) and os.access(path, os.R_OK | os.X_OK)
 
 
-def pick_default_qemu_bin():
+def pick_default_qemu_bin(arch=None):
     """
     Picks the path of a QEMU binary, starting either in the current working
     directory or in the source tree root directory.
 
+    :param arch: the arch to use when looking for a QEMU binary (the target
+                 will match the arch given).  If None (the default) arch
+                 will be the current host system arch (as given by
+                 :func:`os.uname`).
+    :type arch: str
     :returns: the path to the default QEMU binary or None if one could not
               be found
     :rtype: str or None
     """
-    arch = os.uname()[4]
+    if arch is None:
+        arch = os.uname()[4]
     qemu_bin_relative_path = os.path.join("%s-softmmu" % arch,
                                           "qemu-system-%s" % arch)
     if is_readable_executable_file(qemu_bin_relative_path):
@@ -47,8 +53,9 @@ def pick_default_qemu_bin():
 class Test(avocado.Test):
     def setUp(self):
         self.vm = None
+        self.arch = self.params.get('arch')
         self.qemu_bin = self.params.get('qemu_bin',
-                                        default=pick_default_qemu_bin())
+                                        default=pick_default_qemu_bin(self.arch))
         if self.qemu_bin is None:
             self.cancel("No QEMU binary defined or found in the source tree")
         self.vm = QEMUMachine(self.qemu_bin)
-- 
2.20.1

[Qemu-devel] [PATCH 06/18] Acceptance tests: use "arch:" tag to filter target specific tests

[Cleber Rosa]

Currently, the only test that contains some target architecture
information is "boot_linux_console.py" which test contains a "x86_64"
tag.  But that tag is not respected in the default execution, that is,
"make check-acceptance" doesn't do anything with it.

That said, even the target architecture handling currently present in
the "avocado_qemu.Test" class, class is pretty limited.  For instance,
by default, it chooses a target based on the host architecture.

Because the original implementation of the tags feature in Avocado did
not include any time of namespace or "key:val" mechanism, no tag has
relation to another tag.  The new implementation of the tags feature
from version 67.0 onwards, allows "key:val" tags, and because of that,
a test can be classified with a tag in a given key.  For instance, the
new proposed version of the "boot_linux_console.py" test, which
downloads and attempts to run a x86_64 kernel, is now tagged as:

  :avocado: tags=arch:x86_64

This means that it can be filtered (out) when no x86_64 target is
available.  At the same time, tests that don't have a "arch:" tag,
will not be filtered out.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 tests/Makefile.include                 | 3 +++
 tests/acceptance/boot_linux_console.py | 2 +-
 tests/requirements.txt                 | 2 +-
 3 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/tests/Makefile.include b/tests/Makefile.include
index c73298740d..af00c35ff5 100644
--- a/tests/Makefile.include
+++ b/tests/Makefile.include
@@ -957,6 +957,7 @@ TESTS_RESULTS_DIR=$(BUILD_DIR)/tests/results
 # Any number of command separated loggers are accepted.  For more
 # information please refer to "avocado --help".
 AVOCADO_SHOW=app
+AVOCADO_TAGS=$(patsubst %-softmmu,-t arch:%, $(filter %-softmmu,$(TARGET_DIRS)))
 
 PYTHON3 = $(shell $(PYTHON) -c 'import sys; print(1 if sys.version_info >= (3, 0) else 0)')
 ifeq ($(PYTHON3), 1)
@@ -983,6 +984,8 @@ check-acceptance: check-venv $(TESTS_RESULTS_DIR)
 	$(call quiet-command, \
             $(TESTS_VENV_DIR)/bin/python -m avocado \
             --show=$(AVOCADO_SHOW) run --job-results-dir=$(TESTS_RESULTS_DIR) \
+            --filter-by-tags-include-empty --filter-by-tags-include-empty-key \
+            $(AVOCADO_TAGS) \
             --failfast=on $(SRC_PATH)/tests/acceptance, \
             "AVOCADO", "tests/acceptance")
 
diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index 98324f7591..46b20bdfe2 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -19,7 +19,7 @@ class BootLinuxConsole(Test):
     and the kernel command line is properly passed from QEMU to the kernel
 
     :avocado: enable
-    :avocado: tags=x86_64
+    :avocado: tags=arch:x86_64
     """
 
     timeout = 60
diff --git a/tests/requirements.txt b/tests/requirements.txt
index 64c6e27a94..426a9e8a9a 100644
--- a/tests/requirements.txt
+++ b/tests/requirements.txt
@@ -1,4 +1,4 @@
 # Add Python module requirements, one per line, to be installed
 # in the tests/venv Python virtual environment. For more info,
 # refer to: https://pip.pypa.io/en/stable/user_guide/#id1
-avocado-framework==65.0
+avocado-framework==67.0
-- 
2.20.1

[Qemu-devel] [PATCH 07/18] Acceptance tests: look for target architecture in test tags first

[Cleber Rosa]

A test can, optionally, be tagged for one or many architectures.  If a
test has been tagged for a single architecture, there's a high chance
that the test won't run on other architectures.  This changes the
default order of choosing a default target architecture to use based
on the 'arch' tag value first.

The precedence order is for choosing a QEMU binary to use for a test
is now:

 * qemu_bin parameter
 * arch parameter
 * arch tag value (for example, x86_64 if ":avocado: tags=arch:x86_64
   is used)

This means that if one runs:

 $ avocado run -p qemu_bin=/usr/bin/qemu-system-x86_64 test.py

No arch parameter or tag will influence the selection of the QEMU
target binary.  If one runs:

 $ avocado run -p arch=ppc64 test.py

The target binary selection mechanism will attempt to find a binary
such as "ppc64-softmmu/qemu-system-ppc64".  And finally, if one runs
a test that is tagged (in its docstring) with "arch:aarch64":

 $ avocado run aarch64.py

The target binary selection mechanism will attempt to find a binary
such as "aarch64-softmmu/qemu-system-aarch64".

At this time, no provision is made to cancel the execution of tests if
the arch parameter given (manually) does not match the test "arch"
tag, but it may be a useful default behavior to be added in the
future.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 tests/acceptance/avocado_qemu/__init__.py | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/tests/acceptance/avocado_qemu/__init__.py b/tests/acceptance/avocado_qemu/__init__.py
index 7a38851753..ddbdb7b926 100644
--- a/tests/acceptance/avocado_qemu/__init__.py
+++ b/tests/acceptance/avocado_qemu/__init__.py
@@ -53,7 +53,12 @@ def pick_default_qemu_bin(arch=None):
 class Test(avocado.Test):
     def setUp(self):
         self.vm = None
-        self.arch = self.params.get('arch')
+        arches = self.tags.get('arch', [])
+        if len(arches) == 1:
+            arch = arches.pop()
+        else:
+            arch = None
+        self.arch = self.params.get('arch', default=arch)
         self.qemu_bin = self.params.get('qemu_bin',
                                         default=pick_default_qemu_bin(self.arch))
         if self.qemu_bin is None:
-- 
2.20.1

[Qemu-devel] [PATCH 08/18] Boot Linux Console Test: rename the x86_64 after the arch and machine

[Cleber Rosa]

Given that the test is specific to x86_64 and pc, and new tests are
going to be added to the same class, let's rename it accordingly.
Also, let's make the class documentation not architecture specific.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 tests/acceptance/boot_linux_console.py | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index 46b20bdfe2..89df7f6e4f 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -15,16 +15,19 @@ from avocado_qemu import Test
 
 class BootLinuxConsole(Test):
     """
-    Boots a x86_64 Linux kernel and checks that the console is operational
-    and the kernel command line is properly passed from QEMU to the kernel
+    Boots a Linux kernel and checks that the console is operational and the
+    kernel command line is properly passed from QEMU to the kernel
 
     :avocado: enable
-    :avocado: tags=arch:x86_64
     """
 
     timeout = 60
 
-    def test(self):
+    def test_x86_64_pc(self):
+        """
+        :avocado: tags=arch:x86_64
+        :avocado: tags=machine:pc
+        """
         kernel_url = ('https://mirrors.kernel.org/fedora/releases/28/'
                       'Everything/x86_64/os/images/pxeboot/vmlinuz')
         kernel_hash = '238e083e114c48200f80d889f7e32eeb2793e02a'
-- 
2.20.1

[Qemu-devel] [PATCH 09/18] Boot Linux Console Test: update the x86_64 kernel

[Cleber Rosa]

To the stock Fedora 29 kernel, from the Fedora 28.  New tests will be
added using the 29 kernel, so for consistency, let's also update it
here.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 tests/acceptance/boot_linux_console.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index 89df7f6e4f..35b31162d4 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -28,9 +28,9 @@ class BootLinuxConsole(Test):
         :avocado: tags=arch:x86_64
         :avocado: tags=machine:pc
         """
-        kernel_url = ('https://mirrors.kernel.org/fedora/releases/28/'
+        kernel_url = ('https://mirrors.kernel.org/fedora/releases/29/'
                       'Everything/x86_64/os/images/pxeboot/vmlinuz')
-        kernel_hash = '238e083e114c48200f80d889f7e32eeb2793e02a'
+        kernel_hash = '23bebd2680757891cf7adedb033532163a792495'
         kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
 
         self.vm.set_machine('pc')
-- 
2.20.1

[Qemu-devel] [PATCH 10/18] Boot Linux Console Test: refactor the console watcher into utility method

[Cleber Rosa]

This introduces a utility method that monitors the console device and
looks for either a message that signals the test success or failure.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 tests/acceptance/boot_linux_console.py | 30 ++++++++++++++++++--------
 1 file changed, 21 insertions(+), 9 deletions(-)

diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index 35b31162d4..278bb2be3d 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -23,6 +23,25 @@ class BootLinuxConsole(Test):
 
     timeout = 60
 
+    def wait_for_console_pattern(self, success_message,
+                                 failure_message='Kernel panic - not syncing'):
+        """
+        Waits for messages to appear on the console, while logging the content
+
+        :param success_message: if this message appears, test succeeds
+        :param failure_message: if this message appears, test fails
+        """
+        console = self.vm.console_socket.makefile()
+        console_logger = logging.getLogger('console')
+        while True:
+            msg = console.readline()
+            console_logger.debug(msg.strip())
+            if success_message in msg:
+                break
+            if failure_message in msg:
+                fail = 'Failure message found in console: %s' % failure_message
+                self.fail(fail)
+
     def test_x86_64_pc(self):
         """
         :avocado: tags=arch:x86_64
@@ -39,12 +58,5 @@ class BootLinuxConsole(Test):
         self.vm.add_args('-kernel', kernel_path,
                          '-append', kernel_command_line)
         self.vm.launch()
-        console = self.vm.console_socket.makefile()
-        console_logger = logging.getLogger('console')
-        while True:
-            msg = console.readline()
-            console_logger.debug(msg.strip())
-            if 'Kernel command line: %s' % kernel_command_line in msg:
-                break
-            if 'Kernel panic - not syncing' in msg:
-                self.fail("Kernel panic reached")
+        console_pattern = 'Kernel command line: %s' % kernel_command_line
+        self.wait_for_console_pattern(console_pattern)
-- 
2.20.1

[Qemu-devel] [PATCH 11/18] scripts/qemu.py: support adding a console with the default serial device

[Cleber Rosa]

The set_console() utility function traditionally adds a device either
based on the explicitly given device type, or based on the machine type,
a known good type of device.

But, for a number of machine types, it may be impossible or
inconvenient to add the devices my means of "-device" command line
options, and then it may better to just use the "-serial" option and
let QEMU itself, based on the machine type, set the device
accordingly.

To achieve that, the behavior of set_console() now flags the intention
to add a console device on launch(), and if no explicit device type is
given, and there's no definition on CONSOLE_DEV_TYPES, the "-serial"
is going to be added to the QEMU command line, instead of raising
exceptions.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 scripts/qemu.py | 28 +++++++++++++++-------------
 1 file changed, 15 insertions(+), 13 deletions(-)

diff --git a/scripts/qemu.py b/scripts/qemu.py
index ec3567d4e2..88e1608b42 100644
--- a/scripts/qemu.py
+++ b/scripts/qemu.py
@@ -121,6 +121,7 @@ class QEMUMachine(object):
         self._temp_dir = None
         self._launched = False
         self._machine = None
+        self._console_set = False
         self._console_device_type = None
         self._console_address = None
         self._console_socket = None
@@ -240,13 +241,17 @@ class QEMUMachine(object):
                 '-display', 'none', '-vga', 'none']
         if self._machine is not None:
             args.extend(['-machine', self._machine])
-        if self._console_device_type is not None:
+        if self._console_set:
             self._console_address = os.path.join(self._temp_dir,
                                                  self._name + "-console.sock")
             chardev = ('socket,id=console,path=%s,server,nowait' %
                        self._console_address)
-            device = '%s,chardev=console' % self._console_device_type
-            args.extend(['-chardev', chardev, '-device', device])
+            args.extend(['-chardev', chardev])
+            if self._console_device_type is None:
+                args.extend(['-serial', 'chardev:console'])
+            else:
+                device = '%s,chardev=console' % self._console_device_type
+                args.extend(['-device', device])
         return args
 
     def _pre_launch(self):
@@ -479,23 +484,20 @@ class QEMUMachine(object):
         machine launch time, as it depends on the temporary directory
         to be created.
 
-        @param device_type: the device type, such as "isa-serial"
+        @param device_type: the device type, such as "isa-serial".  If
+                            None is given (the default value) a "-serial
+                            chardev:console" command line argument will
+                            be used instead, resorting to the machine's
+                            default device type.
         @raises: QEMUMachineAddDeviceError if the device type is not given
                  and can not be determined.
         """
-        if device_type is None:
-            if self._machine is None:
-                raise QEMUMachineAddDeviceError("Can not add a console device:"
-                                                " QEMU instance without a "
-                                                "defined machine type")
+        self._console_set = True
+        if device_type is None and self._machine is not None:
             for regex, device in CONSOLE_DEV_TYPES.items():
                 if re.match(regex, self._machine):
                     device_type = device
                     break
-            if device_type is None:
-                raise QEMUMachineAddDeviceError("Can not add a console device:"
-                                                " no matching console device "
-                                                "type definition")
         self._console_device_type = device_type
 
     @property
-- 
2.20.1

[Qemu-devel] [PATCH 12/18] Boot Linux Console Test: add a test for mips + malta

[Cleber Rosa]

From: Philippe Mathieu-Daudé <f4bug@amsat.org>

Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
board and verify the serial is working.  Also, it relies on the serial
device set by the machine itself.

If mips is a target being built, "make check-acceptance" will
automatically include this test by the use of the "arch:mips" tags.

Alternatively, this test can be run using:

    $ avocado run -t arch:mips tests/acceptance
    $ avocado run -t machine:malta tests/acceptance
    $ avocado run -t endian:big tests/acceptance

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 .travis.yml                            |  2 +-
 scripts/qemu.py                        |  1 -
 tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
 3 files changed, 21 insertions(+), 2 deletions(-)

diff --git a/.travis.yml b/.travis.yml
index 844d514afa..49f9016e6a 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -187,7 +187,7 @@ matrix:
 
     # Acceptance (Functional) tests
     - env:
-        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu"
+        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
         - TEST_CMD="make check-acceptance"
       addons:
         apt:
diff --git a/scripts/qemu.py b/scripts/qemu.py
index 88e1608b42..ef84b0f843 100644
--- a/scripts/qemu.py
+++ b/scripts/qemu.py
@@ -35,7 +35,6 @@ def kvm_available(target_arch=None):
 #: Maps machine types to the preferred console device types
 CONSOLE_DEV_TYPES = {
     r'^clipper$': 'isa-serial',
-    r'^malta': 'isa-serial',
     r'^(pc.*|q35.*|isapc)$': 'isa-serial',
     r'^(40p|powernv|prep)$': 'isa-serial',
     r'^pseries.*': 'spapr-vty',
diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index 278bb2be3d..0678ec91d2 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -60,3 +60,23 @@ class BootLinuxConsole(Test):
         self.vm.launch()
         console_pattern = 'Kernel command line: %s' % kernel_command_line
         self.wait_for_console_pattern(console_pattern)
+
+    def test_mips_malta(self):
+        """
+        :avocado: tags=arch:mips
+        :avocado: tags=machine:malta
+        :avocado: tags=endian:big
+        """
+        kernel_url = ('http://people.debian.org/~aurel32/qemu/mips/'
+                      'vmlinux-3.2.0-4-4kc-malta')
+        kernel_hash = '592e384a4edc16dade52a6cd5c785c637bcbc9ad'
+        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
+
+        self.vm.set_machine('malta')
+        self.vm.set_console()
+        kernel_command_line = 'console=ttyS0 printk.time=0'
+        self.vm.add_args('-kernel', kernel_path,
+                         '-append', kernel_command_line)
+        self.vm.launch()
+        console_pattern = 'Kernel command line: %s' % kernel_command_line
+        self.wait_for_console_pattern(console_pattern)
-- 
2.20.1

[Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Cleber Rosa]

Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
board and verify the serial is working.

If mips64el is a target being built, "make check-acceptance" will
automatically include this test by the use of the "arch:mips" tags.

Alternatively, this test can be run using:

    $ avocado run -t arch:mips64el tests/acceptance
    $ avocado run -t machine:malta tests/acceptance

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 .travis.yml                            |  2 +-
 tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
 2 files changed, 40 insertions(+), 1 deletion(-)

diff --git a/.travis.yml b/.travis.yml
index 49f9016e6a..28648f7a61 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -187,7 +187,7 @@ matrix:
 
     # Acceptance (Functional) tests
     - env:
-        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
+        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
         - TEST_CMD="make check-acceptance"
       addons:
         apt:
diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index 0678ec91d2..20b845fce1 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -9,8 +9,11 @@
 # later.  See the COPYING file in the top-level directory.
 
 import logging
+import os
 
 from avocado_qemu import Test
+from avocado.utils import process
+from avocado.utils import archive
 
 
 class BootLinuxConsole(Test):
@@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
         self.vm.launch()
         console_pattern = 'Kernel command line: %s' % kernel_command_line
         self.wait_for_console_pattern(console_pattern)
+
+    def test_mips64el_malta(self):
+        """
+        This test requires the ar tool to extract "data.tar.gz" from
+        the Debian package.
+
+        The kernel can be rebuilt using this Debian kernel source [1] and
+        following the instructions on [2].
+
+        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
+        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
+
+        :avocado: tags=arch:mips64el
+        :avocado: tags=machine:malta
+        """
+        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
+                   'pool/main/l/linux-2.6/'
+                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
+        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
+        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
+
+        cwd = os.getcwd()
+        os.chdir(self.workdir)
+        process.run("ar x %s data.tar.gz" % deb_path)
+        archive.extract("data.tar.gz", self.workdir)
+        os.chdir(cwd)
+        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
+
+        self.vm.set_machine('malta')
+        self.vm.set_console()
+        kernel_command_line = 'console=ttyS0 printk.time=0'
+        self.vm.add_args('-kernel', kernel_path,
+                         '-append', kernel_command_line)
+        self.vm.launch()
+        console_pattern = 'Kernel command line: %s' % kernel_command_line
+        self.wait_for_console_pattern(console_pattern)
-- 
2.20.1

[Qemu-devel] [PATCH 14/18] Boot Linux Console Test: add a test for ppc64 + pseries

[Cleber Rosa]

Just like the previous tests, boots a Linux kernel on a ppc64 target
using the pseries machine.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 .travis.yml                            |  2 +-
 scripts/qemu.py                        |  1 -
 tests/acceptance/boot_linux_console.py | 19 +++++++++++++++++++
 3 files changed, 20 insertions(+), 2 deletions(-)

diff --git a/.travis.yml b/.travis.yml
index 28648f7a61..54100eea5a 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -187,7 +187,7 @@ matrix:
 
     # Acceptance (Functional) tests
     - env:
-        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
+        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
         - TEST_CMD="make check-acceptance"
       addons:
         apt:
diff --git a/scripts/qemu.py b/scripts/qemu.py
index ef84b0f843..1531e28fc1 100644
--- a/scripts/qemu.py
+++ b/scripts/qemu.py
@@ -37,7 +37,6 @@ CONSOLE_DEV_TYPES = {
     r'^clipper$': 'isa-serial',
     r'^(pc.*|q35.*|isapc)$': 'isa-serial',
     r'^(40p|powernv|prep)$': 'isa-serial',
-    r'^pseries.*': 'spapr-vty',
     r'^s390-ccw-virtio.*': 'sclpconsole',
     }
 
diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index 20b845fce1..f3ccd23a7a 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -119,3 +119,22 @@ class BootLinuxConsole(Test):
         self.vm.launch()
         console_pattern = 'Kernel command line: %s' % kernel_command_line
         self.wait_for_console_pattern(console_pattern)
+
+    def test_ppc64_pseries(self):
+        """
+        :avocado: tags=arch:ppc64
+        :avocado: tags=machine:pseries
+        """
+        kernel_url = ('http://mirrors.rit.edu/fedora/fedora-secondary/'
+                      'releases/29/Everything/ppc64le/os/ppc/ppc64/vmlinuz')
+        kernel_hash = '3fe04abfc852b66653b8c3c897a59a689270bc77'
+        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
+
+        self.vm.set_machine('pseries')
+        self.vm.set_console()
+        kernel_command_line = 'console=hvc0'
+        self.vm.add_args('-kernel', kernel_path,
+                         '-append', kernel_command_line)
+        self.vm.launch()
+        console_pattern = 'Kernel command line: %s' % kernel_command_line
+        self.wait_for_console_pattern(console_pattern)
-- 
2.20.1

[Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Cleber Rosa]

Just like the previous tests, boots a Linux kernel on a aarch64 target
using the virt machine.

One special option added is the CPU type, given that the kernel
selected fails to boot on the virt machine's default CPU (cortex-a15).

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 .travis.yml                            |  2 +-
 tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
 2 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/.travis.yml b/.travis.yml
index 54100eea5a..595e8c0b6c 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -187,7 +187,7 @@ matrix:
 
     # Acceptance (Functional) tests
     - env:
-        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
+        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
         - TEST_CMD="make check-acceptance"
       addons:
         apt:
diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index f3ccd23a7a..107700b517 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
         self.vm.launch()
         console_pattern = 'Kernel command line: %s' % kernel_command_line
         self.wait_for_console_pattern(console_pattern)
+
+    def test_aarch64_virt(self):
+        """
+        :avocado: tags=arch:aarch64
+        :avocado: tags=machine:virt
+        """
+        kernel_url = ('https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/'
+                      'releases/29/Server/aarch64/os/images/pxeboot/vmlinuz')
+        kernel_hash = '8c73e469fc6ea06a58dc83a628fc695b693b8493'
+        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
+
+        self.vm.set_machine('virt')
+        self.vm.set_console()
+        kernel_command_line = 'console=ttyAMA0'
+        self.vm.add_args('-cpu', 'cortex-a53',
+                         '-kernel', kernel_path,
+                         '-append', kernel_command_line)
+        self.vm.launch()
+        console_pattern = 'Kernel command line: %s' % kernel_command_line
+        self.wait_for_console_pattern(console_pattern)
-- 
2.20.1

[Qemu-devel] [PATCH 16/18] Boot Linux Console Test: add a test for arm + virt

[Cleber Rosa]

Just like the previous tests, boots a Linux kernel on an arm target
using the virt machine.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 .travis.yml                            |  2 +-
 tests/acceptance/boot_linux_console.py | 19 +++++++++++++++++++
 2 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/.travis.yml b/.travis.yml
index 595e8c0b6c..ea7c9828ac 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -187,7 +187,7 @@ matrix:
 
     # Acceptance (Functional) tests
     - env:
-        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
+        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu"
         - TEST_CMD="make check-acceptance"
       addons:
         apt:
diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index 107700b517..00aee3114e 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -158,3 +158,22 @@ class BootLinuxConsole(Test):
         self.vm.launch()
         console_pattern = 'Kernel command line: %s' % kernel_command_line
         self.wait_for_console_pattern(console_pattern)
+
+    def test_arm_virt(self):
+        """
+        :avocado: tags=arch:arm
+        :avocado: tags=machine:virt
+        """
+        kernel_url = ('https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/'
+                      'releases/29/Server/armhfp/os/images/pxeboot/vmlinuz')
+        kernel_hash = 'e9826d741b4fb04cadba8d4824d1ed3b7fb8b4d4'
+        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
+
+        self.vm.set_machine('virt')
+        self.vm.set_console()
+        kernel_command_line = 'console=ttyAMA0'
+        self.vm.add_args('-kernel', kernel_path,
+                         '-append', kernel_command_line)
+        self.vm.launch()
+        console_pattern = 'Kernel command line: %s' % kernel_command_line
+        self.wait_for_console_pattern(console_pattern)
-- 
2.20.1

[Qemu-devel] [PATCH 17/18] Boot Linux Console Test: add a test for s390x + s390-ccw-virtio

[Cleber Rosa]

Just like the previous tests, boots a Linux kernel on a s390x target
using the s390-ccw-virtio machine.

Because it's not possible to have multiple VT220 consoles,
'-nodefaults' is used, so that the one set with set_console() works
correctly.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 .travis.yml                            |  2 +-
 tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
 2 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/.travis.yml b/.travis.yml
index ea7c9828ac..0d5a4b104b 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -187,7 +187,7 @@ matrix:
 
     # Acceptance (Functional) tests
     - env:
-        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu"
+        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu"
         - TEST_CMD="make check-acceptance"
       addons:
         apt:
diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index 00aee3114e..c4d5477d45 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -177,3 +177,23 @@ class BootLinuxConsole(Test):
         self.vm.launch()
         console_pattern = 'Kernel command line: %s' % kernel_command_line
         self.wait_for_console_pattern(console_pattern)
+
+    def test_s390x_s390_ccw_virtio(self):
+        """
+        :avocado: tags=arch:s390x
+        :avocado: tags=machine:s390_ccw_virtio
+        """
+        kernel_url = ('http://mirrors.rit.edu/fedora/fedora-secondary/releases'
+                      '/29/Server/s390x/os/images/kernel.img')
+        kernel_hash = 'e8e8439103ef8053418ef062644ffd46a7919313'
+        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
+
+        self.vm.set_machine('s390-ccw-virtio')
+        self.vm.set_console()
+        kernel_command_line = 'console=sclp0'
+        self.vm.add_args('-nodefaults',
+                         '-kernel', kernel_path,
+                         '-append', kernel_command_line)
+        self.vm.launch()
+        console_pattern = 'Kernel command line: %s' % kernel_command_line
+        self.wait_for_console_pattern(console_pattern)
-- 
2.20.1

[Qemu-devel] [PATCH 18/18] Boot Linux Console Test: add a test for alpha + clipper

[Cleber Rosa]

Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
board and verify the serial is working.  One extra command added to
the QEMU command line is '-vga std', because the kernel used is
known to crash without it.

If alpha is a target being built, "make check-acceptance" will
automatically include this test by the use of the "arch:alpha" tags.

Alternatively, this test can be run using:

    $ avocado run -t arch:alpha tests/acceptance
    $ avocado run -t machine:clipper tests/acceptance

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Cleber Rosa <crosa@redhat.com>
---
 .travis.yml                            |  2 +-
 scripts/qemu.py                        |  1 -
 tests/acceptance/boot_linux_console.py | 22 ++++++++++++++++++++++
 3 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/.travis.yml b/.travis.yml
index 0d5a4b104b..73a113af87 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -187,7 +187,7 @@ matrix:
 
     # Acceptance (Functional) tests
     - env:
-        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu"
+        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu,alpha-softmmu"
         - TEST_CMD="make check-acceptance"
       addons:
         apt:
diff --git a/scripts/qemu.py b/scripts/qemu.py
index 1531e28fc1..a704da418a 100644
--- a/scripts/qemu.py
+++ b/scripts/qemu.py
@@ -34,7 +34,6 @@ def kvm_available(target_arch=None):
 
 #: Maps machine types to the preferred console device types
 CONSOLE_DEV_TYPES = {
-    r'^clipper$': 'isa-serial',
     r'^(pc.*|q35.*|isapc)$': 'isa-serial',
     r'^(40p|powernv|prep)$': 'isa-serial',
     r'^s390-ccw-virtio.*': 'sclpconsole',
diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
index c4d5477d45..a8028a39d4 100644
--- a/tests/acceptance/boot_linux_console.py
+++ b/tests/acceptance/boot_linux_console.py
@@ -197,3 +197,25 @@ class BootLinuxConsole(Test):
         self.vm.launch()
         console_pattern = 'Kernel command line: %s' % kernel_command_line
         self.wait_for_console_pattern(console_pattern)
+
+    def test_alpha_clipper(self):
+        """
+        :avocado: tags=arch:alpha
+        :avocado: tags=machine:clipper
+        """
+        kernel_url = ('http://archive.debian.org/debian/dists/lenny/main/'
+                      'installer-alpha/current/images/cdrom/vmlinuz')
+        kernel_hash = '3a943149335529e2ed3e74d0d787b85fb5671ba3'
+        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
+
+        uncompressed_kernel = archive.uncompress(kernel_path, self.workdir)
+
+        self.vm.set_machine('clipper')
+        self.vm.set_console()
+        kernel_command_line = 'console=ttyS0 printk.time=0'
+        self.vm.add_args('-vga', 'std',
+                         '-kernel', uncompressed_kernel,
+                         '-append', kernel_command_line)
+        self.vm.launch()
+        console_pattern = 'Kernel command line: %s' % kernel_command_line
+        self.wait_for_console_pattern(console_pattern)
-- 
2.20.1

Re: [Qemu-devel] [PATCH 17/18] Boot Linux Console Test: add a test for s390x + s390-ccw-virtio

[Cornelia Huck]

On Thu, 17 Jan 2019 13:56:27 -0500
Cleber Rosa <crosa@redhat.com> wrote:

> Just like the previous tests, boots a Linux kernel on a s390x target
> using the s390-ccw-virtio machine.
> 
> Because it's not possible to have multiple VT220 consoles,
> '-nodefaults' is used, so that the one set with set_console() works
> correctly.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  .travis.yml                            |  2 +-
>  tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>  2 files changed, 21 insertions(+), 1 deletion(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index ea7c9828ac..0d5a4b104b 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>  
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 00aee3114e..c4d5477d45 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -177,3 +177,23 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_s390x_s390_ccw_virtio(self):
> +        """
> +        :avocado: tags=arch:s390x
> +        :avocado: tags=machine:s390_ccw_virtio
> +        """
> +        kernel_url = ('http://mirrors.rit.edu/fedora/fedora-secondary/releases'
> +                      '/29/Server/s390x/os/images/kernel.img')
> +        kernel_hash = 'e8e8439103ef8053418ef062644ffd46a7919313'
> +        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
> +
> +        self.vm.set_machine('s390-ccw-virtio')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=sclp0'
> +        self.vm.add_args('-nodefaults',
> +                         '-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)

Cool, I had been looking at adding some of the tests I'm running here
to this framework as well, and this gives me a good base to work with :)

I'll try to find some time to give it a whirl.

Re: [Qemu-devel] [PATCH 06/18] Acceptance tests: use "arch:" tag to filter target specific tests

[Cornelia Huck]

On Thu, 17 Jan 2019 13:56:16 -0500
Cleber Rosa <crosa@redhat.com> wrote:

> Currently, the only test that contains some target architecture
> information is "boot_linux_console.py" which test contains a "x86_64"
> tag.  But that tag is not respected in the default execution, that is,
> "make check-acceptance" doesn't do anything with it.
> 
> That said, even the target architecture handling currently present in
> the "avocado_qemu.Test" class, class is pretty limited.  For instance,
> by default, it chooses a target based on the host architecture.
> 
> Because the original implementation of the tags feature in Avocado did
> not include any time of namespace or "key:val" mechanism, no tag has
> relation to another tag.  The new implementation of the tags feature
> from version 67.0 onwards, allows "key:val" tags, and because of that,
> a test can be classified with a tag in a given key.  For instance, the
> new proposed version of the "boot_linux_console.py" test, which
> downloads and attempts to run a x86_64 kernel, is now tagged as:
> 
>   :avocado: tags=arch:x86_64
> 
> This means that it can be filtered (out) when no x86_64 target is
> available.  At the same time, tests that don't have a "arch:" tag,
> will not be filtered out.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  tests/Makefile.include                 | 3 +++
>  tests/acceptance/boot_linux_console.py | 2 +-
>  tests/requirements.txt                 | 2 +-
>  3 files changed, 5 insertions(+), 2 deletions(-)
> 

> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 98324f7591..46b20bdfe2 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -19,7 +19,7 @@ class BootLinuxConsole(Test):
>      and the kernel command line is properly passed from QEMU to the kernel
>  
>      :avocado: enable
> -    :avocado: tags=x86_64
> +    :avocado: tags=arch:x86_64
>      """
>  
>      timeout = 60

You probably want to do the same change in virtio_version.py;
otherwise, if I run the acceptance tests on s390x, it will run into
timeouts (it looks like that test is intended to be run with x86
machines anyway.)

Re: [Qemu-devel] [PATCH 17/18] Boot Linux Console Test: add a test for s390x + s390-ccw-virtio

[Cleber Rosa]

On 1/18/19 3:58 AM, Cornelia Huck wrote:
> 
> Cool, I had been looking at adding some of the tests I'm running here
> to this framework as well, and this gives me a good base to work with :)
> 
> I'll try to find some time to give it a whirl.
> 

I can't assist you in finding some time, but for anything else, please
ping me :)

Cheers,
- Cleber.

Re: [Qemu-devel] [PATCH 05/18] Acceptance tests: introduce arch parameter and attribute

[Caio Carrara]

Hi, Cleber.

On Thu, Jan 17, 2019 at 01:56:15PM -0500, Cleber Rosa wrote:
> It's useful to define the architecture that should be used in
> situations such as:
>  * the intended target of the QEMU binary to be used on tests
>  * the architecture of code to be run within the QEMU binary, such
>    as a kernel image or a full blown guest OS image
> 
> This commit introduces both a test parameter and a test instance
> attribute, that will contain such a value.
> 
> Now, when the "arch" test parameter is given, it will influence the
> selection of the default QEMU binary, if one is not given explicitly
> by means of the "qemu_img" parameter.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  docs/devel/testing.rst                    | 18 ++++++++++++++++++
>  tests/acceptance/avocado_qemu/__init__.py | 13 ++++++++++---
>  2 files changed, 28 insertions(+), 3 deletions(-)
> 
> diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
> index 44c9b3ae74..60a16c5834 100644
> --- a/docs/devel/testing.rst
> +++ b/docs/devel/testing.rst
> @@ -689,6 +689,17 @@ vm
>  A QEMUMachine instance, initially configured according to the given
>  ``qemu_bin`` parameter.
>  
> +arch
> +~~~~
> +
> +The architecture that will be used on a number of different
> +scenarios.  For instance, when a QEMU binary is not explicitly given,
> +the one selected will depend on this attribute.
> +
> +The ``arch`` attribute will be set to the test parameter of the same
> +name, and if one is not given explicitly, it will default to the
> +system architecture (as given by ``uname``).

Would you think it's important to make explicit that it's the *host*
system architecture? Just like you did in the docstrings of
pick_default_qemu_bin().

> +
>  qemu_bin
>  ~~~~~~~~
>  
> @@ -711,6 +722,13 @@ like the following:
>  
>    PARAMS (key=qemu_bin, path=*, default=x86_64-softmmu/qemu-system-x86_64) => 'x86_64-softmmu/qemu-system-x86_64
>  
> +arch
> +~~~~
> +
> +The architecture that will be used on a number of different scenarios.
> +This parameter has a direct relation with the ``arch`` attribute.  If
> +not given, it will default to None.
> +
>  qemu_bin
>  ~~~~~~~~
>  
> diff --git a/tests/acceptance/avocado_qemu/__init__.py b/tests/acceptance/avocado_qemu/__init__.py
> index d8d5b48dac..7a38851753 100644
> --- a/tests/acceptance/avocado_qemu/__init__.py
> +++ b/tests/acceptance/avocado_qemu/__init__.py
> @@ -23,16 +23,22 @@ def is_readable_executable_file(path):
>      return os.path.isfile(path) and os.access(path, os.R_OK | os.X_OK)
>  
>  
> -def pick_default_qemu_bin():
> +def pick_default_qemu_bin(arch=None):
>      """
>      Picks the path of a QEMU binary, starting either in the current working
>      directory or in the source tree root directory.
>  
> +    :param arch: the arch to use when looking for a QEMU binary (the target
> +                 will match the arch given).  If None (the default) arch
> +                 will be the current host system arch (as given by
> +                 :func:`os.uname`).
> +    :type arch: str
>      :returns: the path to the default QEMU binary or None if one could not
>                be found
>      :rtype: str or None
>      """
> -    arch = os.uname()[4]
> +    if arch is None:
> +        arch = os.uname()[4]
>      qemu_bin_relative_path = os.path.join("%s-softmmu" % arch,
>                                            "qemu-system-%s" % arch)
>      if is_readable_executable_file(qemu_bin_relative_path):
> @@ -47,8 +53,9 @@ def pick_default_qemu_bin():
>  class Test(avocado.Test):
>      def setUp(self):
>          self.vm = None
> +        self.arch = self.params.get('arch')
>          self.qemu_bin = self.params.get('qemu_bin',
> -                                        default=pick_default_qemu_bin())
> +                                        default=pick_default_qemu_bin(self.arch))

Since it was used the named argument in function declaration I think
it's more consistent also use the argument name here in the function
call: `pick_default_qemu_bin(arch=self.arch)`

>          if self.qemu_bin is None:
>              self.cancel("No QEMU binary defined or found in the source tree")
>          self.vm = QEMUMachine(self.qemu_bin)
> -- 
> 2.20.1
> 

Thanks,
-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 01/18] scripts/qemu.py: log QEMU launch command line

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:11PM -0500, Cleber Rosa wrote:
> Even when the launch of QEMU succeeds, it's useful to have the command
> line recorded.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  scripts/qemu.py | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/scripts/qemu.py b/scripts/qemu.py
> index 6e3b0e6771..ec3567d4e2 100644
> --- a/scripts/qemu.py
> +++ b/scripts/qemu.py
> @@ -312,6 +312,7 @@ class QEMUMachine(object):
>          self._pre_launch()
>          self._qemu_full_args = (self._wrapper + [self._binary] +
>                                  self._base_args() + self._args)
> +        LOG.debug('VM launch command: %r', ' '.join(self._qemu_full_args))
>          self._popen = subprocess.Popen(self._qemu_full_args,
>                                         stdin=devnull,
>                                         stdout=self._qemu_log_file,
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 02/18] Acceptance tests: show avocado test execution by default

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:12PM -0500, Cleber Rosa wrote:
> The current version of the "check-acceptance" target will only show
> one line for execution of all tests.  That's probably OK if the tests
> to be run are quick enough and they're always the same.
> 
> But, there's already one test alone that takes on average ~5 seconds
> to run, we intend to adapt the list of tests to match the user's build
> environment (among other choices).
> 
> Because of that, let's present the default Avocado UI by default.
> Users can always choose a different output by setting the AVOCADO_SHOW
> variable.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  .travis.yml            | 2 +-
>  tests/Makefile.include | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index 93fd0164a0..844d514afa 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -188,7 +188,7 @@ matrix:
>      # Acceptance (Functional) tests
>      - env:
>          - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu"
> -        - TEST_CMD="make AVOCADO_SHOW=app check-acceptance"
> +        - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
>            packages:
> diff --git a/tests/Makefile.include b/tests/Makefile.include
> index f403a6571d..c73298740d 100644
> --- a/tests/Makefile.include
> +++ b/tests/Makefile.include
> @@ -956,7 +956,7 @@ TESTS_RESULTS_DIR=$(BUILD_DIR)/tests/results
>  # Controls the output generated by Avocado when running tests.
>  # Any number of command separated loggers are accepted.  For more
>  # information please refer to "avocado --help".
> -AVOCADO_SHOW=none
> +AVOCADO_SHOW=app
>  
>  PYTHON3 = $(shell $(PYTHON) -c 'import sys; print(1 if sys.version_info >= (3, 0) else 0)')
>  ifeq ($(PYTHON3), 1)
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 03/18] Acceptance tests: improve docstring on pick_default_qemu_bin()

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:13PM -0500, Cleber Rosa wrote:
> Making it clear what is returned by this utility function.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  tests/acceptance/avocado_qemu/__init__.py | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/tests/acceptance/avocado_qemu/__init__.py b/tests/acceptance/avocado_qemu/__init__.py
> index 1e54fd5932..d8d5b48dac 100644
> --- a/tests/acceptance/avocado_qemu/__init__.py
> +++ b/tests/acceptance/avocado_qemu/__init__.py
> @@ -27,6 +27,10 @@ def pick_default_qemu_bin():
>      """
>      Picks the path of a QEMU binary, starting either in the current working
>      directory or in the source tree root directory.
> +
> +    :returns: the path to the default QEMU binary or None if one could not
> +              be found
> +    :rtype: str or None
>      """
>      arch = os.uname()[4]
>      qemu_bin_relative_path = os.path.join("%s-softmmu" % arch,
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 04/18] Acceptance tests: fix doc reference to avocado_qemu directory

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:14PM -0500, Cleber Rosa wrote:
> The "this directory" reference is misleading and confusing, it's a
> leftover from when this text was proposed in a README file inside
> the "tests/acceptance/avocado_qemu" directory.
> 
> When that text was moved to the top level docs directory, the
> reference was not updated.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  docs/devel/testing.rst | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
> index 18e2c0868a..44c9b3ae74 100644
> --- a/docs/devel/testing.rst
> +++ b/docs/devel/testing.rst
> @@ -590,8 +590,9 @@ Alternatively, follow the instructions on this link:
>  Overview
>  --------
>  
> -This directory provides the ``avocado_qemu`` Python module, containing
> -the ``avocado_qemu.Test`` class.  Here's a simple usage example:
> +The ``tests/acceptance/avocado_qemu`` directory provides the
> +``avocado_qemu`` Python module, containing the ``avocado_qemu.Test``
> +class.  Here's a simple usage example:
>  
>  .. code::
>  
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 06/18] Acceptance tests: use "arch:" tag to filter target specific tests

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:16PM -0500, Cleber Rosa wrote:
> Currently, the only test that contains some target architecture
> information is "boot_linux_console.py" which test contains a "x86_64"
> tag.  But that tag is not respected in the default execution, that is,
> "make check-acceptance" doesn't do anything with it.
> 
> That said, even the target architecture handling currently present in
> the "avocado_qemu.Test" class, class is pretty limited.  For instance,
> by default, it chooses a target based on the host architecture.
> 
> Because the original implementation of the tags feature in Avocado did
> not include any time of namespace or "key:val" mechanism, no tag has
> relation to another tag.  The new implementation of the tags feature
> from version 67.0 onwards, allows "key:val" tags, and because of that,
> a test can be classified with a tag in a given key.  For instance, the
> new proposed version of the "boot_linux_console.py" test, which
> downloads and attempts to run a x86_64 kernel, is now tagged as:
> 
>   :avocado: tags=arch:x86_64
> 
> This means that it can be filtered (out) when no x86_64 target is
> available.  At the same time, tests that don't have a "arch:" tag,
> will not be filtered out.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  tests/Makefile.include                 | 3 +++
>  tests/acceptance/boot_linux_console.py | 2 +-
>  tests/requirements.txt                 | 2 +-
>  3 files changed, 5 insertions(+), 2 deletions(-)
> 
{...}
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 07/18] Acceptance tests: look for target architecture in test tags first

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:17PM -0500, Cleber Rosa wrote:
> A test can, optionally, be tagged for one or many architectures.  If a
> test has been tagged for a single architecture, there's a high chance
> that the test won't run on other architectures.  This changes the
> default order of choosing a default target architecture to use based
> on the 'arch' tag value first.
> 
> The precedence order is for choosing a QEMU binary to use for a test
> is now:
> 
>  * qemu_bin parameter
>  * arch parameter
>  * arch tag value (for example, x86_64 if ":avocado: tags=arch:x86_64
>    is used)
> 
> This means that if one runs:
> 
>  $ avocado run -p qemu_bin=/usr/bin/qemu-system-x86_64 test.py
> 
> No arch parameter or tag will influence the selection of the QEMU
> target binary.  If one runs:
> 
>  $ avocado run -p arch=ppc64 test.py
> 
> The target binary selection mechanism will attempt to find a binary
> such as "ppc64-softmmu/qemu-system-ppc64".  And finally, if one runs
> a test that is tagged (in its docstring) with "arch:aarch64":
> 
>  $ avocado run aarch64.py
> 
> The target binary selection mechanism will attempt to find a binary
> such as "aarch64-softmmu/qemu-system-aarch64".
> 
> At this time, no provision is made to cancel the execution of tests if
> the arch parameter given (manually) does not match the test "arch"
> tag, but it may be a useful default behavior to be added in the
> future.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  tests/acceptance/avocado_qemu/__init__.py | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/tests/acceptance/avocado_qemu/__init__.py b/tests/acceptance/avocado_qemu/__init__.py
> index 7a38851753..ddbdb7b926 100644
> --- a/tests/acceptance/avocado_qemu/__init__.py
> +++ b/tests/acceptance/avocado_qemu/__init__.py
> @@ -53,7 +53,12 @@ def pick_default_qemu_bin(arch=None):
>  class Test(avocado.Test):
>      def setUp(self):
>          self.vm = None
> -        self.arch = self.params.get('arch')
> +        arches = self.tags.get('arch', [])
> +        if len(arches) == 1:
> +            arch = arches.pop()
> +        else:
> +            arch = None
> +        self.arch = self.params.get('arch', default=arch)
>          self.qemu_bin = self.params.get('qemu_bin',
>                                          default=pick_default_qemu_bin(self.arch))
>          if self.qemu_bin is None:
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 08/18] Boot Linux Console Test: rename the x86_64 after the arch and machine

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:18PM -0500, Cleber Rosa wrote:
> Given that the test is specific to x86_64 and pc, and new tests are
> going to be added to the same class, let's rename it accordingly.
> Also, let's make the class documentation not architecture specific.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  tests/acceptance/boot_linux_console.py | 11 +++++++----
>  1 file changed, 7 insertions(+), 4 deletions(-)
> 
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 46b20bdfe2..89df7f6e4f 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -15,16 +15,19 @@ from avocado_qemu import Test
>  
>  class BootLinuxConsole(Test):
>      """
> -    Boots a x86_64 Linux kernel and checks that the console is operational
> -    and the kernel command line is properly passed from QEMU to the kernel
> +    Boots a Linux kernel and checks that the console is operational and the
> +    kernel command line is properly passed from QEMU to the kernel
>  
>      :avocado: enable
> -    :avocado: tags=arch:x86_64
>      """
>  
>      timeout = 60
>  
> -    def test(self):
> +    def test_x86_64_pc(self):
> +        """
> +        :avocado: tags=arch:x86_64
> +        :avocado: tags=machine:pc
> +        """
>          kernel_url = ('https://mirrors.kernel.org/fedora/releases/28/'
>                        'Everything/x86_64/os/images/pxeboot/vmlinuz')
>          kernel_hash = '238e083e114c48200f80d889f7e32eeb2793e02a'
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 09/18] Boot Linux Console Test: update the x86_64 kernel

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:19PM -0500, Cleber Rosa wrote:
> To the stock Fedora 29 kernel, from the Fedora 28.  New tests will be
> added using the 29 kernel, so for consistency, let's also update it
> here.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  tests/acceptance/boot_linux_console.py | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 89df7f6e4f..35b31162d4 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -28,9 +28,9 @@ class BootLinuxConsole(Test):
>          :avocado: tags=arch:x86_64
>          :avocado: tags=machine:pc
>          """
> -        kernel_url = ('https://mirrors.kernel.org/fedora/releases/28/'
> +        kernel_url = ('https://mirrors.kernel.org/fedora/releases/29/'
>                        'Everything/x86_64/os/images/pxeboot/vmlinuz')
> -        kernel_hash = '238e083e114c48200f80d889f7e32eeb2793e02a'
> +        kernel_hash = '23bebd2680757891cf7adedb033532163a792495'
>          kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
>  
>          self.vm.set_machine('pc')
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 10/18] Boot Linux Console Test: refactor the console watcher into utility method

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:20PM -0500, Cleber Rosa wrote:
> This introduces a utility method that monitors the console device and
> looks for either a message that signals the test success or failure.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  tests/acceptance/boot_linux_console.py | 30 ++++++++++++++++++--------
>  1 file changed, 21 insertions(+), 9 deletions(-)
> 
{...}
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 11/18] scripts/qemu.py: support adding a console with the default serial device

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:21PM -0500, Cleber Rosa wrote:
> The set_console() utility function traditionally adds a device either
> based on the explicitly given device type, or based on the machine type,
> a known good type of device.
> 
> But, for a number of machine types, it may be impossible or
> inconvenient to add the devices my means of "-device" command line
> options, and then it may better to just use the "-serial" option and
> let QEMU itself, based on the machine type, set the device
> accordingly.
> 
> To achieve that, the behavior of set_console() now flags the intention
> to add a console device on launch(), and if no explicit device type is
> given, and there's no definition on CONSOLE_DEV_TYPES, the "-serial"
> is going to be added to the QEMU command line, instead of raising
> exceptions.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  scripts/qemu.py | 28 +++++++++++++++-------------
>  1 file changed, 15 insertions(+), 13 deletions(-)
> 
{...}
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 12/18] Boot Linux Console Test: add a test for mips + malta

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:22PM -0500, Cleber Rosa wrote:
> From: Philippe Mathieu-Daudé <f4bug@amsat.org>
> 
> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
> board and verify the serial is working.  Also, it relies on the serial
> device set by the machine itself.
> 
> If mips is a target being built, "make check-acceptance" will
> automatically include this test by the use of the "arch:mips" tags.
> 
> Alternatively, this test can be run using:
> 
>     $ avocado run -t arch:mips tests/acceptance
>     $ avocado run -t machine:malta tests/acceptance
>     $ avocado run -t endian:big tests/acceptance
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  .travis.yml                            |  2 +-
>  scripts/qemu.py                        |  1 -
>  tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>  3 files changed, 21 insertions(+), 2 deletions(-)
> 
{...}
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:23PM -0500, Cleber Rosa wrote:
> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
> board and verify the serial is working.
> 
> If mips64el is a target being built, "make check-acceptance" will
> automatically include this test by the use of the "arch:mips" tags.
> 
> Alternatively, this test can be run using:
> 
>     $ avocado run -t arch:mips64el tests/acceptance
>     $ avocado run -t machine:malta tests/acceptance
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  .travis.yml                            |  2 +-
>  tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>  2 files changed, 40 insertions(+), 1 deletion(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index 49f9016e6a..28648f7a61 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>  
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 0678ec91d2..20b845fce1 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -9,8 +9,11 @@
>  # later.  See the COPYING file in the top-level directory.
>  
>  import logging
> +import os
>  
>  from avocado_qemu import Test
> +from avocado.utils import process
> +from avocado.utils import archive
>  
>  
>  class BootLinuxConsole(Test):
> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_mips64el_malta(self):
> +        """
> +        This test requires the ar tool to extract "data.tar.gz" from
> +        the Debian package.
> +
> +        The kernel can be rebuilt using this Debian kernel source [1] and
> +        following the instructions on [2].
> +
> +        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
> +        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
> +
> +        :avocado: tags=arch:mips64el
> +        :avocado: tags=machine:malta
> +        """
> +        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
> +                   'pool/main/l/linux-2.6/'
> +                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
> +
> +        cwd = os.getcwd()
> +        os.chdir(self.workdir)
> +        process.run("ar x %s data.tar.gz" % deb_path)
> +        archive.extract("data.tar.gz", self.workdir)
> +        os.chdir(cwd)
> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
> +
> +        self.vm.set_machine('malta')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=ttyS0 printk.time=0'
> +        self.vm.add_args('-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 14/18] Boot Linux Console Test: add a test for ppc64 + pseries

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:24PM -0500, Cleber Rosa wrote:
> Just like the previous tests, boots a Linux kernel on a ppc64 target
> using the pseries machine.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  .travis.yml                            |  2 +-
>  scripts/qemu.py                        |  1 -
>  tests/acceptance/boot_linux_console.py | 19 +++++++++++++++++++
>  3 files changed, 20 insertions(+), 2 deletions(-)
> 
{...}
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:25PM -0500, Cleber Rosa wrote:
> Just like the previous tests, boots a Linux kernel on a aarch64 target
> using the virt machine.
> 
> One special option added is the CPU type, given that the kernel
> selected fails to boot on the virt machine's default CPU (cortex-a15).
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  .travis.yml                            |  2 +-
>  tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>  2 files changed, 21 insertions(+), 1 deletion(-)
> 
{...}
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 16/18] Boot Linux Console Test: add a test for arm + virt

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:26PM -0500, Cleber Rosa wrote:
> Just like the previous tests, boots a Linux kernel on an arm target
> using the virt machine.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  .travis.yml                            |  2 +-
>  tests/acceptance/boot_linux_console.py | 19 +++++++++++++++++++
>  2 files changed, 20 insertions(+), 1 deletion(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index 595e8c0b6c..ea7c9828ac 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>  
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 107700b517..00aee3114e 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -158,3 +158,22 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_arm_virt(self):
> +        """
> +        :avocado: tags=arch:arm
> +        :avocado: tags=machine:virt
> +        """
> +        kernel_url = ('https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/'
> +                      'releases/29/Server/armhfp/os/images/pxeboot/vmlinuz')
> +        kernel_hash = 'e9826d741b4fb04cadba8d4824d1ed3b7fb8b4d4'
> +        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
> +
> +        self.vm.set_machine('virt')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=ttyAMA0'
> +        self.vm.add_args('-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 17/18] Boot Linux Console Test: add a test for s390x + s390-ccw-virtio

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:27PM -0500, Cleber Rosa wrote:
> Just like the previous tests, boots a Linux kernel on a s390x target
> using the s390-ccw-virtio machine.
> 
> Because it's not possible to have multiple VT220 consoles,
> '-nodefaults' is used, so that the one set with set_console() works
> correctly.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  .travis.yml                            |  2 +-
>  tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>  2 files changed, 21 insertions(+), 1 deletion(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index ea7c9828ac..0d5a4b104b 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>  
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 00aee3114e..c4d5477d45 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -177,3 +177,23 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_s390x_s390_ccw_virtio(self):
> +        """
> +        :avocado: tags=arch:s390x
> +        :avocado: tags=machine:s390_ccw_virtio
> +        """
> +        kernel_url = ('http://mirrors.rit.edu/fedora/fedora-secondary/releases'
> +                      '/29/Server/s390x/os/images/kernel.img')
> +        kernel_hash = 'e8e8439103ef8053418ef062644ffd46a7919313'
> +        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
> +
> +        self.vm.set_machine('s390-ccw-virtio')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=sclp0'
> +        self.vm.add_args('-nodefaults',
> +                         '-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 18/18] Boot Linux Console Test: add a test for alpha + clipper

[Caio Carrara]

On Thu, Jan 17, 2019 at 01:56:28PM -0500, Cleber Rosa wrote:
> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
> board and verify the serial is working.  One extra command added to
> the QEMU command line is '-vga std', because the kernel used is
> known to crash without it.
> 
> If alpha is a target being built, "make check-acceptance" will
> automatically include this test by the use of the "arch:alpha" tags.
> 
> Alternatively, this test can be run using:
> 
>     $ avocado run -t arch:alpha tests/acceptance
>     $ avocado run -t machine:clipper tests/acceptance
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Caio Carrara <ccarrara@redhat.com>

> ---
>  .travis.yml                            |  2 +-
>  scripts/qemu.py                        |  1 -
>  tests/acceptance/boot_linux_console.py | 22 ++++++++++++++++++++++
>  3 files changed, 23 insertions(+), 2 deletions(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index 0d5a4b104b..73a113af87 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>  
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu,alpha-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/scripts/qemu.py b/scripts/qemu.py
> index 1531e28fc1..a704da418a 100644
> --- a/scripts/qemu.py
> +++ b/scripts/qemu.py
> @@ -34,7 +34,6 @@ def kvm_available(target_arch=None):
>  
>  #: Maps machine types to the preferred console device types
>  CONSOLE_DEV_TYPES = {
> -    r'^clipper$': 'isa-serial',
>      r'^(pc.*|q35.*|isapc)$': 'isa-serial',
>      r'^(40p|powernv|prep)$': 'isa-serial',
>      r'^s390-ccw-virtio.*': 'sclpconsole',
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index c4d5477d45..a8028a39d4 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -197,3 +197,25 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_alpha_clipper(self):
> +        """
> +        :avocado: tags=arch:alpha
> +        :avocado: tags=machine:clipper
> +        """
> +        kernel_url = ('http://archive.debian.org/debian/dists/lenny/main/'
> +                      'installer-alpha/current/images/cdrom/vmlinuz')
> +        kernel_hash = '3a943149335529e2ed3e74d0d787b85fb5671ba3'
> +        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
> +
> +        uncompressed_kernel = archive.uncompress(kernel_path, self.workdir)
> +
> +        self.vm.set_machine('clipper')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=ttyS0 printk.time=0'
> +        self.vm.add_args('-vga', 'std',
> +                         '-kernel', uncompressed_kernel,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)
> -- 
> 2.20.1
> 

-- 
Caio Carrara
Software Engineer, Virt Team - Red Hat
ccarrara@redhat.com

Re: [Qemu-devel] [PATCH 00/18] Acceptance Tests: target architecture support

[Aleksandar Markovic]

> From: Cleber Rosa <crosa@redhat.com>

>   class My(Test):
>       def test_nx_cpu_flag(self):
>           """
>           :avocado: tags=arch:x86_64
>           """
>           test_code()

> The value of the "arch" key, in this case, "x86_64" will be used when
> selecting the QEMU binary to use in the test.  At the same time, if
> "x86_64-softmmu" is not a built target, the test will be filtered out
> by "make check-acceptance"[3].

I think, the term "arch" is a little problematic in QEMU parlance. IMHO,
"target" should be used instead. ("arch" is used in Linux kernel community)

The overall structure of the "tags" should be a little different. My
suggestion:

"target"
"isa" (instruction set architecture, determeines how the kernel and rootfs are built)
"cpu"
"machine"

This would allow clear view what a particular acceptance test tests, and will
enforce consistency and clarity in the test organization across the board.

That said, I am very excited about this series.

Regards,
Aleksandar

Re: [Qemu-devel] [PATCH 01/18] scripts/qemu.py: log QEMU launch command line

[Philippe Mathieu-Daudé]

On 1/17/19 7:56 PM, Cleber Rosa wrote:
> Even when the launch of QEMU succeeds, it's useful to have the command
> line recorded.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  scripts/qemu.py | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/scripts/qemu.py b/scripts/qemu.py
> index 6e3b0e6771..ec3567d4e2 100644
> --- a/scripts/qemu.py
> +++ b/scripts/qemu.py
> @@ -312,6 +312,7 @@ class QEMUMachine(object):
>          self._pre_launch()
>          self._qemu_full_args = (self._wrapper + [self._binary] +
>                                  self._base_args() + self._args)
> +        LOG.debug('VM launch command: %r', ' '.join(self._qemu_full_args))

Useful!

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

>          self._popen = subprocess.Popen(self._qemu_full_args,
>                                         stdin=devnull,
>                                         stdout=self._qemu_log_file,
> 

Re: [Qemu-devel] [PATCH 02/18] Acceptance tests: show avocado test execution by default

[Philippe Mathieu-Daudé]

On 1/17/19 7:56 PM, Cleber Rosa wrote:
> The current version of the "check-acceptance" target will only show
> one line for execution of all tests.  That's probably OK if the tests
> to be run are quick enough and they're always the same.
> 
> But, there's already one test alone that takes on average ~5 seconds
> to run, we intend to adapt the list of tests to match the user's build
> environment (among other choices).
> 
> Because of that, let's present the default Avocado UI by default.
> Users can always choose a different output by setting the AVOCADO_SHOW
> variable.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> ---
>  .travis.yml            | 2 +-
>  tests/Makefile.include | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index 93fd0164a0..844d514afa 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -188,7 +188,7 @@ matrix:
>      # Acceptance (Functional) tests
>      - env:
>          - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu"
> -        - TEST_CMD="make AVOCADO_SHOW=app check-acceptance"
> +        - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
>            packages:
> diff --git a/tests/Makefile.include b/tests/Makefile.include
> index f403a6571d..c73298740d 100644
> --- a/tests/Makefile.include
> +++ b/tests/Makefile.include
> @@ -956,7 +956,7 @@ TESTS_RESULTS_DIR=$(BUILD_DIR)/tests/results
>  # Controls the output generated by Avocado when running tests.
>  # Any number of command separated loggers are accepted.  For more
>  # information please refer to "avocado --help".
> -AVOCADO_SHOW=none
> +AVOCADO_SHOW=app
>  
>  PYTHON3 = $(shell $(PYTHON) -c 'import sys; print(1 if sys.version_info >= (3, 0) else 0)')
>  ifeq ($(PYTHON3), 1)
> 

Re: [Qemu-devel] [PATCH 04/18] Acceptance tests: fix doc reference to avocado_qemu directory

[Philippe Mathieu-Daudé]

On 1/17/19 7:56 PM, Cleber Rosa wrote:
> The "this directory" reference is misleading and confusing, it's a
> leftover from when this text was proposed in a README file inside
> the "tests/acceptance/avocado_qemu" directory.
> 
> When that text was moved to the top level docs directory, the
> reference was not updated.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> ---
>  docs/devel/testing.rst | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
> index 18e2c0868a..44c9b3ae74 100644
> --- a/docs/devel/testing.rst
> +++ b/docs/devel/testing.rst
> @@ -590,8 +590,9 @@ Alternatively, follow the instructions on this link:
>  Overview
>  --------
>  
> -This directory provides the ``avocado_qemu`` Python module, containing
> -the ``avocado_qemu.Test`` class.  Here's a simple usage example:
> +The ``tests/acceptance/avocado_qemu`` directory provides the
> +``avocado_qemu`` Python module, containing the ``avocado_qemu.Test``
> +class.  Here's a simple usage example:
>  
>  .. code::
>  
> 

Re: [Qemu-devel] [PATCH 05/18] Acceptance tests: introduce arch parameter and attribute

[Philippe Mathieu-Daudé]

On 1/18/19 3:28 PM, Caio Carrara wrote:
> Hi, Cleber.
> 
> On Thu, Jan 17, 2019 at 01:56:15PM -0500, Cleber Rosa wrote:
>> It's useful to define the architecture that should be used in
>> situations such as:
>>  * the intended target of the QEMU binary to be used on tests
>>  * the architecture of code to be run within the QEMU binary, such
>>    as a kernel image or a full blown guest OS image
>>
>> This commit introduces both a test parameter and a test instance
>> attribute, that will contain such a value.
>>
>> Now, when the "arch" test parameter is given, it will influence the
>> selection of the default QEMU binary, if one is not given explicitly
>> by means of the "qemu_img" parameter.
>>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>  docs/devel/testing.rst                    | 18 ++++++++++++++++++
>>  tests/acceptance/avocado_qemu/__init__.py | 13 ++++++++++---
>>  2 files changed, 28 insertions(+), 3 deletions(-)
>>
>> diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
>> index 44c9b3ae74..60a16c5834 100644
>> --- a/docs/devel/testing.rst
>> +++ b/docs/devel/testing.rst
>> @@ -689,6 +689,17 @@ vm
>>  A QEMUMachine instance, initially configured according to the given
>>  ``qemu_bin`` parameter.
>>  
>> +arch
>> +~~~~
>> +
>> +The architecture that will be used on a number of different
>> +scenarios.  For instance, when a QEMU binary is not explicitly given,
>> +the one selected will depend on this attribute.
>> +
>> +The ``arch`` attribute will be set to the test parameter of the same
>> +name, and if one is not given explicitly, it will default to the
>> +system architecture (as given by ``uname``).
> 
> Would you think it's important to make explicit that it's the *host*
> system architecture? Just like you did in the docstrings of
> pick_default_qemu_bin().

Agreed.

> 
>> +
>>  qemu_bin
>>  ~~~~~~~~
>>  
>> @@ -711,6 +722,13 @@ like the following:
>>  
>>    PARAMS (key=qemu_bin, path=*, default=x86_64-softmmu/qemu-system-x86_64) => 'x86_64-softmmu/qemu-system-x86_64
>>  
>> +arch
>> +~~~~
>> +
>> +The architecture that will be used on a number of different scenarios.
>> +This parameter has a direct relation with the ``arch`` attribute.  If
>> +not given, it will default to None.
>> +
>>  qemu_bin
>>  ~~~~~~~~
>>  
>> diff --git a/tests/acceptance/avocado_qemu/__init__.py b/tests/acceptance/avocado_qemu/__init__.py
>> index d8d5b48dac..7a38851753 100644
>> --- a/tests/acceptance/avocado_qemu/__init__.py
>> +++ b/tests/acceptance/avocado_qemu/__init__.py
>> @@ -23,16 +23,22 @@ def is_readable_executable_file(path):
>>      return os.path.isfile(path) and os.access(path, os.R_OK | os.X_OK)
>>  
>>  
>> -def pick_default_qemu_bin():
>> +def pick_default_qemu_bin(arch=None):
>>      """
>>      Picks the path of a QEMU binary, starting either in the current working
>>      directory or in the source tree root directory.
>>  
>> +    :param arch: the arch to use when looking for a QEMU binary (the target
>> +                 will match the arch given).  If None (the default) arch
>> +                 will be the current host system arch (as given by
>> +                 :func:`os.uname`).

\o/ thanks :)

>> +    :type arch: str
>>      :returns: the path to the default QEMU binary or None if one could not
>>                be found
>>      :rtype: str or None
>>      """
>> -    arch = os.uname()[4]
>> +    if arch is None:
>> +        arch = os.uname()[4]
>>      qemu_bin_relative_path = os.path.join("%s-softmmu" % arch,
>>                                            "qemu-system-%s" % arch)
>>      if is_readable_executable_file(qemu_bin_relative_path):
>> @@ -47,8 +53,9 @@ def pick_default_qemu_bin():
>>  class Test(avocado.Test):
>>      def setUp(self):
>>          self.vm = None
>> +        self.arch = self.params.get('arch')
>>          self.qemu_bin = self.params.get('qemu_bin',
>> -                                        default=pick_default_qemu_bin())
>> +                                        default=pick_default_qemu_bin(self.arch))
> 
> Since it was used the named argument in function declaration I think
> it's more consistent also use the argument name here in the function
> call: `pick_default_qemu_bin(arch=self.arch)`

With Caio's comments addressed:
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> 
>>          if self.qemu_bin is None:
>>              self.cancel("No QEMU binary defined or found in the source tree")
>>          self.vm = QEMUMachine(self.qemu_bin)
>> -- 
>> 2.20.1
>>
> 
> Thanks,
> 

Re: [Qemu-devel] [PATCH 08/18] Boot Linux Console Test: rename the x86_64 after the arch and machine

[Philippe Mathieu-Daudé]

On 1/17/19 7:56 PM, Cleber Rosa wrote:
> Given that the test is specific to x86_64 and pc, and new tests are
> going to be added to the same class, let's rename it accordingly.
> Also, let's make the class documentation not architecture specific.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  tests/acceptance/boot_linux_console.py | 11 +++++++----
>  1 file changed, 7 insertions(+), 4 deletions(-)
> 
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 46b20bdfe2..89df7f6e4f 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -15,16 +15,19 @@ from avocado_qemu import Test
>  
>  class BootLinuxConsole(Test):
>      """
> -    Boots a x86_64 Linux kernel and checks that the console is operational
> -    and the kernel command line is properly passed from QEMU to the kernel
> +    Boots a Linux kernel and checks that the console is operational and the
> +    kernel command line is properly passed from QEMU to the kernel
>  
>      :avocado: enable
> -    :avocado: tags=arch:x86_64
>      """
>  
>      timeout = 60
>  
> -    def test(self):
> +    def test_x86_64_pc(self):
> +        """
> +        :avocado: tags=arch:x86_64
> +        :avocado: tags=machine:pc
> +        """
>          kernel_url = ('https://mirrors.kernel.org/fedora/releases/28/'
>                        'Everything/x86_64/os/images/pxeboot/vmlinuz')
>          kernel_hash = '238e083e114c48200f80d889f7e32eeb2793e02a'
> 

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Re: [Qemu-devel] [PATCH 10/18] Boot Linux Console Test: refactor the console watcher into utility method

[Philippe Mathieu-Daudé]

Hi Cleber,

On 1/17/19 7:56 PM, Cleber Rosa wrote:
> This introduces a utility method that monitors the console device and
> looks for either a message that signals the test success or failure.
> 
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  tests/acceptance/boot_linux_console.py | 30 ++++++++++++++++++--------
>  1 file changed, 21 insertions(+), 9 deletions(-)
> 
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 35b31162d4..278bb2be3d 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -23,6 +23,25 @@ class BootLinuxConsole(Test):
>  
>      timeout = 60
>  
> +    def wait_for_console_pattern(self, success_message,
> +                                 failure_message='Kernel panic - not syncing'):
> +        """
> +        Waits for messages to appear on the console, while logging the content
> +
> +        :param success_message: if this message appears, test succeeds
> +        :param failure_message: if this message appears, test fails
> +        """
> +        console = self.vm.console_socket.makefile()
> +        console_logger = logging.getLogger('console')
> +        while True:
> +            msg = console.readline()
> +            console_logger.debug(msg.strip())
> +            if success_message in msg:
> +                break
> +            if failure_message in msg:
> +                fail = 'Failure message found in console: %s' % failure_message
> +                self.fail(fail)
> +

This helper is more generic than the BootLinuxConsole class, can you
move it out? I'd like to use it in test_uefi_ovmf_x86_64_pc and
test_uefi_armvirtqemu_aarch64_virt.

Anyway this can be a follow-up patch, so regardless:
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

>      def test_x86_64_pc(self):
>          """
>          :avocado: tags=arch:x86_64
> @@ -39,12 +58,5 @@ class BootLinuxConsole(Test):
>          self.vm.add_args('-kernel', kernel_path,
>                           '-append', kernel_command_line)
>          self.vm.launch()
> -        console = self.vm.console_socket.makefile()
> -        console_logger = logging.getLogger('console')
> -        while True:
> -            msg = console.readline()
> -            console_logger.debug(msg.strip())
> -            if 'Kernel command line: %s' % kernel_command_line in msg:
> -                break
> -            if 'Kernel panic - not syncing' in msg:
> -                self.fail("Kernel panic reached")
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)
> 

Re: [Qemu-devel] [PATCH 12/18] Boot Linux Console Test: add a test for mips + malta

[Philippe Mathieu-Daudé]

Hi Cleber,

On 1/17/19 7:56 PM, Cleber Rosa wrote:
> From: Philippe Mathieu-Daudé <f4bug@amsat.org>
> 
> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
> board and verify the serial is working.  Also, it relies on the serial
> device set by the machine itself.
> 
> If mips is a target being built, "make check-acceptance" will
> automatically include this test by the use of the "arch:mips" tags.
> 
> Alternatively, this test can be run using:
> 
>     $ avocado run -t arch:mips tests/acceptance
>     $ avocado run -t machine:malta tests/acceptance
>     $ avocado run -t endian:big tests/acceptance
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  .travis.yml                            |  2 +-
>  scripts/qemu.py                        |  1 -
>  tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>  3 files changed, 21 insertions(+), 2 deletions(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index 844d514afa..49f9016e6a 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>  
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/scripts/qemu.py b/scripts/qemu.py
> index 88e1608b42..ef84b0f843 100644
> --- a/scripts/qemu.py
> +++ b/scripts/qemu.py
> @@ -35,7 +35,6 @@ def kvm_available(target_arch=None):
>  #: Maps machine types to the preferred console device types
>  CONSOLE_DEV_TYPES = {
>      r'^clipper$': 'isa-serial',
> -    r'^malta': 'isa-serial',
>      r'^(pc.*|q35.*|isapc)$': 'isa-serial',
>      r'^(40p|powernv|prep)$': 'isa-serial',
>      r'^pseries.*': 'spapr-vty',
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 278bb2be3d..0678ec91d2 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -60,3 +60,23 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_mips_malta(self):
> +        """
> +        :avocado: tags=arch:mips
> +        :avocado: tags=machine:malta
> +        :avocado: tags=endian:big
> +        """
> +        kernel_url = ('http://people.debian.org/~aurel32/qemu/mips/'
> +                      'vmlinux-3.2.0-4-4kc-malta')
> +        kernel_hash = '592e384a4edc16dade52a6cd5c785c637bcbc9ad'
> +        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)

Alex was concerned about using this "not-so-easily-reproducible" binary:
https://lists.gnu.org/archive/html/qemu-devel/2018-06/msg08441.html

So I sent another patch using a reproducible one, which you reviewed:
https://lists.gnu.org/archive/html/qemu-devel/2018-10/msg02747.html

Do you mind base this series on it?

Thanks,

Phil.

> +
> +        self.vm.set_machine('malta')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=ttyS0 printk.time=0'
> +        self.vm.add_args('-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)
> 

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Philippe Mathieu-Daudé]

Hi Cleber,

On 1/17/19 7:56 PM, Cleber Rosa wrote:
> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
> board and verify the serial is working.
> 
> If mips64el is a target being built, "make check-acceptance" will
> automatically include this test by the use of the "arch:mips" tags.
> 
> Alternatively, this test can be run using:
> 
>     $ avocado run -t arch:mips64el tests/acceptance
>     $ avocado run -t machine:malta tests/acceptance
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  .travis.yml                            |  2 +-
>  tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>  2 files changed, 40 insertions(+), 1 deletion(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index 49f9016e6a..28648f7a61 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>  
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 0678ec91d2..20b845fce1 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -9,8 +9,11 @@
>  # later.  See the COPYING file in the top-level directory.
>  
>  import logging
> +import os
>  
>  from avocado_qemu import Test
> +from avocado.utils import process
> +from avocado.utils import archive
>  
>  
>  class BootLinuxConsole(Test):
> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_mips64el_malta(self):
> +        """
> +        This test requires the ar tool to extract "data.tar.gz" from
> +        the Debian package.
> +
> +        The kernel can be rebuilt using this Debian kernel source [1] and
> +        following the instructions on [2].
> +
> +        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
> +        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
> +
> +        :avocado: tags=arch:mips64el
> +        :avocado: tags=machine:malta
> +        """
> +        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
> +                   'pool/main/l/linux-2.6/'
> +                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
> +

Can you refactor this code ...

> +        cwd = os.getcwd()
> +        os.chdir(self.workdir)
> +        process.run("ar x %s data.tar.gz" % deb_path)
> +        archive.extract("data.tar.gz", self.workdir)
> +        os.chdir(cwd)
> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'

... as an helper function?

Can be follow-up patch, so:
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> +
> +        self.vm.set_machine('malta')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=ttyS0 printk.time=0'
> +        self.vm.add_args('-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)
> 

Re: [Qemu-devel] [PATCH 00/18] Acceptance Tests: target architecture support

[Philippe Mathieu-Daudé]

Hi Aleksandar,

On 1/21/19 11:15 PM, Aleksandar Markovic wrote:
>> From: Cleber Rosa <crosa@redhat.com>
> 
>>   class My(Test):
>>       def test_nx_cpu_flag(self):
>>           """
>>           :avocado: tags=arch:x86_64
>>           """
>>           test_code()
> 
>> The value of the "arch" key, in this case, "x86_64" will be used when
>> selecting the QEMU binary to use in the test.  At the same time, if
>> "x86_64-softmmu" is not a built target, the test will be filtered out
>> by "make check-acceptance"[3].
> 
> I think, the term "arch" is a little problematic in QEMU parlance. IMHO,
> "target" should be used instead. ("arch" is used in Linux kernel community)

Using target_arch/host_arch might be more explicit, but host_arch is not
very relevant regarding Avocado (except to choose the correct TCG
binary), so usually arch implies target_arch.

I mean, it is unlikely you enforce --host_arch on the command line to
run tests, this using --arch instead of --target_arch looks OK to me,
since host_arch can be guessed.

> 
> The overall structure of the "tags" should be a little different. My
> suggestion:
> 
> "target"
> "isa" (instruction set architecture, determeines how the kernel and rootfs are built)

In QEMU, "isa" is implicit with CPU, isnt' it?

I.e. I use:

$ mips64el-softmmu/qemu-system-mips64el -M Malta -cpu mips64dspr2

> "cpu"
> "machine"
> 
> This would allow clear view what a particular acceptance test tests, and will
> enforce consistency and clarity in the test organization across the board.

Maybe the problem you are pointing out is not Avocado test organization
but QEMU CPU organization... (which also bother me with boards able to
use different SoC, with different peripherals or cpus).

What is your idea on passing arch/cpu/isa to start QEMU?

> 
> That said, I am very excited about this series.

Me too :)

Regards,

Phil.

Re: [Qemu-devel] [PATCH 18/18] Boot Linux Console Test: add a test for alpha + clipper

[Philippe Mathieu-Daudé]

On 1/17/19 7:56 PM, Cleber Rosa wrote:
> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
> board and verify the serial is working.  One extra command added to
> the QEMU command line is '-vga std', because the kernel used is
> known to crash without it.
> 
> If alpha is a target being built, "make check-acceptance" will
> automatically include this test by the use of the "arch:alpha" tags.
> 
> Alternatively, this test can be run using:
> 
>     $ avocado run -t arch:alpha tests/acceptance
>     $ avocado run -t machine:clipper tests/acceptance
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  .travis.yml                            |  2 +-
>  scripts/qemu.py                        |  1 -
>  tests/acceptance/boot_linux_console.py | 22 ++++++++++++++++++++++
>  3 files changed, 23 insertions(+), 2 deletions(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index 0d5a4b104b..73a113af87 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>  
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu,alpha-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/scripts/qemu.py b/scripts/qemu.py
> index 1531e28fc1..a704da418a 100644
> --- a/scripts/qemu.py
> +++ b/scripts/qemu.py
> @@ -34,7 +34,6 @@ def kvm_available(target_arch=None):
>  
>  #: Maps machine types to the preferred console device types
>  CONSOLE_DEV_TYPES = {
> -    r'^clipper$': 'isa-serial',

Why this change?

>      r'^(pc.*|q35.*|isapc)$': 'isa-serial',
>      r'^(40p|powernv|prep)$': 'isa-serial',
>      r'^s390-ccw-virtio.*': 'sclpconsole',
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index c4d5477d45..a8028a39d4 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -197,3 +197,25 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_alpha_clipper(self):
> +        """
> +        :avocado: tags=arch:alpha
> +        :avocado: tags=machine:clipper
> +        """
> +        kernel_url = ('http://archive.debian.org/debian/dists/lenny/main/'
> +                      'installer-alpha/current/images/cdrom/vmlinuz')
> +        kernel_hash = '3a943149335529e2ed3e74d0d787b85fb5671ba3'
> +        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
> +
> +        uncompressed_kernel = archive.uncompress(kernel_path, self.workdir)
> +
> +        self.vm.set_machine('clipper')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=ttyS0 printk.time=0'
> +        self.vm.add_args('-vga', 'std',
> +                         '-kernel', uncompressed_kernel,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)
> 

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Philippe Mathieu-Daudé]

On 1/17/19 7:56 PM, Cleber Rosa wrote:
> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
> board and verify the serial is working.
> 
> If mips64el is a target being built, "make check-acceptance" will
> automatically include this test by the use of the "arch:mips" tags.
> 
> Alternatively, this test can be run using:
> 
>     $ avocado run -t arch:mips64el tests/acceptance
>     $ avocado run -t machine:malta tests/acceptance
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  .travis.yml                            |  2 +-
>  tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>  2 files changed, 40 insertions(+), 1 deletion(-)
> 
> diff --git a/.travis.yml b/.travis.yml
> index 49f9016e6a..28648f7a61 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>  
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 0678ec91d2..20b845fce1 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -9,8 +9,11 @@
>  # later.  See the COPYING file in the top-level directory.
>  
>  import logging
> +import os
>  
>  from avocado_qemu import Test
> +from avocado.utils import process
> +from avocado.utils import archive
>  
>  
>  class BootLinuxConsole(Test):
> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_mips64el_malta(self):
> +        """
> +        This test requires the ar tool to extract "data.tar.gz" from
> +        the Debian package.
> +
> +        The kernel can be rebuilt using this Debian kernel source [1] and
> +        following the instructions on [2].
> +
> +        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
> +        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
> +
> +        :avocado: tags=arch:mips64el
> +        :avocado: tags=machine:malta
> +        """
> +        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
> +                   'pool/main/l/linux-2.6/'
> +                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
> +
> +        cwd = os.getcwd()
> +        os.chdir(self.workdir)
> +        process.run("ar x %s data.tar.gz" % deb_path)
> +        archive.extract("data.tar.gz", self.workdir)
> +        os.chdir(cwd)
> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
> +
> +        self.vm.set_machine('malta')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=ttyS0 printk.time=0'

What about having a generic KERNEL_COMMON_COMMAND_LINE='printk.time=0 '
then using
  kernel_command_line = KERNEL_COMMON_COMMAND_LINE + 'console=ttyS0'
in each tests of this series?

> +        self.vm.add_args('-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)
> 

Re: [Qemu-devel] [PATCH 01/18] scripts/qemu.py: log QEMU launch command line

[Alex Bennée]

Cleber Rosa <crosa@redhat.com> writes:

> Even when the launch of QEMU succeeds, it's useful to have the command
> line recorded.
>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  scripts/qemu.py | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/scripts/qemu.py b/scripts/qemu.py
> index 6e3b0e6771..ec3567d4e2 100644
> --- a/scripts/qemu.py
> +++ b/scripts/qemu.py
> @@ -312,6 +312,7 @@ class QEMUMachine(object):
>          self._pre_launch()
>          self._qemu_full_args = (self._wrapper + [self._binary] +
>                                  self._base_args() + self._args)
> +        LOG.debug('VM launch command: %r', ' '.join(self._qemu_full_args))
>          self._popen = subprocess.Popen(self._qemu_full_args,
>                                         stdin=devnull,
>                                         stdout=self._qemu_log_file,


Reviewed-by: Alex Bennée <alex.bennee@linaro.org>

--
Alex Bennée

Re: [Qemu-devel] [PATCH 02/18] Acceptance tests: show avocado test execution by default

[Alex Bennée]

Cleber Rosa <crosa@redhat.com> writes:

> The current version of the "check-acceptance" target will only show
> one line for execution of all tests.  That's probably OK if the tests
> to be run are quick enough and they're always the same.
>
> But, there's already one test alone that takes on average ~5 seconds
> to run, we intend to adapt the list of tests to match the user's build
> environment (among other choices).
>
> Because of that, let's present the default Avocado UI by default.
> Users can always choose a different output by setting the AVOCADO_SHOW
> variable.
>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>

> ---
>  .travis.yml            | 2 +-
>  tests/Makefile.include | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/.travis.yml b/.travis.yml
> index 93fd0164a0..844d514afa 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -188,7 +188,7 @@ matrix:
>      # Acceptance (Functional) tests
>      - env:
>          - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu"
> -        - TEST_CMD="make AVOCADO_SHOW=app check-acceptance"
> +        - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
>            packages:
> diff --git a/tests/Makefile.include b/tests/Makefile.include
> index f403a6571d..c73298740d 100644
> --- a/tests/Makefile.include
> +++ b/tests/Makefile.include
> @@ -956,7 +956,7 @@ TESTS_RESULTS_DIR=$(BUILD_DIR)/tests/results
>  # Controls the output generated by Avocado when running tests.
>  # Any number of command separated loggers are accepted.  For more
>  # information please refer to "avocado --help".
> -AVOCADO_SHOW=none
> +AVOCADO_SHOW=app
>
>  PYTHON3 = $(shell $(PYTHON) -c 'import sys; print(1 if sys.version_info >= (3, 0) else 0)')
>  ifeq ($(PYTHON3), 1)


--
Alex Bennée

Re: [Qemu-devel] [PATCH 14/18] Boot Linux Console Test: add a test for ppc64 + pseries

[Alex Bennée]

Cleber Rosa <crosa@redhat.com> writes:

> Just like the previous tests, boots a Linux kernel on a ppc64 target
> using the pseries machine.

So running this on my rather slow SynQuacer I get:

 (04/16) /home/alex/lsrc/qemu.git/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_ppc64_pseries: INTERRUPTED: Test reported status but did not finish\nRunner error occurred: Timeout reached\nOriginal status: ERROR\n{'name': '04-/home/alex/lsrc/qemu.git/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_ppc64_pseries', 'logdir': '/home/alex/lsrc/qemu.git/te... (60.93 s)

which I'm guessing is a timeout occurring.

I wonder if that means we should:

  a) set timeouts longer for when running on TCG
  or
  b) split tests into TCG and KVM tests and select KVM tests on appropriate HW

The qemu.py code has (slightly flawed) logic for detecting KVM and
passing --enable-kvm. Maybe we should be doing that here?

>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>  .travis.yml                            |  2 +-
>  scripts/qemu.py                        |  1 -
>  tests/acceptance/boot_linux_console.py | 19 +++++++++++++++++++
>  3 files changed, 20 insertions(+), 2 deletions(-)
>
> diff --git a/.travis.yml b/.travis.yml
> index 28648f7a61..54100eea5a 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>
>      # Acceptance (Functional) tests
>      - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
>          - TEST_CMD="make check-acceptance"
>        addons:
>          apt:
> diff --git a/scripts/qemu.py b/scripts/qemu.py
> index ef84b0f843..1531e28fc1 100644
> --- a/scripts/qemu.py
> +++ b/scripts/qemu.py
> @@ -37,7 +37,6 @@ CONSOLE_DEV_TYPES = {
>      r'^clipper$': 'isa-serial',
>      r'^(pc.*|q35.*|isapc)$': 'isa-serial',
>      r'^(40p|powernv|prep)$': 'isa-serial',
> -    r'^pseries.*': 'spapr-vty',
>      r'^s390-ccw-virtio.*': 'sclpconsole',
>      }
>
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 20b845fce1..f3ccd23a7a 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -119,3 +119,22 @@ class BootLinuxConsole(Test):
>          self.vm.launch()
>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>          self.wait_for_console_pattern(console_pattern)
> +
> +    def test_ppc64_pseries(self):
> +        """
> +        :avocado: tags=arch:ppc64
> +        :avocado: tags=machine:pseries
> +        """
> +        kernel_url = ('http://mirrors.rit.edu/fedora/fedora-secondary/'
> +                      'releases/29/Everything/ppc64le/os/ppc/ppc64/vmlinuz')
> +        kernel_hash = '3fe04abfc852b66653b8c3c897a59a689270bc77'
> +        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
> +
> +        self.vm.set_machine('pseries')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=hvc0'
> +        self.vm.add_args('-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)


--
Alex Bennée

Re: [Qemu-devel] [PATCH 05/18] Acceptance tests: introduce arch parameter and attribute

[Cleber Rosa]

On 1/18/19 9:28 AM, Caio Carrara wrote:
> Hi, Cleber.
> 
> On Thu, Jan 17, 2019 at 01:56:15PM -0500, Cleber Rosa wrote:
>> It's useful to define the architecture that should be used in
>> situations such as:
>>  * the intended target of the QEMU binary to be used on tests
>>  * the architecture of code to be run within the QEMU binary, such
>>    as a kernel image or a full blown guest OS image
>>
>> This commit introduces both a test parameter and a test instance
>> attribute, that will contain such a value.
>>
>> Now, when the "arch" test parameter is given, it will influence the
>> selection of the default QEMU binary, if one is not given explicitly
>> by means of the "qemu_img" parameter.
>>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>  docs/devel/testing.rst                    | 18 ++++++++++++++++++
>>  tests/acceptance/avocado_qemu/__init__.py | 13 ++++++++++---
>>  2 files changed, 28 insertions(+), 3 deletions(-)
>>
>> diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
>> index 44c9b3ae74..60a16c5834 100644
>> --- a/docs/devel/testing.rst
>> +++ b/docs/devel/testing.rst
>> @@ -689,6 +689,17 @@ vm
>>  A QEMUMachine instance, initially configured according to the given
>>  ``qemu_bin`` parameter.
>>  
>> +arch
>> +~~~~
>> +
>> +The architecture that will be used on a number of different
>> +scenarios.  For instance, when a QEMU binary is not explicitly given,
>> +the one selected will depend on this attribute.
>> +
>> +The ``arch`` attribute will be set to the test parameter of the same
>> +name, and if one is not given explicitly, it will default to the
>> +system architecture (as given by ``uname``).
> 
> Would you think it's important to make explicit that it's the *host*
> system architecture? Just like you did in the docstrings of
> pick_default_qemu_bin().
> 

Yes, you're right.  Adding that.

Re: [Qemu-devel] [PATCH 05/18] Acceptance tests: introduce arch parameter and attribute

[Cleber Rosa]

On 1/18/19 9:28 AM, Caio Carrara wrote:
> Hi, Cleber.
> 
> On Thu, Jan 17, 2019 at 01:56:15PM -0500, Cleber Rosa wrote:
>> It's useful to define the architecture that should be used in
>> situations such as:
>>  * the intended target of the QEMU binary to be used on tests
>>  * the architecture of code to be run within the QEMU binary, such
>>    as a kernel image or a full blown guest OS image
>>
>> This commit introduces both a test parameter and a test instance
>> attribute, that will contain such a value.
>>
>> Now, when the "arch" test parameter is given, it will influence the
>> selection of the default QEMU binary, if one is not given explicitly
>> by means of the "qemu_img" parameter.
>>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>  docs/devel/testing.rst                    | 18 ++++++++++++++++++
>>  tests/acceptance/avocado_qemu/__init__.py | 13 ++++++++++---
>>  2 files changed, 28 insertions(+), 3 deletions(-)
>>
>> diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
>> index 44c9b3ae74..60a16c5834 100644
>> --- a/docs/devel/testing.rst
>> +++ b/docs/devel/testing.rst
>> @@ -689,6 +689,17 @@ vm
>>  A QEMUMachine instance, initially configured according to the given
>>  ``qemu_bin`` parameter.
>>  
>> +arch
>> +~~~~
>> +
>> +The architecture that will be used on a number of different
>> +scenarios.  For instance, when a QEMU binary is not explicitly given,
>> +the one selected will depend on this attribute.
>> +
>> +The ``arch`` attribute will be set to the test parameter of the same
>> +name, and if one is not given explicitly, it will default to the
>> +system architecture (as given by ``uname``).
> 
> Would you think it's important to make explicit that it's the *host*
> system architecture? Just like you did in the docstrings of
> pick_default_qemu_bin().
> 
>> +
>>  qemu_bin
>>  ~~~~~~~~
>>  
>> @@ -711,6 +722,13 @@ like the following:
>>  
>>    PARAMS (key=qemu_bin, path=*, default=x86_64-softmmu/qemu-system-x86_64) => 'x86_64-softmmu/qemu-system-x86_64
>>  
>> +arch
>> +~~~~
>> +
>> +The architecture that will be used on a number of different scenarios.
>> +This parameter has a direct relation with the ``arch`` attribute.  If
>> +not given, it will default to None.
>> +
>>  qemu_bin
>>  ~~~~~~~~
>>  
>> diff --git a/tests/acceptance/avocado_qemu/__init__.py b/tests/acceptance/avocado_qemu/__init__.py
>> index d8d5b48dac..7a38851753 100644
>> --- a/tests/acceptance/avocado_qemu/__init__.py
>> +++ b/tests/acceptance/avocado_qemu/__init__.py
>> @@ -23,16 +23,22 @@ def is_readable_executable_file(path):
>>      return os.path.isfile(path) and os.access(path, os.R_OK | os.X_OK)
>>  
>>  
>> -def pick_default_qemu_bin():
>> +def pick_default_qemu_bin(arch=None):
>>      """
>>      Picks the path of a QEMU binary, starting either in the current working
>>      directory or in the source tree root directory.
>>  
>> +    :param arch: the arch to use when looking for a QEMU binary (the target
>> +                 will match the arch given).  If None (the default) arch
>> +                 will be the current host system arch (as given by
>> +                 :func:`os.uname`).
>> +    :type arch: str
>>      :returns: the path to the default QEMU binary or None if one could not
>>                be found
>>      :rtype: str or None
>>      """
>> -    arch = os.uname()[4]
>> +    if arch is None:
>> +        arch = os.uname()[4]
>>      qemu_bin_relative_path = os.path.join("%s-softmmu" % arch,
>>                                            "qemu-system-%s" % arch)
>>      if is_readable_executable_file(qemu_bin_relative_path):
>> @@ -47,8 +53,9 @@ def pick_default_qemu_bin():
>>  class Test(avocado.Test):
>>      def setUp(self):
>>          self.vm = None
>> +        self.arch = self.params.get('arch')
>>          self.qemu_bin = self.params.get('qemu_bin',
>> -                                        default=pick_default_qemu_bin())
>> +                                        default=pick_default_qemu_bin(self.arch))
> 
> Since it was used the named argument in function declaration I think
> it's more consistent also use the argument name here in the function
> call: `pick_default_qemu_bin(arch=self.arch)`
> 

Oops, I missed this part of your review in the previous message.
Anyway, I don't see a problem with either approach, so I'm fine changing it.

- Cleber.

Re: [Qemu-devel] [PATCH 06/18] Acceptance tests: use "arch:" tag to filter target specific tests

[Cleber Rosa]

On 1/18/19 5:38 AM, Cornelia Huck wrote:
> On Thu, 17 Jan 2019 13:56:16 -0500
> Cleber Rosa <crosa@redhat.com> wrote:
> 
>> Currently, the only test that contains some target architecture
>> information is "boot_linux_console.py" which test contains a "x86_64"
>> tag.  But that tag is not respected in the default execution, that is,
>> "make check-acceptance" doesn't do anything with it.
>>
>> That said, even the target architecture handling currently present in
>> the "avocado_qemu.Test" class, class is pretty limited.  For instance,
>> by default, it chooses a target based on the host architecture.
>>
>> Because the original implementation of the tags feature in Avocado did
>> not include any time of namespace or "key:val" mechanism, no tag has
>> relation to another tag.  The new implementation of the tags feature
>> from version 67.0 onwards, allows "key:val" tags, and because of that,
>> a test can be classified with a tag in a given key.  For instance, the
>> new proposed version of the "boot_linux_console.py" test, which
>> downloads and attempts to run a x86_64 kernel, is now tagged as:
>>
>>   :avocado: tags=arch:x86_64
>>
>> This means that it can be filtered (out) when no x86_64 target is
>> available.  At the same time, tests that don't have a "arch:" tag,
>> will not be filtered out.
>>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>  tests/Makefile.include                 | 3 +++
>>  tests/acceptance/boot_linux_console.py | 2 +-
>>  tests/requirements.txt                 | 2 +-
>>  3 files changed, 5 insertions(+), 2 deletions(-)
>>
> 
>> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
>> index 98324f7591..46b20bdfe2 100644
>> --- a/tests/acceptance/boot_linux_console.py
>> +++ b/tests/acceptance/boot_linux_console.py
>> @@ -19,7 +19,7 @@ class BootLinuxConsole(Test):
>>      and the kernel command line is properly passed from QEMU to the kernel
>>  
>>      :avocado: enable
>> -    :avocado: tags=x86_64
>> +    :avocado: tags=arch:x86_64
>>      """
>>  
>>      timeout = 60
> 
> You probably want to do the same change in virtio_version.py;
> otherwise, if I run the acceptance tests on s390x, it will run into
> timeouts (it looks like that test is intended to be run with x86
> machines anyway.)
> 

Right, good catch.  Also, the same applies to the "linux_initrd.py"
test, merged after this version.

- Cleber.

Re: [Qemu-devel] [PATCH 10/18] Boot Linux Console Test: refactor the console watcher into utility method

[Cleber Rosa]

On 1/22/19 5:06 AM, Philippe Mathieu-Daudé wrote:
> Hi Cleber,
> 
> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>> This introduces a utility method that monitors the console device and
>> looks for either a message that signals the test success or failure.
>>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>  tests/acceptance/boot_linux_console.py | 30 ++++++++++++++++++--------
>>  1 file changed, 21 insertions(+), 9 deletions(-)
>>
>> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
>> index 35b31162d4..278bb2be3d 100644
>> --- a/tests/acceptance/boot_linux_console.py
>> +++ b/tests/acceptance/boot_linux_console.py
>> @@ -23,6 +23,25 @@ class BootLinuxConsole(Test):
>>  
>>      timeout = 60
>>  
>> +    def wait_for_console_pattern(self, success_message,
>> +                                 failure_message='Kernel panic - not syncing'):
>> +        """
>> +        Waits for messages to appear on the console, while logging the content
>> +
>> +        :param success_message: if this message appears, test succeeds
>> +        :param failure_message: if this message appears, test fails
>> +        """
>> +        console = self.vm.console_socket.makefile()
>> +        console_logger = logging.getLogger('console')
>> +        while True:
>> +            msg = console.readline()
>> +            console_logger.debug(msg.strip())
>> +            if success_message in msg:
>> +                break
>> +            if failure_message in msg:
>> +                fail = 'Failure message found in console: %s' % failure_message
>> +                self.fail(fail)
>> +
> 
> This helper is more generic than the BootLinuxConsole class, can you
> move it out? I'd like to use it in test_uefi_ovmf_x86_64_pc and
> test_uefi_armvirtqemu_aarch64_virt.
> 
> Anyway this can be a follow-up patch, so regardless:
> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> 

Sure thing, I'll work on something on top of this.  Tracking it here in
the mean time:

https://trello.com/c/oJtL1nRJ/78-create-utility-out-of-waitforconsolepattern

- Cleber.

Re: [Qemu-devel] [PATCH 12/18] Boot Linux Console Test: add a test for mips + malta

[Cleber Rosa]

On 1/22/19 5:16 AM, Philippe Mathieu-Daudé wrote:
> Hi Cleber,
> 
> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>> From: Philippe Mathieu-Daudé <f4bug@amsat.org>
>>
>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>> board and verify the serial is working.  Also, it relies on the serial
>> device set by the machine itself.
>>
>> If mips is a target being built, "make check-acceptance" will
>> automatically include this test by the use of the "arch:mips" tags.
>>
>> Alternatively, this test can be run using:
>>
>>     $ avocado run -t arch:mips tests/acceptance
>>     $ avocado run -t machine:malta tests/acceptance
>>     $ avocado run -t endian:big tests/acceptance
>>
>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>  .travis.yml                            |  2 +-
>>  scripts/qemu.py                        |  1 -
>>  tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>>  3 files changed, 21 insertions(+), 2 deletions(-)
>>
>> diff --git a/.travis.yml b/.travis.yml
>> index 844d514afa..49f9016e6a 100644
>> --- a/.travis.yml
>> +++ b/.travis.yml
>> @@ -187,7 +187,7 @@ matrix:
>>  
>>      # Acceptance (Functional) tests
>>      - env:
>> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu"
>> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
>>          - TEST_CMD="make check-acceptance"
>>        addons:
>>          apt:
>> diff --git a/scripts/qemu.py b/scripts/qemu.py
>> index 88e1608b42..ef84b0f843 100644
>> --- a/scripts/qemu.py
>> +++ b/scripts/qemu.py
>> @@ -35,7 +35,6 @@ def kvm_available(target_arch=None):
>>  #: Maps machine types to the preferred console device types
>>  CONSOLE_DEV_TYPES = {
>>      r'^clipper$': 'isa-serial',
>> -    r'^malta': 'isa-serial',
>>      r'^(pc.*|q35.*|isapc)$': 'isa-serial',
>>      r'^(40p|powernv|prep)$': 'isa-serial',
>>      r'^pseries.*': 'spapr-vty',
>> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
>> index 278bb2be3d..0678ec91d2 100644
>> --- a/tests/acceptance/boot_linux_console.py
>> +++ b/tests/acceptance/boot_linux_console.py
>> @@ -60,3 +60,23 @@ class BootLinuxConsole(Test):
>>          self.vm.launch()
>>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>>          self.wait_for_console_pattern(console_pattern)
>> +
>> +    def test_mips_malta(self):
>> +        """
>> +        :avocado: tags=arch:mips
>> +        :avocado: tags=machine:malta
>> +        :avocado: tags=endian:big
>> +        """
>> +        kernel_url = ('http://people.debian.org/~aurel32/qemu/mips/'
>> +                      'vmlinux-3.2.0-4-4kc-malta')
>> +        kernel_hash = '592e384a4edc16dade52a6cd5c785c637bcbc9ad'
>> +        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
> 
> Alex was concerned about using this "not-so-easily-reproducible" binary:
> https://lists.gnu.org/archive/html/qemu-devel/2018-06/msg08441.html
>

I clearly missed that.

> So I sent another patch using a reproducible one, which you reviewed:
> https://lists.gnu.org/archive/html/qemu-devel/2018-10/msg02747.html
> 

And I clearly forgot about that :)

> Do you mind base this series on it?
> 
> Thanks,
> 
> Phil.

Of course!  Thanks for pointing that out.

- Cleber.

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Cleber Rosa]

On 1/22/19 5:19 AM, Philippe Mathieu-Daudé wrote:
> Hi Cleber,
> 
> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>> board and verify the serial is working.
>>
>> If mips64el is a target being built, "make check-acceptance" will
>> automatically include this test by the use of the "arch:mips" tags.
>>
>> Alternatively, this test can be run using:
>>
>>     $ avocado run -t arch:mips64el tests/acceptance
>>     $ avocado run -t machine:malta tests/acceptance
>>
>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>  .travis.yml                            |  2 +-
>>  tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>>  2 files changed, 40 insertions(+), 1 deletion(-)
>>
>> diff --git a/.travis.yml b/.travis.yml
>> index 49f9016e6a..28648f7a61 100644
>> --- a/.travis.yml
>> +++ b/.travis.yml
>> @@ -187,7 +187,7 @@ matrix:
>>  
>>      # Acceptance (Functional) tests
>>      - env:
>> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
>> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>>          - TEST_CMD="make check-acceptance"
>>        addons:
>>          apt:
>> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
>> index 0678ec91d2..20b845fce1 100644
>> --- a/tests/acceptance/boot_linux_console.py
>> +++ b/tests/acceptance/boot_linux_console.py
>> @@ -9,8 +9,11 @@
>>  # later.  See the COPYING file in the top-level directory.
>>  
>>  import logging
>> +import os
>>  
>>  from avocado_qemu import Test
>> +from avocado.utils import process
>> +from avocado.utils import archive
>>  
>>  
>>  class BootLinuxConsole(Test):
>> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>>          self.vm.launch()
>>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>>          self.wait_for_console_pattern(console_pattern)
>> +
>> +    def test_mips64el_malta(self):
>> +        """
>> +        This test requires the ar tool to extract "data.tar.gz" from
>> +        the Debian package.
>> +
>> +        The kernel can be rebuilt using this Debian kernel source [1] and
>> +        following the instructions on [2].
>> +
>> +        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
>> +        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
>> +
>> +        :avocado: tags=arch:mips64el
>> +        :avocado: tags=machine:malta
>> +        """
>> +        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
>> +                   'pool/main/l/linux-2.6/'
>> +                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
>> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
>> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
>> +
> 
> Can you refactor this code ...
> 
>> +        cwd = os.getcwd()
>> +        os.chdir(self.workdir)
>> +        process.run("ar x %s data.tar.gz" % deb_path)
>> +        archive.extract("data.tar.gz", self.workdir)
>> +        os.chdir(cwd)
>> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
> 
> ... as an helper function?
> 

Yes.  In fact, because of the kernel change in the previous patch, I've
added it there.

> Can be follow-up patch, so:
> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> 

In the longer term, I expect to add deb (and RPM, etc) support to the
avocado.utils.archive module:

https://trello.com/c/bQBnvkjv/1437-avocadoutilsarchive-support-rpm-deb-etc


Thanks!
- Cleber.

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Cleber Rosa]

On 1/22/19 5:57 AM, Philippe Mathieu-Daudé wrote:
> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>> board and verify the serial is working.
>>
>> If mips64el is a target being built, "make check-acceptance" will
>> automatically include this test by the use of the "arch:mips" tags.
>>
>> Alternatively, this test can be run using:
>>
>>     $ avocado run -t arch:mips64el tests/acceptance
>>     $ avocado run -t machine:malta tests/acceptance
>>
>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>  .travis.yml                            |  2 +-
>>  tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>>  2 files changed, 40 insertions(+), 1 deletion(-)
>>
>> diff --git a/.travis.yml b/.travis.yml
>> index 49f9016e6a..28648f7a61 100644
>> --- a/.travis.yml
>> +++ b/.travis.yml
>> @@ -187,7 +187,7 @@ matrix:
>>  
>>      # Acceptance (Functional) tests
>>      - env:
>> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
>> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>>          - TEST_CMD="make check-acceptance"
>>        addons:
>>          apt:
>> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
>> index 0678ec91d2..20b845fce1 100644
>> --- a/tests/acceptance/boot_linux_console.py
>> +++ b/tests/acceptance/boot_linux_console.py
>> @@ -9,8 +9,11 @@
>>  # later.  See the COPYING file in the top-level directory.
>>  
>>  import logging
>> +import os
>>  
>>  from avocado_qemu import Test
>> +from avocado.utils import process
>> +from avocado.utils import archive
>>  
>>  
>>  class BootLinuxConsole(Test):
>> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>>          self.vm.launch()
>>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>>          self.wait_for_console_pattern(console_pattern)
>> +
>> +    def test_mips64el_malta(self):
>> +        """
>> +        This test requires the ar tool to extract "data.tar.gz" from
>> +        the Debian package.
>> +
>> +        The kernel can be rebuilt using this Debian kernel source [1] and
>> +        following the instructions on [2].
>> +
>> +        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
>> +        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
>> +
>> +        :avocado: tags=arch:mips64el
>> +        :avocado: tags=machine:malta
>> +        """
>> +        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
>> +                   'pool/main/l/linux-2.6/'
>> +                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
>> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
>> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
>> +
>> +        cwd = os.getcwd()
>> +        os.chdir(self.workdir)
>> +        process.run("ar x %s data.tar.gz" % deb_path)
>> +        archive.extract("data.tar.gz", self.workdir)
>> +        os.chdir(cwd)
>> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
>> +
>> +        self.vm.set_machine('malta')
>> +        self.vm.set_console()
>> +        kernel_command_line = 'console=ttyS0 printk.time=0'
> 
> What about having a generic KERNEL_COMMON_COMMAND_LINE='printk.time=0 '
> then using
>   kernel_command_line = KERNEL_COMMON_COMMAND_LINE + 'console=ttyS0'
> in each tests of this series?
>

Not sure... I do like the fact that it's more uniform, but I don't like
the locality aspect.  Having to refer back to something defined either
in the module or class level for something so small is a big con, IMO.

I'll keep it AS IS for now, and if more people feel like it's a good
idea, or you feel stronger that it's a good idea, I can add it in the
next version.

Thanks!
- Cleber.

>> +        self.vm.add_args('-kernel', kernel_path,
>> +                         '-append', kernel_command_line)
>> +        self.vm.launch()
>> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
>> +        self.wait_for_console_pattern(console_pattern)
>>

Re: [Qemu-devel] [PATCH 14/18] Boot Linux Console Test: add a test for ppc64 + pseries

[Cleber Rosa]

On 1/22/19 11:07 AM, Alex Bennée wrote:
> 
> Cleber Rosa <crosa@redhat.com> writes:
> 
>> Just like the previous tests, boots a Linux kernel on a ppc64 target
>> using the pseries machine.
> 
> So running this on my rather slow SynQuacer I get:
> 
>  (04/16) /home/alex/lsrc/qemu.git/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_ppc64_pseries: INTERRUPTED: Test reported status but did not finish\nRunner error occurred: Timeout reached\nOriginal status: ERROR\n{'name': '04-/home/alex/lsrc/qemu.git/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_ppc64_pseries', 'logdir': '/home/alex/lsrc/qemu.git/te... (60.93 s)
> 
> which I'm guessing is a timeout occurring.
> 

Yes, that's what's happening.  It's hard to pinpoint, and control, the
sluggishness points in such a test running on a different environment.
For this one execution, I do trust your assessment, and it's most
certainly caused by your "slow SynQuacer", spending too much running
emulation code.

But, I'd like to mention that there are other possibilities.  One is
that you're hitting a "asset fetcher bug" that I recently fixed in
Avocado[1] (fix to be available on 68.0, to be released next Monday, Feb
4th).

Even with that bug fixed, I feel like it's unfair to test code to spend
its time waiting to download a file when it's not testing *the file
download itself*.  Because of that, there are plans to add an (optional)
job pre-processing step that will make sure the needed assets are in the
cache ahead of time[2][3].

> I wonder if that means we should:
> 
>   a) set timeouts longer for when running on TCG
>   or
>   b) split tests into TCG and KVM tests and select KVM tests on appropriate HW
> 

I wonder the same, and I believe this falls into a similar scenario
we've seen with the setup of console devices in the QEMUMachine class.
I've started by setting the device types defined at the framework level,
and then reverted to the machine's default devices (using '-serial'),
because the "default" behavior of QEMU is usually what a test writer
wants when not setting something else explicitly.

> The qemu.py code has (slightly flawed) logic for detecting KVM and
> passing --enable-kvm. Maybe we should be doing that here?
> 

I'm not sure.  IMO, the common question is: should we magically (at a
framework level) configure tests based on probed host environment
characteristics?  I feel like we should attempt to minimize that for the
sake of tests being more obvious and more easily reproducible.

And because of that, I'd go, *initially*, with an approach more similar
to your option "b".

Having said that, we don't want to rewrite most tests just to be able to
test with either KVM or TCG, if the tests are not explicitly testing KVM
or TCG.  At this point, using KVM or TCG is test/framework
*configuration*, and in Avocado we hope to solve this by having the
executed tests easily identifiable and reproducible (a test ID will
contain a information about the options passed, and a replay of the job
will apply the same configuration).

For now, I think the best approach is to increase the timeout, because I
think it's much worse to have to deal with false negatives (longer
execution times that don't really mean a failure), than having a test
possibly taking some more time to finish.

And sorry for extremely the long answer!
- Cleber.

[1] - https://github.com/avocado-framework/avocado/pull/2996
[2] -
https://trello.com/c/WPd4FrIy/1479-add-support-to-specify-assets-in-test-docstring
[3] - https://trello.com/c/CKP7YS6G/1481-on-cache-check-for-asset-fetcher

Re: [Qemu-devel] [PATCH 18/18] Boot Linux Console Test: add a test for alpha + clipper

[Cleber Rosa]

On 1/22/19 5:52 AM, Philippe Mathieu-Daudé wrote:
> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>> board and verify the serial is working.  One extra command added to
>> the QEMU command line is '-vga std', because the kernel used is
>> known to crash without it.
>>
>> If alpha is a target being built, "make check-acceptance" will
>> automatically include this test by the use of the "arch:alpha" tags.
>>
>> Alternatively, this test can be run using:
>>
>>     $ avocado run -t arch:alpha tests/acceptance
>>     $ avocado run -t machine:clipper tests/acceptance
>>
>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>  .travis.yml                            |  2 +-
>>  scripts/qemu.py                        |  1 -
>>  tests/acceptance/boot_linux_console.py | 22 ++++++++++++++++++++++
>>  3 files changed, 23 insertions(+), 2 deletions(-)
>>
>> diff --git a/.travis.yml b/.travis.yml
>> index 0d5a4b104b..73a113af87 100644
>> --- a/.travis.yml
>> +++ b/.travis.yml
>> @@ -187,7 +187,7 @@ matrix:
>>  
>>      # Acceptance (Functional) tests
>>      - env:
>> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu"
>> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu,alpha-softmmu"
>>          - TEST_CMD="make check-acceptance"
>>        addons:
>>          apt:
>> diff --git a/scripts/qemu.py b/scripts/qemu.py
>> index 1531e28fc1..a704da418a 100644
>> --- a/scripts/qemu.py
>> +++ b/scripts/qemu.py
>> @@ -34,7 +34,6 @@ def kvm_available(target_arch=None):
>>  
>>  #: Maps machine types to the preferred console device types
>>  CONSOLE_DEV_TYPES = {
>> -    r'^clipper$': 'isa-serial',
> 
> Why this change?
> 

Because we've come to conclusion (I believe :) that's better, whenever
possible, to let QEMU pick the default device type by machine.  The
discussion happened here:

https://lists.gnu.org/archive/html/qemu-devel/2018-10/msg04582.html

- Cleber.

Re: [Qemu-devel] [PATCH 14/18] Boot Linux Console Test: add a test for ppc64 + pseries

[Philippe Mathieu-Daudé]

On 1/31/19 3:37 AM, Cleber Rosa wrote:
> On 1/22/19 11:07 AM, Alex Bennée wrote:
>> Cleber Rosa <crosa@redhat.com> writes:
>>
>>> Just like the previous tests, boots a Linux kernel on a ppc64 target
>>> using the pseries machine.
>>
>> So running this on my rather slow SynQuacer I get:
>>
>>  (04/16) /home/alex/lsrc/qemu.git/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_ppc64_pseries: INTERRUPTED: Test reported status but did not finish\nRunner error occurred: Timeout reached\nOriginal status: ERROR\n{'name': '04-/home/alex/lsrc/qemu.git/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_ppc64_pseries', 'logdir': '/home/alex/lsrc/qemu.git/te... (60.93 s)
>>
>> which I'm guessing is a timeout occurring.
>>
> 
> Yes, that's what's happening.  It's hard to pinpoint, and control, the
> sluggishness points in such a test running on a different environment.
> For this one execution, I do trust your assessment, and it's most
> certainly caused by your "slow SynQuacer", spending too much running
> emulation code.
> 
> But, I'd like to mention that there are other possibilities.  One is
> that you're hitting a "asset fetcher bug" that I recently fixed in
> Avocado[1] (fix to be available on 68.0, to be released next Monday, Feb
> 4th).
> 
> Even with that bug fixed, I feel like it's unfair to test code to spend
> its time waiting to download a file when it's not testing *the file
> download itself*.  Because of that, there are plans to add an (optional)
> job pre-processing step that will make sure the needed assets are in the
> cache ahead of time[2][3].
> 
>> I wonder if that means we should:
>>
>>   a) set timeouts longer for when running on TCG

I hit the same problem with VM tests, and suggested a poor "increase
timeout" patch [1].

Then Peter sent a different patch [2] which happens to inadvertently
Dictionary resolve my problem, since the longer a VM took to boot on the
Cavium ThunderX I have access is 288 seconds, which is closely below the
300 seconds limit =) I understood nobody seemed to really care about
testing the x86 TCG backend this way, so I didn't worry much.

[1] http://lists.nongnu.org/archive/html/qemu-devel/2018-07/msg03416.html
[2] http://lists.nongnu.org/archive/html/qemu-devel/2018-08/msg04977.html

>>   or
>>   b) split tests into TCG and KVM tests and select KVM tests on appropriate HW
>>
> 
> I wonder the same, and I believe this falls into a similar scenario
> we've seen with the setup of console devices in the QEMUMachine class.
> I've started by setting the device types defined at the framework level,
> and then reverted to the machine's default devices (using '-serial'),
> because the "default" behavior of QEMU is usually what a test writer
> wants when not setting something else explicitly.
> 
>> The qemu.py code has (slightly flawed) logic for detecting KVM and
>> passing --enable-kvm. Maybe we should be doing that here?
>>
> 
> I'm not sure.  IMO, the common question is: should we magically (at a
> framework level) configure tests based on probed host environment
> characteristics?  I feel like we should attempt to minimize that for the
> sake of tests being more obvious and more easily reproducible.

I agree we shouldn't randomly test different features, but rather
explicitly add 2 tests (TCG/KVM), and if it is not possible to run a
test, mark it as SKIPPED.

An user with KVM available would then have to run --filter-out=tcg, or
build QEMU with --disable-tcg.

> And because of that, I'd go, *initially*, with an approach more similar
> to your option "b".
> 
> Having said that, we don't want to rewrite most tests just to be able to
> test with either KVM or TCG, if the tests are not explicitly testing KVM
> or TCG.  At this point, using KVM or TCG is test/framework
> *configuration*, and in Avocado we hope to solve this by having the
> executed tests easily identifiable and reproducible (a test ID will
> contain a information about the options passed, and a replay of the job
> will apply the same configuration).
> 
> For now, I think the best approach is to increase the timeout, because I
> think it's much worse to have to deal with false negatives (longer
> execution times that don't really mean a failure), than having a test
> possibly taking some more time to finish.
> 
> And sorry for extremely the long answer!
> - Cleber.
> 
> [1] - https://github.com/avocado-framework/avocado/pull/2996
> [2] -
> https://trello.com/c/WPd4FrIy/1479-add-support-to-specify-assets-in-test-docstring
> [3] - https://trello.com/c/CKP7YS6G/1481-on-cache-check-for-asset-fetcher
> 

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Philippe Mathieu-Daudé]

On 1/31/19 2:26 AM, Cleber Rosa wrote:
> On 1/22/19 5:19 AM, Philippe Mathieu-Daudé wrote:
>> Hi Cleber,
>>
>> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>>> board and verify the serial is working.
>>>
>>> If mips64el is a target being built, "make check-acceptance" will
>>> automatically include this test by the use of the "arch:mips" tags.
>>>
>>> Alternatively, this test can be run using:
>>>
>>>     $ avocado run -t arch:mips64el tests/acceptance
>>>     $ avocado run -t machine:malta tests/acceptance
>>>
>>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>>> ---
>>>  .travis.yml                            |  2 +-
>>>  tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>>>  2 files changed, 40 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/.travis.yml b/.travis.yml
>>> index 49f9016e6a..28648f7a61 100644
>>> --- a/.travis.yml
>>> +++ b/.travis.yml
>>> @@ -187,7 +187,7 @@ matrix:
>>>  
>>>      # Acceptance (Functional) tests
>>>      - env:
>>> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
>>> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>>>          - TEST_CMD="make check-acceptance"
>>>        addons:
>>>          apt:
>>> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
>>> index 0678ec91d2..20b845fce1 100644
>>> --- a/tests/acceptance/boot_linux_console.py
>>> +++ b/tests/acceptance/boot_linux_console.py
>>> @@ -9,8 +9,11 @@
>>>  # later.  See the COPYING file in the top-level directory.
>>>  
>>>  import logging
>>> +import os
>>>  
>>>  from avocado_qemu import Test
>>> +from avocado.utils import process
>>> +from avocado.utils import archive
>>>  
>>>  
>>>  class BootLinuxConsole(Test):
>>> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>>>          self.vm.launch()
>>>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>>>          self.wait_for_console_pattern(console_pattern)
>>> +
>>> +    def test_mips64el_malta(self):
>>> +        """
>>> +        This test requires the ar tool to extract "data.tar.gz" from
>>> +        the Debian package.
>>> +
>>> +        The kernel can be rebuilt using this Debian kernel source [1] and
>>> +        following the instructions on [2].
>>> +
>>> +        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
>>> +        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
>>> +
>>> +        :avocado: tags=arch:mips64el
>>> +        :avocado: tags=machine:malta
>>> +        """
>>> +        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
>>> +                   'pool/main/l/linux-2.6/'
>>> +                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
>>> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
>>> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
>>> +
>>
>> Can you refactor this code ...
>>
>>> +        cwd = os.getcwd()
>>> +        os.chdir(self.workdir)
>>> +        process.run("ar x %s data.tar.gz" % deb_path)
>>> +        archive.extract("data.tar.gz", self.workdir)
>>> +        os.chdir(cwd)
>>> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
>>
>> ... as an helper function?
>>
> 
> Yes.  In fact, because of the kernel change in the previous patch, I've
> added it there.
> 
>> Can be follow-up patch, so:
>> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>>
> 
> In the longer term, I expect to add deb (and RPM, etc) support to the
> avocado.utils.archive module:
> 
> https://trello.com/c/bQBnvkjv/1437-avocadoutilsarchive-support-rpm-deb-etc

Awesome, thanks!

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Philippe Mathieu-Daudé]

On 1/31/19 2:34 AM, Cleber Rosa wrote:
> On 1/22/19 5:57 AM, Philippe Mathieu-Daudé wrote:
>> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>>> board and verify the serial is working.
>>>
>>> If mips64el is a target being built, "make check-acceptance" will
>>> automatically include this test by the use of the "arch:mips" tags.
>>>
>>> Alternatively, this test can be run using:
>>>
>>>     $ avocado run -t arch:mips64el tests/acceptance
>>>     $ avocado run -t machine:malta tests/acceptance
>>>
>>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>>> ---
>>>  .travis.yml                            |  2 +-
>>>  tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>>>  2 files changed, 40 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/.travis.yml b/.travis.yml
>>> index 49f9016e6a..28648f7a61 100644
>>> --- a/.travis.yml
>>> +++ b/.travis.yml
>>> @@ -187,7 +187,7 @@ matrix:
>>>  
>>>      # Acceptance (Functional) tests
>>>      - env:
>>> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
>>> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>>>          - TEST_CMD="make check-acceptance"
>>>        addons:
>>>          apt:
>>> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
>>> index 0678ec91d2..20b845fce1 100644
>>> --- a/tests/acceptance/boot_linux_console.py
>>> +++ b/tests/acceptance/boot_linux_console.py
>>> @@ -9,8 +9,11 @@
>>>  # later.  See the COPYING file in the top-level directory.
>>>  
>>>  import logging
>>> +import os
>>>  
>>>  from avocado_qemu import Test
>>> +from avocado.utils import process
>>> +from avocado.utils import archive
>>>  
>>>  
>>>  class BootLinuxConsole(Test):
>>> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>>>          self.vm.launch()
>>>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>>>          self.wait_for_console_pattern(console_pattern)
>>> +
>>> +    def test_mips64el_malta(self):
>>> +        """
>>> +        This test requires the ar tool to extract "data.tar.gz" from
>>> +        the Debian package.
>>> +
>>> +        The kernel can be rebuilt using this Debian kernel source [1] and
>>> +        following the instructions on [2].
>>> +
>>> +        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
>>> +        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
>>> +
>>> +        :avocado: tags=arch:mips64el
>>> +        :avocado: tags=machine:malta
>>> +        """
>>> +        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
>>> +                   'pool/main/l/linux-2.6/'
>>> +                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
>>> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
>>> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
>>> +
>>> +        cwd = os.getcwd()
>>> +        os.chdir(self.workdir)
>>> +        process.run("ar x %s data.tar.gz" % deb_path)
>>> +        archive.extract("data.tar.gz", self.workdir)
>>> +        os.chdir(cwd)
>>> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
>>> +
>>> +        self.vm.set_machine('malta')
>>> +        self.vm.set_console()
>>> +        kernel_command_line = 'console=ttyS0 printk.time=0'
>>
>> What about having a generic KERNEL_COMMON_COMMAND_LINE='printk.time=0 '
>> then using
>>   kernel_command_line = KERNEL_COMMON_COMMAND_LINE + 'console=ttyS0'
>> in each tests of this series?
>>
> 
> Not sure... I do like the fact that it's more uniform, but I don't like
> the locality aspect.  Having to refer back to something defined either
> in the module or class level for something so small is a big con, IMO.
> 
> I'll keep it AS IS for now, and if more people feel like it's a good
> idea, or you feel stronger that it's a good idea, I can add it in the
> next version.

I'm fine with that, however I'd really like all the Linux tests to use
"printk.time=0" in their kernel cmdline, it really simplify
parsing/diffing the console output.

>>> +        self.vm.add_args('-kernel', kernel_path,
>>> +                         '-append', kernel_command_line)
>>> +        self.vm.launch()
>>> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
>>> +        self.wait_for_console_pattern(console_pattern)
>>>
> 

Re: [Qemu-devel] [PATCH 18/18] Boot Linux Console Test: add a test for alpha + clipper

[Philippe Mathieu-Daudé]

On 1/31/19 3:53 AM, Cleber Rosa wrote:
> On 1/22/19 5:52 AM, Philippe Mathieu-Daudé wrote:
>> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>>> board and verify the serial is working.  One extra command added to
>>> the QEMU command line is '-vga std', because the kernel used is
>>> known to crash without it.
>>>
>>> If alpha is a target being built, "make check-acceptance" will
>>> automatically include this test by the use of the "arch:alpha" tags.
>>>
>>> Alternatively, this test can be run using:
>>>
>>>     $ avocado run -t arch:alpha tests/acceptance
>>>     $ avocado run -t machine:clipper tests/acceptance
>>>
>>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>>> ---
>>>  .travis.yml                            |  2 +-
>>>  scripts/qemu.py                        |  1 -
>>>  tests/acceptance/boot_linux_console.py | 22 ++++++++++++++++++++++
>>>  3 files changed, 23 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/.travis.yml b/.travis.yml
>>> index 0d5a4b104b..73a113af87 100644
>>> --- a/.travis.yml
>>> +++ b/.travis.yml
>>> @@ -187,7 +187,7 @@ matrix:
>>>  
>>>      # Acceptance (Functional) tests
>>>      - env:
>>> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu"
>>> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu,alpha-softmmu"
>>>          - TEST_CMD="make check-acceptance"
>>>        addons:
>>>          apt:
>>> diff --git a/scripts/qemu.py b/scripts/qemu.py
>>> index 1531e28fc1..a704da418a 100644
>>> --- a/scripts/qemu.py
>>> +++ b/scripts/qemu.py
>>> @@ -34,7 +34,6 @@ def kvm_available(target_arch=None):
>>>  
>>>  #: Maps machine types to the preferred console device types
>>>  CONSOLE_DEV_TYPES = {
>>> -    r'^clipper$': 'isa-serial',
>>
>> Why this change?
>>
> 
> Because we've come to conclusion (I believe :) that's better, whenever
> possible, to let QEMU pick the default device type by machine.  The
> discussion happened here:
> 
> https://lists.gnu.org/archive/html/qemu-devel/2018-10/msg04582.html

I was expecting this change to be global (in a single patch previous to
the target-specific patches). If you prefer to do it a target at a time
I'm OK with it but you should add a comment about it in the commit IMHO.

Regards,

Phil.

Re: [Qemu-devel] [PATCH 05/18] Acceptance tests: introduce arch parameter and attribute

[Wainer dos Santos Moschetta]

Hi Cleber,

On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> It's useful to define the architecture that should be used in
> situations such as:
>   * the intended target of the QEMU binary to be used on tests
>   * the architecture of code to be run within the QEMU binary, such
>     as a kernel image or a full blown guest OS image
>
> This commit introduces both a test parameter and a test instance
> attribute, that will contain such a value.
>
> Now, when the "arch" test parameter is given, it will influence the
> selection of the default QEMU binary, if one is not given explicitly
> by means of the "qemu_img" parameter.
>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>   docs/devel/testing.rst                    | 18 ++++++++++++++++++
>   tests/acceptance/avocado_qemu/__init__.py | 13 ++++++++++---
>   2 files changed, 28 insertions(+), 3 deletions(-)
>
> diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
> index 44c9b3ae74..60a16c5834 100644
> --- a/docs/devel/testing.rst
> +++ b/docs/devel/testing.rst
> @@ -689,6 +689,17 @@ vm
>   A QEMUMachine instance, initially configured according to the given
>   ``qemu_bin`` parameter.
>   
> +arch
> +~~~~
> +
> +The architecture that will be used on a number of different
> +scenarios.  For instance, when a QEMU binary is not explicitly given,
> +the one selected will depend on this attribute.
> +
> +The ``arch`` attribute will be set to the test parameter of the same
> +name, and if one is not given explicitly, it will default to the
> +system architecture (as given by ``uname``).

Unless I'm missing something, the arch attribute will be None if I don't 
pass the arch parameter nor set the arch tag.

------
[root@xxx qemu]# cat tests/acceptance/arch.py
from avocado_qemu import Test

class Arch(Test):
     """
     :avocado: enable
     """
     def test(self):
         self.assertEqual(self.arch, "x86_64")
[root@xxx qemu]# ./tests/venv/bin/avocado run tests/acceptance/arch.py
JOB ID     : 032aabb8ec2364944d0e29df3b61931b42f4587c
JOB LOG    : /root/avocado/job-results/job-2019-01-31T08.42-032aabb/job.log
  (1/1) tests/acceptance/arch.py:Arch.test: FAIL: None != 'x86_64' (0.03 s)
RESULTS    : PASS 0 | ERROR 0 | FAIL 1 | SKIP 0 | WARN 0 | INTERRUPT 0 | 
CANCEL 0
JOB TIME   : 0.20 s
[root@xxx qemu]# ./tests/venv/bin/avocado run -p arch=x86_64 
tests/acceptance/arch.py
JOB ID     : 1913a73510cab4b6b118bd7e94e2746011368628
JOB LOG    : /root/avocado/job-results/job-2019-01-31T08.42-1913a73/job.log
  (1/1) tests/acceptance/arch.py:Arch.test: PASS (0.01 s)
RESULTS    : PASS 1 | ERROR 0 | FAIL 0 | SKIP 0 | WARN 0 | INTERRUPT 0 | 
CANCEL 0
JOB TIME   : 0.19 s
[root@xxx qemu]# cat tests/acceptance/arch_tag.py
from avocado_qemu import Test

class Arch(Test):
     """
     :avocado: enable
     :avocado: tags=arch:x86_64
     """
     def test(self):
         self.assertEqual(self.arch, "x86_64")
[root@xxx qemu]# ./tests/venv/bin/avocado run tests/acceptance/arch_tag.py
JOB ID     : bc85be065586b8e9cf371489e1801489910e41b0
JOB LOG    : /root/avocado/job-results/job-2019-01-31T08.43-bc85be0/job.log
  (1/1) tests/acceptance/arch_tag.py:Arch.test: PASS (0.01 s)
RESULTS    : PASS 1 | ERROR 0 | FAIL 0 | SKIP 0 | WARN 0 | INTERRUPT 0 | 
CANCEL 0
JOB TIME   : 0.19 s
------

- Wainer

> +
>   qemu_bin
>   ~~~~~~~~
>   
> @@ -711,6 +722,13 @@ like the following:
>   
>     PARAMS (key=qemu_bin, path=*, default=x86_64-softmmu/qemu-system-x86_64) => 'x86_64-softmmu/qemu-system-x86_64
>   
> +arch
> +~~~~
> +
> +The architecture that will be used on a number of different scenarios.
> +This parameter has a direct relation with the ``arch`` attribute.  If
> +not given, it will default to None.
> +
>   qemu_bin
>   ~~~~~~~~
>   
> diff --git a/tests/acceptance/avocado_qemu/__init__.py b/tests/acceptance/avocado_qemu/__init__.py
> index d8d5b48dac..7a38851753 100644
> --- a/tests/acceptance/avocado_qemu/__init__.py
> +++ b/tests/acceptance/avocado_qemu/__init__.py
> @@ -23,16 +23,22 @@ def is_readable_executable_file(path):
>       return os.path.isfile(path) and os.access(path, os.R_OK | os.X_OK)
>   
>   
> -def pick_default_qemu_bin():
> +def pick_default_qemu_bin(arch=None):
>       """
>       Picks the path of a QEMU binary, starting either in the current working
>       directory or in the source tree root directory.
>   
> +    :param arch: the arch to use when looking for a QEMU binary (the target
> +                 will match the arch given).  If None (the default) arch
> +                 will be the current host system arch (as given by
> +                 :func:`os.uname`).
> +    :type arch: str
>       :returns: the path to the default QEMU binary or None if one could not
>                 be found
>       :rtype: str or None
>       """
> -    arch = os.uname()[4]
> +    if arch is None:
> +        arch = os.uname()[4]
>       qemu_bin_relative_path = os.path.join("%s-softmmu" % arch,
>                                             "qemu-system-%s" % arch)
>       if is_readable_executable_file(qemu_bin_relative_path):
> @@ -47,8 +53,9 @@ def pick_default_qemu_bin():
>   class Test(avocado.Test):
>       def setUp(self):
>           self.vm = None
> +        self.arch = self.params.get('arch')
>           self.qemu_bin = self.params.get('qemu_bin',
> -                                        default=pick_default_qemu_bin())
> +                                        default=pick_default_qemu_bin(self.arch))
>           if self.qemu_bin is None:
>               self.cancel("No QEMU binary defined or found in the source tree")
>           self.vm = QEMUMachine(self.qemu_bin)

Re: [Qemu-devel] [PATCH 00/18] Acceptance Tests: target architecture support

[Cleber Rosa]

On 1/22/19 5:48 AM, Philippe Mathieu-Daudé wrote:
> Hi Aleksandar,
> 
> On 1/21/19 11:15 PM, Aleksandar Markovic wrote:
>>> From: Cleber Rosa <crosa@redhat.com>
>>
>>>   class My(Test):
>>>       def test_nx_cpu_flag(self):
>>>           """
>>>           :avocado: tags=arch:x86_64
>>>           """
>>>           test_code()
>>
>>> The value of the "arch" key, in this case, "x86_64" will be used when
>>> selecting the QEMU binary to use in the test.  At the same time, if
>>> "x86_64-softmmu" is not a built target, the test will be filtered out
>>> by "make check-acceptance"[3].
>>
>> I think, the term "arch" is a little problematic in QEMU parlance. IMHO,
>> "target" should be used instead. ("arch" is used in Linux kernel community)
> 
> Using target_arch/host_arch might be more explicit, but host_arch is not
> very relevant regarding Avocado (except to choose the correct TCG
> binary), so usually arch implies target_arch.
> 
> I mean, it is unlikely you enforce --host_arch on the command line to
> run tests, this using --arch instead of --target_arch looks OK to me,
> since host_arch can be guessed.
> 

Naming things is hard, so this is a valid discussion.  But, I have to
say that I also find "arch" in this context to be descriptive enough.

>>
>> The overall structure of the "tags" should be a little different. My
>> suggestion:
>>
>> "target"
>> "isa" (instruction set architecture, determeines how the kernel and rootfs are built)
> 
> In QEMU, "isa" is implicit with CPU, isnt' it?
> 
> I.e. I use:
> 
> $ mips64el-softmmu/qemu-system-mips64el -M Malta -cpu mips64dspr2
> 
>> "cpu"
>> "machine"
>>
>> This would allow clear view what a particular acceptance test tests, and will
>> enforce consistency and clarity in the test organization across the board.
> 
> Maybe the problem you are pointing out is not Avocado test organization
> but QEMU CPU organization... (which also bother me with boards able to
> use different SoC, with different peripherals or cpus).
> 
> What is your idea on passing arch/cpu/isa to start QEMU?
> 
>>
>> That said, I am very excited about this series.
> 
> Me too :)
> 
> Regards,
> 
> Phil.
> 

Thanks! For the encouragement, help and input!

- Cleber.

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Cleber Rosa]

On 1/31/19 5:26 AM, Philippe Mathieu-Daudé wrote:
> On 1/31/19 2:34 AM, Cleber Rosa wrote:
>> On 1/22/19 5:57 AM, Philippe Mathieu-Daudé wrote:
>>> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>>>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>>>> board and verify the serial is working.
>>>>
>>>> If mips64el is a target being built, "make check-acceptance" will
>>>> automatically include this test by the use of the "arch:mips" tags.
>>>>
>>>> Alternatively, this test can be run using:
>>>>
>>>>     $ avocado run -t arch:mips64el tests/acceptance
>>>>     $ avocado run -t machine:malta tests/acceptance
>>>>
>>>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>>>> ---
>>>>  .travis.yml                            |  2 +-
>>>>  tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>>>>  2 files changed, 40 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/.travis.yml b/.travis.yml
>>>> index 49f9016e6a..28648f7a61 100644
>>>> --- a/.travis.yml
>>>> +++ b/.travis.yml
>>>> @@ -187,7 +187,7 @@ matrix:
>>>>  
>>>>      # Acceptance (Functional) tests
>>>>      - env:
>>>> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
>>>> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>>>>          - TEST_CMD="make check-acceptance"
>>>>        addons:
>>>>          apt:
>>>> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
>>>> index 0678ec91d2..20b845fce1 100644
>>>> --- a/tests/acceptance/boot_linux_console.py
>>>> +++ b/tests/acceptance/boot_linux_console.py
>>>> @@ -9,8 +9,11 @@
>>>>  # later.  See the COPYING file in the top-level directory.
>>>>  
>>>>  import logging
>>>> +import os
>>>>  
>>>>  from avocado_qemu import Test
>>>> +from avocado.utils import process
>>>> +from avocado.utils import archive
>>>>  
>>>>  
>>>>  class BootLinuxConsole(Test):
>>>> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>>>>          self.vm.launch()
>>>>          console_pattern = 'Kernel command line: %s' % kernel_command_line
>>>>          self.wait_for_console_pattern(console_pattern)
>>>> +
>>>> +    def test_mips64el_malta(self):
>>>> +        """
>>>> +        This test requires the ar tool to extract "data.tar.gz" from
>>>> +        the Debian package.
>>>> +
>>>> +        The kernel can be rebuilt using this Debian kernel source [1] and
>>>> +        following the instructions on [2].
>>>> +
>>>> +        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
>>>> +        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
>>>> +
>>>> +        :avocado: tags=arch:mips64el
>>>> +        :avocado: tags=machine:malta
>>>> +        """
>>>> +        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
>>>> +                   'pool/main/l/linux-2.6/'
>>>> +                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
>>>> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
>>>> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
>>>> +
>>>> +        cwd = os.getcwd()
>>>> +        os.chdir(self.workdir)
>>>> +        process.run("ar x %s data.tar.gz" % deb_path)
>>>> +        archive.extract("data.tar.gz", self.workdir)
>>>> +        os.chdir(cwd)
>>>> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
>>>> +
>>>> +        self.vm.set_machine('malta')
>>>> +        self.vm.set_console()
>>>> +        kernel_command_line = 'console=ttyS0 printk.time=0'
>>>
>>> What about having a generic KERNEL_COMMON_COMMAND_LINE='printk.time=0 '
>>> then using
>>>   kernel_command_line = KERNEL_COMMON_COMMAND_LINE + 'console=ttyS0'
>>> in each tests of this series?
>>>
>>
>> Not sure... I do like the fact that it's more uniform, but I don't like
>> the locality aspect.  Having to refer back to something defined either
>> in the module or class level for something so small is a big con, IMO.
>>
>> I'll keep it AS IS for now, and if more people feel like it's a good
>> idea, or you feel stronger that it's a good idea, I can add it in the
>> next version.
> 
> I'm fine with that, however I'd really like all the Linux tests to use
> "printk.time=0" in their kernel cmdline, it really simplify
> parsing/diffing the console output.
> 

OK, I'll include that in v2.

Thanks!
- Cleber.

>>>> +        self.vm.add_args('-kernel', kernel_path,
>>>> +                         '-append', kernel_command_line)
>>>> +        self.vm.launch()
>>>> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
>>>> +        self.wait_for_console_pattern(console_pattern)
>>>>
>>

Re: [Qemu-devel] [PATCH 10/18] Boot Linux Console Test: refactor the console watcher into utility method

[Wainer dos Santos Moschetta]

Hi Cleber,

On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> This introduces a utility method that monitors the console device and
> looks for either a message that signals the test success or failure.
>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>   tests/acceptance/boot_linux_console.py | 30 ++++++++++++++++++--------
>   1 file changed, 21 insertions(+), 9 deletions(-)
>
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 35b31162d4..278bb2be3d 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -23,6 +23,25 @@ class BootLinuxConsole(Test):
>   
>       timeout = 60
>   
> +    def wait_for_console_pattern(self, success_message,
> +                                 failure_message='Kernel panic - not syncing'):
> +        """
> +        Waits for messages to appear on the console, while logging the content
> +
> +        :param success_message: if this message appears, test succeeds
> +        :param failure_message: if this message appears, test fails
> +        """

What's the docstring format used here? I am wondering if :type is optional.

- Wainer

> +        console = self.vm.console_socket.makefile()
> +        console_logger = logging.getLogger('console')
> +        while True:
> +            msg = console.readline()
> +            console_logger.debug(msg.strip())
> +            if success_message in msg:
> +                break
> +            if failure_message in msg:
> +                fail = 'Failure message found in console: %s' % failure_message
> +                self.fail(fail)
> +
>       def test_x86_64_pc(self):
>           """
>           :avocado: tags=arch:x86_64
> @@ -39,12 +58,5 @@ class BootLinuxConsole(Test):
>           self.vm.add_args('-kernel', kernel_path,
>                            '-append', kernel_command_line)
>           self.vm.launch()
> -        console = self.vm.console_socket.makefile()
> -        console_logger = logging.getLogger('console')
> -        while True:
> -            msg = console.readline()
> -            console_logger.debug(msg.strip())
> -            if 'Kernel command line: %s' % kernel_command_line in msg:
> -                break
> -            if 'Kernel panic - not syncing' in msg:
> -                self.fail("Kernel panic reached")
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)

Re: [Qemu-devel] [PATCH 00/18] Acceptance Tests: target architecture support

[no-reply]

Patchew URL: https://patchew.org/QEMU/20190117185628.21862-1-crosa@redhat.com/



Hi,

This series seems to have some coding style problems. See output below for
more information:

Subject: [Qemu-devel] [PATCH 00/18] Acceptance Tests: target architecture support
Type: series
Message-id: 20190117185628.21862-1-crosa@redhat.com

=== TEST SCRIPT BEGIN ===
#!/bin/bash
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
./scripts/checkpatch.pl --mailback base..
=== TEST SCRIPT END ===

Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
Switched to a new branch 'test'
a27efa3fec Boot Linux Console Test: add a test for alpha + clipper
9c5d101f9b Boot Linux Console Test: add a test for s390x + s390-ccw-virtio
7d584982d7 Boot Linux Console Test: add a test for arm + virt
039f09b5e5 Boot Linux Console Test: add a test for aarch64 + virt
a52a2dd405 Boot Linux Console Test: add a test for ppc64 + pseries
1995edd0d0 Boot Linux Console Test: add a test for mips64el + malta
cbd4fb29d7 Boot Linux Console Test: add a test for mips + malta
62ca82a551 scripts/qemu.py: support adding a console with the default serial device
8f31626ddd Boot Linux Console Test: refactor the console watcher into utility method
c257e982f1 Boot Linux Console Test: update the x86_64 kernel
9859ac0034 Boot Linux Console Test: rename the x86_64 after the arch and machine
4fa0bc613a Acceptance tests: look for target architecture in test tags first
6051a67e66 Acceptance tests: use "arch:" tag to filter target specific tests
8e86083ca9 Acceptance tests: introduce arch parameter and attribute
6ec8dc9d2e Acceptance tests: fix doc reference to avocado_qemu directory
5c8a9897ba Acceptance tests: improve docstring on pick_default_qemu_bin()
0e98f722c6 Acceptance tests: show avocado test execution by default
7f31c568c2 scripts/qemu.py: log QEMU launch command line

=== OUTPUT BEGIN ===
1/18 Checking commit 7f31c568c2e0 (scripts/qemu.py: log QEMU launch command line)
2/18 Checking commit 0e98f722c669 (Acceptance tests: show avocado test execution by default)
3/18 Checking commit 5c8a9897ba6a (Acceptance tests: improve docstring on pick_default_qemu_bin())
4/18 Checking commit 6ec8dc9d2eed (Acceptance tests: fix doc reference to avocado_qemu directory)
5/18 Checking commit 8e86083ca9e8 (Acceptance tests: introduce arch parameter and attribute)
WARNING: line over 80 characters
#96: FILE: tests/acceptance/avocado_qemu/__init__.py:58:
+                                        default=pick_default_qemu_bin(self.arch))

total: 0 errors, 1 warnings, 64 lines checked

Patch 5/18 has style problems, please review.  If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
6/18 Checking commit 6051a67e66ea (Acceptance tests: use "arch:" tag to filter target specific tests)
7/18 Checking commit 4fa0bc613a2e (Acceptance tests: look for target architecture in test tags first)
8/18 Checking commit 9859ac003400 (Boot Linux Console Test: rename the x86_64 after the arch and machine)
9/18 Checking commit c257e982f129 (Boot Linux Console Test: update the x86_64 kernel)
10/18 Checking commit 8f31626ddd45 (Boot Linux Console Test: refactor the console watcher into utility method)
11/18 Checking commit 62ca82a551c2 (scripts/qemu.py: support adding a console with the default serial device)
12/18 Checking commit cbd4fb29d7c8 (Boot Linux Console Test: add a test for mips + malta)
13/18 Checking commit 1995edd0d009 (Boot Linux Console Test: add a test for mips64el + malta)
ERROR: line over 90 characters
#66: FILE: tests/acceptance/boot_linux_console.py:95:
+        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official

ERROR: line over 90 characters
#67: FILE: tests/acceptance/boot_linux_console.py:96:
+        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48

total: 2 errors, 0 warnings, 58 lines checked

Patch 13/18 has style problems, please review.  If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.

14/18 Checking commit a52a2dd405ae (Boot Linux Console Test: add a test for ppc64 + pseries)
15/18 Checking commit 039f09b5e5a4 (Boot Linux Console Test: add a test for aarch64 + virt)
16/18 Checking commit 7d584982d75c (Boot Linux Console Test: add a test for arm + virt)
17/18 Checking commit 9c5d101f9bfb (Boot Linux Console Test: add a test for s390x + s390-ccw-virtio)
18/18 Checking commit a27efa3fec59 (Boot Linux Console Test: add a test for alpha + clipper)
=== OUTPUT END ===

Test command exited with code: 1


The full log is available at
http://patchew.org/logs/20190117185628.21862-1-crosa@redhat.com/testing.checkpatch/?type=message.
---
Email generated automatically by Patchew [http://patchew.org/].
Please send your feedback to patchew-devel@redhat.com

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Wainer dos Santos Moschetta]

On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
> board and verify the serial is working.
>
> If mips64el is a target being built, "make check-acceptance" will
> automatically include this test by the use of the "arch:mips" tags.

s/arch:mips/arch:mips64el

>
> Alternatively, this test can be run using:
>
>      $ avocado run -t arch:mips64el tests/acceptance
>      $ avocado run -t machine:malta tests/acceptance
>
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>   .travis.yml                            |  2 +-
>   tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>   2 files changed, 40 insertions(+), 1 deletion(-)
>
> diff --git a/.travis.yml b/.travis.yml
> index 49f9016e6a..28648f7a61 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>   
>       # Acceptance (Functional) tests
>       - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>           - TEST_CMD="make check-acceptance"
>         addons:
>           apt:
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index 0678ec91d2..20b845fce1 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -9,8 +9,11 @@
>   # later.  See the COPYING file in the top-level directory.
>   
>   import logging
> +import os
>   
>   from avocado_qemu import Test
> +from avocado.utils import process
> +from avocado.utils import archive
>   
>   
>   class BootLinuxConsole(Test):
> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>           self.vm.launch()
>           console_pattern = 'Kernel command line: %s' % kernel_command_line
>           self.wait_for_console_pattern(console_pattern)
> +
> +    def test_mips64el_malta(self):
> +        """
> +        This test requires the ar tool to extract "data.tar.gz" from
> +        the Debian package.
> +
> +        The kernel can be rebuilt using this Debian kernel source [1] and
> +        following the instructions on [2].
> +
> +        [1] https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
> +        [2] http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48

There references 1 and 2 are inverted.

- Wainer

> +
> +        :avocado: tags=arch:mips64el
> +        :avocado: tags=machine:malta
> +        """
> +        deb_url = ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
> +                   'pool/main/l/linux-2.6/'
> +                   'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
> +
> +        cwd = os.getcwd()
> +        os.chdir(self.workdir)
> +        process.run("ar x %s data.tar.gz" % deb_path)
> +        archive.extract("data.tar.gz", self.workdir)
> +        os.chdir(cwd)
> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
> +
> +        self.vm.set_machine('malta')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=ttyS0 printk.time=0'
> +        self.vm.add_args('-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)

Re: [Qemu-devel] [PATCH 11/18] scripts/qemu.py: support adding a console with the default serial device

[Wainer dos Santos Moschetta]

Hi Cleber,

me again. :)

On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> The set_console() utility function traditionally adds a device either
> based on the explicitly given device type, or based on the machine type,
> a known good type of device.
>
> But, for a number of machine types, it may be impossible or
> inconvenient to add the devices my means of "-device" command line
> options, and then it may better to just use the "-serial" option and
> let QEMU itself, based on the machine type, set the device
> accordingly.
>
> To achieve that, the behavior of set_console() now flags the intention
> to add a console device on launch(), and if no explicit device type is
> given, and there's no definition on CONSOLE_DEV_TYPES, the "-serial"
> is going to be added to the QEMU command line, instead of raising
> exceptions.
>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>   scripts/qemu.py | 28 +++++++++++++++-------------
>   1 file changed, 15 insertions(+), 13 deletions(-)
>
> diff --git a/scripts/qemu.py b/scripts/qemu.py
> index ec3567d4e2..88e1608b42 100644
> --- a/scripts/qemu.py
> +++ b/scripts/qemu.py
> @@ -121,6 +121,7 @@ class QEMUMachine(object):
>           self._temp_dir = None
>           self._launched = False
>           self._machine = None
> +        self._console_set = False
>           self._console_device_type = None
>           self._console_address = None
>           self._console_socket = None
> @@ -240,13 +241,17 @@ class QEMUMachine(object):
>                   '-display', 'none', '-vga', 'none']
>           if self._machine is not None:
>               args.extend(['-machine', self._machine])
> -        if self._console_device_type is not None:
> +        if self._console_set:
>               self._console_address = os.path.join(self._temp_dir,
>                                                    self._name + "-console.sock")
>               chardev = ('socket,id=console,path=%s,server,nowait' %
>                          self._console_address)
> -            device = '%s,chardev=console' % self._console_device_type
> -            args.extend(['-chardev', chardev, '-device', device])
> +            args.extend(['-chardev', chardev])
> +            if self._console_device_type is None:
> +                args.extend(['-serial', 'chardev:console'])
> +            else:
> +                device = '%s,chardev=console' % self._console_device_type
> +                args.extend(['-device', device])
>           return args
>   
>       def _pre_launch(self):
> @@ -479,23 +484,20 @@ class QEMUMachine(object):
>           machine launch time, as it depends on the temporary directory
>           to be created.
>   
> -        @param device_type: the device type, such as "isa-serial"
> +        @param device_type: the device type, such as "isa-serial".  If
> +                            None is given (the default value) a "-serial
> +                            chardev:console" command line argument will
> +                            be used instead, resorting to the machine's
> +                            default device type.

Shouldn't you mention it will look for device type on CONSOLE_DEV_TYPES 
if device_type is None and machine is not None?

The description on set_console()'s docstring is out-of-sync with current 
implementation too.

- Wainer

>           @raises: QEMUMachineAddDeviceError if the device type is not given
>                    and can not be determined.
>           """
> -        if device_type is None:
> -            if self._machine is None:
> -                raise QEMUMachineAddDeviceError("Can not add a console device:"
> -                                                " QEMU instance without a "
> -                                                "defined machine type")
> +        self._console_set = True
> +        if device_type is None and self._machine is not None:
>               for regex, device in CONSOLE_DEV_TYPES.items():
>                   if re.match(regex, self._machine):
>                       device_type = device
>                       break
> -            if device_type is None:
> -                raise QEMUMachineAddDeviceError("Can not add a console device:"
> -                                                " no matching console device "
> -                                                "type definition")
>           self._console_device_type = device_type
>   
>       @property

Re: [Qemu-devel] [PATCH 05/18] Acceptance tests: introduce arch parameter and attribute

[Cleber Rosa]

On 1/31/19 8:55 AM, Wainer dos Santos Moschetta wrote:
> Hi Cleber,
> 
> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
>> It's useful to define the architecture that should be used in
>> situations such as:
>>   * the intended target of the QEMU binary to be used on tests
>>   * the architecture of code to be run within the QEMU binary, such
>>     as a kernel image or a full blown guest OS image
>>
>> This commit introduces both a test parameter and a test instance
>> attribute, that will contain such a value.
>>
>> Now, when the "arch" test parameter is given, it will influence the
>> selection of the default QEMU binary, if one is not given explicitly
>> by means of the "qemu_img" parameter.
>>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>   docs/devel/testing.rst                    | 18 ++++++++++++++++++
>>   tests/acceptance/avocado_qemu/__init__.py | 13 ++++++++++---
>>   2 files changed, 28 insertions(+), 3 deletions(-)
>>
>> diff --git a/docs/devel/testing.rst b/docs/devel/testing.rst
>> index 44c9b3ae74..60a16c5834 100644
>> --- a/docs/devel/testing.rst
>> +++ b/docs/devel/testing.rst
>> @@ -689,6 +689,17 @@ vm
>>   A QEMUMachine instance, initially configured according to the given
>>   ``qemu_bin`` parameter.
>>   +arch
>> +~~~~
>> +
>> +The architecture that will be used on a number of different
>> +scenarios.  For instance, when a QEMU binary is not explicitly given,
>> +the one selected will depend on this attribute.
>> +
>> +The ``arch`` attribute will be set to the test parameter of the same
>> +name, and if one is not given explicitly, it will default to the
>> +system architecture (as given by ``uname``).
> 
> Unless I'm missing something, the arch attribute will be None if I don't
> pass the arch parameter nor set the arch tag.
> 

I think you're not missing anything, but I was.  Thanks for catching that.

- Cleber.

> ------
> [root@xxx qemu]# cat tests/acceptance/arch.py
> from avocado_qemu import Test
> 
> class Arch(Test):
>     """
>     :avocado: enable
>     """
>     def test(self):
>         self.assertEqual(self.arch, "x86_64")
> [root@xxx qemu]# ./tests/venv/bin/avocado run tests/acceptance/arch.py
> JOB ID     : 032aabb8ec2364944d0e29df3b61931b42f4587c
> JOB LOG    : /root/avocado/job-results/job-2019-01-31T08.42-032aabb/job.log
>  (1/1) tests/acceptance/arch.py:Arch.test: FAIL: None != 'x86_64' (0.03 s)
> RESULTS    : PASS 0 | ERROR 0 | FAIL 1 | SKIP 0 | WARN 0 | INTERRUPT 0 |
> CANCEL 0
> JOB TIME   : 0.20 s
> [root@xxx qemu]# ./tests/venv/bin/avocado run -p arch=x86_64
> tests/acceptance/arch.py
> JOB ID     : 1913a73510cab4b6b118bd7e94e2746011368628
> JOB LOG    : /root/avocado/job-results/job-2019-01-31T08.42-1913a73/job.log
>  (1/1) tests/acceptance/arch.py:Arch.test: PASS (0.01 s)
> RESULTS    : PASS 1 | ERROR 0 | FAIL 0 | SKIP 0 | WARN 0 | INTERRUPT 0 |
> CANCEL 0
> JOB TIME   : 0.19 s
> [root@xxx qemu]# cat tests/acceptance/arch_tag.py
> from avocado_qemu import Test
> 
> class Arch(Test):
>     """
>     :avocado: enable
>     :avocado: tags=arch:x86_64
>     """
>     def test(self):
>         self.assertEqual(self.arch, "x86_64")
> [root@xxx qemu]# ./tests/venv/bin/avocado run tests/acceptance/arch_tag.py
> JOB ID     : bc85be065586b8e9cf371489e1801489910e41b0
> JOB LOG    : /root/avocado/job-results/job-2019-01-31T08.43-bc85be0/job.log
>  (1/1) tests/acceptance/arch_tag.py:Arch.test: PASS (0.01 s)
> RESULTS    : PASS 1 | ERROR 0 | FAIL 0 | SKIP 0 | WARN 0 | INTERRUPT 0 |
> CANCEL 0
> JOB TIME   : 0.19 s
> ------
> 
> - Wainer
> 
>> +
>>   qemu_bin
>>   ~~~~~~~~
>>   @@ -711,6 +722,13 @@ like the following:
>>       PARAMS (key=qemu_bin, path=*,
>> default=x86_64-softmmu/qemu-system-x86_64) =>
>> 'x86_64-softmmu/qemu-system-x86_64
>>   +arch
>> +~~~~
>> +
>> +The architecture that will be used on a number of different scenarios.
>> +This parameter has a direct relation with the ``arch`` attribute.  If
>> +not given, it will default to None.
>> +
>>   qemu_bin
>>   ~~~~~~~~
>>   diff --git a/tests/acceptance/avocado_qemu/__init__.py
>> b/tests/acceptance/avocado_qemu/__init__.py
>> index d8d5b48dac..7a38851753 100644
>> --- a/tests/acceptance/avocado_qemu/__init__.py
>> +++ b/tests/acceptance/avocado_qemu/__init__.py
>> @@ -23,16 +23,22 @@ def is_readable_executable_file(path):
>>       return os.path.isfile(path) and os.access(path, os.R_OK | os.X_OK)
>>     -def pick_default_qemu_bin():
>> +def pick_default_qemu_bin(arch=None):
>>       """
>>       Picks the path of a QEMU binary, starting either in the current
>> working
>>       directory or in the source tree root directory.
>>   +    :param arch: the arch to use when looking for a QEMU binary
>> (the target
>> +                 will match the arch given).  If None (the default) arch
>> +                 will be the current host system arch (as given by
>> +                 :func:`os.uname`).
>> +    :type arch: str
>>       :returns: the path to the default QEMU binary or None if one
>> could not
>>                 be found
>>       :rtype: str or None
>>       """
>> -    arch = os.uname()[4]
>> +    if arch is None:
>> +        arch = os.uname()[4]
>>       qemu_bin_relative_path = os.path.join("%s-softmmu" % arch,
>>                                             "qemu-system-%s" % arch)
>>       if is_readable_executable_file(qemu_bin_relative_path):
>> @@ -47,8 +53,9 @@ def pick_default_qemu_bin():
>>   class Test(avocado.Test):
>>       def setUp(self):
>>           self.vm = None
>> +        self.arch = self.params.get('arch')
>>           self.qemu_bin = self.params.get('qemu_bin',
>> -                                        default=pick_default_qemu_bin())
>> +                                       
>> default=pick_default_qemu_bin(self.arch))
>>           if self.qemu_bin is None:
>>               self.cancel("No QEMU binary defined or found in the
>> source tree")
>>           self.vm = QEMUMachine(self.qemu_bin)
> 

-- 
Cleber Rosa
[ Sr Software Engineer - Virtualization Team - Red Hat ]
[ Avocado Test Framework - avocado-framework.github.io ]
[  7ABB 96EB 8B46 B94D 5E0F  E9BB 657E 8D33 A5F2 09F3  ]

Re: [Qemu-devel] [PATCH 10/18] Boot Linux Console Test: refactor the console watcher into utility method

[Cleber Rosa]

On 1/31/19 12:46 PM, Wainer dos Santos Moschetta wrote:
> Hi Cleber,
> 
> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
>> This introduces a utility method that monitors the console device and
>> looks for either a message that signals the test success or failure.
>>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>   tests/acceptance/boot_linux_console.py | 30 ++++++++++++++++++--------
>>   1 file changed, 21 insertions(+), 9 deletions(-)
>>
>> diff --git a/tests/acceptance/boot_linux_console.py
>> b/tests/acceptance/boot_linux_console.py
>> index 35b31162d4..278bb2be3d 100644
>> --- a/tests/acceptance/boot_linux_console.py
>> +++ b/tests/acceptance/boot_linux_console.py
>> @@ -23,6 +23,25 @@ class BootLinuxConsole(Test):
>>         timeout = 60
>>   +    def wait_for_console_pattern(self, success_message,
>> +                                 failure_message='Kernel panic - not
>> syncing'):
>> +        """
>> +        Waits for messages to appear on the console, while logging
>> the content
>> +
>> +        :param success_message: if this message appears, test succeeds
>> +        :param failure_message: if this message appears, test fails
>> +        """
> 
> What's the docstring format used here? I am wondering if :type is optional.
> 
> - Wainer
> 

I've been using the same Sphinx-based docstring statements we've been
using in Avocado, for the sake of cross reference.  So yes, type is
optional, and go on the same line, or on a different one with ":type
<name of parame>".

But, there may be issues that we don't know as we're not building those
API docs.  IIRC Eduardo Habkost is going to propose a GSOC project
related to building API docs.

- Cleber.

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Wainer dos Santos Moschetta]

On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> Just like the previous tests, boots a Linux kernel on a aarch64 target
> using the virt machine.
>
> One special option added is the CPU type, given that the kernel
> selected fails to boot on the virt machine's default CPU (cortex-a15).
>
> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> ---
>   .travis.yml                            |  2 +-
>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>   2 files changed, 21 insertions(+), 1 deletion(-)
>
> diff --git a/.travis.yml b/.travis.yml
> index 54100eea5a..595e8c0b6c 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -187,7 +187,7 @@ matrix:
>   
>       # Acceptance (Functional) tests
>       - env:
> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
>           - TEST_CMD="make check-acceptance"
>         addons:
>           apt:
> diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py
> index f3ccd23a7a..107700b517 100644
> --- a/tests/acceptance/boot_linux_console.py
> +++ b/tests/acceptance/boot_linux_console.py
> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
>           self.vm.launch()
>           console_pattern = 'Kernel command line: %s' % kernel_command_line
>           self.wait_for_console_pattern(console_pattern)
> +
> +    def test_aarch64_virt(self):

That test case fails on my system (Fedora 29 x86_64). Avocado seems 
unable to kill the VM so it  reaches the timeout.

I compiled QEMU with default configuration:

$ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu 
--target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)

Follows a snippet of the Avocado's job.log file:
----
2019-01-31 14:41:34,912 test             L0602 INFO | START 
07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA 
(filename=output.expected) => NOT FOUND (data sources: variant, test, file)
2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch, 
path=*, default=aarch64) => 'aarch64'
2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS 
(key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) => 
'aarch64-softmmu/qemu-system-aarch64'
2019-01-31 14:41:34,915 download         L0070 INFO | Fetching 
https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz 
-> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL 
"https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz": 
content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT", 
last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command: 
'aarch64-softmmu/qemu-system-aarch64 -chardev 
socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon 
chardev=mon,mode=control -display none -vga none -machine virt -chardev 
socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait 
-serial chardev:console -cpu cortex-a53 -kernel 
/var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute': 
'qmp_capabilities'}
2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000] 
Booting Linux on physical CPU 0x0000000000 [0x410fd034]

(...)

2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000] 
Policy zone: DMA32
2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000] 
Kernel command line: console=ttyAMA0
2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute': 
'quit'}
2019-01-31 14:44:35,636 qemu             L0357 WARNI| qemu received 
signal -9: aarch64-softmmu/qemu-system-aarch64 -chardev 
socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon 
chardev=mon,mode=control -display none -vga none -machine virt -chardev 
socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait 
-serial chardev:console -cpu cortex-a53 -kernel 
/var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0

(...)

2019-01-31 14:44:35,663 runner           L0253 ERROR| ERROR Test 
reported status but did not finish -> TestAbortError: 
07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt.
2019-01-31 14:44:35,664 runner           L0062 ERROR| Runner error 
occurred: Timeout reached
----

- Wainer

> +        """
> +        :avocado: tags=arch:aarch64
> +        :avocado: tags=machine:virt
> +        """
> +        kernel_url = ('https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/'
> +                      'releases/29/Server/aarch64/os/images/pxeboot/vmlinuz')
> +        kernel_hash = '8c73e469fc6ea06a58dc83a628fc695b693b8493'
> +        kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash)
> +
> +        self.vm.set_machine('virt')
> +        self.vm.set_console()
> +        kernel_command_line = 'console=ttyAMA0'
> +        self.vm.add_args('-cpu', 'cortex-a53',
> +                         '-kernel', kernel_path,
> +                         '-append', kernel_command_line)
> +        self.vm.launch()
> +        console_pattern = 'Kernel command line: %s' % kernel_command_line
> +        self.wait_for_console_pattern(console_pattern)

Re: [Qemu-devel] [PATCH 11/18] scripts/qemu.py: support adding a console with the default serial device

[Cleber Rosa]

On 1/31/19 1:49 PM, Wainer dos Santos Moschetta wrote:
> Hi Cleber,
> 
> me again. :)
> 
> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
>> The set_console() utility function traditionally adds a device either
>> based on the explicitly given device type, or based on the machine type,
>> a known good type of device.
>>
>> But, for a number of machine types, it may be impossible or
>> inconvenient to add the devices my means of "-device" command line
>> options, and then it may better to just use the "-serial" option and
>> let QEMU itself, based on the machine type, set the device
>> accordingly.
>>
>> To achieve that, the behavior of set_console() now flags the intention
>> to add a console device on launch(), and if no explicit device type is
>> given, and there's no definition on CONSOLE_DEV_TYPES, the "-serial"
>> is going to be added to the QEMU command line, instead of raising
>> exceptions.
>>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>   scripts/qemu.py | 28 +++++++++++++++-------------
>>   1 file changed, 15 insertions(+), 13 deletions(-)
>>
>> diff --git a/scripts/qemu.py b/scripts/qemu.py
>> index ec3567d4e2..88e1608b42 100644
>> --- a/scripts/qemu.py
>> +++ b/scripts/qemu.py
>> @@ -121,6 +121,7 @@ class QEMUMachine(object):
>>           self._temp_dir = None
>>           self._launched = False
>>           self._machine = None
>> +        self._console_set = False
>>           self._console_device_type = None
>>           self._console_address = None
>>           self._console_socket = None
>> @@ -240,13 +241,17 @@ class QEMUMachine(object):
>>                   '-display', 'none', '-vga', 'none']
>>           if self._machine is not None:
>>               args.extend(['-machine', self._machine])
>> -        if self._console_device_type is not None:
>> +        if self._console_set:
>>               self._console_address = os.path.join(self._temp_dir,
>>                                                    self._name +
>> "-console.sock")
>>               chardev = ('socket,id=console,path=%s,server,nowait' %
>>                          self._console_address)
>> -            device = '%s,chardev=console' % self._console_device_type
>> -            args.extend(['-chardev', chardev, '-device', device])
>> +            args.extend(['-chardev', chardev])
>> +            if self._console_device_type is None:
>> +                args.extend(['-serial', 'chardev:console'])
>> +            else:
>> +                device = '%s,chardev=console' %
>> self._console_device_type
>> +                args.extend(['-device', device])
>>           return args
>>         def _pre_launch(self):
>> @@ -479,23 +484,20 @@ class QEMUMachine(object):
>>           machine launch time, as it depends on the temporary directory
>>           to be created.
>>   -        @param device_type: the device type, such as "isa-serial"
>> +        @param device_type: the device type, such as "isa-serial".  If
>> +                            None is given (the default value) a "-serial
>> +                            chardev:console" command line argument will
>> +                            be used instead, resorting to the machine's
>> +                            default device type.
> 
> Shouldn't you mention it will look for device type on CONSOLE_DEV_TYPES
> if device_type is None and machine is not None?
> 

Yes, you're right.  How about this:

...
        @param device_type: the device type, such as "isa-serial".  If
                            None is given (the default value) a "-serial
                            chardev:console" command line argument will
                            be used instead, resorting to the machine's
                            default device type, if a machine type is set,
                            and a matching entry exists on
CONSOLE_DEV_TYPES.

...

> The description on set_console()'s docstring is out-of-sync with current
> implementation too.
> 

Good catch!  I'm rewriting that as:

...

        This is a convenience method that will either use the provided
        device type, of if not given, it will use the device type set
        on CONSOLE_DEV_TYPES if a machine type is set, and a matching
        entry exists on CONSOLE_DEV_TYPES.

...

How does it sound?

Thanks!
- Cleber.

> - Wainer
> 
>>           @raises: QEMUMachineAddDeviceError if the device type is not
>> given
>>                    and can not be determined.
>>           """
>> -        if device_type is None:
>> -            if self._machine is None:
>> -                raise QEMUMachineAddDeviceError("Can not add a
>> console device:"
>> -                                                " QEMU instance
>> without a "
>> -                                                "defined machine type")
>> +        self._console_set = True
>> +        if device_type is None and self._machine is not None:
>>               for regex, device in CONSOLE_DEV_TYPES.items():
>>                   if re.match(regex, self._machine):
>>                       device_type = device
>>                       break
>> -            if device_type is None:
>> -                raise QEMUMachineAddDeviceError("Can not add a
>> console device:"
>> -                                                " no matching console
>> device "
>> -                                                "type definition")
>>           self._console_device_type = device_type
>>         @property
> 
> 

Re: [Qemu-devel] [PATCH 13/18] Boot Linux Console Test: add a test for mips64el + malta

[Cleber Rosa]

On 1/31/19 1:14 PM, Wainer dos Santos Moschetta wrote:
> 
> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>> board and verify the serial is working.
>>
>> If mips64el is a target being built, "make check-acceptance" will
>> automatically include this test by the use of the "arch:mips" tags.
> 
> s/arch:mips/arch:mips64el
> 

Good catch!

>>
>> Alternatively, this test can be run using:
>>
>>      $ avocado run -t arch:mips64el tests/acceptance
>>      $ avocado run -t machine:malta tests/acceptance
>>
>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>   .travis.yml                            |  2 +-
>>   tests/acceptance/boot_linux_console.py | 39 ++++++++++++++++++++++++++
>>   2 files changed, 40 insertions(+), 1 deletion(-)
>>
>> diff --git a/.travis.yml b/.travis.yml
>> index 49f9016e6a..28648f7a61 100644
>> --- a/.travis.yml
>> +++ b/.travis.yml
>> @@ -187,7 +187,7 @@ matrix:
>>         # Acceptance (Functional) tests
>>       - env:
>> -        - CONFIG="--python=/usr/bin/python3
>> --target-list=x86_64-softmmu,mips-softmmu"
>> +        - CONFIG="--python=/usr/bin/python3
>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu"
>>           - TEST_CMD="make check-acceptance"
>>         addons:
>>           apt:
>> diff --git a/tests/acceptance/boot_linux_console.py
>> b/tests/acceptance/boot_linux_console.py
>> index 0678ec91d2..20b845fce1 100644
>> --- a/tests/acceptance/boot_linux_console.py
>> +++ b/tests/acceptance/boot_linux_console.py
>> @@ -9,8 +9,11 @@
>>   # later.  See the COPYING file in the top-level directory.
>>     import logging
>> +import os
>>     from avocado_qemu import Test
>> +from avocado.utils import process
>> +from avocado.utils import archive
>>       class BootLinuxConsole(Test):
>> @@ -80,3 +83,39 @@ class BootLinuxConsole(Test):
>>           self.vm.launch()
>>           console_pattern = 'Kernel command line: %s' %
>> kernel_command_line
>>           self.wait_for_console_pattern(console_pattern)
>> +
>> +    def test_mips64el_malta(self):
>> +        """
>> +        This test requires the ar tool to extract "data.tar.gz" from
>> +        the Debian package.
>> +
>> +        The kernel can be rebuilt using this Debian kernel source [1]
>> and
>> +        following the instructions on [2].
>> +
>> +        [1]
>> https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official
>>
>> +        [2]
>> http://snapshot.debian.org/package/linux-2.6/2.6.32-48/#linux-source-2.6.32_2.6.32-48
>>
> 
> There references 1 and 2 are inverted.
> 

And yet another one good catch!  Your proofreading is awesome!

- Cleber.

> - Wainer
> 
>> +
>> +        :avocado: tags=arch:mips64el
>> +        :avocado: tags=machine:malta
>> +        """
>> +        deb_url =
>> ('http://snapshot.debian.org/archive/debian/20130217T032700Z/'
>> +                   'pool/main/l/linux-2.6/'
>> +                  
>> 'linux-image-2.6.32-5-5kc-malta_2.6.32-48_mipsel.deb')
>> +        deb_hash = '1aaec92083bf22fda31e0d27fa8d9a388e5fc3d5'
>> +        deb_path = self.fetch_asset(deb_url, asset_hash=deb_hash)
>> +
>> +        cwd = os.getcwd()
>> +        os.chdir(self.workdir)
>> +        process.run("ar x %s data.tar.gz" % deb_path)
>> +        archive.extract("data.tar.gz", self.workdir)
>> +        os.chdir(cwd)
>> +        kernel_path = self.workdir + '/boot/vmlinux-2.6.32-5-5kc-malta'
>> +
>> +        self.vm.set_machine('malta')
>> +        self.vm.set_console()
>> +        kernel_command_line = 'console=ttyS0 printk.time=0'
>> +        self.vm.add_args('-kernel', kernel_path,
>> +                         '-append', kernel_command_line)
>> +        self.vm.launch()
>> +        console_pattern = 'Kernel command line: %s' %
>> kernel_command_line
>> +        self.wait_for_console_pattern(console_pattern)
> 

-- 
Cleber Rosa
[ Sr Software Engineer - Virtualization Team - Red Hat ]
[ Avocado Test Framework - avocado-framework.github.io ]
[  7ABB 96EB 8B46 B94D 5E0F  E9BB 657E 8D33 A5F2 09F3  ]

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Cleber Rosa]

On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
> 
> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
>> Just like the previous tests, boots a Linux kernel on a aarch64 target
>> using the virt machine.
>>
>> One special option added is the CPU type, given that the kernel
>> selected fails to boot on the virt machine's default CPU (cortex-a15).
>>
>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>> ---
>>   .travis.yml                            |  2 +-
>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>>   2 files changed, 21 insertions(+), 1 deletion(-)
>>
>> diff --git a/.travis.yml b/.travis.yml
>> index 54100eea5a..595e8c0b6c 100644
>> --- a/.travis.yml
>> +++ b/.travis.yml
>> @@ -187,7 +187,7 @@ matrix:
>>         # Acceptance (Functional) tests
>>       - env:
>> -        - CONFIG="--python=/usr/bin/python3
>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
>> +        - CONFIG="--python=/usr/bin/python3
>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
>>
>>           - TEST_CMD="make check-acceptance"
>>         addons:
>>           apt:
>> diff --git a/tests/acceptance/boot_linux_console.py
>> b/tests/acceptance/boot_linux_console.py
>> index f3ccd23a7a..107700b517 100644
>> --- a/tests/acceptance/boot_linux_console.py
>> +++ b/tests/acceptance/boot_linux_console.py
>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
>>           self.vm.launch()
>>           console_pattern = 'Kernel command line: %s' %
>> kernel_command_line
>>           self.wait_for_console_pattern(console_pattern)
>> +
>> +    def test_aarch64_virt(self):
> 
> That test case fails on my system (Fedora 29 x86_64). Avocado seems
> unable to kill the VM so it  reaches the timeout.
> 
> I compiled QEMU with default configuration:
> 
> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
> 
> 
> Follows a snippet of the Avocado's job.log file:
> ----
> 2019-01-31 14:41:34,912 test             L0602 INFO | START
> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
> 
> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
> path=*, default=aarch64) => 'aarch64'
> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
> 'aarch64-softmmu/qemu-system-aarch64'
> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
> 'aarch64-softmmu/qemu-system-aarch64 -chardev
> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
> chardev=mon,mode=control -display none -vga none -machine virt -chardev
> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
> -serial chardev:console -cpu cortex-a53 -kernel
> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
> 'qmp_capabilities'}
> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
> 
> (...)
> 
> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> Policy zone: DMA32
> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> Kernel command line: console=ttyAMA0
> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
> 'quit'}

Here, a QMP response like "<<< {'return': {}}" would be expected.

Since I can not reproduce this on my system (or on Travis-CI jobs I've
sent), can you tell me on top of which commit you've applied these patches?

Thanks!
- Cleber.

> 2019-01-31 14:44:35,636 qemu             L0357 WARNI| qemu received
> signal -9: aarch64-softmmu/qemu-system-aarch64 -chardev
> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
> chardev=mon,mode=control -display none -vga none -machine virt -chardev
> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
> -serial chardev:console -cpu cortex-a53 -kernel
> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0
> 
> (...)
> 
> 2019-01-31 14:44:35,663 runner           L0253 ERROR| ERROR Test
> reported status but did not finish -> TestAbortError:
> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt.
> 
> 2019-01-31 14:44:35,664 runner           L0062 ERROR| Runner error
> occurred: Timeout reached
> ----
> 
> - Wainer
> 
>> +        """
>> +        :avocado: tags=arch:aarch64
>> +        :avocado: tags=machine:virt
>> +        """
>> +        kernel_url =
>> ('https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/'
>> +                     
>> 'releases/29/Server/aarch64/os/images/pxeboot/vmlinuz')
>> +        kernel_hash = '8c73e469fc6ea06a58dc83a628fc695b693b8493'
>> +        kernel_path = self.fetch_asset(kernel_url,
>> asset_hash=kernel_hash)
>> +
>> +        self.vm.set_machine('virt')
>> +        self.vm.set_console()
>> +        kernel_command_line = 'console=ttyAMA0'
>> +        self.vm.add_args('-cpu', 'cortex-a53',
>> +                         '-kernel', kernel_path,
>> +                         '-append', kernel_command_line)
>> +        self.vm.launch()
>> +        console_pattern = 'Kernel command line: %s' %
>> kernel_command_line
>> +        self.wait_for_console_pattern(console_pattern)
> 

Re: [Qemu-devel] [PATCH 18/18] Boot Linux Console Test: add a test for alpha + clipper

[Cleber Rosa]

On 1/31/19 5:30 AM, Philippe Mathieu-Daudé wrote:
> On 1/31/19 3:53 AM, Cleber Rosa wrote:
>> On 1/22/19 5:52 AM, Philippe Mathieu-Daudé wrote:
>>> On 1/17/19 7:56 PM, Cleber Rosa wrote:
>>>> Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
>>>> board and verify the serial is working.  One extra command added to
>>>> the QEMU command line is '-vga std', because the kernel used is
>>>> known to crash without it.
>>>>
>>>> If alpha is a target being built, "make check-acceptance" will
>>>> automatically include this test by the use of the "arch:alpha" tags.
>>>>
>>>> Alternatively, this test can be run using:
>>>>
>>>>     $ avocado run -t arch:alpha tests/acceptance
>>>>     $ avocado run -t machine:clipper tests/acceptance
>>>>
>>>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>>>> ---
>>>>  .travis.yml                            |  2 +-
>>>>  scripts/qemu.py                        |  1 -
>>>>  tests/acceptance/boot_linux_console.py | 22 ++++++++++++++++++++++
>>>>  3 files changed, 23 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/.travis.yml b/.travis.yml
>>>> index 0d5a4b104b..73a113af87 100644
>>>> --- a/.travis.yml
>>>> +++ b/.travis.yml
>>>> @@ -187,7 +187,7 @@ matrix:
>>>>  
>>>>      # Acceptance (Functional) tests
>>>>      - env:
>>>> -        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu"
>>>> +        - CONFIG="--python=/usr/bin/python3 --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu,arm-softmmu,s390x-softmmu,alpha-softmmu"
>>>>          - TEST_CMD="make check-acceptance"
>>>>        addons:
>>>>          apt:
>>>> diff --git a/scripts/qemu.py b/scripts/qemu.py
>>>> index 1531e28fc1..a704da418a 100644
>>>> --- a/scripts/qemu.py
>>>> +++ b/scripts/qemu.py
>>>> @@ -34,7 +34,6 @@ def kvm_available(target_arch=None):
>>>>  
>>>>  #: Maps machine types to the preferred console device types
>>>>  CONSOLE_DEV_TYPES = {
>>>> -    r'^clipper$': 'isa-serial',
>>>
>>> Why this change?
>>>
>>
>> Because we've come to conclusion (I believe :) that's better, whenever
>> possible, to let QEMU pick the default device type by machine.  The
>> discussion happened here:
>>
>> https://lists.gnu.org/archive/html/qemu-devel/2018-10/msg04582.html
> 
> I was expecting this change to be global (in a single patch previous to
> the target-specific patches). If you prefer to do it a target at a time
> I'm OK with it but you should add a comment about it in the commit IMHO.
> 
> Regards,
> 
> Phil.
> 

You're right, that's a better approach.  I've made this change part of
the set_console() changes.

Thanks!
- Cleber.

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Cleber Rosa]

On 1/31/19 3:21 PM, Cleber Rosa wrote:
> 
> 
> On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
>>
>> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
>>> Just like the previous tests, boots a Linux kernel on a aarch64 target
>>> using the virt machine.
>>>
>>> One special option added is the CPU type, given that the kernel
>>> selected fails to boot on the virt machine's default CPU (cortex-a15).
>>>
>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>>> ---
>>>   .travis.yml                            |  2 +-
>>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>>>   2 files changed, 21 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/.travis.yml b/.travis.yml
>>> index 54100eea5a..595e8c0b6c 100644
>>> --- a/.travis.yml
>>> +++ b/.travis.yml
>>> @@ -187,7 +187,7 @@ matrix:
>>>         # Acceptance (Functional) tests
>>>       - env:
>>> -        - CONFIG="--python=/usr/bin/python3
>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
>>> +        - CONFIG="--python=/usr/bin/python3
>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
>>>
>>>           - TEST_CMD="make check-acceptance"
>>>         addons:
>>>           apt:
>>> diff --git a/tests/acceptance/boot_linux_console.py
>>> b/tests/acceptance/boot_linux_console.py
>>> index f3ccd23a7a..107700b517 100644
>>> --- a/tests/acceptance/boot_linux_console.py
>>> +++ b/tests/acceptance/boot_linux_console.py
>>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
>>>           self.vm.launch()
>>>           console_pattern = 'Kernel command line: %s' %
>>> kernel_command_line
>>>           self.wait_for_console_pattern(console_pattern)
>>> +
>>> +    def test_aarch64_virt(self):
>>
>> That test case fails on my system (Fedora 29 x86_64). Avocado seems
>> unable to kill the VM so it  reaches the timeout.
>>
>> I compiled QEMU with default configuration:
>>
>> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
>>
>>
>> Follows a snippet of the Avocado's job.log file:
>> ----
>> 2019-01-31 14:41:34,912 test             L0602 INFO | START
>> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
>>
>> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
>> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
>> path=*, default=aarch64) => 'aarch64'
>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
>> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
>> 'aarch64-softmmu/qemu-system-aarch64'
>> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
>> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
>> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
>> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
>> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
>> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
>> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
>> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
>> 'aarch64-softmmu/qemu-system-aarch64 -chardev
>> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
>> chardev=mon,mode=control -display none -vga none -machine virt -chardev
>> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
>> -serial chardev:console -cpu cortex-a53 -kernel
>> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
>> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
>> 'qmp_capabilities'}
>> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
>> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
>> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
>>
>> (...)
>>
>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
>> Policy zone: DMA32
>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
>> Kernel command line: console=ttyAMA0
>> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
>> 'quit'}
> 
> Here, a QMP response like "<<< {'return': {}}" would be expected.
> 
> Since I can not reproduce this on my system (or on Travis-CI jobs I've
> sent), can you tell me on top of which commit you've applied these patches?
> 

I spoke too soon:

https://travis-ci.org/clebergnu/qemu/jobs/487121425#L3033

This looks like a recent regression, and I'm guessing it's not on the
test's side.  I'll try to bisect it and let you know.

Thanks,
- Cleber.

Re: [Qemu-devel] [PATCH 00/18] Acceptance Tests: target architecture support

[Aleksandar Markovic]

>> I think, the term "arch" is a little problematic in QEMU parlance. IMHO,
>> "target" should be used instead. ("arch" is used in Linux kernel community)

> Naming things is hard, so this is a valid discussion.  But, I have to
> say that I also find "arch" in this context to be descriptive enough.

This is not a "naming" problem, but a fundamental terminology convention used in QEMU. I am truly dissapointed that you chose to disrespect it.

Aleksandar

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Cleber Rosa]

On 1/31/19 4:26 PM, Cleber Rosa wrote:
> 
> 
> On 1/31/19 3:21 PM, Cleber Rosa wrote:
>>
>>
>> On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
>>>
>>> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
>>>> Just like the previous tests, boots a Linux kernel on a aarch64 target
>>>> using the virt machine.
>>>>
>>>> One special option added is the CPU type, given that the kernel
>>>> selected fails to boot on the virt machine's default CPU (cortex-a15).
>>>>
>>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
>>>> ---
>>>>   .travis.yml                            |  2 +-
>>>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
>>>>   2 files changed, 21 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/.travis.yml b/.travis.yml
>>>> index 54100eea5a..595e8c0b6c 100644
>>>> --- a/.travis.yml
>>>> +++ b/.travis.yml
>>>> @@ -187,7 +187,7 @@ matrix:
>>>>         # Acceptance (Functional) tests
>>>>       - env:
>>>> -        - CONFIG="--python=/usr/bin/python3
>>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
>>>> +        - CONFIG="--python=/usr/bin/python3
>>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
>>>>
>>>>           - TEST_CMD="make check-acceptance"
>>>>         addons:
>>>>           apt:
>>>> diff --git a/tests/acceptance/boot_linux_console.py
>>>> b/tests/acceptance/boot_linux_console.py
>>>> index f3ccd23a7a..107700b517 100644
>>>> --- a/tests/acceptance/boot_linux_console.py
>>>> +++ b/tests/acceptance/boot_linux_console.py
>>>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
>>>>           self.vm.launch()
>>>>           console_pattern = 'Kernel command line: %s' %
>>>> kernel_command_line
>>>>           self.wait_for_console_pattern(console_pattern)
>>>> +
>>>> +    def test_aarch64_virt(self):
>>>
>>> That test case fails on my system (Fedora 29 x86_64). Avocado seems
>>> unable to kill the VM so it  reaches the timeout.
>>>
>>> I compiled QEMU with default configuration:
>>>
>>> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
>>>
>>>
>>> Follows a snippet of the Avocado's job.log file:
>>> ----
>>> 2019-01-31 14:41:34,912 test             L0602 INFO | START
>>> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
>>>
>>> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
>>> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
>>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
>>> path=*, default=aarch64) => 'aarch64'
>>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
>>> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
>>> 'aarch64-softmmu/qemu-system-aarch64'
>>> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
>>> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
>>> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
>>> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
>>> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
>>> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
>>> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
>>> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
>>> 'aarch64-softmmu/qemu-system-aarch64 -chardev
>>> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
>>> chardev=mon,mode=control -display none -vga none -machine virt -chardev
>>> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
>>> -serial chardev:console -cpu cortex-a53 -kernel
>>> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
>>> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
>>> 'qmp_capabilities'}
>>> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
>>> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
>>> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
>>>
>>> (...)
>>>
>>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
>>> Policy zone: DMA32
>>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
>>> Kernel command line: console=ttyAMA0
>>> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
>>> 'quit'}
>>
>> Here, a QMP response like "<<< {'return': {}}" would be expected.
>>
>> Since I can not reproduce this on my system (or on Travis-CI jobs I've
>> sent), can you tell me on top of which commit you've applied these patches?
>>
> 
> I spoke too soon:
> 
> https://travis-ci.org/clebergnu/qemu/jobs/487121425#L3033
> 
> This looks like a recent regression, and I'm guessing it's not on the
> test's side.  I'll try to bisect it and let you know.
> 

On a fresh environment, I am able to get this reproduced on every 2 of
runs, more or less.  When I hit it, I attached GDB to it, and the
backtrace shows:

Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
warning: Loadable section ".note.gnu.property" outside of ELF segments
warning: Loadable section ".note.gnu.property" outside of ELF segments
__lll_lock_wait () at ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
103     2:      movl    %edx, %eax
(gdb) bt
#0  __lll_lock_wait () at
../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
#1  0x00007fc6ba1a2e09 in __GI___pthread_mutex_lock
(mutex=mutex@entry=0x5615a233d020 <qemu_global_mutex>) at
../nptl/pthread_mutex_lock.c:80
#2  0x00005615a1bb7593 in qemu_mutex_lock_impl (mutex=0x5615a233d020
<qemu_global_mutex>, file=0x5615a1db2d4c "util/main-loop.c", line=236)
at util/qemu-thread-posix.c:66
#3  0x00005615a171125e in qemu_mutex_lock_iothread_impl
(file=file@entry=0x5615a1db2d4c "util/main-loop.c", line=line@entry=236)
at /home/cleber/src/qemu/cpus.c:1849
#4  0x00005615a1bb415d in os_host_main_loop_wait (timeout=<optimized
out>) at util/main-loop.c:236
#5  main_loop_wait (nonblocking=<optimized out>) at util/main-loop.c:497
#6  0x00005615a18fdd39 in main_loop () at vl.c:1928
#7  0x00005615a16c9ee9 in main (argc=<optimized out>, argv=<optimized
out>, envp=<optimized out>) at vl.c:4665

Running it with `taskset -c 1` prevents this issue from happening, which
AFAICT, contributes even further towards this being a QEMU race condition.

I'm CC'ing Peter and Claudio (listed maintainers of aarch64), as this
seems to limited to that target.  Any tips on what to do here?

Thanks,
- Cleber.

> Thanks,
> - Cleber.
> 

Re: [Qemu-devel] [PATCH 00/18] Acceptance Tests: target architecture support

[Cleber Rosa]

On 2/1/19 12:32 AM, Aleksandar Markovic wrote:
>>> I think, the term "arch" is a little problematic in QEMU parlance. IMHO,
>>> "target" should be used instead. ("arch" is used in Linux kernel community)
> 
>> Naming things is hard, so this is a valid discussion.  But, I have to
>> say that I also find "arch" in this context to be descriptive enough.
> 
> This is not a "naming" problem, but a fundamental terminology convention used in QEMU. I am truly dissapointed that you chose to disrespect it.
> 
> Aleksandar
> 

Hi Aleksandar,

I'm basically listening to more experienced people on the matter, as I
clearly lack the background on the "terminology conventions", and thus,
I'm unable to make the best naming choices by myself.

Now, we have had participation from few people, and there doesn't seem
to be consensus.  So, to make it clear, I didn't *choose* to disrespect
it, I chose to respect how I perceived the community input.

If this makes you truly disappointed, please ask more people to give
their input, and I'll gladly steer the direction of this humble ship, I
mean, patch series. :)

Regards,
- Cleber.

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Eduardo Habkost]

On Fri, Feb 01, 2019 at 11:10:31AM -0500, Cleber Rosa wrote:
> 
> 
> On 1/31/19 4:26 PM, Cleber Rosa wrote:
> > 
> > 
> > On 1/31/19 3:21 PM, Cleber Rosa wrote:
> >>
> >>
> >> On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
> >>>
> >>> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> >>>> Just like the previous tests, boots a Linux kernel on a aarch64 target
> >>>> using the virt machine.
> >>>>
> >>>> One special option added is the CPU type, given that the kernel
> >>>> selected fails to boot on the virt machine's default CPU (cortex-a15).
> >>>>
> >>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> >>>> ---
> >>>>   .travis.yml                            |  2 +-
> >>>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
> >>>>   2 files changed, 21 insertions(+), 1 deletion(-)
> >>>>
> >>>> diff --git a/.travis.yml b/.travis.yml
> >>>> index 54100eea5a..595e8c0b6c 100644
> >>>> --- a/.travis.yml
> >>>> +++ b/.travis.yml
> >>>> @@ -187,7 +187,7 @@ matrix:
> >>>>         # Acceptance (Functional) tests
> >>>>       - env:
> >>>> -        - CONFIG="--python=/usr/bin/python3
> >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
> >>>> +        - CONFIG="--python=/usr/bin/python3
> >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
> >>>>
> >>>>           - TEST_CMD="make check-acceptance"
> >>>>         addons:
> >>>>           apt:
> >>>> diff --git a/tests/acceptance/boot_linux_console.py
> >>>> b/tests/acceptance/boot_linux_console.py
> >>>> index f3ccd23a7a..107700b517 100644
> >>>> --- a/tests/acceptance/boot_linux_console.py
> >>>> +++ b/tests/acceptance/boot_linux_console.py
> >>>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
> >>>>           self.vm.launch()
> >>>>           console_pattern = 'Kernel command line: %s' %
> >>>> kernel_command_line
> >>>>           self.wait_for_console_pattern(console_pattern)
> >>>> +
> >>>> +    def test_aarch64_virt(self):
> >>>
> >>> That test case fails on my system (Fedora 29 x86_64). Avocado seems
> >>> unable to kill the VM so it  reaches the timeout.
> >>>
> >>> I compiled QEMU with default configuration:
> >>>
> >>> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
> >>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
> >>>
> >>>
> >>> Follows a snippet of the Avocado's job.log file:
> >>> ----
> >>> 2019-01-31 14:41:34,912 test             L0602 INFO | START
> >>> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
> >>>
> >>> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
> >>> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
> >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
> >>> path=*, default=aarch64) => 'aarch64'
> >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
> >>> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
> >>> 'aarch64-softmmu/qemu-system-aarch64'
> >>> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
> >>> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
> >>> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
> >>> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
> >>> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
> >>> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
> >>> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
> >>> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
> >>> 'aarch64-softmmu/qemu-system-aarch64 -chardev
> >>> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
> >>> chardev=mon,mode=control -display none -vga none -machine virt -chardev
> >>> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
> >>> -serial chardev:console -cpu cortex-a53 -kernel
> >>> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
> >>> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
> >>> 'qmp_capabilities'}
> >>> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
> >>> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
> >>> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
> >>>
> >>> (...)
> >>>
> >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> >>> Policy zone: DMA32
> >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> >>> Kernel command line: console=ttyAMA0
> >>> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
> >>> 'quit'}
> >>
> >> Here, a QMP response like "<<< {'return': {}}" would be expected.
> >>
> >> Since I can not reproduce this on my system (or on Travis-CI jobs I've
> >> sent), can you tell me on top of which commit you've applied these patches?
> >>
> > 
> > I spoke too soon:
> > 
> > https://travis-ci.org/clebergnu/qemu/jobs/487121425#L3033
> > 
> > This looks like a recent regression, and I'm guessing it's not on the
> > test's side.  I'll try to bisect it and let you know.
> > 
> 
> On a fresh environment, I am able to get this reproduced on every 2 of
> runs, more or less.  When I hit it, I attached GDB to it, and the
> backtrace shows:
> 
> Thread debugging using libthread_db enabled]
> Using host libthread_db library "/lib64/libthread_db.so.1".
> warning: Loadable section ".note.gnu.property" outside of ELF segments
> warning: Loadable section ".note.gnu.property" outside of ELF segments
> __lll_lock_wait () at ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> 103     2:      movl    %edx, %eax
> (gdb) bt
> #0  __lll_lock_wait () at
> ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> #1  0x00007fc6ba1a2e09 in __GI___pthread_mutex_lock
> (mutex=mutex@entry=0x5615a233d020 <qemu_global_mutex>) at
> ../nptl/pthread_mutex_lock.c:80
> #2  0x00005615a1bb7593 in qemu_mutex_lock_impl (mutex=0x5615a233d020
> <qemu_global_mutex>, file=0x5615a1db2d4c "util/main-loop.c", line=236)
> at util/qemu-thread-posix.c:66
> #3  0x00005615a171125e in qemu_mutex_lock_iothread_impl
> (file=file@entry=0x5615a1db2d4c "util/main-loop.c", line=line@entry=236)
> at /home/cleber/src/qemu/cpus.c:1849
> #4  0x00005615a1bb415d in os_host_main_loop_wait (timeout=<optimized
> out>) at util/main-loop.c:236
> #5  main_loop_wait (nonblocking=<optimized out>) at util/main-loop.c:497
> #6  0x00005615a18fdd39 in main_loop () at vl.c:1928
> #7  0x00005615a16c9ee9 in main (argc=<optimized out>, argv=<optimized
> out>, envp=<optimized out>) at vl.c:4665

Tip: run "thread apply all bt" so you can get a backtrace of all
threads.


> 
> Running it with `taskset -c 1` prevents this issue from happening, which
> AFAICT, contributes even further towards this being a QEMU race condition.
> 
> I'm CC'ing Peter and Claudio (listed maintainers of aarch64), as this
> seems to limited to that target.  Any tips on what to do here?

I am hitting this on Travis, too, and I finally could reproduce
it locally,

The guest is still writing on the serial console, but nobody is
reading the data on the other side.  A VCPU thread is stuck
inside the EAGAIN/nanosleep loop at qemu_chr_write_buffer(),
holding the QEMU global lock.

Thread 4 (Thread 0x7f2e45fff700 (LWP 6461)):
#0  0x00007f2e4ec03500 in nanosleep () at /lib64/libpthread.so.0
#1  0x00007f2e4fb229d7 in g_usleep () at /lib64/libglib-2.0.so.0
#2  0x0000559a4e7ca4c9 in qemu_chr_write_buffer (s=s@entry=0x559a502d0ac0, buf=buf@entry=0x7f2e45ffdd90 "7", len=1, offset=offset@entry=0x7f2e45ffdd60, write_all=true) at chardev/char.c:115
#3  0x0000559a4e7ca78f in qemu_chr_write (s=0x559a502d0ac0, buf=buf@entry=0x7f2e45ffdd90 "7", len=len@entry=1, write_all=write_all@entry=true) at chardev/char.c:148
#4  0x0000559a4e7cc7e2 in qemu_chr_fe_write_all (be=be@entry=0x559a504b4c50, buf=buf@entry=0x7f2e45ffdd90 "7", len=len@entry=1) at chardev/char-fe.c:53
#5  0x0000559a4e58f320 in pl011_write (opaque=0x559a504b47d0, offset=0, value=55, size=<optimized out>) at hw/char/pl011.c:183
#6  0x0000559a4e325121 in memory_region_write_accessor (mr=0x559a504b4ae0, addr=0, value=<optimized out>, size=2, shift=<optimized out>, mask=<optimized out>, attrs=...) at /home/ehabkost/rh/proj/virt/qemu/memory.c:503
#7  0x0000559a4e322cd6 in access_with_adjusted_size (addr=addr@entry=0, value=value@entry=0x7f2e45ffded8, size=size@entry=2, access_size_min=<optimized out>, access_size_max=<optimized out>, access_fn=access_fn@entry=
    0x559a4e3250a0 <memory_region_write_accessor>, mr=0x559a504b4ae0, attrs=...) at /home/ehabkost/rh/proj/virt/qemu/memory.c:569
#8  0x0000559a4e32763f in memory_region_dispatch_write (mr=mr@entry=0x559a504b4ae0, addr=addr@entry=0, data=<optimized out>, data@entry=55, size=size@entry=2, attrs=...) at /home/ehabkost/rh/proj/virt/qemu/memory.c:1497
#9  0x0000559a4e338708 in io_writex (env=env@entry=0x559a503d5620, mmu_idx=mmu_idx@entry=1, val=val@entry=55, addr=addr@entry=18446462598867529728, retaddr=139836732143069, size=2, iotlbentry=<optimized out>, iotlbentry=<optimized out>)
    at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cputlb.c:945
#10 0x0000559a4e33d203 in store_helper (big_endian=false, size=2, retaddr=<optimized out>, oi=<optimized out>, val=55, addr=18446462598867529728, env=0x559a503d5620) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cputlb.c:1544
#11 0x0000559a4e33d203 in helper_le_stw_mmu (env=0x559a503d5620, addr=18446462598867529728, val=55, oi=<optimized out>, retaddr=139836732143069) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cputlb.c:1636
#12 0x00007f2e46bef5dd in code_gen_buffer ()
#13 0x0000559a4e352381 in cpu_tb_exec (itb=<optimized out>, cpu=0x559a503d5620) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cpu-exec.c:171
#14 0x0000559a4e352381 in cpu_loop_exec_tb (tb_exit=<synthetic pointer>, last_tb=<synthetic pointer>, tb=<optimized out>, cpu=0x559a503d5620) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cpu-exec.c:618
#15 0x0000559a4e352381 in cpu_exec (cpu=cpu@entry=0x559a503cd360) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cpu-exec.c:729
#16 0x0000559a4e30ea0f in tcg_cpu_exec (cpu=0x559a503cd360) at /home/ehabkost/rh/proj/virt/qemu/cpus.c:1434
#17 0x0000559a4e310b6b in qemu_tcg_cpu_thread_fn (arg=arg@entry=0x559a503cd360) at /home/ehabkost/rh/proj/virt/qemu/cpus.c:1743
#18 0x0000559a4e83669a in qemu_thread_start (args=<optimized out>) at util/qemu-thread-posix.c:502
#19 0x00007f2e4ebf958e in start_thread () at /lib64/libpthread.so.0
#20 0x00007f2e4eb266f3 in clone () at /lib64/libc.so.6


For reference, this is the QEMU command line:

aarch64-softmmu/qemu-system-aarch64 -chardev socket,id=mon,path=/var/tmp/tmpxnkcjvf0/qemu-6453-monitor.sock -mon chardev=mon,mode=control -display none -vga none -machine virt -chardev socket,id=console,path=/var/tmp/tmpxnkcjvf0/qemu-6453-console.sock,server,nowait -serial chardev:console -cpu cortex-a53 -kernel /home/ehabkost/rh/proj/virt/qemu/tests/venv/data/cache/by_location/e959d0e1dd72e77653e218e666198db1f3d0c213/vmlinuz -append printk.time=0 console=ttyAMA0



-- 
Eduardo

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Eduardo Habkost]

On Fri, Jun 07, 2019 at 12:26:48AM -0300, Eduardo Habkost wrote:
> On Fri, Feb 01, 2019 at 11:10:31AM -0500, Cleber Rosa wrote:
> > 
> > 
> > On 1/31/19 4:26 PM, Cleber Rosa wrote:
> > > 
> > > 
> > > On 1/31/19 3:21 PM, Cleber Rosa wrote:
> > >>
> > >>
> > >> On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
> > >>>
> > >>> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> > >>>> Just like the previous tests, boots a Linux kernel on a aarch64 target
> > >>>> using the virt machine.
> > >>>>
> > >>>> One special option added is the CPU type, given that the kernel
> > >>>> selected fails to boot on the virt machine's default CPU (cortex-a15).
> > >>>>
> > >>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> > >>>> ---
> > >>>>   .travis.yml                            |  2 +-
> > >>>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
> > >>>>   2 files changed, 21 insertions(+), 1 deletion(-)
> > >>>>
> > >>>> diff --git a/.travis.yml b/.travis.yml
> > >>>> index 54100eea5a..595e8c0b6c 100644
> > >>>> --- a/.travis.yml
> > >>>> +++ b/.travis.yml
> > >>>> @@ -187,7 +187,7 @@ matrix:
> > >>>>         # Acceptance (Functional) tests
> > >>>>       - env:
> > >>>> -        - CONFIG="--python=/usr/bin/python3
> > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
> > >>>> +        - CONFIG="--python=/usr/bin/python3
> > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
> > >>>>
> > >>>>           - TEST_CMD="make check-acceptance"
> > >>>>         addons:
> > >>>>           apt:
> > >>>> diff --git a/tests/acceptance/boot_linux_console.py
> > >>>> b/tests/acceptance/boot_linux_console.py
> > >>>> index f3ccd23a7a..107700b517 100644
> > >>>> --- a/tests/acceptance/boot_linux_console.py
> > >>>> +++ b/tests/acceptance/boot_linux_console.py
> > >>>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
> > >>>>           self.vm.launch()
> > >>>>           console_pattern = 'Kernel command line: %s' %
> > >>>> kernel_command_line
> > >>>>           self.wait_for_console_pattern(console_pattern)
> > >>>> +
> > >>>> +    def test_aarch64_virt(self):
> > >>>
> > >>> That test case fails on my system (Fedora 29 x86_64). Avocado seems
> > >>> unable to kill the VM so it  reaches the timeout.
> > >>>
> > >>> I compiled QEMU with default configuration:
> > >>>
> > >>> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
> > >>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
> > >>>
> > >>>
> > >>> Follows a snippet of the Avocado's job.log file:
> > >>> ----
> > >>> 2019-01-31 14:41:34,912 test             L0602 INFO | START
> > >>> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
> > >>>
> > >>> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
> > >>> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
> > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
> > >>> path=*, default=aarch64) => 'aarch64'
> > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
> > >>> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
> > >>> 'aarch64-softmmu/qemu-system-aarch64'
> > >>> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
> > >>> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
> > >>> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
> > >>> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
> > >>> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
> > >>> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
> > >>> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
> > >>> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
> > >>> 'aarch64-softmmu/qemu-system-aarch64 -chardev
> > >>> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
> > >>> chardev=mon,mode=control -display none -vga none -machine virt -chardev
> > >>> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
> > >>> -serial chardev:console -cpu cortex-a53 -kernel
> > >>> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
> > >>> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
> > >>> 'qmp_capabilities'}
> > >>> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
> > >>> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
> > >>> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
> > >>>
> > >>> (...)
> > >>>
> > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > >>> Policy zone: DMA32
> > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > >>> Kernel command line: console=ttyAMA0
> > >>> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
> > >>> 'quit'}
> > >>
> > >> Here, a QMP response like "<<< {'return': {}}" would be expected.
> > >>
> > >> Since I can not reproduce this on my system (or on Travis-CI jobs I've
> > >> sent), can you tell me on top of which commit you've applied these patches?
> > >>
> > > 
> > > I spoke too soon:
> > > 
> > > https://travis-ci.org/clebergnu/qemu/jobs/487121425#L3033
> > > 
> > > This looks like a recent regression, and I'm guessing it's not on the
> > > test's side.  I'll try to bisect it and let you know.
> > > 
> > 
> > On a fresh environment, I am able to get this reproduced on every 2 of
> > runs, more or less.  When I hit it, I attached GDB to it, and the
> > backtrace shows:
> > 
> > Thread debugging using libthread_db enabled]
> > Using host libthread_db library "/lib64/libthread_db.so.1".
> > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > __lll_lock_wait () at ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > 103     2:      movl    %edx, %eax
> > (gdb) bt
> > #0  __lll_lock_wait () at
> > ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > #1  0x00007fc6ba1a2e09 in __GI___pthread_mutex_lock
> > (mutex=mutex@entry=0x5615a233d020 <qemu_global_mutex>) at
> > ../nptl/pthread_mutex_lock.c:80
> > #2  0x00005615a1bb7593 in qemu_mutex_lock_impl (mutex=0x5615a233d020
> > <qemu_global_mutex>, file=0x5615a1db2d4c "util/main-loop.c", line=236)
> > at util/qemu-thread-posix.c:66
> > #3  0x00005615a171125e in qemu_mutex_lock_iothread_impl
> > (file=file@entry=0x5615a1db2d4c "util/main-loop.c", line=line@entry=236)
> > at /home/cleber/src/qemu/cpus.c:1849
> > #4  0x00005615a1bb415d in os_host_main_loop_wait (timeout=<optimized
> > out>) at util/main-loop.c:236
> > #5  main_loop_wait (nonblocking=<optimized out>) at util/main-loop.c:497
> > #6  0x00005615a18fdd39 in main_loop () at vl.c:1928
> > #7  0x00005615a16c9ee9 in main (argc=<optimized out>, argv=<optimized
> > out>, envp=<optimized out>) at vl.c:4665
> 
> Tip: run "thread apply all bt" so you can get a backtrace of all
> threads.
> 
> 
> > 
> > Running it with `taskset -c 1` prevents this issue from happening, which
> > AFAICT, contributes even further towards this being a QEMU race condition.
> > 
> > I'm CC'ing Peter and Claudio (listed maintainers of aarch64), as this
> > seems to limited to that target.  Any tips on what to do here?
> 
> I am hitting this on Travis, too, and I finally could reproduce
> it locally,
> 
> The guest is still writing on the serial console, but nobody is
> reading the data on the other side.  A VCPU thread is stuck
> inside the EAGAIN/nanosleep loop at qemu_chr_write_buffer(),
> holding the QEMU global lock.

Experimental fix below.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
---
 python/qemu/__init__.py | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/python/qemu/__init__.py b/python/qemu/__init__.py
index 81d9657ec0..4a691f34da 100644
--- a/python/qemu/__init__.py
+++ b/python/qemu/__init__.py
@@ -274,10 +274,6 @@ class QEMUMachine(object):
 
         self._qemu_log_path = None
 
-        if self._console_socket is not None:
-            self._console_socket.close()
-            self._console_socket = None
-
         if self._temp_dir is not None:
             shutil.rmtree(self._temp_dir)
             self._temp_dir = None
@@ -336,6 +332,14 @@ class QEMUMachine(object):
         """
         Terminate the VM and clean up
         """
+
+        # If we keep the console socket open, we may deadlock waiting
+        # for QEMU to exit, while QEMU is waiting for the socket to
+        # become writeable.
+        if self._console_socket is not None:
+            self._console_socket.close()
+            self._console_socket = None
+
         if self.is_running():
             try:
                 self._qmp.cmd('quit')
-- 
2.18.0.rc1.1.g3f1ff2140

-- 
Eduardo

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Laszlo Ersek]

On 06/07/19 05:26, Eduardo Habkost wrote:

> Thread 4 (Thread 0x7f2e45fff700 (LWP 6461)):
> #0  0x00007f2e4ec03500 in nanosleep () at /lib64/libpthread.so.0
> #1  0x00007f2e4fb229d7 in g_usleep () at /lib64/libglib-2.0.so.0
> #2  0x0000559a4e7ca4c9 in qemu_chr_write_buffer (s=s@entry=0x559a502d0ac0, buf=buf@entry=0x7f2e45ffdd90 "7", len=1, offset=offset@entry=0x7f2e45ffdd60, write_all=true) at chardev/char.c:115
> #3  0x0000559a4e7ca78f in qemu_chr_write (s=0x559a502d0ac0, buf=buf@entry=0x7f2e45ffdd90 "7", len=len@entry=1, write_all=write_all@entry=true) at chardev/char.c:148
> #4  0x0000559a4e7cc7e2 in qemu_chr_fe_write_all (be=be@entry=0x559a504b4c50, buf=buf@entry=0x7f2e45ffdd90 "7", len=len@entry=1) at chardev/char-fe.c:53
> #5  0x0000559a4e58f320 in pl011_write (opaque=0x559a504b47d0, offset=0, value=55, size=<optimized out>) at hw/char/pl011.c:183
> #6  0x0000559a4e325121 in memory_region_write_accessor (mr=0x559a504b4ae0, addr=0, value=<optimized out>, size=2, shift=<optimized out>, mask=<optimized out>, attrs=...) at /home/ehabkost/rh/proj/virt/qemu/memory.c:503
> #7  0x0000559a4e322cd6 in access_with_adjusted_size (addr=addr@entry=0, value=value@entry=0x7f2e45ffded8, size=size@entry=2, access_size_min=<optimized out>, access_size_max=<optimized out>, access_fn=access_fn@entry=
>     0x559a4e3250a0 <memory_region_write_accessor>, mr=0x559a504b4ae0, attrs=...) at /home/ehabkost/rh/proj/virt/qemu/memory.c:569
> #8  0x0000559a4e32763f in memory_region_dispatch_write (mr=mr@entry=0x559a504b4ae0, addr=addr@entry=0, data=<optimized out>, data@entry=55, size=size@entry=2, attrs=...) at /home/ehabkost/rh/proj/virt/qemu/memory.c:1497
> #9  0x0000559a4e338708 in io_writex (env=env@entry=0x559a503d5620, mmu_idx=mmu_idx@entry=1, val=val@entry=55, addr=addr@entry=18446462598867529728, retaddr=139836732143069, size=2, iotlbentry=<optimized out>, iotlbentry=<optimized out>)
>     at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cputlb.c:945
> #10 0x0000559a4e33d203 in store_helper (big_endian=false, size=2, retaddr=<optimized out>, oi=<optimized out>, val=55, addr=18446462598867529728, env=0x559a503d5620) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cputlb.c:1544
> #11 0x0000559a4e33d203 in helper_le_stw_mmu (env=0x559a503d5620, addr=18446462598867529728, val=55, oi=<optimized out>, retaddr=139836732143069) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cputlb.c:1636
> #12 0x00007f2e46bef5dd in code_gen_buffer ()
> #13 0x0000559a4e352381 in cpu_tb_exec (itb=<optimized out>, cpu=0x559a503d5620) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cpu-exec.c:171
> #14 0x0000559a4e352381 in cpu_loop_exec_tb (tb_exit=<synthetic pointer>, last_tb=<synthetic pointer>, tb=<optimized out>, cpu=0x559a503d5620) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cpu-exec.c:618
> #15 0x0000559a4e352381 in cpu_exec (cpu=cpu@entry=0x559a503cd360) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cpu-exec.c:729
> #16 0x0000559a4e30ea0f in tcg_cpu_exec (cpu=0x559a503cd360) at /home/ehabkost/rh/proj/virt/qemu/cpus.c:1434
> #17 0x0000559a4e310b6b in qemu_tcg_cpu_thread_fn (arg=arg@entry=0x559a503cd360) at /home/ehabkost/rh/proj/virt/qemu/cpus.c:1743
> #18 0x0000559a4e83669a in qemu_thread_start (args=<optimized out>) at util/qemu-thread-posix.c:502
> #19 0x00007f2e4ebf958e in start_thread () at /lib64/libpthread.so.0
> #20 0x00007f2e4eb266f3 in clone () at /lib64/libc.so.6

See also <https://bugzilla.redhat.com/show_bug.cgi?id=1661940>.

$ git show 6ab3fc32ea64 -- hw/char/pl011.c

Thanks
Laszlo

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Cleber Rosa]

On Fri, Jun 07, 2019 at 12:26:48AM -0300, Eduardo Habkost wrote:
> On Fri, Feb 01, 2019 at 11:10:31AM -0500, Cleber Rosa wrote:
> > 
> > 
> > On 1/31/19 4:26 PM, Cleber Rosa wrote:
> > > 
> > > 
> > > On 1/31/19 3:21 PM, Cleber Rosa wrote:
> > >>
> > >>
> > >> On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
> > >>>
> > >>> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> > >>>> Just like the previous tests, boots a Linux kernel on a aarch64 target
> > >>>> using the virt machine.
> > >>>>
> > >>>> One special option added is the CPU type, given that the kernel
> > >>>> selected fails to boot on the virt machine's default CPU (cortex-a15).
> > >>>>
> > >>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> > >>>> ---
> > >>>>   .travis.yml                            |  2 +-
> > >>>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
> > >>>>   2 files changed, 21 insertions(+), 1 deletion(-)
> > >>>>
> > >>>> diff --git a/.travis.yml b/.travis.yml
> > >>>> index 54100eea5a..595e8c0b6c 100644
> > >>>> --- a/.travis.yml
> > >>>> +++ b/.travis.yml
> > >>>> @@ -187,7 +187,7 @@ matrix:
> > >>>>         # Acceptance (Functional) tests
> > >>>>       - env:
> > >>>> -        - CONFIG="--python=/usr/bin/python3
> > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
> > >>>> +        - CONFIG="--python=/usr/bin/python3
> > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
> > >>>>
> > >>>>           - TEST_CMD="make check-acceptance"
> > >>>>         addons:
> > >>>>           apt:
> > >>>> diff --git a/tests/acceptance/boot_linux_console.py
> > >>>> b/tests/acceptance/boot_linux_console.py
> > >>>> index f3ccd23a7a..107700b517 100644
> > >>>> --- a/tests/acceptance/boot_linux_console.py
> > >>>> +++ b/tests/acceptance/boot_linux_console.py
> > >>>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
> > >>>>           self.vm.launch()
> > >>>>           console_pattern = 'Kernel command line: %s' %
> > >>>> kernel_command_line
> > >>>>           self.wait_for_console_pattern(console_pattern)
> > >>>> +
> > >>>> +    def test_aarch64_virt(self):
> > >>>
> > >>> That test case fails on my system (Fedora 29 x86_64). Avocado seems
> > >>> unable to kill the VM so it  reaches the timeout.
> > >>>
> > >>> I compiled QEMU with default configuration:
> > >>>
> > >>> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
> > >>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
> > >>>
> > >>>
> > >>> Follows a snippet of the Avocado's job.log file:
> > >>> ----
> > >>> 2019-01-31 14:41:34,912 test             L0602 INFO | START
> > >>> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
> > >>>
> > >>> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
> > >>> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
> > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
> > >>> path=*, default=aarch64) => 'aarch64'
> > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
> > >>> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
> > >>> 'aarch64-softmmu/qemu-system-aarch64'
> > >>> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
> > >>> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
> > >>> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
> > >>> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
> > >>> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
> > >>> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
> > >>> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
> > >>> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
> > >>> 'aarch64-softmmu/qemu-system-aarch64 -chardev
> > >>> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
> > >>> chardev=mon,mode=control -display none -vga none -machine virt -chardev
> > >>> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
> > >>> -serial chardev:console -cpu cortex-a53 -kernel
> > >>> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
> > >>> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
> > >>> 'qmp_capabilities'}
> > >>> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
> > >>> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
> > >>> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
> > >>>
> > >>> (...)
> > >>>
> > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > >>> Policy zone: DMA32
> > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > >>> Kernel command line: console=ttyAMA0
> > >>> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
> > >>> 'quit'}
> > >>
> > >> Here, a QMP response like "<<< {'return': {}}" would be expected.
> > >>
> > >> Since I can not reproduce this on my system (or on Travis-CI jobs I've
> > >> sent), can you tell me on top of which commit you've applied these patches?
> > >>
> > > 
> > > I spoke too soon:
> > > 
> > > https://travis-ci.org/clebergnu/qemu/jobs/487121425#L3033
> > > 
> > > This looks like a recent regression, and I'm guessing it's not on the
> > > test's side.  I'll try to bisect it and let you know.
> > > 
> > 
> > On a fresh environment, I am able to get this reproduced on every 2 of
> > runs, more or less.  When I hit it, I attached GDB to it, and the
> > backtrace shows:
> > 
> > Thread debugging using libthread_db enabled]
> > Using host libthread_db library "/lib64/libthread_db.so.1".
> > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > __lll_lock_wait () at ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > 103     2:      movl    %edx, %eax
> > (gdb) bt
> > #0  __lll_lock_wait () at
> > ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > #1  0x00007fc6ba1a2e09 in __GI___pthread_mutex_lock
> > (mutex=mutex@entry=0x5615a233d020 <qemu_global_mutex>) at
> > ../nptl/pthread_mutex_lock.c:80
> > #2  0x00005615a1bb7593 in qemu_mutex_lock_impl (mutex=0x5615a233d020
> > <qemu_global_mutex>, file=0x5615a1db2d4c "util/main-loop.c", line=236)
> > at util/qemu-thread-posix.c:66
> > #3  0x00005615a171125e in qemu_mutex_lock_iothread_impl
> > (file=file@entry=0x5615a1db2d4c "util/main-loop.c", line=line@entry=236)
> > at /home/cleber/src/qemu/cpus.c:1849
> > #4  0x00005615a1bb415d in os_host_main_loop_wait (timeout=<optimized
> > out>) at util/main-loop.c:236
> > #5  main_loop_wait (nonblocking=<optimized out>) at util/main-loop.c:497
> > #6  0x00005615a18fdd39 in main_loop () at vl.c:1928
> > #7  0x00005615a16c9ee9 in main (argc=<optimized out>, argv=<optimized
> > out>, envp=<optimized out>) at vl.c:4665
> 
> Tip: run "thread apply all bt" so you can get a backtrace of all
> threads.
> 
>

Nice, thanks!

> > 
> > Running it with `taskset -c 1` prevents this issue from happening, which
> > AFAICT, contributes even further towards this being a QEMU race condition.
> > 
> > I'm CC'ing Peter and Claudio (listed maintainers of aarch64), as this
> > seems to limited to that target.  Any tips on what to do here?
> 
> I am hitting this on Travis, too, and I finally could reproduce
> it locally,
> 
> The guest is still writing on the serial console, but nobody is
> reading the data on the other side.  A VCPU thread is stuck
> inside the EAGAIN/nanosleep loop at qemu_chr_write_buffer(),
> holding the QEMU global lock.
>

This is awesome, because it really looks like a perfect explanation
for the behavior/reproducer/workaround that I described here:

 https://bugs.launchpad.net/qemu/+bug/1829779

> Thread 4 (Thread 0x7f2e45fff700 (LWP 6461)):
> #0  0x00007f2e4ec03500 in nanosleep () at /lib64/libpthread.so.0
> #1  0x00007f2e4fb229d7 in g_usleep () at /lib64/libglib-2.0.so.0
> #2  0x0000559a4e7ca4c9 in qemu_chr_write_buffer (s=s@entry=0x559a502d0ac0, buf=buf@entry=0x7f2e45ffdd90 "7", len=1, offset=offset@entry=0x7f2e45ffdd60, write_all=true) at chardev/char.c:115
> #3  0x0000559a4e7ca78f in qemu_chr_write (s=0x559a502d0ac0, buf=buf@entry=0x7f2e45ffdd90 "7", len=len@entry=1, write_all=write_all@entry=true) at chardev/char.c:148
> #4  0x0000559a4e7cc7e2 in qemu_chr_fe_write_all (be=be@entry=0x559a504b4c50, buf=buf@entry=0x7f2e45ffdd90 "7", len=len@entry=1) at chardev/char-fe.c:53
> #5  0x0000559a4e58f320 in pl011_write (opaque=0x559a504b47d0, offset=0, value=55, size=<optimized out>) at hw/char/pl011.c:183
> #6  0x0000559a4e325121 in memory_region_write_accessor (mr=0x559a504b4ae0, addr=0, value=<optimized out>, size=2, shift=<optimized out>, mask=<optimized out>, attrs=...) at /home/ehabkost/rh/proj/virt/qemu/memory.c:503
> #7  0x0000559a4e322cd6 in access_with_adjusted_size (addr=addr@entry=0, value=value@entry=0x7f2e45ffded8, size=size@entry=2, access_size_min=<optimized out>, access_size_max=<optimized out>, access_fn=access_fn@entry=
>     0x559a4e3250a0 <memory_region_write_accessor>, mr=0x559a504b4ae0, attrs=...) at /home/ehabkost/rh/proj/virt/qemu/memory.c:569
> #8  0x0000559a4e32763f in memory_region_dispatch_write (mr=mr@entry=0x559a504b4ae0, addr=addr@entry=0, data=<optimized out>, data@entry=55, size=size@entry=2, attrs=...) at /home/ehabkost/rh/proj/virt/qemu/memory.c:1497
> #9  0x0000559a4e338708 in io_writex (env=env@entry=0x559a503d5620, mmu_idx=mmu_idx@entry=1, val=val@entry=55, addr=addr@entry=18446462598867529728, retaddr=139836732143069, size=2, iotlbentry=<optimized out>, iotlbentry=<optimized out>)
>     at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cputlb.c:945
> #10 0x0000559a4e33d203 in store_helper (big_endian=false, size=2, retaddr=<optimized out>, oi=<optimized out>, val=55, addr=18446462598867529728, env=0x559a503d5620) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cputlb.c:1544
> #11 0x0000559a4e33d203 in helper_le_stw_mmu (env=0x559a503d5620, addr=18446462598867529728, val=55, oi=<optimized out>, retaddr=139836732143069) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cputlb.c:1636
> #12 0x00007f2e46bef5dd in code_gen_buffer ()
> #13 0x0000559a4e352381 in cpu_tb_exec (itb=<optimized out>, cpu=0x559a503d5620) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cpu-exec.c:171
> #14 0x0000559a4e352381 in cpu_loop_exec_tb (tb_exit=<synthetic pointer>, last_tb=<synthetic pointer>, tb=<optimized out>, cpu=0x559a503d5620) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cpu-exec.c:618
> #15 0x0000559a4e352381 in cpu_exec (cpu=cpu@entry=0x559a503cd360) at /home/ehabkost/rh/proj/virt/qemu/accel/tcg/cpu-exec.c:729
> #16 0x0000559a4e30ea0f in tcg_cpu_exec (cpu=0x559a503cd360) at /home/ehabkost/rh/proj/virt/qemu/cpus.c:1434
> #17 0x0000559a4e310b6b in qemu_tcg_cpu_thread_fn (arg=arg@entry=0x559a503cd360) at /home/ehabkost/rh/proj/virt/qemu/cpus.c:1743
> #18 0x0000559a4e83669a in qemu_thread_start (args=<optimized out>) at util/qemu-thread-posix.c:502
> #19 0x00007f2e4ebf958e in start_thread () at /lib64/libpthread.so.0
> #20 0x00007f2e4eb266f3 in clone () at /lib64/libc.so.6
> 
> 
> For reference, this is the QEMU command line:
> 
> aarch64-softmmu/qemu-system-aarch64 -chardev socket,id=mon,path=/var/tmp/tmpxnkcjvf0/qemu-6453-monitor.sock -mon chardev=mon,mode=control -display none -vga none -machine virt -chardev socket,id=console,path=/var/tmp/tmpxnkcjvf0/qemu-6453-console.sock,server,nowait -serial chardev:console -cpu cortex-a53 -kernel /home/ehabkost/rh/proj/virt/qemu/tests/venv/data/cache/by_location/e959d0e1dd72e77653e218e666198db1f3d0c213/vmlinuz -append printk.time=0 console=ttyAMA0
> 
> 
> 
> -- 
> Eduardo

Besides the command line, I hope it's also easy to use the following
reproducer:

 https://github.com/clebergnu/qemu/commit/c778e28c24030c4a36548b714293b319f4bf18df

It should apply cleanly to QEMU master, and instructions on how to run
it are documented in the commit message.

Thanks,
- Cleber.

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Cleber Rosa]

On Fri, Jun 07, 2019 at 12:42:14AM -0300, Eduardo Habkost wrote:
> On Fri, Jun 07, 2019 at 12:26:48AM -0300, Eduardo Habkost wrote:
> > On Fri, Feb 01, 2019 at 11:10:31AM -0500, Cleber Rosa wrote:
> > > 
> > > 
> > > On 1/31/19 4:26 PM, Cleber Rosa wrote:
> > > > 
> > > > 
> > > > On 1/31/19 3:21 PM, Cleber Rosa wrote:
> > > >>
> > > >>
> > > >> On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
> > > >>>
> > > >>> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> > > >>>> Just like the previous tests, boots a Linux kernel on a aarch64 target
> > > >>>> using the virt machine.
> > > >>>>
> > > >>>> One special option added is the CPU type, given that the kernel
> > > >>>> selected fails to boot on the virt machine's default CPU (cortex-a15).
> > > >>>>
> > > >>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> > > >>>> ---
> > > >>>>   .travis.yml                            |  2 +-
> > > >>>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
> > > >>>>   2 files changed, 21 insertions(+), 1 deletion(-)
> > > >>>>
> > > >>>> diff --git a/.travis.yml b/.travis.yml
> > > >>>> index 54100eea5a..595e8c0b6c 100644
> > > >>>> --- a/.travis.yml
> > > >>>> +++ b/.travis.yml
> > > >>>> @@ -187,7 +187,7 @@ matrix:
> > > >>>>         # Acceptance (Functional) tests
> > > >>>>       - env:
> > > >>>> -        - CONFIG="--python=/usr/bin/python3
> > > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
> > > >>>> +        - CONFIG="--python=/usr/bin/python3
> > > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
> > > >>>>
> > > >>>>           - TEST_CMD="make check-acceptance"
> > > >>>>         addons:
> > > >>>>           apt:
> > > >>>> diff --git a/tests/acceptance/boot_linux_console.py
> > > >>>> b/tests/acceptance/boot_linux_console.py
> > > >>>> index f3ccd23a7a..107700b517 100644
> > > >>>> --- a/tests/acceptance/boot_linux_console.py
> > > >>>> +++ b/tests/acceptance/boot_linux_console.py
> > > >>>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
> > > >>>>           self.vm.launch()
> > > >>>>           console_pattern = 'Kernel command line: %s' %
> > > >>>> kernel_command_line
> > > >>>>           self.wait_for_console_pattern(console_pattern)
> > > >>>> +
> > > >>>> +    def test_aarch64_virt(self):
> > > >>>
> > > >>> That test case fails on my system (Fedora 29 x86_64). Avocado seems
> > > >>> unable to kill the VM so it  reaches the timeout.
> > > >>>
> > > >>> I compiled QEMU with default configuration:
> > > >>>
> > > >>> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
> > > >>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
> > > >>>
> > > >>>
> > > >>> Follows a snippet of the Avocado's job.log file:
> > > >>> ----
> > > >>> 2019-01-31 14:41:34,912 test             L0602 INFO | START
> > > >>> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
> > > >>>
> > > >>> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
> > > >>> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
> > > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
> > > >>> path=*, default=aarch64) => 'aarch64'
> > > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
> > > >>> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
> > > >>> 'aarch64-softmmu/qemu-system-aarch64'
> > > >>> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
> > > >>> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
> > > >>> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
> > > >>> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
> > > >>> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
> > > >>> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
> > > >>> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
> > > >>> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
> > > >>> 'aarch64-softmmu/qemu-system-aarch64 -chardev
> > > >>> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
> > > >>> chardev=mon,mode=control -display none -vga none -machine virt -chardev
> > > >>> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
> > > >>> -serial chardev:console -cpu cortex-a53 -kernel
> > > >>> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
> > > >>> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
> > > >>> 'qmp_capabilities'}
> > > >>> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
> > > >>> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > >>> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
> > > >>>
> > > >>> (...)
> > > >>>
> > > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > >>> Policy zone: DMA32
> > > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > >>> Kernel command line: console=ttyAMA0
> > > >>> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
> > > >>> 'quit'}
> > > >>
> > > >> Here, a QMP response like "<<< {'return': {}}" would be expected.
> > > >>
> > > >> Since I can not reproduce this on my system (or on Travis-CI jobs I've
> > > >> sent), can you tell me on top of which commit you've applied these patches?
> > > >>
> > > > 
> > > > I spoke too soon:
> > > > 
> > > > https://travis-ci.org/clebergnu/qemu/jobs/487121425#L3033
> > > > 
> > > > This looks like a recent regression, and I'm guessing it's not on the
> > > > test's side.  I'll try to bisect it and let you know.
> > > > 
> > > 
> > > On a fresh environment, I am able to get this reproduced on every 2 of
> > > runs, more or less.  When I hit it, I attached GDB to it, and the
> > > backtrace shows:
> > > 
> > > Thread debugging using libthread_db enabled]
> > > Using host libthread_db library "/lib64/libthread_db.so.1".
> > > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > > __lll_lock_wait () at ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > > 103     2:      movl    %edx, %eax
> > > (gdb) bt
> > > #0  __lll_lock_wait () at
> > > ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > > #1  0x00007fc6ba1a2e09 in __GI___pthread_mutex_lock
> > > (mutex=mutex@entry=0x5615a233d020 <qemu_global_mutex>) at
> > > ../nptl/pthread_mutex_lock.c:80
> > > #2  0x00005615a1bb7593 in qemu_mutex_lock_impl (mutex=0x5615a233d020
> > > <qemu_global_mutex>, file=0x5615a1db2d4c "util/main-loop.c", line=236)
> > > at util/qemu-thread-posix.c:66
> > > #3  0x00005615a171125e in qemu_mutex_lock_iothread_impl
> > > (file=file@entry=0x5615a1db2d4c "util/main-loop.c", line=line@entry=236)
> > > at /home/cleber/src/qemu/cpus.c:1849
> > > #4  0x00005615a1bb415d in os_host_main_loop_wait (timeout=<optimized
> > > out>) at util/main-loop.c:236
> > > #5  main_loop_wait (nonblocking=<optimized out>) at util/main-loop.c:497
> > > #6  0x00005615a18fdd39 in main_loop () at vl.c:1928
> > > #7  0x00005615a16c9ee9 in main (argc=<optimized out>, argv=<optimized
> > > out>, envp=<optimized out>) at vl.c:4665
> > 
> > Tip: run "thread apply all bt" so you can get a backtrace of all
> > threads.
> > 
> > 
> > > 
> > > Running it with `taskset -c 1` prevents this issue from happening, which
> > > AFAICT, contributes even further towards this being a QEMU race condition.
> > > 
> > > I'm CC'ing Peter and Claudio (listed maintainers of aarch64), as this
> > > seems to limited to that target.  Any tips on what to do here?
> > 
> > I am hitting this on Travis, too, and I finally could reproduce
> > it locally,
> > 
> > The guest is still writing on the serial console, but nobody is
> > reading the data on the other side.  A VCPU thread is stuck
> > inside the EAGAIN/nanosleep loop at qemu_chr_write_buffer(),
> > holding the QEMU global lock.
> 
> Experimental fix below.
> 
> Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
> ---
>  python/qemu/__init__.py | 12 ++++++++----
>  1 file changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/python/qemu/__init__.py b/python/qemu/__init__.py
> index 81d9657ec0..4a691f34da 100644
> --- a/python/qemu/__init__.py
> +++ b/python/qemu/__init__.py
> @@ -274,10 +274,6 @@ class QEMUMachine(object):
>  
>          self._qemu_log_path = None
>  
> -        if self._console_socket is not None:
> -            self._console_socket.close()
> -            self._console_socket = None
> -
>          if self._temp_dir is not None:
>              shutil.rmtree(self._temp_dir)
>              self._temp_dir = None
> @@ -336,6 +332,14 @@ class QEMUMachine(object):
>          """
>          Terminate the VM and clean up
>          """
> +
> +        # If we keep the console socket open, we may deadlock waiting
> +        # for QEMU to exit, while QEMU is waiting for the socket to
> +        # become writeable.
> +        if self._console_socket is not None:
> +            self._console_socket.close()
> +            self._console_socket = None
> +

Right, this is somewhat equivalent to the following "workaround":

   https://github.com/clebergnu/qemu/commit/e1713f3b91972ad57c089f276c54db3f3fa63423

I could not tell at the moment, though, if closing (or shutting down)
the console socket was the appropriate fix.

What I see is that Rick's commit pointed to by Lazlo is dated from
2016, and it says that a virtio-console fix is necessary.  I'd imagine
that, if a fix was ever proposed, it'd have been incorporated in the
F29 kernel used in this test... and that this workaround/fix would
not be the best solution.

But, I'm mostly attempting to navigate this using my senses , few hard
data and even less background :).

Regards,
- Cleber.

>          if self.is_running():
>              try:
>                  self._qmp.cmd('quit')
> -- 
> 2.18.0.rc1.1.g3f1ff2140
> 
> -- 
> Eduardo

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Eduardo Habkost]

CCing Daniel, who wrote commit 6ab3fc32ea64.

On Fri, Jun 07, 2019 at 11:44:32AM -0400, Cleber Rosa wrote:
> On Fri, Jun 07, 2019 at 12:42:14AM -0300, Eduardo Habkost wrote:
> > On Fri, Jun 07, 2019 at 12:26:48AM -0300, Eduardo Habkost wrote:
> > > On Fri, Feb 01, 2019 at 11:10:31AM -0500, Cleber Rosa wrote:
> > > > 
> > > > 
> > > > On 1/31/19 4:26 PM, Cleber Rosa wrote:
> > > > > 
> > > > > 
> > > > > On 1/31/19 3:21 PM, Cleber Rosa wrote:
> > > > >>
> > > > >>
> > > > >> On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
> > > > >>>
> > > > >>> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> > > > >>>> Just like the previous tests, boots a Linux kernel on a aarch64 target
> > > > >>>> using the virt machine.
> > > > >>>>
> > > > >>>> One special option added is the CPU type, given that the kernel
> > > > >>>> selected fails to boot on the virt machine's default CPU (cortex-a15).
> > > > >>>>
> > > > >>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> > > > >>>> ---
> > > > >>>>   .travis.yml                            |  2 +-
> > > > >>>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
> > > > >>>>   2 files changed, 21 insertions(+), 1 deletion(-)
> > > > >>>>
> > > > >>>> diff --git a/.travis.yml b/.travis.yml
> > > > >>>> index 54100eea5a..595e8c0b6c 100644
> > > > >>>> --- a/.travis.yml
> > > > >>>> +++ b/.travis.yml
> > > > >>>> @@ -187,7 +187,7 @@ matrix:
> > > > >>>>         # Acceptance (Functional) tests
> > > > >>>>       - env:
> > > > >>>> -        - CONFIG="--python=/usr/bin/python3
> > > > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
> > > > >>>> +        - CONFIG="--python=/usr/bin/python3
> > > > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
> > > > >>>>
> > > > >>>>           - TEST_CMD="make check-acceptance"
> > > > >>>>         addons:
> > > > >>>>           apt:
> > > > >>>> diff --git a/tests/acceptance/boot_linux_console.py
> > > > >>>> b/tests/acceptance/boot_linux_console.py
> > > > >>>> index f3ccd23a7a..107700b517 100644
> > > > >>>> --- a/tests/acceptance/boot_linux_console.py
> > > > >>>> +++ b/tests/acceptance/boot_linux_console.py
> > > > >>>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
> > > > >>>>           self.vm.launch()
> > > > >>>>           console_pattern = 'Kernel command line: %s' %
> > > > >>>> kernel_command_line
> > > > >>>>           self.wait_for_console_pattern(console_pattern)
> > > > >>>> +
> > > > >>>> +    def test_aarch64_virt(self):
> > > > >>>
> > > > >>> That test case fails on my system (Fedora 29 x86_64). Avocado seems
> > > > >>> unable to kill the VM so it  reaches the timeout.
> > > > >>>
> > > > >>> I compiled QEMU with default configuration:
> > > > >>>
> > > > >>> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
> > > > >>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
> > > > >>>
> > > > >>>
> > > > >>> Follows a snippet of the Avocado's job.log file:
> > > > >>> ----
> > > > >>> 2019-01-31 14:41:34,912 test             L0602 INFO | START
> > > > >>> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
> > > > >>>
> > > > >>> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
> > > > >>> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
> > > > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
> > > > >>> path=*, default=aarch64) => 'aarch64'
> > > > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
> > > > >>> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
> > > > >>> 'aarch64-softmmu/qemu-system-aarch64'
> > > > >>> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
> > > > >>> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
> > > > >>> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
> > > > >>> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
> > > > >>> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
> > > > >>> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
> > > > >>> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
> > > > >>> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
> > > > >>> 'aarch64-softmmu/qemu-system-aarch64 -chardev
> > > > >>> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
> > > > >>> chardev=mon,mode=control -display none -vga none -machine virt -chardev
> > > > >>> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
> > > > >>> -serial chardev:console -cpu cortex-a53 -kernel
> > > > >>> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
> > > > >>> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
> > > > >>> 'qmp_capabilities'}
> > > > >>> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
> > > > >>> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > > >>> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
> > > > >>>
> > > > >>> (...)
> > > > >>>
> > > > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > > >>> Policy zone: DMA32
> > > > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > > >>> Kernel command line: console=ttyAMA0
> > > > >>> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
> > > > >>> 'quit'}
> > > > >>
> > > > >> Here, a QMP response like "<<< {'return': {}}" would be expected.
> > > > >>
> > > > >> Since I can not reproduce this on my system (or on Travis-CI jobs I've
> > > > >> sent), can you tell me on top of which commit you've applied these patches?
> > > > >>
> > > > > 
> > > > > I spoke too soon:
> > > > > 
> > > > > https://travis-ci.org/clebergnu/qemu/jobs/487121425#L3033
> > > > > 
> > > > > This looks like a recent regression, and I'm guessing it's not on the
> > > > > test's side.  I'll try to bisect it and let you know.
> > > > > 
> > > > 
> > > > On a fresh environment, I am able to get this reproduced on every 2 of
> > > > runs, more or less.  When I hit it, I attached GDB to it, and the
> > > > backtrace shows:
> > > > 
> > > > Thread debugging using libthread_db enabled]
> > > > Using host libthread_db library "/lib64/libthread_db.so.1".
> > > > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > > > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > > > __lll_lock_wait () at ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > > > 103     2:      movl    %edx, %eax
> > > > (gdb) bt
> > > > #0  __lll_lock_wait () at
> > > > ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > > > #1  0x00007fc6ba1a2e09 in __GI___pthread_mutex_lock
> > > > (mutex=mutex@entry=0x5615a233d020 <qemu_global_mutex>) at
> > > > ../nptl/pthread_mutex_lock.c:80
> > > > #2  0x00005615a1bb7593 in qemu_mutex_lock_impl (mutex=0x5615a233d020
> > > > <qemu_global_mutex>, file=0x5615a1db2d4c "util/main-loop.c", line=236)
> > > > at util/qemu-thread-posix.c:66
> > > > #3  0x00005615a171125e in qemu_mutex_lock_iothread_impl
> > > > (file=file@entry=0x5615a1db2d4c "util/main-loop.c", line=line@entry=236)
> > > > at /home/cleber/src/qemu/cpus.c:1849
> > > > #4  0x00005615a1bb415d in os_host_main_loop_wait (timeout=<optimized
> > > > out>) at util/main-loop.c:236
> > > > #5  main_loop_wait (nonblocking=<optimized out>) at util/main-loop.c:497
> > > > #6  0x00005615a18fdd39 in main_loop () at vl.c:1928
> > > > #7  0x00005615a16c9ee9 in main (argc=<optimized out>, argv=<optimized
> > > > out>, envp=<optimized out>) at vl.c:4665
> > > 
> > > Tip: run "thread apply all bt" so you can get a backtrace of all
> > > threads.
> > > 
> > > 
> > > > 
> > > > Running it with `taskset -c 1` prevents this issue from happening, which
> > > > AFAICT, contributes even further towards this being a QEMU race condition.
> > > > 
> > > > I'm CC'ing Peter and Claudio (listed maintainers of aarch64), as this
> > > > seems to limited to that target.  Any tips on what to do here?
> > > 
> > > I am hitting this on Travis, too, and I finally could reproduce
> > > it locally,
> > > 
> > > The guest is still writing on the serial console, but nobody is
> > > reading the data on the other side.  A VCPU thread is stuck
> > > inside the EAGAIN/nanosleep loop at qemu_chr_write_buffer(),
> > > holding the QEMU global lock.
> > 
> > Experimental fix below.
> > 
> > Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
> > ---
> >  python/qemu/__init__.py | 12 ++++++++----
> >  1 file changed, 8 insertions(+), 4 deletions(-)
> > 
> > diff --git a/python/qemu/__init__.py b/python/qemu/__init__.py
> > index 81d9657ec0..4a691f34da 100644
> > --- a/python/qemu/__init__.py
> > +++ b/python/qemu/__init__.py
> > @@ -274,10 +274,6 @@ class QEMUMachine(object):
> >  
> >          self._qemu_log_path = None
> >  
> > -        if self._console_socket is not None:
> > -            self._console_socket.close()
> > -            self._console_socket = None
> > -
> >          if self._temp_dir is not None:
> >              shutil.rmtree(self._temp_dir)
> >              self._temp_dir = None
> > @@ -336,6 +332,14 @@ class QEMUMachine(object):
> >          """
> >          Terminate the VM and clean up
> >          """
> > +
> > +        # If we keep the console socket open, we may deadlock waiting
> > +        # for QEMU to exit, while QEMU is waiting for the socket to
> > +        # become writeable.
> > +        if self._console_socket is not None:
> > +            self._console_socket.close()
> > +            self._console_socket = None
> > +
> 
> Right, this is somewhat equivalent to the following "workaround":
> 
>    https://github.com/clebergnu/qemu/commit/e1713f3b91972ad57c089f276c54db3f3fa63423
> 
> I could not tell at the moment, though, if closing (or shutting down)
> the console socket was the appropriate fix.
> 
> What I see is that Rick's commit pointed to by Lazlo is dated from
> 2016, and it says that a virtio-console fix is necessary.  I'd imagine
> that, if a fix was ever proposed, it'd have been incorporated in the
> F29 kernel used in this test... and that this workaround/fix would
> not be the best solution.

If I understood this correctly, fixing the guest driver wouldn't
help here.  The commit mentioned by Laszlo (6ab3fc32ea64 "hw:
replace most use of qemu_chr_fe_write with qemu_chr_fe_write_all")
fixes data loss bugs but introduces the potential for deadlocks
like the one we are seeing.

Unless we replace the existing ~30 qemu_chr_fe_write_all() calls
in device code, we need to make sure we are constantly reading
data from the console socket.

> 
> But, I'm mostly attempting to navigate this using my senses , few hard
> data and even less background :).
> 
> Regards,
> - Cleber.
> 
> >          if self.is_running():
> >              try:
> >                  self._qmp.cmd('quit')
> > -- 
> > 2.18.0.rc1.1.g3f1ff2140
> > 
> > -- 
> > Eduardo

-- 
Eduardo

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Daniel P. Berrangé]

On Fri, Jun 07, 2019 at 03:58:57PM -0300, Eduardo Habkost wrote:
> CCing Daniel, who wrote commit 6ab3fc32ea64.
> 
> On Fri, Jun 07, 2019 at 11:44:32AM -0400, Cleber Rosa wrote:
> > On Fri, Jun 07, 2019 at 12:42:14AM -0300, Eduardo Habkost wrote:
> > > On Fri, Jun 07, 2019 at 12:26:48AM -0300, Eduardo Habkost wrote:
> > > > On Fri, Feb 01, 2019 at 11:10:31AM -0500, Cleber Rosa wrote:
> > > > > 
> > > > > 
> > > > > On 1/31/19 4:26 PM, Cleber Rosa wrote:
> > > > > > 
> > > > > > 
> > > > > > On 1/31/19 3:21 PM, Cleber Rosa wrote:
> > > > > >>
> > > > > >>
> > > > > >> On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
> > > > > >>>
> > > > > >>> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> > > > > >>>> Just like the previous tests, boots a Linux kernel on a aarch64 target
> > > > > >>>> using the virt machine.
> > > > > >>>>
> > > > > >>>> One special option added is the CPU type, given that the kernel
> > > > > >>>> selected fails to boot on the virt machine's default CPU (cortex-a15).
> > > > > >>>>
> > > > > >>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> > > > > >>>> ---
> > > > > >>>>   .travis.yml                            |  2 +-
> > > > > >>>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
> > > > > >>>>   2 files changed, 21 insertions(+), 1 deletion(-)
> > > > > >>>>
> > > > > >>>> diff --git a/.travis.yml b/.travis.yml
> > > > > >>>> index 54100eea5a..595e8c0b6c 100644
> > > > > >>>> --- a/.travis.yml
> > > > > >>>> +++ b/.travis.yml
> > > > > >>>> @@ -187,7 +187,7 @@ matrix:
> > > > > >>>>         # Acceptance (Functional) tests
> > > > > >>>>       - env:
> > > > > >>>> -        - CONFIG="--python=/usr/bin/python3
> > > > > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
> > > > > >>>> +        - CONFIG="--python=/usr/bin/python3
> > > > > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
> > > > > >>>>
> > > > > >>>>           - TEST_CMD="make check-acceptance"
> > > > > >>>>         addons:
> > > > > >>>>           apt:
> > > > > >>>> diff --git a/tests/acceptance/boot_linux_console.py
> > > > > >>>> b/tests/acceptance/boot_linux_console.py
> > > > > >>>> index f3ccd23a7a..107700b517 100644
> > > > > >>>> --- a/tests/acceptance/boot_linux_console.py
> > > > > >>>> +++ b/tests/acceptance/boot_linux_console.py
> > > > > >>>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
> > > > > >>>>           self.vm.launch()
> > > > > >>>>           console_pattern = 'Kernel command line: %s' %
> > > > > >>>> kernel_command_line
> > > > > >>>>           self.wait_for_console_pattern(console_pattern)
> > > > > >>>> +
> > > > > >>>> +    def test_aarch64_virt(self):
> > > > > >>>
> > > > > >>> That test case fails on my system (Fedora 29 x86_64). Avocado seems
> > > > > >>> unable to kill the VM so it  reaches the timeout.
> > > > > >>>
> > > > > >>> I compiled QEMU with default configuration:
> > > > > >>>
> > > > > >>> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
> > > > > >>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
> > > > > >>>
> > > > > >>>
> > > > > >>> Follows a snippet of the Avocado's job.log file:
> > > > > >>> ----
> > > > > >>> 2019-01-31 14:41:34,912 test             L0602 INFO | START
> > > > > >>> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
> > > > > >>>
> > > > > >>> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
> > > > > >>> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
> > > > > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
> > > > > >>> path=*, default=aarch64) => 'aarch64'
> > > > > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
> > > > > >>> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
> > > > > >>> 'aarch64-softmmu/qemu-system-aarch64'
> > > > > >>> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
> > > > > >>> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
> > > > > >>> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
> > > > > >>> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
> > > > > >>> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
> > > > > >>> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
> > > > > >>> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
> > > > > >>> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
> > > > > >>> 'aarch64-softmmu/qemu-system-aarch64 -chardev
> > > > > >>> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
> > > > > >>> chardev=mon,mode=control -display none -vga none -machine virt -chardev
> > > > > >>> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
> > > > > >>> -serial chardev:console -cpu cortex-a53 -kernel
> > > > > >>> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
> > > > > >>> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
> > > > > >>> 'qmp_capabilities'}
> > > > > >>> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
> > > > > >>> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > > > >>> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
> > > > > >>>
> > > > > >>> (...)
> > > > > >>>
> > > > > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > > > >>> Policy zone: DMA32
> > > > > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > > > >>> Kernel command line: console=ttyAMA0
> > > > > >>> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
> > > > > >>> 'quit'}
> > > > > >>
> > > > > >> Here, a QMP response like "<<< {'return': {}}" would be expected.
> > > > > >>
> > > > > >> Since I can not reproduce this on my system (or on Travis-CI jobs I've
> > > > > >> sent), can you tell me on top of which commit you've applied these patches?
> > > > > >>
> > > > > > 
> > > > > > I spoke too soon:
> > > > > > 
> > > > > > https://travis-ci.org/clebergnu/qemu/jobs/487121425#L3033
> > > > > > 
> > > > > > This looks like a recent regression, and I'm guessing it's not on the
> > > > > > test's side.  I'll try to bisect it and let you know.
> > > > > > 
> > > > > 
> > > > > On a fresh environment, I am able to get this reproduced on every 2 of
> > > > > runs, more or less.  When I hit it, I attached GDB to it, and the
> > > > > backtrace shows:
> > > > > 
> > > > > Thread debugging using libthread_db enabled]
> > > > > Using host libthread_db library "/lib64/libthread_db.so.1".
> > > > > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > > > > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > > > > __lll_lock_wait () at ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > > > > 103     2:      movl    %edx, %eax
> > > > > (gdb) bt
> > > > > #0  __lll_lock_wait () at
> > > > > ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > > > > #1  0x00007fc6ba1a2e09 in __GI___pthread_mutex_lock
> > > > > (mutex=mutex@entry=0x5615a233d020 <qemu_global_mutex>) at
> > > > > ../nptl/pthread_mutex_lock.c:80
> > > > > #2  0x00005615a1bb7593 in qemu_mutex_lock_impl (mutex=0x5615a233d020
> > > > > <qemu_global_mutex>, file=0x5615a1db2d4c "util/main-loop.c", line=236)
> > > > > at util/qemu-thread-posix.c:66
> > > > > #3  0x00005615a171125e in qemu_mutex_lock_iothread_impl
> > > > > (file=file@entry=0x5615a1db2d4c "util/main-loop.c", line=line@entry=236)
> > > > > at /home/cleber/src/qemu/cpus.c:1849
> > > > > #4  0x00005615a1bb415d in os_host_main_loop_wait (timeout=<optimized
> > > > > out>) at util/main-loop.c:236
> > > > > #5  main_loop_wait (nonblocking=<optimized out>) at util/main-loop.c:497
> > > > > #6  0x00005615a18fdd39 in main_loop () at vl.c:1928
> > > > > #7  0x00005615a16c9ee9 in main (argc=<optimized out>, argv=<optimized
> > > > > out>, envp=<optimized out>) at vl.c:4665
> > > > 
> > > > Tip: run "thread apply all bt" so you can get a backtrace of all
> > > > threads.
> > > > 
> > > > 
> > > > > 
> > > > > Running it with `taskset -c 1` prevents this issue from happening, which
> > > > > AFAICT, contributes even further towards this being a QEMU race condition.
> > > > > 
> > > > > I'm CC'ing Peter and Claudio (listed maintainers of aarch64), as this
> > > > > seems to limited to that target.  Any tips on what to do here?
> > > > 
> > > > I am hitting this on Travis, too, and I finally could reproduce
> > > > it locally,
> > > > 
> > > > The guest is still writing on the serial console, but nobody is
> > > > reading the data on the other side.  A VCPU thread is stuck
> > > > inside the EAGAIN/nanosleep loop at qemu_chr_write_buffer(),
> > > > holding the QEMU global lock.
> > > 
> > > Experimental fix below.
> > > 
> > > Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
> > > ---
> > >  python/qemu/__init__.py | 12 ++++++++----
> > >  1 file changed, 8 insertions(+), 4 deletions(-)
> > > 
> > > diff --git a/python/qemu/__init__.py b/python/qemu/__init__.py
> > > index 81d9657ec0..4a691f34da 100644
> > > --- a/python/qemu/__init__.py
> > > +++ b/python/qemu/__init__.py
> > > @@ -274,10 +274,6 @@ class QEMUMachine(object):
> > >  
> > >          self._qemu_log_path = None
> > >  
> > > -        if self._console_socket is not None:
> > > -            self._console_socket.close()
> > > -            self._console_socket = None
> > > -
> > >          if self._temp_dir is not None:
> > >              shutil.rmtree(self._temp_dir)
> > >              self._temp_dir = None
> > > @@ -336,6 +332,14 @@ class QEMUMachine(object):
> > >          """
> > >          Terminate the VM and clean up
> > >          """
> > > +
> > > +        # If we keep the console socket open, we may deadlock waiting
> > > +        # for QEMU to exit, while QEMU is waiting for the socket to
> > > +        # become writeable.
> > > +        if self._console_socket is not None:
> > > +            self._console_socket.close()
> > > +            self._console_socket = None
> > > +
> > 
> > Right, this is somewhat equivalent to the following "workaround":
> > 
> >    https://github.com/clebergnu/qemu/commit/e1713f3b91972ad57c089f276c54db3f3fa63423
> > 
> > I could not tell at the moment, though, if closing (or shutting down)
> > the console socket was the appropriate fix.
> > 
> > What I see is that Rick's commit pointed to by Lazlo is dated from
> > 2016, and it says that a virtio-console fix is necessary.  I'd imagine
> > that, if a fix was ever proposed, it'd have been incorporated in the
> > F29 kernel used in this test... and that this workaround/fix would
> > not be the best solution.
> 
> If I understood this correctly, fixing the guest driver wouldn't
> help here.  The commit mentioned by Laszlo (6ab3fc32ea64 "hw:
> replace most use of qemu_chr_fe_write with qemu_chr_fe_write_all")
> fixes data loss bugs but introduces the potential for deadlocks
> like the one we are seeing.
> 
> Unless we replace the existing ~30 qemu_chr_fe_write_all() calls
> in device code, we need to make sure we are constantly reading
> data from the console socket.

Yes, you must *always* have something reading from the chardev backend.
This really sucks, but it is preferrable to letting data end up in
/dev/null.

If this is being a problem for tests then consider it motivation to
fix the root cause problem and make the devices properly do async
I/O for chardevs.  Only a handful of them properly do this right
now.


Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

Re: [Qemu-devel] [PATCH 15/18] Boot Linux Console Test: add a test for aarch64 + virt

[Cleber Rosa]

On Mon, Jun 10, 2019 at 09:53:03AM +0100, Daniel P. Berrangé wrote:
> On Fri, Jun 07, 2019 at 03:58:57PM -0300, Eduardo Habkost wrote:
> > CCing Daniel, who wrote commit 6ab3fc32ea64.
> > 
> > On Fri, Jun 07, 2019 at 11:44:32AM -0400, Cleber Rosa wrote:
> > > On Fri, Jun 07, 2019 at 12:42:14AM -0300, Eduardo Habkost wrote:
> > > > On Fri, Jun 07, 2019 at 12:26:48AM -0300, Eduardo Habkost wrote:
> > > > > On Fri, Feb 01, 2019 at 11:10:31AM -0500, Cleber Rosa wrote:
> > > > > > 
> > > > > > 
> > > > > > On 1/31/19 4:26 PM, Cleber Rosa wrote:
> > > > > > > 
> > > > > > > 
> > > > > > > On 1/31/19 3:21 PM, Cleber Rosa wrote:
> > > > > > >>
> > > > > > >>
> > > > > > >> On 1/31/19 3:02 PM, Wainer dos Santos Moschetta wrote:
> > > > > > >>>
> > > > > > >>> On 01/17/2019 04:56 PM, Cleber Rosa wrote:
> > > > > > >>>> Just like the previous tests, boots a Linux kernel on a aarch64 target
> > > > > > >>>> using the virt machine.
> > > > > > >>>>
> > > > > > >>>> One special option added is the CPU type, given that the kernel
> > > > > > >>>> selected fails to boot on the virt machine's default CPU (cortex-a15).
> > > > > > >>>>
> > > > > > >>>> Signed-off-by: Cleber Rosa <crosa@redhat.com>
> > > > > > >>>> ---
> > > > > > >>>>   .travis.yml                            |  2 +-
> > > > > > >>>>   tests/acceptance/boot_linux_console.py | 20 ++++++++++++++++++++
> > > > > > >>>>   2 files changed, 21 insertions(+), 1 deletion(-)
> > > > > > >>>>
> > > > > > >>>> diff --git a/.travis.yml b/.travis.yml
> > > > > > >>>> index 54100eea5a..595e8c0b6c 100644
> > > > > > >>>> --- a/.travis.yml
> > > > > > >>>> +++ b/.travis.yml
> > > > > > >>>> @@ -187,7 +187,7 @@ matrix:
> > > > > > >>>>         # Acceptance (Functional) tests
> > > > > > >>>>       - env:
> > > > > > >>>> -        - CONFIG="--python=/usr/bin/python3
> > > > > > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu"
> > > > > > >>>> +        - CONFIG="--python=/usr/bin/python3
> > > > > > >>>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu"
> > > > > > >>>>
> > > > > > >>>>           - TEST_CMD="make check-acceptance"
> > > > > > >>>>         addons:
> > > > > > >>>>           apt:
> > > > > > >>>> diff --git a/tests/acceptance/boot_linux_console.py
> > > > > > >>>> b/tests/acceptance/boot_linux_console.py
> > > > > > >>>> index f3ccd23a7a..107700b517 100644
> > > > > > >>>> --- a/tests/acceptance/boot_linux_console.py
> > > > > > >>>> +++ b/tests/acceptance/boot_linux_console.py
> > > > > > >>>> @@ -138,3 +138,23 @@ class BootLinuxConsole(Test):
> > > > > > >>>>           self.vm.launch()
> > > > > > >>>>           console_pattern = 'Kernel command line: %s' %
> > > > > > >>>> kernel_command_line
> > > > > > >>>>           self.wait_for_console_pattern(console_pattern)
> > > > > > >>>> +
> > > > > > >>>> +    def test_aarch64_virt(self):
> > > > > > >>>
> > > > > > >>> That test case fails on my system (Fedora 29 x86_64). Avocado seems
> > > > > > >>> unable to kill the VM so it  reaches the timeout.
> > > > > > >>>
> > > > > > >>> I compiled QEMU with default configuration:
> > > > > > >>>
> > > > > > >>> $ configure --python=/usr/bin/python3 --target-list=x86_64-softmmu
> > > > > > >>> --target-list=x86_64-softmmu,mips-softmmu,mips64el-softmmu,ppc64-softmmu,aarch64-softmmu)
> > > > > > >>>
> > > > > > >>>
> > > > > > >>> Follows a snippet of the Avocado's job.log file:
> > > > > > >>> ----
> > > > > > >>> 2019-01-31 14:41:34,912 test             L0602 INFO | START
> > > > > > >>> 07-/root/src/qemu/tests/acceptance/boot_linux_console.py:BootLinuxConsole.test_aarch64_virt
> > > > > > >>>
> > > > > > >>> 2019-01-31 14:41:34,912 test             L0298 DEBUG| DATA
> > > > > > >>> (filename=output.expected) => NOT FOUND (data sources: variant, test, file)
> > > > > > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS (key=arch,
> > > > > > >>> path=*, default=aarch64) => 'aarch64'
> > > > > > >>> 2019-01-31 14:41:34,913 parameters       L0146 DEBUG| PARAMS
> > > > > > >>> (key=qemu_bin, path=*, default=aarch64-softmmu/qemu-system-aarch64) =>
> > > > > > >>> 'aarch64-softmmu/qemu-system-aarch64'
> > > > > > >>> 2019-01-31 14:41:34,915 download         L0070 INFO | Fetching
> > > > > > >>> https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz
> > > > > > >>> -> /var/lib/avocado/data/cache/by_name/vmlinuz.3upct2pr
> > > > > > >>> 2019-01-31 14:41:35,490 download         L0054 DEBUG| Retrieved URL
> > > > > > >>> "https://sjc.edge.kernel.org/fedora-buffet/fedora/linux/releases/29/Server/aarch64/os/images/pxeboot/vmlinuz":
> > > > > > >>> content-length 8623423, date: "Thu, 31 Jan 2019 19:41:35 GMT",
> > > > > > >>> last-modified: "Sun, 21 Oct 2018 00:43:09 GMT"
> > > > > > >>> 2019-01-31 14:41:41,765 qemu             L0317 DEBUG| VM launch command:
> > > > > > >>> 'aarch64-softmmu/qemu-system-aarch64 -chardev
> > > > > > >>> socket,id=mon,path=/var/tmp/tmpizirkcud/qemu-32609-monitor.sock -mon
> > > > > > >>> chardev=mon,mode=control -display none -vga none -machine virt -chardev
> > > > > > >>> socket,id=console,path=/var/tmp/tmpizirkcud/qemu-32609-console.sock,server,nowait
> > > > > > >>> -serial chardev:console -cpu cortex-a53 -kernel
> > > > > > >>> /var/lib/avocado/data/cache/by_name/vmlinuz -append console=ttyAMA0'
> > > > > > >>> 2019-01-31 14:41:41,779 qmp              L0167 DEBUG| >>> {'execute':
> > > > > > >>> 'qmp_capabilities'}
> > > > > > >>> 2019-01-31 14:41:41,931 qmp              L0175 DEBUG| <<< {'return': {}}
> > > > > > >>> 2019-01-31 14:41:42,830 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > > > > >>> Booting Linux on physical CPU 0x0000000000 [0x410fd034]
> > > > > > >>>
> > > > > > >>> (...)
> > > > > > >>>
> > > > > > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > > > > >>> Policy zone: DMA32
> > > > > > >>> 2019-01-31 14:41:42,833 boot_linux_conso L0041 DEBUG| [    0.000000]
> > > > > > >>> Kernel command line: console=ttyAMA0
> > > > > > >>> 2019-01-31 14:41:42,833 qmp              L0167 DEBUG| >>> {'execute':
> > > > > > >>> 'quit'}
> > > > > > >>
> > > > > > >> Here, a QMP response like "<<< {'return': {}}" would be expected.
> > > > > > >>
> > > > > > >> Since I can not reproduce this on my system (or on Travis-CI jobs I've
> > > > > > >> sent), can you tell me on top of which commit you've applied these patches?
> > > > > > >>
> > > > > > > 
> > > > > > > I spoke too soon:
> > > > > > > 
> > > > > > > https://travis-ci.org/clebergnu/qemu/jobs/487121425#L3033
> > > > > > > 
> > > > > > > This looks like a recent regression, and I'm guessing it's not on the
> > > > > > > test's side.  I'll try to bisect it and let you know.
> > > > > > > 
> > > > > > 
> > > > > > On a fresh environment, I am able to get this reproduced on every 2 of
> > > > > > runs, more or less.  When I hit it, I attached GDB to it, and the
> > > > > > backtrace shows:
> > > > > > 
> > > > > > Thread debugging using libthread_db enabled]
> > > > > > Using host libthread_db library "/lib64/libthread_db.so.1".
> > > > > > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > > > > > warning: Loadable section ".note.gnu.property" outside of ELF segments
> > > > > > __lll_lock_wait () at ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > > > > > 103     2:      movl    %edx, %eax
> > > > > > (gdb) bt
> > > > > > #0  __lll_lock_wait () at
> > > > > > ../sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:103
> > > > > > #1  0x00007fc6ba1a2e09 in __GI___pthread_mutex_lock
> > > > > > (mutex=mutex@entry=0x5615a233d020 <qemu_global_mutex>) at
> > > > > > ../nptl/pthread_mutex_lock.c:80
> > > > > > #2  0x00005615a1bb7593 in qemu_mutex_lock_impl (mutex=0x5615a233d020
> > > > > > <qemu_global_mutex>, file=0x5615a1db2d4c "util/main-loop.c", line=236)
> > > > > > at util/qemu-thread-posix.c:66
> > > > > > #3  0x00005615a171125e in qemu_mutex_lock_iothread_impl
> > > > > > (file=file@entry=0x5615a1db2d4c "util/main-loop.c", line=line@entry=236)
> > > > > > at /home/cleber/src/qemu/cpus.c:1849
> > > > > > #4  0x00005615a1bb415d in os_host_main_loop_wait (timeout=<optimized
> > > > > > out>) at util/main-loop.c:236
> > > > > > #5  main_loop_wait (nonblocking=<optimized out>) at util/main-loop.c:497
> > > > > > #6  0x00005615a18fdd39 in main_loop () at vl.c:1928
> > > > > > #7  0x00005615a16c9ee9 in main (argc=<optimized out>, argv=<optimized
> > > > > > out>, envp=<optimized out>) at vl.c:4665
> > > > > 
> > > > > Tip: run "thread apply all bt" so you can get a backtrace of all
> > > > > threads.
> > > > > 
> > > > > 
> > > > > > 
> > > > > > Running it with `taskset -c 1` prevents this issue from happening, which
> > > > > > AFAICT, contributes even further towards this being a QEMU race condition.
> > > > > > 
> > > > > > I'm CC'ing Peter and Claudio (listed maintainers of aarch64), as this
> > > > > > seems to limited to that target.  Any tips on what to do here?
> > > > > 
> > > > > I am hitting this on Travis, too, and I finally could reproduce
> > > > > it locally,
> > > > > 
> > > > > The guest is still writing on the serial console, but nobody is
> > > > > reading the data on the other side.  A VCPU thread is stuck
> > > > > inside the EAGAIN/nanosleep loop at qemu_chr_write_buffer(),
> > > > > holding the QEMU global lock.
> > > > 
> > > > Experimental fix below.
> > > > 
> > > > Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
> > > > ---
> > > >  python/qemu/__init__.py | 12 ++++++++----
> > > >  1 file changed, 8 insertions(+), 4 deletions(-)
> > > > 
> > > > diff --git a/python/qemu/__init__.py b/python/qemu/__init__.py
> > > > index 81d9657ec0..4a691f34da 100644
> > > > --- a/python/qemu/__init__.py
> > > > +++ b/python/qemu/__init__.py
> > > > @@ -274,10 +274,6 @@ class QEMUMachine(object):
> > > >  
> > > >          self._qemu_log_path = None
> > > >  
> > > > -        if self._console_socket is not None:
> > > > -            self._console_socket.close()
> > > > -            self._console_socket = None
> > > > -
> > > >          if self._temp_dir is not None:
> > > >              shutil.rmtree(self._temp_dir)
> > > >              self._temp_dir = None
> > > > @@ -336,6 +332,14 @@ class QEMUMachine(object):
> > > >          """
> > > >          Terminate the VM and clean up
> > > >          """
> > > > +
> > > > +        # If we keep the console socket open, we may deadlock waiting
> > > > +        # for QEMU to exit, while QEMU is waiting for the socket to
> > > > +        # become writeable.
> > > > +        if self._console_socket is not None:
> > > > +            self._console_socket.close()
> > > > +            self._console_socket = None
> > > > +
> > > 
> > > Right, this is somewhat equivalent to the following "workaround":
> > > 
> > >    https://github.com/clebergnu/qemu/commit/e1713f3b91972ad57c089f276c54db3f3fa63423
> > > 
> > > I could not tell at the moment, though, if closing (or shutting down)
> > > the console socket was the appropriate fix.
> > > 
> > > What I see is that Rick's commit pointed to by Lazlo is dated from
> > > 2016, and it says that a virtio-console fix is necessary.  I'd imagine
> > > that, if a fix was ever proposed, it'd have been incorporated in the
> > > F29 kernel used in this test... and that this workaround/fix would
> > > not be the best solution.
> > 
> > If I understood this correctly, fixing the guest driver wouldn't
> > help here.  The commit mentioned by Laszlo (6ab3fc32ea64 "hw:
> > replace most use of qemu_chr_fe_write with qemu_chr_fe_write_all")
> > fixes data loss bugs but introduces the potential for deadlocks
> > like the one we are seeing.
> > 
> > Unless we replace the existing ~30 qemu_chr_fe_write_all() calls
> > in device code, we need to make sure we are constantly reading
> > data from the console socket.
> 
> Yes, you must *always* have something reading from the chardev backend.
> This really sucks, but it is preferrable to letting data end up in
> /dev/null.
> 
> If this is being a problem for tests then consider it motivation to
> fix the root cause problem and make the devices properly do async
> I/O for chardevs.  Only a handful of them properly do this right
> now.
>
> 
> Regards,
> Daniel
> -- 
> |: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
> |: https://libvirt.org         -o-            https://fstop138.berrange.com :|
> |: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

Daniel,

Thanks for the explanation.  I don't feel completely capable of tackling
the root cause, but I'll give it a go anyway.

Eduardo,

Looks like we'll need this type of workaround now anyways.  As seen in
the reproducer I wrote, a test can cause this deadlock before it gets
to the test tearDown(), and either implicitly or explicitly gets to
QEMUMachine::shutdown().

I would also like to avoid exposing the user (test writer) to that
kind of situation though, which seems like what you tried to do.  But,
maybe the most honest approach right now is to make a explicit call to
read from and close the socket, paired with documentation/comment about
why that is being done?  This would be similar to the "XXX" comments on
6ab3fc32ea64.

Regards!
- Cleber.