Introducing a new NUMA affinity, FORM2, requires a new mechanism to
switch between affinity modes after CAS. Also, we want FORM2 data
structures and functions to be completely separated from the existing
FORM1 code, allowing us to avoid adding new code that inherits the
existing complexity of FORM1.
The idea of switching values used by the write_dt() functions in
spapr_numa.c was already introduced in the previous patch, and
the same approach will be used when dealing with the FORM1 and FORM2
arrays.
We can accomplish that by that by renaming the existing numa_assoc_array
to FORM1_assoc_array, which now is used exclusively to handle FORM1 affinity
data. A new helper get_associativity() is then introduced to be used by the
write_dt() functions to retrieve the current ibm,associativity array of
a given node, after considering affinity selection that might have been
done during CAS. All code that was using numa_assoc_array now needs to
retrieve the array by calling this function.
This will allow for an easier plug of FORM2 data later on.
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20210920174947.556324-5-danielhb413@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The next preliminary step to introduce NUMA FORM2 affinity is to make
the existing code independent of FORM1 macros and values, i.e.
MAX_DISTANCE_REF_POINTS, NUMA_ASSOC_SIZE and VCPU_ASSOC_SIZE. This patch
accomplishes that by doing the following:
- move the NUMA related macros from spapr.h to spapr_numa.c where they
are used. spapr.h gets instead a 'NUMA_NODES_MAX_NUM' macro that is used
to refer to the maximum number of NUMA nodes, including GPU nodes, that
the machine can support;
- MAX_DISTANCE_REF_POINTS and NUMA_ASSOC_SIZE are renamed to
FORM1_DIST_REF_POINTS and FORM1_NUMA_ASSOC_SIZE. These FORM1 specific
macros are used in FORM1 init functions;
- code that uses MAX_DISTANCE_REF_POINTS now retrieves the
max_dist_ref_points value using get_max_dist_ref_points().
NUMA_ASSOC_SIZE is replaced by get_numa_assoc_size() and VCPU_ASSOC_SIZE
is replaced by get_vcpu_assoc_size(). These functions are used by the
generic device tree functions and h_home_node_associativity() and will
allow them to switch between FORM1 and FORM2 without changing their core
logic.
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20210920174947.556324-4-danielhb413@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
When first introduced, 'legacy_numa' was a way to refer to guests that
either wouldn't be affected by associativity domain calculations, namely
the ones with only 1 NUMA node, and pre 5.2 guests that shouldn't be
affected by it because it would be an userspace change. Calling these
cases 'legacy_numa' was a convenient way to label these cases.
We're about to introduce a new NUMA affinity, FORM2, and this concept
of 'legacy_numa' is now a bit misleading because, although it is called
'legacy' it is in fact a FORM1 exclusive contraint.
This patch removes spapr_machine_using_legacy_numa() and open code the
conditions in each caller. While we're at it, move the chunk inside
spapr_numa_FORM1_affinity_init() that sets all numa_assoc_array domains
with 'node_id' to spapr_numa_define_FORM1_domains(). This chunk was
being executed if !pre_5_2_numa_associativity and num_nodes => 1, the
same conditions in which spapr_numa_define_FORM1_domains() is called
shortly after.
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20210920174947.556324-3-danielhb413@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The upcoming FORM2 NUMA affinity will support asymmetric NUMA topologies
and doesn't need be concerned with all the legacy support for older
pseries FORM1 guests.
We're also not going to calculate associativity domains based on numa
distance (via spapr_numa_define_associativity_domains) since the
distances will be written directly into new DT properties.
Let's split FORM1 code into its own functions to allow for easier
insertion of FORM2 logic later on.
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20210920174947.556324-2-danielhb413@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Add a Host Radix field (hr) in DisasContext with LPCR[HR] value to allow
us to decide between Radix and HPT while validating instructions
arguments. Note that PowerISA v3.1 does not require LPCR[HR] and PATE.HR
to match if the thread is in ultravisor/hypervisor real addressing mode,
so ctx->hr may be invalid if ctx->hv and ctx->dr are set.
Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20210917114751.206845-2-matheus.ferst@eldorado.org.br>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Linux Kernel 5.12 is now unisolating CPU DRCs in the device_removal
error path, signalling that the hotunplug process wasn't successful.
This allow us to send a DEVICE_UNPLUG_GUEST_ERROR in drc_unisolate_logical()
to signal this error to the management layer.
We also have another error path in spapr_memory_unplug_rollback() for
configured LMB DRCs. Kernels older than 5.13 will not unisolate the LMBs
in the hotunplug error path, but it will reconfigure them. Let's send
the DEVICE_UNPLUG_GUEST_ERROR event in that code path as well to cover the
case of older kernels.
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20210907004755.424931-7-danielhb413@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
At this moment we only provide one event to report a hotunplug error,
MEM_UNPLUG_ERROR. As of Linux kernel 5.12 and QEMU 6.0.0, the pseries
machine is now able to report unplug errors for other device types, such
as CPUs.
Instead of creating a (device_type)_UNPLUG_ERROR for each new device,
create a generic DEVICE_UNPLUG_GUEST_ERROR event that can be used by all
guest side unplug errors in the future. This event has a similar API as
the existing DEVICE_DELETED event, always providing the QOM path of the
device and dev->id if there's any.
With this new generic event, MEM_UNPLUG_ERROR is now marked as deprecated.
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20210907004755.424931-6-danielhb413@gmail.com>
[dwg: Correct missing ')' in stubs/qdev.c]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The error_report() call in drc_unisolate_logical() is not considering
that drc->dev->id can be NULL, and the underlying functions error_report()
calls to do its job (vprintf(), g_strdup_printf() ...) has undefined
behavior when trying to handle "%s" with NULL arguments.
Besides, there is no utility into reporting that an unknown device was
rejected by the guest.
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20210907004755.424931-4-danielhb413@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
qapi_event_send_mem_unplug_error() deals with @device being NULL by
replacing it with an empty string ("") when emitting the event. Aside
from the fact that this behavior (qapi visitor mapping NULL pointer to
"") can be patched/changed someday, there's also the lack of utility
that the event brings to listeners, e.g. "a memory unplug error happened
somewhere".
In theory we should just avoit emitting this event at all if dev->id is
NULL, but this would be an incompatible change to existing guests.
Instead, let's make the forementioned behavior explicit: if dev->id is
NULL, pass an empty string to qapi_event_send_mem_unplug_error().
Suggested-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20210907004755.424931-2-danielhb413@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
According to the ISA, CR should be set based on the source value, and
not on the packed decimal result.
The way this was implemented would cause GT, LT and EQ to be set
incorrectly when the source value was too large and the 31 least
significant digits of the packed decimal result ended up being all zero.
This would happen for source values of +/-10^31, +/-10^32, etc.
The new implementation fixes this and also skips the result calculation
altogether in case of src overflow.
Signed-off-by: Luis Pires <luis.pires@eldorado.org.br>
Message-Id: <20210823150235.35759-1-luis.pires@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
This also fixes a small skiboot/skiroot typo and removes the links to
the specific POWER8 and POWER9 images since the firmware images can be
used to run all machines.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20210902130928.528803-2-clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Integration testing patches
- More Linux kernel record/replay tests (Pavel Dovgalyuk)
- Various fixes (Willian Rampazzo, Cleber Rosa)
- Split machine_ppc.py per machine (David Gibson)
- Add AVOCADO_TESTS command line environment variable (Willian Rampazzo)
- Test PowerPC PowerNV 8/9 machines (Cédric Le Goater)
# gpg: Signature made Mon 27 Sep 2021 18:24:03 BST
# gpg: using RSA key FAABE75E12917221DCFD6BB2E3E32C2CDEADC0DE
# gpg: Good signature from "Philippe Mathieu-Daudé (F4BUG) <f4bug@amsat.org>" [full]
# Primary key fingerprint: FAAB E75E 1291 7221 DCFD 6BB2 E3E3 2C2C DEAD C0DE
* remotes/philmd/tags/integration-testing-20210927: (22 commits)
tests/acceptance: Test powernv machines
tests/acceptance/ppc_prep_40p.py: clean up unused import
tests/acceptance/ppc_prep_40p.py: NetBSD 7.1.2 location update
tests/acceptance/boot_xen.py: removed unused import
acceptance/tests/vnc.py: use explicit syntax for enabling passwords
Acceptance Tests: improve check-acceptance description
qemu: Split machine_ppc.py acceptance tests
tests/Makefile: add AVOCADO_TESTS option to make check-acceptance
docs/devel/testing: add instruction to run a single acceptance test
tests/Makefile: allow control over tags during check-acceptance
avocado_qemu: fix inheritance order on LinuxTest class
avocado_qemu: explicitly return None to avoid R1710
avocado_qemu: tweak ssh connect method
avocado_qemu: fix import module based on isort
avocado_qemu: standardize super() call following PEP3135
Acceptance Tests: add standard clean up at test tearDown()
Acceptance tests: add myself as a reviewer for the acceptance tests
tests/acceptance: Linux boot test for record/replay
tests/acceptance: add replay kernel test for alpha
tests/acceptance: add replay kernel test for nios2
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The NetBSD-7.1.2-prep.iso is no longer available on the CDN, but it's
still available in the archive.
Let's update its location so that users without the file on cache can
still fetch it and run the test.
Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210924185506.2542588-15-crosa@redhat.com>
The "check-acceptance" make rule won't necessarily run *all* available
tests, because it employs a filter based on the currently configured
targets. This change in the description of the rule makes that
behavior extra clear.
Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210924185506.2542588-3-crosa@redhat.com>
machine_ppc.py contains tests for 3 different ppc based machine types. It
is listed in MAINTAINERS along with the PPC TCG cpu code. That's not
really accurate though, since it's really more about testing those machines
than the CPUs.
Therefore, split it up into separate files for the separate machine types,
and list those along with their machine types in MAINTAINERS.
Suggested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210927044808.73391-2-david@gibson.dropbear.id.au>
Add the possibility of running all the tests from a single file, or
multiple files, running a single test within a file or multiple tests
within multiple files using `make check-acceptance` and the
AVOCADO_TESTS environment variable.
Suggested-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210923161141.232208-4-willianr@redhat.com>
Although it is possible to run a specific test using the avocado
command-line, a user may want to use a specific tag while running the
``make check-acceptance`` during the development or debugging.
This allows using the AVOCADO_TAGS environment variable where the user
takes total control of which tests should run based on the tags defined.
This also makes the check-acceptance command flexible to restrict tests
based on tags while running on CI.
e.g.:
AVOCADO_TAGS="foo bar baz" make check-acceptance
Signed-off-by: Willian Rampazzo <willianr@redhat.com>
Tested-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210923161141.232208-2-willianr@redhat.com>
The linter is complaining the `pick_default_qemu_bin` is not explicitly
returning None. Fix it to explicitly return None and avoid R1710
inconsistent-return-statements.
Signed-off-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210920204932.94132-6-willianr@redhat.com>
The current implementation will crash if the connection fails as the
`time` module is not imported. Fix the import problem. While here,
tweaks the connection to wait progressively when the connection fails.
Signed-off-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
[PMD: Reworded description]
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210920204932.94132-5-willianr@redhat.com>
PEP3135 states when calling super(), there is no need to use arguments.
This changes the calls on avocado_qemu to standardize according to
PEP3135 and avoid warnings from linters.
Signed-off-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210920204932.94132-3-willianr@redhat.com>
The avocado.Test class, used as the basis of the avocado_qemu.Test
class, performs a clean of temporary directories up as part of its own
tearDown() implementation.
But the avocado_qemu.Test class is currently missing the same clean
up, as it implemented its own tearDown() method without resorting to
the upper class behavior.
This brings avocado_qemu.Test behavior in sync with the standard
avocado.Test behavior and prevents temporary directories from
cluttering the test results directory (unless instructed to do so with
Avocado's "--keep-tmp" option).
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Cleber Rosa <crosa@redhat.com>
[willianr: respin to new Python super format]
Signed-off-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210920204932.94132-2-willianr@redhat.com>
This patch adds a test for record/replay, which boots Linux
image from the disk and interacts with the network.
The idea and code of this test is borrowed from boot_linux.py
This test includes only x86_64 platform. Other platforms and
machines will be added later after testing and improving
record/replay to completely support them.
Each test consists of the following phases:
- downloading the disk image
- recording the execution
- replaying the execution
Replay does not validates the output, but waits until QEMU
finishes the execution. This is reasonable, because
QEMU usually hangs when replay goes wrong.
Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgaluk@ispras.ru>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <162737554047.1735673.13133593401566029378.stgit@pasha-ThinkPad-X280>
This patch adds record/replay test which boots Linux
kernel on alpha platform. The test uses kernel binaries
taken from boot_linux_console test.
Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <162737553482.1735673.10021851966976933952.stgit@pasha-ThinkPad-X280>
This patch adds record/replay test which boots Linux
kernel on nios2 platform. The test uses kernel binaries
taken from boot_linux_console test.
Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <162737552919.1735673.12493523185952280539.stgit@pasha-ThinkPad-X280>
Daniel Henrique Barboza