OSDN Git Service

qmiga/qemu.git
2 years agotarget/loongarch: Add fixed point bit instruction translation
Song Gao [Mon, 6 Jun 2022 12:42:56 +0000 (20:42 +0800)]
target/loongarch: Add fixed point bit instruction translation

This includes:
- EXT.W.{B/H}
- CL{O/Z}.{W/D}, CT{O/Z}.{W/D}
- BYTEPICK.{W/D}
- REVB.{2H/4H/2W/D}
- REVH.{2W/D}
- BITREV.{4B/8B}, BITREV.{W/D}
- BSTRINS.{W/D}, BSTRPICK.{W/D}
- MASKEQZ, MASKNEZ

Signed-off-by: Song Gao <gaosong@loongson.cn>
Signed-off-by: Xiaojuan Yang <yangxiaojuan@loongson.cn>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220606124333.2060567-7-yangxiaojuan@loongson.cn>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotarget/loongarch: Add fixed point shift instruction translation
Song Gao [Mon, 6 Jun 2022 12:42:55 +0000 (20:42 +0800)]
target/loongarch: Add fixed point shift instruction translation

This includes:
- SLL.W, SRL.W, SRA.W, ROTR.W
- SLLI.W, SRLI.W, SRAI.W, ROTRI.W
- SLL.D, SRL.D, SRA.D, ROTR.D
- SLLI.D, SRLI.D, SRAI.D, ROTRI.D

Signed-off-by: Song Gao <gaosong@loongson.cn>
Signed-off-by: Xiaojuan Yang <yangxiaojuan@loongson.cn>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220606124333.2060567-6-yangxiaojuan@loongson.cn>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotarget/loongarch: Add fixed point arithmetic instruction translation
Song Gao [Mon, 6 Jun 2022 12:42:54 +0000 (20:42 +0800)]
target/loongarch: Add fixed point arithmetic instruction translation

This includes:
- ADD.{W/D}, SUB.{W/D}
- ADDI.{W/D}, ADDU16ID
- ALSL.{W[U]/D}
- LU12I.W, LU32I.D LU52I.D
- SLT[U], SLT[U]I
- PCADDI, PCADDU12I, PCADDU18I, PCALAU12I
- AND, OR, NOR, XOR, ANDN, ORN
- MUL.{W/D}, MULH.{W[U]/D[U]}
- MULW.D.W[U]
- DIV.{W[U]/D[U]}, MOD.{W[U]/D[U]}
- ANDI, ORI, XORI

Signed-off-by: Song Gao <gaosong@loongson.cn>
Signed-off-by: Xiaojuan Yang <yangxiaojuan@loongson.cn>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220606124333.2060567-5-yangxiaojuan@loongson.cn>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotarget/loongarch: Add main translation routines
Song Gao [Mon, 6 Jun 2022 12:42:53 +0000 (20:42 +0800)]
target/loongarch: Add main translation routines

This patch adds main translation routines and
basic functions for translation.

Signed-off-by: Song Gao <gaosong@loongson.cn>
Signed-off-by: Xiaojuan Yang <yangxiaojuan@loongson.cn>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220606124333.2060567-4-yangxiaojuan@loongson.cn>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotarget/loongarch: Add core definition
Song Gao [Mon, 6 Jun 2022 12:42:52 +0000 (20:42 +0800)]
target/loongarch: Add core definition

This patch adds target state header, target definitions
and initialization routines.

Signed-off-by: Song Gao <gaosong@loongson.cn>
Signed-off-by: Xiaojuan Yang <yangxiaojuan@loongson.cn>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220606124333.2060567-3-yangxiaojuan@loongson.cn>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotarget/loongarch: Add README
Song Gao [Mon, 6 Jun 2022 12:42:51 +0000 (20:42 +0800)]
target/loongarch: Add README

This patch gives an introduction to the LoongArch target.

Signed-off-by: Song Gao <gaosong@loongson.cn>
Signed-off-by: Xiaojuan Yang <yangxiaojuan@loongson.cn>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220606124333.2060567-2-yangxiaojuan@loongson.cn>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agoMerge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging
Richard Henderson [Mon, 6 Jun 2022 14:57:14 +0000 (07:57 -0700)]
Merge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging

* prepare to expand usage of test venv
* fix CPUID when passing through host cache information
* a20 fix
* SGX fix
* generate per-target modinfo
* replay cleanups and simplifications
* "make modules" target

# -----BEGIN PGP SIGNATURE-----
#
# iQFIBAABCAAyFiEE8TM4V0tmI4mGbHaCv/vSX3jHroMFAmKeCV0UHHBib256aW5p
# QHJlZGhhdC5jb20ACgkQv/vSX3jHroNVWQgArVaF9Q/RWd1jJ2EUoWhy/GKsrdBs
# CHmSzrYzF/0cqD4OtrjTRyUP72Kywf4/aP4GYtLsjSYRUyriqS4nm5yEkDsME+pH
# 0adATIumtGHBPjVJyPUux3H1cbPUmsxbE03ZIWUK65M/IM2m46hmPDTWVx6ZSXMN
# VnSQdTUeZdxY26D8zfLp+oufqmfgaPT19zURcMZlvWFJ830t+1spfZwH0RGjDlfZ
# ffNVMPcco1qreNFggNIV9sKtoLb8A9PxpFJPV9OtBONA5UGxZ7xhjjZY9jCI4Bxa
# xAd60nXeoAwiOP7LE2bEwXbPku3xgy2e0LNXhqblMGNGOjdlRlwfL/FhFw==
# =C9N3
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 06 Jun 2022 07:04:13 AM PDT
# gpg:                using RSA key F13338574B662389866C7682BFFBD25F78C7AE83
# gpg:                issuer "pbonzini@redhat.com"
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [undefined]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [undefined]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* tag 'for-upstream' of https://gitlab.com/bonzini/qemu: (29 commits)
  meson: qga: do not use deprecated meson.build_root()
  configure: remove reference to removed option
  regenerate meson-buildoptions.sh
  tests: run 'device-crash-test' from tests/venv
  tests: add python3-venv to debian10.docker
  tests: use tests/venv to run basevm.py-based scripts
  tests: install "qemu" namespace package into venv
  tests: add quiet-venv-pip macro
  tests: silence pip upgrade warnings during venv creation
  tests: use python3 as the python executable name
  tests: add "TESTS_PYTHON" variable to Makefile
  python: update for mypy 0.950
  x86: cpu: fixup number of addressable IDs for logical processors sharing cache
  x86: cpu: make sure number of addressable IDs for processor cores meets the spec
  tests/Makefile.include: Fix 'make check-help' output
  tests/avocado: add replay Linux test for Aarch64 machines
  tests/avocado: add replay Linux tests for virtio machine
  tests/avocado: update replay_linux test
  docs: move replay docs to docs/system/replay.rst
  docs: convert docs/devel/replay page to rst
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agomeson: qga: do not use deprecated meson.build_root()
Paolo Bonzini [Mon, 6 Jun 2022 14:03:38 +0000 (16:03 +0200)]
meson: qga: do not use deprecated meson.build_root()

The function will return the build root of the parent project if called from a
subproject; that is irrelevant for QEMU's usage but rarely desirable, and
therefore the function was deprecated and replaced by two functions
project_build_root() and global_build_root().  Replace it with the former.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoconfigure: remove reference to removed option
Paolo Bonzini [Mon, 6 Jun 2022 10:44:57 +0000 (12:44 +0200)]
configure: remove reference to removed option

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoregenerate meson-buildoptions.sh
Paolo Bonzini [Mon, 6 Jun 2022 10:44:45 +0000 (12:44 +0200)]
regenerate meson-buildoptions.sh

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests: run 'device-crash-test' from tests/venv
John Snow [Thu, 26 May 2022 00:09:21 +0000 (20:09 -0400)]
tests: run 'device-crash-test' from tests/venv

Remove the sys.path hacking from device-crash-test, and add in a little
user-friendly message for anyone who was used to running this script
directly from the source tree.

Modify the GitLab job recipes to create the tests/venv first, then run
device-crash-test from that venv.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220526000921.1581503-10-jsnow@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests: add python3-venv to debian10.docker
John Snow [Thu, 26 May 2022 00:09:20 +0000 (20:09 -0400)]
tests: add python3-venv to debian10.docker

This is needed to be able to add a venv-building step to 'make check';
the clang-user job in particular needs this to be able to run
check-unit.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220526000921.1581503-9-jsnow@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests: use tests/venv to run basevm.py-based scripts
John Snow [Thu, 26 May 2022 00:09:19 +0000 (20:09 -0400)]
tests: use tests/venv to run basevm.py-based scripts

This patch co-opts the virtual environment being used by avocado tests
to also run the basevm.py tests. This is being done in preparation for
for the qemu.qmp package being removed from qemu.git.

As part of the change, remove any sys.path() hacks and treat "qemu" as a
normal third-party import.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220526000921.1581503-8-jsnow@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests: install "qemu" namespace package into venv
John Snow [Thu, 26 May 2022 00:09:18 +0000 (20:09 -0400)]
tests: install "qemu" namespace package into venv

This patch adds the "qemu" namespace package to the $build/tests/venv
directory. It does so in "editable" mode, which means that changes to
the source python directory will actively be reflected by the venv.

This patch also then removes any sys.path hacking from the avocado test
scripts directly. By doing this, the environment of where to find these
packages is managed entirely by the virtual environment and not by the
scripts themselves.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220526000921.1581503-7-jsnow@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests: add quiet-venv-pip macro
John Snow [Thu, 26 May 2022 00:09:17 +0000 (20:09 -0400)]
tests: add quiet-venv-pip macro

Factor out the "test venv pip" macro; rewrite the "check-venv" rule to
be a little more compact. Replace the "PIP" pseudo-command output with
"VENVPIP" to make it 1% more clear that we are talking about using pip
to install something into a venv.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220526000921.1581503-6-jsnow@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests: silence pip upgrade warnings during venv creation
John Snow [Thu, 26 May 2022 00:09:16 +0000 (20:09 -0400)]
tests: silence pip upgrade warnings during venv creation

Turn off the nag warning coaxing us to upgrade pip. It's not really that
interesting to see in CI logs, and as long as nothing is broken --
nothing is broken.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220526000921.1581503-5-jsnow@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests: use python3 as the python executable name
John Snow [Thu, 26 May 2022 00:09:15 +0000 (20:09 -0400)]
tests: use python3 as the python executable name

Use "python3" instead of "python" as per PEP0394:
https://peps.python.org/pep-0394/

This should always be defined (in a venv, at least!), matching the
preferred python shebang of "#!/usr/bin/env python3".

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220526000921.1581503-4-jsnow@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests: add "TESTS_PYTHON" variable to Makefile
John Snow [Thu, 26 May 2022 00:09:14 +0000 (20:09 -0400)]
tests: add "TESTS_PYTHON" variable to Makefile

This is a convenience feature: $(PYTHON) points to the Python executable
we were instructed to use by the configure script. We use that Python to
create a virtual environment with the "check-venv" target in
tests/Makefile.include.

$(TESTS_PYTHON) points to the Python executable belonging to the virtual
environment tied to the build. This Python executable is a symlink to
the binary used to create the venv, which will be the version provided
at configure time.

Using $(TESTS_PYTHON) therefore uses the $(PYTHON) executable, but with
paths modified to use packages installed to the venv.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220526000921.1581503-3-jsnow@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agopython: update for mypy 0.950
John Snow [Thu, 26 May 2022 00:09:13 +0000 (20:09 -0400)]
python: update for mypy 0.950

typeshed (included in mypy) recently updated to improve the typing for
WriteTransport objects. I was working around this, but now there's a
version where I shouldn't work around it.

Unfortunately this creates some minor ugliness if I want to support both
pre- and post-0.950 versions. For now, for my sanity, just disable the
unused-ignores warning.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220526000921.1581503-2-jsnow@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agox86: cpu: fixup number of addressable IDs for logical processors sharing cache
Igor Mammedov [Tue, 24 May 2022 15:10:20 +0000 (11:10 -0400)]
x86: cpu: fixup number of addressable IDs for logical processors sharing cache

When QEMU is started with '-cpu host,host-cache-info=on', it will
passthrough host's number of logical processors sharing cache and
number of processor cores in the physical package. QEMU already
fixes up the later to correctly reflect number of configured cores
for VM, however number of logical processors sharing cache is still
comes from host CPU, which confuses guest started with:

       -machine q35,accel=kvm \
       -cpu host,host-cache-info=on,l3-cache=off \
       -smp 20,sockets=2,dies=1,cores=10,threads=1  \
       -numa node,nodeid=0,memdev=ram-node0 \
       -numa node,nodeid=1,memdev=ram-node1 \
       -numa cpu,socket-id=0,node-id=0 \
       -numa cpu,socket-id=1,node-id=1

on 2 socket Xeon 4210R host with 10 cores per socket
with CPUID[04H]:
      ...
        --- cache 3 ---
      cache type                           = unified cache (3)
      cache level                          = 0x3 (3)
      self-initializing cache level        = true
      fully associative cache              = false
      maximum IDs for CPUs sharing cache   = 0x1f (31)
      maximum IDs for cores in pkg         = 0xf (15)
      ...
that doesn't match number of logical processors VM was
configured with and as result RHEL 9.0 guest complains:

   sched: CPU #10's llc-sibling CPU #0 is not on the same node! [node: 1 != 0]. Ignoring dependency.
   WARNING: CPU: 10 PID: 0 at arch/x86/kernel/smpboot.c:421 topology_sane.isra.0+0x67/0x80
   ...
   Call Trace:
     set_cpu_sibling_map+0x176/0x590
     start_secondary+0x5b/0x150
     secondary_startup_64_no_verify+0xc2/0xcb

Fix it by capping max number of logical processors to vcpus/socket
as it was configured, which fixes the issue.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2088311
Message-Id: <20220524151020.2541698-3-imammedo@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agox86: cpu: make sure number of addressable IDs for processor cores meets the spec
Igor Mammedov [Tue, 24 May 2022 15:10:19 +0000 (11:10 -0400)]
x86: cpu: make sure number of addressable IDs for processor cores meets the spec

Accourding Intel's CPUID[EAX=04H] resulting bits 31 - 26 in EAX
should be:
"
 **** The nearest power-of-2 integer that is not smaller than (1 + EAX[31:26]) is the number of unique
    Core_IDs reserved for addressing different processor cores in a physical package. Core ID is a subset of
    bits of the initial APIC ID.
"

ensure that values stored in EAX[31-26] always meets this condition.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20220524151020.2541698-2-imammedo@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests/Makefile.include: Fix 'make check-help' output
Dario Faggioli [Fri, 27 May 2022 15:30:54 +0000 (17:30 +0200)]
tests/Makefile.include: Fix 'make check-help' output

Since commit 3d2f73ef75e ("build: use "meson test" as the test harness"),
check-report.tap is no more, and we have check-report.junit.xml.

Update the output of 'make check-help', which was still listing
'check-report.tap', accordingly.

Fixes: 3d2f73ef75e
Signed-off-by: Dario Faggioli <dfaggioli@suse.com>
Message-Id: <165366545439.6869.11633009118019728798.stgit@work>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests/avocado: add replay Linux test for Aarch64 machines
Pavel Dovgalyuk [Fri, 27 May 2022 10:46:53 +0000 (13:46 +0300)]
tests/avocado: add replay Linux test for Aarch64 machines

This patch adds two tests for replaying Linux boot process
on Aarch64 platform.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Message-Id: <165364841373.688121.8868079200312201658.stgit@pasha-ThinkPad-X280>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests/avocado: add replay Linux tests for virtio machine
Pavel Dovgalyuk [Fri, 27 May 2022 10:46:48 +0000 (13:46 +0300)]
tests/avocado: add replay Linux tests for virtio machine

This patch adds two tests for replaying Linux boot process
on x86_64 virtio platform.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Message-Id: <165364840811.688121.11931681195199516354.stgit@pasha-ThinkPad-X280>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotests/avocado: update replay_linux test
Pavel Dovgalyuk [Fri, 27 May 2022 10:46:42 +0000 (13:46 +0300)]
tests/avocado: update replay_linux test

This patch updates replay_linux test to make it compatible with
new LinuxTest class.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Message-Id: <165364840253.688121.10404266209986316381.stgit@pasha-ThinkPad-X280>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agodocs: move replay docs to docs/system/replay.rst
Pavel Dovgalyuk [Fri, 27 May 2022 10:46:36 +0000 (13:46 +0300)]
docs: move replay docs to docs/system/replay.rst

This patch adds replay description page, converting prior
text from docs/replay.txt.
The text was also updated and some sections were moved
to devel part of the docs.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <165364839601.688121.5131456980322853233.stgit@pasha-ThinkPad-X280>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agodocs: convert docs/devel/replay page to rst
Pavel Dovgalyuk [Fri, 27 May 2022 10:46:30 +0000 (13:46 +0300)]
docs: convert docs/devel/replay page to rst

This patch converts prior .txt replay devel documentation to .rst.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <165364839013.688121.11935249420738873044.stgit@pasha-ThinkPad-X280>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoreplay: simplify async event processing
Pavel Dovgalyuk [Fri, 27 May 2022 10:46:23 +0000 (13:46 +0300)]
replay: simplify async event processing

This patch joins replay event id and async event id into single byte in the log.
It makes processing a bit faster and log a bit smaller.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
--

v2: minor enum fixes (suggested by Richard Henderson)
Message-Id: <165364838393.688121.8191379555130516329.stgit@pasha-ThinkPad-X280>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoreplay: rewrite async event handling
Pavel Dovgalyuk [Fri, 27 May 2022 10:46:18 +0000 (13:46 +0300)]
replay: rewrite async event handling

This patch decouples checkpoints and async events.
It was a tricky part of replay implementation. Now it becomes
much simpler and easier to maintain.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <165364837856.688121.8785039478408995979.stgit@pasha-ThinkPad-X280>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoreplay: notify vCPU when BH is scheduled
Pavel Dovgalyuk [Fri, 27 May 2022 10:46:13 +0000 (13:46 +0300)]
replay: notify vCPU when BH is scheduled

vCPU execution should be suspended when new BH is scheduled.
This is needed to avoid guest timeouts caused by the long cycles
of the execution. In replay mode execution may hang when
vCPU sleeps and block event comes to the queue.
This patch adds notification which wakes up vCPU or interrupts
execution of guest code.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
--

v2: changed first_cpu to current_cpu (suggested by Richard Henderson)
v4: moved vCPU notification to aio_bh_enqueue (suggested by Paolo Bonzini)
Message-Id: <165364837317.688121.17680519919871405281.stgit@pasha-ThinkPad-X280>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoreplay: fix event queue flush for qemu shutdown
Pavel Dovgalyuk [Fri, 27 May 2022 10:46:07 +0000 (13:46 +0300)]
replay: fix event queue flush for qemu shutdown

This patch fixes event queue flush in the case of emulator
shutdown. replay_finish_events should be called when replay_mode
is not cleared.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <165364836758.688121.7959245442743676491.stgit@pasha-ThinkPad-X280>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agomodules: generates per-target modinfo
Jose R. Ziviani [Fri, 27 May 2022 22:20:35 +0000 (00:20 +0200)]
modules: generates per-target modinfo

This patch changes the way modinfo is generated and built. Instead of
one modinfo.c it generates one modinfo-<target>-softmmu.c per target. It
aims a fine-tune control of modules by configuring Kconfig.

Signed-off-by: Jose R. Ziviani <jziviani@suse.de>
Signed-off-by: Dario Faggioli <dfaggioli@suse.com>
Message-Id: <165369003038.5857.13084289285185196779.stgit@work>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agomodules: introduces module_kconfig directive
Jose R. Ziviani [Fri, 27 May 2022 22:20:23 +0000 (00:20 +0200)]
modules: introduces module_kconfig directive

module_kconfig is a new directive that should be used with module_obj
whenever that module depends on the Kconfig to be enabled.

When the module is enabled in Kconfig we are sure that its dependencies
will be enabled as well, thus the module will be loaded without any
problem.

The correct way to use module_kconfig is by passing the Kconfig option
to module_kconfig (or the *config-devices.mak without CONFIG_).

Signed-off-by: Jose R. Ziviani <jziviani@suse.de>
Signed-off-by: Dario Faggioli <dfaggioli@suse.com>
Message-Id: <165369002370.5857.12150544416563557322.stgit@work>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotarget/i386: Fix wrong count setting
Yang Zhong [Mon, 30 May 2022 13:18:34 +0000 (09:18 -0400)]
target/i386: Fix wrong count setting

The previous patch used wrong count setting with index value, which got wrong
value from CPUID(EAX=12,ECX=0):EAX. So the SGX1 instruction can't be exposed
to VM and the SGX decice can't work in VM.

Fixes: d19d6ffa0710 ("target/i386: introduce helper to access supported CPUID")

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Message-Id: <20220530131834.1222801-1-yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agobuild: add a "make modules" target
Paolo Bonzini [Mon, 30 May 2022 12:24:34 +0000 (14:24 +0200)]
build: add a "make modules" target

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agotarget/i386/tcg: Fix masking of real-mode addresses with A20 bit
Stephen Michael Jothen [Wed, 25 May 2022 15:33:36 +0000 (17:33 +0200)]
target/i386/tcg: Fix masking of real-mode addresses with A20 bit

The correct A20 masking is done if paging is enabled (protected mode) but it
seems to have been forgotten in real mode. For example from the AMD64 APM Vol. 2
section 1.2.4:

> If the sum of the segment base and effective address carries over into bit 20,
> that bit can be optionally truncated to mimic the 20-bit address wrapping of the
> 8086 processor by using the A20M# input signal to mask the A20 address bit.

Most BIOSes will enable the A20 line on boot, but I found by disabling the A20 line
afterwards, the correct wrapping wasn't taking place.

`handle_mmu_fault' in target/i386/tcg/sysemu/excp_helper.c seems to be the culprit.
In real mode, it fills the TLB with the raw unmasked address. However, for the
protected mode, the `mmu_translate' function does the correct A20 masking.

The fix then should be to just apply the A20 mask in the first branch of the if
statement.

Signed-off-by: Stephen Michael Jothen <sjothen@gmail.com>
Message-Id: <Yo5MUMSz80jXtvt9@air-old.local>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoMerge tag 'nvme-next-pull-request' of git://git.infradead.org/qemu-nvme into staging
Richard Henderson [Fri, 3 Jun 2022 21:14:24 +0000 (14:14 -0700)]
Merge tag 'nvme-next-pull-request' of git://git.infradead.org/qemu-nvme into staging

hw/nvme updates

# -----BEGIN PGP SIGNATURE-----
#
# iQEzBAABCAAdFiEEUigzqnXi3OaiR2bATeGvMW1PDekFAmKaZmgACgkQTeGvMW1P
# DenI7wgAxY4QtRlUnufzaZqcoi+affFTKlKm0JYKZm/Ldxt2RtHoWxRZDLLIUp8B
# 4XAlIGJw7VwrafEtSkx4K6cSyKluMJ9Ax8pNd03sEweXBBfdhNizspPprp+Jm9P9
# hRcH8kSiBp5B451cORBlgmoHguWeWawe1r66uFLTCbEMtfQQNaxNVsTsgAsOvtwv
# XsjLVFVKGNDWXGRta+lzu4seNNuzfucsAmKWUjg5HN38rstY7XxfLVMzt8ORcwjk
# oNmQuy3JiKujdPVhE5PVgNRZkigwoDt3hDA1QTncGTBUoA/CtaB5SK9EhcJ5xJVI
# EHv99S9LQ8ng5BJC2pUSU32yRkaNOQ==
# =XTXH
# -----END PGP SIGNATURE-----
# gpg: Signature made Fri 03 Jun 2022 12:52:08 PM PDT
# gpg:                using RSA key 522833AA75E2DCE6A24766C04DE1AF316D4F0DE9
# gpg: Good signature from "Klaus Jensen <its@irrelevant.dk>" [unknown]
# gpg:                 aka "Klaus Jensen <k.jensen@samsung.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: DDCA 4D9C 9EF9 31CC 3468  4272 63D5 6FC5 E55D A838
#      Subkey fingerprint: 5228 33AA 75E2 DCE6 A247  66C0 4DE1 AF31 6D4F 0DE9

* tag 'nvme-next-pull-request' of git://git.infradead.org/qemu-nvme:
  hw/nvme: add new command abort case
  hw/nvme: deprecate the use-intel-id compatibility parameter
  hw/nvme: bump firmware revision
  hw/nvme: do not report null uuid
  hw/nvme: do not auto-generate uuid
  hw/nvme: do not auto-generate eui64
  hw/nvme: enforce common serial per subsystem
  hw/nvme: fix smart aen
  hw/nvme: fix copy cmd for pi enabled namespaces
  hw/nvme: add missing return statement
  hw/nvme: fix narrowing conversion

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agohw/nvme: add new command abort case
Dmitry Tikhov [Wed, 20 Apr 2022 08:20:44 +0000 (11:20 +0300)]
hw/nvme: add new command abort case

NVMe command set specification for end-to-end data protection formatted
namespace states:

    o If the Reference Tag Check bit of the PRCHK field is set to ‘1’ and
      the namespace is formatted for Type 3 protection, then the
      controller:
          ▪ should not compare the protection Information Reference Tag
            field to the computed reference tag; and
          ▪ may ignore the ILBRT and EILBRT fields. If a command is
            aborted as a result of the Reference Tag Check bit of the
            PRCHK field being set to ‘1’, then that command should be
            aborted with a status code of Invalid Protection Information,
            but may be aborted with a status code of Invalid Field in
            Command.

Currently qemu compares reftag in the nvme_dif_prchk function whenever
Reference Tag Check bit is set in the command. For type 3 namespaces
however, caller of nvme_dif_prchk - nvme_dif_check does not increment
reftag for each subsequent logical block. That way commands incorporating
more than one logical block for type 3 formatted namespaces with reftag
check bit set, always fail with End-to-end Reference Tag Check Error.
Comply with spec by handling case of set Reference Tag Check
bit in the type 3 formatted namespace.

Fixes: 146f720c5563 ("hw/block/nvme: end-to-end data protection")
Signed-off-by: Dmitry Tikhov <d.tihov@yadro.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: deprecate the use-intel-id compatibility parameter
Klaus Jensen [Tue, 19 Apr 2022 11:24:23 +0000 (13:24 +0200)]
hw/nvme: deprecate the use-intel-id compatibility parameter

Since version 5.2 commit 6eb7a071292a ("hw/block/nvme: change controller
pci id"), the emulated NVMe controller has defaulted to a non-Intel PCI
identifier.

Deprecate the compatibility parameter so we can get rid of it once and
for all.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: bump firmware revision
Klaus Jensen [Fri, 29 Apr 2022 08:33:36 +0000 (10:33 +0200)]
hw/nvme: bump firmware revision

The Linux kernel quirks the QEMU NVMe controller pretty heavily because
of the namespace identifier mess. Since this is now fixed, bump the
firmware revision number to allow the quirk to be disabled for this
revision.

As of now, bump the firmware revision number to be equal to the QEMU
release version number.

Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: do not report null uuid
Klaus Jensen [Fri, 29 Apr 2022 08:33:35 +0000 (10:33 +0200)]
hw/nvme: do not report null uuid

Do not report the "null uuid" (all zeros) in the namespace
identification descriptors.

Reported-by: Luis Chamberlain <mcgrof@kernel.org>
Reported-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: do not auto-generate uuid
Klaus Jensen [Fri, 29 Apr 2022 08:33:34 +0000 (10:33 +0200)]
hw/nvme: do not auto-generate uuid

Do not default to generate an UUID for namespaces if it is not
explicitly specified.

This is a technically a breaking change in behavior. However, since the
UUID changes on every VM launch, it is not spec compliant and is of
little use since the UUID cannot be used reliably anyway and the
behavior prior to this patch must be considered buggy.

Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: do not auto-generate eui64
Klaus Jensen [Fri, 29 Apr 2022 08:33:33 +0000 (10:33 +0200)]
hw/nvme: do not auto-generate eui64

We cannot provide auto-generated unique or persistent namespace
identifiers (EUI64, NGUID, UUID) easily. Since 6.1, namespaces have been
assigned a generated EUI64 of the form "52:54:00:<namespace counter>".
This is will be unique within a QEMU instance, but not globally.

Revert that this is assigned automatically and immediately deprecate the
compatibility parameter. Users can opt-in to this with the
`eui64-default=on` device parameter or set it explicitly with
`eui64=UINT64`.

Cc: libvir-list@redhat.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: enforce common serial per subsystem
Klaus Jensen [Fri, 29 Apr 2022 08:33:32 +0000 (10:33 +0200)]
hw/nvme: enforce common serial per subsystem

The Identify Controller Serial Number (SN) is the serial number for the
NVM subsystem and must be the same across all controller in the NVM
subsystem.

Enforce this.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: fix smart aen
Klaus Jensen [Thu, 5 May 2022 22:21:47 +0000 (00:21 +0200)]
hw/nvme: fix smart aen

Pass the right constant to nvme_smart_event(). The NVME_AER* values hold
the bit position in the SMART byte, not the shifted value that we expect
it to be in nvme_smart_event().

Fixes: c62720f137df ("hw/block/nvme: trigger async event during injecting smart warning")
Acked-by: zhenwei pi <pizhenwei@bytedance.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: fix copy cmd for pi enabled namespaces
Dmitry Tikhov [Thu, 21 Apr 2022 10:51:58 +0000 (13:51 +0300)]
hw/nvme: fix copy cmd for pi enabled namespaces

Current implementation have problem in the read part of copy command.
Because there is no metadata mangling before nvme_dif_check invocation,
reftag error could be thrown for blocks of namespace that have not been
previously written to.

Signed-off-by: Dmitry Tikhov <d.tihov@yadro.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: add missing return statement
Dmitry Tikhov [Fri, 15 Apr 2022 20:48:32 +0000 (23:48 +0300)]
hw/nvme: add missing return statement

Since there is no return after nvme_dsm_cb invocation, metadata
associated with non-zero block range is currently zeroed. Also this
behaviour leads to segfault since we schedule iocb->bh two times.
First when entering nvme_dsm_cb with iocb->idx == iocb->nr and
second because of missing return on call stack unwinding by calling
blk_aio_pwrite_zeroes and subsequent nvme_dsm_cb callback.

Fixes: d7d1474fd85d ("hw/nvme: reimplement dsm to allow cancellation")
Signed-off-by: Dmitry Tikhov <d.tihov@yadro.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agohw/nvme: fix narrowing conversion
Dmitry Tikhov [Tue, 12 Apr 2022 08:59:09 +0000 (11:59 +0300)]
hw/nvme: fix narrowing conversion

Since nlbas is of type int, it does not work with large namespace size
values, e.g., 9 TB size of file backing namespace and 8 byte metadata
with 4096 bytes lbasz gives negative nlbas value, which is later
promoted to negative int64_t type value and results in negative
ns->moff which breaks namespace

Signed-off-by: Dmitry Tikhov <ddtikhov@gmail.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
2 years agoMerge tag 'pull-request-2022-06-03' of https://gitlab.com/thuth/qemu into staging
Richard Henderson [Fri, 3 Jun 2022 13:43:38 +0000 (06:43 -0700)]
Merge tag 'pull-request-2022-06-03' of https://gitlab.com/thuth/qemu into staging

* s390x storage key improvements for KVM
* Some cosmetics for s390x
* Update MAINTAINERS entries
* Improve some spots wrt memory handling in the qtests
* Clean up the "-display sdl" parameter parsing

# -----BEGIN PGP SIGNATURE-----
#
# iQJFBAABCAAvFiEEJ7iIR+7gJQEY8+q5LtnXdP5wLbUFAmKZpQURHHRodXRoQHJl
# ZGhhdC5jb20ACgkQLtnXdP5wLbXQ1Q//TVyCnWe+kAwwMljS9tebnVwZ1C3wrH68
# KJxeeWNxIdCLMJetZa3V8lYkLuLKojELi24+qAFH+//RWoEIBFvhVlZffWL/7NHw
# KAYy8ZWHeGui29hYgAxlKFC8GSCdsGD+PJU6E0C7wN7R6UVCvPfxPk+MQXjbt/Tn
# UGgb0Ysp4MxEKk3i2t3DR49n7KniUD3ZGZimGhAWfPwe+tPP52Ijczheenv2/R4J
# ip0bc+xo2zUwa6VPBdncnZtOhyXTCNeTxsJcnxnjFVmo9scP2aFOFrV43PvLs0E7
# l2yhQKwXbqIMyaJ+/aFULuOOLL3w2fdOu6TzsMlWiRFF5UmbJjj/8V7Vq1GELupZ
# 2QU0wSJMlNPteRlHUMD0Tz8nG/r/ffYtOAoK3eZAG0tIWV/3xcTIGt0TZuV1VgzY
# M8nNhQTdweCtx5y4sbncEO6qmF5F8u8hbZ7DZCFRCKYkLkC3sv+QDiF1Ys9Iu1MM
# wy7uLcrQZ9rPZ96x1aPKgYMHtrV98yN5om6/CnFVXKvAYX+pAp2of9iTSEAWhRKi
# sDktJ39uCdkdgxnNKBQe6g0cug6mlnpoCM9PSR0zuXuTjFTVZxJBgw7p6Uyofccv
# jL8T6/uR3xQMJ/VDuo0UZLYa4bYnzCc/gJmarLiwBxhykzHAthobBwaXSsLTZ+9J
# 1DpfGpSKDDQ=
# =3Rhl
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 02 Jun 2022 11:07:01 PM PDT
# gpg:                using RSA key 27B88847EEE0250118F3EAB92ED9D774FE702DB5
# gpg:                issuer "thuth@redhat.com"
# gpg: Good signature from "Thomas Huth <th.huth@gmx.de>" [undefined]
# gpg:                 aka "Thomas Huth <thuth@redhat.com>" [undefined]
# gpg:                 aka "Thomas Huth <th.huth@posteo.de>" [unknown]
# gpg:                 aka "Thomas Huth <huth@tuxfamily.org>" [undefined]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 27B8 8847 EEE0 2501 18F3  EAB9 2ED9 D774 FE70 2DB5

* tag 'pull-request-2022-06-03' of https://gitlab.com/thuth/qemu:
  ui: Remove deprecated options "-sdl" and "-curses"
  ui: Switch "-display sdl" to use the QAPI parser
  ui: Remove deprecated parameters of the "-display sdl" option
  tests/qtest: use g_autofree for test_server_create_chr
  qtest/npcm7xx_pwm-test: Fix memory leak in mft_qom_set
  MAINTAINERS: Update maintainers for Guest x86 HAXM CPUs
  MAINTAINERS: Change my email address
  tests/tcg: Test overflow conditions
  MAINTAINERS: Update s390 vhost entries
  target/s390x: kvm: Honor storage keys during emulation
  hw/s390x/s390-virtio-ccw: Improve the machine description string
  s390: Typo fix FLOATING_POINT_SUPPPORT_ENH

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agoui: Remove deprecated options "-sdl" and "-curses"
Thomas Huth [Thu, 19 May 2022 15:56:25 +0000 (17:56 +0200)]
ui: Remove deprecated options "-sdl" and "-curses"

We have "-sdl" and "-curses", but no "-gtk" and no "-cocoa" ...
these old-style options are rather confusing than helpful nowadays.
Now that the deprecation period is over, let's remove them, so we
get a cleaner interface (where "-display" is the only way to select
the user interface).

Message-Id: <20220519155625.1414365-4-thuth@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agoui: Switch "-display sdl" to use the QAPI parser
Thomas Huth [Thu, 19 May 2022 15:56:24 +0000 (17:56 +0200)]
ui: Switch "-display sdl" to use the QAPI parser

The "-display sdl" option still uses a hand-crafted parser for its
parameters since we didn't want to drag an interface we considered
somewhat flawed into the QAPI schema. Since the flaws are gone now,
it's time to QAPIfy.

This introduces the new "DisplaySDL" QAPI struct that is used to hold
the parameters that are unique to the SDL display. The only specific
parameter is currently "grab-mod" that is used to specify the required
modifier keys to escape from the mouse grabbing mode.

Message-Id: <20220519155625.1414365-3-thuth@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agoui: Remove deprecated parameters of the "-display sdl" option
Thomas Huth [Thu, 19 May 2022 15:56:23 +0000 (17:56 +0200)]
ui: Remove deprecated parameters of the "-display sdl" option

Dropping these deprecated parameters simplifies further refactoring
(e.g. QAPIfication is easier without underscores in the name).

Message-Id: <20220519155625.1414365-2-thuth@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agotests/qtest: use g_autofree for test_server_create_chr
Alex Bennée [Tue, 24 May 2022 15:40:52 +0000 (16:40 +0100)]
tests/qtest: use g_autofree for test_server_create_chr

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220524154056.2896913-12-alex.bennee@linaro.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agoqtest/npcm7xx_pwm-test: Fix memory leak in mft_qom_set
Miaoqian Lin [Tue, 31 May 2022 08:09:21 +0000 (12:09 +0400)]
qtest/npcm7xx_pwm-test: Fix memory leak in mft_qom_set

g_strdup_printf() allocated memory for path, we should free it with
g_free() when no longer needed.

Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Reviewed-by: Hao Wu <wuhaotsh@google.com>
Message-Id: <20220531080921.4704-1-linmq006@gmail.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agoMAINTAINERS: Update maintainers for Guest x86 HAXM CPUs
Wenchao Wang [Tue, 15 Mar 2022 12:51:22 +0000 (20:51 +0800)]
MAINTAINERS: Update maintainers for Guest x86 HAXM CPUs

Clean up the maintainer list.

Reviewed-by: Hang Yuan <hang.yuan@intel.com>
Signed-off-by: Wenchao Wang <wenchao.wang@intel.com>
Message-Id: <DM6PR11MB4090A58ACCA4AD8C752AEEA587199@DM6PR11MB4090.namprd11.prod.outlook.com>
[thuth: Note: Colin Xu's address bounces]
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agoMAINTAINERS: Change my email address
Hailiang Zhang [Tue, 14 Dec 2021 07:54:24 +0000 (15:54 +0800)]
MAINTAINERS: Change my email address

The zhang.zhanghailiang@huawei.com email address has been
stopped. Change it to my new email address.

Signed-off-by: Hailiang Zhang <zhanghailiang@xfusion.com>
Message-Id: <20211214075424.6920-1-zhanghailiang@xfusion.com>
Acked-by: Gonglei <arei.gonglei@huawei.com>
Acked-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agotests/tcg: Test overflow conditions
Gautam Agrawal [Tue, 31 May 2022 18:35:24 +0000 (00:05 +0530)]
tests/tcg: Test overflow conditions

Add a test to check for overflow conditions in s390x.
This patch is based on the following patches :
* https://git.qemu.org/?p=qemu.git;a=commitdiff;h=5a2e67a691501
* https://git.qemu.org/?p=qemu.git;a=commitdiff;h=fc6e0d0f2db51

Signed-off-by: Gautam Agrawal <gautamnagrawal@gmail.com>
Message-Id: <20220531183524.40948-1-gautamnagrawal@gmail.com>
[thuth: Move overflow.c to tests/tcg/multiarch/ to make it generic]
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agoMAINTAINERS: Update s390 vhost entries
Eric Farman [Wed, 25 May 2022 14:58:14 +0000 (16:58 +0200)]
MAINTAINERS: Update s390 vhost entries

Commit 7a523d96a0 ("virtio-ccw: move vhost_ccw_scsi to a separate file")
introduced a new file hw/s390x/vhost-scsi-ccw.c, which received a
couple comments [1][2] to update MAINTAINERS that were missed.

Fix that by making the vhost CCW entries a wildcard.

[1] https://lore.kernel.org/r/d8d2bbd5021076bdba444d31a6da74f507baede3.camel@linux.ibm.com/
[2] https://lore.kernel.org/r/87k0c4gb9f.fsf@redhat.com/

Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20220525145814.2750501-1-farman@linux.ibm.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agotarget/s390x: kvm: Honor storage keys during emulation
Janis Schoetterl-Glausch [Fri, 6 May 2022 15:39:56 +0000 (17:39 +0200)]
target/s390x: kvm: Honor storage keys during emulation

Storage key controlled protection is currently not honored when
emulating instructions.
If available, enable key protection for the MEM_OP ioctl, thereby
enabling it for the s390_cpu_virt_mem_* functions, when using kvm.
As a result, the emulation of the following instructions honors storage
keys:

* CLP
   The Synch I/O CLP command would need special handling in order
   to support storage keys, but is currently not supported.
* CHSC
Performing commands asynchronously would require special
handling, but commands are currently always synchronous.
* STSI
* TSCH
Must (and does) not change channel if terminated due to
protection.
* MSCH
Suppressed on protection, works because fetching instruction.
* SSCH
Suppressed on protection, works because fetching instruction.
* STSCH
* STCRW
Suppressed on protection, this works because no partial store is
possible, because the operand cannot span multiple pages.
* PCISTB
* MPCIFC
* STPCIFC

Signed-off-by: Janis Schoetterl-Glausch <scgl@linux.ibm.com>
Message-Id: <20220506153956.2217601-3-scgl@linux.ibm.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agohw/s390x/s390-virtio-ccw: Improve the machine description string
Thomas Huth [Fri, 6 May 2022 06:50:26 +0000 (08:50 +0200)]
hw/s390x/s390-virtio-ccw: Improve the machine description string

The machine name already contains the words "ccw" and "virtio", so
using "VirtIO-ccw" in the description likely does not really help
the average user to get an idea what this machine type is about.
Thus let's switch to "Virtual s390x machine" now, since "virtual
machine" should be a familiar term, and "s390x" signals that this
is about 64-bit guests (unlike S390 which could mean that it is
31-bit only).
Also expand "v" to "version", since this makes it easier to use
this macro also with non-numeric machine names in downstream.

Message-Id: <20220506065026.513590-1-thuth@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agos390: Typo fix FLOATING_POINT_SUPPPORT_ENH
Dr. David Alan Gilbert [Mon, 23 May 2022 11:51:23 +0000 (12:51 +0100)]
s390: Typo fix FLOATING_POINT_SUPPPORT_ENH

One less P needed.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20220523115123.150340-1-dgilbert@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2 years agoMerge tag 'pull-tcg-20220602' of https://gitlab.com/rth7680/qemu into staging
Richard Henderson [Thu, 2 Jun 2022 15:13:36 +0000 (08:13 -0700)]
Merge tag 'pull-tcg-20220602' of https://gitlab.com/rth7680/qemu into staging

Add tcg_gen_mov_ptr.
Fix tcg/i386 encoding of avx512 vpsraq.
Fix tcg/aarch64 handling of out-of-range shli.

# -----BEGIN PGP SIGNATURE-----
#
# iQFRBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAmKY0xodHHJpY2hhcmQu
# aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV8nhwf+LTq5foaCunKI9yeQ
# o78A1BbSFzSUd9hlzHt4MDWlDKsF+0WUKJ2kkgUTXNFUdTgpvQUKCaTuqHpt+LfQ
# o8Gfbbd2eYixz/utkJ+PE3xUqNR5Wnh6XfuSFlk1ib4x2Wztr7+Mm+szoRcbIXR4
# 5pqZxGVQ9XVFPaIQ6cnb6IQe68ky/p9ejoG1SPiphlLt5/zY7Fqsicfn7NFePK5t
# IxfcLjapCokLKP3AeioSf3l96sT1OWT6lQLeammEipMH0MuR/jTd3Ayx0nV0RwVz
# FoSV+JP+CCyRu+C1yQ+ImkxMQhgBfu+T19CBIzimIEx8sn+8Nby+oi2CvNx8mBWj
# Z3dQlg==
# =sVH7
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 02 Jun 2022 08:11:22 AM PDT
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [ultimate]

* tag 'pull-tcg-20220602' of https://gitlab.com/rth7680/qemu:
  tcg/aarch64: Fix illegal insn from out-of-range shli
  tcg/i386: Fix encoding of OPC_VPSRAQ for INDEX_op_sars_vec
  tcg: Add tcg_gen_mov_ptr

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/aarch64: Fix illegal insn from out-of-range shli
Richard Henderson [Thu, 2 Jun 2022 01:04:30 +0000 (01:04 +0000)]
tcg/aarch64: Fix illegal insn from out-of-range shli

The masking in tcg_out_shl was incorrect, producing an
illegal instruction, rather than merely unspecified results
for the out-of-range shift.

Tested-by: Joel Stanley <joel@jms.id.au>
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1051
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg/i386: Fix encoding of OPC_VPSRAQ for INDEX_op_sars_vec
Richard Henderson [Wed, 1 Jun 2022 23:22:19 +0000 (16:22 -0700)]
tcg/i386: Fix encoding of OPC_VPSRAQ for INDEX_op_sars_vec

We wanted the VPSRAQ variant with the scalar vector shift operand,
not the variant with an immediate operand.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1022
Fixes: 47b331b2a8da ("tcg/i386: Implement avx512 scalar shift")
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotcg: Add tcg_gen_mov_ptr
Richard Henderson [Mon, 16 May 2022 23:10:51 +0000 (16:10 -0700)]
tcg: Add tcg_gen_mov_ptr

Add an interface to perform moves between TCGv_ptr.

Reviewed-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agoMerge tag 'm68k-for-7.1-pull-request' of https://github.com/vivier/qemu-m68k into...
Richard Henderson [Thu, 2 Jun 2022 13:30:24 +0000 (06:30 -0700)]
Merge tag 'm68k-for-7.1-pull-request' of https://github.com/vivier/qemu-m68k into staging

m68k pull request 20220602

- Fixes and cleanup
- Implement TRAP opcodes
- Enable halt on 68060

# -----BEGIN PGP SIGNATURE-----
#
# iQJGBAABCAAwFiEEzS913cjjpNwuT1Fz8ww4vT8vvjwFAmKYpeQSHGxhdXJlbnRA
# dml2aWVyLmV1AAoJEPMMOL0/L748U+AP/i6EYidZmelIEqOwZTwwzxreF5bTZmAP
# v0Hxt3Tef3PWJpLnoCXCsd4othCO3PgHcwtrLff+bkWRl0Wt5CYcq+tTu2im7fIN
# zM7RSO00Pt/va7Ss7Ej8d5P5l7uuFqcBFytnitbsNrvHNK4cQ9PVmOkPnJZe0lYt
# vA3pUk7giE1KV+/s78Z4VD5CbvwpTRQpDCPDvba7oIP2E9mOELajKtYGh7gvPthx
# hrG2L5Ou4rYWxJkpZ0mNyYvoPuGRmzgPImdaDMTPLjEYNJMnnqGCRm+ANtzNk+jy
# d/fE/xJ41xvPAt4Q29yCp0vITuRF468M/elp5hQr/rHc6xtitLCi57FhduY9PuL/
# zCMXytgFtnU1C9XhDI/FtQhQxpvEKkZmEJRrAnsuHQKLrHlGoofjBU3whHqfx1zG
# qw/cdYqx/RUKKxvmoTbk76doaqfVQvBIx2nB6CsHF3pqOpQETK5TYeId49GCkwgR
# 4DmBPL1RZZpkYxi1KEKprcJWMj1l29UTa2dJ+kt9T2YACRm7MYQurP8OCGoHFIX4
# MOr3vdxaqSRU+mE2lWLZWupkZyzFrG/khHSB7A9htTomgbfZLfc0YkHX5kOkHQNq
# k4ymLpf16F94aau568HVQO8UZV+1FedtRwJL2EWVqkzKri9rSCCeI8I27HVLjwLP
# YzrHwsMVsjgl
# =T1g0
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 02 Jun 2022 04:58:28 AM PDT
# gpg:                using RSA key CD2F75DDC8E3A4DC2E4F5173F30C38BD3F2FBE3C
# gpg:                issuer "laurent@vivier.eu"
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>" [undefined]
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>" [undefined]
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>" [undefined]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* tag 'm68k-for-7.1-pull-request' of https://github.com/vivier/qemu-m68k:
  target/m68k: Mark helper_raise_exception as noreturn
  linux-user/strace: Adjust get_thread_area for m68k
  linux-user/strace: Use is_error in print_syscall_err
  tests/tcg/m68k: Add trap.c
  target/m68k: Implement FTRAPcc
  target/m68k: Implement TRAPV
  target/m68k: Implement TPF in terms of TRAPcc
  target/m68k: Implement TRAPcc
  target/m68k: Fix stack frame for EXCP_ILLEGAL
  target/m68k: Fix address argument for EXCP_TRACE
  target/m68k: Fix pc, c flag, and address argument for EXCP_DIV0
  target/m68k: Fix address argument for EXCP_CHK
  target/m68k: Remove retaddr in m68k_interrupt_all
  linux-user/m68k: Handle EXCP_TRAP1 through EXCP_TRAP15
  target/m68k: Fix coding style in m68k_interrupt_all
  target/m68k: Switch over exception type in m68k_interrupt_all
  target/m68k: Raise the TRAPn exception with the correct pc
  target/m68k: Enable halt insn for 68060
  target/m68k: Clear mach in m68k_cpu_disas_set_info

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agotarget/m68k: Mark helper_raise_exception as noreturn
Richard Henderson [Thu, 2 Jun 2022 01:34:01 +0000 (18:34 -0700)]
target/m68k: Mark helper_raise_exception as noreturn

Also mark raise_exception_ra and raise_exception, lest we
generate a warning about helper_raise_exception returning.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-18-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agolinux-user/strace: Adjust get_thread_area for m68k
Richard Henderson [Thu, 2 Jun 2022 01:34:00 +0000 (18:34 -0700)]
linux-user/strace: Adjust get_thread_area for m68k

Unlike i386, m68k get_thread_area has no arguments.

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-17-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agolinux-user/strace: Use is_error in print_syscall_err
Richard Henderson [Thu, 2 Jun 2022 01:33:59 +0000 (18:33 -0700)]
linux-user/strace: Use is_error in print_syscall_err

Errors are not all negative numbers: use is_error.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-16-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotests/tcg/m68k: Add trap.c
Richard Henderson [Thu, 2 Jun 2022 01:33:58 +0000 (18:33 -0700)]
tests/tcg/m68k: Add trap.c

Test various trap instructions: chk, div, trap, trapv, trapcc, ftrapcc,
and the signals and addresses that we expect from them.

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-15-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Implement FTRAPcc
Richard Henderson [Thu, 2 Jun 2022 01:33:57 +0000 (18:33 -0700)]
target/m68k: Implement FTRAPcc

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-14-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Implement TRAPV
Richard Henderson [Thu, 2 Jun 2022 01:33:56 +0000 (18:33 -0700)]
target/m68k: Implement TRAPV

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-13-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Implement TPF in terms of TRAPcc
Richard Henderson [Thu, 2 Jun 2022 01:33:55 +0000 (18:33 -0700)]
target/m68k: Implement TPF in terms of TRAPcc

TPF stands for "trap false", and is a long-form nop for ColdFire.
Re-use the immediate consumption code from trapcc; the insn will
already expand to a nop because of the TCG_COND_NEVER test
within do_trapcc.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-12-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Implement TRAPcc
Richard Henderson [Thu, 2 Jun 2022 01:33:54 +0000 (18:33 -0700)]
target/m68k: Implement TRAPcc

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/754
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-11-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Fix stack frame for EXCP_ILLEGAL
Richard Henderson [Thu, 2 Jun 2022 01:33:53 +0000 (18:33 -0700)]
target/m68k: Fix stack frame for EXCP_ILLEGAL

According to the M68040 Users Manual, section 8.4.1, Four word
stack frame (format 0), includes Illegal Instruction.  Use the
correct frame format, which does not use the ADDR argument.

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-10-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Fix address argument for EXCP_TRACE
Richard Henderson [Thu, 2 Jun 2022 01:33:52 +0000 (18:33 -0700)]
target/m68k: Fix address argument for EXCP_TRACE

According to the M68040 Users Manual, section 8.4.3,
Six word stack frame (format 2), Trace (and others) is
supposed to record the next insn in PC and the address
of the trapping instruction in ADDRESS.

Create gen_raise_exception_format2 to record the trapping
pc in env->mmu.ar.  Update m68k_interrupt_all to pass the
value to do_stack_frame.  Update cpu_loop to handle EXCP_TRACE.

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-9-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Fix pc, c flag, and address argument for EXCP_DIV0
Richard Henderson [Thu, 2 Jun 2022 01:33:51 +0000 (18:33 -0700)]
target/m68k: Fix pc, c flag, and address argument for EXCP_DIV0

According to the M68040 Users Manual, section 8.4.3,
Six word stack frame (format 2), Zero Div (and others)
is supposed to record the next insn in PC and the
address of the trapping instruction in ADDRESS.

While the N, Z and V flags are documented to be undefine on DIV0,
the C flag is documented as always cleared.

Update helper_div* to take the instruction length as an argument
and use raise_exception_format2.  Hoist the reset of the C flag
above the division by zero check.

Update m68k_interrupt_all to pass mmu.ar to do_stack_frame.
Update cpu_loop to pass mmu.ar to siginfo.si_addr, as the
kernel does in trap_c().

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-8-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Fix address argument for EXCP_CHK
Richard Henderson [Thu, 2 Jun 2022 01:33:50 +0000 (18:33 -0700)]
target/m68k: Fix address argument for EXCP_CHK

According to the M68040 Users Manual, section 8.4.3,
Six word stack frame (format 2), CHK, CHK2 (and others)
are supposed to record the next insn in PC and the
address of the trapping instruction in ADDRESS.

Create a raise_exception_format2 function to centralize recording
of the trapping pc in mmu.ar, plus advancing to the next insn.

Update m68k_interrupt_all to pass mmu.ar to do_stack_frame.
Update cpu_loop to pass mmu.ar to siginfo.si_addr, as the
kernel does in trap_c().

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-7-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Remove retaddr in m68k_interrupt_all
Richard Henderson [Thu, 2 Jun 2022 01:33:49 +0000 (18:33 -0700)]
target/m68k: Remove retaddr in m68k_interrupt_all

The only value this variable holds is now env->pc.

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-6-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agolinux-user/m68k: Handle EXCP_TRAP1 through EXCP_TRAP15
Richard Henderson [Thu, 2 Jun 2022 01:33:48 +0000 (18:33 -0700)]
linux-user/m68k: Handle EXCP_TRAP1 through EXCP_TRAP15

These are raised by guest instructions, and should not
fall through into the default abort case.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-5-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Fix coding style in m68k_interrupt_all
Richard Henderson [Thu, 2 Jun 2022 01:33:47 +0000 (18:33 -0700)]
target/m68k: Fix coding style in m68k_interrupt_all

Add parenthesis around & vs &&.

Remove assignment to sr in function call argument -- note that
sr is unused after the call, so the assignment was never needed,
only the result of the & expression.

Suggested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-4-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Switch over exception type in m68k_interrupt_all
Richard Henderson [Thu, 2 Jun 2022 01:33:46 +0000 (18:33 -0700)]
target/m68k: Switch over exception type in m68k_interrupt_all

Replace an if ladder with a switch for clarity.

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-3-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agotarget/m68k: Raise the TRAPn exception with the correct pc
Richard Henderson [Thu, 2 Jun 2022 01:33:45 +0000 (18:33 -0700)]
target/m68k: Raise the TRAPn exception with the correct pc

Rather than adjust the PC in all of the consumers, raise
the exception with the correct PC in the first place.

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220602013401.303699-2-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2 years agoMerge tag 'pull-testing-next-010622-3' of https://github.com/stsquad/qemu into staging
Richard Henderson [Wed, 1 Jun 2022 19:15:30 +0000 (12:15 -0700)]
Merge tag 'pull-testing-next-010622-3' of https://github.com/stsquad/qemu into staging

Various testing updates

  - fix some gitlab container dependencies
  - report meson test results via JUnit
  - fix meson display of enabled cross compilers
  - convert more cross build containers to lcitool and Debian 11
  - re-factor cross compiler detection
  - use test cross-compilers for building ROMs
  - disable CI runs by default (see docs)
  - fix some broken links in development documentation

# -----BEGIN PGP SIGNATURE-----
#
# iQEzBAABCgAdFiEEZoWumedRZ7yvyN81+9DbCVqeKkQFAmKXqXYACgkQ+9DbCVqe
# KkTaogf+PpslFcqKzcT+CY2zkmen7+531fSHQ2oP6+lMPSuIDXi7jn/d/IXQ1gcI
# BfCvnAsictBzBqCPNhKYd8zxZwdzSYT3QtRWR4D86G26tyL5Zsuw+HrXuja/af7A
# RDULBl8HGKuZKY83QI1wTekjamfdFQeWn1t9051np5mPt0buPe4S3xELl0A/QGX+
# f3H9hqYeQ+tb11a5KXEsbWrQqKRCNrjFKI2iKyz5rZa73Go0ODjHrtTfBccBNJv4
# TBFxAW/XU+AUTI3vgtAsBIonEz7AcJyqlMffnx93GIttuChn5KRIIMPuz7Oztn89
# OFHtGajs3DSEWjfEg9xncCaL7EFasA==
# =9Iod
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 01 Jun 2022 11:01:26 AM PDT
# gpg:                using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [undefined]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* tag 'pull-testing-next-010622-3' of https://github.com/stsquad/qemu: (33 commits)
  docs/devel: clean-up the CI links in the docs
  gitlab: don't run CI jobs in forks by default
  gitlab: convert build/container jobs to .base_job_template
  gitlab: convert static checks to .base_job_template
  gitlab: convert Cirrus jobs to .base_job_template
  gitlab: introduce a common base job template
  configure: remove unused variables from config-host.mak
  configure: enable cross compilation of vof
  configure: enable cross-compilation of optionrom
  configure: enable cross-compilation of s390-ccw
  configure: move symlink configuration earlier
  configure: include more binutils in tests/tcg makefile
  configure: introduce --cross-prefix-*=
  configure: handle host compiler in probe_target_compiler
  configure: add missing cross compiler fallbacks
  tests/tcg: merge configure.sh back into main configure script
  tests/tcg: correct target CPU for sparc32
  configure, meson: move symlinking of ROMs to meson
  build: do a full build before running TCG tests
  build: add a more generic way to specify make->ninja dependencies
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2 years agodocs/devel: clean-up the CI links in the docs
Alex Bennée [Fri, 27 May 2022 15:36:03 +0000 (16:36 +0100)]
docs/devel: clean-up the CI links in the docs

There where some broken links so fix those up with proper references
to the devel docs. I also did a little light copy-editing to reflect
the current state and broke up a paragraph to reduce the "wall of
text" effect.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220527153603.887929-34-alex.bennee@linaro.org>

2 years agogitlab: don't run CI jobs in forks by default
Daniel P. Berrangé [Fri, 27 May 2022 15:36:02 +0000 (16:36 +0100)]
gitlab: don't run CI jobs in forks by default

To preserve CI shared runner credits we don't want to run
pipelines on every push.

This sets up the config so that pipelines are never created
for contributors by default. To override this the QEMU_CI
variable can be set to a non-zero value. If set to 1, the
pipeline will be created but all jobs will remain manually
started. The contributor can selectively run jobs that they
care about. If set to 2, the pipeline will be created and
all jobs will immediately start.

This behavior can be controlled using push variables

  git push -o ci.variable=QEMU_CI=1

To make this more convenient define an alias

   git config --local alias.push-ci "push -o ci.variable=QEMU_CI=1"
   git config --local alias.push-ci-now "push -o ci.variable=QEMU_CI=2"

Which lets you run

  git push-ci

to create the pipeline, or

  git push-ci-now

to create and run the pipeline

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220526110705.59952-6-berrange@redhat.com>
[AJB: fix typo, replicate alias tips in ci.rst]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220527153603.887929-33-alex.bennee@linaro.org>

2 years agogitlab: convert build/container jobs to .base_job_template
Daniel P. Berrangé [Fri, 27 May 2022 15:36:01 +0000 (16:36 +0100)]
gitlab: convert build/container jobs to .base_job_template

This converts the main build and container jobs to use the
base job rules, defining the following new variables

 - QEMU_JOB_SKIPPED - jobs that are known to be currently
   broken and should not be run. Can still be manually
   launched if desired.

 - QEMU_JOB_AVOCADO - jobs that run the Avocado integration
   test harness.

 - QEMU_JOB_PUBLISH - jobs that publish content after the
   branch is merged upstream

As build-tools-and-docs runs on master we declare the requirement of
building amd64-debian-container optional as it should already exits
once we merge.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220526110705.59952-5-berrange@redhat.com>
[AJB: fix upstream typo, mention optional container req]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220527153603.887929-32-alex.bennee@linaro.org>

2 years agogitlab: convert static checks to .base_job_template
Daniel P. Berrangé [Fri, 27 May 2022 15:36:00 +0000 (16:36 +0100)]
gitlab: convert static checks to .base_job_template

This folds the static checks into using the base job
template rules, introducing one new variable

 - QEMU_JOB_ONLY_FORKS - a job that should never run
   on an upstream pipeline. The information it reports
   is only applicable to contributors in a pre-submission
   scenario, not time of merge.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220526110705.59952-4-berrange@redhat.com>
[AJB: fix typo]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220527153603.887929-31-alex.bennee@linaro.org>

2 years agogitlab: convert Cirrus jobs to .base_job_template
Daniel P. Berrangé [Fri, 27 May 2022 15:35:59 +0000 (16:35 +0100)]
gitlab: convert Cirrus jobs to .base_job_template

This folds the Cirrus job rules into the base job
template, introducing two new variables

  - QEMU_JOB_CIRRUS - identifies the job as making
    use of Cirrus CI via cirrus-run

  - QEMU_JOB_OPTIONAL - identifies the job as one
    that is not run by default, primarily due to
    resource constraints. It can be manually invoked
    by users if they wish to validate that scenario.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220526110705.59952-3-berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220527153603.887929-30-alex.bennee@linaro.org>

2 years agogitlab: introduce a common base job template
Daniel P. Berrangé [Fri, 27 May 2022 15:35:58 +0000 (16:35 +0100)]
gitlab: introduce a common base job template

Currently job rules are spread across the various templates
and jobs, making it hard to understand exactly what runs in
what scenario. This leads to inconsistency in the rules and
increased maint burden.

The intent is that we introduce a common '.base_job_template'
which will have a general purpose 'rules:' block. No other
template or job should define 'rules:', but instead they must
rely on the inherited rules. To allow behaviour to be tweaked,
rules will be influenced by a number of variables with the
naming scheme 'QEMU_JOB_nnnn'.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220526110705.59952-2-berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220527153603.887929-29-alex.bennee@linaro.org>

2 years agoconfigure: remove unused variables from config-host.mak
Paolo Bonzini [Fri, 27 May 2022 15:35:57 +0000 (16:35 +0100)]
configure: remove unused variables from config-host.mak

The only compiler variable that is still needed is $(CC), for
contrib/plugins/Makefile.  All firmware builds have their own
config-host.mak file.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220517092616.1272238-17-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220527153603.887929-28-alex.bennee@linaro.org>

2 years agoconfigure: enable cross compilation of vof
Paolo Bonzini [Fri, 27 May 2022 15:35:56 +0000 (16:35 +0100)]
configure: enable cross compilation of vof

While container-based cross compilers are not supported, this already
makes it possible to build vof on any machine that has an installation
of GCC and binutils for 32- or 64-bit PowerPC.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220517092616.1272238-16-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220527153603.887929-27-alex.bennee@linaro.org>

2 years agoconfigure: enable cross-compilation of optionrom
Paolo Bonzini [Fri, 27 May 2022 15:35:55 +0000 (16:35 +0100)]
configure: enable cross-compilation of optionrom

While container-based cross compilers are not supported, this already makes
it possible to build x86 optionroms on any machine that has an installation
of GCC and binutils for 32- or 64-bit x86.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220517092616.1272238-15-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220527153603.887929-26-alex.bennee@linaro.org>

2 years agoconfigure: enable cross-compilation of s390-ccw
Paolo Bonzini [Fri, 27 May 2022 15:35:54 +0000 (16:35 +0100)]
configure: enable cross-compilation of s390-ccw

While container-based cross compilers are not supported, this already makes
it possible to build s390-ccw on any machine that has s390x GCC and binutils
installed.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220517092616.1272238-14-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Acked-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220527153603.887929-25-alex.bennee@linaro.org>

2 years agoconfigure: move symlink configuration earlier
Paolo Bonzini [Fri, 27 May 2022 15:35:53 +0000 (16:35 +0100)]
configure: move symlink configuration earlier

Ensure that the pc-bios/optionrom and pc-bios/s390-ccw directory
exist at the time when we'll write out the compiler configuration
for them.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220517092616.1272238-13-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220527153603.887929-24-alex.bennee@linaro.org>

2 years agoconfigure: include more binutils in tests/tcg makefile
Paolo Bonzini [Fri, 27 May 2022 15:35:52 +0000 (16:35 +0100)]
configure: include more binutils in tests/tcg makefile

Firmware builds require paths to all the binutils; it is not enough to
use only cc, or even as/ld as in the case of tests/tcg/tricore.
Adjust the cross-compiler configurator to detect also ar, nm, objcopy,
ranlib and strip.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220517092616.1272238-12-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220527153603.887929-23-alex.bennee@linaro.org>

2 years agoconfigure: introduce --cross-prefix-*=
Paolo Bonzini [Fri, 27 May 2022 15:35:51 +0000 (16:35 +0100)]
configure: introduce --cross-prefix-*=

Also in preparation for handling more binaries from the cross binutils,
support an option --cross-prefix-ARCH.  All cross_cc_* defaults are
replaced with cross_prefix_*; the cross_cc_* fallbacks are extended
to the cross-compilation prefix, but the compiler fallbacks remain
as well.  This way, for example, --cross-cc-arm=arm-linux-gnueabihf-clang
also applies to armeb binaries.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220517092616.1272238-11-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220527153603.887929-22-alex.bennee@linaro.org>

2 years agoconfigure: handle host compiler in probe_target_compiler
Paolo Bonzini [Fri, 27 May 2022 15:35:50 +0000 (16:35 +0100)]
configure: handle host compiler in probe_target_compiler

In preparation for handling more binaries than just cc, handle
the case of "probe_target_compiler $cpu" directly in the function,
setting the target_* variables based on the ones that are used to
build QEMU.  The clang check also needs to be moved after this
fallback.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220517092616.1272238-10-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220527153603.887929-21-alex.bennee@linaro.org>

2 years agoconfigure: add missing cross compiler fallbacks
Paolo Bonzini [Fri, 27 May 2022 15:35:49 +0000 (16:35 +0100)]
configure: add missing cross compiler fallbacks

The arm compiler can be used for armeb, and the sparc64 compiler
can be used for sparc.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220517092616.1272238-9-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220527153603.887929-20-alex.bennee@linaro.org>

2 years agotests/tcg: merge configure.sh back into main configure script
Paolo Bonzini [Fri, 27 May 2022 15:35:48 +0000 (16:35 +0100)]
tests/tcg: merge configure.sh back into main configure script

tests/tcg/configure.sh has a complicated story.

In the beginning its code ran as part of the creation of config-target.mak
files, and that is where it placed the information on the target compiler.
However, probing for the buildability of TCG tests required multiple
inclusions of config-target.mak in the _main_ Makefile (not in
Makefile.target, which took care of building the QEMU executables in
the pre-Meson era), which polluted the namespace.

Thus, it was moved to a separate directory.  It created small config-*.mak
files in $(BUILD_DIR)/tests/tcg.  Those were also included multiple
times, but at least they were small and manageable; this was also an
important step in disentangling the TCG tests from Makefile.target.

Since then, Meson has allowed the configure script to go on a diet.
A few compilation tests survive (mostly for sanitizers) but these days
it mostly takes care of command line parsing, looking for tools, and
setting up the environment for Meson to do its stuff.

It's time to extend configure with the capability to build for more
than just one target: not just tests, but also firmware.  As a first
step, integrate all the logic to find cross compilers in the configure
script, and move tests/tcg/configure.sh back there (though as a
separate loop, not integrated in the one that generates target
configurations for Meson).

tests/tcg is actually very close to being buildable as a standalone
project, so I actually expect the compiler tests to move back to
tests/tcg, as a "configure" script of sorts which would run at Make
time after the docker images are built.  The GCC tree has a similar idea
of doing only bare-bones tree-wide configuration and leaving the rest
for Make time.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220517092616.1272238-8-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20220527153603.887929-19-alex.bennee@linaro.org>