util/oslib-posix: Avoid creating a single thread with MADV_POPULATE_WRITE

Let's simplify the case when we only want a single thread and don't have
to mess with signal handlers.

Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20211217134611.31172-6-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

util/oslib-posix: Don't create too many threads with small memory or little pages

Let's limit the number of threads to something sane, especially that
- We don't have more threads than the number of pages we have
- We don't have threads that initialize small (< 64 MiB) memory

Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20211217134611.31172-5-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

util/oslib-posix: Introduce and use MemsetContext for touch_all_pages()

Let's minimize the number of global variables to prepare for
os_mem_prealloc() getting called concurrently and make the code a bit
easier to read.

The only consumer that really needs a global variable is the sigbus
handler, which will require protection via a mutex in the future either way
as we cannot concurrently mess with the SIGBUS handler.

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20211217134611.31172-4-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

util/oslib-posix: Support MADV_POPULATE_WRITE for os_mem_prealloc()

Let's sense support and use it for preallocation. MADV_POPULATE_WRITE
does not require a SIGBUS handler, doesn't actually touch page content,
and avoids context switches; it is, therefore, faster and easier to handle
than our current approach.

While MADV_POPULATE_WRITE is, in general, faster than manual
prefaulting, and especially faster with 4k pages, there is still value in
prefaulting using multiple threads to speed up preallocation.

More details on MADV_POPULATE_WRITE can be found in the Linux commits
4ca9b3859dac ("mm/madvise: introduce MADV_POPULATE_(READ|WRITE) to prefault
page tables") and eb2faa513c24 ("mm/madvise: report SIGBUS as -EFAULT for
MADV_POPULATE_(READ|WRITE)"), and in the man page proposal [1].

This resolves the TODO in do_touch_pages().

In the future, we might want to look into using fallocate(), eventually
combined with MADV_POPULATE_READ, when dealing with shared file/fd
mappings and not caring about memory bindings.

[1] https://lkml.kernel.org/r/20210816081922.5155-1-david@redhat.com

Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20211217134611.31172-3-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

util/oslib-posix: Let touch_all_pages() return an error

Let's prepare touch_all_pages() for returning differing errors. Return
an error from the thread and report the last processed error.

Translate SIGBUS to -EFAULT, as a SIGBUS can mean all different kind of
things (memory error, read error, out of memory). When allocating memory
fails via the current SIGBUS-based mechanism, we'll get:
    os_mem_prealloc: preallocating memory failed: Bad address

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20211217134611.31172-2-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/vhost-user-blk: turn on VIRTIO_BLK_F_SIZE_MAX feature for virtio blk device

Turn on pre-defined feature VIRTIO_BLK_F_SIZE_MAX for virtio blk device to
avoid guest DMA request sizes which are too large for hardware spec.

Signed-off-by: Andy Pei <andy.pei@intel.com>
Message-Id: <1641202092-149677-1-git-send-email-andy.pei@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Raphael Norwitz <raphael.norwitz@nutanix.com>

hw/i386: expose a "smbios-entry-point-type" PC machine property

The i440fx and Q35 machine types are both hardcoded to use the
legacy SMBIOS 2.1 (32-bit) entry point. This is a sensible
conservative choice because SeaBIOS only supports SMBIOS 2.1

EDK2, however, can also support SMBIOS 3.0 (64-bit) entry points,
and QEMU already uses this on the ARM virt machine type.

This adds a property to allow the choice of SMBIOS entry point
versions For example to opt in to 64-bit SMBIOS entry point:

   $QEMU -machine q35,smbios-entry-point-type=64

Based on a patch submitted by Daniel Berrangé.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20211026151100.1691925-4-ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

hw/smbios: Use qapi for SmbiosEntryPointType

This prepares for exposing the SMBIOS entry point type as a
machine property on x86.

Based on a patch from Daniel P. Berrangé.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20211026151100.1691925-3-ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Markus Armbruster <armbru@redhat.com>

smbios: Rename SMBIOS_ENTRY_POINT_* enums

Rename the enums to match the naming style used by QAPI, and to
use "32" and "64" instead of "20" and "31".  This will allow us
to more easily move the enum to the QAPI schema later.

About the naming choice: "SMBIOS 2.1 entry point"/"SMBIOS 3.0
entry point" and "32-bit entry point"/"64-bit entry point" are
synonymous in the SMBIOS specification.  However, the phrases
"32-bit entry point" and "64-bit entry point" are used more often.

The new names also avoid confusion between the entry point format
and the actual SMBIOS version reported in the entry point
structure.  For example: currently the 32-bit entry point
actually report SMBIOS 2.8 support, not 2.1.

Based on portions of a patch submitted by Daniel P. Berrangé.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20211026151100.1691925-2-ehabkost@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pcie_aer: Don't trigger a LSI if none are defined

Skip triggering an LSI when the AER root error status is updated if no
LSI is defined for the device. We can have a root bridge with no LSI,
MSI and MSI-X defined, for example on POWER systems.

Signed-off-by: Frederic Barrat <fbarrat@linux.ibm.com>
Message-Id: <20211116170133.724751-4-fbarrat@linux.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>

pci: Export the pci_intx() function

Move the pci_intx() definition to the PCI header file, so that it can
be called from other PCI files. It is used by the next patch.

Signed-off-by: Frederic Barrat <fbarrat@linux.ibm.com>
Message-Id: <20211116170133.724751-3-fbarrat@linux.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>

vhost-user-blk: propagate error return from generic vhost

Fix the only callsite that doesn't propagate the error code from the
generic vhost code.

Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20211111153354.18807-11-rvkagan@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>

vhost: stick to -errno error return convention

The generic vhost code expects that many of the VhostOps methods in the
respective backends set errno on errors.  However, none of the existing
backends actually bothers to do so.  In a number of those methods errno
from the failed call is clobbered by successful later calls to some
library functions; on a few code paths the generic vhost code then
negates and returns that errno, thus making failures look as successes
to the caller.

As a result, in certain scenarios (e.g. live migration) the device
doesn't notice the first failure and goes on through its state
transitions as if everything is ok, instead of taking recovery actions
(break and reestablish the vhost-user connection, cancel migration, etc)
before it's too late.

To fix this, consolidate on the convention to return negated errno on
failures throughout generic vhost, and use it for error propagation.

Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20211111153354.18807-10-rvkagan@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-user: stick to -errno error return convention

VhostOps methods in user_ops are not very consistent in their error
returns: some return negated errno while others just -1.

Make sure all of them consistently return negated errno.  This also
helps error propagation from the functions being called inside.
Besides, this synchronizes the error return convention with the other
two vhost backends, kernel and vdpa, and will therefore allow for
consistent error propagation in the generic vhost code (in a followup
patch).

Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20211111153354.18807-9-rvkagan@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-vdpa: stick to -errno error return convention

Almost all VhostOps methods in vdpa_ops follow the convention of
returning negated errno on error.

Adjust the few that don't.  To that end, rework vhost_vdpa_add_status to
check if setting of the requested status bits has succeeded and return
the respective error code it hasn't, and propagate the error codes
wherever it's appropriate.

Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20211111153354.18807-8-rvkagan@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-backend: stick to -errno error return convention

Almost all VhostOps methods in kernel_ops follow the convention of
returning negated errno on error.

Adjust the only one that doesn't.

Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20211111153354.18807-7-rvkagan@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

vhost-backend: avoid overflow on memslots_limit

Fix the (hypothetical) potential problem when the value parsed out of
the vhost module parameter in sysfs overflows the return value from
vhost_kernel_memslots_limit.

Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20211111153354.18807-6-rvkagan@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

chardev/char-socket: tcp_chr_sync_read: don't clobber errno

After the return from tcp_chr_recv, tcp_chr_sync_read calls into a
function which eventually makes a system call and may clobber errno.

Make a copy of errno right after tcp_chr_recv and restore the errno on
return from tcp_chr_sync_read.

Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20211111153354.18807-4-rvkagan@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

chardev/char-socket: tcp_chr_recv: don't clobber errno

tcp_chr_recv communicates the specific error condition to the caller via
errno.  However, after setting it, it may call into some system calls or
library functions which can clobber the errno.

Avoid this by moving the errno assignment to the end of the function.

Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20211111153354.18807-3-rvkagan@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

vhost-user-blk: reconnect on any error during realize

vhost-user-blk realize only attempts to reconnect if the previous
connection attempt failed on "a problem with the connection and not an
error related to the content (which would fail again the same way in the
next attempt)".

However this distinction is very subtle, and may be inadvertently broken
if the code changes somewhere deep down the stack and a new error gets
propagated up to here.

OTOH now that the number of reconnection attempts is limited it seems
harmless to try reconnecting on any error.

So relax the condition of whether to retry connecting to check for any
error.

This patch amends a527e312b5 "vhost-user-blk: Implement reconnection
during realize".

Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20211111153354.18807-2-rvkagan@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>

trace-events,pci: unify trace events format

Unify format used by trace_pci_update_mappings_del(),
trace_pci_update_mappings_add(), trace_pci_cfg_write() and
trace_pci_cfg_read() to print the device name and bus number,
slot number and function number.

For instance:

  pci_cfg_read virtio-net-pci 00:0 @0x20 -> 0xffffc00c
  pci_cfg_write virtio-net-pci 00:0 @0x20 <- 0xfea0000c
  pci_update_mappings_del d=0x555810b92330 01:00.0 4,0xffffc000+0x4000
  pci_update_mappings_add d=0x555810b92330 01:00.0 4,0xfea00000+0x4000

becomes

  pci_cfg_read virtio-net-pci 01:00.0 @0x20 -> 0xffffc00c
  pci_cfg_write virtio-net-pci 01:00.0 @0x20 <- 0xfea0000c
  pci_update_mappings_del virtio-net-pci 01:00.0 4,0xffffc000+0x4000
  pci_update_mappings_add virtio-net-pci 01:00.0 4,0xfea00000+0x4000

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20211105192541.655831-1-lvivier@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-pci: add support for configure interrupt

Add support for configure interrupt, The process is used kvm_irqfd_assign
to set the gsi to kernel. When the configure notifier was signal by
host, qemu will inject a msix interrupt to guest

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-11-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mmio: add support for configure interrupt

Add configure interrupt support for virtio-mmio bus. This
interrupt will be working while the backend is vhost-vdpa

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-10-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-net: add support for configure interrupt

Add functions to support configure interrupt in virtio_net
The functions are config_pending and config_mask, while
this input idx is VIRTIO_CONFIG_IRQ_IDX will check the
function of configure interrupt.

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-9-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost: add support for configure interrupt

Add functions to support configure interrupt.
The configure interrupt process will start in vhost_dev_start
and stop in vhost_dev_stop.

Also add the functions to support vhost_config_pending and
vhost_config_mask, for masked_config_notifier, we only
use the notifier saved in vq 0.

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-8-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio: add support for configure interrupt

Add the functions to support the configure interrupt in virtio
The function virtio_config_guest_notifier_read will notify the
guest if there is an configure interrupt.
The function virtio_config_set_guest_notifier_fd_handler is
to set the fd hander for the notifier

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-7-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-vdpa: add support for config interrupt

Add new call back function in vhost-vdpa, this function will
set the event fd to kernel. This function will be called
in the vhost_dev_start and vhost_dev_stop

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-6-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost: introduce new VhostOps vhost_set_config_call

This patch introduces new VhostOps vhost_set_config_call. This function allows the
vhost to set the event fd to kernel

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-5-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-pci: decouple the single vector from the interrupt process

To reuse the interrupt process in configure interrupt
Need to decouple the single vector from the interrupt process. Add new function
kvm_virtio_pci_vector_use_one and _release_one. These functions are use
for the single vector, the whole process will finish in a loop for the vq number.

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-4-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-pci: decouple notifier from interrupt process

To reuse the notifier process in configure interrupt.
Use the virtio_pci_get_notifier function to get the notifier.
the INPUT of this function is the IDX, the OUTPUT is notifier and
the vector

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-3-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio: introduce macro IRTIO_CONFIG_IRQ_IDX

To support configure interrupt for vhost-vdpa
Introduce VIRTIO_CONFIG_IRQ_IDX -1 as configure interrupt's queue index,
Then we can reuse the functions guest_notifier_mask and guest_notifier_pending.
Add the check of queue index in these drivers, if the driver does not support
configure interrupt, the function will just return

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20211104164827.21911-2-lulu@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

acpi: validate hotplug selector on access

When bus is looked up on a pci write, we didn't
validate that the lookup succeeded.
Fuzzers thus can trigger QEMU crash by dereferencing the NULL
bus pointer.

Fixes: b32bd763a1 ("pci: introduce acpi-index property for PCI device")
Fixes: CVE-2021-4158
Cc: "Igor Mammedov" <imammedo@redhat.com>
Fixes: https://gitlab.com/qemu-project/qemu/-/issues/770
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Ani Sinha <ani@anisinha.ca>

virtio-mem: Don't skip alignment checks when warning about block size

If we warn about the block size being smaller than the default, we skip
some alignment checks.

This can currently only fail on x86-64, when specifying a block size of
1 MiB, however, we detect the THP size of 2 MiB.

Fixes: 228957fea3a9 ("virtio-mem: Probe THP size to determine default block size")
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20211011173305.13778-1-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Merge tag 'pull-request-2022-01-05' of https://gitlab.com/thuth/qemu into staging

* Add compat machines for 7.0
* Some minor qtest and unit test improvements
* Remove -no-quit option
* Fixes for the docs

# gpg: Signature made Wed 05 Jan 2022 02:10:49 AM PST
# 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-2022-01-05' of https://gitlab.com/thuth/qemu:
  docs/tools/qemu-trace-stap.rst: Do not hard-code the QEMU binary name
  gitlab-ci: Enable docs in the centos job
  docs/sphinx: fix compatibility with sphinx < 1.8
  qemu-options: Remove the deprecated -no-quit option
  tests/unit/test-util-sockets: Use g_file_open_tmp() to create temp file
  tests/qtest/hd-geo-test: Check for the lsi53c895a controller before using it
  tests/qtest/test-x86-cpuid-compat: Check for machines before using them
  hw: Add compat machines for 7.0

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

docs/tools/qemu-trace-stap.rst: Do not hard-code the QEMU binary name

In downstream, we want to use a different name for the QEMU binary,
and some people might also use the docs for non-x86 binaries, that's
why we already created the |qemu_system| placeholder in the past.
Use it now in the stap trace doc, too.

Message-Id: <20220104103319.179870-1-thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>

gitlab-ci: Enable docs in the centos job

We just ran into a problem that the docs don't build on RHEL8 / CentOS 8
anymore. Seems like these distros are using one of the oldest Sphinx
versions that we still have to support. Thus enable the docs build in
the CI on CentOS so that such bugs don't slip in so easily again.

Message-Id: <20220104091240.160867-1-thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

docs/sphinx: fix compatibility with sphinx < 1.8

SphinxDirective was added with sphinx 1.8 (2018-09-13).

Reported-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220104074649.1712440-1-marcandre.lureau@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

qemu-options: Remove the deprecated -no-quit option

This option was just a wrapper around the -display ...,window-close=off
parameter, and the name "no-quit" is rather confusing compared to
"window-close" (since there are still other means to quit the emulator),
so let's remove this now.

Message-Id: <20211215082417.180735-1-thuth@redhat.com>
Acked-by: Michal Prívozník <mprivozn@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

tests/unit/test-util-sockets: Use g_file_open_tmp() to create temp file

Similarly to commit e63ed64c6d1 ("tests/qtest/virtio-net-failover:
Use g_file_open_tmp() to create temporary file"), avoid calling
g_test_rand_int() before g_test_init(): use g_file_open_tmp().

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20211224234504.3413370-1-philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

tests/qtest/hd-geo-test: Check for the lsi53c895a controller before using it

The lsi53c895a SCSI controller might have been disabled in the target
binary, so let's check for its availability first before using it.

Message-Id: <20211222153600.976588-1-thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

tests/qtest/test-x86-cpuid-compat: Check for machines before using them

The user might have disabled the pc-i440fx machine type (or it's older
versions, like done in downstream RHEL) in the QEMU binary, so let's
better check whether the machine types are available before using them.

Message-Id: <20211222153923.1000420-1-thuth@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

hw: Add compat machines for 7.0

Add 7.0 machine types for arm/i440fx/q35/s390x/spapr.

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20211217143948.289995-1-cohuck@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

common-user: Really fix i386 calls to safe_syscall_set_errno_tail

Brown bag time: offset 0 from esp is the return address,
offset 4 is the first argument.

Fixes: d7478d4229f0 ("common-user: Fix tail calls to safe_syscall_set_errno_tail")
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge tag 'pull-tcg-20220104' of https://gitlab.com/rth7680/qemu into staging

Fix for safe_syscall_base.
Fix for folding of vector add/sub.
Fix build on loongarch64 with gcc 8.
Remove decl for qemu_run_machine_init_done_notifiers.

# gpg: Signature made Tue 04 Jan 2022 04:39:35 PM PST
# 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-20220104' of https://gitlab.com/rth7680/qemu:
  common-user: Fix tail calls to safe_syscall_set_errno_tail
  sysemu: Cleanup qemu_run_machine_init_done_notifiers()
  linux-user: Fix trivial build error on loongarch64 hosts
  tcg/optimize: Fix folding of vector ops

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

common-user: Fix tail calls to safe_syscall_set_errno_tail

For the ABIs in which the syscall return register is not
also the first function argument register, move the errno
value into the correct place.

Fixes: a3310c0397e2 ("linux-user: Move syscall error detection into safe_syscall_base")
Reported-by: Laurent Vivier <laurent@vivier.eu>
Tested-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: <20220104190454.542225-1-richard.henderson@linaro.org>

sysemu: Cleanup qemu_run_machine_init_done_notifiers()

Remove qemu_run_machine_init_done_notifiers() since no implementation
and user.

Fixes: f66dc8737c9 ("vl: move all generic initialization out of vl.c")
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220104024136.1433545-1-xiaoyao.li@intel.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Fix trivial build error on loongarch64 hosts

When building using GCC 8.3.0 on loongarch64 (Loongnix) we get:

  In file included from ../linux-user/signal.c:33:
  ../linux-user/host/loongarch64/host-signal.h: In function ‘host_signal_write’:
  ../linux-user/host/loongarch64/host-signal.h:57:9: error: a label can only be part of a statement and a declaration is not a statement
         uint32_t sel = (insn >> 15) & 0b11111111111;
         ^~~~~~~~

We don't use the 'sel' variable more than once, so drop it.

Meson output for the record:

  Host machine cpu family: loongarch64
  Host machine cpu: loongarch64
  C compiler for the host machine: cc (gcc 8.3.0 "cc (Loongnix 8.3.0-6.lnd.vec.27) 8.3.0")
  C linker for the host machine: cc ld.bfd 2.31.1-system

Fixes: ad812c3bd65 ("linux-user: Implement CPU-specific signal handler for loongarch64 hosts")
Reported-by: Song Gao <gaosong@loongson.cn>
Suggested-by: Song Gao <gaosong@loongson.cn>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: WANG Xuerui <git@xen0n.name>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220104215027.2180972-1-f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/optimize: Fix folding of vector ops

Bitwise operations are easy to fold, because the operation is
identical regardless of element size.  But add and sub need
extra element size info that is not currently propagated.

Fixes: 2f9f08ba43d
Cc: qemu-stable@nongnu.org
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/799
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge tag 'pull-ppc-20220104' of https://github.com/legoater/qemu into staging

ppc 7.0 queue:

* Cleanup of PowerNV PHBs (Daniel and Cedric)
* Cleanup and fixes for PPC405 machine (Cedric)
* Fix for xscvspdpn (Matheus)
* Rework of powerpc exception handling 1/n (Fabiano)
* Optimisation for PMU (Richard and Daniel)

# gpg: Signature made Mon 03 Jan 2022 11:04:06 PM PST
# gpg:                using RSA key A0F66548F04895EBFE6B0B6051A343C7CFFBECA1
# gpg: Good signature from "Cédric Le Goater <clg@kaod.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: A0F6 6548 F048 95EB FE6B  0B60 51A3 43C7 CFFB ECA1

* tag 'pull-ppc-20220104' of https://github.com/legoater/qemu: (26 commits)
  target/ppc: do not call hreg_compute_hflags() in helper_store_mmcr0()
  target/ppc: Use env->pnc_cyc_cnt
  target/ppc: Rewrite pmu_increment_insns
  target/ppc: Cache per-pmc insn and cycle count settings
  target/ppc: powerpc_excp: Stop passing excp_model around
  target/ppc: powerpc_excp: Move system call vectored code together
  target/ppc: powerpc_excp: Set vector earlier
  target/ppc: powerpc_excp: Add excp_vectors bounds check
  target/ppc: powerpc_excp: Set alternate SRRs directly
  target/ppc: do not silence snan in xscvspdpn
  ppc/ppc405: Dump specific registers
  ppc/ppc405: Introduce a store helper for SPR_40x_PID
  ppc/ppc405: Fix timer initialization
  ppc/ppc405: Rework ppc_40x_timers_init() to use a PowerPCCPU
  ppc/ppc405: Restore TCR and STR write handlers
  ppc/ppc405: Activate MMU logs
  ppc/ppc4xx: Convert printfs()
  target/ppc: Print out literal exception names in logs
  target/ppc: Remove static inline
  target/ppc: Check effective address validity
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/ppc: do not call hreg_compute_hflags() in helper_store_mmcr0()

MMCR0 writes will change only MMCR0 bits which are used to calculate
HFLAGS_PMCC0, HFLAGS_PMCC1 and HFLAGS_INSN_CNT hflags. No other machine
register will be changed during this operation. This means that
hreg_compute_hflags() is overkill for what we need to do.

pmu_update_summaries() is already updating HFLAGS_INSN_CNT without
calling hreg_compure_hflags(). Let's do the same for the other 2 MMCR0
hflags.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220103224746.167831-5-danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: Use env->pnc_cyc_cnt

Use the cached pmc_cyc_cnt value in pmu_update_cycles
and pmc_update_overflow_timer.  This leaves pmc_get_event
and pmc_is_inactive unused, so remove them.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220103224746.167831-4-danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: Rewrite pmu_increment_insns

Use the cached pmc_ins_cnt value.  Unroll the loop over the
different PMC counters.  Treat the PMC4 run-latch specially.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220103224746.167831-3-danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: Cache per-pmc insn and cycle count settings

This is the combination of frozen bit and counter type, on a per
counter basis. So far this is only used by HFLAGS_INSN_CNT, but
will be used more later.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
[danielhb: fixed PMC4 cyc_cnt shift, insn run latch code,
           MMCR0_FC handling, "PMC[1-6]" comment]
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220103224746.167831-2-danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: powerpc_excp: Stop passing excp_model around

We can just access it directly in powerpc_excp.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
[ clg: Took into account removal of inline ]
Message-Id: <20211229165751.3774248-6-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: powerpc_excp: Move system call vectored code together

Now that 'vector' is known before calling the interrupt-specific setup
code, we can move all of the scv setup into one place.

No functional change intended.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20211229165751.3774248-5-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: powerpc_excp: Set vector earlier

None of the interrupt setup code touches 'vector', so we can move it
earlier in the function. This will allow us to later move the System
Call Vectored setup that is on the top level into the
POWERPC_EXCP_SYSCALL_VECTORED code block.

This patch also moves the verification for when 'excp' does not have
an address associated with it. We now bail a little earlier when that
is the case. This should not cause any visible effects.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20211229165751.3774248-4-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: powerpc_excp: Add excp_vectors bounds check

The next patch will start accessing the excp_vectors array earlier in
the function, so add a bounds check as first thing here.

This converts the empty return on POWERPC_EXCP_NONE to an error. This
exception number never reaches this function and if it does it
probably means something else went wrong up the line.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20211229165751.3774248-3-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: powerpc_excp: Set alternate SRRs directly

There are currently only two interrupts that use alternate SRRs, so
let them write to them directly during the setup code.

No functional change intended.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20211229165751.3774248-2-farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: do not silence snan in xscvspdpn

The non-signalling versions of VSX scalar convert to shorter/longer
precision insns doesn't silence SNaNs in the hardware. To better match
this behavior, use the non-arithmatic conversion of helper_todouble
instead of float32_to_float64. A test is added to prevent future
regressions.

Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Message-Id: <20211228120310.1957990-1-matheus.ferst@eldorado.org.br>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/ppc405: Dump specific registers

Rework slightly ppc_cpu_dump_state() to replace the various 'if'
statements with a 'switch'.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20211222064025.1541490-9-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220103063441.3424853-10-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/ppc405: Introduce a store helper for SPR_40x_PID

The PID SPR of the 405 CPU contains the translation ID of the TLB
which is a 8-bit field. Enforce the mask with a store helper.

Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20211222064025.1541490-8-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220103063441.3424853-9-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/ppc405: Fix timer initialization

Timers are already initialized in ppc4xx_init(). No need to do it a
second time with a wrong set.

Fixes: d715ea961254 ("PPC: 405: Fix ppc405ep initialization")
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20211222064025.1541490-7-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220103063441.3424853-8-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/ppc405: Rework ppc_40x_timers_init() to use a PowerPCCPU

This is a small cleanup to ease reading. It includes the removal of a
check done on the returned value of g_malloc0(), which can not fail.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20211222064025.1541490-6-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220103063441.3424853-7-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/ppc405: Restore TCR and STR write handlers

The 405 timers were broken when booke support was added. Assumption
was made that the register numbers were the same but it's not :

    SPR_BOOKE_TSR         (0x150)
    SPR_BOOKE_TCR         (0x154)
    SPR_40x_TSR           (0x3D8)
    SPR_40x_TCR           (0x3DA)

Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Fixes: ddd1055b07fd ("PPC: booke timers")
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20211222064025.1541490-5-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220103063441.3424853-6-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/ppc405: Activate MMU logs

There is no need to deactivate MMU logging at compile time. Remove all
use of defines. Only keep DUMP_PAGE_TABLES for another series since
page tables could be dumped from the monitor.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20211222064025.1541490-4-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220103063441.3424853-5-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/ppc4xx: Convert printfs()

Use a QEMU log primitive for errors and trace events for debug.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.drobear.id.au>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20211222064025.1541490-3-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220103063441.3424853-4-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: Print out literal exception names in logs

It facilitates reading the logs when mask CPU_LOG_INT is activated. We
should do the same for error codes.

Cc: Fabiano Rosas <farosas@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20211222064025.1541490-2-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220103063441.3424853-3-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: Remove static inline

The compiler should know better how to inline code if necessary.

Suggested-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220103063441.3424853-2-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: Check effective address validity

For Radix translation, the EA range is 64-bits. when EA(2:11) are
nonzero, a segment interrupt should occur.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Frederic Barrat <fbarrat@linux.ibm.com>
Message-Id: <20211231073122.3183583-1-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

target/ppc: Improve logging in Radix MMU

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20211222071002.1568894-1-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

pnv_phb4.c: do not set 'root-bus' as bus name

This change has the same motivation as the one done for pnv-phb3-root-bus
buses previously. Defaulting every bus to 'root-bus' makes it impossible to attach
root ports to specific buses and it doesn't allow for custom bus
naming because we're ignoring the 'id' value when registering the root
bus.

After this patch, creating pnv-phb4 devices with 'id' being set will
result in the following qtree:

qemu-system-ppc64 -m 4G -machine powernv9,accel=tcg \
   -device pnv-phb4,chip-id=0,index=0,id=pcie.0 \
   -device pnv-phb4,chip-id=1,index=4,id=pcie.1

bus: main-system-bus
  type System
  dev: pnv-phb4, id "pcie.1"
    index = 4 (0x4)
    chip-id = 1 (0x1)
    version = 704374636546 (0xa400000002)
    device-id = 1217 (0x4c1)
    x-config-reg-migration-enabled = true
    bypass-iommu = false
    bus: pcie.1
      type pnv-phb4-root-bus
  dev: pnv-phb4, id "pcie.0"
    index = 0 (0x0)
    chip-id = 0 (0x0)
    version = 704374636546 (0xa400000002)
    device-id = 1217 (0x4c1)
    x-config-reg-migration-enabled = true
    bypass-iommu = false
    bus: pcie.0
      type pnv-phb4-root-bus

And without setting any ids:

qemu-system-ppc64 -m 4G -machine powernv9,accel=tcg \
   -device pnv-phb4,chip-id=0,index=0,id=pcie.0 \
   -device pnv-phb4,chip-id=1,index=4,id=pcie.1

bus: main-system-bus
  type System
  dev: pnv-phb4, id ""
    index = 4 (0x4)
    chip-id = 1 (0x1)
    version = 704374636546 (0xa400000002)
    device-id = 1217 (0x4c1)
    x-config-reg-migration-enabled = true
    bypass-iommu = false
    bus: pnv-phb4-root-bus.1
      type pnv-phb4-root-bus
  dev: pnv-phb4, id ""
    index = 0 (0x0)
    chip-id = 0 (0x0)
    version = 704374636546 (0xa400000002)
    device-id = 1217 (0x4c1)
    x-config-reg-migration-enabled = true
    bypass-iommu = false
    bus: pnv-phb4-root-bus.0
      type pnv-phb4-root-bus

Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20211228193806.1198496-17-danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

pnv_phb3.c: do not set 'root-bus' as bus name

All pnv-phb3-root-bus buses are being created as 'root-bus'. This
makes it impossible to, for example, add a pnv-phb3-root-port in
a specific root bus, since they all have the same name. By default
the device will be parented by the pnv-phb3 device that precedeced it in
the QEMU command line.

Moreover, this doesn't all for custom bus naming. Libvirt, for instance,
likes to name these buses as 'pcie.N', where 'N' is the index value of
the controller in the domain XML, by using the 'id' command line
attribute. At this moment this is also being ignored - the created root
bus will always be named 'root-bus'.

This patch fixes both scenarios by removing the 'root-bus' name from the
pci_register_root_bus() call. If an "id" is provided, use that.
Otherwise use 'NULL' as bus name. The 'NULL' value will be handled in
qbus_init_internal() and it will defaulted as lowercase bus type + the
global bus_id value.

After this path we can define the bus name by using the 'id' attribute:

qemu-system-ppc64 -m 4G -machine powernv8,accel=tcg \
    -device pnv-phb3,chip-id=0,index=1,id=pcie.0

  dev: pnv-phb3, id "pcie.0"
    index = 1 (0x1)
    chip-id = 0 (0x0)
    x-config-reg-migration-enabled = true
    bypass-iommu = false
    bus: pcie.0
      type pnv-phb3-root-bus

And without an 'id' we will have the following default:

qemu-system-ppc64 -m 4G -machine powernv8,accel=tcg \
    -device pnv-phb3,chip-id=0,index=1

  dev: pnv-phb3, id ""
    index = 1 (0x1)
    chip-id = 0 (0x0)
    x-config-reg-migration-enabled = true
    bypass-iommu = false
    bus: pnv-phb3-root-bus.0
      type pnv-phb3-root-bus

Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20211228193806.1198496-3-danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/pnv: Remove the PHB4 "device-id" property

It's unused.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20211222063817.1541058-4-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/pnv: Remove PHB4 reset handler

The PHB4 reset handler was preparing ground for PHB5 to set
appropriately the device id. We don't need it for the PHB4 since the
device id is already set in the root port complex. PH5 will introduce
its own.

"device-id" property is now useless. It should be removed.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20211222063817.1541058-3-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

ppc/pnv: Change the maximum of PHB3 devices for Power8NVL

The POWER8 processors with a NVLink logic unit have 4 PHB3 devices per
chip.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20211222063817.1541058-2-clg@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>

Merge tag 'pull-misc-20220103' of https://gitlab.com/rth7680/qemu into staging

Fix some meson conversion breakage
Disable check-python-tox
Fix emulation of hppa STBY insn

# gpg: Signature made Mon 03 Jan 2022 09:31:48 AM PST
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [ultimate]

* tag 'pull-misc-20220103' of https://gitlab.com/rth7680/qemu:
  gitlab: Disable check-python-tox
  target/hppa: Fix atomic_store_3 for STBY
  tests/tcg: Unconditionally use 90 second timeout
  tests/tcg: Use $cpu in configure.sh
  meson: Unify mips and mips64 in host_arch

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

gitlab: Disable check-python-tox

Set this test to be manually run, until failures can be fixed.

Suggested-by: John Snow <jsnow@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge tag 'machine-core-20211231' of https://github.com/philmd/qemu into staging

Machine core patches

- Clarify qdev_connect_gpio_out() documentation
- Rework test-smp-parse tests following QOM style
- Introduce CPU cluster topology support (Yanan Wang)
- MAINTAINERS updates (Yanan Wang, Li Zhijian, myself)

# gpg: Signature made Fri 31 Dec 2021 04:45:35 AM PST
# gpg:                using RSA key FAABE75E12917221DCFD6BB2E3E32C2CDEADC0DE
# gpg: Good signature from "Philippe Mathieu-Daudé (F4BUG) <f4bug@amsat.org>" [full]

* tag 'machine-core-20211231' of https://github.com/philmd/qemu:
  MAINTAINERS: email address change
  MAINTAINERS: Change philmd's email address
  MAINTAINERS: Self-recommended as reviewer of "Machine core"
  tests/unit/test-smp-parse: Keep default MIN/MAX CPUs in machine_base_class_init
  tests/unit/test-smp-parse: No need to explicitly zero MachineClass members
  tests/unit/test-smp-parse: Add testcases for CPU clusters
  hw/core/machine: Introduce CPU cluster topology support
  qemu-options: Improve readability of SMP related Docs
  hw/core: Rename smp_parse() -> machine_parse_smp_config()
  tests/unit/test-smp-parse: Constify some pointer/struct
  tests/unit/test-smp-parse: Simplify pointer to compound literal use
  tests/unit/test-smp-parse: Add 'smp-generic-valid' machine type
  tests/unit/test-smp-parse: Add 'smp-generic-invalid' machine type
  tests/unit/test-smp-parse: Add 'smp-with-dies' machine type
  tests/unit/test-smp-parse: Split the 'generic' test in valid / invalid
  tests/unit/test-smp-parse: Pass machine type as argument to tests
  hw/qdev: Rename qdev_connect_gpio_out*() 'input_pin' parameter
  hw/qdev: Correct qdev_connect_gpio_out_named() documentation
  hw/qdev: Correct qdev_init_gpio_out_named() documentation
  hw/qdev: Cosmetic around documentation

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

MAINTAINERS: email address change

Fujitsu's mail service has migrated to O365 months ago, the
lizhijian@cn.fujitsu.com address will stop working on 2022-06-01,
change it to my new email address lizhijian@fujitsu.com.

Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Acked-by: Zhang Chen <chen.zhang@intel.com>
Message-Id: <20211231050901.360-1-lizhijian@cn.fujitsu.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

MAINTAINERS: Change philmd's email address

The philmd@redhat.com email address will stop working on
2022-01-01, change it to my personal email address.

Update .mailmap in case anyone wants to send me an email
because of some past commit I authored.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211231000759.707519-1-philmd@redhat.com>

MAINTAINERS: Self-recommended as reviewer of "Machine core"

I've built interests in the generic machine subsystem and
have also been working on projects related to this part,
self-recommand myself as a reviewer so that I can help to
review some patches familiar to me, and have a chance to
learn more continuously.

Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211228092221.21068-8-wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

tests/unit/test-smp-parse: Keep default MIN/MAX CPUs in machine_base_class_init

Most machine types in test-smp-parse will be OK to have the default
MIN/MAX CPUs except "smp-generic-invalid", let's keep the default
values in machine_base_class_init which will be inherited. And if
we hope a different value for a specific machine, modify it in its
own initialization function.

Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211228092221.21068-7-wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

tests/unit/test-smp-parse: No need to explicitly zero MachineClass members

The default value of the MachineClass members is 0, which
means we don't have to explicitly zero them. Also the value
of "mc->smp_props.prefer_sockets" will be taken care of by
smp_parse_test(), we don't necessarily need the statement
in machine_base_class_init() either.

Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211228092221.21068-6-wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

tests/unit/test-smp-parse: Add testcases for CPU clusters

Add testcases for parsing of the four-level CPU topology hierarchy,
ie sockets/clusters/cores/threads, which will be supported on ARM
virt machines.

Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211228092221.21068-5-wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

hw/core/machine: Introduce CPU cluster topology support

The new Cluster-Aware Scheduling support has landed in Linux 5.16,
which has been proved to benefit the scheduling performance (e.g.
load balance and wake_affine strategy) on both x86_64 and AArch64.

So now in Linux 5.16 we have four-level arch-neutral CPU topology
definition like below and a new scheduler level for clusters.
struct cpu_topology {
    int thread_id;
    int core_id;
    int cluster_id;
    int package_id;
    int llc_id;
    cpumask_t thread_sibling;
    cpumask_t core_sibling;
    cpumask_t cluster_sibling;
    cpumask_t llc_sibling;
}

A cluster generally means a group of CPU cores which share L2 cache
or other mid-level resources, and it is the shared resources that
is used to improve scheduler's behavior. From the point of view of
the size range, it's between CPU die and CPU core. For example, on
some ARM64 Kunpeng servers, we have 6 clusters in each NUMA node,
and 4 CPU cores in each cluster. The 4 CPU cores share a separate
L2 cache and a L3 cache tag, which brings cache affinity advantage.

In virtualization, on the Hosts which have pClusters (physical
clusters), if we can design a vCPU topology with cluster level for
guest kernel and have a dedicated vCPU pinning. A Cluster-Aware
Guest kernel can also make use of the cache affinity of CPU clusters
to gain similar scheduling performance.

This patch adds infrastructure for CPU cluster level topology
configuration and parsing, so that the user can specify cluster
parameter if their machines support it.

Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
Message-Id: <20211228092221.21068-3-wangyanan55@huawei.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
[PMD: Added '(since 7.0)' to @clusters in qapi/machine.json]
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

qemu-options: Improve readability of SMP related Docs

We have a description in qemu-options.hx for each CPU topology
parameter to explain what it exactly means, and also an extra
declaration for the target-specific one, e.g. "for PC only"
when describing "dies", and "for PC, it's on one die" when
describing "cores".

Now we are going to introduce one more non-generic parameter
"clusters", it will make the Doc less readable and  if we still
continue to use the legacy way to describe it.

So let's at first make two tweaks of the Docs to improve the
readability and also scalability:
1) In the -help text: Delete the extra specific declaration and
   describe each topology parameter level by level. Then add a
   note to declare that different machines may support different
   subsets and the actual meaning of the supported parameters
   will vary accordingly.
2) In the rST text: List all the sub-hierarchies currently
   supported in QEMU, and correspondingly give an example of
   -smp configuration for each of them.

Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211228092221.21068-2-wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

hw/core: Rename smp_parse() -> machine_parse_smp_config()

All methods related to MachineState are prefixed with "machine_".
smp_parse() does not need to be an exception. Rename it and
const'ify the SMPConfiguration argument, since it doesn't need
to be modified.

Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Tested-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211216132015.815493-9-philmd@redhat.com>

tests/unit/test-smp-parse: Constify some pointer/struct

Declare structures const when we don't need to modify
them at runtime.

Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Tested-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211216132015.815493-8-philmd@redhat.com>

tests/unit/test-smp-parse: Simplify pointer to compound literal use

We can simply use a local variable (and pass its pointer) instead
of a pointer to a compound literal.

Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Tested-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211216132015.815493-7-philmd@redhat.com>

tests/unit/test-smp-parse: Add 'smp-generic-valid' machine type

Keep the common TYPE_MACHINE class initialization in
machine_base_class_init(), make it abstract, and move
the non-common code to a new class: "smp-generic-valid".

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Message-Id: <20211216132015.815493-6-philmd@redhat.com>

tests/unit/test-smp-parse: Add 'smp-generic-invalid' machine type

Avoid modifying the MachineClass internals by adding the
'smp-generic-invalid' machine, which inherits from TYPE_MACHINE.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Message-Id: <20211216132015.815493-5-philmd@redhat.com>

tests/unit/test-smp-parse: Add 'smp-with-dies' machine type

Avoid modifying the MachineClass internals by adding the
'smp-with-dies' machine, which inherits from TYPE_MACHINE.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Tested-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211216132015.815493-4-philmd@redhat.com>

tests/unit/test-smp-parse: Split the 'generic' test in valid / invalid

Split the 'generic' test in two tests: 'valid' and 'invalid'.
This will allow us to remove the hack which modifies the
MachineClass internal state.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211216132015.815493-3-philmd@redhat.com>

tests/unit/test-smp-parse: Pass machine type as argument to tests

Use g_test_add_data_func() instead of g_test_add_func() so we can
pass the machine type to the tests (we will soon have different
machine types).

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211216132015.815493-2-philmd@redhat.com>

hw/qdev: Rename qdev_connect_gpio_out*() 'input_pin' parameter

@pin is an input where we connect a device output.
Rename it @input_pin to simplify the documentation.

Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20211218130437.1516929-5-f4bug@amsat.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

hw/qdev: Correct qdev_connect_gpio_out_named() documentation

qdev_connect_gpio_out_named() is described as qdev_connect_gpio_out(),
and referring to itself in an endless loop, which is confusing. Fix.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Message-Id: <20211218130437.1516929-4-f4bug@amsat.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

hw/qdev: Correct qdev_init_gpio_out_named() documentation

qdev_init_gpio_out_named() is described as qdev_init_gpio_out(),
and referring to itself in an endless loop, which is confusing. Fix.

Reported-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Message-Id: <20211218130437.1516929-3-f4bug@amsat.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

hw/qdev: Cosmetic around documentation

Add empty lines to have a clearer distinction between different
functions declarations.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Message-Id: <20211218130437.1516929-2-f4bug@amsat.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

target/hppa: Fix atomic_store_3 for STBY

The parallel version of STBY did not take host endianness into
account, and also computed the incorrect address for STBY_E.

Bswap twice to handle the merge and store.  Compute mask inside
the function rather than as a parameter.  Force align the address,
rather than subtracting one.

Generalize the function to system mode by using probe_access().

Cc: qemu-stable@nongnu.org
Tested-by: Helge Deller <deller@gmx.de>
Reported-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tests/tcg: Unconditionally use 90 second timeout

The cross-i386-tci test has timeouts because we're no longer
applying the timeout that we desired.  Hack around it.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Fixes: 23a77b2d18b8 ("build-system: clean up TCG/TCI configury")
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>