hw/net: Declare link using static DEFINE_PROP_LINK() macro

Declare link statically using DEFINE_PROP_LINK().

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20231017140150.44995-7-philmd@linaro.org>

hw/dma: Declare link using static DEFINE_PROP_LINK() macro

Declare link statically using DEFINE_PROP_LINK().

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20231017140150.44995-6-philmd@linaro.org>

hw/scsi/virtio-scsi: Use VIRTIO_SCSI_COMMON() macro

Access QOM parent with the proper QOM VIRTIO_SCSI_COMMON() macro.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20231017140150.44995-5-philmd@linaro.org>

hw/display/virtio-gpu: Use VIRTIO_DEVICE() macro

Access QOM parent with the proper QOM VIRTIO_DEVICE() macro.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20231017140150.44995-4-philmd@linaro.org>

hw/block/vhost-user-blk: Use DEVICE() / VIRTIO_DEVICE() macros

Access QOM parent with the proper QOM [VIRTIO_]DEVICE() macros.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20231017140150.44995-3-philmd@linaro.org>

hw/virtio/virtio-pmem: Replace impossible check by assertion

The get_memory_region() handler is used when (un)plugging the
device, which can only occur *after* it is realized.

virtio_pmem_realize() ensure the instance can not be realized
without 'memdev'. Remove the superfluous check, replacing it
by an assertion.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Message-Id: <20231017140150.44995-2-philmd@linaro.org>

hw/s390x/css-bridge: Realize sysbus device before accessing it

qbus_new() should not be called on unrealized device.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20231019071611.98885-10-philmd@linaro.org>

hw/isa: Realize ISA bridge device before accessing it

qbus_new() should not be called on unrealized device.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20231019071611.98885-9-philmd@linaro.org>

hw/arm/virt: Realize ARM_GICV2M sysbus device before accessing it

sysbus_mmio_map() should not be called on unrealized device.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20231018141151.87466-8-philmd@linaro.org>

hw/acpi: Realize ACPI_GED sysbus device before accessing it

sysbus_mmio_map() should not be called on unrealized device.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20231018141151.87466-7-philmd@linaro.org>

hw/pci-host/bonito: Do not use SysBus API to map local MMIO region

There is no point in exposing an internal MMIO region via
SysBus and directly mapping it in the very same device.

Just map it without using the SysBus API.

Transformation done using the following coccinelle script:

  @@
  expression sbdev;
  expression index;
  expression addr;
  expression subregion;
  @@
  -    sysbus_init_mmio(sbdev, subregion);
       ... when != sbdev
  -    sysbus_mmio_map(sbdev, index, addr);
  +    memory_region_add_subregion(get_system_memory(), addr, subregion);

and manually adding the local 'host_mem' variable to
avoid multiple calls to get_system_memory().

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20231019071611.98885-6-philmd@linaro.org>

hw/misc/allwinner-dramc: Do not use SysBus API to map local MMIO region

There is no point in exposing an internal MMIO region via
SysBus and directly mapping it in the very same device.

Just map it without using the SysBus API.

Transformation done using the following coccinelle script:

  @@
  expression sbdev;
  expression index;
  expression addr;
  expression subregion;
  @@
  -    sysbus_init_mmio(sbdev, subregion);
       ... when != sbdev
  -    sysbus_mmio_map(sbdev, index, addr);
  +    memory_region_add_subregion(get_system_memory(),
  +                                addr, subregion);

  @@
  expression priority;
  @@
  -    sysbus_init_mmio(sbdev, subregion);
       ... when != sbdev
  -    sysbus_mmio_map_overlap(sbdev, index, addr, priority);
  +    memory_region_add_subregion_overlap(get_system_memory(),
  +                                        addr,
  +                                        subregion, priority);

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20231019071611.98885-5-philmd@linaro.org>

hw/misc/allwinner-dramc: Move sysbus_mmio_map call from init -> realize

In order to make the next commit trivial, move the sysbus_init_mmio()
call in allwinner_r40_dramc_init() just before the corresponding
sysbus_mmio_map_overlap() call in allwinner_r40_dramc_realize().

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20231019071611.98885-4-philmd@linaro.org>

hw/i386/intel_iommu: Do not use SysBus API to map local MMIO region

There is no point in exposing an internal MMIO region via
SysBus and directly mapping it in the very same device.

Just map it without using the SysBus API.

Transformation done using the following coccinelle script:

  @@
  expression sbdev;
  expression index;
  expression addr;
  expression subregion;
  @@
  -    sysbus_init_mmio(sbdev, subregion);
       ... when != sbdev
  -    sysbus_mmio_map(sbdev, index, addr);
  +    memory_region_add_subregion(get_system_memory(), addr, subregion);

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20231018141151.87466-3-philmd@linaro.org>

hw/i386/amd_iommu: Do not use SysBus API to map local MMIO region

There is no point in exposing an internal MMIO region via
SysBus and directly mapping it in the very same device.

Just map it without using the SysBus API.

Transformation done using the following coccinelle script:

  @@
  expression sbdev;
  expression index;
  expression addr;
  expression subregion;
  @@
  -    sysbus_init_mmio(sbdev, subregion);
       ... when != sbdev
  -    sysbus_mmio_map(sbdev, index, addr);
  +    memory_region_add_subregion(get_system_memory(), addr, subregion);

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20231018141151.87466-2-philmd@linaro.org>

hw/audio/pcspk: Inline pcspk_init()

pcspk_init() is a legacy init function, inline and remove it.

Since the device is realized using &error_fatal, use the same
error for setting the "pit" link.

Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20231019073307.99608-1-philmd@linaro.org>

hw/intc/spapr_xive: Do not use SysBus API to map local MMIO region

There is no point in exposing an internal MMIO region via
SysBus and directly mapping it in the very same device.

Just map it without using the SysBus API.

Transformation done using the following coccinelle script:

  @@
  expression sbdev;
  expression index;
  expression addr;
  expression subregion;
  @@
  -    sysbus_init_mmio(sbdev, subregion);
       ... when != sbdev
  -    sysbus_mmio_map(sbdev, index, addr);
  +    memory_region_add_subregion(get_system_memory(), addr, subregion);

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20231019131647.19690-8-philmd@linaro.org>

hw/intc/spapr_xive: Move sysbus_init_mmio() calls around

In order to make the next commit trivial, move sysbus_init_mmio()
calls just before the corresponding sysbus_mmio_map() calls.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20231019131647.19690-7-philmd@linaro.org>

hw/ppc/pnv: Do not use SysBus API to map local MMIO region

There is no point in exposing an internal MMIO region via
SysBus and directly mapping it in the very same device.

Just map it without using the SysBus API.

Transformation done using the following coccinelle script:

  @@
  expression sbdev;
  expression index;
  expression addr;
  expression subregion;
  @@
  -    sysbus_init_mmio(sbdev, subregion);
       ... when != sbdev
  -    sysbus_mmio_map(sbdev, index, addr);
  +    memory_region_add_subregion(get_system_memory(), addr, subregion);

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20231019131647.19690-6-philmd@linaro.org>

hw/ppc/pnv_xscom: Do not use SysBus API to map local MMIO region

There is no point in exposing an internal MMIO region via
SysBus and directly mapping it in the very same device.

Just map it without using the SysBus API.

Transformation done using the following coccinelle script:

  @@
  expression sbdev;
  expression index;
  expression addr;
  expression subregion;
  @@
  -    sysbus_init_mmio(sbdev, subregion);
       ... when != sbdev
  -    sysbus_mmio_map(sbdev, index, addr);
  +    memory_region_add_subregion(get_system_memory(), addr, subregion);

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20231019131647.19690-5-philmd@linaro.org>

hw/ppc/pnv_xscom: Move sysbus_mmio_map() call within pnv_xscom_init()

In order to make the next commit trivial, move sysbus_init_mmio()
calls just before the corresponding sysbus_mmio_map() calls.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20231019131647.19690-4-philmd@linaro.org>

hw/ppc/pnv_xscom: Rename pnv_xscom_realize(Error **) -> pnv_xscom_init()

pnv_xscom_realize() is not used to *realize* QDev object, rename
it as pnv_xscom_init(). The Error** argument is unused: remove it.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20231019131647.19690-3-philmd@linaro.org>

hw/ppc/spapr_vio: Realize SPAPR_VIO_BRIDGE device before accessing it

qbus_new() should not be called on unrealized device.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20231019131647.19690-2-philmd@linaro.org>

hw/intc/apic: Use ERRP_GUARD() in apic_common_realize()

APICCommonClass::realize() is a DeviceRealize() handler which
take an Error** parameter and can fail. Do not proceed further
on failure.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Message-Id: <20231003082728.83496-2-philmd@linaro.org>

hw/s390x: Clean up global variable shadowing in quiesce_powerdown_req()

Fix:

  hw/s390x/sclpquiesce.c:90:22: error: declaration shadows a variable in the global scope [-Werror,-Wshadow]
      QuiesceNotifier *qn = container_of(n, QuiesceNotifier, notifier);
                       ^
  hw/s390x/sclpquiesce.c:86:3: note: previous declaration is here
  } qn;
    ^

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Message-Id: <20231010115048.11856-7-philmd@linaro.org>

hw/pci: Clean up global variable shadowing of address_space_io variable

Fix:

  hw/pci/pci.c:504:54: error: declaration shadows a variable in the global scope [-Werror,-Wshadow]
                                         MemoryRegion *address_space_io,
                                                       ^
  hw/pci/pci.c:533:38: error: declaration shadows a variable in the global scope [-Werror,-Wshadow]
                         MemoryRegion *address_space_io,
                                       ^
  hw/pci/pci.c:543:40: error: declaration shadows a variable in the global scope [-Werror,-Wshadow]
                           MemoryRegion *address_space_io,
                                         ^
  hw/pci/pci.c:590:45: error: declaration shadows a variable in the global scope [-Werror,-Wshadow]
                                MemoryRegion *address_space_io,
                                              ^
  include/exec/address-spaces.h:35:21: note: previous declaration is here
  extern AddressSpace address_space_io;
                      ^

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20231010115048.11856-6-philmd@linaro.org>

hw/acpi/pcihp: Clean up global variable shadowing in acpi_pcihp_init()

Fix:

  hw/acpi/pcihp.c:499:36: error: declaration shadows a variable in the global scope [-Werror,-Wshadow]
                       MemoryRegion *address_space_io,
                                     ^
  include/exec/address-spaces.h:35:21: note: previous declaration is here
  extern AddressSpace address_space_io;
                      ^

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Ani Sinha <anisinha@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20231010115048.11856-5-philmd@linaro.org>

hw/sparc64/ebus: Access memory regions via pci_address_space_io()

PCI functions are plugged on a PCI bus. They can only access
external memory regions via the bus.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20231011185954.10337-5-philmd@linaro.org>

hw/pci-host/sh_pcic: Replace magic value by proper definition

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Message-Id: <20231012041237.22281-4-philmd@linaro.org>

hw/pci-host/sh_pcic: Correct PCI host / devfn#0 function names

Host bridge device and PCI function #0 are inverted.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Message-Id: <20231012041237.22281-3-philmd@linaro.org>

hw/pci-host/sh_pcic: Declare CPU QOM types using DEFINE_TYPES() macro

When multiple QOM types are registered in the same file,
it is simpler to use the the DEFINE_TYPES() macro. In
particular because type array declared with such macro
are easier to review.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Message-Id: <20231012041237.22281-2-philmd@linaro.org>

hw/misc/mips_itu: Make MIPSITUState target agnostic

When prototyping a heterogenous machine including the ITU,
we get:

  include/hw/misc/mips_itu.h:76:5: error: unknown type name 'MIPSCPU'
      MIPSCPU *cpu0;
      ^

MIPSCPU is declared in the target specific "cpu.h" header,
but we don't want to include it, because "cpu.h" is target
specific and its inclusion taints all files including
"mips_itu.h", which become target specific too. We can
however use the 'ArchCPU *' type in the public header.
By keeping the TYPE_MIPS_CPU QOM type check in the link
property declaration, QOM core code will still check the
property is a correct MIPS CPU.

TYPE_MIPS_ITU is still built per-(MIPS)target, but its header
can now be included by other targets.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20231009171443.12145-4-philmd@linaro.org>

hw/misc/mips_itu: Declare itc_reconfigure() in 'hw/misc/mips_itu.h'

We already provide "hw/misc/mips_itu.h" to declare prototype
related to MIPSITUState. Move itc_reconfigure() declaration
there.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20231009171443.12145-3-philmd@linaro.org>

hw/mips: Merge 'hw/mips/cpudevs.h' with 'target/mips/cpu.h'

"hw/mips/cpudevs.h" contains declarations which are specific
to the MIPS architecture; it doesn't make sense for these to
be called from a non-MIPS architecture. Move the declarations
to "target/mips/cpu.h".

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20231009171443.12145-2-philmd@linaro.org>

hw/mips/malta: Use sdram_type enum from 'hw/i2c/smbus_eeprom.h'

Since commit 93198b6cad ("i2c: Split smbus into parts") the SDRAM
types are enumerated as sdram_type in "hw/i2c/smbus_eeprom.h".

Using the enum removes this global shadow warning:

  hw/mips/malta.c:209:12: error: declaration shadows a variable in the global scope [-Werror,-Wshadow]
      enum { SDR = 0x4, DDR2 = 0x8 } type;
             ^
  include/hw/i2c/smbus_eeprom.h:33:19: note: previous declaration is here
  enum sdram_type { SDR = 0x4, DDR = 0x7, DDR2 = 0x8 };
                    ^

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: BALATON Zoltan <balaton@eik.bme.hu>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20231009092127.49778-1-philmd@linaro.org>

hw/sd/sdhci: Block Size Register bits [14:12] is lost

Block Size Register bits [14:12] is SDMA Buffer Boundary, it is missed
in register write, but it is needed in SDMA transfer. e.g. it will be
used in sdhci_sdma_transfer_multi_blocks to calculate boundary_ variables.

Missing this field will cause wrong operation for different SDMA Buffer
Boundary settings.

Fixes: d7dfca0807 ("hw/sdhci: introduce standard SD host controller")
Fixes: dfba99f17f ("hw/sdhci: Fix DMA Transfer Block Size field")
Signed-off-by: Lu Gao <lu.gao@verisilicon.com>
Signed-off-by: Jianxian Wen <jianxian.wen@verisilicon.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-ID: <20220321055618.4026-1-lu.gao@verisilicon.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

memory: follow Error API guidelines

Return true/false on success/failure.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20231009075310.153617-1-marcandre.lureau@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

memory: drop needless argument

The argument is unused since commit bdc44640c ("cpu: Use QTAILQ for CPU list").

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20231009075231.150568-1-marcandre.lureau@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

buildsys: Only display Objective-C information when Objective-C is used

When configuring with '--disable-cocoa --disable-coreaudio'
on Darwin, we get:

 meson.build:4081:58: ERROR: Tried to access compiler for language "objc", not specified for host machine.
 meson.build:4097:47: ERROR: Tried to access unknown option 'objc_args'.

Instead of unconditionally display Objective-C informations
on Darwin, display them when Objective-C is discovered.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-Id: <20231009093812.52915-1-philmd@linaro.org>

cutils: Fix get_relocated_path on Windows

get_relocated_path() did not have error handling for PathCchSkipRoot()
because a path given to get_relocated_path() was expected to be a valid
path containing a drive letter or UNC server/share path elements on
Windows, but sometimes it turned out otherwise.

The paths passed to get_relocated_path() are defined by macros generated
by Meson. Meson in turn uses a prefix given by the configure script to
generate them. For Windows, the script passes /qemu as a prefix to
Meson by default.

As documented in docs/about/build-platforms.rst, typically MSYS2 is used
for the build system, but it is also possible to use Linux as well. When
MSYS2 is used, its Bash variant recognizes /qemu as a MSYS2 path, and
converts it to a Windows path, adding the MSYS2 prefix including a drive
letter or UNC server/share path elements. Such a conversion does not
happen on a shell on Linux however, and /qemu will be passed as is in
the case.

Implement a proper error handling of PathCchSkipRoot() in
get_relocated_path() so that it can handle a path without a drive letter
or UNC server/share path elements.

Reported-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20231005064726.6945-1-akihiko.odaki@daynix.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

MAINTAINERS: Add include/hw/openrisc/ to the OpenRISC section

hw/openrisc/ is already listed here, so we should mention
the folder for the headers here, too.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20231017154242.234133-1-thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

MAINTAINERS: Add include/hw/intc/loongson_liointc.h to the Loongson-3 virt section

The corresponding .c file is already listed here, so we should
mention the header here, too.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Acked-by: Song Gao <gaosong@loongson.cn>
Message-ID: <20231017153345.233807-1-thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

MAINTAINERS: Add hw/input/lasips2.c to the HPPA machine section

hw/input/lasips2.c and the corresponding header include/hw/input/lasips2.h
are only used by the HPPA machine, so add them to the corresponding section
in the MAINTAINERS file.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-ID: <20231017151933.213780-1-thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

MAINTAINERS: Split vt82c686 out of fuloong2e

The VIA south bridges are now mostly used by other machines not just
fuloong2e so split off into a separate section and take maintainership.

Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Message-ID: <20231015141517.219317456A7@zero.eik.bme.hu>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

mailmap: update email addresses for Luc Michel

Map my old and now invalid work email addresses to my personal one.

Signed-off-by: Luc Michel <luc@lmichel.fr>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20230929140326.2056658-2-luc@lmichel.fr>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

Merge tag 'pull-lu-20231018' of https://gitlab.com/rth7680/qemu into staging

linux-user/mips: fix abort on integer overflow
linux-user/sh4: Fix crashes on signal delivery
linux-user/loongarch: Enable LSX/LASX in HWCAP
linux-user: Fixes for zero_bss
linux-user: Propagate failure in mmap_reserve_or_unmap back to target_munmap
linux-user: Detect and report host crashes
linux-user: Remap guest SIGABRT

# -----BEGIN PGP SIGNATURE-----
#
# iQFRBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAmUwapYdHHJpY2hhcmQu
# aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV92UAf/RSsFWwCBAqt1WKIK
# 7/7F8AF7WW1Hhjy3bHLjNnzgsDeWYfdIVxMGfF9IYKrYMeEqFeBeQ+vcOe9LTAvW
# fEZkA//V+LosiYCwtVGBXyCbeXYxoONMp/taRv6lVHoqVU7aSlbXsYqwePcUtPWq
# r/V+Ru5vssqMueBdE9+E53JPewGPVw8xQE+xGgd1TZIeHWgegZHBzKWVap/3noey
# dKjTig3yxXXg1gQJLCRw+a6bl8oCl4vEluGsLzh5P8aV1imjvFPXkR2w1vQOC3ws
# 8DmyMcPEcsY4D2WLeTAGMheURLRzs5141nT0fQCOB4yzO/I8zYFIG12xCxgWoQkg
# kW9XZw==
# =5oFh
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 18 Oct 2023 16:30:30 PDT
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* tag 'pull-lu-20231018' of https://gitlab.com/rth7680/qemu:
  linux-user: Remap guest SIGABRT
  linux-user: Detect and report host SIGILL, SIGFPE, SIGTRAP
  linux-user: Split out host_sig{segv,bus}_handler
  linux-user: Simplify signal_init
  linux-user: Map unsupported signals to an out-of-bounds value
  linux-user: Only register handlers for core_dump_signal by default
  linux-user: Detect and report host crashes
  linux-user: Exit not abort in die_with_backtrace
  linux-user: Split out die_with_signal
  linux-user: Propagate failure in mmap_reserve_or_unmap back to target_munmap
  linux-user/elfload: Enable LSX/LASX in HWCAP for LoongArch
  linux-user/sh4: Fix crashes on signal delivery
  linux-user/mips: fix abort on integer overflow
  linux-user: Fixes for zero_bss

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Merge tag 'migration-20231018-pull-request' of https://gitlab.com/juan.quintela/qemu into staging

Migration Pull request (20231018)

In this pull request:
- RDMA cleanups
- compression cleanups

CI: https://gitlab.com/juan.quintela/qemu/-/pipelines/1040780020

Please apply.

PD.  I tried to get the deprecated bits integrated, but I broke
     qemu-iotests duer to blk warning.  Will resend it.

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEGJn/jt6/WMzuA0uC9IfvGFhy1yMFAmUvrjQACgkQ9IfvGFhy
# 1yODnQ/+OKaOQMAEtJsJ1B67394VRjpGDd0K47U3uewJJ26XRMUy4uw0zeGYdiBc
# VFjrX1NJu4jRZBOdhRzZQLZU9wDEGY/8zIL/sJB55X/gv1EysDB3IrNCWosNL8SS
# weeYu9qkxsB5aJfM4Lp6XnPIplb7PIMSqX380sUGcK7uVLo3x3H8PgFxQszG5ZMs
# 8OqhOdxZ8jPc7gUOxPcA0n/L6pJcfnuK1/8Vlf5wbkdD+lyVCs0QDTSgX8AnS5hd
# DniV2nMFkVvNkOhDG3X8qr8FyjyQ9eyJxxw/2Nt+0201UIiCirz3U2T6EMejCYOv
# LtIaaNaNHmEw5OdfSBhEjXOy7gHavcn+/LlUexYJQSiB/CXrdgh3jpSFmlAzcAY2
# Si514BRty6WX43f+698PSAKF4XaRnPGtvVCv7ubrFb1qVrg8DTEnYXNO+LadqSBS
# bu1TpRK1iVgKnApQN1SQr26MRAzU+U+yqz/MB9QzIGwonM2TEWCF6c5Sqq6/RK8S
# IIDu4s/NTx0wVWrR2rAZv335ANGa7oa1z2LykxcXmBJozqWAOgo6wzZJJ3klcpjZ
# Li39m2jzx36oSKqudYANxj2Ds0MvXStnd2ZX9mQiEB++S2SY2Z8he57HESQwNzf0
# 0Z61eqdHr7th4zfLz/akiiZnRs66A+6/LFRFpoKedSAABE/sKY8=
# =gNbf
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 18 Oct 2023 03:06:44 PDT
# gpg:                using RSA key 1899FF8EDEBF58CCEE034B82F487EF185872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>" [full]
# gpg:                 aka "Juan Quintela <quintela@trasno.org>" [full]
# Primary key fingerprint: 1899 FF8E DEBF 58CC EE03  4B82 F487 EF18 5872 D723

* tag 'migration-20231018-pull-request' of https://gitlab.com/juan.quintela/qemu:
  migration: save_zero_page() can take block through pss
  migration: control_save_page() can take block through pss
  migration: save_compress_page() can take block through pss
  migration: Print block status when needed
  migration: Use "i" as an for index in ram-compress.c
  migration: Simplify decompress_data_with_multi_threads()
  migration: Move update_compress_threads_counts() to ram-compress.c
  migration: Create ram_compressed_pages()
  migration: Create populate_compress()
  migration: Move compression_counters cleanup ram-compress.c
  migration: RDMA is not compatible with anything else

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

linux-user: Remap guest SIGABRT

Distinguish host SIGABRT from guest SIGABRT by mapping
the guest signal onto one of the host RT signals.

This prevents a cycle by which a host assertion failure
is caught and handled by host_signal_handler, queued for
the guest, and then we attempt to continue past the
host abort.  What happens next depends on the host libc,
but is neither good nor helpful.

Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Detect and report host SIGILL, SIGFPE, SIGTRAP

These signals, when not spoofed via kill(), are always bugs.
Use die_from_signal to report this sensibly.

Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Split out host_sig{segv,bus}_handler

Make host_signal_handler slightly easier to read.

Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Simplify signal_init

Install the host signal handler at the same time we are
probing the target signals for SIG_IGN/SIG_DFL.  Ignore
unmapped target signals.

Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Map unsupported signals to an out-of-bounds value

Do not return a valid signal number in one domain
when given an invalid signal number in the other domain.

Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Only register handlers for core_dump_signal by default

The set of fatal signals is really immaterial.  If one arrives,
and is unhandled, then the qemu process dies and the parent gets
the correct signal.

It is only for those signals which we would like to perform a
guest core dump instead of a host core dump that we need to catch.

Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Detect and report host crashes

If there is an internal program error in the qemu source code which
raises SIGSEGV or SIGBUS, we currently assume the signal belongs to
the guest.  With an artificial error introduced, we will now print

   QEMU internal SIGSEGV {code=MAPERR, addr=(nil)}

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20230812164314.352131-1-deller@gmx.de>
[rth: Use in_code_gen_buffer and die_with_signal; drop backtrace]
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Exit not abort in die_with_backtrace

This line is supposed to be unreachable, but if we're going to
have it at all, SIGABRT via abort() is subject to the same signal
peril that created this function in the first place.

We can _exit immediately without peril.

Acked-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Split out die_with_signal

Because we trap so many signals for use by the guest,
we have to take extra steps to exit properly.

Acked-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Propagate failure in mmap_reserve_or_unmap back to target_munmap

Do not assert success, but return any failure received.
Additionally, fix the method of earlier error return in target_munmap.

Reported-by: Andreas Schwab <schwab@suse.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user/elfload: Enable LSX/LASX in HWCAP for LoongArch

Since support for LSX and LASX is landed in QEMU recently, we can update
HWCAPS accordingly.

Signed-off-by: Jiajie Chen <c@jia.je>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20231001085315.1692667-1-c@jia.je>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user/sh4: Fix crashes on signal delivery

sh4 uses gUSA (general UserSpace Atomicity) to provide atomicity on CPUs
that don't have atomic instructions. A gUSA region that adds 1 to an
atomic variable stored in @R2 looks like this:

  4004b6:       03 c7           mova    4004c4 <gusa+0x10>,r0
  4004b8:       f3 61           mov     r15,r1
  4004ba:       09 00           nop
  4004bc:       fa ef           mov     #-6,r15
  4004be:       22 63           mov.l   @r2,r3
  4004c0:       01 73           add     #1,r3
  4004c2:       32 22           mov.l   r3,@r2
  4004c4:       13 6f           mov     r1,r15

R0 contains a pointer to the end of the gUSA region
R1 contains the saved stack pointer
R15 contains negative length of the gUSA region

When this region is interrupted by a signal, the kernel detects if
R15 >= -128U. If yes, the kernel rolls back PC to the beginning of the
region and restores SP by copying R1 to R15.

The problem happens if we are interrupted by a signal at address 4004c4.
R15 still holds the value -6, but the atomic value was already written by
an instruction at address 4004c2. In this situation we can't undo the
gUSA. The function unwind_gusa does nothing, the signal handler attempts
to push a signal frame to the address -6 and crashes.

This patch fixes it, so that if we are interrupted at the last instruction
in a gUSA region, we copy R1 to R15 to restore the correct stack pointer
and avoid crashing.

There's another bug: if we are interrupted in a delay slot, we save the
address of the instruction in the delay slot. We must save the address of
the previous instruction.

Cc: qemu-stable@nongnu.org
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Reviewed-by: Yoshinori Sato <ysato@users.sourcefoege.jp>
Message-Id: <b16389f7-6c62-70b7-59b3-87533c0bcc@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user/mips: fix abort on integer overflow

QEMU mips userspace emulation crashes with "qemu: unhandled CPU exception
0x15 - aborting" when one of the integer arithmetic instructions detects
an overflow.

This patch fixes it so that it delivers SIGFPE with FPE_INTOVF instead.

Cc: qemu-stable@nongnu.org
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Message-Id: <3ef979a8-3ee1-eb2d-71f7-d788ff88dd11@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user: Fixes for zero_bss

The previous change, 2d385be6152, assumed !PAGE_VALID meant that
the page would be unmapped by the elf image.  However, since we
reserved the entire image space via mmap, PAGE_VALID will always
be set.  Instead, assume PROT_NONE for the same condition.

Furthermore, assume bss is only ever present for writable segments,
and that there is no page overlap between PT_LOAD segments.
Instead of an assert, return false to indicate failure.

Cc: qemu-stable@nongnu.org
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1854
Fixes: 2d385be6152 ("linux-user: Do not adjust zero_bss for host page size")
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge tag 'pull-vfio-20231018' of https://github.com/legoater/qemu into staging

vfio queue:

* Support for VFIODisplay migration with ramfb
* Preliminary work for IOMMUFD support

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEoPZlSPBIlev+awtgUaNDx8/77KEFAmUvlEYACgkQUaNDx8/7
# 7KFlaw//X2053de2eTdo38/UMSzi5ACWWn2j1iGQZf/3+J2LcdlixZarZr/2DN56
# 4axmwF6+GKozt5+EnvWtgodDn6U9iyMNaAB3CGBHFHsH8uqKeZd/Ii754q4Rcmy9
# ZufBOPWm9Ff7s2MMFiAZvso75jP2wuwVEe1YPRjeJnsNSNIJ6WZfemh3Sl96yRBb
# r38uqzqetKwl7HziMMWP3yb8v+dU8A9bqI1hf1FZGttfFz3XA+pmjXKA6XxdfiZF
# AAotu5x9w86a08sAlr/qVsZFLR37oQykkXM0D840DafJDyr5fbJiq8cwfOjMw9+D
# w6+udRm5KoBWPsvb/T3dR88GRMO22PChjH9Vjl51TstMNhdTxuKJTKhhSoUFZbXV
# 8CMjwfALk5ggIOyCk1LRd04ed+9qkqgcbw1Guy5pYnyPnY/X6XurxxaxS6Gemgtn
# UvgRYhSjio+LgHLO77IVkWJMooTEPzUTty2Zxa7ldbbE+utPUtsmac9+1m2pnpqk
# 5VQmB074QnsJuvf+7HPU6vYCzQWoXHsH1UY/A0fF7MPedNUAbVYzKrdGPyqEMqHy
# xbilAIaS3oO0pMT6kUpRv5c5vjbwkx94Nf/ii8fQVjWzPfCcaF3yEfaam62jMUku
# stySaRpavKIx2oYLlucBqeKaBGaUofk13gGTQlsFs8pKCOAV7r4=
# =s0fN
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 18 Oct 2023 04:16:06 EDT
# gpg:                using RSA key A0F66548F04895EBFE6B0B6051A343C7CFFBECA1
# gpg: Good signature from "Cédric Le Goater <clg@redhat.com>" [unknown]
# gpg:                 aka "Cédric Le Goater <clg@kaod.org>" [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: A0F6 6548 F048 95EB FE6B  0B60 51A3 43C7 CFFB ECA1

* tag 'pull-vfio-20231018' of https://github.com/legoater/qemu: (22 commits)
  hw/vfio: add ramfb migration support
  ramfb-standalone: add migration support
  ramfb: add migration support
  vfio/pci: Remove vfio_detach_device from vfio_realize error path
  vfio/ccw: Remove redundant definition of TYPE_VFIO_CCW
  vfio/ap: Remove pointless apdev variable
  vfio/pci: Fix a potential memory leak in vfio_listener_region_add
  vfio/common: Move legacy VFIO backend code into separate container.c
  vfio/common: Introduce a global VFIODevice list
  vfio/common: Store the parent container in VFIODevice
  vfio/common: Introduce a per container device list
  vfio/common: Move VFIO reset handler registration to a group agnostic function
  vfio/ccw: Use vfio_[attach/detach]_device
  vfio/ap: Use vfio_[attach/detach]_device
  vfio/platform: Use vfio_[attach/detach]_device
  vfio/pci: Introduce vfio_[attach/detach]_device
  vfio/common: Extract out vfio_kvm_device_[add/del]_fd
  vfio/common: Introduce vfio_container_add|del_section_window()
  vfio/common: Propagate KVM_SET_DEVICE_ATTR error if any
  vfio/common: Move IOMMU agnostic helpers to a separate file
  ...

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Merge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging

* build system and Python cleanups
* fix netbsd VM build
* allow non-relocatable installs
* allow using command line options to configure qemu-ga
* target/i386: check intercept for XSETBV
* target/i386: fix CPUID_HT exposure

# -----BEGIN PGP SIGNATURE-----
#
# iQFIBAABCAAyFiEE8TM4V0tmI4mGbHaCv/vSX3jHroMFAmUvkQQUHHBib256aW5p
# QHJlZGhhdC5jb20ACgkQv/vSX3jHroM3pQgArXCsmnsjlng1chjCvKnIuVmaTYZ5
# aC9pcx7TlyM0+XWtTN0NQhFt71Te+3ioReXIQRvy5O68RNbEkiu8LXfOJhWAHbWk
# vZVtzHQuOZVizeZtUruKlDaw0nZ8bg+NI4aGLs6rs3WphEAM+tiLnZJ0BouiedKS
# e/COB/Hqjok+Ntksbfv5q7XpWjwQB0y2073vM1Mcf0ToOWFLFdL7x0SZ3hxyYlYl
# eoefp/8kbWeUWA7HuoOKmpiLIxmKnY7eXp+UCvdnEhnSce9sCxpn2nzqqLuPItTK
# V3GrJ2//+lrekPHyQvb8IjUMUrPOmzf8GadIE0tkfdHjEP72IsHk0VX81A==
# =rPte
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 18 Oct 2023 04:02:12 EDT
# gpg:                using RSA key F13338574B662389866C7682BFFBD25F78C7AE83
# gpg:                issuer "pbonzini@redhat.com"
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]
# 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: (32 commits)
  configure: define "pkg-config" in addition to "pkgconfig"
  meson: add a note on why we use config_host for program paths
  meson-buildoptions: document the data at the top
  configure, meson: use command line options to configure qemu-ga
  configure: unify handling of several Debian cross containers
  configure: move environment-specific defaults to config-meson.cross
  configure: move target-specific defaults to an external machine file
  configure: remove some dead cruft
  configure: clean up PIE option handling
  configure: clean up plugin option handling
  configure, tests/tcg: simplify GDB conditionals
  tests/tcg/arm: move non-SVE tests out of conditional
  hw/remote: move stub vfu_object_set_bus_irq out of stubs/
  hw/xen: cleanup sourcesets
  configure: clean up handling of CFI option
  meson, cutils: allow non-relocatable installs
  meson: do not use set10
  meson: do not build shaders by default
  tracetool: avoid invalid escape in Python string
  tests/vm: avoid invalid escape in Python string
  ...

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

migration: save_zero_page() can take block through pss

Reviewed-by: Lukas Straub <lukasstraub2@web.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration: control_save_page() can take block through pss

Reviewed-by: Lukas Straub <lukasstraub2@web.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration: save_compress_page() can take block through pss

Reviewed-by: Lukas Straub <lukasstraub2@web.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>

hw/vfio: add ramfb migration support

Add a "VFIODisplay" subsection whenever "x-ramfb-migrate" is turned on.

Turn it off by default on machines <= 8.1 for compatibility reasons.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
[ clg:  - checkpatch fixes
   - improved warn_report() in vfio_realize() ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>

ramfb-standalone: add migration support

Add a "ramfb-dev" section whenever "x-migrate" is turned on. Turn it off
by default on machines <= 8.1 for compatibility reasons.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

ramfb: add migration support

Implementing RAMFB migration is quite straightforward. One caveat is to
treat the whole RAMFBCfg as a blob, since that's what is exposed to the
guest directly. This avoid having to fiddle with endianness issues if we
were to migrate fields individually as integers.

The devices using RAMFB will have to include ramfb_vmstate in their
migration description.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/pci: Remove vfio_detach_device from vfio_realize error path

In vfio_realize, on the error path, we currently call
vfio_detach_device() after a successful vfio_attach_device.
While this looks natural, vfio_instance_finalize also induces
a vfio_detach_device(), and it seems to be the right place
instead as other resources are released there which happen
to be a prerequisite to a successful UNSET_CONTAINER.

So let's rely on the finalize vfio_detach_device call to free
all the relevant resources.

Fixes: a28e06621170 ("vfio/pci: Introduce vfio_[attach/detach]_device")
Reported-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/ccw: Remove redundant definition of TYPE_VFIO_CCW

No functional changes.

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/ap: Remove pointless apdev variable

No need to double-cast, call VFIO_AP_DEVICE() on DeviceState.

No functional changes.

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Tony Krowiak <akrowiak@linux.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/pci: Fix a potential memory leak in vfio_listener_region_add

When there is an failure in vfio_listener_region_add() and the section
belongs to a ram device, there is an inaccurate error report which should
never be related to vfio_dma_map failure. The memory holding err is also
incrementally leaked in each failure.

Fix it by reporting the real error and free it.

Fixes: 567b5b309ab ("vfio/pci: Relax DMA map errors for MMIO regions")
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Move legacy VFIO backend code into separate container.c

Move all the code really dependent on the legacy VFIO container/group
into a separate file: container.c. What does remain in common.c is
the code related to VFIOAddressSpace, MemoryListeners, migration and
all other general operations.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Introduce a global VFIODevice list

Some functions iterate over all the VFIODevices. This is currently
achieved by iterating over all groups/devices. Let's
introduce a global list of VFIODevices simplifying that scan.

This will also be useful while migrating to IOMMUFD by hiding the
group specificity.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Suggested-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Store the parent container in VFIODevice

let's store the parent contaienr within the VFIODevice.
This simplifies the logic in vfio_viommu_preset() and
brings the benefice to hide the group specificity which
is useful for IOMMUFD migration.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Introduce a per container device list

Several functions need to iterate over the VFIO devices attached to
a given container.  This is currently achieved by iterating over the
groups attached to the container and then over the devices in the group.
Let's introduce a per container device list that simplifies this
search.

Per container list is used in below functions:
vfio_devices_all_dirty_tracking
vfio_devices_all_device_dirty_tracking
vfio_devices_all_running_and_mig_active
vfio_devices_dma_logging_stop
vfio_devices_dma_logging_start
vfio_devices_query_dirty_bitmap

This will also ease the migration of IOMMUFD by hiding the group
specificity.

Suggested-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Move VFIO reset handler registration to a group agnostic function

Move the reset handler registration/unregistration to a place that is not
group specific. vfio_[get/put]_address_space are the best places for that
purpose.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/ccw: Use vfio_[attach/detach]_device

Let the vfio-ccw device use vfio_attach_device() and
vfio_detach_device(), hence hiding the details of the used
IOMMU backend.

Note that the migration reduces the following trace
"vfio: subchannel %s has already been attached" (featuring
cssid.ssid.devid) into "device is already attached"

Also now all the devices have been migrated to use the new
vfio_attach_device/vfio_detach_device API, let's turn the
legacy functions into static functions, local to container.c.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/ap: Use vfio_[attach/detach]_device

Let the vfio-ap device use vfio_attach_device() and
vfio_detach_device(), hence hiding the details of the used
IOMMU backend.

We take the opportunity to use g_path_get_basename() which
is prefered, as suggested by
3e015d815b ("use g_path_get_basename instead of basename")

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Matthew Rosato <mjrosato@linux.ibm.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/platform: Use vfio_[attach/detach]_device

Let the vfio-platform device use vfio_attach_device() and
vfio_detach_device(), hence hiding the details of the used
IOMMU backend.

Drop the trace event for vfio-platform as we have similar
one in vfio_attach_device.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/pci: Introduce vfio_[attach/detach]_device

We want the VFIO devices to be able to use two different
IOMMU backends, the legacy VFIO one and the new iommufd one.

Introduce vfio_[attach/detach]_device which aim at hiding the
underlying IOMMU backend (IOCTLs, datatypes, ...).

Once vfio_attach_device completes, the device is attached
to a security context and its fd can be used. Conversely
When vfio_detach_device completes, the device has been
detached from the security context.

At the moment only the implementation based on the legacy
container/group exists. Let's use it from the vfio-pci device.
Subsequent patches will handle other devices.

We also take benefit of this patch to properly free
vbasedev->name on failure.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Extract out vfio_kvm_device_[add/del]_fd

Introduce two new helpers, vfio_kvm_device_[add/del]_fd
which take as input a file descriptor which can be either a group fd or
a cdev fd. This uses the new KVM_DEV_VFIO_FILE VFIO KVM device group,
which aliases to the legacy KVM_DEV_VFIO_GROUP.

vfio_kvm_device_[add/del]_group then call those new helpers.

Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Introduce vfio_container_add|del_section_window()

Introduce helper functions that isolate the code used for
VFIO_SPAPR_TCE_v2_IOMMU.

Those helpers hide implementation details beneath the container object
and make the vfio_listener_region_add/del() implementations more
readable. No code change intended.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Propagate KVM_SET_DEVICE_ATTR error if any

In the VFIO_SPAPR_TCE_v2_IOMMU container case, when
KVM_SET_DEVICE_ATTR fails, we currently don't propagate the
error as we do on the vfio_spapr_create_window() failure
case. Let's align the code. Take the opportunity to
reword the error message and make it more explicit.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Move IOMMU agnostic helpers to a separate file

Move low-level iommu agnostic helpers to a separate helpers.c
file. They relate to regions, interrupts, device/region
capabilities and etc.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

linux-headers: Add iommufd.h

Since commit da3c22c74a3c ("linux-headers: Update to Linux v6.6-rc1"),
linux-headers has been updated to v6.6-rc1.

As previous patch added iommufd.h to update-linux-headers.sh,
run the script again against TAG v6.6-rc1 to have iommufd.h included.

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

scripts/update-linux-headers: Add iommufd.h

Update the script to import iommufd.h

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

configure: define "pkg-config" in addition to "pkgconfig"

Meson used to allow both "pkgconfig" and "pkg-config" entries in machine
files; the former was used for dependency lookup and the latter
was used as return value for "find_program('pkg-config')", which is a less
common use-case and one that QEMU does not need.

This inconsistency is going to be fixed by Meson 1.3, which will deprecate
"pkgconfig" in favor of "pkg-config" (the less common one, but it makes
sense because it matches the name of the binary). For backward
compatibility it is still allowed to define both, so do that in the
configure-generated machine file.

Related: https://github.com/mesonbuild/meson/pull/12385
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson: add a note on why we use config_host for program paths

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson-buildoptions: document the data at the top

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: use command line options to configure qemu-ga

Preserve the functionality of the environment variables, but
allow using the command line instead.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: unify handling of several Debian cross containers

The Debian and GNU architecture names match very often, even though
there are common cases (32-bit Arm or 64-bit x86) where they do not
and other cases in which the GNU triplet is actually a quadruplet.
But it is still possible to group the common case into a single
case inside probe_target_compiler.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: move environment-specific defaults to config-meson.cross

Store the -Werror and SMBD defaults in the machine file, which still allows
them to be overridden on the command line and enables automatic parsing
of the related options.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: move target-specific defaults to an external machine file

Enable Windows-specific defaults with a machine file, so that related
options can be automatically parsed and included in the help message.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: remove some dead cruft

print_error is only invoked in one place, and $git is unused.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: clean up PIE option handling

Keep together all the conditions that lead to disabling PIE.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: clean up plugin option handling

Keep together all the conditions that lead to disabling plugins, and
remove now-dead code.

Since the option was not in SKIP_OPTIONS, it was present twice in
the help message, both from configure and from meson-buildoptions.sh.
Remove the duplication and take the occasion to document the option as
autodetected, which it is.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, tests/tcg: simplify GDB conditionals

Unify HAVE_GDB_BIN (currently in config-host.mak) and
HOST_GDB_SUPPORTS_ARCH into a single GDB variable in
config-target.mak.

Reviewed-by: Emmanouil Pitsidianakis <manos.pitsidianakis@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

tests/tcg/arm: move non-SVE tests out of conditional

test-aes, sha1-vector and sha512-vector need not be conditional on
$(CROSS_CC_HAS_SVE), reorganize the "if"s to move them outside.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>