Re: [PATCH v4 0/9] user-mode: Prune build dependencies (part 2)

[Philippe Mathieu-Daudé]

On 9/12/20 11:16 AM, Paolo Bonzini wrote:
> On 10/09/20 17:22, Philippe Mathieu-Daudé wrote:
>> This is the second part of a series reducing user-mode
>> dependencies. By stripping out unused code, the build
>> and testing time is reduced (as is space used by objects).
>>
>> Part 2:
>> - Extract code not related to user-mode from qapi/misc.json
>>
>> Since v3:
>> - Rebased (Meson)
>> - Addressed Markus comments:
>>   - Use g_assert_not_reached() in stub
>>   - Add commit descriptions
>>   - Extract ACPI / PCI
>>   - MEMORY not extracted as no feedback from Igor/MST
>>
>> Since v2:
>> - Addressed Igor review comments (move query-uuid to machine)
>> - Addressed David review comments (fix balloon documentation)
>>
>> Since v1:
>> - Do not extract the 'add_client' command (elmarco)
>>
>> CI:
>> https://gitlab.com/philmd/qemu/-/pipelines/188367944
>>
>> Philippe Mathieu-Daudé (9):
>>   target/i386: Restrict X86CPUFeatureWord to X86 targets
>>   qapi: Restrict LostTickPolicy enum to machine code
>>   qapi: Correct balloon documentation
>>   qapi: Restrict balloon-related commands to machine code
>>   qapi: Restrict query-vm-generation-id command to machine code
>>   qapi: Restrict query-uuid command to machine code
>>   qapi: Restrict device memory commands to machine code
>>   qapi: Extract ACPI commands to 'acpi.json'
>>   qapi: Extract PCI commands to 'pci.json'
>>
>>  qapi/acpi.json                       | 141 +++++
>>  qapi/machine-target.json             |  45 ++
>>  qapi/machine.json                    | 430 ++++++++++++--
>>  qapi/misc.json                       | 820 ---------------------------
>>  qapi/pci.json                        | 316 +++++++++++
>>  qapi/qapi-schema.json                |   2 +
>>  include/hw/acpi/acpi_dev_interface.h |   2 +-
>>  include/hw/mem/memory-device.h       |   2 +-
>>  include/hw/rtc/mc146818rtc.h         |   2 +-
>>  include/hw/virtio/virtio-pmem.h      |   2 +-
>>  include/sysemu/balloon.h             |   2 +-
>>  block/iscsi.c                        |   2 +-
>>  hw/acpi/core.c                       |   2 +-
>>  hw/acpi/cpu.c                        |   2 +-
>>  hw/acpi/memory_hotplug.c             |   3 +-
>>  hw/acpi/vmgenid.c                    |   2 +-
>>  hw/core/qdev-properties.c            |   1 +
>>  hw/i386/kvm/i8254.c                  |   2 +-
>>  hw/pci/pci-stub.c                    |   2 +-
>>  hw/pci/pci.c                         |   2 +-
>>  hw/virtio/virtio-balloon.c           |   2 +-
>>  hw/virtio/virtio-mem-pci.c           |   1 +
>>  monitor/hmp-cmds.c                   |   2 +
>>  monitor/qmp-cmds.c                   |   1 +
>>  softmmu/balloon.c                    |   2 +-
>>  stubs/uuid.c                         |   2 +-
>>  stubs/vmgenid.c                      |   2 +-
>>  target/i386/cpu.c                    |   2 +-
>>  target/i386/machine-stub.c           |  23 +
>>  MAINTAINERS                          |   2 +
>>  qapi/meson.build                     |   2 +
>>  target/i386/meson.build              |   1 +
>>  32 files changed, 944 insertions(+), 880 deletions(-)
>>  create mode 100644 qapi/acpi.json
>>  create mode 100644 qapi/pci.json
>>  create mode 100644 target/i386/machine-stub.c
>>
> 
> I'm not sure about the first.  It's a bit ugly to have a property whose
> getter aborts.

This is (IIUC) the change Markus requested in v3 (see [*]).

Commit description: "add a stub to satisfy the linker."

This stub should be never used on user-mode => abort.

[*] https://www.mail-archive.com/qemu-block@nongnu.org/msg67560.html

> 
> For patches 2-9 though:
> 
> Acked-by: Paolo Bonzini <pbonzini@redhat.com>

Thanks!

> 
> Paolo
>