Skip to contentSkip to conversation

microvm: add acpi support

PATCH Superseded Gerd Hoffmann 5 participants 61 messages May 20, 2020 – May 28, 2020

v3 of 8 Superseded Latest is v8 →

lore
Share & tools
lore
#0 lore raw

v3 cover microvm: add acpi support

I know that not supporting ACPI in microvm is intentional. If you still don't want ACPI this is perfectly fine, you can use the usual -no-acpi switch to toggle ACPI support.

These are the advantages you are going to loose then:

(1) virtio-mmio device discovery without command line hacks (tweaking the command line is a problem when not using direct kernel boot).
(2) Better IO-APIC support, we can use IRQ lines 16-23.
(3) ACPI power button (aka powerdown request) works.
(4) machine poweroff (aka S5 state) works.

Together with seabios patches for virtio-mmio support this allows to boot standard fedora images (cloud, coreos, workstation live) with the microvm machine type.

git branch for testing (including updated seabios):
https://git.kraxel.org/cgit/qemu/log/?h=sirius/microvm

changes in v2:
* some acpi cleanups are an separate patch series now.
* switched to hw reduced acpi & generic event device.
* misc fixes here and there.

changes in v3:
* depeds on "[PATCH v6 00/16] acpi: i386 tweaks" series.
* renamed qboot to qboot.bin
* updated seabios to master branch snapshot.
- this version boots fine with rtc=off
* generic event device tweaks (Igor's comments).
* make SMP work.
* add RfC patches to turn off acpi by default for microvm.
* misc fixes here and there.

cheers,
Gerd

Gerd Hoffmann (22):
microvm: name qboot binary qboot.bin [testing] seabios: update submodule to master snapshot [testing] seabios: update config & build rules [testing] seabios: update binaries to master snapshot
acpi: make build_madt() more generic.
acpi: create acpi-common.c and move madt code
acpi: madt: skip pci override on pci-less systems.
acpi: fadt: add hw-reduced sleep register support
acpi: ged: rename event memory region
acpi: ged: add control regs
acpi: ged: add x86 device variant.
acpi: move acpi_dsdt_add_power_button() to ged
x86: coldplug cpus
microvm: use 2G split unconditionally
microvm: make virtio irq base runtime configurable microvm/acpi: add minimal acpi support microvm/acpi: add acpi_dsdt_add_virtio() for x86 microvm/acpi: use GSI 16-23 for virtio microvm/acpi: use seabios with acpi=on microvm/acpi: disable virtio-mmio cmdline hack [RfC] acpi: add per machine type acpi default [RfC] acpi: flip default to off for microvm

 hw/i386/acpi-common.h                  |  15 ++
 hw/i386/acpi-microvm.h                 |   6 +
 include/hw/acpi/acpi-defs.h            |   2 +
 include/hw/acpi/generic_event_device.h |  14 +-
 include/hw/i386/microvm.h              |  11 +-
 include/hw/i386/x86.h                  |   1 +
 hw/acpi/aml-build.c                    |   4 +-
 hw/acpi/generic_event_device.c         |  68 ++++++-
 hw/arm/virt-acpi-build.c               |   8 -
 hw/i386/acpi-build.c                   | 126 +------------
 hw/i386/acpi-common.c                  | 156 +++++++++++++++++
 hw/i386/acpi-microvm.c                 | 234 +++++++++++++++++++++++++
 hw/i386/generic_event_device_x86.c     |  36 ++++
 hw/i386/microvm.c                      |  78 ++++++---
 hw/i386/x86.c                          |  32 +++-
 hw/i386/Kconfig                        |   1 +
 hw/i386/Makefile.objs                  |   3 +
 pc-bios/bios-256k.bin                  | Bin 262144 -> 262144 bytes
 pc-bios/bios-microvm.bin               | Bin 65536 -> 131072 bytes
 pc-bios/bios.bin                       | Bin 131072 -> 131072 bytes
 pc-bios/qboot.bin                      | Bin 0 -> 65536 bytes
 roms/Makefile                          |  11 +-
 roms/config.seabios-128k               |   3 +
 roms/config.seabios-microvm            |  26 +++
 roms/seabios                           |   2 +-
 25 files changed, 664 insertions(+), 173 deletions(-)
 create mode 100644 hw/i386/acpi-common.h
 create mode 100644 hw/i386/acpi-microvm.h
 create mode 100644 hw/i386/acpi-common.c
 create mode 100644 hw/i386/acpi-microvm.c
 create mode 100644 hw/i386/generic_event_device_x86.c
 create mode 100644 pc-bios/qboot.bin
 create mode 100644 roms/config.seabios-microvm
Signature · 1 line
2.18.4
#1 lore raw

PATCH v3 1/22 microvm: name qboot binary qboot.bin

qboot isn't a bios and shouldnt be named that way.

---
 hw/i386/microvm.c                       |   4 ++--
 pc-bios/{bios-microvm.bin => qboot.bin} | Bin
 roms/Makefile                           |   6 +++---
 3 files changed, 5 insertions(+), 5 deletions(-)
 rename pc-bios/{bios-microvm.bin => qboot.bin} (100%)
hw/i386/microvm.c +2 -2
index 937db10ae6a5..867d3d652145 100644
@@ -47,7 +47,7 @@
 #include "kvm_i386.h"
 #include "hw/xen/start_info.h"
 
-#define MICROVM_BIOS_FILENAME "bios-microvm.bin"
+#define MICROVM_QBOOT_FILENAME "qboot.bin"
 
 static void microvm_set_rtc(MicrovmMachineState *mms, ISADevice *s)
 {
@@ -158,7 +158,7 @@ static void microvm_devices_init(MicrovmMachineState *mms)
     }
 
     if (bios_name == NULL) {
-        bios_name = MICROVM_BIOS_FILENAME;
+        bios_name = MICROVM_QBOOT_FILENAME;
     }
     x86_bios_rom_init(get_system_memory(), true);
 }
pc-bios/qboot.bin renamed from pc-bios/bios-microvm.bin +0 -0
similarity index 100%
rename from pc-bios/bios-microvm.bin
rename to pc-bios/qboot.bin
roms/Makefile +3 -3
index f9acf39954dc..d6d4948bfc31 100644
@@ -68,7 +68,7 @@ default help:
 	@echo "  opensbi64-virt     -- update OpenSBI for 64-bit virt machine"
 	@echo "  opensbi32-sifive_u -- update OpenSBI for 32-bit sifive_u machine"
 	@echo "  opensbi64-sifive_u -- update OpenSBI for 64-bit sifive_u machine"
-	@echo "  bios-microvm       -- update bios-microvm.bin (qboot)"
+	@echo "  qboot              -- update qboot"
 	@echo "  clean              -- delete the files generated by the previous" \
 	                              "build targets"
 
@@ -194,9 +194,9 @@ opensbi64-sifive_u:
 		PLATFORM="sifive/fu540"
 	cp opensbi/build/platform/sifive/fu540/firmware/fw_jump.bin ../pc-bios/opensbi-riscv64-sifive_u-fw_jump.bin
 
-bios-microvm:
+qboot:
 	$(MAKE) -C qboot
-	cp qboot/bios.bin ../pc-bios/bios-microvm.bin
+	cp qboot/bios.bin ../pc-bios/qboot.bin
 
 clean:
 	rm -rf seabios/.config seabios/out seabios/builds
Signature · 1 line
2.18.4
#2 lore raw

Re: [PATCH v3 01/22] microvm: name qboot binary qboot.bin

Hi Gerd,

Quoted 1 line · Gerd Hoffmann

qboot isn't a bios and shouldnt be named that way.

Still it resides in a directory named pc-bios =)

It is used later in x86_bios_rom_init() as:

    bios = g_malloc(sizeof(*bios));
    memory_region_init_ram(bios, NULL, "pc.bios", bios_size, &error_fatal);
    if (!isapc_ram_fw) {
        memory_region_set_readonly(bios, true);
    }
    ret = rom_add_file_fixed(bios_name, (uint32_t)(-bios_size), -1);
    if (ret != 0) {
    bios_error:
        fprintf(stderr, "qemu: could not load PC BIOS '%s'\n", bios_name);
        exit(1);
    }
    g_free(filename);

Many 'bios' occurrences.

Should it be called 'qboot.rom' instead?

Quoted 61 lines
---
  hw/i386/microvm.c                       |   4 ++--
  pc-bios/{bios-microvm.bin => qboot.bin} | Bin
  roms/Makefile                           |   6 +++---
  3 files changed, 5 insertions(+), 5 deletions(-)

rename pc-bios/{bios-microvm.bin => qboot.bin} (100%)

hw/i386/microvm.c +2 -2
index 937db10ae6a5..867d3d652145 100644
@@ -47,7 +47,7 @@
  #include "kvm_i386.h"
  #include "hw/xen/start_info.h"
 
-#define MICROVM_BIOS_FILENAME "bios-microvm.bin"
+#define MICROVM_QBOOT_FILENAME "qboot.bin"
 
  static void microvm_set_rtc(MicrovmMachineState *mms, ISADevice *s)
  {
@@ -158,7 +158,7 @@ static void microvm_devices_init(MicrovmMachineState *mms)
      }
 
      if (bios_name == NULL) {
-        bios_name = MICROVM_BIOS_FILENAME;
+        bios_name = MICROVM_QBOOT_FILENAME;
      }
      x86_bios_rom_init(get_system_memory(), true);
  }
pc-bios/qboot.bin renamed from pc-bios/bios-microvm.bin +0 -0
similarity index 100%
rename from pc-bios/bios-microvm.bin
rename to pc-bios/qboot.bin
roms/Makefile +3 -3
index f9acf39954dc..d6d4948bfc31 100644
@@ -68,7 +68,7 @@ default help:
  	@echo "  opensbi64-virt     -- update OpenSBI for 64-bit virt machine"
  	@echo "  opensbi32-sifive_u -- update OpenSBI for 32-bit sifive_u machine"
  	@echo "  opensbi64-sifive_u -- update OpenSBI for 64-bit sifive_u machine"
-	@echo "  bios-microvm       -- update bios-microvm.bin (qboot)"
+	@echo "  qboot              -- update qboot"
  	@echo "  clean              -- delete the files generated by the previous" \
  	                              "build targets"
 
@@ -194,9 +194,9 @@ opensbi64-sifive_u:
  		PLATFORM="sifive/fu540"
  	cp opensbi/build/platform/sifive/fu540/firmware/fw_jump.bin ../pc-bios/opensbi-riscv64-sifive_u-fw_jump.bin
 
-bios-microvm:
+qboot:
  	$(MAKE) -C qboot
-	cp qboot/bios.bin ../pc-bios/bios-microvm.bin
+	cp qboot/bios.bin ../pc-bios/qboot.bin
 
  clean:
  	rm -rf seabios/.config seabios/out seabios/builds
#3 lore raw

Re: [PATCH v3 01/22] microvm: name qboot binary qboot.bin

Quoted 6 lines · Philippe Mathieu-Daudé

Hi Gerd,

Quoted 1 line · Gerd Hoffmann

qboot isn't a bios and shouldnt be named that way.

Still it resides in a directory named pc-bios =)

Yep, that should be named "firmware" but changing that is beyond the scope of this patch ;)

Quoted 15 lines

It is used later in x86_bios_rom_init() as:

    bios = g_malloc(sizeof(*bios));
    memory_region_init_ram(bios, NULL, "pc.bios", bios_size, &error_fatal);
    if (!isapc_ram_fw) {
        memory_region_set_readonly(bios, true);
    }
    ret = rom_add_file_fixed(bios_name, (uint32_t)(-bios_size), -1);
    if (ret != 0) {
    bios_error:
        fprintf(stderr, "qemu: could not load PC BIOS '%s'\n", bios_name);
        exit(1);
    }
    g_free(filename);
Quoted 3 lines

Many 'bios' occurrences.

Should it be called 'qboot.rom' instead?

I don't care much, I only want bios*.bin for seabios which actually is a complete bios implementation ...

take care,
Gerd

#4 lore raw

Re: [PATCH v3 01/22] microvm: name qboot binary qboot.bin

Quoted 26 lines · Gerd Hoffmann
Quoted 6 lines · Philippe Mathieu-Daudé

Hi Gerd,

Quoted 1 line · Gerd Hoffmann

qboot isn't a bios and shouldnt be named that way.

Still it resides in a directory named pc-bios =)

Yep, that should be named "firmware" but changing that is beyond the scope of this patch ;)

Quoted 15 lines

It is used later in x86_bios_rom_init() as:

    bios = g_malloc(sizeof(*bios));
    memory_region_init_ram(bios, NULL, "pc.bios", bios_size, &error_fatal);
    if (!isapc_ram_fw) {
        memory_region_set_readonly(bios, true);
    }
    ret = rom_add_file_fixed(bios_name, (uint32_t)(-bios_size), -1);
    if (ret != 0) {
    bios_error:
        fprintf(stderr, "qemu: could not load PC BIOS '%s'\n", bios_name);
        exit(1);
    }
    g_free(filename);
Quoted 6 lines
Quoted 3 lines

Many 'bios' occurrences.

Should it be called 'qboot.rom' instead?

I don't care much, I only want bios*.bin for seabios which actually is a complete bios implementation ...

OK.

Preferably named 'qboot.rom':

#5 lore raw

PATCH v3 2/22 seabios: update submodule to master snapshot

In reply to Gerd Hoffmann
---
 roms/seabios | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
roms/seabios +1 -1
index f21b5a4aeb02..b8eda1319544 160000
@@ -1 +1 @@
-Subproject commit f21b5a4aeb020f2a5e2c6503f906a9349dd2f069
+Subproject commit b8eda131954452bb5a236100a6572fe8f27d8021
Signature · 1 line
2.18.4
#6 lore raw

PATCH v3 3/22 seabios: update config & build rules

In reply to Gerd Hoffmann
---
 roms/Makefile               |  5 ++++-
 roms/config.seabios-128k    |  3 +++
 roms/config.seabios-microvm | 26 ++++++++++++++++++++++++++
 3 files changed, 33 insertions(+), 1 deletion(-)
 create mode 100644 roms/config.seabios-microvm
roms/Makefile +4 -1
index d6d4948bfc31..f7cdad027374 100644
@@ -72,9 +72,12 @@ default help:
 	@echo "  clean              -- delete the files generated by the previous" \
 	                              "build targets"
 
-bios: build-seabios-config-seabios-128k build-seabios-config-seabios-256k
+bios: build-seabios-config-seabios-128k \
+		build-seabios-config-seabios-256k \
+		build-seabios-config-seabios-microvm
 	cp seabios/builds/seabios-128k/bios.bin ../pc-bios/bios.bin
 	cp seabios/builds/seabios-256k/bios.bin ../pc-bios/bios-256k.bin
+	cp seabios/builds/seabios-microvm/bios.bin ../pc-bios/bios-microvm.bin
 
 vgabios seavgabios: $(patsubst %,seavgabios-%,$(vgabios_variants))
 
roms/config.seabios-128k +3 -0
index c43912bf9de4..d18c802c46e9 100644
@@ -11,8 +11,11 @@ CONFIG_USB_UAS=n
 CONFIG_SDCARD=n
 CONFIG_TCGBIOS=n
 CONFIG_MPT_SCSI=n
+CONFIG_ESP_SCSI=n
+CONFIG_MEGASAS=n
 CONFIG_PVSCSI=n
 CONFIG_NVME=n
 CONFIG_USE_SMM=n
 CONFIG_VGAHOOKS=n
 CONFIG_HOST_BIOS_GEOMETRY=n
+CONFIG_ACPI_PARSE=n
roms/config.seabios-microvm new file +26 -0
new file mode 100644
index 000000000000..a253e2edc6ec
@@ -0,0 +1,26 @@
+CONFIG_QEMU=y
+CONFIG_QEMU_HARDWARE=y
+CONFIG_PERMIT_UNALIGNED_PCIROM=y
+CONFIG_ROM_SIZE=128
+CONFIG_XEN=n
+CONFIG_BOOTSPLASH=n
+CONFIG_ATA=n
+CONFIG_AHCI=n
+CONFIG_SDCARD=n
+CONFIG_PVSCSI=n
+CONFIG_ESP_SCSI=n
+CONFIG_LSI_SCSI=n
+CONFIG_MEGASAS=n
+CONFIG_MPT_SCSI=n
+CONFIG_FLOPPY=n
+CONFIG_FLASH_FLOPPY=n
+CONFIG_NVME=n
+CONFIG_PS2PORT=n
+CONFIG_USB=n
+CONFIG_LPT=n
+CONFIG_RTC_TIMER=n
+CONFIG_USE_SMM=n
+CONFIG_PMTIMER=n
+CONFIG_TCGBIOS=n
+CONFIG_HARDWARE_IRQ=n
+CONFIG_ACPI_PARSE=y
Signature · 1 line
2.18.4
#7 lore raw

PATCH v3 4/22 seabios: update binaries to master snapshot

In reply to Gerd Hoffmann
---
 pc-bios/bios-256k.bin    | Bin 262144 -> 262144 bytes
 pc-bios/bios-microvm.bin | Bin 0 -> 131072 bytes
 pc-bios/bios.bin         | Bin 131072 -> 131072 bytes
 3 files changed, 0 insertions(+), 0 deletions(-)
 create mode 100644 pc-bios/bios-microvm.bin
pc-bios/bios-256k.bin binary +0 -0
index 4af7f7d5a91bcc19905e1742b0706e034e2a655c..f3471dafea16d7ce43deade009954afa2e60ab35 100644
GIT binary patch
pc-bios/bios-microvm.bin new file binary +0 -0
new file mode 100644
index 0000000000000000000000000000000000000000..fff24bec590dbe9b882d6d4c0f8fd5b5c3086ce8
GIT binary patch
pc-bios/bios.bin binary +0 -0
index 67874c31b8b0284eb292deb2c9ebc191fa2d73b6..e7ed9c9347333631fb59d125bad3017f5fae4af8 100644
GIT binary patch
Signature · 1 line
2.18.4
#8 lore raw

PATCH v3 5/22 acpi: make build_madt() more generic.

In reply to Gerd Hoffmann

Remove PCMachineState dependency from build_madt(). Pass AcpiDeviceIf as separate argument instead of depending on PCMachineState->acpi_dev.

---
 hw/i386/acpi-build.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
hw/i386/acpi-build.c +7 -7
index 1afb47b09ee9..4f6418d60b05 100644
@@ -366,14 +366,13 @@ void pc_madt_cpu_entry(AcpiDeviceIf *adev, int uid,
 }
 
 static void
-build_madt(GArray *table_data, BIOSLinker *linker, PCMachineState *pcms)
+build_madt(GArray *table_data, BIOSLinker *linker,
+           X86MachineState *x86ms, AcpiDeviceIf *adev)
 {
-    MachineClass *mc = MACHINE_GET_CLASS(pcms);
-    X86MachineState *x86ms = X86_MACHINE(pcms);
-    const CPUArchIdList *apic_ids = mc->possible_cpu_arch_ids(MACHINE(pcms));
+    MachineClass *mc = MACHINE_GET_CLASS(x86ms);
+    const CPUArchIdList *apic_ids = mc->possible_cpu_arch_ids(MACHINE(x86ms));
     int madt_start = table_data->len;
-    AcpiDeviceIfClass *adevc = ACPI_DEVICE_IF_GET_CLASS(pcms->acpi_dev);
-    AcpiDeviceIf *adev = ACPI_DEVICE_IF(pcms->acpi_dev);
+    AcpiDeviceIfClass *adevc = ACPI_DEVICE_IF_GET_CLASS(adev);
     bool x2apic_mode = false;
 
     AcpiMultipleApicTable *madt;
@@ -2534,7 +2533,8 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
     aml_len += tables_blob->len - fadt;
 
     acpi_add_table(table_offsets, tables_blob);
-    build_madt(tables_blob, tables->linker, pcms);
+    build_madt(tables_blob, tables->linker, x86ms,
+               ACPI_DEVICE_IF(pcms->acpi_dev));
 
     vmgenid_dev = find_vmgenid_dev();
     if (vmgenid_dev) {
Signature · 1 line
2.18.4
#9 lore raw

PATCH v3 6/22 acpi: create acpi-common.c and move madt code

In reply to Gerd Hoffmann

We'll need madt support for microvm.

---
 hw/i386/acpi-common.h |  14 ++++
 hw/i386/acpi-build.c  | 126 +---------------------------------
 hw/i386/acpi-common.c | 152 ++++++++++++++++++++++++++++++++++++++++++
 hw/i386/Makefile.objs |   1 +
 4 files changed, 170 insertions(+), 123 deletions(-)
 create mode 100644 hw/i386/acpi-common.h
 create mode 100644 hw/i386/acpi-common.c
hw/i386/acpi-common.h new file +14 -0
new file mode 100644
index 000000000000..c30e461f1854
@@ -0,0 +1,14 @@
+#ifndef HW_I386_ACPI_COMMON_H
+#define HW_I386_ACPI_COMMON_H
+#include "include/hw/acpi/acpi_dev_interface.h"
+
+#include "include/hw/acpi/bios-linker-loader.h"
+#include "include/hw/i386/x86.h"
+
+/* Default IOAPIC ID */
+#define ACPI_BUILD_IOAPIC_ID 0x0
+
+void acpi_build_madt(GArray *table_data, BIOSLinker *linker,
+                     X86MachineState *x86ms, AcpiDeviceIf *adev);
+
+#endif
hw/i386/acpi-build.c +3 -123

150 lines not on this page · raw message

hw/i386/acpi-common.c new file +152 -0

152 lines not on this page · raw message

hw/i386/Makefile.objs +1 -0
index 8ce1b265335b..6abc74551a72 100644
@@ -16,4 +16,5 @@ obj-$(CONFIG_VMMOUSE) += vmmouse.o
 obj-$(CONFIG_PC) += port92.o
 
 obj-y += kvmvapic.o
+obj-$(CONFIG_ACPI) += acpi-common.o
 obj-$(CONFIG_PC) += acpi-build.o
Signature · 1 line
2.18.4
Reviewed-by on v3 6/22
#10 lore raw

Re: [PATCH v3 06/22] acpi: create acpi-common.c and move madt code

Quoted 355 lines · Gerd Hoffmann

We'll need madt support for microvm.

---
  hw/i386/acpi-common.h |  14 ++++
  hw/i386/acpi-build.c  | 126 +---------------------------------
  hw/i386/acpi-common.c | 152 ++++++++++++++++++++++++++++++++++++++++++
  hw/i386/Makefile.objs |   1 +
  4 files changed, 170 insertions(+), 123 deletions(-)

create mode 100644 hw/i386/acpi-common.h
create mode 100644 hw/i386/acpi-common.c

hw/i386/acpi-common.h new file +14 -0
new file mode 100644
index 000000000000..c30e461f1854
@@ -0,0 +1,14 @@
+#ifndef HW_I386_ACPI_COMMON_H
+#define HW_I386_ACPI_COMMON_H
+#include "include/hw/acpi/acpi_dev_interface.h"
+
+#include "include/hw/acpi/bios-linker-loader.h"
+#include "include/hw/i386/x86.h"
+
+/* Default IOAPIC ID */
+#define ACPI_BUILD_IOAPIC_ID 0x0
+
+void acpi_build_madt(GArray *table_data, BIOSLinker *linker,
+                     X86MachineState *x86ms, AcpiDeviceIf *adev);
+
+#endif
hw/i386/acpi-build.c +3 -123
index 4f6418d60b05..c0fdd539a640 100644
@@ -24,6 +24,7 @@
  #include "qapi/error.h"
  #include "qapi/qmp/qnum.h"
  #include "acpi-build.h"
+#include "acpi-common.h"
  #include "qemu/bitmap.h"
  #include "qemu/error-report.h"
  #include "hw/pci/pci.h"
@@ -89,9 +90,6 @@
  #define ACPI_BUILD_DPRINTF(fmt, ...)
  #endif
 
-/* Default IOAPIC ID */
-#define ACPI_BUILD_IOAPIC_ID 0x0
-
  typedef struct AcpiPmInfo {
      bool s3_disabled;
      bool s4_disabled;
@@ -327,124 +325,6 @@ build_facs(GArray *table_data)
      facs->length = cpu_to_le32(sizeof(*facs));
  }
 
-void pc_madt_cpu_entry(AcpiDeviceIf *adev, int uid,
-                       const CPUArchIdList *apic_ids, GArray *entry)
-{
-    uint32_t apic_id = apic_ids->cpus[uid].arch_id;
-
-    /* ACPI spec says that LAPIC entry for non present
-     * CPU may be omitted from MADT or it must be marked
-     * as disabled. However omitting non present CPU from
-     * MADT breaks hotplug on linux. So possible CPUs
-     * should be put in MADT but kept disabled.
-     */
-    if (apic_id < 255) {
-        AcpiMadtProcessorApic *apic = acpi_data_push(entry, sizeof *apic);
-
-        apic->type = ACPI_APIC_PROCESSOR;
-        apic->length = sizeof(*apic);
-        apic->processor_id = uid;
-        apic->local_apic_id = apic_id;
-        if (apic_ids->cpus[uid].cpu != NULL) {
-            apic->flags = cpu_to_le32(1);
-        } else {
-            apic->flags = cpu_to_le32(0);
-        }
-    } else {
-        AcpiMadtProcessorX2Apic *apic = acpi_data_push(entry, sizeof *apic);
-
-        apic->type = ACPI_APIC_LOCAL_X2APIC;
-        apic->length = sizeof(*apic);
-        apic->uid = cpu_to_le32(uid);
-        apic->x2apic_id = cpu_to_le32(apic_id);
-        if (apic_ids->cpus[uid].cpu != NULL) {
-            apic->flags = cpu_to_le32(1);
-        } else {
-            apic->flags = cpu_to_le32(0);
-        }
-    }
-}
-
-static void
-build_madt(GArray *table_data, BIOSLinker *linker,
-           X86MachineState *x86ms, AcpiDeviceIf *adev)
-{
-    MachineClass *mc = MACHINE_GET_CLASS(x86ms);
-    const CPUArchIdList *apic_ids = mc->possible_cpu_arch_ids(MACHINE(x86ms));
-    int madt_start = table_data->len;
-    AcpiDeviceIfClass *adevc = ACPI_DEVICE_IF_GET_CLASS(adev);
-    bool x2apic_mode = false;
-
-    AcpiMultipleApicTable *madt;
-    AcpiMadtIoApic *io_apic;
-    AcpiMadtIntsrcovr *intsrcovr;
-    int i;
-
-    madt = acpi_data_push(table_data, sizeof *madt);
-    madt->local_apic_address = cpu_to_le32(APIC_DEFAULT_ADDRESS);
-    madt->flags = cpu_to_le32(1);
-
-    for (i = 0; i < apic_ids->len; i++) {
-        adevc->madt_cpu(adev, i, apic_ids, table_data);
-        if (apic_ids->cpus[i].arch_id > 254) {
-            x2apic_mode = true;
-        }
-    }
-
-    io_apic = acpi_data_push(table_data, sizeof *io_apic);
-    io_apic->type = ACPI_APIC_IO;
-    io_apic->length = sizeof(*io_apic);
-    io_apic->io_apic_id = ACPI_BUILD_IOAPIC_ID;
-    io_apic->address = cpu_to_le32(IO_APIC_DEFAULT_ADDRESS);
-    io_apic->interrupt = cpu_to_le32(0);
-
-    if (x86ms->apic_xrupt_override) {
-        intsrcovr = acpi_data_push(table_data, sizeof *intsrcovr);
-        intsrcovr->type   = ACPI_APIC_XRUPT_OVERRIDE;
-        intsrcovr->length = sizeof(*intsrcovr);
-        intsrcovr->source = 0;
-        intsrcovr->gsi    = cpu_to_le32(2);
-        intsrcovr->flags  = cpu_to_le16(0); /* conforms to bus specifications */
-    }
-    for (i = 1; i < 16; i++) {
-#define ACPI_BUILD_PCI_IRQS ((1<<5) | (1<<9) | (1<<10) | (1<<11))
-        if (!(ACPI_BUILD_PCI_IRQS & (1 << i))) {
-            /* No need for a INT source override structure. */
-            continue;
-        }
-        intsrcovr = acpi_data_push(table_data, sizeof *intsrcovr);
-        intsrcovr->type   = ACPI_APIC_XRUPT_OVERRIDE;
-        intsrcovr->length = sizeof(*intsrcovr);
-        intsrcovr->source = i;
-        intsrcovr->gsi    = cpu_to_le32(i);
-        intsrcovr->flags  = cpu_to_le16(0xd); /* active high, level triggered */
-    }
-
-    if (x2apic_mode) {
-        AcpiMadtLocalX2ApicNmi *local_nmi;
-
-        local_nmi = acpi_data_push(table_data, sizeof *local_nmi);
-        local_nmi->type   = ACPI_APIC_LOCAL_X2APIC_NMI;
-        local_nmi->length = sizeof(*local_nmi);
-        local_nmi->uid    = 0xFFFFFFFF; /* all processors */
-        local_nmi->flags  = cpu_to_le16(0);
-        local_nmi->lint   = 1; /* ACPI_LINT1 */
-    } else {
-        AcpiMadtLocalNmi *local_nmi;
-
-        local_nmi = acpi_data_push(table_data, sizeof *local_nmi);
-        local_nmi->type         = ACPI_APIC_LOCAL_NMI;
-        local_nmi->length       = sizeof(*local_nmi);
-        local_nmi->processor_id = 0xff; /* all processors */
-        local_nmi->flags        = cpu_to_le16(0);
-        local_nmi->lint         = 1; /* ACPI_LINT1 */
-    }
-
-    build_header(linker, table_data,
-                 (void *)(table_data->data + madt_start), "APIC",
-                 table_data->len - madt_start, 1, NULL, NULL);
-}
-
  static void build_append_pcihp_notify_entry(Aml *method, int slot)
  {
      Aml *if_ctx;
@@ -2533,8 +2413,8 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
      aml_len += tables_blob->len - fadt;
 
      acpi_add_table(table_offsets, tables_blob);
-    build_madt(tables_blob, tables->linker, x86ms,
-               ACPI_DEVICE_IF(pcms->acpi_dev));
+    acpi_build_madt(tables_blob, tables->linker, x86ms,
+                    ACPI_DEVICE_IF(pcms->acpi_dev));
 
      vmgenid_dev = find_vmgenid_dev();
      if (vmgenid_dev) {
hw/i386/acpi-common.c new file +152 -0
new file mode 100644
index 000000000000..5caca16a0b59
@@ -0,0 +1,152 @@
+/* Support for generating ACPI tables and passing them to Guests
+ *
+ * Copyright (C) 2008-2010  Kevin O'Connor <[email protected]>
+ * Copyright (C) 2006 Fabrice Bellard
+ * Copyright (C) 2013 Red Hat Inc
+ *
+ * Author: Michael S. Tsirkin <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "qemu/osdep.h"
+#include "qapi/error.h"
+
+#include "exec/memory.h"
+#include "hw/acpi/acpi.h"
+#include "hw/acpi/aml-build.h"
+#include "hw/acpi/utils.h"
+#include "hw/i386/pc.h"
+#include "target/i386/cpu.h"
+
+#include "acpi-build.h"
+#include "acpi-common.h"
+
+void pc_madt_cpu_entry(AcpiDeviceIf *adev, int uid,
+                       const CPUArchIdList *apic_ids, GArray *entry)
+{
+    uint32_t apic_id = apic_ids->cpus[uid].arch_id;
+
+    /* ACPI spec says that LAPIC entry for non present
+     * CPU may be omitted from MADT or it must be marked
+     * as disabled. However omitting non present CPU from
+     * MADT breaks hotplug on linux. So possible CPUs
+     * should be put in MADT but kept disabled.
+     */
+    if (apic_id < 255) {
+        AcpiMadtProcessorApic *apic = acpi_data_push(entry, sizeof *apic);
+
+        apic->type = ACPI_APIC_PROCESSOR;
+        apic->length = sizeof(*apic);
+        apic->processor_id = uid;
+        apic->local_apic_id = apic_id;
+        if (apic_ids->cpus[uid].cpu != NULL) {
+            apic->flags = cpu_to_le32(1);
+        } else {
+            apic->flags = cpu_to_le32(0);
+        }
+    } else {
+        AcpiMadtProcessorX2Apic *apic = acpi_data_push(entry, sizeof *apic);
+
+        apic->type = ACPI_APIC_LOCAL_X2APIC;
+        apic->length = sizeof(*apic);
+        apic->uid = cpu_to_le32(uid);
+        apic->x2apic_id = cpu_to_le32(apic_id);
+        if (apic_ids->cpus[uid].cpu != NULL) {
+            apic->flags = cpu_to_le32(1);
+        } else {
+            apic->flags = cpu_to_le32(0);
+        }
+    }
+}
+
+void acpi_build_madt(GArray *table_data, BIOSLinker *linker,
+                     X86MachineState *x86ms, AcpiDeviceIf *adev)
+{
+    MachineClass *mc = MACHINE_GET_CLASS(x86ms);
+    const CPUArchIdList *apic_ids = mc->possible_cpu_arch_ids(MACHINE(x86ms));
+    int madt_start = table_data->len;
+    AcpiDeviceIfClass *adevc = ACPI_DEVICE_IF_GET_CLASS(adev);
+    bool x2apic_mode = false;
+
+    AcpiMultipleApicTable *madt;
+    AcpiMadtIoApic *io_apic;
+    AcpiMadtIntsrcovr *intsrcovr;
+    int i;
+
+    madt = acpi_data_push(table_data, sizeof *madt);
+    madt->local_apic_address = cpu_to_le32(APIC_DEFAULT_ADDRESS);
+    madt->flags = cpu_to_le32(1);
+
+    for (i = 0; i < apic_ids->len; i++) {
+        adevc->madt_cpu(adev, i, apic_ids, table_data);
+        if (apic_ids->cpus[i].arch_id > 254) {
+            x2apic_mode = true;
+        }
+    }
+
+    io_apic = acpi_data_push(table_data, sizeof *io_apic);
+    io_apic->type = ACPI_APIC_IO;
+    io_apic->length = sizeof(*io_apic);
+    io_apic->io_apic_id = ACPI_BUILD_IOAPIC_ID;
+    io_apic->address = cpu_to_le32(IO_APIC_DEFAULT_ADDRESS);
+    io_apic->interrupt = cpu_to_le32(0);
+
+    if (x86ms->apic_xrupt_override) {
+        intsrcovr = acpi_data_push(table_data, sizeof *intsrcovr);
+        intsrcovr->type   = ACPI_APIC_XRUPT_OVERRIDE;
+        intsrcovr->length = sizeof(*intsrcovr);
+        intsrcovr->source = 0;
+        intsrcovr->gsi    = cpu_to_le32(2);
+        intsrcovr->flags  = cpu_to_le16(0); /* conforms to bus specifications */
+    }
+    for (i = 1; i < 16; i++) {
+#define ACPI_BUILD_PCI_IRQS ((1<<5) | (1<<9) | (1<<10) | (1<<11))
+        if (!(ACPI_BUILD_PCI_IRQS & (1 << i))) {
+            /* No need for a INT source override structure. */
+            continue;
+        }
+        intsrcovr = acpi_data_push(table_data, sizeof *intsrcovr);
+        intsrcovr->type   = ACPI_APIC_XRUPT_OVERRIDE;
+        intsrcovr->length = sizeof(*intsrcovr);
+        intsrcovr->source = i;
+        intsrcovr->gsi    = cpu_to_le32(i);
+        intsrcovr->flags  = cpu_to_le16(0xd); /* active high, level triggered */
+    }
+
+    if (x2apic_mode) {
+        AcpiMadtLocalX2ApicNmi *local_nmi;
+
+        local_nmi = acpi_data_push(table_data, sizeof *local_nmi);
+        local_nmi->type   = ACPI_APIC_LOCAL_X2APIC_NMI;
+        local_nmi->length = sizeof(*local_nmi);
+        local_nmi->uid    = 0xFFFFFFFF; /* all processors */
+        local_nmi->flags  = cpu_to_le16(0);
+        local_nmi->lint   = 1; /* ACPI_LINT1 */
+    } else {
+        AcpiMadtLocalNmi *local_nmi;
+
+        local_nmi = acpi_data_push(table_data, sizeof *local_nmi);
+        local_nmi->type         = ACPI_APIC_LOCAL_NMI;
+        local_nmi->length       = sizeof(*local_nmi);
+        local_nmi->processor_id = 0xff; /* all processors */
+        local_nmi->flags        = cpu_to_le16(0);
+        local_nmi->lint         = 1; /* ACPI_LINT1 */
+    }
+
+    build_header(linker, table_data,
+                 (void *)(table_data->data + madt_start), "APIC",
+                 table_data->len - madt_start, 1, NULL, NULL);
+}
+
hw/i386/Makefile.objs +0 -0
index 8ce1b265335b..6abc74551a72 100644

hw/i386/Makefile.objs obj-$(CONFIG_VMMOUSE) += vmmouse.o · line 20

Quoted 5 lines
hw/i386/Makefile.objs +1 -0
@@ -16,4 +16,5 @@ obj-$(CONFIG_VMMOUSE) += vmmouse.o
  obj-$(CONFIG_PC) += port92.o
 
  obj-y += kvmvapic.o
+obj-$(CONFIG_ACPI) += acpi-common.o
  obj-$(CONFIG_PC) += acpi-build.o
#11 lore raw

PATCH v3 7/22 acpi: madt: skip pci override on pci-less systems.

In reply to Gerd Hoffmann

Needed for microvm.

---
 hw/i386/acpi-common.h |  3 ++-
 hw/i386/acpi-build.c  |  2 +-
 hw/i386/acpi-common.c | 26 +++++++++++++++-----------
 3 files changed, 18 insertions(+), 13 deletions(-)
hw/i386/acpi-common.h +2 -1
index c30e461f1854..9cac18dddf5b 100644
@@ -9,6 +9,7 @@
 #define ACPI_BUILD_IOAPIC_ID 0x0
 
 void acpi_build_madt(GArray *table_data, BIOSLinker *linker,
-                     X86MachineState *x86ms, AcpiDeviceIf *adev);
+                     X86MachineState *x86ms, AcpiDeviceIf *adev,
+                     bool has_pci);
 
 #endif
hw/i386/acpi-build.c +1 -1
index c0fdd539a640..583bb58d6f0c 100644
@@ -2414,7 +2414,7 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
 
     acpi_add_table(table_offsets, tables_blob);
     acpi_build_madt(tables_blob, tables->linker, x86ms,
-                    ACPI_DEVICE_IF(pcms->acpi_dev));
+                    ACPI_DEVICE_IF(pcms->acpi_dev), true);
 
     vmgenid_dev = find_vmgenid_dev();
     if (vmgenid_dev) {
hw/i386/acpi-common.c +15 -11
index 5caca16a0b59..ab9b00581a15 100644
@@ -72,7 +72,8 @@ void pc_madt_cpu_entry(AcpiDeviceIf *adev, int uid,
 }
 
 void acpi_build_madt(GArray *table_data, BIOSLinker *linker,
-                     X86MachineState *x86ms, AcpiDeviceIf *adev)
+                     X86MachineState *x86ms, AcpiDeviceIf *adev,
+                     bool has_pci)
 {
     MachineClass *mc = MACHINE_GET_CLASS(x86ms);
     const CPUArchIdList *apic_ids = mc->possible_cpu_arch_ids(MACHINE(x86ms));
@@ -111,18 +112,21 @@ void acpi_build_madt(GArray *table_data, BIOSLinker *linker,
         intsrcovr->gsi    = cpu_to_le32(2);
         intsrcovr->flags  = cpu_to_le16(0); /* conforms to bus specifications */
     }
-    for (i = 1; i < 16; i++) {
+
+    if (has_pci) {
+        for (i = 1; i < 16; i++) {
 #define ACPI_BUILD_PCI_IRQS ((1<<5) | (1<<9) | (1<<10) | (1<<11))
-        if (!(ACPI_BUILD_PCI_IRQS & (1 << i))) {
-            /* No need for a INT source override structure. */
-            continue;
+            if (!(ACPI_BUILD_PCI_IRQS & (1 << i))) {
+                /* No need for a INT source override structure. */
+                continue;
+            }
+            intsrcovr = acpi_data_push(table_data, sizeof *intsrcovr);
+            intsrcovr->type   = ACPI_APIC_XRUPT_OVERRIDE;
+            intsrcovr->length = sizeof(*intsrcovr);
+            intsrcovr->source = i;
+            intsrcovr->gsi    = cpu_to_le32(i);
+            intsrcovr->flags  = cpu_to_le16(0xd); /* active high, level triggered */
         }
-        intsrcovr = acpi_data_push(table_data, sizeof *intsrcovr);
-        intsrcovr->type   = ACPI_APIC_XRUPT_OVERRIDE;
-        intsrcovr->length = sizeof(*intsrcovr);
-        intsrcovr->source = i;
-        intsrcovr->gsi    = cpu_to_le32(i);
-        intsrcovr->flags  = cpu_to_le16(0xd); /* active high, level triggered */
     }
 
     if (x2apic_mode) {
Signature · 1 line
2.18.4
#12 lore raw

PATCH v3 8/22 acpi: fadt: add hw-reduced sleep register support

In reply to Gerd Hoffmann

Add fields to struct AcpiFadtData and update build_fadt() to properly generate sleep register entries.

---
 include/hw/acpi/acpi-defs.h | 2 ++
 hw/acpi/aml-build.c         | 4 ++--
 2 files changed, 4 insertions(+), 2 deletions(-)
include/hw/acpi/acpi-defs.h +2 -0
index c13327fa7867..3be9ab504968 100644
@@ -88,6 +88,8 @@ typedef struct AcpiFadtData {
     struct AcpiGenericAddress pm_tmr;    /* PM_TMR_BLK */
     struct AcpiGenericAddress gpe0_blk;  /* GPE0_BLK */
     struct AcpiGenericAddress reset_reg; /* RESET_REG */
+    struct AcpiGenericAddress sleep_ctl; /* SLEEP_CONTROL_REG */
+    struct AcpiGenericAddress sleep_sts; /* SLEEP_STATUS_REG */
     uint8_t reset_val;         /* RESET_VALUE */
     uint8_t  rev;              /* Revision */
     uint32_t flags;            /* Flags */
hw/acpi/aml-build.c +2 -2
index 3681ec6e3d88..a894e2581fa2 100644
@@ -1865,9 +1865,9 @@ void build_fadt(GArray *tbl, BIOSLinker *linker, const AcpiFadtData *f,
     }
 
     /* SLEEP_CONTROL_REG */
-    build_append_gas(tbl, AML_AS_SYSTEM_MEMORY, 0 , 0, 0, 0);
+    build_append_gas_from_struct(tbl, &f->sleep_ctl);
     /* SLEEP_STATUS_REG */
-    build_append_gas(tbl, AML_AS_SYSTEM_MEMORY, 0 , 0, 0, 0);
+    build_append_gas_from_struct(tbl, &f->sleep_sts);
 
     /* TODO: extra fields need to be added to support revisions above rev5 */
     assert(f->rev == 5);
Signature · 1 line
2.18.4
#13 lore raw

PATCH v3 9/22 acpi: ged: rename event memory region

In reply to Gerd Hoffmann

Rename memory region and callbacks and ops to carry "evt" in the name because a second region will be added shortly.

---
 include/hw/acpi/generic_event_device.h |  2 +-
 hw/acpi/generic_event_device.c         | 16 ++++++++--------
 2 files changed, 9 insertions(+), 9 deletions(-)
include/hw/acpi/generic_event_device.h +1 -1
index 83917de02425..90a9180db572 100644
@@ -86,7 +86,7 @@
 #define ACPI_GED_NVDIMM_HOTPLUG_EVT 0x4
 
 typedef struct GEDState {
-    MemoryRegion io;
+    MemoryRegion evt;
     uint32_t     sel;
 } GEDState;
 
hw/acpi/generic_event_device.c +8 -8
index b1cbdd86b66c..1cb34111e59a 100644
@@ -142,7 +142,7 @@ void build_ged_aml(Aml *table, const char *name, HotplugHandler *hotplug_dev,
 }
 
 /* Memory read by the GED _EVT AML dynamic method */
-static uint64_t ged_read(void *opaque, hwaddr addr, unsigned size)
+static uint64_t ged_evt_read(void *opaque, hwaddr addr, unsigned size)
 {
     uint64_t val = 0;
     GEDState *ged_st = opaque;
@@ -161,14 +161,14 @@ static uint64_t ged_read(void *opaque, hwaddr addr, unsigned size)
 }
 
 /* Nothing is expected to be written to the GED memory region */
-static void ged_write(void *opaque, hwaddr addr, uint64_t data,
-                      unsigned int size)
+static void ged_evt_write(void *opaque, hwaddr addr, uint64_t data,
+                          unsigned int size)
 {
 }
 
-static const MemoryRegionOps ged_ops = {
-    .read = ged_read,
-    .write = ged_write,
+static const MemoryRegionOps ged_evt_ops = {
+    .read = ged_evt_read,
+    .write = ged_evt_write,
     .endianness = DEVICE_LITTLE_ENDIAN,
     .valid = {
         .min_access_size = 4,
@@ -287,9 +287,9 @@ static void acpi_ged_initfn(Object *obj)
     SysBusDevice *sbd = SYS_BUS_DEVICE(obj);
     GEDState *ged_st = &s->ged_state;
 
-    memory_region_init_io(&ged_st->io, obj, &ged_ops, ged_st,
+    memory_region_init_io(&ged_st->evt, obj, &ged_evt_ops, ged_st,
                           TYPE_ACPI_GED, ACPI_GED_EVT_SEL_LEN);
-    sysbus_init_mmio(sbd, &ged_st->io);
+    sysbus_init_mmio(sbd, &ged_st->evt);
 
     sysbus_init_irq(sbd, &s->irq);
 
Signature · 1 line
2.18.4
Reviewed-by on v3 9/22
#14 lore raw

Re: [PATCH v3 09/22] acpi: ged: rename event memory region

Quoted 58 lines · Gerd Hoffmann

Rename memory region and callbacks and ops to carry "evt" in the name because a second region will be added shortly.

---
  include/hw/acpi/generic_event_device.h |  2 +-
  hw/acpi/generic_event_device.c         | 16 ++++++++--------
  2 files changed, 9 insertions(+), 9 deletions(-)
include/hw/acpi/generic_event_device.h +1 -1
index 83917de02425..90a9180db572 100644
@@ -86,7 +86,7 @@
  #define ACPI_GED_NVDIMM_HOTPLUG_EVT 0x4
 
  typedef struct GEDState {
-    MemoryRegion io;
+    MemoryRegion evt;
      uint32_t     sel;
  } GEDState;
 
hw/acpi/generic_event_device.c +6 -6
index b1cbdd86b66c..1cb34111e59a 100644
@@ -142,7 +142,7 @@ void build_ged_aml(Aml *table, const char *name, HotplugHandler *hotplug_dev,
  }
 
  /* Memory read by the GED _EVT AML dynamic method */
-static uint64_t ged_read(void *opaque, hwaddr addr, unsigned size)
+static uint64_t ged_evt_read(void *opaque, hwaddr addr, unsigned size)
  {
      uint64_t val = 0;
      GEDState *ged_st = opaque;
@@ -161,14 +161,14 @@ static uint64_t ged_read(void *opaque, hwaddr addr, unsigned size)
  }
 
  /* Nothing is expected to be written to the GED memory region */
-static void ged_write(void *opaque, hwaddr addr, uint64_t data,
-                      unsigned int size)
+static void ged_evt_write(void *opaque, hwaddr addr, uint64_t data,
+                          unsigned int size)
  {
  }
 
-static const MemoryRegionOps ged_ops = {
-    .read = ged_read,
-    .write = ged_write,
+static const MemoryRegionOps ged_evt_ops = {
+    .read = ged_evt_read,
+    .write = ged_evt_write,
      .endianness = DEVICE_LITTLE_ENDIAN,
      .valid = {
          .min_access_size = 4,
@@ -287,9 +287,9 @@ static void acpi_ged_initfn(Object *obj)
      SysBusDevice *sbd = SYS_BUS_DEVICE(obj);

hw/acpi/generic_event_device.c static void acpi_ged_initfn(Object *obj) · line 295

Quoted 10 lines
hw/acpi/generic_event_device.c +2 -2
      GEDState *ged_st = &s->ged_state;
 
-    memory_region_init_io(&ged_st->io, obj, &ged_ops, ged_st,
+    memory_region_init_io(&ged_st->evt, obj, &ged_evt_ops, ged_st,
                            TYPE_ACPI_GED, ACPI_GED_EVT_SEL_LEN);
-    sysbus_init_mmio(sbd, &ged_st->io);
+    sysbus_init_mmio(sbd, &ged_st->evt);
 
      sysbus_init_irq(sbd, &s->irq);
 
Reviewed-by on v3 9/22
#15 lore raw

Re: [PATCH v3 09/22] acpi: ged: rename event memory region

In reply to Gerd Hoffmann
Quoted 4 lines · On Wed, 20 May 2020 15:19:50 +0200 Gerd Hoffmann <[email protected]> wrote:

Rename memory region and callbacks and ops to carry "evt" in the name because a second region will be added shortly.

Quoted 63 lines
---
 include/hw/acpi/generic_event_device.h |  2 +-
 hw/acpi/generic_event_device.c         | 16 ++++++++--------
 2 files changed, 9 insertions(+), 9 deletions(-)
(quoted diff) +0 -0

b/include/hw/acpi/generic_event_device.h index 83917de02425..90a9180db572 100644 --- a/include/hw/acpi/generic_event_device.h +++ b/include/hw/acpi/generic_event_device.h @@ -86,7 +86,7 @@
#define ACPI_GED_NVDIMM_HOTPLUG_EVT 0x4

typedef struct GEDState {

include/hw/acpi/generic_event_device.h +1 -1
-    MemoryRegion io;
+    MemoryRegion evt;
     uint32_t     sel;
 } GEDState;
(quoted diff) +0 -0

b/hw/acpi/generic_event_device.c index b1cbdd86b66c..1cb34111e59a 100644 --- a/hw/acpi/generic_event_device.c
+++ b/hw/acpi/generic_event_device.c

hw/acpi/generic_event_device.c +8 -8
@@ -142,7 +142,7 @@ void build_ged_aml(Aml *table, const char *name,
 HotplugHandler *hotplug_dev, }
 
 /* Memory read by the GED _EVT AML dynamic method */
-static uint64_t ged_read(void *opaque, hwaddr addr, unsigned size)
+static uint64_t ged_evt_read(void *opaque, hwaddr addr, unsigned
 size) {
     uint64_t val = 0;
     GEDState *ged_st = opaque;
@@ -161,14 +161,14 @@ static uint64_t ged_read(void *opaque, hwaddr
 addr, unsigned size) }
 
 /* Nothing is expected to be written to the GED memory region */
-static void ged_write(void *opaque, hwaddr addr, uint64_t data,
-                      unsigned int size)
+static void ged_evt_write(void *opaque, hwaddr addr, uint64_t data,
+                          unsigned int size)
 {
 }
 
-static const MemoryRegionOps ged_ops = {
-    .read = ged_read,
-    .write = ged_write,
+static const MemoryRegionOps ged_evt_ops = {
+    .read = ged_evt_read,
+    .write = ged_evt_write,
     .endianness = DEVICE_LITTLE_ENDIAN,
     .valid = {
         .min_access_size = 4,
@@ -287,9 +287,9 @@ static void acpi_ged_initfn(Object *obj)
     SysBusDevice *sbd = SYS_BUS_DEVICE(obj);
     GEDState *ged_st = &s->ged_state;
 
-    memory_region_init_io(&ged_st->io, obj, &ged_ops, ged_st,
+    memory_region_init_io(&ged_st->evt, obj, &ged_evt_ops, ged_st,
                           TYPE_ACPI_GED, ACPI_GED_EVT_SEL_LEN);
-    sysbus_init_mmio(sbd, &ged_st->io);
+    sysbus_init_mmio(sbd, &ged_st->evt);
 
     sysbus_init_irq(sbd, &s->irq);
 
#16 lore raw

PATCH v3 10/22 acpi: ged: add control regs

In reply to Gerd Hoffmann

Add control regs (sleep, reset) for hw-reduced acpi.

---
 include/hw/acpi/generic_event_device.h |  7 ++++
 hw/acpi/generic_event_device.c         | 44 ++++++++++++++++++++++++++
 2 files changed, 51 insertions(+)
include/hw/acpi/generic_event_device.h +7 -0
index 90a9180db572..474c92198080 100644
@@ -72,6 +72,12 @@
 #define ACPI_GED_EVT_SEL_OFFSET    0x0
 #define ACPI_GED_EVT_SEL_LEN       0x4
 
+#define ACPI_GED_REG_SLEEP_CTL     0x00
+#define ACPI_GED_REG_SLEEP_STS     0x01
+#define ACPI_GED_REG_RESET         0x02
+#define   ACPI_GED_RESET_VALUE     0x42
+#define ACPI_GED_REG_COUNT         0x03
+
 #define GED_DEVICE      "GED"
 #define AML_GED_EVT_REG "EREG"
 #define AML_GED_EVT_SEL "ESEL"
@@ -87,6 +93,7 @@
 
 typedef struct GEDState {
     MemoryRegion evt;
+    MemoryRegion regs;
     uint32_t     sel;
 } GEDState;
 
hw/acpi/generic_event_device.c +44 -0

62 lines not on this page · raw message

Signature · 1 line
2.18.4
#17 lore raw

Re: [PATCH v3 10/22] acpi: ged: add control regs

Quoted 3 lines · On Wed, 20 May 2020 15:19:51 +0200 Gerd Hoffmann <[email protected]> wrote:

Add control regs (sleep, reset) for hw-reduced acpi.

this should be acompanied by docs update
docs/specs/acpi_hw_reduced_hotplug.rst
to document new registers and their semantics and probaly rename file to something more appropriate

Quoted 99 lines
---
 include/hw/acpi/generic_event_device.h |  7 ++++
 hw/acpi/generic_event_device.c         | 44 ++++++++++++++++++++++++++
 2 files changed, 51 insertions(+)
include/hw/acpi/generic_event_device.h +7 -0
index 90a9180db572..474c92198080 100644
@@ -72,6 +72,12 @@
 #define ACPI_GED_EVT_SEL_OFFSET    0x0
 #define ACPI_GED_EVT_SEL_LEN       0x4
 
+#define ACPI_GED_REG_SLEEP_CTL     0x00
+#define ACPI_GED_REG_SLEEP_STS     0x01
+#define ACPI_GED_REG_RESET         0x02
+#define   ACPI_GED_RESET_VALUE     0x42
+#define ACPI_GED_REG_COUNT         0x03
+
 #define GED_DEVICE      "GED"
 #define AML_GED_EVT_REG "EREG"
 #define AML_GED_EVT_SEL "ESEL"
@@ -87,6 +93,7 @@
 
 typedef struct GEDState {
     MemoryRegion evt;
+    MemoryRegion regs;
     uint32_t     sel;
 } GEDState;
 
hw/acpi/generic_event_device.c +44 -0
index 1cb34111e59a..3072fb4ab32b 100644
@@ -20,6 +20,7 @@
 #include "hw/qdev-properties.h"
 #include "migration/vmstate.h"
 #include "qemu/error-report.h"
+#include "sysemu/runstate.h"
 
 static const uint32_t ged_supported_events[] = {
     ACPI_GED_MEM_HOTPLUG_EVT,
@@ -176,6 +177,45 @@ static const MemoryRegionOps ged_evt_ops = {
     },
 };
 
+static uint64_t ged_regs_read(void *opaque, hwaddr addr, unsigned size)
+{
+    return 0;
+}
+
+static void ged_regs_write(void *opaque, hwaddr addr, uint64_t data,
+                           unsigned int size)
+{
+    bool slp_en;
+    int slp_typ;
+
+    switch (addr) {
+    case ACPI_GED_REG_SLEEP_CTL:
+        slp_typ = (data >> 2) & 0x07;
+        slp_en  = (data >> 5) & 0x01;
+        if (slp_en && slp_typ == 5) {
+            qemu_system_shutdown_request(SHUTDOWN_CAUSE_GUEST_SHUTDOWN);
+        }
+        return;
+    case ACPI_GED_REG_SLEEP_STS:
+        return;
+    case ACPI_GED_REG_RESET:
+        if (data == ACPI_GED_RESET_VALUE) {
+            qemu_system_reset_request(SHUTDOWN_CAUSE_GUEST_SHUTDOWN);
+        }
+        return;
+    }
+}
+
+static const MemoryRegionOps ged_regs_ops = {
+    .read = ged_regs_read,
+    .write = ged_regs_write,
+    .endianness = DEVICE_LITTLE_ENDIAN,
+    .valid = {
+        .min_access_size = 1,
+        .max_access_size = 1,
+    },
+};
+
 static void acpi_ged_device_plug_cb(HotplugHandler *hotplug_dev,
                                     DeviceState *dev, Error **errp)
 {
@@ -305,6 +345,10 @@ static void acpi_ged_initfn(Object *obj)
      sysbus_init_mmio(sbd, &s->container_memhp);
      acpi_memory_hotplug_init(&s->container_memhp, OBJECT(dev),
                               &s->memhp_state, 0);
+
+    memory_region_init_io(&ged_st->regs, obj, &ged_regs_ops, ged_st,
+                          TYPE_ACPI_GED "-regs", ACPI_GED_REG_COUNT);
+    sysbus_init_mmio(sbd, &ged_st->regs);
 }
 
 static void acpi_ged_class_init(ObjectClass *class, void *data)
#18 lore raw

PATCH v3 11/22 acpi: ged: add x86 device variant.

In reply to Gerd Hoffmann

Set AcpiDeviceIfClass->madt_cpu,
otherwise identical to TYPE_ACPI_GED.

---
 include/hw/acpi/generic_event_device.h |  4 +++
 hw/i386/generic_event_device_x86.c     | 36 ++++++++++++++++++++++++++
 hw/i386/Makefile.objs                  |  1 +
 3 files changed, 41 insertions(+)
 create mode 100644 hw/i386/generic_event_device_x86.c
include/hw/acpi/generic_event_device.h +4 -0
index 474c92198080..9687cb524258 100644
@@ -69,6 +69,10 @@
 #define ACPI_GED(obj) \
     OBJECT_CHECK(AcpiGedState, (obj), TYPE_ACPI_GED)
 
+#define TYPE_ACPI_GED_X86 "acpi-ged-x86"
+#define ACPI_GED_X86(obj) \
+    OBJECT_CHECK(AcpiGedX86State, (obj), TYPE_ACPI_GED_X86)
+
 #define ACPI_GED_EVT_SEL_OFFSET    0x0
 #define ACPI_GED_EVT_SEL_LEN       0x4
 
hw/i386/generic_event_device_x86.c new file +36 -0
new file mode 100644
index 000000000000..e26fb02a2ef6
@@ -0,0 +1,36 @@
+/*
+ * x86 variant of the generic event device for hw reduced acpi
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/acpi/generic_event_device.h"
+#include "hw/i386/pc.h"
+
+static void acpi_ged_x86_class_init(ObjectClass *class, void *data)
+{
+    AcpiDeviceIfClass *adevc = ACPI_DEVICE_IF_CLASS(class);
+
+    adevc->madt_cpu = pc_madt_cpu_entry;
+}
+
+static const TypeInfo acpi_ged_x86_info = {
+    .name          = TYPE_ACPI_GED_X86,
+    .parent        = TYPE_ACPI_GED,
+    .class_init    = acpi_ged_x86_class_init,
+    .interfaces = (InterfaceInfo[]) {
+        { TYPE_HOTPLUG_HANDLER },
+        { TYPE_ACPI_DEVICE_IF },
+        { }
+    }
+};
+
+static void acpi_ged_x86_register_types(void)
+{
+    type_register_static(&acpi_ged_x86_info);
+}
+
+type_init(acpi_ged_x86_register_types)
hw/i386/Makefile.objs +1 -0
index 6abc74551a72..622739305882 100644
@@ -17,4 +17,5 @@ obj-$(CONFIG_PC) += port92.o
 
 obj-y += kvmvapic.o
 obj-$(CONFIG_ACPI) += acpi-common.o
+obj-$(CONFIG_ACPI_HW_REDUCED) += generic_event_device_x86.o
 obj-$(CONFIG_PC) += acpi-build.o
Signature · 1 line
2.18.4
Reviewed-by on v3 11/22
#19 lore raw

Re: [PATCH v3 11/22] acpi: ged: add x86 device variant.

Quoted 4 lines · On Wed, 20 May 2020 15:19:52 +0200 Gerd Hoffmann <[email protected]> wrote:

Set AcpiDeviceIfClass->madt_cpu,
otherwise identical to TYPE_ACPI_GED.

Quoted 74 lines
---
 include/hw/acpi/generic_event_device.h |  4 +++
 hw/i386/generic_event_device_x86.c     | 36 ++++++++++++++++++++++++++
 hw/i386/Makefile.objs                  |  1 +
 3 files changed, 41 insertions(+)
 create mode 100644 hw/i386/generic_event_device_x86.c
include/hw/acpi/generic_event_device.h +4 -0
index 474c92198080..9687cb524258 100644
@@ -69,6 +69,10 @@
 #define ACPI_GED(obj) \
     OBJECT_CHECK(AcpiGedState, (obj), TYPE_ACPI_GED)
 
+#define TYPE_ACPI_GED_X86 "acpi-ged-x86"
+#define ACPI_GED_X86(obj) \
+    OBJECT_CHECK(AcpiGedX86State, (obj), TYPE_ACPI_GED_X86)
+
 #define ACPI_GED_EVT_SEL_OFFSET    0x0
 #define ACPI_GED_EVT_SEL_LEN       0x4
 
hw/i386/generic_event_device_x86.c new file +36 -0
new file mode 100644
index 000000000000..e26fb02a2ef6
@@ -0,0 +1,36 @@
+/*
+ * x86 variant of the generic event device for hw reduced acpi
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/acpi/generic_event_device.h"
+#include "hw/i386/pc.h"
+
+static void acpi_ged_x86_class_init(ObjectClass *class, void *data)
+{
+    AcpiDeviceIfClass *adevc = ACPI_DEVICE_IF_CLASS(class);
+
+    adevc->madt_cpu = pc_madt_cpu_entry;
+}
+
+static const TypeInfo acpi_ged_x86_info = {
+    .name          = TYPE_ACPI_GED_X86,
+    .parent        = TYPE_ACPI_GED,
+    .class_init    = acpi_ged_x86_class_init,
+    .interfaces = (InterfaceInfo[]) {
+        { TYPE_HOTPLUG_HANDLER },
+        { TYPE_ACPI_DEVICE_IF },
+        { }
+    }
+};
+
+static void acpi_ged_x86_register_types(void)
+{
+    type_register_static(&acpi_ged_x86_info);
+}
+
+type_init(acpi_ged_x86_register_types)
hw/i386/Makefile.objs +1 -0
index 6abc74551a72..622739305882 100644
@@ -17,4 +17,5 @@ obj-$(CONFIG_PC) += port92.o
 
 obj-y += kvmvapic.o
 obj-$(CONFIG_ACPI) += acpi-common.o
+obj-$(CONFIG_ACPI_HW_REDUCED) += generic_event_device_x86.o
 obj-$(CONFIG_PC) += acpi-build.o
#20 lore raw

PATCH v3 12/22 acpi: move acpi_dsdt_add_power_button() to ged

In reply to Gerd Hoffmann

Allow reuse for microvm.

---
 include/hw/acpi/generic_event_device.h | 1 +
 hw/acpi/generic_event_device.c         | 8 ++++++++
 hw/arm/virt-acpi-build.c               | 8 --------
 3 files changed, 9 insertions(+), 8 deletions(-)
include/hw/acpi/generic_event_device.h +1 -0
index 9687cb524258..ebed1ffe4231 100644
@@ -113,5 +113,6 @@ typedef struct AcpiGedState {
 
 void build_ged_aml(Aml *table, const char* name, HotplugHandler *hotplug_dev,
                    uint32_t ged_irq, AmlRegionSpace rs, hwaddr ged_base);
+void acpi_dsdt_add_power_button(Aml *scope);
 
 #endif
hw/acpi/generic_event_device.c +8 -0
index 3072fb4ab32b..cde50da5b3ca 100644
@@ -142,6 +142,14 @@ void build_ged_aml(Aml *table, const char *name, HotplugHandler *hotplug_dev,
     aml_append(table, dev);
 }
 
+void acpi_dsdt_add_power_button(Aml *scope)
+{
+    Aml *dev = aml_device(ACPI_POWER_BUTTON_DEVICE);
+    aml_append(dev, aml_name_decl("_HID", aml_string("PNP0C0C")));
+    aml_append(dev, aml_name_decl("_UID", aml_int(0)));
+    aml_append(scope, dev);
+}
+
 /* Memory read by the GED _EVT AML dynamic method */
 static uint64_t ged_evt_read(void *opaque, hwaddr addr, unsigned size)
 {
hw/arm/virt-acpi-build.c +0 -8
index 1b0a584c7b62..49eacc622219 100644
@@ -354,14 +354,6 @@ static void acpi_dsdt_add_gpio(Aml *scope, const MemMapEntry *gpio_memmap,
     aml_append(scope, dev);
 }
 
-static void acpi_dsdt_add_power_button(Aml *scope)
-{
-    Aml *dev = aml_device(ACPI_POWER_BUTTON_DEVICE);
-    aml_append(dev, aml_name_decl("_HID", aml_string("PNP0C0C")));
-    aml_append(dev, aml_name_decl("_UID", aml_int(0)));
-    aml_append(scope, dev);
-}
-
 static void
 build_iort(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
 {
Signature · 1 line
2.18.4
Reviewed-by on v3 12/22
#21 lore raw

Re: [PATCH v3 12/22] acpi: move acpi_dsdt_add_power_button() to ged

Quoted 49 lines · Gerd Hoffmann

Allow reuse for microvm.

---
  include/hw/acpi/generic_event_device.h | 1 +
  hw/acpi/generic_event_device.c         | 8 ++++++++
  hw/arm/virt-acpi-build.c               | 8 --------
  3 files changed, 9 insertions(+), 8 deletions(-)
include/hw/acpi/generic_event_device.h +1 -0
index 9687cb524258..ebed1ffe4231 100644
@@ -113,5 +113,6 @@ typedef struct AcpiGedState {
 
  void build_ged_aml(Aml *table, const char* name, HotplugHandler *hotplug_dev,
                     uint32_t ged_irq, AmlRegionSpace rs, hwaddr ged_base);
+void acpi_dsdt_add_power_button(Aml *scope);
 
  #endif
hw/acpi/generic_event_device.c +8 -0
index 3072fb4ab32b..cde50da5b3ca 100644
@@ -142,6 +142,14 @@ void build_ged_aml(Aml *table, const char *name, HotplugHandler *hotplug_dev,
      aml_append(table, dev);
  }
 
+void acpi_dsdt_add_power_button(Aml *scope)
+{
+    Aml *dev = aml_device(ACPI_POWER_BUTTON_DEVICE);
+    aml_append(dev, aml_name_decl("_HID", aml_string("PNP0C0C")));
+    aml_append(dev, aml_name_decl("_UID", aml_int(0)));
+    aml_append(scope, dev);
+}
+
  /* Memory read by the GED _EVT AML dynamic method */
  static uint64_t ged_evt_read(void *opaque, hwaddr addr, unsigned size)
  {
hw/arm/virt-acpi-build.c +0 -1
index 1b0a584c7b62..49eacc622219 100644
@@ -354,14 +354,6 @@ static void acpi_dsdt_add_gpio(Aml *scope, const MemMapEntry *gpio_memmap,
      aml_append(scope, dev);
  }
 
-static void acpi_dsdt_add_power_button(Aml *scope)

hw/arm/virt-acpi-build.c static void acpi_dsdt_add_gpio(Aml *scope, const MemMapEntry *gpio_memmap, · line 359

Quoted 10 lines
hw/arm/virt-acpi-build.c +0 -7
-{
-    Aml *dev = aml_device(ACPI_POWER_BUTTON_DEVICE);
-    aml_append(dev, aml_name_decl("_HID", aml_string("PNP0C0C")));
-    aml_append(dev, aml_name_decl("_UID", aml_int(0)));
-    aml_append(scope, dev);
-}
-
  static void
  build_iort(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
  {
Reviewed-by on v3 12/22
#22 lore raw

Re: [PATCH v3 12/22] acpi: move acpi_dsdt_add_power_button() to ged

In reply to Gerd Hoffmann
Quoted 3 lines · On Wed, 20 May 2020 15:19:53 +0200 Gerd Hoffmann <[email protected]> wrote:

Allow reuse for microvm.

Quoted 55 lines
---
 include/hw/acpi/generic_event_device.h | 1 +
 hw/acpi/generic_event_device.c         | 8 ++++++++
 hw/arm/virt-acpi-build.c               | 8 --------
 3 files changed, 9 insertions(+), 8 deletions(-)
include/hw/acpi/generic_event_device.h +1 -0
index 9687cb524258..ebed1ffe4231 100644
@@ -113,5 +113,6 @@ typedef struct AcpiGedState {
 
 void build_ged_aml(Aml *table, const char* name, HotplugHandler *hotplug_dev,
                    uint32_t ged_irq, AmlRegionSpace rs, hwaddr ged_base);
+void acpi_dsdt_add_power_button(Aml *scope);
 
 #endif
hw/acpi/generic_event_device.c +8 -0
index 3072fb4ab32b..cde50da5b3ca 100644
@@ -142,6 +142,14 @@ void build_ged_aml(Aml *table, const char *name, HotplugHandler *hotplug_dev,
     aml_append(table, dev);
 }
 
+void acpi_dsdt_add_power_button(Aml *scope)
+{
+    Aml *dev = aml_device(ACPI_POWER_BUTTON_DEVICE);
+    aml_append(dev, aml_name_decl("_HID", aml_string("PNP0C0C")));
+    aml_append(dev, aml_name_decl("_UID", aml_int(0)));
+    aml_append(scope, dev);
+}
+
 /* Memory read by the GED _EVT AML dynamic method */
 static uint64_t ged_evt_read(void *opaque, hwaddr addr, unsigned size)
 {
hw/arm/virt-acpi-build.c +0 -8
index 1b0a584c7b62..49eacc622219 100644
@@ -354,14 +354,6 @@ static void acpi_dsdt_add_gpio(Aml *scope, const MemMapEntry *gpio_memmap,
     aml_append(scope, dev);
 }
 
-static void acpi_dsdt_add_power_button(Aml *scope)
-{
-    Aml *dev = aml_device(ACPI_POWER_BUTTON_DEVICE);
-    aml_append(dev, aml_name_decl("_HID", aml_string("PNP0C0C")));
-    aml_append(dev, aml_name_decl("_UID", aml_int(0)));
-    aml_append(scope, dev);
-}
-
 static void
 build_iort(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
 {
#23 lore raw

PATCH v3 13/22 x86: coldplug cpus

In reply to Gerd Hoffmann

microvm needs this, for correct madt creation
(make sure enable bit is set).

---
 hw/i386/x86.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
hw/i386/x86.c +11 -0
index 7a3bc7ab6639..ac7a0a958781 100644
@@ -118,6 +118,7 @@ uint32_t x86_cpu_apic_id_from_index(X86MachineState *x86ms,
 
 void x86_cpu_new(X86MachineState *x86ms, int64_t apic_id, Error **errp)
 {
+    MachineClass *mc = MACHINE_GET_CLASS(x86ms);
     Object *cpu = NULL;
     Error *local_err = NULL;
 
@@ -126,6 +127,16 @@ void x86_cpu_new(X86MachineState *x86ms, int64_t apic_id, Error **errp)
     object_property_set_uint(cpu, apic_id, "apic-id", &local_err);
     object_property_set_bool(cpu, true, "realized", &local_err);
 
+    if (!mc->has_hotpluggable_cpus) {
+        /* coldplug cpu */
+        MachineState *ms = MACHINE(x86ms);
+        int i = 0;
+        while (ms->possible_cpus->cpus[i].arch_id != apic_id) {
+            i++;
+        }
+        ms->possible_cpus->cpus[i].cpu = cpu;
+    }
+
     object_unref(cpu);
     error_propagate(errp, local_err);
 }
Signature · 1 line
2.18.4
#24 lore raw

Re: [PATCH v3 13/22] x86: coldplug cpus

Quoted 23 lines · On Wed, 20 May 2020 15:19:54 +0200 Gerd Hoffmann <[email protected]> wrote:

microvm needs this, for correct madt creation
(make sure enable bit is set).

---
 hw/i386/x86.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
hw/i386/x86.c +1 -0
index 7a3bc7ab6639..ac7a0a958781 100644
@@ -118,6 +118,7 @@ uint32_t x86_cpu_apic_id_from_index(X86MachineState *x86ms,
 
 void x86_cpu_new(X86MachineState *x86ms, int64_t apic_id, Error **errp)
 {
+    MachineClass *mc = MACHINE_GET_CLASS(x86ms);
     Object *cpu = NULL;
     Error *local_err = NULL;
 
@@ -126,6 +127,16 @@ void x86_cpu_new(X86MachineState *x86ms, int64_t apic_id, Error **errp)
     object_property_set_uint(cpu, apic_id, "apic-id", &local_err);
     object_property_set_bool(cpu, true, "realized", &local_err);

hw/i386/x86.c void x86_cpu_new(X86MachineState *x86ms, int64_t apic_id, Error **errp) · line 138

Quoted 10 lines
hw/i386/x86.c +9 -0
 
+    if (!mc->has_hotpluggable_cpus) {
+        /* coldplug cpu */
+        MachineState *ms = MACHINE(x86ms);
+        int i = 0;
+        while (ms->possible_cpus->cpus[i].arch_id != apic_id) {
+            i++;
+        }
+        ms->possible_cpus->cpus[i].cpu = cpu;
+    }

I'm thinking about removing x86_cpu_new alogether (in favor of device_add), so I'd preffer avoid adding direct access to possible_cpus here.

that's the job of pc_cpu_plug (and possibly pc_cpu_pre_plug) in combination with static const TypeInfo pc_machine_info = {
.name = TYPE_PC_MACHINE,
...

    .interfaces = (InterfaceInfo[]) {
         { TYPE_HOTPLUG_HANDLER },
         { }
    },
};

I'd reuse common code if possible in microvm or make microvm specific simplified version of it if it makes job easier.

Quoted 3 lines
hw/i386/x86.c +0 -0
     object_unref(cpu);
     error_propagate(errp, local_err);
 }
#25 lore raw

PATCH v3 14/22 microvm: use 2G split unconditionally

In reply to Gerd Hoffmann

Looks like the logiv was copied over from q35.

q35 does this for backward compatibility, there is no reason to do this on microvm though. So split @ 2G unconditionally.

---
 hw/i386/microvm.c | 16 +---------------
 1 file changed, 1 insertion(+), 15 deletions(-)
hw/i386/microvm.c +1 -15
index 867d3d652145..b8f0d3283758 100644
@@ -170,23 +170,9 @@ static void microvm_memory_init(MicrovmMachineState *mms)
     MemoryRegion *ram_below_4g, *ram_above_4g;
     MemoryRegion *system_memory = get_system_memory();
     FWCfgState *fw_cfg;
-    ram_addr_t lowmem;
+    ram_addr_t lowmem = 0x80000000; /* 2G */
     int i;
 
-    /*
-     * Check whether RAM fits below 4G (leaving 1/2 GByte for IO memory
-     * and 256 Mbytes for PCI Express Enhanced Configuration Access Mapping
-     * also known as MMCFG).
-     * If it doesn't, we need to split it in chunks below and above 4G.
-     * In any case, try to make sure that guest addresses aligned at
-     * 1G boundaries get mapped to host addresses aligned at 1G boundaries.
-     */
-    if (machine->ram_size >= 0xb0000000) {
-        lowmem = 0x80000000;
-    } else {
-        lowmem = 0xb0000000;
-    }
-
     /*
      * Handle the machine opt max-ram-below-4g.  It is basically doing
      * min(qemu limit, user limit).
Signature · 1 line
2.18.4
Reviewed-by on v3 14/22
#26 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

Quoted 1 line · Gerd Hoffmann

Looks like the logiv was copied over from q35.

Typo 'logiv' -> 'logic'.

Quoted 3 lines

q35 does this for backward compatibility, there is no reason to do this on microvm though. So split @ 2G unconditionally.

Yes please!

Quoted 36 lines
---
  hw/i386/microvm.c | 16 +---------------
  1 file changed, 1 insertion(+), 15 deletions(-)
hw/i386/microvm.c +1 -15
index 867d3d652145..b8f0d3283758 100644
@@ -170,23 +170,9 @@ static void microvm_memory_init(MicrovmMachineState *mms)
      MemoryRegion *ram_below_4g, *ram_above_4g;
      MemoryRegion *system_memory = get_system_memory();
      FWCfgState *fw_cfg;
-    ram_addr_t lowmem;
+    ram_addr_t lowmem = 0x80000000; /* 2G */
      int i;
 
-    /*
-     * Check whether RAM fits below 4G (leaving 1/2 GByte for IO memory
-     * and 256 Mbytes for PCI Express Enhanced Configuration Access Mapping
-     * also known as MMCFG).
-     * If it doesn't, we need to split it in chunks below and above 4G.
-     * In any case, try to make sure that guest addresses aligned at
-     * 1G boundaries get mapped to host addresses aligned at 1G boundaries.
-     */
-    if (machine->ram_size >= 0xb0000000) {
-        lowmem = 0x80000000;
-    } else {
-        lowmem = 0xb0000000;
-    }
-
      /*
       * Handle the machine opt max-ram-below-4g.  It is basically doing
       * min(qemu limit, user limit).
Reviewed-by on v3 14/22
#27 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

In reply to Gerd Hoffmann
Quoted 6 lines · On Wed, 20 May 2020 15:19:55 +0200 Gerd Hoffmann <[email protected]> wrote:

Looks like the logiv was copied over from q35.

q35 does this for backward compatibility, there is no reason to do this on microvm though. So split @ 2G unconditionally.

Quoted 33 lines
---
 hw/i386/microvm.c | 16 +---------------
 1 file changed, 1 insertion(+), 15 deletions(-)
hw/i386/microvm.c +1 -15
index 867d3d652145..b8f0d3283758 100644
@@ -170,23 +170,9 @@ static void microvm_memory_init(MicrovmMachineState *mms)
     MemoryRegion *ram_below_4g, *ram_above_4g;
     MemoryRegion *system_memory = get_system_memory();
     FWCfgState *fw_cfg;
-    ram_addr_t lowmem;
+    ram_addr_t lowmem = 0x80000000; /* 2G */
     int i;
 
-    /*
-     * Check whether RAM fits below 4G (leaving 1/2 GByte for IO memory
-     * and 256 Mbytes for PCI Express Enhanced Configuration Access Mapping
-     * also known as MMCFG).
-     * If it doesn't, we need to split it in chunks below and above 4G.
-     * In any case, try to make sure that guest addresses aligned at
-     * 1G boundaries get mapped to host addresses aligned at 1G boundaries.
-     */
-    if (machine->ram_size >= 0xb0000000) {
-        lowmem = 0x80000000;
-    } else {
-        lowmem = 0xb0000000;
-    }
-
     /*
      * Handle the machine opt max-ram-below-4g.  It is basically doing
      * min(qemu limit, user limit).
#28 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

In reply to Gerd Hoffmann
Quoted 4 lines · On Wed, 20 May 2020 15:19:55 +0200 Gerd Hoffmann <[email protected]> wrote:

Looks like the logiv was copied over from q35.

q35 does this for backward compatibility, there is no reason to do this on microvm though. So split @ 2G unconditionally.

not related to your ACPI rework, but just an idea for future of microvm

I wonder if we should carry over all this fixed RAM layout legacy from pc/q35 with a bunch of knobs to tweak it (along with complicated logic).

Can we just re-use pc-dimms for main RAM and let user specify RAM layout the way they wish?

Quoted 34 lines
---
 hw/i386/microvm.c | 16 +---------------
 1 file changed, 1 insertion(+), 15 deletions(-)
hw/i386/microvm.c +1 -15
index 867d3d652145..b8f0d3283758 100644
@@ -170,23 +170,9 @@ static void microvm_memory_init(MicrovmMachineState *mms)
     MemoryRegion *ram_below_4g, *ram_above_4g;
     MemoryRegion *system_memory = get_system_memory();
     FWCfgState *fw_cfg;
-    ram_addr_t lowmem;
+    ram_addr_t lowmem = 0x80000000; /* 2G */
     int i;
 
-    /*
-     * Check whether RAM fits below 4G (leaving 1/2 GByte for IO memory
-     * and 256 Mbytes for PCI Express Enhanced Configuration Access Mapping
-     * also known as MMCFG).
-     * If it doesn't, we need to split it in chunks below and above 4G.
-     * In any case, try to make sure that guest addresses aligned at
-     * 1G boundaries get mapped to host addresses aligned at 1G boundaries.
-     */
-    if (machine->ram_size >= 0xb0000000) {
-        lowmem = 0x80000000;
-    } else {
-        lowmem = 0xb0000000;
-    }
-
     /*
      * Handle the machine opt max-ram-below-4g.  It is basically doing
      * min(qemu limit, user limit).
#29 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

Quoted 5 lines · Igor Mammedov
Quoted 2 lines · On Wed, 20 May 2020 15:19:55 +0200 Gerd Hoffmann <[email protected]> wrote:

Looks like the logiv was copied over from q35.

Quoted 7 lines
Quoted 2 lines

q35 does this for backward compatibility, there is no reason to do this on microvm though. So split @ 2G unconditionally.

not related to your ACPI rework, but just an idea for future of microvm

I wonder if we should carry over all this fixed RAM layout legacy from pc/q35 with a bunch of knobs to tweak it (along with complicated logic).

Well, I think we can (should) drop max-ram-below-4g too. There is no reason to use that with microvm, other that shooting yourself into the foot (by making mmio overlap with ram).

With that being gone too there isn't much logic left ...

take care,
Gerd

#30 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

Quoted 11 lines · On Mon, 25 May 2020 13:45:08 +0200 Gerd Hoffmann <[email protected]> wrote:
Quoted 10 lines · Igor Mammedov
Quoted 4 lines · On Wed, 20 May 2020 15:19:55 +0200 Gerd Hoffmann <[email protected]> wrote:

Looks like the logiv was copied over from q35.

q35 does this for backward compatibility, there is no reason to do this on microvm though. So split @ 2G unconditionally.

not related to your ACPI rework, but just an idea for future of microvm

Quoted 8 lines
Quoted 2 lines

I wonder if we should carry over all this fixed RAM layout legacy from pc/q35 with a bunch of knobs to tweak it (along with complicated logic).

Well, I think we can (should) drop max-ram-below-4g too. There is no reason to use that with microvm, other that shooting yourself into the foot (by making mmio overlap with ram).

With that being gone too there isn't much logic left ...

I wonder if we need 2G split for microvm at all?
Can we map 1 contiguous big blob from 0 GPA and overlay bios & other x86 TOLUD stuff?

Quoted 3 lines

take care,
Gerd

#31 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

Quoted 8 lines
Quoted 5 lines

Well, I think we can (should) drop max-ram-below-4g too. There is no reason to use that with microvm, other that shooting yourself into the foot (by making mmio overlap with ram).

With that being gone too there isn't much logic left ...

I wonder if we need 2G split for microvm at all?
Can we map 1 contiguous big blob from 0 GPA and overlay bios & other x86 TOLUD stuff?

I think it would work, but it has some drawbacks:

(1) we loose a bit of memory.
(2) we loose a gigabyte page.
(3) we wouldn't have guard pages (unused address space) between between ram and mmio space.

take care,
Gerd

#32 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

Quoted 4 lines · On Tue, 26 May 2020 06:48:39 +0200 Gerd Hoffmann <[email protected]> wrote:
Quoted 4 lines
Quoted 4 lines

Well, I think we can (should) drop max-ram-below-4g too. There is no reason to use that with microvm, other that shooting yourself into the foot (by making mmio overlap with ram).

Quoted 8 lines
Quoted 4 lines
Quoted 1 line

With that being gone too there isn't much logic left ...

I wonder if we need 2G split for microvm at all?
Can we map 1 contiguous big blob from 0 GPA and overlay bios & other x86 TOLUD stuff?

I think it would work, but it has some drawbacks:

(1) we loose a bit of memory.

it's probably not a big enough to care about, we do similar ovarlay mapping on pc/q35 at the beginning of RAM

Quoted 1 line

(2) we loose a gigabyte page.

I'm not sure waht exactly we loose in this case.
Lets assume we allocating guest 5G of continuous RAM using 1G huge pages, in this case I'd think that on host side MMIO overlay won't affect RAM blob on guest side pagetables will be fragmented due to MMIO holes, but guest still could use huge pages smaller ones in fragmented area and 1G where there is no fragmentation.

Quoted 2 lines

(3) we wouldn't have guard pages (unused address space) between between ram and mmio space.

if it's holes' mmio,then do we really need them (access is going to be terminated either in always valid RAM or in valid mmio hole)?

Quoted 5 lines

take care,
Gerd

#33 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

Quoted 6 lines · Igor Mammedov
Quoted 1 line

(2) we loose a gigabyte page.

I'm not sure waht exactly we loose in this case.
Lets assume we allocating guest 5G of continuous RAM using 1G huge pages, in this case I'd think that on host side MMIO overlay won't affect RAM blob on guest side pagetables will be fragmented due to MMIO holes, but guest still could use huge pages smaller ones in fragmented area and 1G where there is no fragmentation.

Access to the 3G-4G area would not be able to use 1G EPT pages.

But why use 2G split instead of 3G? There's only very little MMIO and no PCI hole (including no huge MMCONFIG BAR) on microvm.

Paolo

#34 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

Quoted 9 lines · On Wed, 27 May 2020 15:06:28 +0200 Paolo Bonzini <[email protected]> wrote:
Quoted 6 lines · Igor Mammedov
Quoted 1 line

(2) we loose a gigabyte page.

I'm not sure waht exactly we loose in this case.
Lets assume we allocating guest 5G of continuous RAM using 1G huge pages, in this case I'd think that on host side MMIO overlay won't affect RAM blob on guest side pagetables will be fragmented due to MMIO holes, but guest still could use huge pages smaller ones in fragmented area and 1G where there is no fragmentation.

Access to the 3G-4G area would not be able to use 1G EPT pages.

Could it use 2Mb pages instead of 1Gb?
Do we really care about 1 gigabyte huge page in microvm intended usecase?
(fast starting VMs for microservices like FaaS, which unlikely would use much memory to begin with)

Quoted 5 lines

But why use 2G split instead of 3G? There's only very little MMIO and no PCI hole (including no huge MMCONFIG BAR) on microvm.

Paolo

#35 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

Quoted 11 lines · On Wed, 27 May 2020 16:26:46 +0200 Igor Mammedov <[email protected]> wrote:
Quoted 8 lines · On Wed, 27 May 2020 15:06:28 +0200 Paolo Bonzini <[email protected]> wrote:
Quoted 6 lines · Igor Mammedov
Quoted 1 line

(2) we loose a gigabyte page.

I'm not sure waht exactly we loose in this case.
Lets assume we allocating guest 5G of continuous RAM using 1G huge pages, in this case I'd think that on host side MMIO overlay won't affect RAM blob on guest side pagetables will be fragmented due to MMIO holes, but guest still could use huge pages smaller ones in fragmented area and 1G where there is no fragmentation.

Quoted 4 lines
Quoted 1 line

Access to the 3G-4G area would not be able to use 1G EPT pages.

Could it use 2Mb pages instead of 1Gb?
Do we really care about 1 gigabyte huge page in microvm intended usecase?
(fast starting VMs for microservices like FaaS, which unlikely would use much memory to begin with)

my interest in having single memory region, is in possibility of drop in conversion to [nv|pc-dimm] later on without breaking ABI. (I'm not sure that we actually need it though)

Quoted 7 lines
Quoted 5 lines

But why use 2G split instead of 3G? There's only very little MMIO and no PCI hole (including no huge MMCONFIG BAR) on microvm.

Paolo

#36 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

In reply to Igor Mammedov
Quoted 5 lines · Igor Mammedov
Quoted 2 lines · On Wed, 27 May 2020 15:06:28 +0200 Paolo Bonzini <[email protected]> wrote:
Quoted 1 line · Igor Mammedov
Quoted 1 line

(2) we loose a gigabyte page.

Quoted 8 lines
Quoted 7 lines
Quoted 5 lines

I'm not sure waht exactly we loose in this case.
Lets assume we allocating guest 5G of continuous RAM using 1G huge pages, in this case I'd think that on host side MMIO overlay won't affect RAM blob on guest side pagetables will be fragmented due to MMIO holes, but guest still could use huge pages smaller ones in fragmented area and 1G where there is no fragmentation.

Access to the 3G-4G area would not be able to use 1G EPT pages.

Could it use 2Mb pages instead of 1Gb?

Yes, probably a mix of 2 MiB pages and 4 KiB pages around the memslot splits.

Quoted 2 lines

Do we really care about 1 gigabyte huge page in microvm intended usecase?
(fast starting VMs for microservices like FaaS, which unlikely would use much memory to begin with)

I honestly don't think it's measurable, but at least in theory we care because such workloads could have more TLB misses (relative to the execution time) than long-lasting VMs.

Paolo

#37 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

In reply to Paolo Bonzini

Hi,

Quoted 2 lines

But why use 2G split instead of 3G? There's only very little MMIO and no PCI hole (including no huge MMCONFIG BAR) on microvm.

Yes, we can go for 3G, we are indeed not short on address space ;)

take care,
Gerd

#38 lore raw

Re: [PATCH v3 14/22] microvm: use 2G split unconditionally

In reply to Igor Mammedov

Hi,

Quoted 3 lines
Quoted 1 line

(1) we loose a bit of memory.

it's probably not a big enough to care about, we do similar ovarlay mapping on pc/q35 at the beginning of RAM

Yes, shouldn't be too much.

Quoted 2 lines
Quoted 1 line

(2) we loose a gigabyte page.

I'm not sure waht exactly we loose in this case.

The 1G page for 0xc0000000 -> 0xffffffff (as explained by paolo).

Quoted 4 lines
Quoted 2 lines

(3) we wouldn't have guard pages (unused address space) between between ram and mmio space.

if it's holes' mmio,then do we really need them (access is going to be terminated either in always valid RAM or in valid mmio hole)?

Not required, but more robust. Less likely that the guest touches mmio by accident.

I'd expect it also requires some e820 hacks.

cheers,
Gerd

#39 lore raw

PATCH v3 15/22 microvm: make virtio irq base runtime configurable

In reply to Gerd Hoffmann
---
 include/hw/i386/microvm.h |  2 +-
 hw/i386/microvm.c         | 11 +++++++----
 2 files changed, 8 insertions(+), 5 deletions(-)
include/hw/i386/microvm.h +1 -1
index ba68d1f22bb3..2709306c4016 100644
@@ -27,7 +27,6 @@
 
 /* Platform virtio definitions */
 #define VIRTIO_MMIO_BASE      0xc0000000
-#define VIRTIO_IRQ_BASE       5
 #define VIRTIO_NUM_TRANSPORTS 8
 #define VIRTIO_CMDLINE_MAXLEN 64
 
@@ -57,6 +56,7 @@ typedef struct {
     bool auto_kernel_cmdline;
 
     /* Machine state */
+    uint32_t virtio_irq_base;
     bool kernel_cmdline_fixed;
 } MicrovmMachineState;
 
hw/i386/microvm.c +7 -4
index b8f0d3283758..ea4460714a99 100644
@@ -121,10 +121,11 @@ static void microvm_devices_init(MicrovmMachineState *mms)
 
     kvmclock_create();
 
+    mms->virtio_irq_base = 8;
     for (i = 0; i < VIRTIO_NUM_TRANSPORTS; i++) {
         sysbus_create_simple("virtio-mmio",
                              VIRTIO_MMIO_BASE + i * 512,
-                             x86ms->gsi[VIRTIO_IRQ_BASE + i]);
+                             x86ms->gsi[mms->virtio_irq_base + i]);
     }
 
     /* Optional and legacy devices */
@@ -246,7 +247,7 @@ static void microvm_memory_init(MicrovmMachineState *mms)
     x86ms->ioapic_as = &address_space_memory;
 }
 
-static gchar *microvm_get_mmio_cmdline(gchar *name)
+static gchar *microvm_get_mmio_cmdline(gchar *name, uint32_t virtio_irq_base)
 {
     gchar *cmdline;
     gchar *separator;
@@ -266,7 +267,7 @@ static gchar *microvm_get_mmio_cmdline(gchar *name)
     ret = g_snprintf(cmdline, VIRTIO_CMDLINE_MAXLEN,
                      " virtio_mmio.device=512@0x%lx:%ld",
                      VIRTIO_MMIO_BASE + index * 512,
-                     VIRTIO_IRQ_BASE + index);
+                     virtio_irq_base + index);
     if (ret < 0 || ret >= VIRTIO_CMDLINE_MAXLEN) {
         g_free(cmdline);
         return NULL;
@@ -278,6 +279,7 @@ static gchar *microvm_get_mmio_cmdline(gchar *name)
 static void microvm_fix_kernel_cmdline(MachineState *machine)
 {
     X86MachineState *x86ms = X86_MACHINE(machine);
+    MicrovmMachineState *mms = MICROVM_MACHINE(machine);
     BusState *bus;
     BusChild *kid;
     char *cmdline;
@@ -301,7 +303,8 @@ static void microvm_fix_kernel_cmdline(MachineState *machine)
             BusState *mmio_bus = &mmio_virtio_bus->parent_obj;
 
             if (!QTAILQ_EMPTY(&mmio_bus->children)) {
-                gchar *mmio_cmdline = microvm_get_mmio_cmdline(mmio_bus->name);
+                gchar *mmio_cmdline = microvm_get_mmio_cmdline
+                    (mmio_bus->name, mms->virtio_irq_base);
                 if (mmio_cmdline) {
                     char *newcmd = g_strjoin(NULL, cmdline, mmio_cmdline, NULL);
                     g_free(mmio_cmdline);
Signature · 1 line
2.18.4
#40 lore raw

Re: [PATCH v3 15/22] microvm: make virtio irq base runtime configurable

Quoted 32 lines · Gerd Hoffmann
---
  include/hw/i386/microvm.h |  2 +-
  hw/i386/microvm.c         | 11 +++++++----
  2 files changed, 8 insertions(+), 5 deletions(-)
include/hw/i386/microvm.h +1 -1
index ba68d1f22bb3..2709306c4016 100644
@@ -27,7 +27,6 @@
 
  /* Platform virtio definitions */
  #define VIRTIO_MMIO_BASE      0xc0000000
-#define VIRTIO_IRQ_BASE       5
  #define VIRTIO_NUM_TRANSPORTS 8
  #define VIRTIO_CMDLINE_MAXLEN 64
 
@@ -57,6 +56,7 @@ typedef struct {
      bool auto_kernel_cmdline;
 
      /* Machine state */
+    uint32_t virtio_irq_base;
      bool kernel_cmdline_fixed;
  } MicrovmMachineState;
 
hw/i386/microvm.c +0 -0
index b8f0d3283758..ea4460714a99 100644

hw/i386/microvm.c static void microvm_devices_init(MicrovmMachineState *mms) · line 124

Quoted 4 lines
hw/i386/microvm.c +1 -0
@@ -121,10 +121,11 @@ static void microvm_devices_init(MicrovmMachineState *mms)
 
      kvmclock_create();
 
+    mms->virtio_irq_base = 8;

Hmm you made it runtime configurable indeed, but you also changed it from 5 to 8. Worth a different patch, or at least mentioning in commit description.

Quoted 45 lines
hw/i386/microvm.c +6 -4
      for (i = 0; i < VIRTIO_NUM_TRANSPORTS; i++) {
          sysbus_create_simple("virtio-mmio",
                               VIRTIO_MMIO_BASE + i * 512,
-                             x86ms->gsi[VIRTIO_IRQ_BASE + i]);
+                             x86ms->gsi[mms->virtio_irq_base + i]);
      }
 
      /* Optional and legacy devices */
@@ -246,7 +247,7 @@ static void microvm_memory_init(MicrovmMachineState *mms)
      x86ms->ioapic_as = &address_space_memory;
  }
 
-static gchar *microvm_get_mmio_cmdline(gchar *name)
+static gchar *microvm_get_mmio_cmdline(gchar *name, uint32_t virtio_irq_base)
  {
      gchar *cmdline;
      gchar *separator;
@@ -266,7 +267,7 @@ static gchar *microvm_get_mmio_cmdline(gchar *name)
      ret = g_snprintf(cmdline, VIRTIO_CMDLINE_MAXLEN,
                       " virtio_mmio.device=512@0x%lx:%ld",
                       VIRTIO_MMIO_BASE + index * 512,
-                     VIRTIO_IRQ_BASE + index);
+                     virtio_irq_base + index);
      if (ret < 0 || ret >= VIRTIO_CMDLINE_MAXLEN) {
          g_free(cmdline);
          return NULL;
@@ -278,6 +279,7 @@ static gchar *microvm_get_mmio_cmdline(gchar *name)
  static void microvm_fix_kernel_cmdline(MachineState *machine)
  {
      X86MachineState *x86ms = X86_MACHINE(machine);
+    MicrovmMachineState *mms = MICROVM_MACHINE(machine);
      BusState *bus;
      BusChild *kid;
      char *cmdline;
@@ -301,7 +303,8 @@ static void microvm_fix_kernel_cmdline(MachineState *machine)
              BusState *mmio_bus = &mmio_virtio_bus->parent_obj;
 
              if (!QTAILQ_EMPTY(&mmio_bus->children)) {
-                gchar *mmio_cmdline = microvm_get_mmio_cmdline(mmio_bus->name);
+                gchar *mmio_cmdline = microvm_get_mmio_cmdline
+                    (mmio_bus->name, mms->virtio_irq_base);
                  if (mmio_cmdline) {
                      char *newcmd = g_strjoin(NULL, cmdline, mmio_cmdline, NULL);
                      g_free(mmio_cmdline);
#41 lore raw

Re: [PATCH v3 15/22] microvm: make virtio irq base runtime configurable

Hi,

Quoted 6 lines

include/hw/i386/microvm.h

Quoted 3 lines
include/hw/i386/microvm.h +1 -1
-#define VIRTIO_IRQ_BASE       5
 
+    mms->virtio_irq_base = 8;

Hmm you made it runtime configurable indeed, but you also changed it from 5 to 8. Worth a different patch, or at least mentioning in commit description.

Oops, that wasn't intentional. I'll change it back to 5.

take care,
Gerd

#42 lore raw

Re: [PATCH v3 15/22] microvm: make virtio irq base runtime configurable

Quoted 10 lines · Gerd Hoffmann

Hi,

Quoted 6 lines

include/hw/i386/microvm.h

Quoted 3 lines
include/hw/i386/microvm.h +1 -1
-#define VIRTIO_IRQ_BASE       5
 
+    mms->virtio_irq_base = 8;

Hmm you made it runtime configurable indeed, but you also changed it from 5 to 8. Worth a different patch, or at least mentioning in commit description.

Oops, that wasn't intentional. I'll change it back to 5.

OK. Using 5:

#43 lore raw

PATCH v3 16/22 microvm/acpi: add minimal acpi support

In reply to Gerd Hoffmann

$subject says all. Can be controlled using -M microvm,acpi=on/off.

---
 hw/i386/acpi-microvm.h    |   6 ++
 include/hw/i386/microvm.h |   9 ++
 hw/i386/acpi-microvm.c    | 182 ++++++++++++++++++++++++++++++++++++++
 hw/i386/microvm.c         |  37 ++++++++
 hw/i386/Kconfig           |   1 +
 hw/i386/Makefile.objs     |   1 +
 6 files changed, 236 insertions(+)
 create mode 100644 hw/i386/acpi-microvm.h
 create mode 100644 hw/i386/acpi-microvm.c
hw/i386/acpi-microvm.h new file +6 -0
new file mode 100644
index 000000000000..6a6c2967102b
@@ -0,0 +1,6 @@
+#ifndef HW_I386_ACPI_MICROVM_H
+#define HW_I386_ACPI_MICROVM_H
+
+void acpi_setup_microvm(void);
+
+#endif
include/hw/i386/microvm.h +9 -0
index 2709306c4016..ebd7f49d5de8 100644
@@ -24,12 +24,18 @@
 
 #include "hw/boards.h"
 #include "hw/i386/x86.h"
+#include "hw/acpi/acpi_dev_interface.h"
 
 /* Platform virtio definitions */
 #define VIRTIO_MMIO_BASE      0xc0000000
 #define VIRTIO_NUM_TRANSPORTS 8
 #define VIRTIO_CMDLINE_MAXLEN 64
 
+#define GED_MMIO_BASE         0xc1000000
+#define GED_MMIO_BASE_MEMHP   (GED_MMIO_BASE + 0x100)
+#define GED_MMIO_BASE_REGS    (GED_MMIO_BASE + 0x200)
+#define GED_MMIO_IRQ          9
+
 /* Machine type options */
 #define MICROVM_MACHINE_PIT                 "pit"
 #define MICROVM_MACHINE_PIC                 "pic"
@@ -58,6 +64,9 @@ typedef struct {
     /* Machine state */
     uint32_t virtio_irq_base;
     bool kernel_cmdline_fixed;
+    Notifier machine_done;
+    Notifier powerdown_req;
+    AcpiDeviceIf *acpi_dev;
 } MicrovmMachineState;
 
 #define TYPE_MICROVM_MACHINE   MACHINE_TYPE_NAME("microvm")
hw/i386/acpi-microvm.c new file +182 -0

182 lines not on this page · raw message

hw/i386/microvm.c +37 -0
index ea4460714a99..b38321112baa 100644
@@ -26,6 +26,8 @@
 #include "sysemu/cpus.h"
 #include "sysemu/numa.h"
 #include "sysemu/reset.h"
+#include "sysemu/runstate.h"
+#include "acpi-microvm.h"
 
 #include "hw/loader.h"
 #include "hw/irq.h"
@@ -41,6 +43,8 @@
 #include "hw/i386/e820_memory_layout.h"
 #include "hw/i386/fw_cfg.h"
 #include "hw/virtio/virtio-mmio.h"
+#include "hw/acpi/acpi.h"
+#include "hw/acpi/generic_event_device.h"
 
 #include "cpu.h"
 #include "elf.h"
@@ -129,6 +133,17 @@ static void microvm_devices_init(MicrovmMachineState *mms)
     }
 
     /* Optional and legacy devices */
+    if (x86_machine_is_acpi_enabled(x86ms)) {
+        DeviceState *dev = qdev_create(NULL, TYPE_ACPI_GED_X86);
+        qdev_prop_set_uint32(dev, "ged-event", ACPI_GED_PWR_DOWN_EVT);
+        sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, GED_MMIO_BASE);
+        /* sysbus_mmio_map(SYS_BUS_DEVICE(dev), 1, GED_MMIO_BASE_MEMHP); */
+        sysbus_mmio_map(SYS_BUS_DEVICE(dev), 2, GED_MMIO_BASE_REGS);
+        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0,
+                           x86ms->gsi[GED_MMIO_IRQ]);
+        qdev_init_nofail(dev);
+        mms->acpi_dev = ACPI_DEVICE_IF(dev);
+    }
 
     if (mms->pic == ON_OFF_AUTO_ON || mms->pic == ON_OFF_AUTO_AUTO) {
         qemu_irq *i8259;
@@ -457,6 +472,23 @@ static void microvm_machine_set_auto_kernel_cmdline(Object *obj, bool value,
     mms->auto_kernel_cmdline = value;
 }
 
+static void microvm_machine_done(Notifier *notifier, void *data)
+{
+    acpi_setup_microvm();
+}
+
+static void microvm_powerdown_req(Notifier *notifier, void *data)
+{
+    MicrovmMachineState *mms = container_of(notifier, MicrovmMachineState,
+                                            powerdown_req);
+
+    if (mms->acpi_dev) {
+        Object *obj = OBJECT(mms->acpi_dev);
+        AcpiDeviceIfClass *adevc = ACPI_DEVICE_IF_GET_CLASS(obj);
+        adevc->send_event(mms->acpi_dev, ACPI_POWER_DOWN_STATUS);
+    }
+}
+
 static void microvm_machine_initfn(Object *obj)
 {
     MicrovmMachineState *mms = MICROVM_MACHINE(obj);
@@ -471,6 +503,11 @@ static void microvm_machine_initfn(Object *obj)
 
     /* State */
     mms->kernel_cmdline_fixed = false;
+
+    mms->machine_done.notify = microvm_machine_done;
+    qemu_add_machine_init_done_notifier(&mms->machine_done);
+    mms->powerdown_req.notify = microvm_powerdown_req;
+    qemu_register_powerdown_notifier(&mms->powerdown_req);
 }
 
 static void microvm_class_init(ObjectClass *oc, void *data)
hw/i386/Kconfig +1 -0
index c93f32f6579d..be746bcb49eb 100644
@@ -102,6 +102,7 @@ config MICROVM
     select I8259
     select MC146818RTC
     select VIRTIO_MMIO
+    select ACPI_HW_REDUCED
 
 config X86_IOMMU
     bool
hw/i386/Makefile.objs +1 -0
index 622739305882..bbb2fe78f3cd 100644
@@ -19,3 +19,4 @@ obj-y += kvmvapic.o
 obj-$(CONFIG_ACPI) += acpi-common.o
 obj-$(CONFIG_ACPI_HW_REDUCED) += generic_event_device_x86.o
 obj-$(CONFIG_PC) += acpi-build.o
+obj-$(CONFIG_MICROVM) += acpi-microvm.o
Signature · 1 line
2.18.4
Reviewed-by on v3 16/22
#44 lore raw

Re: [PATCH v3 16/22] microvm/acpi: add minimal acpi support

Quoted 1 line · On Wed, 20 May 2020 15:19:57 +0200 Gerd Hoffmann <[email protected]> wrote:

$subject says all. Can be controlled using -M microvm,acpi=on/off.

thanks for really nice and clean impl. There are a couple minor notes below, but otherwise

Quoted 64 lines
---
 hw/i386/acpi-microvm.h    |   6 ++
 include/hw/i386/microvm.h |   9 ++
 hw/i386/acpi-microvm.c    | 182 ++++++++++++++++++++++++++++++++++++++
 hw/i386/microvm.c         |  37 ++++++++
 hw/i386/Kconfig           |   1 +
 hw/i386/Makefile.objs     |   1 +
 6 files changed, 236 insertions(+)
 create mode 100644 hw/i386/acpi-microvm.h
 create mode 100644 hw/i386/acpi-microvm.c
hw/i386/acpi-microvm.h new file +6 -0
new file mode 100644
index 000000000000..6a6c2967102b
@@ -0,0 +1,6 @@
+#ifndef HW_I386_ACPI_MICROVM_H
+#define HW_I386_ACPI_MICROVM_H
+
+void acpi_setup_microvm(void);
+
+#endif
include/hw/i386/microvm.h +9 -0
index 2709306c4016..ebd7f49d5de8 100644
@@ -24,12 +24,18 @@
 
 #include "hw/boards.h"
 #include "hw/i386/x86.h"
+#include "hw/acpi/acpi_dev_interface.h"
 
 /* Platform virtio definitions */
 #define VIRTIO_MMIO_BASE      0xc0000000
 #define VIRTIO_NUM_TRANSPORTS 8
 #define VIRTIO_CMDLINE_MAXLEN 64
 
+#define GED_MMIO_BASE         0xc1000000
+#define GED_MMIO_BASE_MEMHP   (GED_MMIO_BASE + 0x100)
+#define GED_MMIO_BASE_REGS    (GED_MMIO_BASE + 0x200)
+#define GED_MMIO_IRQ          9
+
 /* Machine type options */
 #define MICROVM_MACHINE_PIT                 "pit"
 #define MICROVM_MACHINE_PIC                 "pic"
@@ -58,6 +64,9 @@ typedef struct {
     /* Machine state */
     uint32_t virtio_irq_base;
     bool kernel_cmdline_fixed;
+    Notifier machine_done;
+    Notifier powerdown_req;
+    AcpiDeviceIf *acpi_dev;
 } MicrovmMachineState;
 
 #define TYPE_MICROVM_MACHINE   MACHINE_TYPE_NAME("microvm")
hw/i386/acpi-microvm.c new file +0 -0
new file mode 100644
index 000000000000..d2dd024909cc

hw/i386/acpi-microvm.c line 7

Quoted 7 lines
hw/i386/acpi-microvm.c new file +7 -0
@@ -0,0 +1,182 @@
+/* Support for generating ACPI tables and passing them to Guests
+ *
+ * Copyright (C) 2008-2010  Kevin O'Connor <[email protected]>
+ * Copyright (C) 2006 Fabrice Bellard
+ * Copyright (C) 2013 Red Hat Inc
+ *
+ * Author: Michael S. Tsirkin <[email protected]>

since you wrote it s/^^^/Gerd/

Quoted 201 lines
hw/i386/acpi-microvm.c new file +175 -0
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "qemu/osdep.h"
+#include "qapi/error.h"
+
+#include "exec/memory.h"
+#include "hw/acpi/acpi.h"
+#include "hw/acpi/aml-build.h"
+#include "hw/acpi/bios-linker-loader.h"
+#include "hw/acpi/generic_event_device.h"
+#include "hw/acpi/utils.h"
+#include "hw/boards.h"
+#include "hw/i386/fw_cfg.h"
+#include "hw/i386/microvm.h"
+
+#include "acpi-common.h"
+#include "acpi-microvm.h"
+
+static void
+build_dsdt_microvm(GArray *table_data, BIOSLinker *linker,
+                   MicrovmMachineState *mms)
+{
+    X86MachineState *x86ms = X86_MACHINE(mms);
+    Aml *dsdt, *sb_scope, *scope, *pkg;
+    bool ambiguous;
+    Object *isabus;
+
+    isabus = object_resolve_path_type("", TYPE_ISA_BUS, &ambiguous);
+    assert(isabus);
+    assert(!ambiguous);
+
+    dsdt = init_aml_allocator();
+
+    /* Reserve space for header */
+    acpi_data_push(dsdt->buf, sizeof(AcpiTableHeader));
+
+    sb_scope = aml_scope("_SB");
+    fw_cfg_add_acpi_dsdt(sb_scope, x86ms->fw_cfg);
+    isa_build_aml(ISA_BUS(isabus), sb_scope);
+    build_ged_aml(sb_scope, GED_DEVICE, HOTPLUG_HANDLER(mms->acpi_dev),
+                  GED_MMIO_IRQ, AML_SYSTEM_MEMORY, GED_MMIO_BASE);
+    acpi_dsdt_add_power_button(sb_scope);
+    aml_append(dsdt, sb_scope);
+
+    scope = aml_scope("\\");
+    pkg = aml_package(4);
+    aml_append(pkg, aml_int(5)); /* SLEEP_CONTROL_REG.SLP_TYP */
+    aml_append(pkg, aml_int(0)); /* ignored */
+    aml_append(pkg, aml_int(0)); /* reserved */
+    aml_append(pkg, aml_int(0)); /* reserved */
+    aml_append(scope, aml_name_decl("_S5", pkg));
+    aml_append(dsdt, scope);
+
+    /* copy AML table into ACPI tables blob and patch header there */
+    g_array_append_vals(table_data, dsdt->buf->data, dsdt->buf->len);
+    build_header(linker, table_data,
+        (void *)(table_data->data + table_data->len - dsdt->buf->len),
+        "DSDT", dsdt->buf->len, 5, NULL, NULL);
+    free_aml_allocator();
+}
+
+static void acpi_build_microvm(AcpiBuildTables *tables,
+                               MicrovmMachineState *mms)
+{
+    MachineState *machine = MACHINE(mms);
+    GArray *table_offsets;
+    GArray *tables_blob = tables->table_data;
+    unsigned dsdt, xsdt;
+    AcpiFadtData pmfadt = {
+        .rev = 5,
+        .minor_ver = 1,
+        .flags = ((1 << ACPI_FADT_F_HW_REDUCED_ACPI) |
+                  (1 << ACPI_FADT_F_RESET_REG_SUP)),
+        .sleep_ctl = {
+            .space_id = AML_AS_SYSTEM_MEMORY,
+            .bit_width = 8,
+            .address = GED_MMIO_BASE_REGS + ACPI_GED_REG_SLEEP_CTL,
+        },
+        .sleep_sts = {
+            .space_id = AML_AS_SYSTEM_MEMORY,
+            .bit_width = 8,
+            .address = GED_MMIO_BASE_REGS + ACPI_GED_REG_SLEEP_STS,
+        },
+        .reset_reg = {
+            .space_id = AML_AS_SYSTEM_MEMORY,
+            .bit_width = 8,
+            .address = GED_MMIO_BASE_REGS + ACPI_GED_REG_RESET,
+        },
+        .reset_val = ACPI_GED_RESET_VALUE,
+    };
+
+    table_offsets = g_array_new(false, true /* clear */,
+                                        sizeof(uint32_t));
+    bios_linker_loader_alloc(tables->linker,
+                             ACPI_BUILD_TABLE_FILE, tables_blob,
+                             64 /* Ensure FACS is aligned */,
+                             false /* high memory */);
+
+    dsdt = tables_blob->len;
+    build_dsdt_microvm(tables_blob, tables->linker, mms);
+
+    pmfadt.dsdt_tbl_offset = &dsdt;
+    pmfadt.xdsdt_tbl_offset = &dsdt;
+    acpi_add_table(table_offsets, tables_blob);
+    build_fadt(tables_blob, tables->linker, &pmfadt, NULL, NULL);
+
+    acpi_add_table(table_offsets, tables_blob);
+    acpi_build_madt(tables_blob, tables->linker, X86_MACHINE(machine),
+                    mms->acpi_dev, false);
+
+    xsdt = tables_blob->len;
+    build_xsdt(tables_blob, tables->linker, table_offsets, NULL, NULL);
+
+    /* RSDP is in FSEG memory, so allocate it separately */
+    {
+        AcpiRsdpData rsdp_data = {
+            .revision = 2,
+            .oem_id = ACPI_BUILD_APPNAME6,
+            .xsdt_tbl_offset = &xsdt,
+            .rsdt_tbl_offset = NULL,
+        };
+        build_rsdp(tables->rsdp, tables->linker, &rsdp_data);
+    }
+
+    /* Cleanup memory that's no longer used. */
+    g_array_free(table_offsets, true);
+}
+
+static void acpi_build_no_update(void *build_opaque)
+{
+    /* nothing, microvm tables don't change at runtime */
+}
+
+void acpi_setup_microvm(void)
+{
+    MicrovmMachineState *mms = MICROVM_MACHINE(qdev_get_machine());
+    X86MachineState *x86ms = X86_MACHINE(mms);
+    AcpiBuildTables tables;
+
+    assert(x86ms->fw_cfg);
+
+    if (!x86_machine_is_acpi_enabled(x86ms)) {
+        return;
+    }
+
+    acpi_build_tables_init(&tables);
+    acpi_build_microvm(&tables, mms);
+
+    /* Now expose it all to Guest */
+    acpi_add_rom_blob(acpi_build_no_update, NULL,
+                      tables.table_data,
+                      ACPI_BUILD_TABLE_FILE,
+                      ACPI_BUILD_TABLE_MAX_SIZE);
+    acpi_add_rom_blob(acpi_build_no_update, NULL,
+                      tables.linker->cmd_blob,
+                      "etc/table-loader", 0);
+    acpi_add_rom_blob(acpi_build_no_update, NULL,
+                      tables.rsdp,
+                      ACPI_BUILD_RSDP_FILE, 0);
+
+    acpi_build_tables_cleanup(&tables, false);
+}
hw/i386/microvm.c +4 -0
index ea4460714a99..b38321112baa 100644
@@ -26,6 +26,8 @@
 #include "sysemu/cpus.h"
 #include "sysemu/numa.h"
 #include "sysemu/reset.h"
+#include "sysemu/runstate.h"
+#include "acpi-microvm.h"
 
 #include "hw/loader.h"
 #include "hw/irq.h"
@@ -41,6 +43,8 @@
 #include "hw/i386/e820_memory_layout.h"
 #include "hw/i386/fw_cfg.h"
 #include "hw/virtio/virtio-mmio.h"
+#include "hw/acpi/acpi.h"
+#include "hw/acpi/generic_event_device.h"
 
 #include "cpu.h"
 #include "elf.h"
@@ -129,6 +133,17 @@ static void microvm_devices_init(MicrovmMachineState *mms)
     }
 
     /* Optional and legacy devices */

hw/i386/microvm.c static void microvm_devices_init(MicrovmMachineState *mms) · line 145

Quoted 10 lines
hw/i386/microvm.c +10 -0
+    if (x86_machine_is_acpi_enabled(x86ms)) {
+        DeviceState *dev = qdev_create(NULL, TYPE_ACPI_GED_X86);
+        qdev_prop_set_uint32(dev, "ged-event", ACPI_GED_PWR_DOWN_EVT);
+        sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, GED_MMIO_BASE);
+        /* sysbus_mmio_map(SYS_BUS_DEVICE(dev), 1, GED_MMIO_BASE_MEMHP); */
+        sysbus_mmio_map(SYS_BUS_DEVICE(dev), 2, GED_MMIO_BASE_REGS);
+        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0,
+                           x86ms->gsi[GED_MMIO_IRQ]);
+        qdev_init_nofail(dev);
+        mms->acpi_dev = ACPI_DEVICE_IF(dev);

considering all x86 machines have it, maybe move acpi_dev field to parent type?

hw/i386/microvm.c static void microvm_machine_set_auto_kernel_cmdline(Object *obj, bool value, · line 477

Quoted 11 lines
hw/i386/microvm.c +4 -0
+    }
 
     if (mms->pic == ON_OFF_AUTO_ON || mms->pic == ON_OFF_AUTO_AUTO) {
         qemu_irq *i8259;
@@ -457,6 +472,23 @@ static void microvm_machine_set_auto_kernel_cmdline(Object *obj, bool value,
     mms->auto_kernel_cmdline = value;
 }
 
+static void microvm_machine_done(Notifier *notifier, void *data)
+{
+    acpi_setup_microvm();

pass MicrovmMachineState down to acpi_setup_microvm() the same way you are doing below? (and avoid qdev_get_machine() in acpi_setup_microvm())

Quoted 50 lines
hw/i386/microvm.c +19 -0
+}
+
+static void microvm_powerdown_req(Notifier *notifier, void *data)
+{
+    MicrovmMachineState *mms = container_of(notifier, MicrovmMachineState,
+                                            powerdown_req);
+
+    if (mms->acpi_dev) {
+        Object *obj = OBJECT(mms->acpi_dev);
+        AcpiDeviceIfClass *adevc = ACPI_DEVICE_IF_GET_CLASS(obj);
+        adevc->send_event(mms->acpi_dev, ACPI_POWER_DOWN_STATUS);
+    }
+}
+
 static void microvm_machine_initfn(Object *obj)
 {
     MicrovmMachineState *mms = MICROVM_MACHINE(obj);
@@ -471,6 +503,11 @@ static void microvm_machine_initfn(Object *obj)
 
     /* State */
     mms->kernel_cmdline_fixed = false;
+
+    mms->machine_done.notify = microvm_machine_done;
+    qemu_add_machine_init_done_notifier(&mms->machine_done);
+    mms->powerdown_req.notify = microvm_powerdown_req;
+    qemu_register_powerdown_notifier(&mms->powerdown_req);
 }
 
 static void microvm_class_init(ObjectClass *oc, void *data)
hw/i386/Kconfig +1 -0
index c93f32f6579d..be746bcb49eb 100644
@@ -102,6 +102,7 @@ config MICROVM
     select I8259
     select MC146818RTC
     select VIRTIO_MMIO
+    select ACPI_HW_REDUCED
 
 config X86_IOMMU
     bool
hw/i386/Makefile.objs +1 -0
index 622739305882..bbb2fe78f3cd 100644
@@ -19,3 +19,4 @@ obj-y += kvmvapic.o
 obj-$(CONFIG_ACPI) += acpi-common.o
 obj-$(CONFIG_ACPI_HW_REDUCED) += generic_event_device_x86.o
 obj-$(CONFIG_PC) += acpi-build.o
+obj-$(CONFIG_MICROVM) += acpi-microvm.o
#45 lore raw

PATCH v3 17/22 microvm/acpi: add acpi_dsdt_add_virtio() for x86

In reply to Gerd Hoffmann

Makes x86 linux kernel find virtio-mmio devices automatically.

---
 hw/i386/acpi-microvm.c | 52 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 52 insertions(+)
hw/i386/acpi-microvm.c +52 -0
index d2dd024909cc..12060939ff6c 100644
@@ -21,6 +21,7 @@
  */
 
 #include "qemu/osdep.h"
+#include "qemu/cutils.h"
 #include "qapi/error.h"
 
 #include "exec/memory.h"
@@ -32,10 +33,60 @@
 #include "hw/boards.h"
 #include "hw/i386/fw_cfg.h"
 #include "hw/i386/microvm.h"
+#include "hw/virtio/virtio-mmio.h"
 
 #include "acpi-common.h"
 #include "acpi-microvm.h"
 
+static void acpi_dsdt_add_virtio(Aml *scope,
+                                 MicrovmMachineState *mms)
+{
+    gchar *separator;
+    long int index;
+    BusState *bus;
+    BusChild *kid;
+
+    bus = sysbus_get_default();
+    QTAILQ_FOREACH(kid, &bus->children, sibling) {
+        DeviceState *dev = kid->child;
+        Object *obj = object_dynamic_cast(OBJECT(dev), TYPE_VIRTIO_MMIO);
+
+        if (obj) {
+            VirtIOMMIOProxy *mmio = VIRTIO_MMIO(obj);
+            VirtioBusState *mmio_virtio_bus = &mmio->bus;
+            BusState *mmio_bus = &mmio_virtio_bus->parent_obj;
+
+            if (QTAILQ_EMPTY(&mmio_bus->children)) {
+                continue;
+            }
+            separator = g_strrstr(mmio_bus->name, ".");
+            if (!separator) {
+                continue;
+            }
+            if (qemu_strtol(separator + 1, NULL, 10, &index) != 0) {
+                continue;
+            }
+
+            uint32_t irq = mms->virtio_irq_base + index;
+            hwaddr base = VIRTIO_MMIO_BASE + index * 512;
+            hwaddr size = 512;
+
+            Aml *dev = aml_device("VR%02u", (unsigned)index);
+            aml_append(dev, aml_name_decl("_HID", aml_string("LNRO0005")));
+            aml_append(dev, aml_name_decl("_UID", aml_int(index)));
+            aml_append(dev, aml_name_decl("_CCA", aml_int(1)));
+
+            Aml *crs = aml_resource_template();
+            aml_append(crs, aml_memory32_fixed(base, size, AML_READ_WRITE));
+            aml_append(crs,
+                       aml_interrupt(AML_CONSUMER, AML_LEVEL, AML_ACTIVE_HIGH,
+                                     AML_EXCLUSIVE, &irq, 1));
+            aml_append(dev, aml_name_decl("_CRS", crs));
+            aml_append(scope, dev);
+        }
+    }
+}
+
 static void
 build_dsdt_microvm(GArray *table_data, BIOSLinker *linker,
                    MicrovmMachineState *mms)
@@ -60,6 +111,7 @@ build_dsdt_microvm(GArray *table_data, BIOSLinker *linker,
     build_ged_aml(sb_scope, GED_DEVICE, HOTPLUG_HANDLER(mms->acpi_dev),
                   GED_MMIO_IRQ, AML_SYSTEM_MEMORY, GED_MMIO_BASE);
     acpi_dsdt_add_power_button(sb_scope);
+    acpi_dsdt_add_virtio(sb_scope, mms);
     aml_append(dsdt, sb_scope);
 
     scope = aml_scope("\\");
Signature · 1 line
2.18.4
Reviewed-by on v3 17/22
#46 lore raw

Re: [PATCH v3 17/22] microvm/acpi: add acpi_dsdt_add_virtio() for x86

Quoted 4 lines · On Wed, 20 May 2020 15:19:58 +0200 Gerd Hoffmann <[email protected]> wrote:

Makes x86 linux kernel find virtio-mmio devices automatically.

Quoted 85 lines
---
 hw/i386/acpi-microvm.c | 52 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 52 insertions(+)
hw/i386/acpi-microvm.c +52 -0
index d2dd024909cc..12060939ff6c 100644
@@ -21,6 +21,7 @@
  */
 
 #include "qemu/osdep.h"
+#include "qemu/cutils.h"
 #include "qapi/error.h"
 
 #include "exec/memory.h"
@@ -32,10 +33,60 @@
 #include "hw/boards.h"
 #include "hw/i386/fw_cfg.h"
 #include "hw/i386/microvm.h"
+#include "hw/virtio/virtio-mmio.h"
 
 #include "acpi-common.h"
 #include "acpi-microvm.h"
 
+static void acpi_dsdt_add_virtio(Aml *scope,
+                                 MicrovmMachineState *mms)
+{
+    gchar *separator;
+    long int index;
+    BusState *bus;
+    BusChild *kid;
+
+    bus = sysbus_get_default();
+    QTAILQ_FOREACH(kid, &bus->children, sibling) {
+        DeviceState *dev = kid->child;
+        Object *obj = object_dynamic_cast(OBJECT(dev), TYPE_VIRTIO_MMIO);
+
+        if (obj) {
+            VirtIOMMIOProxy *mmio = VIRTIO_MMIO(obj);
+            VirtioBusState *mmio_virtio_bus = &mmio->bus;
+            BusState *mmio_bus = &mmio_virtio_bus->parent_obj;
+
+            if (QTAILQ_EMPTY(&mmio_bus->children)) {
+                continue;
+            }
+            separator = g_strrstr(mmio_bus->name, ".");
+            if (!separator) {
+                continue;
+            }
+            if (qemu_strtol(separator + 1, NULL, 10, &index) != 0) {
+                continue;
+            }
+
+            uint32_t irq = mms->virtio_irq_base + index;
+            hwaddr base = VIRTIO_MMIO_BASE + index * 512;
+            hwaddr size = 512;
+
+            Aml *dev = aml_device("VR%02u", (unsigned)index);
+            aml_append(dev, aml_name_decl("_HID", aml_string("LNRO0005")));
+            aml_append(dev, aml_name_decl("_UID", aml_int(index)));
+            aml_append(dev, aml_name_decl("_CCA", aml_int(1)));
+
+            Aml *crs = aml_resource_template();
+            aml_append(crs, aml_memory32_fixed(base, size, AML_READ_WRITE));
+            aml_append(crs,
+                       aml_interrupt(AML_CONSUMER, AML_LEVEL, AML_ACTIVE_HIGH,
+                                     AML_EXCLUSIVE, &irq, 1));
+            aml_append(dev, aml_name_decl("_CRS", crs));
+            aml_append(scope, dev);
+        }
+    }
+}
+
 static void
 build_dsdt_microvm(GArray *table_data, BIOSLinker *linker,
                    MicrovmMachineState *mms)
@@ -60,6 +111,7 @@ build_dsdt_microvm(GArray *table_data, BIOSLinker *linker,
     build_ged_aml(sb_scope, GED_DEVICE, HOTPLUG_HANDLER(mms->acpi_dev),
                   GED_MMIO_IRQ, AML_SYSTEM_MEMORY, GED_MMIO_BASE);
     acpi_dsdt_add_power_button(sb_scope);
+    acpi_dsdt_add_virtio(sb_scope, mms);
     aml_append(dsdt, sb_scope);
 
     scope = aml_scope("\\");
#47 lore raw

PATCH v3 18/22 microvm/acpi: use GSI 16-23 for virtio

In reply to Gerd Hoffmann

With ACPI enabled and IO-APIC being properly declared in the ACPI tables we can use interrupt lines 16-23 for virtio and avoid shared interrupts.

With acpi disabled we continue to use lines 8-15.

---
 hw/i386/microvm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
hw/i386/microvm.c +1 -1
index b38321112baa..7d8174809ae1 100644
@@ -125,7 +125,7 @@ static void microvm_devices_init(MicrovmMachineState *mms)
 
     kvmclock_create();
 
-    mms->virtio_irq_base = 8;
+    mms->virtio_irq_base = x86_machine_is_acpi_enabled(x86ms) ? 16 : 8;
     for (i = 0; i < VIRTIO_NUM_TRANSPORTS; i++) {
         sysbus_create_simple("virtio-mmio",
                              VIRTIO_MMIO_BASE + i * 512,
Signature · 1 line
2.18.4
Reviewed-by on v3 18/22
#48 lore raw

Re: [PATCH v3 18/22] microvm/acpi: use GSI 16-23 for virtio

Quoted 7 lines · On Wed, 20 May 2020 15:19:59 +0200 Gerd Hoffmann <[email protected]> wrote:

With ACPI enabled and IO-APIC being properly declared in the ACPI tables we can use interrupt lines 16-23 for virtio and avoid shared interrupts.

With acpi disabled we continue to use lines 8-15.

Quoted 17 lines
---
 hw/i386/microvm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
hw/i386/microvm.c +1 -1
index b38321112baa..7d8174809ae1 100644
@@ -125,7 +125,7 @@ static void microvm_devices_init(MicrovmMachineState *mms)
 
     kvmclock_create();
 
-    mms->virtio_irq_base = 8;
+    mms->virtio_irq_base = x86_machine_is_acpi_enabled(x86ms) ? 16 : 8;
     for (i = 0; i < VIRTIO_NUM_TRANSPORTS; i++) {
         sysbus_create_simple("virtio-mmio",
                              VIRTIO_MMIO_BASE + i * 512,
#49 lore raw

PATCH v3 19/22 microvm/acpi: use seabios with acpi=on

In reply to Gerd Hoffmann

With acpi=off continue to use qboot.

---
 hw/i386/microvm.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
hw/i386/microvm.c +4 -1
index 7d8174809ae1..ddb8570a2915 100644
@@ -52,6 +52,7 @@
 #include "hw/xen/start_info.h"
 
 #define MICROVM_QBOOT_FILENAME "qboot.bin"
+#define MICROVM_BIOS_FILENAME  "bios-microvm.bin"
 
 static void microvm_set_rtc(MicrovmMachineState *mms, ISADevice *s)
 {
@@ -174,7 +175,9 @@ static void microvm_devices_init(MicrovmMachineState *mms)
     }
 
     if (bios_name == NULL) {
-        bios_name = MICROVM_QBOOT_FILENAME;
+        bios_name = x86_machine_is_acpi_enabled(x86ms)
+            ? MICROVM_BIOS_FILENAME
+            : MICROVM_QBOOT_FILENAME;
     }
     x86_bios_rom_init(get_system_memory(), true);
 }
Signature · 1 line
2.18.4
Reviewed-by on v3 19/22
#50 lore raw

Re: [PATCH v3 19/22] microvm/acpi: use seabios with acpi=on

Quoted 3 lines · On Wed, 20 May 2020 15:20:00 +0200 Gerd Hoffmann <[email protected]> wrote:

With acpi=off continue to use qboot.

Quoted 27 lines
---
 hw/i386/microvm.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
hw/i386/microvm.c +4 -1
index 7d8174809ae1..ddb8570a2915 100644
@@ -52,6 +52,7 @@
 #include "hw/xen/start_info.h"
 
 #define MICROVM_QBOOT_FILENAME "qboot.bin"
+#define MICROVM_BIOS_FILENAME  "bios-microvm.bin"
 
 static void microvm_set_rtc(MicrovmMachineState *mms, ISADevice *s)
 {
@@ -174,7 +175,9 @@ static void microvm_devices_init(MicrovmMachineState *mms)
     }
 
     if (bios_name == NULL) {
-        bios_name = MICROVM_QBOOT_FILENAME;
+        bios_name = x86_machine_is_acpi_enabled(x86ms)
+            ? MICROVM_BIOS_FILENAME
+            : MICROVM_QBOOT_FILENAME;
     }
     x86_bios_rom_init(get_system_memory(), true);
 }
#51 lore raw

PATCH v3 20/22 microvm/acpi: disable virtio-mmio cmdline hack

In reply to Gerd Hoffmann

... in case we are using ACPI.

---
 hw/i386/microvm.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
hw/i386/microvm.c +2 -1
index ddb8570a2915..602c6a8f75f3 100644
@@ -362,7 +362,8 @@ static void microvm_machine_reset(MachineState *machine)
     CPUState *cs;
     X86CPU *cpu;
 
-    if (machine->kernel_filename != NULL &&
+    if (!x86_machine_is_acpi_enabled(X86_MACHINE(machine)) &&
+        machine->kernel_filename != NULL &&
         mms->auto_kernel_cmdline && !mms->kernel_cmdline_fixed) {
         microvm_fix_kernel_cmdline(machine);
         mms->kernel_cmdline_fixed = true;
Signature · 1 line
2.18.4
#52 lore raw

RFC PATCH v3 21/22 acpi: add per machine type acpi default

In reply to Gerd Hoffmann

Allow setting acpi default value for each machine type.

---
 include/hw/i386/x86.h |  1 +
 hw/i386/x86.c         | 21 ++++++++++++++++++---
 2 files changed, 19 insertions(+), 3 deletions(-)
include/hw/i386/x86.h +1 -0
index b52285481687..d2fffa8252ff 100644
@@ -37,6 +37,7 @@ typedef struct {
     bool save_tsc_khz;
     /* Enables contiguous-apic-ID mode */
     bool compat_apic_id_mode;
+    bool acpi_default;
 } X86MachineClass;
 
 typedef struct {
hw/i386/x86.c +18 -3
index ac7a0a958781..3a56a157c5f0 100644
@@ -940,10 +940,22 @@ static void x86_machine_set_smm(Object *obj, Visitor *v, const char *name,
 
 bool x86_machine_is_acpi_enabled(X86MachineState *x86ms)
 {
-    if (x86ms->acpi == ON_OFF_AUTO_OFF) {
-        return false;
+    X86MachineClass *x86mc = X86_MACHINE_GET_CLASS(x86ms);
+    bool enabled;
+
+    switch (x86ms->acpi) {
+    case ON_OFF_AUTO_ON:
+        enabled = true;
+        break;
+    case ON_OFF_AUTO_OFF:
+        enabled = false;
+        break;
+    case ON_OFF_AUTO_AUTO:
+    default:
+        enabled = x86mc->acpi_default;
+        break;
     }
-    return true;
+    return enabled;
 }
 
 static void x86_machine_get_acpi(Object *obj, Visitor *v, const char *name,
@@ -991,6 +1003,9 @@ static void x86_machine_class_init(ObjectClass *oc, void *data)
     x86mc->save_tsc_khz = true;
     nc->nmi_monitor_handler = x86_nmi;
 
+    /* acpi is on by default */
+    x86mc->acpi_default = true;
+
     object_class_property_add(oc, X86_MACHINE_MAX_RAM_BELOW_4G, "size",
         x86_machine_get_max_ram_below_4g, x86_machine_set_max_ram_below_4g,
         NULL, NULL);
Signature · 1 line
2.18.4
#53 lore raw

Re: [PATCH v3 21/22] [RfC] acpi: add per machine type acpi default

Quoted 37 lines · On Wed, 20 May 2020 15:20:02 +0200 Gerd Hoffmann <[email protected]> wrote:

Allow setting acpi default value for each machine type.

---
 include/hw/i386/x86.h |  1 +
 hw/i386/x86.c         | 21 ++++++++++++++++++---
 2 files changed, 19 insertions(+), 3 deletions(-)
include/hw/i386/x86.h +1 -0
index b52285481687..d2fffa8252ff 100644
@@ -37,6 +37,7 @@ typedef struct {
     bool save_tsc_khz;
     /* Enables contiguous-apic-ID mode */
     bool compat_apic_id_mode;
+    bool acpi_default;
 } X86MachineClass;
 
 typedef struct {
hw/i386/x86.c +7 -2
index ac7a0a958781..3a56a157c5f0 100644
@@ -940,10 +940,22 @@ static void x86_machine_set_smm(Object *obj, Visitor *v, const char *name,
 
 bool x86_machine_is_acpi_enabled(X86MachineState *x86ms)
 {
-    if (x86ms->acpi == ON_OFF_AUTO_OFF) {
-        return false;
+    X86MachineClass *x86mc = X86_MACHINE_GET_CLASS(x86ms);
+    bool enabled;
+
+    switch (x86ms->acpi) {
+    case ON_OFF_AUTO_ON:
+        enabled = true;
+        break;

hw/i386/x86.c static void x86_machine_set_smm(Object *obj, Visitor *v, const char *name, · line 958

Quoted 10 lines
hw/i386/x86.c +8 -1
+    case ON_OFF_AUTO_OFF:
+        enabled = false;
+        break;
+    case ON_OFF_AUTO_AUTO:
+    default:
+        enabled = x86mc->acpi_default;
+        break;
     }
-    return true;
+    return enabled;

why not make x86ms->acpi a boolean and move logic from here to x86_machine_[get|set]_acpi() to avoid extra helper, the external users would use property acessesors and internal (microvm code) can just access x86ms->acpi directly

Quoted 13 lines
hw/i386/x86.c +3 -0
 }
 
 static void x86_machine_get_acpi(Object *obj, Visitor *v, const char *name,
@@ -991,6 +1003,9 @@ static void x86_machine_class_init(ObjectClass *oc, void *data)
     x86mc->save_tsc_khz = true;
     nc->nmi_monitor_handler = x86_nmi;
 
+    /* acpi is on by default */
+    x86mc->acpi_default = true;
+
     object_class_property_add(oc, X86_MACHINE_MAX_RAM_BELOW_4G, "size",
         x86_machine_get_max_ram_below_4g, x86_machine_set_max_ram_below_4g,
         NULL, NULL);
#54 lore raw

RFC PATCH v3 22/22 acpi: flip default to off for microvm

In reply to Gerd Hoffmann
---
 hw/i386/microvm.c | 4 ++++
 1 file changed, 4 insertions(+)
hw/i386/microvm.c +4 -0
index 602c6a8f75f3..b5c445b5403b 100644
@@ -517,6 +517,7 @@ static void microvm_machine_initfn(Object *obj)
 static void microvm_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    X86MachineClass *x86mc = X86_MACHINE_CLASS(oc);
 
     mc->init = microvm_machine_state_init;
 
@@ -537,6 +538,9 @@ static void microvm_class_init(ObjectClass *oc, void *data)
     /* Machine class handlers */
     mc->reset = microvm_machine_reset;
 
+    /* acpi is off by default */
+    x86mc->acpi_default = false;
+
     object_class_property_add(oc, MICROVM_MACHINE_PIC, "OnOffAuto",
                               microvm_machine_get_pic,
                               microvm_machine_set_pic,
Signature · 1 line
2.18.4
#55 lore raw

Re: [PATCH v3 22/22] [RfC] acpi: flip default to off for microvm

Quoted 18 lines · On Wed, 20 May 2020 15:20:03 +0200 Gerd Hoffmann <[email protected]> wrote:
---
 hw/i386/microvm.c | 4 ++++
 1 file changed, 4 insertions(+)
hw/i386/microvm.c +1 -0
index 602c6a8f75f3..b5c445b5403b 100644
@@ -517,6 +517,7 @@ static void microvm_machine_initfn(Object *obj)
 static void microvm_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    X86MachineClass *x86mc = X86_MACHINE_CLASS(oc);
 
     mc->init = microvm_machine_state_init;
 

hw/i386/microvm.c static void microvm_class_init(ObjectClass *oc, void *data) · line 542

Quoted 5 lines
hw/i386/microvm.c +2 -0
@@ -537,6 +538,9 @@ static void microvm_class_init(ObjectClass *oc, void *data)
     /* Machine class handlers */
     mc->reset = microvm_machine_reset;
 
+    /* acpi is off by default */
+    x86mc->acpi_default = false;

Since we don't care about compat stuff for microvm yet,
I'd preffer default 'on', as in long run it allows to boot larger range of guests without need for tuning QEMU and guest configuration
(i.e. it's the least troublesome case for ordinary user)

Quoted 4 lines
hw/i386/microvm.c +1 -0
+
     object_class_property_add(oc, MICROVM_MACHINE_PIC, "OnOffAuto",
                               microvm_machine_get_pic,
                               microvm_machine_set_pic,
#56 lore raw

Re: [PATCH v3 22/22] [RfC] acpi: flip default to off for microvm

Quoted 24 lines · Igor Mammedov
Quoted 21 lines · On Wed, 20 May 2020 15:20:03 +0200 Gerd Hoffmann <[email protected]> wrote:
---
 hw/i386/microvm.c | 4 ++++
 1 file changed, 4 insertions(+)
hw/i386/microvm.c +1 -0
index 602c6a8f75f3..b5c445b5403b 100644
@@ -517,6 +517,7 @@ static void microvm_machine_initfn(Object *obj)
 static void microvm_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    X86MachineClass *x86mc = X86_MACHINE_CLASS(oc);
 
     mc->init = microvm_machine_state_init;
 
@@ -537,6 +538,9 @@ static void microvm_class_init(ObjectClass *oc, void *data)
     /* Machine class handlers */
     mc->reset = microvm_machine_reset;
 
Quoted 7 lines

hw/i386/microvm.c static void microvm_class_init(ObjectClass *oc, void *data) · line 542

Quoted 2 lines
hw/i386/microvm.c +2 -0
+    /* acpi is off by default */
+    x86mc->acpi_default = false;

Since we don't care about compat stuff for microvm yet,
I'd preffer default 'on', as in long run it allows to boot larger range of guests without need for tuning QEMU and guest configuration
(i.e. it's the least troublesome case for ordinary user)

I fully agree and I'll happily drop patches 21+22 and have microvm default to acpi=on. See also:
https://lists.gnu.org/archive/html/qemu-devel/2020-05/msg01315.html

Older patch versions got multiple review comments asking for acpi=off by default, so I've added these two patches as RfC to show how this could be done. RfC because (1) the topic is controversial and (2) these are not final. I'd move 21/22 up in the series and squash 22/22 into the "add acpi to microvm" patch in case we decide for acpi=off as default.

take care,
Gerd

#57 lore raw
In reply to Gerd Hoffmann
Patchew URL: https://patchew.org/QEMU/[email protected]/

Patchew URL: https://patchew.org/QEMU/[email protected]/

Hi,

This series failed the docker-quick@centos7 build test. Please find the testing commands and their output below. If you have Docker installed, you can probably reproduce it locally.

=== TEST SCRIPT BEGIN ===
#!/bin/bash
make docker-image-centos7 V=1 NETWORK=1
time make docker-test-quick@centos7 SHOW_ENV=1 J=14 NETWORK=1 === TEST SCRIPT END ===

CC aarch64-softmmu/hw/arm/integratorcp.o
CC x86_64-softmmu/hw/i386/kvm/i8254.o
/tmp/qemu-test/src/hw/i386/acpi-microvm.c: In function 'build_dsdt_microvm':
/tmp/qemu-test/src/hw/i386/acpi-microvm.c:109:5: error: implicit declaration of function 'fw_cfg_add_acpi_dsdt' [-Werror=implicit-function-declaration]

     fw_cfg_add_acpi_dsdt(sb_scope, x86ms->fw_cfg);
     ^

/tmp/qemu-test/src/hw/i386/acpi-microvm.c:109:5: error: nested extern declaration of 'fw_cfg_add_acpi_dsdt' [-Werror=nested-externs]
cc1: all warnings being treated as errors
make[1]: *** [hw/i386/acpi-microvm.o] Error 1
make[1]: *** Waiting for unfinished jobs....
CC aarch64-softmmu/hw/arm/mainstone.o
CC aarch64-softmmu/hw/arm/microbit.o
---

  CC      aarch64-softmmu/target/arm/gdbstub64.o
  CC      aarch64-softmmu/target/arm/machine.o
  CC      aarch64-softmmu/target/arm/arch_dump.o

make: *** [x86_64-softmmu/all] Error 2
make: *** Waiting for unfinished jobs....
CC aarch64-softmmu/target/arm/monitor.o
CC aarch64-softmmu/target/arm/arm-powerctl.o
---
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['sudo', '-n', 'docker', 'run', '--label', 'com.qemu.instance.uuid=5cfa0900fad74d3e96f029386896f751', '-u', '1001', '--security-opt', 'seccomp=unconfined', '--rm', '-e', 'TARGET_LIST=', '-e', 'EXTRA_CONFIGURE_OPTS=', '-e', 'V=', '-e', 'J=14', '-e', 'DEBUG=', '-e', 'SHOW_ENV=1', '-e', 'CCACHE_DIR=/var/tmp/ccache', '-v', '/home/patchew/.cache/qemu-docker-ccache:/var/tmp/ccache:z', '-v', '/var/tmp/patchew-tester-tmp-612gx0r0/src/docker-src.2020-05-20-11.43.56.32095:/var/tmp/qemu:z,ro', 'qemu:centos7', '/var/tmp/qemu/run', 'test-quick']' returned non-zero exit status 2. filter=--filter=label=com.qemu.instance.uuid=5cfa0900fad74d3e96f029386896f751
make[1]: *** [docker-run] Error 1
make[1]: Leaving directory `/var/tmp/patchew-tester-tmp-612gx0r0/src'
make: *** [docker-run-test-quick@centos7] Error 2

real 2m31.578s
user 0m8.620s

The full log is available at
http://patchew.org/logs/[email protected]/testing.docker-quick@centos7/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]

#58 lore raw
In reply to Gerd Hoffmann
Patchew URL: https://patchew.org/QEMU/[email protected]/

Patchew URL: https://patchew.org/QEMU/[email protected]/

Hi,

This series seems to have some coding style problems. See output below for more information:

Message-id: [email protected]
Subject: [PATCH v3 00/22] microvm: add acpi support
Type: series

=== TEST SCRIPT BEGIN ===
#!/bin/bash
git rev-parse base > /dev/null || exit 0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
./scripts/checkpatch.pl --mailback base..
=== TEST SCRIPT END ===

From https://github.com/patchew-project/qemu
* [new tag] patchew/[email protected] -> patchew/[email protected]
Switched to a new branch 'test'
7280d50 acpi: flip default to off for microvm 0c46cb7 acpi: add per machine type acpi default 3932159 microvm/acpi: disable virtio-mmio cmdline hack 88c8990 microvm/acpi: use seabios with acpi=on 00de58b microvm/acpi: use GSI 16-23 for virtio f442b20 microvm/acpi: add acpi_dsdt_add_virtio() for x86 50bd765 microvm/acpi: add minimal acpi support b44d3ee2 microvm: make virtio irq base runtime configurable 8473717 microvm: use 2G split unconditionally 9a8690b x86: coldplug cpus
873496e acpi: move acpi_dsdt_add_power_button() to ged 570b71e acpi: ged: add x86 device variant.
42157e1 acpi: ged: add control regs
963229e acpi: ged: rename event memory region ffcb835 acpi: fadt: add hw-reduced sleep register support 628859e acpi: madt: skip pci override on pci-less systems. d7248a7 acpi: create acpi-common.c and move madt code 33254f5 acpi: make build_madt() more generic. 447375a seabios: update binaries to master snapshot f474461 seabios: update config & build rules 583a45c seabios: update submodule to master snapshot ce0b4a0 microvm: name qboot binary qboot.bin

=== OUTPUT BEGIN ===
1/22 Checking commit ce0b4a0ec4ce (microvm: name qboot binary qboot.bin)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? #36:
rename from pc-bios/bios-microvm.bin

total: 0 errors, 1 warnings, 35 lines checked

Patch 1/22 has style problems, please review. If any of these errors are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
2/22 Checking commit 583a45cbd0ff (seabios: update submodule to master snapshot) 3/22 Checking commit f474461987ec (seabios: update config & build rules)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? #45:
new file mode 100644

total: 0 errors, 1 warnings, 50 lines checked

Patch 3/22 has style problems, please review. If any of these errors are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
4/22 Checking commit 447375a74a06 (seabios: update binaries to master snapshot) 5/22 Checking commit 33254f57d62c (acpi: make build_madt() more generic.) 6/22 Checking commit d7248a7438fe (acpi: create acpi-common.c and move madt code)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? #182:
new file mode 100644

WARNING: Block comments use a leading /* on a separate line #187: FILE: hw/i386/acpi-common.c:1:
+/* Support for generating ACPI tables and passing them to Guests

WARNING: Block comments use a leading /* on a separate line #227: FILE: hw/i386/acpi-common.c:41:
+ /* ACPI spec says that LAPIC entry for non present

ERROR: spaces required around that '<<' (ctx:VxV)
#301: FILE: hw/i386/acpi-common.c:115:
+#define ACPI_BUILD_PCI_IRQS ((1<<5) | (1<<9) | (1<<10) | (1<<11))

                                ^

ERROR: spaces required around that '<<' (ctx:VxV)
#301: FILE: hw/i386/acpi-common.c:115:
+#define ACPI_BUILD_PCI_IRQS ((1<<5) | (1<<9) | (1<<10) | (1<<11))

                                         ^

ERROR: spaces required around that '<<' (ctx:VxV)
#301: FILE: hw/i386/acpi-common.c:115:
+#define ACPI_BUILD_PCI_IRQS ((1<<5) | (1<<9) | (1<<10) | (1<<11))

                                                  ^

ERROR: spaces required around that '<<' (ctx:VxV)
#301: FILE: hw/i386/acpi-common.c:115:
+#define ACPI_BUILD_PCI_IRQS ((1<<5) | (1<<9) | (1<<10) | (1<<11))

                                                            ^

total: 4 errors, 3 warnings, 321 lines checked

Patch 6/22 has style problems, please review. If any of these errors are false positives report them to the maintainer, see CHECKPATCH in MAINTAINERS.

7/22 Checking commit 628859e5cd4e (acpi: madt: skip pci override on pci-less systems.)
WARNING: line over 80 characters
#62: FILE: hw/i386/acpi-common.c:128:
+ intsrcovr->flags = cpu_to_le16(0xd); /* active high, level triggered */

total: 0 errors, 1 warnings, 56 lines checked

Patch 7/22 has style problems, please review. If any of these errors are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
8/22 Checking commit ffcb8350661b (acpi: fadt: add hw-reduced sleep register support) 9/22 Checking commit 963229ef8a9a (acpi: ged: rename event memory region) 10/22 Checking commit 42157e12a08c (acpi: ged: add control regs)
11/22 Checking commit 570b71e5b8fa (acpi: ged: add x86 device variant.)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? #24:
new file mode 100644

total: 0 errors, 1 warnings, 51 lines checked

Patch 11/22 has style problems, please review. If any of these errors are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
12/22 Checking commit 873496e2d411 (acpi: move acpi_dsdt_add_power_button() to ged) 13/22 Checking commit 9a8690b252cc (x86: coldplug cpus)
14/22 Checking commit 8473717fafed (microvm: use 2G split unconditionally) 15/22 Checking commit b44d3ee2c4d2 (microvm: make virtio irq base runtime configurable) 16/22 Checking commit 50bd7654bd78 (microvm/acpi: add minimal acpi support)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? #34:
new file mode 100644

WARNING: Block comments use a leading /* on a separate line #39: FILE: hw/i386/acpi-microvm.c:1:
+/* Support for generating ACPI tables and passing them to Guests

WARNING: Block comments use a leading /* on a separate line #150: FILE: hw/i386/acpi-microvm.c:112:
+ table_offsets = g_array_new(false, true /* clear */,

WARNING: Block comments use a leading /* on a separate line #154: FILE: hw/i386/acpi-microvm.c:116:
+ 64 /* Ensure FACS is aligned */,

WARNING: Block comments use a leading /* on a separate line #155: FILE: hw/i386/acpi-microvm.c:117:
+ false /* high memory */);

total: 0 errors, 5 warnings, 293 lines checked

Patch 16/22 has style problems, please review. If any of these errors are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
17/22 Checking commit f442b20044f3 (microvm/acpi: add acpi_dsdt_add_virtio() for x86) 18/22 Checking commit 00de58be2c9d (microvm/acpi: use GSI 16-23 for virtio) 19/22 Checking commit 88c8990ffa2f (microvm/acpi: use seabios with acpi=on) 20/22 Checking commit 3932159d5e0c (microvm/acpi: disable virtio-mmio cmdline hack) 21/22 Checking commit 0c46cb7b51d7 (acpi: add per machine type acpi default) 22/22 Checking commit 7280d5050ee8 (acpi: flip default to off for microvm) === OUTPUT END ===

Test command exited with code: 1

The full log is available at
http://patchew.org/logs/[email protected]/testing.checkpatch/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]

#59 lore raw
In reply to Gerd Hoffmann
Patchew URL: https://patchew.org/QEMU/[email protected]/

Patchew URL: https://patchew.org/QEMU/[email protected]/

Hi,

This series failed the asan build test. Please find the testing commands and their output below. If you have Docker installed, you can probably reproduce it locally.

=== TEST SCRIPT BEGIN ===
#!/bin/bash
export ARCH=x86_64
make docker-image-fedora V=1 NETWORK=1
time make docker-test-debug@fedora TARGET_LIST=x86_64-softmmu J=14 NETWORK=1 === TEST SCRIPT END ===

  CC      x86_64-softmmu/hw/i386/xen/xen_pvdevice.o
  CC      x86_64-softmmu/hw/i386/xen/xen-hvm.o
  CC      x86_64-softmmu/hw/i386/xen/xen-mapcache.o
/tmp/qemu-test/src/hw/i386/acpi-microvm.c:109:5: error: implicit declaration of function 'fw_cfg_add_acpi_dsdt' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
    fw_cfg_add_acpi_dsdt(sb_scope, x86ms->fw_cfg);
    ^

/tmp/qemu-test/src/hw/i386/acpi-microvm.c:109:5: error: this function declaration is not a prototype [-Werror,-Wstrict-prototypes] 2 errors generated.
make[1]: *** [/tmp/qemu-test/src/rules.mak:69: hw/i386/acpi-microvm.o] Error 1
make[1]: *** Waiting for unfinished jobs....
make: *** [Makefile:527: x86_64-softmmu/all] Error 2
Traceback (most recent call last):
File "./tests/docker/docker.py", line 664, in <module>

    sys.exit(main())

---
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['sudo', '-n', 'docker', 'run', '--label', 'com.qemu.instance.uuid=3620f0cf6349445fb42a098ed6ef40c8', '-u', '1001', '--security-opt', 'seccomp=unconfined', '--rm', '-e', 'TARGET_LIST=x86_64-softmmu', '-e', 'EXTRA_CONFIGURE_OPTS=', '-e', 'V=', '-e', 'J=14', '-e', 'DEBUG=', '-e', 'SHOW_ENV=', '-e', 'CCACHE_DIR=/var/tmp/ccache', '-v', '/home/patchew/.cache/qemu-docker-ccache:/var/tmp/ccache:z', '-v', '/var/tmp/patchew-tester-tmp-lc0gao83/src/docker-src.2020-05-20-11.47.05.9825:/var/tmp/qemu:z,ro', 'qemu:fedora', '/var/tmp/qemu/run', 'test-debug']' returned non-zero exit status 2. filter=--filter=label=com.qemu.instance.uuid=3620f0cf6349445fb42a098ed6ef40c8
make[1]: *** [docker-run] Error 1
make[1]: Leaving directory `/var/tmp/patchew-tester-tmp-lc0gao83/src'
make: *** [docker-run-test-debug@fedora] Error 2

real 3m56.264s
user 0m8.551s

The full log is available at
http://patchew.org/logs/[email protected]/testing.asan/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]

#60 lore raw
In reply to Gerd Hoffmann
Patchew URL: https://patchew.org/QEMU/[email protected]/

Patchew URL: https://patchew.org/QEMU/[email protected]/

Hi,

This series failed the docker-mingw@fedora build test. Please find the testing commands and their output below. If you have Docker installed, you can probably reproduce it locally.

=== TEST SCRIPT BEGIN ===
#! /bin/bash
export ARCH=x86_64
make docker-image-fedora V=1 NETWORK=1
time make docker-test-mingw@fedora J=14 NETWORK=1
=== TEST SCRIPT END ===

CC x86_64-softmmu/target/i386/mpx_helper.o
CC x86_64-softmmu/target/i386/seg_helper.o
/tmp/qemu-test/src/hw/i386/acpi-microvm.c: In function 'build_dsdt_microvm':
/tmp/qemu-test/src/hw/i386/acpi-microvm.c:109:5: error: implicit declaration of function 'fw_cfg_add_acpi_dsdt'; did you mean 'fw_cfg_add_file'? [-Werror=implicit-function-declaration]

     fw_cfg_add_acpi_dsdt(sb_scope, x86ms->fw_cfg);
     ^~~~~~~~~~~~~~~~~~~~
     fw_cfg_add_file

/tmp/qemu-test/src/hw/i386/acpi-microvm.c:109:5: error: nested extern declaration of 'fw_cfg_add_acpi_dsdt' [-Werror=nested-externs]
cc1: all warnings being treated as errors
CC x86_64-softmmu/target/i386/smm_helper.o
make[1]: *** [/tmp/qemu-test/src/rules.mak:69: hw/i386/acpi-microvm.o] Error 1
make[1]: *** Waiting for unfinished jobs....
CC aarch64-softmmu/hw/arm/digic_boards.o
CC aarch64-softmmu/hw/arm/exynos4_boards.o
---

  CC      aarch64-softmmu/hw/arm/vexpress.o
  CC      aarch64-softmmu/hw/arm/xilinx_zynq.o
  CC      aarch64-softmmu/hw/arm/sabrelite.o

make: *** [Makefile:527: x86_64-softmmu/all] Error 2
make: *** Waiting for unfinished jobs....
CC aarch64-softmmu/hw/arm/armv7m.o
CC aarch64-softmmu/hw/arm/exynos4210.o
---
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['sudo', '-n', 'docker', 'run', '--label', 'com.qemu.instance.uuid=87e79745f2ae42ae974eb9b247d4af76', '-u', '1003', '--security-opt', 'seccomp=unconfined', '--rm', '-e', 'TARGET_LIST=', '-e', 'EXTRA_CONFIGURE_OPTS=', '-e', 'V=', '-e', 'J=14', '-e', 'DEBUG=', '-e', 'SHOW_ENV=', '-e', 'CCACHE_DIR=/var/tmp/ccache', '-v', '/home/patchew2/.cache/qemu-docker-ccache:/var/tmp/ccache:z', '-v', '/var/tmp/patchew-tester-tmp-43fhva9t/src/docker-src.2020-05-20-11.50.45.16713:/var/tmp/qemu:z,ro', 'qemu:fedora', '/var/tmp/qemu/run', 'test-mingw']' returned non-zero exit status 2. filter=--filter=label=com.qemu.instance.uuid=87e79745f2ae42ae974eb9b247d4af76
make[1]: *** [docker-run] Error 1
make[1]: Leaving directory `/var/tmp/patchew-tester-tmp-43fhva9t/src'
make: *** [docker-run-test-mingw@fedora] Error 2

real 3m1.622s
user 0m7.423s

The full log is available at
http://patchew.org/logs/[email protected]/testing.docker-mingw@fedora/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]