[PULL v1 00/12] Xen queue

[PULL v1 00/12] Xen queue

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

The following changes since commit e638d685ec2a0700fb9529cbd1b2823ac4120c53:

  Open 9.2 development tree (2024-09-03 09:18:43 -0700)

are available in the Git repository at:

  https://gitlab.com/edgar.iglesias/qemu.git tags/edgar/xen-queue-2024-09-04.for-upstream

for you to fetch changes up to 0b57c8160a2a6c833cfb1d958f08385c4391ab70:

  docs/system/i386: xenpvh: Add a basic description (2024-09-04 16:50:43 +0200)

----------------------------------------------------------------
Edgars Xen queue.

----------------------------------------------------------------
Edgar E. Iglesias (12):
      MAINTAINERS: Add docs/system/arm/xenpvh.rst
      hw/arm: xenpvh: Update file header to use SPDX
      hw/arm: xenpvh: Tweak machine description
      hw/arm: xenpvh: Add support for SMP guests
      hw/arm: xenpvh: Remove double-negation in warning
      hw/arm: xenpvh: Move stubbed functions to xen-stubs.c
      hw/arm: xenpvh: Break out a common PVH machine
      hw/arm: xenpvh: Rename xen_arm.c -> xen-pvh.c
      hw/arm: xenpvh: Reverse virtio-mmio creation order
      hw/xen: pvh-common: Add support for creating PCIe/GPEX
      hw/i386/xen: Add a Xen PVH x86 machine
      docs/system/i386: xenpvh: Add a basic description

 MAINTAINERS                     |   2 +
 docs/system/i386/xenpvh.rst     |  49 ++++++
 docs/system/target-i386.rst     |   1 +
 hw/arm/meson.build              |   5 +-
 hw/arm/trace-events             |   5 -
 hw/arm/xen-pvh.c                |  89 ++++++++++
 hw/arm/xen-stubs.c              |  32 ++++
 hw/arm/xen_arm.c                | 267 -----------------------------
 hw/i386/xen/meson.build         |   1 +
 hw/i386/xen/xen-pvh.c           | 121 ++++++++++++++
 hw/xen/meson.build              |   1 +
 hw/xen/trace-events             |   4 +
 hw/xen/xen-pvh-common.c         | 362 ++++++++++++++++++++++++++++++++++++++++
 include/hw/xen/xen-pvh-common.h |  88 ++++++++++
 14 files changed, 754 insertions(+), 273 deletions(-)
 create mode 100644 docs/system/i386/xenpvh.rst
 create mode 100644 hw/arm/xen-pvh.c
 create mode 100644 hw/arm/xen-stubs.c
 delete mode 100644 hw/arm/xen_arm.c
 create mode 100644 hw/i386/xen/xen-pvh.c
 create mode 100644 hw/xen/xen-pvh-common.c
 create mode 100644 include/hw/xen/xen-pvh-common.h

-- 
2.43.0

[PULL v1 01/12] MAINTAINERS: Add docs/system/arm/xenpvh.rst

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Acked-by: Stefano Stabellini <sstabellini@kernel.org>
---
 MAINTAINERS | 1 +
 1 file changed, 1 insertion(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index 3584d6a6c6..c2fb0c2f42 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -559,6 +559,7 @@ F: include/hw/xen/
 F: include/sysemu/xen.h
 F: include/sysemu/xen-mapcache.h
 F: stubs/xen-hw-stub.c
+F: docs/system/arm/xenpvh.rst
 
 Guest CPU Cores (NVMM)
 ----------------------
-- 
2.43.0

[PULL v1 02/12] hw/arm: xenpvh: Update file header to use SPDX

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Update file header to use SPDX and remove stray empty
comment line.

No functional changes.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Acked-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/arm/xen_arm.c | 19 +------------------
 1 file changed, 1 insertion(+), 18 deletions(-)

diff --git a/hw/arm/xen_arm.c b/hw/arm/xen_arm.c
index 6fad829ede..766a194fa1 100644
--- a/hw/arm/xen_arm.c
+++ b/hw/arm/xen_arm.c
@@ -1,24 +1,7 @@
 /*
  * QEMU ARM Xen PVH Machine
  *
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to deal
- * in the Software without restriction, including without limitation the rights
- * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
- * copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
- * THE SOFTWARE.
+ * SPDX-License-Identifier: MIT
  */
 
 #include "qemu/osdep.h"
-- 
2.43.0

[PULL v1 03/12] hw/arm: xenpvh: Tweak machine description

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Tweak machine description to better express that this is
a Xen PVH machine for ARM.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/arm/xen_arm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/hw/arm/xen_arm.c b/hw/arm/xen_arm.c
index 766a194fa1..5f75cc3779 100644
--- a/hw/arm/xen_arm.c
+++ b/hw/arm/xen_arm.c
@@ -216,7 +216,7 @@ static void xen_arm_machine_class_init(ObjectClass *oc, void *data)
 {
 
     MachineClass *mc = MACHINE_CLASS(oc);
-    mc->desc = "Xen Para-virtualized PC";
+    mc->desc = "Xen PVH ARM machine";
     mc->init = xen_arm_init;
     mc->max_cpus = 1;
     mc->default_machine_opts = "accel=xen";
-- 
2.43.0

[PULL v1 04/12] hw/arm: xenpvh: Add support for SMP guests

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Add SMP support for Xen PVH ARM guests.
Create ms->smp.max_cpus ioreq servers to handle hotplug.

Note that ms->smp.max_cpus will be passed to us by the
user (Xen tools) set to the guests maxvcpus.

The value in mc->max_cpus is an absolute maximum for the
-smp option and won't be used to setup ioreq servers unless
the user explicitly specifies it with -smp.

If the user doesn't pass -smp on the command-line, smp.cpus
and smp.max_cpus will default to 1.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/arm/xen_arm.c | 23 +++++++++++++++++++++--
 1 file changed, 21 insertions(+), 2 deletions(-)

diff --git a/hw/arm/xen_arm.c b/hw/arm/xen_arm.c
index 5f75cc3779..fda65d0d8d 100644
--- a/hw/arm/xen_arm.c
+++ b/hw/arm/xen_arm.c
@@ -173,7 +173,7 @@ static void xen_arm_init(MachineState *machine)
 
     xen_init_ram(machine);
 
-    xen_register_ioreq(xam->state, machine->smp.cpus, &xen_memory_listener);
+    xen_register_ioreq(xam->state, machine->smp.max_cpus, &xen_memory_listener);
 
     xen_create_virtio_mmio_devices(xam);
 
@@ -218,7 +218,26 @@ static void xen_arm_machine_class_init(ObjectClass *oc, void *data)
     MachineClass *mc = MACHINE_CLASS(oc);
     mc->desc = "Xen PVH ARM machine";
     mc->init = xen_arm_init;
-    mc->max_cpus = 1;
+
+    /*
+     * mc->max_cpus holds the MAX value allowed in the -smp command-line opts.
+     *
+     * 1. If users don't pass any -smp option:
+     *   ms->smp.cpus will default to 1.
+     *   ms->smp.max_cpus will default to 1.
+     *
+     * 2. If users pass -smp X:
+     *   ms->smp.cpus will be set to X.
+     *   ms->smp.max_cpus will also be set to X.
+     *
+     * 3. If users pass -smp X,maxcpus=Y:
+     *   ms->smp.cpus will be set to X.
+     *   ms->smp.max_cpus will be set to Y.
+     *
+     * In scenarios 2 and 3, if X or Y are set to something larger than
+     * mc->max_cpus, QEMU will bail out with an error message.
+     */
+    mc->max_cpus = GUEST_MAX_VCPUS;
     mc->default_machine_opts = "accel=xen";
     /* Set explicitly here to make sure that real ram_size is passed */
     mc->default_ram_size = 0;
-- 
2.43.0

[PULL v1 05/12] hw/arm: xenpvh: Remove double-negation in warning

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/arm/xen_arm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/hw/arm/xen_arm.c b/hw/arm/xen_arm.c
index fda65d0d8d..16b3f00992 100644
--- a/hw/arm/xen_arm.c
+++ b/hw/arm/xen_arm.c
@@ -165,7 +165,7 @@ static void xen_arm_init(MachineState *machine)
     xam->state =  g_new0(XenIOState, 1);
 
     if (machine->ram_size == 0) {
-        warn_report("%s non-zero ram size not specified. QEMU machine started"
+        warn_report("%s: ram size not specified. QEMU machine started"
                     " without IOREQ (no emulated devices including virtio)",
                     MACHINE_CLASS(object_get_class(OBJECT(machine)))->desc);
         return;
-- 
2.43.0

[PULL v1 06/12] hw/arm: xenpvh: Move stubbed functions to xen-stubs.c

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/arm/meson.build |  5 ++++-
 hw/arm/xen-stubs.c | 32 ++++++++++++++++++++++++++++++++
 hw/arm/xen_arm.c   | 20 --------------------
 3 files changed, 36 insertions(+), 21 deletions(-)
 create mode 100644 hw/arm/xen-stubs.c

diff --git a/hw/arm/meson.build b/hw/arm/meson.build
index 0c07ab522f..074612b40c 100644
--- a/hw/arm/meson.build
+++ b/hw/arm/meson.build
@@ -59,7 +59,10 @@ arm_ss.add(when: 'CONFIG_FSL_IMX7', if_true: files('fsl-imx7.c', 'mcimx7d-sabre.
 arm_ss.add(when: 'CONFIG_ARM_SMMUV3', if_true: files('smmuv3.c'))
 arm_ss.add(when: 'CONFIG_FSL_IMX6UL', if_true: files('fsl-imx6ul.c', 'mcimx6ul-evk.c'))
 arm_ss.add(when: 'CONFIG_NRF51_SOC', if_true: files('nrf51_soc.c'))
-arm_ss.add(when: 'CONFIG_XEN', if_true: files('xen_arm.c'))
+arm_ss.add(when: 'CONFIG_XEN', if_true: files(
+  'xen-stubs.c',
+  'xen_arm.c',
+))
 
 system_ss.add(when: 'CONFIG_ARM_SMMUV3', if_true: files('smmu-common.c'))
 system_ss.add(when: 'CONFIG_CHEETAH', if_true: files('palm.c'))
diff --git a/hw/arm/xen-stubs.c b/hw/arm/xen-stubs.c
new file mode 100644
index 0000000000..4ac6a56a96
--- /dev/null
+++ b/hw/arm/xen-stubs.c
@@ -0,0 +1,32 @@
+/*
+ * Stubs for unimplemented Xen functions for ARM.
+ *
+ * SPDX-License-Identifier: MIT
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/error-report.h"
+#include "qapi/qapi-commands-migration.h"
+#include "hw/boards.h"
+#include "sysemu/sysemu.h"
+#include "hw/xen/xen-hvm-common.h"
+#include "hw/xen/arch_hvm.h"
+
+void arch_handle_ioreq(XenIOState *state, ioreq_t *req)
+{
+    hw_error("Invalid ioreq type 0x%x\n", req->type);
+    return;
+}
+
+void arch_xen_set_memory(XenIOState *state, MemoryRegionSection *section,
+                         bool add)
+{
+}
+
+void xen_hvm_modified_memory(ram_addr_t start, ram_addr_t length)
+{
+}
+
+void qmp_xen_set_global_dirty_log(bool enable, Error **errp)
+{
+}
diff --git a/hw/arm/xen_arm.c b/hw/arm/xen_arm.c
index 16b3f00992..f0868e7be5 100644
--- a/hw/arm/xen_arm.c
+++ b/hw/arm/xen_arm.c
@@ -115,26 +115,6 @@ static void xen_init_ram(MachineState *machine)
     memory_region_add_subregion(sysmem, XEN_GRANT_ADDR_OFF, &xen_grants);
 }
 
-void arch_handle_ioreq(XenIOState *state, ioreq_t *req)
-{
-    hw_error("Invalid ioreq type 0x%x\n", req->type);
-
-    return;
-}
-
-void arch_xen_set_memory(XenIOState *state, MemoryRegionSection *section,
-                         bool add)
-{
-}
-
-void xen_hvm_modified_memory(ram_addr_t start, ram_addr_t length)
-{
-}
-
-void qmp_xen_set_global_dirty_log(bool enable, Error **errp)
-{
-}
-
 #ifdef CONFIG_TPM
 static void xen_enable_tpm(XenArmState *xam)
 {
-- 
2.43.0

[PULL v1 07/12] hw/arm: xenpvh: Break out a common PVH machine

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Break out a common Xen PVH machine in preparation for
adding a x86 Xen PVH machine.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/arm/trace-events             |   5 -
 hw/arm/xen_arm.c                | 198 +++--------------------
 hw/xen/meson.build              |   1 +
 hw/xen/trace-events             |   4 +
 hw/xen/xen-pvh-common.c         | 275 ++++++++++++++++++++++++++++++++
 include/hw/xen/xen-pvh-common.h |  59 +++++++
 6 files changed, 358 insertions(+), 184 deletions(-)
 create mode 100644 hw/xen/xen-pvh-common.c
 create mode 100644 include/hw/xen/xen-pvh-common.h

diff --git a/hw/arm/trace-events b/hw/arm/trace-events
index be6c8f720b..c64ad344bd 100644
--- a/hw/arm/trace-events
+++ b/hw/arm/trace-events
@@ -68,10 +68,5 @@ z2_aer915_send_too_long(int8_t msg) "message too long (%i bytes)"
 z2_aer915_send(uint8_t reg, uint8_t value) "reg %d value 0x%02x"
 z2_aer915_event(int8_t event, int8_t len) "i2c event =0x%x len=%d bytes"
 
-# xen_arm.c
-xen_create_virtio_mmio_devices(int i, int irq, uint64_t base) "Created virtio-mmio device %d: irq %d base 0x%"PRIx64
-xen_init_ram(uint64_t machine_ram_size) "Initialized xen ram with size 0x%"PRIx64
-xen_enable_tpm(uint64_t addr) "Connected tpmdev at address 0x%"PRIx64
-
 # bcm2838.c
 bcm2838_gic_set_irq(int irq, int level) "gic irq:%d lvl:%d"
diff --git a/hw/arm/xen_arm.c b/hw/arm/xen_arm.c
index f0868e7be5..04cb9855af 100644
--- a/hw/arm/xen_arm.c
+++ b/hw/arm/xen_arm.c
@@ -7,44 +7,12 @@
 #include "qemu/osdep.h"
 #include "qemu/error-report.h"
 #include "qapi/qapi-commands-migration.h"
-#include "qapi/visitor.h"
 #include "hw/boards.h"
-#include "hw/irq.h"
-#include "hw/sysbus.h"
-#include "sysemu/block-backend.h"
-#include "sysemu/tpm_backend.h"
 #include "sysemu/sysemu.h"
-#include "hw/xen/xen-hvm-common.h"
-#include "sysemu/tpm.h"
+#include "hw/xen/xen-pvh-common.h"
 #include "hw/xen/arch_hvm.h"
-#include "trace.h"
 
 #define TYPE_XEN_ARM  MACHINE_TYPE_NAME("xenpvh")
-OBJECT_DECLARE_SIMPLE_TYPE(XenArmState, XEN_ARM)
-
-static const MemoryListener xen_memory_listener = {
-    .region_add = xen_region_add,
-    .region_del = xen_region_del,
-    .log_start = NULL,
-    .log_stop = NULL,
-    .log_sync = NULL,
-    .log_global_start = NULL,
-    .log_global_stop = NULL,
-    .priority = MEMORY_LISTENER_PRIORITY_ACCEL,
-};
-
-struct XenArmState {
-    /*< private >*/
-    MachineState parent;
-
-    XenIOState *state;
-
-    struct {
-        uint64_t tpm_base_addr;
-    } cfg;
-};
-
-static MemoryRegion ram_lo, ram_hi;
 
 /*
  * VIRTIO_MMIO_DEV_SIZE is imported from tools/libs/light/libxl_arm.c under Xen
@@ -57,147 +25,26 @@ static MemoryRegion ram_lo, ram_hi;
 #define NR_VIRTIO_MMIO_DEVICES   \
    (GUEST_VIRTIO_MMIO_SPI_LAST - GUEST_VIRTIO_MMIO_SPI_FIRST)
 
-static void xen_set_irq(void *opaque, int irq, int level)
-{
-    if (xendevicemodel_set_irq_level(xen_dmod, xen_domid, irq, level)) {
-        error_report("xendevicemodel_set_irq_level failed");
-    }
-}
-
-static void xen_create_virtio_mmio_devices(XenArmState *xam)
-{
-    int i;
-
-    for (i = 0; i < NR_VIRTIO_MMIO_DEVICES; i++) {
-        hwaddr base = GUEST_VIRTIO_MMIO_BASE + i * VIRTIO_MMIO_DEV_SIZE;
-        qemu_irq irq = qemu_allocate_irq(xen_set_irq, NULL,
-                                         GUEST_VIRTIO_MMIO_SPI_FIRST + i);
-
-        sysbus_create_simple("virtio-mmio", base, irq);
-
-        trace_xen_create_virtio_mmio_devices(i,
-                                             GUEST_VIRTIO_MMIO_SPI_FIRST + i,
-                                             base);
-    }
-}
-
-static void xen_init_ram(MachineState *machine)
+static void xen_arm_instance_init(Object *obj)
 {
-    MemoryRegion *sysmem = get_system_memory();
-    ram_addr_t block_len, ram_size[GUEST_RAM_BANKS];
-
-    trace_xen_init_ram(machine->ram_size);
-    if (machine->ram_size <= GUEST_RAM0_SIZE) {
-        ram_size[0] = machine->ram_size;
-        ram_size[1] = 0;
-        block_len = GUEST_RAM0_BASE + ram_size[0];
-    } else {
-        ram_size[0] = GUEST_RAM0_SIZE;
-        ram_size[1] = machine->ram_size - GUEST_RAM0_SIZE;
-        block_len = GUEST_RAM1_BASE + ram_size[1];
-    }
+    XenPVHMachineState *s = XEN_PVH_MACHINE(obj);
 
-    memory_region_init_ram(&xen_memory, NULL, "xen.ram", block_len,
-                           &error_fatal);
+    /* Default values.  */
+    s->cfg.ram_low = (MemMapEntry) { GUEST_RAM0_BASE, GUEST_RAM0_SIZE };
+    s->cfg.ram_high = (MemMapEntry) { GUEST_RAM1_BASE, GUEST_RAM1_SIZE };
 
-    memory_region_init_alias(&ram_lo, NULL, "xen.ram.lo", &xen_memory,
-                             GUEST_RAM0_BASE, ram_size[0]);
-    memory_region_add_subregion(sysmem, GUEST_RAM0_BASE, &ram_lo);
-    if (ram_size[1] > 0) {
-        memory_region_init_alias(&ram_hi, NULL, "xen.ram.hi", &xen_memory,
-                                 GUEST_RAM1_BASE, ram_size[1]);
-        memory_region_add_subregion(sysmem, GUEST_RAM1_BASE, &ram_hi);
-    }
-
-    /* Setup support for grants.  */
-    memory_region_init_ram(&xen_grants, NULL, "xen.grants", block_len,
-                           &error_fatal);
-    memory_region_add_subregion(sysmem, XEN_GRANT_ADDR_OFF, &xen_grants);
+    s->cfg.virtio_mmio_num = NR_VIRTIO_MMIO_DEVICES;
+    s->cfg.virtio_mmio_irq_base = GUEST_VIRTIO_MMIO_SPI_FIRST;
+    s->cfg.virtio_mmio = (MemMapEntry) { GUEST_VIRTIO_MMIO_BASE,
+                                         VIRTIO_MMIO_DEV_SIZE };
 }
 
-#ifdef CONFIG_TPM
-static void xen_enable_tpm(XenArmState *xam)
-{
-    Error *errp = NULL;
-    DeviceState *dev;
-    SysBusDevice *busdev;
-
-    TPMBackend *be = qemu_find_tpm_be("tpm0");
-    if (be == NULL) {
-        error_report("Couldn't find tmp0 backend");
-        return;
-    }
-    dev = qdev_new(TYPE_TPM_TIS_SYSBUS);
-    object_property_set_link(OBJECT(dev), "tpmdev", OBJECT(be), &errp);
-    object_property_set_str(OBJECT(dev), "tpmdev", be->id, &errp);
-    busdev = SYS_BUS_DEVICE(dev);
-    sysbus_realize_and_unref(busdev, &error_fatal);
-    sysbus_mmio_map(busdev, 0, xam->cfg.tpm_base_addr);
-
-    trace_xen_enable_tpm(xam->cfg.tpm_base_addr);
-}
-#endif
-
-static void xen_arm_init(MachineState *machine)
-{
-    XenArmState *xam = XEN_ARM(machine);
-
-    xam->state =  g_new0(XenIOState, 1);
-
-    if (machine->ram_size == 0) {
-        warn_report("%s: ram size not specified. QEMU machine started"
-                    " without IOREQ (no emulated devices including virtio)",
-                    MACHINE_CLASS(object_get_class(OBJECT(machine)))->desc);
-        return;
-    }
-
-    xen_init_ram(machine);
-
-    xen_register_ioreq(xam->state, machine->smp.max_cpus, &xen_memory_listener);
-
-    xen_create_virtio_mmio_devices(xam);
-
-#ifdef CONFIG_TPM
-    if (xam->cfg.tpm_base_addr) {
-        xen_enable_tpm(xam);
-    } else {
-        warn_report("tpm-base-addr is not provided. TPM will not be enabled");
-    }
-#endif
-}
-
-#ifdef CONFIG_TPM
-static void xen_arm_get_tpm_base_addr(Object *obj, Visitor *v,
-                                      const char *name, void *opaque,
-                                      Error **errp)
-{
-    XenArmState *xam = XEN_ARM(obj);
-    uint64_t value = xam->cfg.tpm_base_addr;
-
-    visit_type_uint64(v, name, &value, errp);
-}
-
-static void xen_arm_set_tpm_base_addr(Object *obj, Visitor *v,
-                                      const char *name, void *opaque,
-                                      Error **errp)
-{
-    XenArmState *xam = XEN_ARM(obj);
-    uint64_t value;
-
-    if (!visit_type_uint64(v, name, &value, errp)) {
-        return;
-    }
-
-    xam->cfg.tpm_base_addr = value;
-}
-#endif
-
 static void xen_arm_machine_class_init(ObjectClass *oc, void *data)
 {
-
+    XenPVHMachineClass *xpc = XEN_PVH_MACHINE_CLASS(oc);
     MachineClass *mc = MACHINE_CLASS(oc);
+
     mc->desc = "Xen PVH ARM machine";
-    mc->init = xen_arm_init;
 
     /*
      * mc->max_cpus holds the MAX value allowed in the -smp command-line opts.
@@ -218,27 +65,20 @@ static void xen_arm_machine_class_init(ObjectClass *oc, void *data)
      * mc->max_cpus, QEMU will bail out with an error message.
      */
     mc->max_cpus = GUEST_MAX_VCPUS;
-    mc->default_machine_opts = "accel=xen";
-    /* Set explicitly here to make sure that real ram_size is passed */
-    mc->default_ram_size = 0;
 
-#ifdef CONFIG_TPM
-    object_class_property_add(oc, "tpm-base-addr", "uint64_t",
-                              xen_arm_get_tpm_base_addr,
-                              xen_arm_set_tpm_base_addr,
-                              NULL, NULL);
-    object_class_property_set_description(oc, "tpm-base-addr",
-                                          "Set Base address for TPM device.");
+    /* List of supported features known to work on PVH ARM.  */
+    xpc->has_tpm = true;
+    xpc->has_virtio_mmio = true;
 
-    machine_class_allow_dynamic_sysbus_dev(mc, TYPE_TPM_TIS_SYSBUS);
-#endif
+    xen_pvh_class_setup_common_props(xpc);
 }
 
 static const TypeInfo xen_arm_machine_type = {
     .name = TYPE_XEN_ARM,
-    .parent = TYPE_MACHINE,
+    .parent = TYPE_XEN_PVH_MACHINE,
     .class_init = xen_arm_machine_class_init,
-    .instance_size = sizeof(XenArmState),
+    .instance_size = sizeof(XenPVHMachineState),
+    .instance_init = xen_arm_instance_init,
 };
 
 static void xen_arm_machine_register_types(void)
diff --git a/hw/xen/meson.build b/hw/xen/meson.build
index d887fa9ba4..4a486e3673 100644
--- a/hw/xen/meson.build
+++ b/hw/xen/meson.build
@@ -15,6 +15,7 @@ xen_specific_ss = ss.source_set()
 xen_specific_ss.add(files(
   'xen-mapcache.c',
   'xen-hvm-common.c',
+  'xen-pvh-common.c',
 ))
 if have_xen_pci_passthrough
   xen_specific_ss.add(files(
diff --git a/hw/xen/trace-events b/hw/xen/trace-events
index d1b27f6c11..a07fe41c6d 100644
--- a/hw/xen/trace-events
+++ b/hw/xen/trace-events
@@ -64,6 +64,10 @@ destroy_hvm_domain_cannot_acquire_handle(void) "Cannot acquire xenctrl handle"
 destroy_hvm_domain_failed_action(const char *action, int sts, char *errno_s) "xc_domain_shutdown failed to issue %s, sts %d, %s"
 destroy_hvm_domain_action(int xen_domid, const char *action) "Issued domain %d %s"
 
+# xen-pvh-common.c
+xen_create_virtio_mmio_devices(int i, int irq, uint64_t base) "Created virtio-mmio device %d: irq %d base 0x%"PRIx64
+xen_enable_tpm(uint64_t addr) "Connected tpmdev at address 0x%"PRIx64
+
 # xen-mapcache.c
 xen_map_cache(uint64_t phys_addr) "want 0x%"PRIx64
 xen_remap_bucket(uint64_t index) "index 0x%"PRIx64
diff --git a/hw/xen/xen-pvh-common.c b/hw/xen/xen-pvh-common.c
new file mode 100644
index 0000000000..880e8143d7
--- /dev/null
+++ b/hw/xen/xen-pvh-common.c
@@ -0,0 +1,275 @@
+/*
+ * QEMU Xen PVH machine - common code.
+ *
+ * Copyright (c) 2024 Advanced Micro Devices, Inc.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/error-report.h"
+#include "qapi/error.h"
+#include "qapi/visitor.h"
+#include "hw/boards.h"
+#include "hw/irq.h"
+#include "hw/sysbus.h"
+#include "sysemu/sysemu.h"
+#include "sysemu/tpm.h"
+#include "sysemu/tpm_backend.h"
+#include "hw/xen/xen-pvh-common.h"
+#include "trace.h"
+
+static const MemoryListener xen_memory_listener = {
+    .region_add = xen_region_add,
+    .region_del = xen_region_del,
+    .log_start = NULL,
+    .log_stop = NULL,
+    .log_sync = NULL,
+    .log_global_start = NULL,
+    .log_global_stop = NULL,
+    .priority = MEMORY_LISTENER_PRIORITY_ACCEL,
+};
+
+static void xen_pvh_init_ram(XenPVHMachineState *s,
+                             MemoryRegion *sysmem)
+{
+    MachineState *ms = MACHINE(s);
+    ram_addr_t block_len, ram_size[2];
+
+    if (ms->ram_size <= s->cfg.ram_low.size) {
+        ram_size[0] = ms->ram_size;
+        ram_size[1] = 0;
+        block_len = s->cfg.ram_low.base + ram_size[0];
+    } else {
+        ram_size[0] = s->cfg.ram_low.size;
+        ram_size[1] = ms->ram_size - s->cfg.ram_low.size;
+        block_len = s->cfg.ram_high.base + ram_size[1];
+    }
+
+    memory_region_init_ram(&xen_memory, NULL, "xen.ram", block_len,
+                           &error_fatal);
+
+    memory_region_init_alias(&s->ram.low, NULL, "xen.ram.lo", &xen_memory,
+                             s->cfg.ram_low.base, ram_size[0]);
+    memory_region_add_subregion(sysmem, s->cfg.ram_low.base, &s->ram.low);
+    if (ram_size[1] > 0) {
+        memory_region_init_alias(&s->ram.high, NULL, "xen.ram.hi", &xen_memory,
+                                 s->cfg.ram_high.base, ram_size[1]);
+        memory_region_add_subregion(sysmem, s->cfg.ram_high.base, &s->ram.high);
+    }
+
+    /* Setup support for grants.  */
+    memory_region_init_ram(&xen_grants, NULL, "xen.grants", block_len,
+                           &error_fatal);
+    memory_region_add_subregion(sysmem, XEN_GRANT_ADDR_OFF, &xen_grants);
+}
+
+static void xen_set_irq(void *opaque, int irq, int level)
+{
+    if (xendevicemodel_set_irq_level(xen_dmod, xen_domid, irq, level)) {
+        error_report("xendevicemodel_set_irq_level failed");
+    }
+}
+
+static void xen_create_virtio_mmio_devices(XenPVHMachineState *s)
+{
+    int i;
+
+    for (i = 0; i < s->cfg.virtio_mmio_num; i++) {
+        hwaddr base = s->cfg.virtio_mmio.base + i * s->cfg.virtio_mmio.size;
+        qemu_irq irq = qemu_allocate_irq(xen_set_irq, NULL,
+                                         s->cfg.virtio_mmio_irq_base + i);
+
+        sysbus_create_simple("virtio-mmio", base, irq);
+
+        trace_xen_create_virtio_mmio_devices(i,
+                                             s->cfg.virtio_mmio_irq_base + i,
+                                             base);
+    }
+}
+
+#ifdef CONFIG_TPM
+static void xen_enable_tpm(XenPVHMachineState *s)
+{
+    Error *errp = NULL;
+    DeviceState *dev;
+    SysBusDevice *busdev;
+
+    TPMBackend *be = qemu_find_tpm_be("tpm0");
+    if (be == NULL) {
+        error_report("Couldn't find tmp0 backend");
+        return;
+    }
+    dev = qdev_new(TYPE_TPM_TIS_SYSBUS);
+    object_property_set_link(OBJECT(dev), "tpmdev", OBJECT(be), &errp);
+    object_property_set_str(OBJECT(dev), "tpmdev", be->id, &errp);
+    busdev = SYS_BUS_DEVICE(dev);
+    sysbus_realize_and_unref(busdev, &error_fatal);
+    sysbus_mmio_map(busdev, 0, s->cfg.tpm.base);
+
+    trace_xen_enable_tpm(s->cfg.tpm.base);
+}
+#endif
+
+static void xen_pvh_init(MachineState *ms)
+{
+    XenPVHMachineState *s = XEN_PVH_MACHINE(ms);
+    XenPVHMachineClass *xpc = XEN_PVH_MACHINE_GET_CLASS(s);
+    MemoryRegion *sysmem = get_system_memory();
+
+    if (ms->ram_size == 0) {
+        warn_report("%s: ram size not specified. QEMU machine started"
+                    " without IOREQ (no emulated devices including virtio)",
+                    MACHINE_CLASS(object_get_class(OBJECT(ms)))->desc);
+        return;
+    }
+
+    xen_pvh_init_ram(s, sysmem);
+    xen_register_ioreq(&s->ioreq, ms->smp.max_cpus, &xen_memory_listener);
+
+    if (s->cfg.virtio_mmio_num) {
+        xen_create_virtio_mmio_devices(s);
+    }
+
+#ifdef CONFIG_TPM
+    if (xpc->has_tpm) {
+        if (s->cfg.tpm.base) {
+            xen_enable_tpm(s);
+        } else {
+            warn_report("tpm-base-addr is not set. TPM will not be enabled");
+        }
+    }
+#endif
+
+    /* Call the implementation specific init.  */
+    if (xpc->init) {
+        xpc->init(ms);
+    }
+}
+
+#define XEN_PVH_PROP_MEMMAP_SETTER(n, f)                                   \
+static void xen_pvh_set_ ## n ## _ ## f(Object *obj, Visitor *v,           \
+                                       const char *name, void *opaque,     \
+                                       Error **errp)                       \
+{                                                                          \
+    XenPVHMachineState *xp = XEN_PVH_MACHINE(obj);                         \
+    uint64_t value;                                                        \
+                                                                           \
+    if (!visit_type_size(v, name, &value, errp)) {                         \
+        return;                                                            \
+    }                                                                      \
+    xp->cfg.n.f = value;                                                   \
+}
+
+#define XEN_PVH_PROP_MEMMAP_GETTER(n, f)                                   \
+static void xen_pvh_get_ ## n ## _ ## f(Object *obj, Visitor *v,           \
+                                       const char *name, void *opaque,     \
+                                       Error **errp)                       \
+{                                                                          \
+    XenPVHMachineState *xp = XEN_PVH_MACHINE(obj);                         \
+    uint64_t value = xp->cfg.n.f;                                          \
+                                                                           \
+    visit_type_uint64(v, name, &value, errp);                              \
+}
+
+#define XEN_PVH_PROP_MEMMAP_BASE(n)        \
+    XEN_PVH_PROP_MEMMAP_SETTER(n, base)    \
+    XEN_PVH_PROP_MEMMAP_GETTER(n, base)    \
+
+#define XEN_PVH_PROP_MEMMAP_SIZE(n)        \
+    XEN_PVH_PROP_MEMMAP_SETTER(n, size)    \
+    XEN_PVH_PROP_MEMMAP_GETTER(n, size)
+
+#define XEN_PVH_PROP_MEMMAP(n)             \
+    XEN_PVH_PROP_MEMMAP_BASE(n)            \
+    XEN_PVH_PROP_MEMMAP_SIZE(n)
+
+XEN_PVH_PROP_MEMMAP(ram_low)
+XEN_PVH_PROP_MEMMAP(ram_high)
+/* TPM only has a base-addr option.  */
+XEN_PVH_PROP_MEMMAP_BASE(tpm)
+XEN_PVH_PROP_MEMMAP(virtio_mmio)
+
+void xen_pvh_class_setup_common_props(XenPVHMachineClass *xpc)
+{
+    ObjectClass *oc = OBJECT_CLASS(xpc);
+    MachineClass *mc = MACHINE_CLASS(xpc);
+
+#define OC_MEMMAP_PROP_BASE(c, prop_name, name)                           \
+do {                                                                      \
+    object_class_property_add(c, prop_name "-base", "uint64_t",           \
+                              xen_pvh_get_ ## name ## _base,              \
+                              xen_pvh_set_ ## name ## _base, NULL, NULL); \
+    object_class_property_set_description(oc, prop_name "-base",          \
+                              "Set base address for " prop_name);         \
+} while (0)
+
+#define OC_MEMMAP_PROP_SIZE(c, prop_name, name)                           \
+do {                                                                      \
+    object_class_property_add(c, prop_name "-size", "uint64_t",           \
+                              xen_pvh_get_ ## name ## _size,              \
+                              xen_pvh_set_ ## name ## _size, NULL, NULL); \
+    object_class_property_set_description(oc, prop_name "-size",          \
+                              "Set memory range size for " prop_name);    \
+} while (0)
+
+#define OC_MEMMAP_PROP(c, prop_name, name)                                \
+do {                                                                      \
+        OC_MEMMAP_PROP_BASE(c, prop_name, name);                          \
+        OC_MEMMAP_PROP_SIZE(c, prop_name, name);                          \
+} while (0)
+
+    /*
+     * We provide memmap properties to allow Xen to move things to other
+     * addresses for example when users need to accomodate the memory-map
+     * for 1:1 mapped devices/memory.
+     */
+    OC_MEMMAP_PROP(oc, "ram-low", ram_low);
+    OC_MEMMAP_PROP(oc, "ram-high", ram_high);
+
+    if (xpc->has_virtio_mmio) {
+        OC_MEMMAP_PROP(oc, "virtio-mmio", virtio_mmio);
+    }
+
+#ifdef CONFIG_TPM
+    if (xpc->has_tpm) {
+        object_class_property_add(oc, "tpm-base-addr", "uint64_t",
+                                  xen_pvh_get_tpm_base,
+                                  xen_pvh_set_tpm_base,
+                                  NULL, NULL);
+        object_class_property_set_description(oc, "tpm-base-addr",
+                                  "Set Base address for TPM device.");
+
+        machine_class_allow_dynamic_sysbus_dev(mc, TYPE_TPM_TIS_SYSBUS);
+    }
+#endif
+}
+
+static void xen_pvh_class_init(ObjectClass *oc, void *data)
+{
+    MachineClass *mc = MACHINE_CLASS(oc);
+
+    mc->init = xen_pvh_init;
+
+    mc->desc = "Xen PVH machine";
+    mc->max_cpus = 1;
+    mc->default_machine_opts = "accel=xen";
+    /* Set to zero to make sure that the real ram size is passed. */
+    mc->default_ram_size = 0;
+}
+
+static const TypeInfo xen_pvh_info = {
+    .name = TYPE_XEN_PVH_MACHINE,
+    .parent = TYPE_MACHINE,
+    .abstract = true,
+    .instance_size = sizeof(XenPVHMachineState),
+    .class_size = sizeof(XenPVHMachineClass),
+    .class_init = xen_pvh_class_init,
+};
+
+static void xen_pvh_register_types(void)
+{
+    type_register_static(&xen_pvh_info);
+}
+
+type_init(xen_pvh_register_types);
diff --git a/include/hw/xen/xen-pvh-common.h b/include/hw/xen/xen-pvh-common.h
new file mode 100644
index 0000000000..77fd98b9fe
--- /dev/null
+++ b/include/hw/xen/xen-pvh-common.h
@@ -0,0 +1,59 @@
+/*
+ * QEMU Xen PVH machine - common code.
+ *
+ * Copyright (c) 2024 Advanced Micro Devices, Inc.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#ifndef XEN_PVH_COMMON_H__
+#define XEN_PVH_COMMON_H__
+
+#include <assert.h>
+#include "hw/sysbus.h"
+#include "hw/hw.h"
+#include "hw/xen/xen-hvm-common.h"
+#include "hw/pci-host/gpex.h"
+
+#define TYPE_XEN_PVH_MACHINE MACHINE_TYPE_NAME("xen-pvh-base")
+OBJECT_DECLARE_TYPE(XenPVHMachineState, XenPVHMachineClass,
+                    XEN_PVH_MACHINE)
+
+struct XenPVHMachineClass {
+    MachineClass parent;
+
+    /* PVH implementation specific init.  */
+    void (*init)(MachineState *state);
+
+    /*
+     * Each implementation can optionally enable features that it
+     * supports and are known to work.
+     */
+    bool has_tpm;
+    bool has_virtio_mmio;
+};
+
+struct XenPVHMachineState {
+    /*< private >*/
+    MachineState parent;
+
+    XenIOState ioreq;
+
+    struct {
+        MemoryRegion low;
+        MemoryRegion high;
+    } ram;
+
+    struct {
+        MemMapEntry ram_low, ram_high;
+        MemMapEntry tpm;
+
+        /* Virtio-mmio */
+        MemMapEntry virtio_mmio;
+        uint32_t virtio_mmio_num;
+        uint32_t virtio_mmio_irq_base;
+    } cfg;
+};
+
+void xen_pvh_class_setup_common_props(XenPVHMachineClass *xpc);
+#endif
-- 
2.43.0

[PULL v1 08/12] hw/arm: xenpvh: Rename xen_arm.c -> xen-pvh.c

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Rename xen_arm.c -> xen-pvh.c to better express that this
is a PVH machine and to align with x86 HVM and future PVH
machine filenames:
hw/i386/xen/xen-hvm.c
hw/i386/xen/xen-pvh.c (in preparation)

No functional changes.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/arm/meson.build              | 2 +-
 hw/arm/{xen_arm.c => xen-pvh.c} | 0
 2 files changed, 1 insertion(+), 1 deletion(-)
 rename hw/arm/{xen_arm.c => xen-pvh.c} (100%)

diff --git a/hw/arm/meson.build b/hw/arm/meson.build
index 074612b40c..4059d0be2e 100644
--- a/hw/arm/meson.build
+++ b/hw/arm/meson.build
@@ -61,7 +61,7 @@ arm_ss.add(when: 'CONFIG_FSL_IMX6UL', if_true: files('fsl-imx6ul.c', 'mcimx6ul-e
 arm_ss.add(when: 'CONFIG_NRF51_SOC', if_true: files('nrf51_soc.c'))
 arm_ss.add(when: 'CONFIG_XEN', if_true: files(
   'xen-stubs.c',
-  'xen_arm.c',
+  'xen-pvh.c',
 ))
 
 system_ss.add(when: 'CONFIG_ARM_SMMUV3', if_true: files('smmu-common.c'))
diff --git a/hw/arm/xen_arm.c b/hw/arm/xen-pvh.c
similarity index 100%
rename from hw/arm/xen_arm.c
rename to hw/arm/xen-pvh.c
-- 
2.43.0

[PULL v1 09/12] hw/arm: xenpvh: Reverse virtio-mmio creation order

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

We've been creating the virtio-mmio devices in forwards order
but since the qbus lists prepend (rather than append) entries,
the virtio busses end up with decreasing base address order.

Xen enables virtio-mmio nodes in forwards order so there's been
a missmatch. So far, we've been working around this with an
out-of-tree patch to Xen.

This reverses the order making sure the virtio busses end up
ordered with increasing base addresses avoiding the need to
patch Xen.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Acked-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/xen/xen-pvh-common.c | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/hw/xen/xen-pvh-common.c b/hw/xen/xen-pvh-common.c
index 880e8143d7..295f920442 100644
--- a/hw/xen/xen-pvh-common.c
+++ b/hw/xen/xen-pvh-common.c
@@ -75,7 +75,18 @@ static void xen_create_virtio_mmio_devices(XenPVHMachineState *s)
 {
     int i;
 
-    for (i = 0; i < s->cfg.virtio_mmio_num; i++) {
+    /*
+     * We create the transports in reverse order. Since qbus_realize()
+     * prepends (not appends) new child buses, the decrementing loop below will
+     * create a list of virtio-mmio buses with increasing base addresses.
+     *
+     * When a -device option is processed from the command line,
+     * qbus_find_recursive() picks the next free virtio-mmio bus in forwards
+     * order.
+     *
+     * This is what the Xen tools expect.
+     */
+    for (i = s->cfg.virtio_mmio_num - 1; i >= 0; i--) {
         hwaddr base = s->cfg.virtio_mmio.base + i * s->cfg.virtio_mmio.size;
         qemu_irq irq = qemu_allocate_irq(xen_set_irq, NULL,
                                          s->cfg.virtio_mmio_irq_base + i);
-- 
2.43.0

[PULL v1 10/12] hw/xen: pvh-common: Add support for creating PCIe/GPEX

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Add support for optionally creating a PCIe/GPEX controller.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/xen/xen-pvh-common.c         | 76 +++++++++++++++++++++++++++++++++
 include/hw/xen/xen-pvh-common.h | 29 +++++++++++++
 2 files changed, 105 insertions(+)

diff --git a/hw/xen/xen-pvh-common.c b/hw/xen/xen-pvh-common.c
index 295f920442..28d7168446 100644
--- a/hw/xen/xen-pvh-common.c
+++ b/hw/xen/xen-pvh-common.c
@@ -122,6 +122,64 @@ static void xen_enable_tpm(XenPVHMachineState *s)
 }
 #endif
 
+/*
+ * We use the GPEX PCIe controller with its internal INTX PCI interrupt
+ * swizzling. This swizzling is emulated in QEMU and routes all INTX
+ * interrupts from endpoints down to only 4 INTX interrupts.
+ * See include/hw/pci/pci.h : pci_swizzle()
+ */
+static inline void xenpvh_gpex_init(XenPVHMachineState *s,
+                                    XenPVHMachineClass *xpc,
+                                    MemoryRegion *sysmem)
+{
+    MemoryRegion *ecam_reg;
+    MemoryRegion *mmio_reg;
+    DeviceState *dev;
+    int i;
+
+    object_initialize_child(OBJECT(s), "gpex", &s->pci.gpex,
+                            TYPE_GPEX_HOST);
+    dev = DEVICE(&s->pci.gpex);
+    sysbus_realize_and_unref(SYS_BUS_DEVICE(dev), &error_fatal);
+
+    ecam_reg = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0);
+    memory_region_add_subregion(sysmem, s->cfg.pci_ecam.base, ecam_reg);
+
+    mmio_reg = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 1);
+
+    if (s->cfg.pci_mmio.size) {
+        memory_region_init_alias(&s->pci.mmio_alias, OBJECT(dev), "pcie-mmio",
+                                 mmio_reg,
+                                 s->cfg.pci_mmio.base, s->cfg.pci_mmio.size);
+        memory_region_add_subregion(sysmem, s->cfg.pci_mmio.base,
+                                    &s->pci.mmio_alias);
+    }
+
+    if (s->cfg.pci_mmio_high.size) {
+        memory_region_init_alias(&s->pci.mmio_high_alias, OBJECT(dev),
+                "pcie-mmio-high",
+                mmio_reg, s->cfg.pci_mmio_high.base, s->cfg.pci_mmio_high.size);
+        memory_region_add_subregion(sysmem, s->cfg.pci_mmio_high.base,
+                &s->pci.mmio_high_alias);
+    }
+
+    /*
+     * PVH implementations with PCI enabled must provide set_pci_intx_irq()
+     * and optionally an implementation of set_pci_link_route().
+     */
+    assert(xpc->set_pci_intx_irq);
+
+    for (i = 0; i < GPEX_NUM_IRQS; i++) {
+        qemu_irq irq = qemu_allocate_irq(xpc->set_pci_intx_irq, s, i);
+
+        sysbus_connect_irq(SYS_BUS_DEVICE(dev), i, irq);
+        gpex_set_irq_num(GPEX_HOST(dev), i, s->cfg.pci_intx_irq_base + i);
+        if (xpc->set_pci_link_route) {
+            xpc->set_pci_link_route(i, s->cfg.pci_intx_irq_base + i);
+        }
+    }
+}
+
 static void xen_pvh_init(MachineState *ms)
 {
     XenPVHMachineState *s = XEN_PVH_MACHINE(ms);
@@ -152,6 +210,15 @@ static void xen_pvh_init(MachineState *ms)
     }
 #endif
 
+    /* Non-zero pci-ecam-size enables PCI.  */
+    if (s->cfg.pci_ecam.size) {
+        if (s->cfg.pci_ecam.size != 256 * MiB) {
+            error_report("pci-ecam-size only supports values 0 or 0x10000000");
+            exit(EXIT_FAILURE);
+        }
+        xenpvh_gpex_init(s, xpc, sysmem);
+    }
+
     /* Call the implementation specific init.  */
     if (xpc->init) {
         xpc->init(ms);
@@ -200,6 +267,9 @@ XEN_PVH_PROP_MEMMAP(ram_high)
 /* TPM only has a base-addr option.  */
 XEN_PVH_PROP_MEMMAP_BASE(tpm)
 XEN_PVH_PROP_MEMMAP(virtio_mmio)
+XEN_PVH_PROP_MEMMAP(pci_ecam)
+XEN_PVH_PROP_MEMMAP(pci_mmio)
+XEN_PVH_PROP_MEMMAP(pci_mmio_high)
 
 void xen_pvh_class_setup_common_props(XenPVHMachineClass *xpc)
 {
@@ -242,6 +312,12 @@ do {                                                                      \
         OC_MEMMAP_PROP(oc, "virtio-mmio", virtio_mmio);
     }
 
+    if (xpc->has_pci) {
+        OC_MEMMAP_PROP(oc, "pci-ecam", pci_ecam);
+        OC_MEMMAP_PROP(oc, "pci-mmio", pci_mmio);
+        OC_MEMMAP_PROP(oc, "pci-mmio-high", pci_mmio_high);
+    }
+
 #ifdef CONFIG_TPM
     if (xpc->has_tpm) {
         object_class_property_add(oc, "tpm-base-addr", "uint64_t",
diff --git a/include/hw/xen/xen-pvh-common.h b/include/hw/xen/xen-pvh-common.h
index 77fd98b9fe..bc09eea936 100644
--- a/include/hw/xen/xen-pvh-common.h
+++ b/include/hw/xen/xen-pvh-common.h
@@ -25,10 +25,29 @@ struct XenPVHMachineClass {
     /* PVH implementation specific init.  */
     void (*init)(MachineState *state);
 
+    /*
+     * set_pci_intx_irq - Deliver INTX irqs to the guest.
+     *
+     * @opaque: pointer to XenPVHMachineState.
+     * @irq: IRQ after swizzling, between 0-3.
+     * @level: IRQ level.
+     */
+    void (*set_pci_intx_irq)(void *opaque, int irq, int level);
+
+    /*
+     * set_pci_link_route: - optional implementation call to setup
+     * routing between INTX IRQ (0 - 3) and GSI's.
+     *
+     * @line: line the INTx line (0 => A .. 3 => B)
+     * @irq: GSI
+     */
+    int (*set_pci_link_route)(uint8_t line, uint8_t irq);
+
     /*
      * Each implementation can optionally enable features that it
      * supports and are known to work.
      */
+    bool has_pci;
     bool has_tpm;
     bool has_virtio_mmio;
 };
@@ -44,6 +63,12 @@ struct XenPVHMachineState {
         MemoryRegion high;
     } ram;
 
+    struct {
+        GPEXHost gpex;
+        MemoryRegion mmio_alias;
+        MemoryRegion mmio_high_alias;
+    } pci;
+
     struct {
         MemMapEntry ram_low, ram_high;
         MemMapEntry tpm;
@@ -52,6 +77,10 @@ struct XenPVHMachineState {
         MemMapEntry virtio_mmio;
         uint32_t virtio_mmio_num;
         uint32_t virtio_mmio_irq_base;
+
+        /* PCI */
+        MemMapEntry pci_ecam, pci_mmio, pci_mmio_high;
+        uint32_t pci_intx_irq_base;
     } cfg;
 };
 
-- 
2.43.0

[PULL v1 11/12] hw/i386/xen: Add a Xen PVH x86 machine

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Add a Xen PVH x86 machine based on the abstract PVH Machine.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/i386/xen/meson.build |   1 +
 hw/i386/xen/xen-pvh.c   | 121 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 122 insertions(+)
 create mode 100644 hw/i386/xen/xen-pvh.c

diff --git a/hw/i386/xen/meson.build b/hw/i386/xen/meson.build
index 3f0df8bc07..c73c62b8e3 100644
--- a/hw/i386/xen/meson.build
+++ b/hw/i386/xen/meson.build
@@ -4,6 +4,7 @@ i386_ss.add(when: 'CONFIG_XEN', if_true: files(
 ))
 i386_ss.add(when: ['CONFIG_XEN', xen], if_true: files(
   'xen-hvm.c',
+  'xen-pvh.c',
 ))
 
 i386_ss.add(when: 'CONFIG_XEN_BUS', if_true: files(
diff --git a/hw/i386/xen/xen-pvh.c b/hw/i386/xen/xen-pvh.c
new file mode 100644
index 0000000000..45645667e9
--- /dev/null
+++ b/hw/i386/xen/xen-pvh.c
@@ -0,0 +1,121 @@
+/*
+ * QEMU Xen PVH x86 Machine
+ *
+ * Copyright (c) 2024 Advanced Micro Devices, Inc.
+ * Written by Edgar E. Iglesias <edgar.iglesias@amd.com>
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/error-report.h"
+#include "hw/boards.h"
+#include "sysemu/sysemu.h"
+#include "hw/xen/arch_hvm.h"
+#include <xen/hvm/hvm_info_table.h>
+#include "hw/xen/xen-pvh-common.h"
+
+#define TYPE_XEN_PVH_X86  MACHINE_TYPE_NAME("xenpvh")
+OBJECT_DECLARE_SIMPLE_TYPE(XenPVHx86State, XEN_PVH_X86)
+
+struct XenPVHx86State {
+    /*< private >*/
+    XenPVHMachineState parent;
+
+    DeviceState **cpu;
+};
+
+static DeviceState *xen_pvh_cpu_new(MachineState *ms,
+                                    int64_t apic_id)
+{
+    Object *cpu = object_new(ms->cpu_type);
+
+    object_property_add_child(OBJECT(ms), "cpu[*]", cpu);
+    object_property_set_uint(cpu, "apic-id", apic_id, &error_fatal);
+    qdev_realize(DEVICE(cpu), NULL, &error_fatal);
+    object_unref(cpu);
+
+    return DEVICE(cpu);
+}
+
+static void xen_pvh_init(MachineState *ms)
+{
+    XenPVHx86State *xp = XEN_PVH_X86(ms);
+    int i;
+
+    /* Create dummy cores. This will indirectly create the APIC MSI window.  */
+    xp->cpu = g_malloc(sizeof xp->cpu[0] * ms->smp.max_cpus);
+    for (i = 0; i < ms->smp.max_cpus; i++) {
+        xp->cpu[i] = xen_pvh_cpu_new(ms, i);
+    }
+}
+
+static void xen_pvh_instance_init(Object *obj)
+{
+    XenPVHMachineState *s = XEN_PVH_MACHINE(obj);
+
+    /* Default values.  */
+    s->cfg.ram_low = (MemMapEntry) { 0x0, 0x80000000U };
+    s->cfg.ram_high = (MemMapEntry) { 0xC000000000ULL, 0x4000000000ULL };
+    s->cfg.pci_intx_irq_base = 16;
+}
+
+/*
+ * Deliver INTX interrupts to Xen guest.
+ */
+static void xen_pvh_set_pci_intx_irq(void *opaque, int irq, int level)
+{
+    /*
+     * Since QEMU emulates all of the swizziling
+     * We don't want Xen to do any additional swizzling in
+     * xen_set_pci_intx_level() so we always set device to 0.
+     */
+    if (xen_set_pci_intx_level(xen_domid, 0, 0, 0, irq, level)) {
+        error_report("xendevicemodel_set_pci_intx_level failed");
+    }
+}
+
+static void xen_pvh_machine_class_init(ObjectClass *oc, void *data)
+{
+    XenPVHMachineClass *xpc = XEN_PVH_MACHINE_CLASS(oc);
+    MachineClass *mc = MACHINE_CLASS(oc);
+
+    mc->desc = "Xen PVH x86 machine";
+    mc->default_cpu_type = TARGET_DEFAULT_CPU_TYPE;
+
+    /* mc->max_cpus holds the MAX value allowed in the -smp cmd-line opts. */
+    mc->max_cpus = HVM_MAX_VCPUS;
+
+    /* We have an implementation specific init to create CPU objects.  */
+    xpc->init = xen_pvh_init;
+
+    /*
+     * PCI INTX routing.
+     *
+     * We describe the mapping between the 4 INTX interrupt and GSIs
+     * using xen_set_pci_link_route(). xen_pvh_set_pci_intx_irq is
+     * used to deliver the interrupt.
+     */
+    xpc->set_pci_intx_irq = xen_pvh_set_pci_intx_irq;
+    xpc->set_pci_link_route = xen_set_pci_link_route;
+
+    /* List of supported features known to work on PVH x86.  */
+    xpc->has_pci = true;
+
+    xen_pvh_class_setup_common_props(xpc);
+}
+
+static const TypeInfo xen_pvh_x86_machine_type = {
+    .name = TYPE_XEN_PVH_X86,
+    .parent = TYPE_XEN_PVH_MACHINE,
+    .class_init = xen_pvh_machine_class_init,
+    .instance_init = xen_pvh_instance_init,
+    .instance_size = sizeof(XenPVHx86State),
+};
+
+static void xen_pvh_machine_register_types(void)
+{
+    type_register_static(&xen_pvh_x86_machine_type);
+}
+
+type_init(xen_pvh_machine_register_types)
-- 
2.43.0

[PULL v1 12/12] docs/system/i386: xenpvh: Add a basic description

[Edgar E. Iglesias]

From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 MAINTAINERS                 |  1 +
 docs/system/i386/xenpvh.rst | 49 +++++++++++++++++++++++++++++++++++++
 docs/system/target-i386.rst |  1 +
 3 files changed, 51 insertions(+)
 create mode 100644 docs/system/i386/xenpvh.rst

diff --git a/MAINTAINERS b/MAINTAINERS
index c2fb0c2f42..c14ac014e2 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -560,6 +560,7 @@ F: include/sysemu/xen.h
 F: include/sysemu/xen-mapcache.h
 F: stubs/xen-hw-stub.c
 F: docs/system/arm/xenpvh.rst
+F: docs/system/i386/xenpvh.rst
 
 Guest CPU Cores (NVMM)
 ----------------------
diff --git a/docs/system/i386/xenpvh.rst b/docs/system/i386/xenpvh.rst
new file mode 100644
index 0000000000..354250f073
--- /dev/null
+++ b/docs/system/i386/xenpvh.rst
@@ -0,0 +1,49 @@
+Xen PVH machine (``xenpvh``)
+=========================================
+
+Xen supports a spectrum of types of guests that vary in how they depend
+on HW virtualization features, emulation models and paravirtualization.
+PVH is a mode that uses HW virtualization features (like HVM) but tries
+to avoid emulation models and instead use passthrough or
+paravirtualized devices.
+
+QEMU can be used to provide PV virtio devices on an emulated PCIe controller.
+That is the purpose of this minimal machine.
+
+Supported devices
+-----------------
+
+The x86 Xen PVH QEMU machine provide the following devices:
+
+- RAM
+- GPEX host bridge
+- virtio-pci devices
+
+The idea is to only connect virtio-pci devices but in theory any compatible
+PCI device model will work depending on Xen and guest support.
+
+Running
+-------
+
+The Xen tools will typically construct a command-line and launch QEMU
+for you when needed. But here's an example of what it can look like in
+case you need to construct one manually:
+
+.. code-block:: console
+
+    qemu-system-i386 -xen-domid 3 -no-shutdown        \
+      -chardev socket,id=libxl-cmd,path=/var/run/xen/qmp-libxl-3,server=on,wait=off \
+      -mon chardev=libxl-cmd,mode=control             \
+      -chardev socket,id=libxenstat-cmd,path=/var/run/xen/qmp-libxenstat-3,server=on,wait=off \
+      -mon chardev=libxenstat-cmd,mode=control        \
+      -nodefaults                                     \
+      -no-user-config                                 \
+      -xen-attach -name g0                            \
+      -vnc none                                       \
+      -display none                                   \
+      -device virtio-net-pci,id=nic0,netdev=net0,mac=00:16:3e:5c:81:78 \
+      -netdev type=tap,id=net0,ifname=vif3.0-emu,br=xenbr0,script=no,downscript=no \
+      -smp 4,maxcpus=4                                \
+      -nographic                                      \
+      -machine xenpvh,ram-low-base=0,ram-low-size=2147483648,ram-high-base=4294967296,ram-high-size=2147483648,pci-ecam-base=824633720832,pci-ecam-size=268435456,pci-mmio-base=4026531840,pci-mmio-size=33554432,pci-mmio-high-base=824902156288,pci-mmio-high-size=68719476736 \
+      -m 4096
diff --git a/docs/system/target-i386.rst b/docs/system/target-i386.rst
index 1b8a1f248a..23e84e3ba7 100644
--- a/docs/system/target-i386.rst
+++ b/docs/system/target-i386.rst
@@ -26,6 +26,7 @@ Architectural features
    i386/cpu
    i386/hyperv
    i386/xen
+   i386/xenpvh
    i386/kvm-pv
    i386/sgx
    i386/amd-memory-encryption
-- 
2.43.0

Re: [PULL v1 00/12] Xen queue

[Peter Maydell]

On Wed, 4 Sept 2024 at 17:15, Edgar E. Iglesias
<edgar.iglesias@gmail.com> wrote:
>
> From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>
>
> The following changes since commit e638d685ec2a0700fb9529cbd1b2823ac4120c53:
>
>   Open 9.2 development tree (2024-09-03 09:18:43 -0700)
>
> are available in the Git repository at:
>
>   https://gitlab.com/edgar.iglesias/qemu.git tags/edgar/xen-queue-2024-09-04.for-upstream
>
> for you to fetch changes up to 0b57c8160a2a6c833cfb1d958f08385c4391ab70:
>
>   docs/system/i386: xenpvh: Add a basic description (2024-09-04 16:50:43 +0200)
>
> ----------------------------------------------------------------
> Edgars Xen queue.
>
> ----------------------------------------------------------------
> Edgar E. Iglesias (12):
>       MAINTAINERS: Add docs/system/arm/xenpvh.rst
>       hw/arm: xenpvh: Update file header to use SPDX
>       hw/arm: xenpvh: Tweak machine description
>       hw/arm: xenpvh: Add support for SMP guests
>       hw/arm: xenpvh: Remove double-negation in warning
>       hw/arm: xenpvh: Move stubbed functions to xen-stubs.c
>       hw/arm: xenpvh: Break out a common PVH machine
>       hw/arm: xenpvh: Rename xen_arm.c -> xen-pvh.c
>       hw/arm: xenpvh: Reverse virtio-mmio creation order
>       hw/xen: pvh-common: Add support for creating PCIe/GPEX
>       hw/i386/xen: Add a Xen PVH x86 machine
>       docs/system/i386: xenpvh: Add a basic description


Applied, thanks.

Please update the changelog at https://wiki.qemu.org/ChangeLog/9.2
for any user-visible changes.

-- PMM

Re: [PULL v1 10/12] hw/xen: pvh-common: Add support for creating PCIe/GPEX

[Philippe Mathieu-Daudé]

Hi Edgar,

On 4/9/24 18:15, Edgar E. Iglesias wrote:
> From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>
> 
> Add support for optionally creating a PCIe/GPEX controller.
> 
> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
> Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
> ---
>   hw/xen/xen-pvh-common.c         | 76 +++++++++++++++++++++++++++++++++
>   include/hw/xen/xen-pvh-common.h | 29 +++++++++++++
>   2 files changed, 105 insertions(+)


> +/*
> + * We use the GPEX PCIe controller with its internal INTX PCI interrupt
> + * swizzling. This swizzling is emulated in QEMU and routes all INTX
> + * interrupts from endpoints down to only 4 INTX interrupts.
> + * See include/hw/pci/pci.h : pci_swizzle()
> + */
> +static inline void xenpvh_gpex_init(XenPVHMachineState *s,
> +                                    XenPVHMachineClass *xpc,
> +                                    MemoryRegion *sysmem)
> +{
> +    MemoryRegion *ecam_reg;
> +    MemoryRegion *mmio_reg;
> +    DeviceState *dev;
> +    int i;
> +
> +    object_initialize_child(OBJECT(s), "gpex", &s->pci.gpex,
> +                            TYPE_GPEX_HOST);
> +    dev = DEVICE(&s->pci.gpex);
> +    sysbus_realize_and_unref(SYS_BUS_DEVICE(dev), &error_fatal);
> +
> +    ecam_reg = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0);
> +    memory_region_add_subregion(sysmem, s->cfg.pci_ecam.base, ecam_reg);
> +
> +    mmio_reg = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 1);
> +
> +    if (s->cfg.pci_mmio.size) {
> +        memory_region_init_alias(&s->pci.mmio_alias, OBJECT(dev), "pcie-mmio",
> +                                 mmio_reg,
> +                                 s->cfg.pci_mmio.base, s->cfg.pci_mmio.size);
> +        memory_region_add_subregion(sysmem, s->cfg.pci_mmio.base,
> +                                    &s->pci.mmio_alias);
> +    }
> +
> +    if (s->cfg.pci_mmio_high.size) {
> +        memory_region_init_alias(&s->pci.mmio_high_alias, OBJECT(dev),
> +                "pcie-mmio-high",
> +                mmio_reg, s->cfg.pci_mmio_high.base, s->cfg.pci_mmio_high.size);
> +        memory_region_add_subregion(sysmem, s->cfg.pci_mmio_high.base,
> +                &s->pci.mmio_high_alias);
> +    }
> +
> +    /*
> +     * PVH implementations with PCI enabled must provide set_pci_intx_irq()
> +     * and optionally an implementation of set_pci_link_route().
> +     */
> +    assert(xpc->set_pci_intx_irq);
> +
> +    for (i = 0; i < GPEX_NUM_IRQS; i++) {
> +        qemu_irq irq = qemu_allocate_irq(xpc->set_pci_intx_irq, s, i);
> +
> +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), i, irq);
> +        gpex_set_irq_num(GPEX_HOST(dev), i, s->cfg.pci_intx_irq_base + i);
> +        if (xpc->set_pci_link_route) {
> +            xpc->set_pci_link_route(i, s->cfg.pci_intx_irq_base + i);
> +        }
> +    }
> +}

Some Kconfig selector seems missing here:

/usr/bin/ld: libqemu-aarch64-softmmu.a.p/hw_xen_xen-pvh-common.c.o: in 
function `xenpvh_gpex_init':
hw/xen/xen-pvh-common.c:174: undefined reference to `gpex_set_irq_num'
/usr/bin/ld: libqemu-aarch64-softmmu.a.p/hw_xen_xen-hvm-common.c.o: in 
function `pci_dev_bus_num':
include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
/usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
/usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
/usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
/usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
/usr/bin/ld: libqemu-aarch64-softmmu.a.p/hw_xen_xen-hvm-common.c.o: in 
function `cpu_ioreq_config':
hw/xen/xen-hvm-common.c:412: undefined reference to 
`pci_host_config_read_common'
/usr/bin/ld: hw/xen/xen-hvm-common.c:428: undefined reference to 
`pci_host_config_read_common'
/usr/bin/ld: hw/xen/xen-hvm-common.c:438: undefined reference to 
`pci_host_config_write_common'

The current 'XEN' key represents both the "accelerator" part and
the common Xen HW, which isn't helping to follow. Anyhow, this
snippet fixes the build issue:

-- >8 --
diff --git a/accel/Kconfig b/accel/Kconfig
index 794e0d18d2..4263cab722 100644
--- a/accel/Kconfig
+++ b/accel/Kconfig
@@ -16,4 +16,5 @@ config KVM
  config XEN
      bool
      select FSDEV_9P if VIRTFS
+    select PCI_EXPRESS_GENERIC_BRIDGE
      select XEN_BUS
---

I'll post a patch later.

Regards,

Phil.

Re: [PULL v1 10/12] hw/xen: pvh-common: Add support for creating PCIe/GPEX

[Edgar E. Iglesias]

[-- Attachment #1: Type: text/plain, Size: 4873 bytes --]

On Tue, Feb 18, 2025 at 6:02 AM Philippe Mathieu-Daudé <philmd@linaro.org>
wrote:

> Hi Edgar,
>
> On 4/9/24 18:15, Edgar E. Iglesias wrote:
> > From: "Edgar E. Iglesias" <edgar.iglesias@amd.com>
> >
> > Add support for optionally creating a PCIe/GPEX controller.
> >
> > Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
> > Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
> > ---
> >   hw/xen/xen-pvh-common.c         | 76 +++++++++++++++++++++++++++++++++
> >   include/hw/xen/xen-pvh-common.h | 29 +++++++++++++
> >   2 files changed, 105 insertions(+)
>
>
> > +/*
> > + * We use the GPEX PCIe controller with its internal INTX PCI interrupt
> > + * swizzling. This swizzling is emulated in QEMU and routes all INTX
> > + * interrupts from endpoints down to only 4 INTX interrupts.
> > + * See include/hw/pci/pci.h : pci_swizzle()
> > + */
> > +static inline void xenpvh_gpex_init(XenPVHMachineState *s,
> > +                                    XenPVHMachineClass *xpc,
> > +                                    MemoryRegion *sysmem)
> > +{
> > +    MemoryRegion *ecam_reg;
> > +    MemoryRegion *mmio_reg;
> > +    DeviceState *dev;
> > +    int i;
> > +
> > +    object_initialize_child(OBJECT(s), "gpex", &s->pci.gpex,
> > +                            TYPE_GPEX_HOST);
> > +    dev = DEVICE(&s->pci.gpex);
> > +    sysbus_realize_and_unref(SYS_BUS_DEVICE(dev), &error_fatal);
> > +
> > +    ecam_reg = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0);
> > +    memory_region_add_subregion(sysmem, s->cfg.pci_ecam.base, ecam_reg);
> > +
> > +    mmio_reg = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 1);
> > +
> > +    if (s->cfg.pci_mmio.size) {
> > +        memory_region_init_alias(&s->pci.mmio_alias, OBJECT(dev),
> "pcie-mmio",
> > +                                 mmio_reg,
> > +                                 s->cfg.pci_mmio.base,
> s->cfg.pci_mmio.size);
> > +        memory_region_add_subregion(sysmem, s->cfg.pci_mmio.base,
> > +                                    &s->pci.mmio_alias);
> > +    }
> > +
> > +    if (s->cfg.pci_mmio_high.size) {
> > +        memory_region_init_alias(&s->pci.mmio_high_alias, OBJECT(dev),
> > +                "pcie-mmio-high",
> > +                mmio_reg, s->cfg.pci_mmio_high.base,
> s->cfg.pci_mmio_high.size);
> > +        memory_region_add_subregion(sysmem, s->cfg.pci_mmio_high.base,
> > +                &s->pci.mmio_high_alias);
> > +    }
> > +
> > +    /*
> > +     * PVH implementations with PCI enabled must provide
> set_pci_intx_irq()
> > +     * and optionally an implementation of set_pci_link_route().
> > +     */
> > +    assert(xpc->set_pci_intx_irq);
> > +
> > +    for (i = 0; i < GPEX_NUM_IRQS; i++) {
> > +        qemu_irq irq = qemu_allocate_irq(xpc->set_pci_intx_irq, s, i);
> > +
> > +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), i, irq);
> > +        gpex_set_irq_num(GPEX_HOST(dev), i, s->cfg.pci_intx_irq_base +
> i);
> > +        if (xpc->set_pci_link_route) {
> > +            xpc->set_pci_link_route(i, s->cfg.pci_intx_irq_base + i);
> > +        }
> > +    }
> > +}
>
> Some Kconfig selector seems missing here:
>
> /usr/bin/ld: libqemu-aarch64-softmmu.a.p/hw_xen_xen-pvh-common.c.o: in
> function `xenpvh_gpex_init':
> hw/xen/xen-pvh-common.c:174: undefined reference to `gpex_set_irq_num'
> /usr/bin/ld: libqemu-aarch64-softmmu.a.p/hw_xen_xen-hvm-common.c.o: in
> function `pci_dev_bus_num':
> include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: include/hw/pci/pci.h:337: undefined reference to `pci_bus_num'
> /usr/bin/ld: libqemu-aarch64-softmmu.a.p/hw_xen_xen-hvm-common.c.o: in
> function `cpu_ioreq_config':
> hw/xen/xen-hvm-common.c:412: undefined reference to
> `pci_host_config_read_common'
> /usr/bin/ld: hw/xen/xen-hvm-common.c:428: undefined reference to
> `pci_host_config_read_common'
> /usr/bin/ld: hw/xen/xen-hvm-common.c:438: undefined reference to
> `pci_host_config_write_common'
>
> The current 'XEN' key represents both the "accelerator" part and
> the common Xen HW, which isn't helping to follow. Anyhow, this
> snippet fixes the build issue:
>
> -- >8 --
> diff --git a/accel/Kconfig b/accel/Kconfig
> index 794e0d18d2..4263cab722 100644
> --- a/accel/Kconfig
> +++ b/accel/Kconfig
> @@ -16,4 +16,5 @@ config KVM
>   config XEN
>       bool
>       select FSDEV_9P if VIRTFS
> +    select PCI_EXPRESS_GENERIC_BRIDGE
>       select XEN_BUS
> ---
>
> I'll post a patch later.
>


Sounds good, thanks Phil!

Cheers,
Edgar


>
> Regards,
>
> Phil.
>

[-- Attachment #2: Type: text/html, Size: 6476 bytes --]