Skip to contentSkip to conversation

Extract build_mcfg

PATCH Superseded Wei Yang 5 participants 20 messages Apr 19, 2019 – May 20, 2019

v4 of 5 Superseded Latest is v5 →

lore
Share & tools
lore

Wei Yang

author
#0 lore raw

v4 cover Extract build_mcfg

This patch set tries to generalize MCFG table build process. And it is based on one un-merged patch from Igor, which is included in this serials.

v3->v4:
* adjust comment to give more information about MCFG table

v2->v3:
* Includes the un-merged patch from Igor
* use build_append_foo() API to construct MCFG

Igor Mammedov (1):
q35: acpi: do not create dummy MCFG table

Wei Yang (5):
hw/arm/virt-acpi-build: remove unnecessary variable mcfg_start i386, acpi: remove mcfg_ prefix in AcpiMcfgInfo members hw/arm/virt-acpi-build: pass AcpiMcfgInfo to build_mcfg() hw/acpi: Consolidate build_mcfg to pci.c
acpi: pci: use build_append_foo() API to construct MCFG

 default-configs/arm-softmmu.mak  |  1 +
 default-configs/i386-softmmu.mak |  1 +
 hw/acpi/Kconfig                  |  4 +++
 hw/acpi/Makefile.objs            |  1 +
 hw/acpi/pci.c                    | 55 ++++++++++++++++++++++++++++++++
 hw/arm/virt-acpi-build.c         | 31 +++++-------------
 hw/i386/acpi-build.c             | 44 ++++---------------------
 include/hw/acpi/acpi-defs.h      | 18 -----------
 include/hw/acpi/pci.h            | 34 ++++++++++++++++++++
 9 files changed, 111 insertions(+), 78 deletions(-)
 create mode 100644 hw/acpi/pci.c
 create mode 100644 include/hw/acpi/pci.h
Signature · 1 line
2.19.1
author
#1 lore raw

PATCH v4 1/6 q35: acpi: do not create dummy MCFG table

From: Igor Mammedov <[email protected]>

Dummy table (with signature "QEMU") creation came from original SeaBIOS codebase. And QEMU would have to keep it around if there were Q35 machine that depended on keeping ACPI tables blob constant size. Luckily there were no versioned Q35 machine types before commit:
(since 2.3) a1666142db acpi-build: make ROMs RAM blocks resizeable
which obsoleted need to keep ACPI tables blob the same size on source/destination.

Considering the 1st versioned machine is pc-q35-2.4, the dummy table is not really necessary and it's safe to drop it without breaking cross version migration in both directions unconditionally.

---
 hw/i386/acpi-build.c | 18 ++++--------------
 1 file changed, 4 insertions(+), 14 deletions(-)
hw/i386/acpi-build.c +4 -14
index b17d4a711d..d009176072 100644
@@ -2401,7 +2401,6 @@ static void
 build_mcfg_q35(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
 {
     AcpiTableMcfg *mcfg;
-    const char *sig;
     int len = sizeof(*mcfg) + 1 * sizeof(mcfg->allocation[0]);
 
     mcfg = acpi_data_push(table_data, len);
@@ -2411,19 +2410,7 @@ build_mcfg_q35(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
     mcfg->allocation[0].start_bus_number = 0;
     mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->mcfg_size - 1);
 
-    /* MCFG is used for ECAM which can be enabled or disabled by guest.
-     * To avoid table size changes (which create migration issues),
-     * always create the table even if there are no allocations,
-     * but set the signature to a reserved value in this case.
-     * ACPI spec requires OSPMs to ignore such tables.
-     */
-    if (info->mcfg_base == PCIE_BASE_ADDR_UNMAPPED) {
-        /* Reserved signature: ignored by OSPM */
-        sig = "QEMU";
-    } else {
-        sig = "MCFG";
-    }
-    build_header(linker, table_data, (void *)mcfg, sig, len, 1, NULL, NULL);
+    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
 }
 
 /*
@@ -2592,6 +2579,9 @@ static bool acpi_get_mcfg(AcpiMcfgInfo *mcfg)
     }
     mcfg->mcfg_base = qnum_get_uint(qobject_to(QNum, o));
     qobject_unref(o);
+    if (mcfg->mcfg_base == PCIE_BASE_ADDR_UNMAPPED) {
+        return false;
+    }
 
     o = object_property_get_qobject(pci_host, PCIE_HOST_MCFG_SIZE, NULL);
     assert(o);
Signature · 1 line
2.19.1
author
#2 lore raw

PATCH v4 2/6 hw/arm/virt-acpi-build: remove unnecessary variable mcfg_start

In reply to Wei Yang

mcfg_start points to the start of MCFG table and is used in build_header. While this information could be derived from mcfg.

This patch removes the unnecessary variable mcfg_start.

---
 hw/arm/virt-acpi-build.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
hw/arm/virt-acpi-build.c +1 -3
index 857989362a..e09e7eff8d 100644
@@ -552,7 +552,6 @@ build_mcfg(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
     const MemMapEntry *memmap = vms->memmap;
     int ecam_id = VIRT_ECAM_ID(vms->highmem_ecam);
     int len = sizeof(*mcfg) + sizeof(mcfg->allocation[0]);
-    int mcfg_start = table_data->len;
 
     mcfg = acpi_data_push(table_data, len);
     mcfg->allocation[0].address = cpu_to_le64(memmap[ecam_id].base);
@@ -563,8 +562,7 @@ build_mcfg(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
     mcfg->allocation[0].end_bus_number =
         PCIE_MMCFG_BUS(memmap[ecam_id].size - 1);
 
-    build_header(linker, table_data, (void *)(table_data->data + mcfg_start),
-                 "MCFG", table_data->len - mcfg_start, 1, NULL, NULL);
+    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
 }
 
 /* GTDT */
Signature · 1 line
2.19.1
author
#3 lore raw

PATCH v4 3/6 i386, acpi: remove mcfg_ prefix in AcpiMcfgInfo members

In reply to Wei Yang

This is obvious the member in AcpiMcfgInfo describe MCFG's property.

Remove the mcfg_ prefix.

---
 hw/i386/acpi-build.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
hw/i386/acpi-build.c +7 -7
index d009176072..f0d27bffd6 100644
@@ -87,8 +87,8 @@
 #define ACPI_BUILD_IOAPIC_ID 0x0
 
 typedef struct AcpiMcfgInfo {
-    uint64_t mcfg_base;
-    uint32_t mcfg_size;
+    uint64_t base;
+    uint32_t size;
 } AcpiMcfgInfo;
 
 typedef struct AcpiPmInfo {
@@ -2404,11 +2404,11 @@ build_mcfg_q35(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
     int len = sizeof(*mcfg) + 1 * sizeof(mcfg->allocation[0]);
 
     mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(info->mcfg_base);
+    mcfg->allocation[0].address = cpu_to_le64(info->base);
     /* Only a single allocation so no need to play with segments */
     mcfg->allocation[0].pci_segment = cpu_to_le16(0);
     mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->mcfg_size - 1);
+    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
 
     build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
 }
@@ -2577,15 +2577,15 @@ static bool acpi_get_mcfg(AcpiMcfgInfo *mcfg)
     if (!o) {
         return false;
     }
-    mcfg->mcfg_base = qnum_get_uint(qobject_to(QNum, o));
+    mcfg->base = qnum_get_uint(qobject_to(QNum, o));
     qobject_unref(o);
-    if (mcfg->mcfg_base == PCIE_BASE_ADDR_UNMAPPED) {
+    if (mcfg->base == PCIE_BASE_ADDR_UNMAPPED) {
         return false;
     }
 
     o = object_property_get_qobject(pci_host, PCIE_HOST_MCFG_SIZE, NULL);
     assert(o);
-    mcfg->mcfg_size = qnum_get_uint(qobject_to(QNum, o));
+    mcfg->size = qnum_get_uint(qobject_to(QNum, o));
     qobject_unref(o);
     return true;
 }
Signature · 1 line
2.19.1
author
#4 lore raw

PATCH v4 4/6 hw/arm/virt-acpi-build: pass AcpiMcfgInfo to build_mcfg()

In reply to Wei Yang

To build MCFG, two information is necessary:

* bus number
* base address

Abstract these two information to AcpiMcfgInfo so that build_mcfg and build_mcfg_q35 will have the same declaration.

---
v3:
* move AcpiMcfgInfo to pci.h
v2:
* for arm platform, construct a AcpiMcfgInfo directly

---
 hw/arm/virt-acpi-build.c | 18 +++++++++++-------
 hw/i386/acpi-build.c     |  6 +-----
 include/hw/acpi/pci.h    | 33 +++++++++++++++++++++++++++++++++
 3 files changed, 45 insertions(+), 12 deletions(-)
 create mode 100644 include/hw/acpi/pci.h
hw/arm/virt-acpi-build.c +11 -7
index e09e7eff8d..ebddcde596 100644
@@ -40,6 +40,7 @@
 #include "hw/hw.h"
 #include "hw/acpi/aml-build.h"
 #include "hw/acpi/utils.h"
+#include "hw/acpi/pci.h"
 #include "hw/pci/pcie_host.h"
 #include "hw/pci/pci.h"
 #include "hw/arm/virt.h"
@@ -546,21 +547,18 @@ build_srat(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
 }
 
 static void
-build_mcfg(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
+build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
 {
     AcpiTableMcfg *mcfg;
-    const MemMapEntry *memmap = vms->memmap;
-    int ecam_id = VIRT_ECAM_ID(vms->highmem_ecam);
     int len = sizeof(*mcfg) + sizeof(mcfg->allocation[0]);
 
     mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(memmap[ecam_id].base);
+    mcfg->allocation[0].address = cpu_to_le64(info->base);
 
     /* Only a single allocation so no need to play with segments */
     mcfg->allocation[0].pci_segment = cpu_to_le16(0);
     mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number =
-        PCIE_MMCFG_BUS(memmap[ecam_id].size - 1);
+    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
 
     build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
 }
@@ -801,7 +799,13 @@ void virt_acpi_build(VirtMachineState *vms, AcpiBuildTables *tables)
     build_gtdt(tables_blob, tables->linker, vms);
 
     acpi_add_table(table_offsets, tables_blob);
-    build_mcfg(tables_blob, tables->linker, vms);
+    {
+        AcpiMcfgInfo mcfg = {
+           .base = vms->memmap[VIRT_ECAM_ID(vms->highmem_ecam)].base,
+           .size = vms->memmap[VIRT_ECAM_ID(vms->highmem_ecam)].size,
+        };
+        build_mcfg(tables_blob, tables->linker, &mcfg);
+    }
 
     acpi_add_table(table_offsets, tables_blob);
     build_spcr(tables_blob, tables->linker, vms);
hw/i386/acpi-build.c +1 -5
index f0d27bffd6..c2de7f4b01 100644
@@ -58,6 +58,7 @@
 
 #include "hw/acpi/aml-build.h"
 #include "hw/acpi/utils.h"
+#include "hw/acpi/pci.h"
 
 #include "qom/qom-qobject.h"
 #include "hw/i386/amd_iommu.h"
@@ -86,11 +87,6 @@
 /* Default IOAPIC ID */
 #define ACPI_BUILD_IOAPIC_ID 0x0
 
-typedef struct AcpiMcfgInfo {
-    uint64_t base;
-    uint32_t size;
-} AcpiMcfgInfo;
-
 typedef struct AcpiPmInfo {
     bool s3_disabled;
     bool s4_disabled;
include/hw/acpi/pci.h new file +33 -0
new file mode 100644
index 0000000000..124af7d32a
@@ -0,0 +1,33 @@
+/*
+ * Support for generating PCI related ACPI tables and passing them to Guests
+ *
+ * Copyright (C) 2006 Fabrice Bellard
+ * Copyright (C) 2008-2010  Kevin O'Connor <[email protected]>
+ * Copyright (C) 2013-2019 Red Hat Inc
+ * Copyright (C) 2019 Intel Corporation
+ *
+ * Author: Wei Yang <[email protected]>
+ * 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/>.
+ */
+#ifndef HW_ACPI_PCI_H
+#define HW_ACPI_PCI_H
+
+typedef struct AcpiMcfgInfo {
+    uint64_t base;
+    uint32_t size;
+} AcpiMcfgInfo;
+
+#endif
Signature · 1 line
2.19.1
author
#5 lore raw

PATCH v4 5/6 hw/acpi: Consolidate build_mcfg to pci.c

In reply to Wei Yang

Now we have two identical build_mcfg functions.

Consolidate them in acpi/pci.c.

---
v3:
* adjust changelog based on Igor's suggestion

---
 default-configs/arm-softmmu.mak  |  1 +
 default-configs/i386-softmmu.mak |  1 +
 hw/acpi/Kconfig                  |  4 +++
 hw/acpi/Makefile.objs            |  1 +
 hw/acpi/pci.c                    | 46 ++++++++++++++++++++++++++++++++
 hw/arm/virt-acpi-build.c         | 17 ------------
 hw/i386/acpi-build.c             | 18 +------------
 include/hw/acpi/pci.h            |  1 +
 8 files changed, 55 insertions(+), 34 deletions(-)
 create mode 100644 hw/acpi/pci.c
default-configs/arm-softmmu.mak +1 -0
index 613d19a06d..8f2796e195 100644
@@ -144,6 +144,7 @@ CONFIG_XIO3130=y
 CONFIG_IOH3420=y
 CONFIG_I82801B11=y
 CONFIG_ACPI=y
+CONFIG_ACPI_PCI=y
 CONFIG_ARM_VIRT=y
 CONFIG_SMBIOS=y
 CONFIG_ASPEED_SOC=y
default-configs/i386-softmmu.mak +1 -0
index ba3fb3ff50..cd5ea391e8 100644
@@ -25,3 +25,4 @@
 CONFIG_ISAPC=y
 CONFIG_I440FX=y
 CONFIG_Q35=y
+CONFIG_ACPI_PCI=y
hw/acpi/Kconfig +4 -0
index eca3beed75..7265843cc3 100644
@@ -23,6 +23,10 @@ config ACPI_NVDIMM
     bool
     depends on ACPI
 
+config ACPI_PCI
+    bool
+    depends on ACPI
+
 config ACPI_VMGENID
     bool
     default y
hw/acpi/Makefile.objs +1 -0
index ba93c5b64a..9bb2101e3b 100644
@@ -11,6 +11,7 @@ common-obj-$(call lnot,$(CONFIG_ACPI_X86)) += acpi-stub.o
 common-obj-y += acpi_interface.o
 common-obj-y += bios-linker-loader.o
 common-obj-y += aml-build.o utils.o
+common-obj-$(CONFIG_ACPI_PCI) += pci.o
 common-obj-$(CONFIG_TPM) += tpm.o
 
 common-obj-$(CONFIG_IPMI) += ipmi.o
hw/acpi/pci.c new file +46 -0

46 lines not on this page · raw message

hw/arm/virt-acpi-build.c +0 -17
index ebddcde596..e3353de9e4 100644
@@ -546,23 +546,6 @@ build_srat(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
                  "SRAT", table_data->len - srat_start, 3, NULL, NULL);
 }
 
-static void
-build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
-{
-    AcpiTableMcfg *mcfg;
-    int len = sizeof(*mcfg) + sizeof(mcfg->allocation[0]);
-
-    mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(info->base);
-
-    /* Only a single allocation so no need to play with segments */
-    mcfg->allocation[0].pci_segment = cpu_to_le16(0);
-    mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
-
-    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
-}
-
 /* GTDT */
 static void
 build_gtdt(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
hw/i386/acpi-build.c +1 -17
index c2de7f4b01..29980bb3f4 100644
@@ -2393,22 +2393,6 @@ build_srat(GArray *table_data, BIOSLinker *linker, MachineState *machine)
                  table_data->len - srat_start, 1, NULL, NULL);
 }
 
-static void
-build_mcfg_q35(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
-{
-    AcpiTableMcfg *mcfg;
-    int len = sizeof(*mcfg) + 1 * sizeof(mcfg->allocation[0]);
-
-    mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(info->base);
-    /* Only a single allocation so no need to play with segments */
-    mcfg->allocation[0].pci_segment = cpu_to_le16(0);
-    mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
-
-    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
-}
-
 /*
  * VT-d spec 8.1 DMA Remapping Reporting Structure
  * (version Oct. 2014 or later)
@@ -2678,7 +2662,7 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
     }
     if (acpi_get_mcfg(&mcfg)) {
         acpi_add_table(table_offsets, tables_blob);
-        build_mcfg_q35(tables_blob, tables->linker, &mcfg);
+        build_mcfg(tables_blob, tables->linker, &mcfg);
     }
     if (x86_iommu_get_default()) {
         IommuType IOMMUType = x86_iommu_get_type();
include/hw/acpi/pci.h +1 -0
index 124af7d32a..8bbd32cf45 100644
@@ -30,4 +30,5 @@ typedef struct AcpiMcfgInfo {
     uint32_t size;
 } AcpiMcfgInfo;
 
+void build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info);
 #endif
Signature · 1 line
2.19.1
#6 lore raw

Re: [Qemu-arm] [PATCH v4 5/6] hw/acpi: Consolidate build_mcfg to pci.c

Quoted 49 lines · Wei Yang

Now we have two identical build_mcfg functions.

Consolidate them in acpi/pci.c.

---
v3:
* adjust changelog based on Igor's suggestion

---
 default-configs/arm-softmmu.mak  |  1 +
 default-configs/i386-softmmu.mak |  1 +
 hw/acpi/Kconfig                  |  4 +++
 hw/acpi/Makefile.objs            |  1 +
 hw/acpi/pci.c                    | 46 ++++++++++++++++++++++++++++++++
 hw/arm/virt-acpi-build.c         | 17 ------------
 hw/i386/acpi-build.c             | 18 +------------
 include/hw/acpi/pci.h            |  1 +
 8 files changed, 55 insertions(+), 34 deletions(-)
 create mode 100644 hw/acpi/pci.c
default-configs/arm-softmmu.mak +1 -0
index 613d19a06d..8f2796e195 100644
@@ -144,6 +144,7 @@ CONFIG_XIO3130=y
 CONFIG_IOH3420=y
 CONFIG_I82801B11=y
 CONFIG_ACPI=y
+CONFIG_ACPI_PCI=y
 CONFIG_ARM_VIRT=y
 CONFIG_SMBIOS=y
 CONFIG_ASPEED_SOC=y
default-configs/i386-softmmu.mak +1 -0
index ba3fb3ff50..cd5ea391e8 100644
@@ -25,3 +25,4 @@
 CONFIG_ISAPC=y
 CONFIG_I440FX=y
 CONFIG_Q35=y
+CONFIG_ACPI_PCI=y
hw/acpi/Kconfig +0 -0
index eca3beed75..7265843cc3 100644

hw/acpi/Kconfig config ACPI_NVDIMM · line 28

Quoted 6 lines
hw/acpi/Kconfig +3 -0
@@ -23,6 +23,10 @@ config ACPI_NVDIMM
     bool
     depends on ACPI
 
+config ACPI_PCI
+    bool
+    depends on ACPI

Shouldn't this be "depends on ACPI && PCI"?

Quoted 143 lines
hw/acpi/Kconfig +1 -0
+
 config ACPI_VMGENID
     bool
     default y
hw/acpi/Makefile.objs +1 -0
index ba93c5b64a..9bb2101e3b 100644
@@ -11,6 +11,7 @@ common-obj-$(call lnot,$(CONFIG_ACPI_X86)) += acpi-stub.o
 common-obj-y += acpi_interface.o
 common-obj-y += bios-linker-loader.o
 common-obj-y += aml-build.o utils.o
+common-obj-$(CONFIG_ACPI_PCI) += pci.o
 common-obj-$(CONFIG_TPM) += tpm.o
 
 common-obj-$(CONFIG_IPMI) += ipmi.o
hw/acpi/pci.c new file +46 -0
new file mode 100644
index 0000000000..fa0fa30bb9
@@ -0,0 +1,46 @@
+/*
+ * Support for generating PCI related ACPI tables and passing them to Guests
+ *
+ * Copyright (C) 2006 Fabrice Bellard
+ * Copyright (C) 2008-2010  Kevin O'Connor <[email protected]>
+ * Copyright (C) 2013-2019 Red Hat Inc
+ * Copyright (C) 2019 Intel Corporation
+ *
+ * Author: Wei Yang <[email protected]>
+ * 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 "hw/acpi/aml-build.h"
+#include "hw/acpi/pci.h"
+#include "hw/pci/pcie_host.h"
+
+void build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
+{
+    AcpiTableMcfg *mcfg;
+    int len = sizeof(*mcfg) + sizeof(mcfg->allocation[0]);
+
+    mcfg = acpi_data_push(table_data, len);
+    mcfg->allocation[0].address = cpu_to_le64(info->base);
+
+    /* Only a single allocation so no need to play with segments */
+    mcfg->allocation[0].pci_segment = cpu_to_le16(0);
+    mcfg->allocation[0].start_bus_number = 0;
+    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
+
+    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
+}
+
hw/arm/virt-acpi-build.c +0 -17
index ebddcde596..e3353de9e4 100644
@@ -546,23 +546,6 @@ build_srat(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
                  "SRAT", table_data->len - srat_start, 3, NULL, NULL);
 }
 
-static void
-build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
-{
-    AcpiTableMcfg *mcfg;
-    int len = sizeof(*mcfg) + sizeof(mcfg->allocation[0]);
-
-    mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(info->base);
-
-    /* Only a single allocation so no need to play with segments */
-    mcfg->allocation[0].pci_segment = cpu_to_le16(0);
-    mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
-
-    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
-}
-
 /* GTDT */
 static void
 build_gtdt(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
hw/i386/acpi-build.c +1 -17
index c2de7f4b01..29980bb3f4 100644
@@ -2393,22 +2393,6 @@ build_srat(GArray *table_data, BIOSLinker *linker, MachineState *machine)
                  table_data->len - srat_start, 1, NULL, NULL);
 }
 
-static void
-build_mcfg_q35(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
-{
-    AcpiTableMcfg *mcfg;
-    int len = sizeof(*mcfg) + 1 * sizeof(mcfg->allocation[0]);
-
-    mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(info->base);
-    /* Only a single allocation so no need to play with segments */
-    mcfg->allocation[0].pci_segment = cpu_to_le16(0);
-    mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
-
-    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
-}
-
 /*
  * VT-d spec 8.1 DMA Remapping Reporting Structure
  * (version Oct. 2014 or later)
@@ -2678,7 +2662,7 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
     }
     if (acpi_get_mcfg(&mcfg)) {
         acpi_add_table(table_offsets, tables_blob);
-        build_mcfg_q35(tables_blob, tables->linker, &mcfg);
+        build_mcfg(tables_blob, tables->linker, &mcfg);
     }
     if (x86_iommu_get_default()) {
         IommuType IOMMUType = x86_iommu_get_type();
include/hw/acpi/pci.h +1 -0
index 124af7d32a..8bbd32cf45 100644
@@ -30,4 +30,5 @@ typedef struct AcpiMcfgInfo {
     uint32_t size;
 } AcpiMcfgInfo;
 
+void build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info);
 #endif
author
#7 lore raw

Re: [Qemu-arm] [PATCH v4 5/6] hw/acpi: Consolidate build_mcfg to pci.c

Quoted 53 lines · Philippe Mathieu-Daudé
Quoted 52 lines · Wei Yang

Now we have two identical build_mcfg functions.

Consolidate them in acpi/pci.c.

---
v3:
* adjust changelog based on Igor's suggestion

---
 default-configs/arm-softmmu.mak  |  1 +
 default-configs/i386-softmmu.mak |  1 +
 hw/acpi/Kconfig                  |  4 +++
 hw/acpi/Makefile.objs            |  1 +
 hw/acpi/pci.c                    | 46 ++++++++++++++++++++++++++++++++
 hw/arm/virt-acpi-build.c         | 17 ------------
 hw/i386/acpi-build.c             | 18 +------------
 include/hw/acpi/pci.h            |  1 +
 8 files changed, 55 insertions(+), 34 deletions(-)
 create mode 100644 hw/acpi/pci.c
default-configs/arm-softmmu.mak +1 -0
index 613d19a06d..8f2796e195 100644
@@ -144,6 +144,7 @@ CONFIG_XIO3130=y
 CONFIG_IOH3420=y
 CONFIG_I82801B11=y
 CONFIG_ACPI=y
+CONFIG_ACPI_PCI=y
 CONFIG_ARM_VIRT=y
 CONFIG_SMBIOS=y
 CONFIG_ASPEED_SOC=y
default-configs/i386-softmmu.mak +1 -0
index ba3fb3ff50..cd5ea391e8 100644
@@ -25,3 +25,4 @@
 CONFIG_ISAPC=y
 CONFIG_I440FX=y
 CONFIG_Q35=y
+CONFIG_ACPI_PCI=y
hw/acpi/Kconfig +0 -0
index eca3beed75..7265843cc3 100644
@@ -23,6 +23,10 @@ config ACPI_NVDIMM
     bool
     depends on ACPI
 
Quoted 6 lines

hw/acpi/Kconfig config ACPI_NVDIMM · line 28

Quoted 3 lines
hw/acpi/Kconfig +3 -0
+config ACPI_PCI
+    bool
+    depends on ACPI

Shouldn't this be "depends on ACPI && PCI"?

I think you are right. Let me fix this.

Thanks

Signature · 2 lines
Wei Yang
Help you, Help me
author
#8 lore raw

PATCH v4 6/6 acpi: pci: use build_append_foo() API to construct MCFG

In reply to Wei Yang

build_append_foo() API doesn't need explicit endianness conversions which eliminates a source of errors and it makes build_mcfg() look like declarative definition of MCFG table in ACPI spec, which makes it easy to review.

---
 hw/acpi/pci.c               | 33 +++++++++++++++++++++------------
 include/hw/acpi/acpi-defs.h | 18 ------------------
 2 files changed, 21 insertions(+), 30 deletions(-)
hw/acpi/pci.c +21 -12
index fa0fa30bb9..341805e786 100644
@@ -30,17 +30,26 @@
 
 void build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
 {
-    AcpiTableMcfg *mcfg;
-    int len = sizeof(*mcfg) + sizeof(mcfg->allocation[0]);
-
-    mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(info->base);
-
-    /* Only a single allocation so no need to play with segments */
-    mcfg->allocation[0].pci_segment = cpu_to_le16(0);
-    mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
-
-    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
+    int mcfg_start = table_data->len;
+
+    acpi_data_push(table_data, sizeof(AcpiTableHeader));
+
+    /*
+     * PCI Firmware Specification, Revision 3.0
+     * 4.1.2 MCFG Table Description.
+     */
+    /* Base address, processor-relative */
+    build_append_int_noprefix(table_data, info->base, 8);
+    /* PCI segment group number */
+    build_append_int_noprefix(table_data, 0, 2);
+    /* Starting PCI Bus number */
+    build_append_int_noprefix(table_data, 0, 1);
+    /* Final PCI Bus number */
+    build_append_int_noprefix(table_data, PCIE_MMCFG_BUS(info->size - 1), 1);
+    /* Reserved */
+    build_append_int_noprefix(table_data, 0, 4);
+
+    build_header(linker, table_data, (void *)(table_data->data + mcfg_start),
+                 "MCFG", table_data->len - mcfg_start, 1, NULL, NULL);
 }
 
include/hw/acpi/acpi-defs.h +0 -18
index f9aa4bd398..57a3f58b0c 100644
@@ -449,24 +449,6 @@ struct AcpiSratProcessorGiccAffinity {
 
 typedef struct AcpiSratProcessorGiccAffinity AcpiSratProcessorGiccAffinity;
 
-/* PCI fw r3.0 MCFG table. */
-/* Subtable */
-struct AcpiMcfgAllocation {
-    uint64_t address;                /* Base address, processor-relative */
-    uint16_t pci_segment;            /* PCI segment group number */
-    uint8_t start_bus_number;       /* Starting PCI Bus number */
-    uint8_t end_bus_number;         /* Final PCI Bus number */
-    uint32_t reserved;
-} QEMU_PACKED;
-typedef struct AcpiMcfgAllocation AcpiMcfgAllocation;
-
-struct AcpiTableMcfg {
-    ACPI_TABLE_HEADER_DEF;
-    uint8_t reserved[8];
-    AcpiMcfgAllocation allocation[0];
-} QEMU_PACKED;
-typedef struct AcpiTableMcfg AcpiTableMcfg;
-
 /*
  * TCPA Description Table
  *
Signature · 1 line
2.19.1
#9 lore raw

Re: [Qemu-arm] [PATCH v4 6/6] acpi: pci: use build_append_foo() API to construct MCFG

Quoted 9 lines · Wei Yang

build_append_foo() API doesn't need explicit endianness conversions which eliminates a source of errors and it makes build_mcfg() look like declarative definition of MCFG table in ACPI spec, which makes it easy to review.

Causes a regression with an invalid MCFG produced. Dropped.

Quoted 79 lines
---
 hw/acpi/pci.c               | 33 +++++++++++++++++++++------------
 include/hw/acpi/acpi-defs.h | 18 ------------------
 2 files changed, 21 insertions(+), 30 deletions(-)
hw/acpi/pci.c +21 -12
index fa0fa30bb9..341805e786 100644
@@ -30,17 +30,26 @@
 
 void build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
 {
-    AcpiTableMcfg *mcfg;
-    int len = sizeof(*mcfg) + sizeof(mcfg->allocation[0]);
-
-    mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(info->base);
-
-    /* Only a single allocation so no need to play with segments */
-    mcfg->allocation[0].pci_segment = cpu_to_le16(0);
-    mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
-
-    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
+    int mcfg_start = table_data->len;
+
+    acpi_data_push(table_data, sizeof(AcpiTableHeader));
+
+    /*
+     * PCI Firmware Specification, Revision 3.0
+     * 4.1.2 MCFG Table Description.
+     */
+    /* Base address, processor-relative */
+    build_append_int_noprefix(table_data, info->base, 8);
+    /* PCI segment group number */
+    build_append_int_noprefix(table_data, 0, 2);
+    /* Starting PCI Bus number */
+    build_append_int_noprefix(table_data, 0, 1);
+    /* Final PCI Bus number */
+    build_append_int_noprefix(table_data, PCIE_MMCFG_BUS(info->size - 1), 1);
+    /* Reserved */
+    build_append_int_noprefix(table_data, 0, 4);
+
+    build_header(linker, table_data, (void *)(table_data->data + mcfg_start),
+                 "MCFG", table_data->len - mcfg_start, 1, NULL, NULL);
 }
 
include/hw/acpi/acpi-defs.h +0 -18
index f9aa4bd398..57a3f58b0c 100644
@@ -449,24 +449,6 @@ struct AcpiSratProcessorGiccAffinity {
 
 typedef struct AcpiSratProcessorGiccAffinity AcpiSratProcessorGiccAffinity;
 
-/* PCI fw r3.0 MCFG table. */
-/* Subtable */
-struct AcpiMcfgAllocation {
-    uint64_t address;                /* Base address, processor-relative */
-    uint16_t pci_segment;            /* PCI segment group number */
-    uint8_t start_bus_number;       /* Starting PCI Bus number */
-    uint8_t end_bus_number;         /* Final PCI Bus number */
-    uint32_t reserved;
-} QEMU_PACKED;
-typedef struct AcpiMcfgAllocation AcpiMcfgAllocation;
-
-struct AcpiTableMcfg {
-    ACPI_TABLE_HEADER_DEF;
-    uint8_t reserved[8];
-    AcpiMcfgAllocation allocation[0];
-} QEMU_PACKED;
-typedef struct AcpiTableMcfg AcpiTableMcfg;
-
 /*
  * TCPA Description Table
  *
Signature · 1 line
2.19.1
#10 lore raw

Re: [Qemu-arm] [PATCH v4 6/6] acpi: pci: use build_append_foo() API to construct MCFG

Quoted 34 lines · Michael S. Tsirkin
Quoted 9 lines · Wei Yang

build_append_foo() API doesn't need explicit endianness conversions which eliminates a source of errors and it makes build_mcfg() look like declarative definition of MCFG table in ACPI spec, which makes it easy to review.

Causes a regression with an invalid MCFG produced. Dropped.

Quoted 20 lines
---
 hw/acpi/pci.c               | 33 +++++++++++++++++++++------------
 include/hw/acpi/acpi-defs.h | 18 ------------------
 2 files changed, 21 insertions(+), 30 deletions(-)
hw/acpi/pci.c +0 -7
index fa0fa30bb9..341805e786 100644
@@ -30,17 +30,26 @@
 
 void build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
 {
-    AcpiTableMcfg *mcfg;
-    int len = sizeof(*mcfg) + sizeof(mcfg->allocation[0]);
-
-    mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(info->base);
-
-    /* Only a single allocation so no need to play with segments */
Quoted 8 lines

hw/acpi/pci.c line 35

Quoted 8 lines
hw/acpi/pci.c +3 -5
-    mcfg->allocation[0].pci_segment = cpu_to_le16(0);
-    mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
-
-    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
+    int mcfg_start = table_data->len;
+
+    acpi_data_push(table_data, sizeof(AcpiTableHeader));

Ah, it seems we missed AcpiTableMcfg.reserved[8]...:

        build_append_int_noprefix(table_data, 0, 8);
Quoted 28 lines
Quoted 28 lines
hw/acpi/pci.c +18 -0
+
+    /*
+     * PCI Firmware Specification, Revision 3.0
+     * 4.1.2 MCFG Table Description.
+     */
+    /* Base address, processor-relative */
+    build_append_int_noprefix(table_data, info->base, 8);
+    /* PCI segment group number */
+    build_append_int_noprefix(table_data, 0, 2);
+    /* Starting PCI Bus number */
+    build_append_int_noprefix(table_data, 0, 1);
+    /* Final PCI Bus number */
+    build_append_int_noprefix(table_data, PCIE_MMCFG_BUS(info->size - 1), 1);
+    /* Reserved */
+    build_append_int_noprefix(table_data, 0, 4);
+
+    build_header(linker, table_data, (void *)(table_data->data + mcfg_start),
+                 "MCFG", table_data->len - mcfg_start, 1, NULL, NULL);
 }
 
include/hw/acpi/acpi-defs.h +0 -0
index f9aa4bd398..57a3f58b0c 100644
@@ -449,24 +449,6 @@ struct AcpiSratProcessorGiccAffinity {
 
 typedef struct AcpiSratProcessorGiccAffinity AcpiSratProcessorGiccAffinity;
 
Quoted 14 lines
Quoted 4 lines
include/hw/acpi/acpi-defs.h +0 -4
-/* PCI fw r3.0 MCFG table. */
-/* Subtable */
-struct AcpiMcfgAllocation {
-    uint64_t address;                /* Base address, processor-relative */

include/hw/acpi/acpi-defs.h struct AcpiSratProcessorGiccAffinity { · line 465

Quoted 10 lines
include/hw/acpi/acpi-defs.h +0 -10
-    uint16_t pci_segment;            /* PCI segment group number */
-    uint8_t start_bus_number;       /* Starting PCI Bus number */
-    uint8_t end_bus_number;         /* Final PCI Bus number */
-    uint32_t reserved;
-} QEMU_PACKED;
-typedef struct AcpiMcfgAllocation AcpiMcfgAllocation;
-
-struct AcpiTableMcfg {
-    ACPI_TABLE_HEADER_DEF;
-    uint8_t reserved[8];
           --------^

Thanks Michael for testing...

Wei, can you add a MCFG test in tests/bios-tables-test.c?

Regards,

Phil.

Quoted 9 lines
Quoted 9 lines
include/hw/acpi/acpi-defs.h +0 -4
-    AcpiMcfgAllocation allocation[0];
-} QEMU_PACKED;
-typedef struct AcpiTableMcfg AcpiTableMcfg;
-
 /*
  * TCPA Description Table
  *
Signature · 1 line
2.19.1
author
#11 lore raw

Re: [Qemu-arm] [PATCH v4 6/6] acpi: pci: use build_append_foo() API to construct MCFG

Quoted 46 lines · Philippe Mathieu-Daudé
Quoted 34 lines · Michael S. Tsirkin
Quoted 9 lines · Wei Yang

build_append_foo() API doesn't need explicit endianness conversions which eliminates a source of errors and it makes build_mcfg() look like declarative definition of MCFG table in ACPI spec, which makes it easy to review.

Causes a regression with an invalid MCFG produced. Dropped.

Quoted 20 lines
---
 hw/acpi/pci.c               | 33 +++++++++++++++++++++------------
 include/hw/acpi/acpi-defs.h | 18 ------------------
 2 files changed, 21 insertions(+), 30 deletions(-)
hw/acpi/pci.c +0 -7
index fa0fa30bb9..341805e786 100644
@@ -30,17 +30,26 @@
 
 void build_mcfg(GArray *table_data, BIOSLinker *linker, AcpiMcfgInfo *info)
 {
-    AcpiTableMcfg *mcfg;
-    int len = sizeof(*mcfg) + sizeof(mcfg->allocation[0]);
-
-    mcfg = acpi_data_push(table_data, len);
-    mcfg->allocation[0].address = cpu_to_le64(info->base);
-
-    /* Only a single allocation so no need to play with segments */
Quoted 8 lines

hw/acpi/pci.c line 35

Quoted 8 lines
hw/acpi/pci.c +3 -5
-    mcfg->allocation[0].pci_segment = cpu_to_le16(0);
-    mcfg->allocation[0].start_bus_number = 0;
-    mcfg->allocation[0].end_bus_number = PCIE_MMCFG_BUS(info->size - 1);
-
-    build_header(linker, table_data, (void *)mcfg, "MCFG", len, 1, NULL, NULL);
+    int mcfg_start = table_data->len;
+
+    acpi_data_push(table_data, sizeof(AcpiTableHeader));

Ah, it seems we missed AcpiTableMcfg.reserved[8]...:

Quoted 2 lines
       build_append_int_noprefix(table_data, 0, 8);

Hmm... you are right. Shame on me.

Quoted 45 lines
Quoted 28 lines
Quoted 28 lines
hw/acpi/pci.c +18 -0
+
+    /*
+     * PCI Firmware Specification, Revision 3.0
+     * 4.1.2 MCFG Table Description.
+     */
+    /* Base address, processor-relative */
+    build_append_int_noprefix(table_data, info->base, 8);
+    /* PCI segment group number */
+    build_append_int_noprefix(table_data, 0, 2);
+    /* Starting PCI Bus number */
+    build_append_int_noprefix(table_data, 0, 1);
+    /* Final PCI Bus number */
+    build_append_int_noprefix(table_data, PCIE_MMCFG_BUS(info->size - 1), 1);
+    /* Reserved */
+    build_append_int_noprefix(table_data, 0, 4);
+
+    build_header(linker, table_data, (void *)(table_data->data + mcfg_start),
+                 "MCFG", table_data->len - mcfg_start, 1, NULL, NULL);
 }
 
include/hw/acpi/acpi-defs.h +0 -0
index f9aa4bd398..57a3f58b0c 100644
@@ -449,24 +449,6 @@ struct AcpiSratProcessorGiccAffinity {
 
 typedef struct AcpiSratProcessorGiccAffinity AcpiSratProcessorGiccAffinity;
 
Quoted 14 lines
Quoted 4 lines
include/hw/acpi/acpi-defs.h +0 -4
-/* PCI fw r3.0 MCFG table. */
-/* Subtable */
-struct AcpiMcfgAllocation {
-    uint64_t address;                /* Base address, processor-relative */

include/hw/acpi/acpi-defs.h struct AcpiSratProcessorGiccAffinity { · line 465

Quoted 10 lines
include/hw/acpi/acpi-defs.h +0 -10
-    uint16_t pci_segment;            /* PCI segment group number */
-    uint8_t start_bus_number;       /* Starting PCI Bus number */
-    uint8_t end_bus_number;         /* Final PCI Bus number */
-    uint32_t reserved;
-} QEMU_PACKED;
-typedef struct AcpiMcfgAllocation AcpiMcfgAllocation;
-
-struct AcpiTableMcfg {
-    ACPI_TABLE_HEADER_DEF;
-    uint8_t reserved[8];
          --------^
Quoted 4 lines

Thanks Michael for testing...

Wei, can you add a MCFG test in tests/bios-tables-test.c?

Sure, let me prepare the test.

Quoted 13 lines

Regards,

Phil.

Quoted 9 lines
Quoted 9 lines
include/hw/acpi/acpi-defs.h +0 -4
-    AcpiMcfgAllocation allocation[0];
-} QEMU_PACKED;
-typedef struct AcpiTableMcfg AcpiTableMcfg;
-
 /*
  * TCPA Description Table
  *
Signature · 1 line
2.19.1
Signature · 2 lines
Wei Yang
Help you, Help me
author
#12 lore raw

Re: [Qemu-arm] [PATCH v4 6/6] acpi: pci: use build_append_foo() API to construct MCFG

In reply to Philippe Mathieu-Daudé
Quoted 5 lines · Philippe Mathieu-Daudé

Thanks Michael for testing...

Wei, can you add a MCFG test in tests/bios-tables-test.c?

I took a look into the test, current q35 has already has a reference MCFG in tests/data/acpi/q35/MCFG.

And there would be a warning message when reserved[8] is missed.

    /x86_64/acpi/q35/bridge: acpi-test: Warning! MCFG mismatch.

Is this enough? Or what more information prefer to add?

Quoted 13 lines

Regards,

Phil.

Quoted 9 lines
Quoted 9 lines
include/hw/acpi/acpi-defs.h +0 -4
-    AcpiMcfgAllocation allocation[0];
-} QEMU_PACKED;
-typedef struct AcpiTableMcfg AcpiTableMcfg;
-
 /*
  * TCPA Description Table
  *
Signature · 1 line
2.19.1
Signature · 2 lines
Wei Yang
Help you, Help me
#13 lore raw

Re: [Qemu-devel] [PATCH v4 6/6] acpi: pci: use build_append_foo() API to construct MCFG

Quoted 8 lines · Wei Yang
Quoted 5 lines · Philippe Mathieu-Daudé

Thanks Michael for testing...

Wei, can you add a MCFG test in tests/bios-tables-test.c?

Quoted 8 lines

I took a look into the test, current q35 has already has a reference MCFG in tests/data/acpi/q35/MCFG.

And there would be a warning message when reserved[8] is missed.

    /x86_64/acpi/q35/bridge: acpi-test: Warning! MCFG mismatch.

Is this enough? Or what more information prefer to add?

Well, the test has to fail for any mismatch (not a simple warning).

A mismatch failure seems to be enough IMHO.

Quoted 13 lines
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
include/hw/acpi/acpi-defs.h +0 -5
-    AcpiMcfgAllocation allocation[0];
-} QEMU_PACKED;
-typedef struct AcpiTableMcfg AcpiTableMcfg;
-
 /*
  * TCPA Description Table
  *
--

2.19.1

--
Wei Yang
Help you, Help me

#14 lore raw

Re: [Qemu-arm] [Qemu-devel] [PATCH v4 6/6] acpi: pci: use build_append_foo() API to construct MCFG

Quoted 14 lines · On Thu, 16 May 2019 13:01:31 +0200 Philippe Mathieu-Daudé <[email protected]>…
Quoted 13 lines · Wei Yang
Quoted 5 lines · Philippe Mathieu-Daudé

Thanks Michael for testing...

Wei, can you add a MCFG test in tests/bios-tables-test.c?

I took a look into the test, current q35 has already has a reference MCFG in tests/data/acpi/q35/MCFG.

And there would be a warning message when reserved[8] is missed.

Quoted 7 lines
Quoted 3 lines
    /x86_64/acpi/q35/bridge: acpi-test: Warning! MCFG mismatch.

Is this enough? Or what more information prefer to add?

Well, the test has to fail for any mismatch (not a simple warning).

A mismatch failure seems to be enough IMHO.

Warning is sufficient, we do not fail ACPI tests on mismatch. It was a policy decision for APCI tests as far as I remember. We might reconsider it in the future but it shouldn't affect this patch.

Quoted 15 lines
Quoted 14 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
include/hw/acpi/acpi-defs.h +0 -5
-    AcpiMcfgAllocation allocation[0];
-} QEMU_PACKED;
-typedef struct AcpiTableMcfg AcpiTableMcfg;
-
 /*
  * TCPA Description Table
  *
--

2.19.1

--
Wei Yang
Help you, Help me

#15 lore raw

Re: [Qemu-arm] [Qemu-devel] [PATCH v4 6/6] acpi: pci: use build_append_foo() API to construct MCFG

Quoted 19 lines · Igor Mammedov
Quoted 16 lines · On Thu, 16 May 2019 13:01:31 +0200 Philippe Mathieu-Daudé <[email protected]>…
Quoted 15 lines · Wei Yang
Quoted 5 lines · Philippe Mathieu-Daudé

Thanks Michael for testing...

Wei, can you add a MCFG test in tests/bios-tables-test.c?

I took a look into the test, current q35 has already has a reference MCFG in tests/data/acpi/q35/MCFG.

And there would be a warning message when reserved[8] is missed.

    /x86_64/acpi/q35/bridge: acpi-test: Warning! MCFG mismatch.
Quoted 8 lines
Quoted 5 lines
Quoted 1 line

Is this enough? Or what more information prefer to add?

Well, the test has to fail for any mismatch (not a simple warning).

A mismatch failure seems to be enough IMHO.

Warning is sufficient, we do not fail ACPI tests on mismatch. It was a policy decision for APCI tests as far as I remember. We might reconsider it in the future but it shouldn't affect this patch.

Yes. And the reason is that conflicts in binary expected files are impossible to resolve. So it's important that we can fix expected files after a patch that changes them.

I actually have an idea for a better way to fix this:
a special list of "warn on mismatch" files.

A patch changing tables will add the changed tables to the list. Then maintainer knows to inspec the diff manually
and re-generate expected files, and remove the
changed tables from the list.

Another thing we should do is drop dependency on IASL:

if IASL is present we should use it to show diff to simplify debugging but at this point a verbatim difference is good enough if IASL is not installed.

And I agree 100%: all this is a subject for a separate patch(set).

Quoted 16 lines
Quoted 16 lines
Quoted 15 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
include/hw/acpi/acpi-defs.h +0 -5
-    AcpiMcfgAllocation allocation[0];
-} QEMU_PACKED;
-typedef struct AcpiTableMcfg AcpiTableMcfg;
-
 /*
  * TCPA Description Table
  *
--

2.19.1

--
Wei Yang
Help you, Help me

author
#16 lore raw

Re: [Qemu-arm] [PATCH v4 6/6] acpi: pci: use build_append_foo() API to construct MCFG

In reply to Michael S. Tsirkin
Quoted 6 lines · Michael S. Tsirkin
Quoted 5 lines · Wei Yang

build_append_foo() API doesn't need explicit endianness conversions which eliminates a source of errors and it makes build_mcfg() look like declarative definition of MCFG table in ACPI spec, which makes it easy to review.

Quoted 8 lines
Quoted 4 lines

Causes a regression with an invalid MCFG produced. Dropped.

Any more information about this regression? How to reproduce it?

Signature · 2 lines
Wei Yang
Help you, Help me
#17 lore raw
In reply to Wei Yang

Hi, Igor

You would take this one? Or what should I do next?

Quoted 36 lines · Wei Yang

This patch set tries to generalize MCFG table build process. And it is based on one un-merged patch from Igor, which is included in this serials.

v3->v4:
* adjust comment to give more information about MCFG table

v2->v3:
* Includes the un-merged patch from Igor
* use build_append_foo() API to construct MCFG

Igor Mammedov (1):
q35: acpi: do not create dummy MCFG table

Wei Yang (5):
hw/arm/virt-acpi-build: remove unnecessary variable mcfg_start i386, acpi: remove mcfg_ prefix in AcpiMcfgInfo members hw/arm/virt-acpi-build: pass AcpiMcfgInfo to build_mcfg() hw/acpi: Consolidate build_mcfg to pci.c
acpi: pci: use build_append_foo() API to construct MCFG

default-configs/arm-softmmu.mak  |  1 +
default-configs/i386-softmmu.mak |  1 +
hw/acpi/Kconfig                  |  4 +++
hw/acpi/Makefile.objs            |  1 +
hw/acpi/pci.c                    | 55 ++++++++++++++++++++++++++++++++
hw/arm/virt-acpi-build.c         | 31 +++++-------------
hw/i386/acpi-build.c             | 44 ++++---------------------
include/hw/acpi/acpi-defs.h      | 18 -----------
include/hw/acpi/pci.h            | 34 ++++++++++++++++++++
9 files changed, 111 insertions(+), 78 deletions(-)

create mode 100644 hw/acpi/pci.c
create mode 100644 include/hw/acpi/pci.h

Signature · 1 line
2.19.1
Signature · 2 lines
Wei Yang
Help you, Help me
#18 lore raw

I merged this will send pull request soon.

Quoted 45 lines · Wei Yang

Hi, Igor

You would take this one? Or what should I do next?

Quoted 36 lines · Wei Yang

This patch set tries to generalize MCFG table build process. And it is based on one un-merged patch from Igor, which is included in this serials.

v3->v4:
* adjust comment to give more information about MCFG table

v2->v3:
* Includes the un-merged patch from Igor
* use build_append_foo() API to construct MCFG

Igor Mammedov (1):
q35: acpi: do not create dummy MCFG table

Wei Yang (5):
hw/arm/virt-acpi-build: remove unnecessary variable mcfg_start i386, acpi: remove mcfg_ prefix in AcpiMcfgInfo members hw/arm/virt-acpi-build: pass AcpiMcfgInfo to build_mcfg() hw/acpi: Consolidate build_mcfg to pci.c
acpi: pci: use build_append_foo() API to construct MCFG

default-configs/arm-softmmu.mak  |  1 +
default-configs/i386-softmmu.mak |  1 +
hw/acpi/Kconfig                  |  4 +++
hw/acpi/Makefile.objs            |  1 +
hw/acpi/pci.c                    | 55 ++++++++++++++++++++++++++++++++
hw/arm/virt-acpi-build.c         | 31 +++++-------------
hw/i386/acpi-build.c             | 44 ++++---------------------
include/hw/acpi/acpi-defs.h      | 18 -----------
include/hw/acpi/pci.h            | 34 ++++++++++++++++++++
9 files changed, 111 insertions(+), 78 deletions(-)

create mode 100644 hw/acpi/pci.c
create mode 100644 include/hw/acpi/pci.h

Signature · 1 line
2.19.1
Signature · 2 lines
Wei Yang
Help you, Help me
#19 lore raw
Quoted 2 lines · Michael S. Tsirkin

I merged this will send pull request soon.

Ah, Thanks :-)

Signature · 2 lines
Wei Yang
Help you, Help me