Skip to contentSkip to conversation

Introduce model for IBM's FSI

PATCH Superseded Ninad Palsule 5 participants 30 messages Oct 21, 2023 – Oct 27, 2023

Subsystems (2)

v6 of 12 Superseded Latest is v12 →

lore
Share & tools
lore
#0 lore raw

v6 cover Introduce model for IBM's FSI

Hello,

Please review the patch-set version 6.
I have incorporated review comments from Cedric, Daniel and Thomas.

Ninad Palsule (10):
hw/fsi: Introduce IBM's Local bus
hw/fsi: Introduce IBM's scratchpad
hw/fsi: Introduce IBM's cfam,fsi-slave
hw/fsi: Introduce IBM's FSI
hw/fsi: IBM's On-chip Peripheral Bus
hw/fsi: Aspeed APB2OPB interface
hw/arm: Hook up FSI module in AST2600
hw/fsi: Added qtest
hw/fsi: Added FSI documentation
hw/fsi: Update MAINTAINER list

 MAINTAINERS                        |   8 +
 docs/specs/fsi.rst                 | 141 +++++++++++++++
 meson.build                        |   1 +
 hw/fsi/trace.h                     |   1 +
 include/hw/arm/aspeed_soc.h        |   4 +
 include/hw/fsi/aspeed-apb2opb.h    |  33 ++++
 include/hw/fsi/cfam.h              |  34 ++++
 include/hw/fsi/engine-scratchpad.h |  32 ++++
 include/hw/fsi/fsi-master.h        |  30 ++++
 include/hw/fsi/fsi-slave.h         |  29 +++
 include/hw/fsi/fsi.h               |  37 ++++
 include/hw/fsi/lbus.h              |  43 +++++
 include/hw/fsi/opb.h               |  43 +++++
 hw/arm/aspeed_ast2600.c            |  19 ++
 hw/fsi/aspeed-apb2opb.c            | 280 +++++++++++++++++++++++++++++
 hw/fsi/cfam.c                      | 173 ++++++++++++++++++
 hw/fsi/engine-scratchpad.c         |  93 ++++++++++
 hw/fsi/fsi-master.c                | 201 +++++++++++++++++++++
 hw/fsi/fsi-slave.c                 |  96 ++++++++++
 hw/fsi/fsi.c                       |  25 +++
 hw/fsi/lbus.c                      |  76 ++++++++
 hw/fsi/opb.c                       | 164 +++++++++++++++++
 tests/qtest/fsi-test.c             | 207 +++++++++++++++++++++
 hw/Kconfig                         |   1 +
 hw/arm/Kconfig                     |   1 +
 hw/fsi/Kconfig                     |  23 +++
 hw/fsi/meson.build                 |   6 +
 hw/fsi/trace-events                |  21 +++
 hw/meson.build                     |   1 +
 tests/qtest/meson.build            |   1 +
 30 files changed, 1824 insertions(+)
 create mode 100644 docs/specs/fsi.rst
 create mode 100644 hw/fsi/trace.h
 create mode 100644 include/hw/fsi/aspeed-apb2opb.h
 create mode 100644 include/hw/fsi/cfam.h
 create mode 100644 include/hw/fsi/engine-scratchpad.h
 create mode 100644 include/hw/fsi/fsi-master.h
 create mode 100644 include/hw/fsi/fsi-slave.h
 create mode 100644 include/hw/fsi/fsi.h
 create mode 100644 include/hw/fsi/lbus.h
 create mode 100644 include/hw/fsi/opb.h
 create mode 100644 hw/fsi/aspeed-apb2opb.c
 create mode 100644 hw/fsi/cfam.c
 create mode 100644 hw/fsi/engine-scratchpad.c
 create mode 100644 hw/fsi/fsi-master.c
 create mode 100644 hw/fsi/fsi-slave.c
 create mode 100644 hw/fsi/fsi.c
 create mode 100644 hw/fsi/lbus.c
 create mode 100644 hw/fsi/opb.c
 create mode 100644 tests/qtest/fsi-test.c
 create mode 100644 hw/fsi/Kconfig
 create mode 100644 hw/fsi/meson.build
 create mode 100644 hw/fsi/trace-events
Signature · 1 line
2.39.2
#1 lore raw

PATCH v6 1/10 hw/fsi: Introduce IBM's Local bus

This is a part of patchset where IBM's Flexible Service Interface is introduced.

The LBUS is modelled to maintain the qdev bus hierarchy and to take advantage of the object model to automatically generate the CFAM configuration block. The configuration block presents engines in the order they are attached to the CFAM's LBUS. Engine implementations should subclass the LBusDevice and set the 'config' member of LBusDeviceClass to match the engine's type.

---
v2:
- Incorporated Joel's review comments.
v5:
- Incorporated review comments by Cedric. v6:
- Incorporated review comments by Cedric & Daniel.

---
 include/hw/fsi/lbus.h | 43 ++++++++++++++++++++++++
 hw/fsi/lbus.c         | 76 +++++++++++++++++++++++++++++++++++++++++++
 hw/Kconfig            |  1 +
 hw/fsi/Kconfig        |  2 ++
 hw/fsi/meson.build    |  1 +
 hw/meson.build        |  1 +
 6 files changed, 124 insertions(+)
 create mode 100644 include/hw/fsi/lbus.h
 create mode 100644 hw/fsi/lbus.c
 create mode 100644 hw/fsi/Kconfig
 create mode 100644 hw/fsi/meson.build
include/hw/fsi/lbus.h new file +43 -0
new file mode 100644
index 0000000000..4fa696bbdb
@@ -0,0 +1,43 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Local bus and connected device structures.
+ */
+#ifndef FSI_LBUS_H
+#define FSI_LBUS_H
+
+#include "exec/memory.h"
+#include "hw/qdev-core.h"
+
+#define TYPE_FSI_LBUS_DEVICE "fsi.lbus.device"
+OBJECT_DECLARE_TYPE(FSILBusDevice, FSILBusDeviceClass, FSI_LBUS_DEVICE)
+
+#define FSI_LBUS_MEM_REGION_SIZE  (2 * 1024 * 1024)
+#define FSI_LBUSDEV_IOMEM_SIZE    0x400
+
+typedef struct FSILBusDevice {
+    DeviceState parent;
+
+    MemoryRegion iomem;
+    uint32_t address;
+} FSILBusDevice;
+
+typedef struct FSILBusDeviceClass {
+    DeviceClass parent;
+
+    uint32_t config;
+} FSILBusDeviceClass;
+
+#define TYPE_FSI_LBUS "fsi.lbus"
+OBJECT_DECLARE_SIMPLE_TYPE(FSILBus, FSI_LBUS)
+
+typedef struct FSILBus {
+    BusState bus;
+
+    MemoryRegion mr;
+} FSILBus;
+
+DeviceState *lbus_create_device(FSILBus *bus, const char *type, uint32_t addr);
+int lbus_add_device(FSILBus *bus, FSILBusDevice *dev);
+#endif /* FSI_LBUS_H */
hw/fsi/lbus.c new file +76 -0

76 lines not on this page · raw message

hw/Kconfig +1 -0
index 9ca7b38c31..2c00936c28 100644
@@ -9,6 +9,7 @@ source core/Kconfig
 source cxl/Kconfig
 source display/Kconfig
 source dma/Kconfig
+source fsi/Kconfig
 source gpio/Kconfig
 source hyperv/Kconfig
 source i2c/Kconfig
hw/fsi/Kconfig new file +2 -0
new file mode 100644
index 0000000000..e650c660f0
@@ -0,0 +1,2 @@
+config FSI_LBUS
+    bool
hw/fsi/meson.build new file +1 -0
new file mode 100644
index 0000000000..4074d3a7d2
@@ -0,0 +1 @@
+system_ss.add(when: 'CONFIG_FSI_LBUS', if_true: files('lbus.c'))
hw/meson.build +1 -0
index f01fac4617..463d702683 100644
@@ -44,6 +44,7 @@ subdir('virtio')
 subdir('watchdog')
 subdir('xen')
 subdir('xenpv')
+subdir('fsi')
 
 subdir('alpha')
 subdir('arm')
Signature · 1 line
2.39.2
#2 lore raw

Re: [PATCH v6 01/10] hw/fsi: Introduce IBM's Local bus

Hi Ninad and Andrew,

Quoted 34 lines · Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

The LBUS is modelled to maintain the qdev bus hierarchy and to take advantage of the object model to automatically generate the CFAM configuration block. The configuration block presents engines in the order they are attached to the CFAM's LBUS. Engine implementations should subclass the LBusDevice and set the 'config' member of LBusDeviceClass to match the engine's type.

---
v2:
- Incorporated Joel's review comments.
v5:
- Incorporated review comments by Cedric. v6:
- Incorporated review comments by Cedric & Daniel.

---
  include/hw/fsi/lbus.h | 43 ++++++++++++++++++++++++
  hw/fsi/lbus.c         | 76 +++++++++++++++++++++++++++++++++++++++++++
  hw/Kconfig            |  1 +
  hw/fsi/Kconfig        |  2 ++
  hw/fsi/meson.build    |  1 +
  hw/meson.build        |  1 +
  6 files changed, 124 insertions(+)

create mode 100644 include/hw/fsi/lbus.h
create mode 100644 hw/fsi/lbus.c
create mode 100644 hw/fsi/Kconfig
create mode 100644 hw/fsi/meson.build

include/hw/fsi/lbus.h new file +1 -0
+#define TYPE_FSI_LBUS_DEVICE "fsi.lbus.device"

include/hw/fsi/lbus.h line 23

Quoted 10 lines
include/hw/fsi/lbus.h new file +10 -0
+OBJECT_DECLARE_TYPE(FSILBusDevice, FSILBusDeviceClass, FSI_LBUS_DEVICE)
+
+#define FSI_LBUS_MEM_REGION_SIZE  (2 * 1024 * 1024)
+#define FSI_LBUSDEV_IOMEM_SIZE    0x400
+
+typedef struct FSILBusDevice {
+    DeviceState parent;
+
+    MemoryRegion iomem;
+    uint32_t address;

[1] 32-bit address,

Quoted 30 lines
(quoted diff) +2 -0
+} FSILBusDevice;
+
hw/fsi/lbus.c new file +20 -0
new file mode 100644
index 0000000000..50d926dbe2
@@ -0,0 +1,76 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Local bus where FSI slaves are connected
+ */
+
+#include "qemu/osdep.h"
+#include "qapi/error.h"
+#include "hw/fsi/lbus.h"
+
+#include "hw/qdev-properties.h"
+
+static void lbus_init(Object *o)
+{
+    FSILBus *lbus = FSI_LBUS(o);
+
+    memory_region_init(&lbus->mr, OBJECT(lbus), TYPE_FSI_LBUS,
+                       FSI_LBUS_MEM_REGION_SIZE - FSI_LBUSDEV_IOMEM_SIZE);
+}

hw/fsi/lbus.c line 30

Quoted 10 lines
hw/fsi/lbus.c new file +10 -0
+
+static const TypeInfo lbus_info = {
+    .name = TYPE_FSI_LBUS,
+    .parent = TYPE_BUS,
+    .instance_init = lbus_init,
+    .instance_size = sizeof(FSILBus),
+};
+
+static Property lbus_device_props[] = {
+    DEFINE_PROP_UINT32("address", FSILBusDevice, address, 0),

[2] 32-bit address,

hw/fsi/lbus.c line 41

Quoted 11 lines
hw/fsi/lbus.c new file +11 -0
+    DEFINE_PROP_END_OF_LIST(),
+};
+
+DeviceState *lbus_create_device(FSILBus *bus, const char *type, uint32_t addr)
+{
+    DeviceState *ds;
+    BusState *state = BUS(bus);
+    FSILBusDevice *dev;
+
+    ds = qdev_new(type);
+    qdev_prop_set_uint8(ds, "address", addr);

[3] set 8-bit address but [1] and [2] declare as 32-bit.

hw/fsi/lbus.c line 42

Quoted 1 line
hw/fsi/lbus.c new file +1 -0
+    qdev_realize_and_unref(ds, state, &error_fatal);

If you pass the bus as argument here, ...

hw/fsi/lbus.c line 48

Quoted 6 lines
hw/fsi/lbus.c new file +6 -0
+
+    dev = FSI_LBUS_DEVICE(ds);
+    memory_region_add_subregion(&bus->mr, dev->address,
+                                &dev->iomem);
+
+    qdev_set_parent_bus(ds, state, &error_abort);

... why do you need that call here?

hw/fsi/lbus.c line 51

Quoted 3 lines
hw/fsi/lbus.c new file +3 -0
+
+    return ds;
+}

Thanks,

Phil.

#3 lore raw

Re: [PATCH v6 01/10] hw/fsi: Introduce IBM's Local bus

Hello Philippe,

Quoted 94 lines · Philippe Mathieu-Daudé

Hi Ninad and Andrew,

Quoted 34 lines · Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

The LBUS is modelled to maintain the qdev bus hierarchy and to take advantage of the object model to automatically generate the CFAM configuration block. The configuration block presents engines in the order they are attached to the CFAM's LBUS. Engine implementations should subclass the LBusDevice and set the 'config' member of LBusDeviceClass to match the engine's type.

---
v2:
- Incorporated Joel's review comments.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric & Daniel.
---
  include/hw/fsi/lbus.h | 43 ++++++++++++++++++++++++ hw/fsi/lbus.c         | 76 +++++++++++++++++++++++++++++++++++++++++++ hw/Kconfig            |  1 +
  hw/fsi/Kconfig        |  2 ++
  hw/fsi/meson.build    |  1 +
  hw/meson.build        |  1 +

  6 files changed, 124 insertions(+)

  create mode 100644 include/hw/fsi/lbus.h
  create mode 100644 hw/fsi/lbus.c
  create mode 100644 hw/fsi/Kconfig
  create mode 100644 hw/fsi/meson.build

include/hw/fsi/lbus.h new file +1 -0
+#define TYPE_FSI_LBUS_DEVICE "fsi.lbus.device"

include/hw/fsi/lbus.h line 23

Quoted 10 lines
include/hw/fsi/lbus.h new file +10 -0
+OBJECT_DECLARE_TYPE(FSILBusDevice, FSILBusDeviceClass, FSI_LBUS_DEVICE)
+
+#define FSI_LBUS_MEM_REGION_SIZE  (2 * 1024 * 1024)
+#define FSI_LBUSDEV_IOMEM_SIZE    0x400
+
+typedef struct FSILBusDevice {
+    DeviceState parent;
+
+    MemoryRegion iomem;
+    uint32_t address;

[1] 32-bit address,

Quoted 31 lines
(quoted diff) +2 -0
+} FSILBusDevice;
+
hw/fsi/lbus.c new file +19 -0
new file mode 100644
index 0000000000..50d926dbe2
@@ -0,0 +1,76 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Local bus where FSI slaves are connected
+ */
+
+#include "qemu/osdep.h"
+#include "qapi/error.h"
+#include "hw/fsi/lbus.h"
+
+#include "hw/qdev-properties.h"
+
+static void lbus_init(Object *o)
+{
+    FSILBus *lbus = FSI_LBUS(o);
+
+    memory_region_init(&lbus->mr, OBJECT(lbus), TYPE_FSI_LBUS,
+                       FSI_LBUS_MEM_REGION_SIZE - 

FSI_LBUSDEV_IOMEM_SIZE);

hw/fsi/lbus.c new file +1 -0
+}

hw/fsi/lbus.c line 30

Quoted 10 lines
hw/fsi/lbus.c new file +10 -0
+
+static const TypeInfo lbus_info = {
+    .name = TYPE_FSI_LBUS,
+    .parent = TYPE_BUS,
+    .instance_init = lbus_init,
+    .instance_size = sizeof(FSILBus),
+};
+
+static Property lbus_device_props[] = {
+    DEFINE_PROP_UINT32("address", FSILBusDevice, address, 0),

[2] 32-bit address,

Quoted 14 lines

hw/fsi/lbus.c line 41

Quoted 12 lines

+    DEFINE_PROP_END_OF_LIST(),

hw/fsi/lbus.c new file +3 -0
+};
+
+DeviceState *lbus_create_device(FSILBus *bus, const char *type, 

uint32_t addr)

hw/fsi/lbus.c new file +7 -0
+{
+    DeviceState *ds;
+    BusState *state = BUS(bus);
+    FSILBusDevice *dev;
+
+    ds = qdev_new(type);
+    qdev_prop_set_uint8(ds, "address", addr);

[3] set 8-bit address but [1] and [2] declare as 32-bit.

Good catch. Fixed it.

Quoted 5 lines
Quoted 2 lines

+    qdev_realize_and_unref(ds, state, &error_fatal);

If you pass the bus as argument here, ...

Quoted 8 lines

hw/fsi/lbus.c line 48

Quoted 6 lines
hw/fsi/lbus.c new file +6 -0
+
+    dev = FSI_LBUS_DEVICE(ds);
+    memory_region_add_subregion(&bus->mr, dev->address,
+                                &dev->iomem);
+
+    qdev_set_parent_bus(ds, state, &error_abort);

... why do you need that call here?

You are right. qdev_realize() call qdev_set_parent_bus() so I removed this call. I will publish the change in version 7.

Thanks for the review.

Regards,

Ninad

Quoted 8 lines

hw/fsi/lbus.c line 51

Quoted 4 lines
hw/fsi/lbus.c new file +3 -0
+
+    return ds;
+}

Thanks,

Phil.

#4 lore raw

PATCH v6 2/10 hw/fsi: Introduce IBM's scratchpad

In reply to Ninad Palsule

This is a part of patchset where scratchpad is introduced.

The scratchpad provides a set of non-functional registers. The firmware is free to use them, hardware does not support any special management support. The scratchpad registers can be read or written from LBUS slave.

In this model, The LBUS device is parent for the scratchpad.

---
v2:
- Incorporated Joel's review comments.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Daniel.

---
 meson.build                        |  1 +
 hw/fsi/trace.h                     |  1 +
 include/hw/fsi/engine-scratchpad.h | 32 ++++++++++
 include/hw/fsi/fsi.h               | 16 +++++
 hw/fsi/engine-scratchpad.c         | 93 ++++++++++++++++++++++++++++++
 hw/fsi/Kconfig                     |  4 ++
 hw/fsi/meson.build                 |  1 +
 hw/fsi/trace-events                |  2 +
 8 files changed, 150 insertions(+)
 create mode 100644 hw/fsi/trace.h
 create mode 100644 include/hw/fsi/engine-scratchpad.h
 create mode 100644 include/hw/fsi/fsi.h
 create mode 100644 hw/fsi/engine-scratchpad.c
 create mode 100644 hw/fsi/trace-events
meson.build +1 -0
index 4961c82a6b..31078f0bb3 100644
@@ -3256,6 +3256,7 @@ if have_system
     'hw/char',
     'hw/display',
     'hw/dma',
+    'hw/fsi',
     'hw/hyperv',
     'hw/i2c',
     'hw/i386',
hw/fsi/trace.h new file +1 -0
new file mode 100644
index 0000000000..ee67c7fb04
@@ -0,0 +1 @@
+#include "trace/trace-hw_fsi.h"
include/hw/fsi/engine-scratchpad.h new file +32 -0
new file mode 100644
index 0000000000..959a8a6677
@@ -0,0 +1,32 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM scratchpad engne
+ */
+#ifndef FSI_ENGINE_SCRATCHPAD_H
+#define FSI_ENGINE_SCRATCHPAD_H
+
+#include "hw/fsi/lbus.h"
+#include "hw/fsi/fsi.h"
+
+#define ENGINE_CONFIG_NEXT              BE_BIT(0)
+#define ENGINE_CONFIG_VPD               BE_BIT(1)
+#define ENGINE_CONFIG_SLOTS             BE_GENMASK(8, 15)
+#define ENGINE_CONFIG_VERSION           BE_GENMASK(16, 19)
+#define ENGINE_CONFIG_TYPE              BE_GENMASK(20, 27)
+#define   ENGINE_CONFIG_TYPE_PEEK       (0x02 << 4)
+#define   ENGINE_CONFIG_TYPE_FSI        (0x03 << 4)
+#define   ENGINE_CONFIG_TYPE_SCRATCHPAD (0x06 << 4)
+#define ENGINE_CONFIG_CRC              BE_GENMASK(28, 31)
+
+#define TYPE_FSI_SCRATCHPAD "fsi.scratchpad"
+#define SCRATCHPAD(obj) OBJECT_CHECK(FSIScratchPad, (obj), TYPE_FSI_SCRATCHPAD)
+
+typedef struct FSIScratchPad {
+        FSILBusDevice parent;
+
+        uint32_t reg;
+} FSIScratchPad;
+
+#endif /* FSI_ENGINE_SCRATCHPAD_H */
include/hw/fsi/fsi.h new file +16 -0
new file mode 100644
index 0000000000..e65f26f17b
@@ -0,0 +1,16 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Flexible Service Interface
+ */
+#ifndef FSI_FSI_H
+#define FSI_FSI_H
+
+/* Bitwise operations at the word level. */
+#define BE_BIT(x)                          BIT(31 - (x))
+#define GENMASK(t, b) \
+    (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) - 1))
+#define BE_GENMASK(t, b)                   GENMASK(BE_BIT(t), BE_BIT(b))
+
+#endif
hw/fsi/engine-scratchpad.c new file +93 -0

93 lines not on this page · raw message

hw/fsi/Kconfig +4 -0
index e650c660f0..f7c7fd1b28 100644
@@ -1,2 +1,6 @@
+config FSI_SCRATCHPAD
+    bool
+    select FSI_LBUS
+
 config FSI_LBUS
     bool
hw/fsi/meson.build +1 -0
index 4074d3a7d2..d45a98c223 100644
@@ -1 +1,2 @@
 system_ss.add(when: 'CONFIG_FSI_LBUS', if_true: files('lbus.c'))
+system_ss.add(when: 'CONFIG_FSI_SCRATCHPAD', if_true: files('engine-scratchpad.c'))
hw/fsi/trace-events new file +2 -0
new file mode 100644
index 0000000000..c5753e2791
@@ -0,0 +1,2 @@
+fsi_scratchpad_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_scratchpad_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
Signature · 1 line
2.39.2
#5 lore raw

Re: [PATCH v6 02/10] hw/fsi: Introduce IBM's scratchpad

Quoted 44 lines · Ninad Palsule

This is a part of patchset where scratchpad is introduced.

The scratchpad provides a set of non-functional registers. The firmware is free to use them, hardware does not support any special management support. The scratchpad registers can be read or written from LBUS slave.

In this model, The LBUS device is parent for the scratchpad.

---
v2:
- Incorporated Joel's review comments.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Daniel.

---
  meson.build                        |  1 +
  hw/fsi/trace.h                     |  1 +
  include/hw/fsi/engine-scratchpad.h | 32 ++++++++++
  include/hw/fsi/fsi.h               | 16 +++++
  hw/fsi/engine-scratchpad.c         | 93 ++++++++++++++++++++++++++++++
  hw/fsi/Kconfig                     |  4 ++
  hw/fsi/meson.build                 |  1 +
  hw/fsi/trace-events                |  2 +
  8 files changed, 150 insertions(+)

create mode 100644 hw/fsi/trace.h
create mode 100644 include/hw/fsi/engine-scratchpad.h create mode 100644 include/hw/fsi/fsi.h create mode 100644 hw/fsi/engine-scratchpad.c create mode 100644 hw/fsi/trace-events

include/hw/fsi/fsi.h new file +3 -0
new file mode 100644
index 0000000000..e65f26f17b
@@ -0,0 +1,16 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.

include/hw/fsi/fsi.h line 13

Quoted 10 lines
include/hw/fsi/fsi.h new file +10 -0
+ *
+ * IBM Flexible Service Interface
+ */
+#ifndef FSI_FSI_H
+#define FSI_FSI_H
+
+/* Bitwise operations at the word level. */
+#define BE_BIT(x)                          BIT(31 - (x))
+#define GENMASK(t, b) \
+    (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) - 1))

Please use MAKE_64BIT_MASK() from "qemu/bitops.h".

Quoted 3 lines
include/hw/fsi/fsi.h new file +3 -0
+#define BE_GENMASK(t, b)                   GENMASK(BE_BIT(t), BE_BIT(b))
+
+#endif
#6 lore raw

Re: [PATCH v6 02/10] hw/fsi: Introduce IBM's scratchpad

Hello Philippe,

Quoted 49 lines · Philippe Mathieu-Daudé
Quoted 48 lines · Ninad Palsule

This is a part of patchset where scratchpad is introduced.

The scratchpad provides a set of non-functional registers. The firmware is free to use them, hardware does not support any special management support. The scratchpad registers can be read or written from LBUS slave.

In this model, The LBUS device is parent for the scratchpad.

---
v2:
- Incorporated Joel's review comments.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Daniel.
---
  meson.build                        |  1 +
  hw/fsi/trace.h                     |  1 + include/hw/fsi/engine-scratchpad.h | 32 ++++++++++ include/hw/fsi/fsi.h               | 16 +++++ hw/fsi/engine-scratchpad.c         | 93 ++++++++++++++++++++++++++++++ hw/fsi/Kconfig                     |  4 ++ hw/fsi/meson.build                 |  1 + hw/fsi/trace-events                |  2 +

  8 files changed, 150 insertions(+)

  create mode 100644 hw/fsi/trace.h
  create mode 100644 include/hw/fsi/engine-scratchpad.h create mode 100644 include/hw/fsi/fsi.h create mode 100644 hw/fsi/engine-scratchpad.c create mode 100644 hw/fsi/trace-events

include/hw/fsi/fsi.h new file +7 -0
new file mode 100644
index 0000000000..e65f26f17b
@@ -0,0 +1,16 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Flexible Service Interface
+ */
+#ifndef FSI_FSI_H
Quoted 8 lines

include/hw/fsi/fsi.h line 13

Quoted 6 lines
include/hw/fsi/fsi.h new file +6 -0
+#define FSI_FSI_H
+
+/* Bitwise operations at the word level. */
+#define BE_BIT(x)                          BIT(31 - (x))
+#define GENMASK(t, b) \
+    (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) - 1))

Please use MAKE_64BIT_MASK() from "qemu/bitops.h".

The GENMASK and MAKE_64BIT_MASK macros are invoke differently.

GENMASK is invoked with bit t and bit b (t:b) and it provides the mask and

MAKE_64BIT_MASK uses shift and length.

Thanks for the review.

Regards,

Ninad

Quoted 5 lines
Quoted 4 lines
include/hw/fsi/fsi.h new file +1 -0
+#define BE_GENMASK(t, b)                   GENMASK(BE_BIT(t), 

BE_BIT(b))

include/hw/fsi/fsi.h new file +2 -0
+
+#endif
#7 lore raw

Re: [PATCH v6 02/10] hw/fsi: Introduce IBM's scratchpad

Quoted 61 lines · Ninad Palsule

Hello Philippe,

Quoted 49 lines · Philippe Mathieu-Daudé
Quoted 48 lines · Ninad Palsule

This is a part of patchset where scratchpad is introduced.

The scratchpad provides a set of non-functional registers. The firmware is free to use them, hardware does not support any special management support. The scratchpad registers can be read or written from LBUS slave.

In this model, The LBUS device is parent for the scratchpad.

---
v2:
- Incorporated Joel's review comments.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Daniel.
---
  meson.build                        |  1 +
  hw/fsi/trace.h                     |  1 + include/hw/fsi/engine-scratchpad.h | 32 ++++++++++ include/hw/fsi/fsi.h               | 16 +++++ hw/fsi/engine-scratchpad.c         | 93 ++++++++++++++++++++++++++++++ hw/fsi/Kconfig                     |  4 ++ hw/fsi/meson.build                 |  1 + hw/fsi/trace-events                |  2 +

  8 files changed, 150 insertions(+)

  create mode 100644 hw/fsi/trace.h
  create mode 100644 include/hw/fsi/engine-scratchpad.h create mode 100644 include/hw/fsi/fsi.h create mode 100644 hw/fsi/engine-scratchpad.c create mode 100644 hw/fsi/trace-events

include/hw/fsi/fsi.h new file +7 -0
new file mode 100644
index 0000000000..e65f26f17b
@@ -0,0 +1,16 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Flexible Service Interface
+ */
+#ifndef FSI_FSI_H
Quoted 8 lines

include/hw/fsi/fsi.h line 13

Quoted 6 lines
include/hw/fsi/fsi.h new file +6 -0
+#define FSI_FSI_H
+
+/* Bitwise operations at the word level. */
+#define BE_BIT(x)                          BIT(31 - (x))
+#define GENMASK(t, b) \
+    (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) - 1))

Please use MAKE_64BIT_MASK() from "qemu/bitops.h".

Quoted 5 lines

The GENMASK and MAKE_64BIT_MASK macros are invoke differently.

GENMASK is invoked with bit t and bit b (t:b) and it provides the mask and

MAKE_64BIT_MASK uses shift and length.

Don't we have:

#define GENMASK(t, b) MAKE_64BIT_MASK(t, b - t + 1)

?

Quoted 12 lines

Thanks for the review.

Regards,

Ninad

Quoted 5 lines
Quoted 4 lines
include/hw/fsi/fsi.h new file +1 -0
+#define BE_GENMASK(t, b)                   GENMASK(BE_BIT(t), 

BE_BIT(b))

include/hw/fsi/fsi.h new file +2 -0
+
+#endif
#8 lore raw

Re: [PATCH v6 02/10] hw/fsi: Introduce IBM's scratchpad

Quoted 67 lines · Philippe Mathieu-Daudé
Quoted 66 lines · Ninad Palsule

Hello Philippe,

Quoted 50 lines · Philippe Mathieu-Daudé
Quoted 49 lines · Ninad Palsule

This is a part of patchset where scratchpad is introduced.

The scratchpad provides a set of non-functional registers. The firmware is free to use them, hardware does not support any special management support. The scratchpad registers can be read or written from LBUS slave.

In this model, The LBUS device is parent for the scratchpad.

---
v2:
- Incorporated Joel's review comments.
v5:
- Incorporated review comments by Cedric. v6:
- Incorporated review comments by Daniel.
---
  meson.build                        |  1 + hw/fsi/trace.h                     |  1 + include/hw/fsi/engine-scratchpad.h | 32 ++++++++++ include/hw/fsi/fsi.h               | 16 +++++ hw/fsi/engine-scratchpad.c         | 93
++++++++++++++++++++++++++++++
  hw/fsi/Kconfig                     |  4 ++ hw/fsi/meson.build                 |  1 + hw/fsi/trace-events                |  2 +

  8 files changed, 150 insertions(+)

  create mode 100644 hw/fsi/trace.h
  create mode 100644 include/hw/fsi/engine-scratchpad.h create mode 100644 include/hw/fsi/fsi.h create mode 100644 hw/fsi/engine-scratchpad.c create mode 100644 hw/fsi/trace-events

include/hw/fsi/fsi.h new file +7 -0
new file mode 100644
index 0000000000..e65f26f17b
@@ -0,0 +1,16 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Flexible Service Interface
+ */
+#ifndef FSI_FSI_H
Quoted 8 lines

include/hw/fsi/fsi.h line 13

Quoted 6 lines
include/hw/fsi/fsi.h new file +6 -0
+#define FSI_FSI_H
+
+/* Bitwise operations at the word level. */
+#define BE_BIT(x)                          BIT(31 - (x))
+#define GENMASK(t, b) \
+    (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) - 1))

Please use MAKE_64BIT_MASK() from "qemu/bitops.h".

The GENMASK and MAKE_64BIT_MASK macros are invoke differently.

GENMASK is invoked with bit t and bit b (t:b) and it provides the mask and

Quoted 7 lines
Quoted 1 line

MAKE_64BIT_MASK uses shift and length.

Don't we have:

#define GENMASK(t, b) MAKE_64BIT_MASK(t, b - t + 1)

?

You are right. I am able to use this macro. I have removed some unused macros.

Thanks for the review.

Regards,

Ninad

Quoted 14 lines
Quoted 13 lines

Thanks for the review.

Regards,

Ninad

Quoted 5 lines
Quoted 4 lines
include/hw/fsi/fsi.h new file +1 -0
+#define BE_GENMASK(t, b)                   GENMASK(BE_BIT(t), 

BE_BIT(b))

include/hw/fsi/fsi.h new file +2 -0
+
+#endif
#9 lore raw

PATCH v6 3/10 hw/fsi: Introduce IBM's cfam,fsi-slave

In reply to Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

The Common FRU Access Macro (CFAM), an address space containing various "engines" that drive accesses on busses internal and external to the POWER chip. Examples include the SBEFIFO and I2C masters. The engines hang off of an internal Local Bus (LBUS) which is described by the CFAM configuration block.

The FSI slave: The slave is the terminal point of the FSI bus for FSI symbols addressed to it. Slaves can be cascaded off of one another. The slave's configuration registers appear in address space of the CFAM to which it is attached.

---
v2:
- Incorporated Joel's review comments.
v3:
- Incorporated Thomas Huth's review comments.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric & Daniel

---
 include/hw/fsi/cfam.h      |  34 ++++++++
 include/hw/fsi/fsi-slave.h |  29 +++++++
 include/hw/fsi/fsi.h       |  21 +++++
 hw/fsi/cfam.c              | 173 +++++++++++++++++++++++++++++++++++++
 hw/fsi/fsi-slave.c         |  96 ++++++++++++++++++++
 hw/fsi/Kconfig             |   9 ++
 hw/fsi/meson.build         |   2 +
 hw/fsi/trace-events        |   7 ++
 8 files changed, 371 insertions(+)
 create mode 100644 include/hw/fsi/cfam.h
 create mode 100644 include/hw/fsi/fsi-slave.h
 create mode 100644 hw/fsi/cfam.c
 create mode 100644 hw/fsi/fsi-slave.c
include/hw/fsi/cfam.h new file +34 -0
new file mode 100644
index 0000000000..842a3bad0c
@@ -0,0 +1,34 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Common FRU Access Macro
+ */
+#ifndef FSI_CFAM_H
+#define FSI_CFAM_H
+
+#include "exec/memory.h"
+
+#include "hw/fsi/fsi-slave.h"
+#include "hw/fsi/lbus.h"
+
+#define TYPE_FSI_CFAM "cfam"
+#define FSI_CFAM(obj) OBJECT_CHECK(FSICFAMState, (obj), TYPE_FSI_CFAM)
+
+/* P9-ism */
+#define CFAM_CONFIG_NR_REGS 0x28
+
+typedef struct FSICFAMState {
+    /* < private > */
+    FSISlaveState parent;
+
+    /* CFAM config address space */
+    MemoryRegion config_iomem;
+
+    MemoryRegion mr;
+    AddressSpace as;
+
+    FSILBus lbus;
+} FSICFAMState;
+
+#endif /* FSI_CFAM_H */
include/hw/fsi/fsi-slave.h new file +29 -0
new file mode 100644
index 0000000000..f5f23f4457
@@ -0,0 +1,29 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Flexible Service Interface slave
+ */
+#ifndef FSI_FSI_SLAVE_H
+#define FSI_FSI_SLAVE_H
+
+#include "exec/memory.h"
+#include "hw/qdev-core.h"
+
+#include "hw/fsi/lbus.h"
+
+#include <stdint.h>
+
+#define TYPE_FSI_SLAVE "fsi.slave"
+OBJECT_DECLARE_SIMPLE_TYPE(FSISlaveState, FSI_SLAVE)
+
+#define FSI_SLAVE_CONTROL_NR_REGS ((0x40 >> 2) + 1)
+
+typedef struct FSISlaveState {
+    DeviceState parent;
+
+    MemoryRegion iomem;
+    uint32_t regs[FSI_SLAVE_CONTROL_NR_REGS];
+} FSISlaveState;
+
+#endif /* FSI_FSI_H */
include/hw/fsi/fsi.h +21 -0
index e65f26f17b..14bc19279f 100644
@@ -7,10 +7,31 @@
 #ifndef FSI_FSI_H
 #define FSI_FSI_H
 
+#include "hw/qdev-core.h"
+
+/*
+ * TODO: Maybe unwind this dependency with const links? Store a
+ * pointer in FSIBus?
+ */
+#include "hw/fsi/cfam.h"
+
 /* Bitwise operations at the word level. */
 #define BE_BIT(x)                          BIT(31 - (x))
 #define GENMASK(t, b) \
     (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) - 1))
 #define BE_GENMASK(t, b)                   GENMASK(BE_BIT(t), BE_BIT(b))
 
+#define TYPE_FSI_BUS "fsi.bus"
+OBJECT_DECLARE_SIMPLE_TYPE(FSIBus, FSI_BUS)
+
+/* TODO: Figure out what's best with a point-to-point bus */
+typedef struct FSISlaveState FSISlaveState;
+
+typedef struct FSIBus {
+    BusState bus;
+
+    /* XXX: It's point-to-point, just instantiate the slave directly for now */
+    FSICFAMState slave;
+} FSIBus;
+
 #endif
hw/fsi/cfam.c new file +173 -0

173 lines not on this page · raw message

hw/fsi/fsi-slave.c new file +96 -0

96 lines not on this page · raw message

hw/fsi/Kconfig +9 -0
index f7c7fd1b28..8d712e77ed 100644
@@ -1,3 +1,12 @@
+config FSI_CFAM
+    bool
+    select FSI
+    select FSI_SCRATCHPAD
+    select FSI_LBUS
+
+config FSI
+    bool
+
 config FSI_SCRATCHPAD
     bool
     select FSI_LBUS
hw/fsi/meson.build +2 -0
index d45a98c223..a9e7cd4099 100644
@@ -1,2 +1,4 @@
 system_ss.add(when: 'CONFIG_FSI_LBUS', if_true: files('lbus.c'))
 system_ss.add(when: 'CONFIG_FSI_SCRATCHPAD', if_true: files('engine-scratchpad.c'))
+system_ss.add(when: 'CONFIG_FSI_CFAM', if_true: files('cfam.c'))
+system_ss.add(when: 'CONFIG_FSI', if_true: files('fsi-slave.c'))
hw/fsi/trace-events +7 -0
index c5753e2791..b57b2dcc86 100644
@@ -1,2 +1,9 @@
 fsi_scratchpad_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
 fsi_scratchpad_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
+fsi_cfam_config_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_cfam_config_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
+fsi_cfam_unimplemented_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_cfam_unimplemented_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
+fsi_cfam_config_write_noaddr(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
+fsi_slave_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_slave_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
Signature · 1 line
2.39.2
#10 lore raw

PATCH v6 4/10 hw/fsi: Introduce IBM's FSI

In reply to Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

This commit models the FSI bus. CFAM is hanging out of FSI bus. The bus is model such a way that it is embedded inside the FSI master which is a bus controller.

The FSI master: A controller in the platform service processor (e.g. BMC) driving CFAM engine accesses into the POWER chip. At the hardware level FSI is a bit-based protocol supporting synchronous and DMA-driven accesses of engines in a CFAM.

---
v2:
- Incorporated review comments by Joel
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric & Daniel

---
 include/hw/fsi/fsi-master.h |  30 ++++++
 hw/fsi/fsi-master.c         | 202 ++++++++++++++++++++++++++++++++++++
 hw/fsi/fsi.c                |  25 +++++
 hw/fsi/meson.build          |   2 +-
 hw/fsi/trace-events         |   2 +
 5 files changed, 260 insertions(+), 1 deletion(-)
 create mode 100644 include/hw/fsi/fsi-master.h
 create mode 100644 hw/fsi/fsi-master.c
 create mode 100644 hw/fsi/fsi.c
include/hw/fsi/fsi-master.h new file +30 -0
new file mode 100644
index 0000000000..847078919c
@@ -0,0 +1,30 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2019 IBM Corp.
+ *
+ * IBM Flexible Service Interface Master
+ */
+#ifndef FSI_FSI_MASTER_H
+#define FSI_FSI_MASTER_H
+
+#include "exec/memory.h"
+#include "hw/qdev-core.h"
+#include "hw/fsi/fsi.h"
+
+#define TYPE_FSI_MASTER "fsi.master"
+OBJECT_DECLARE_SIMPLE_TYPE(FSIMasterState, FSI_MASTER)
+
+#define FSI_MASTER_NR_REGS ((0x2e0 >> 2) + 1)
+
+typedef struct FSIMasterState {
+    DeviceState parent;
+    MemoryRegion iomem;
+    MemoryRegion opb2fsi;
+
+    FSIBus bus;
+
+    uint32_t regs[FSI_MASTER_NR_REGS];
+} FSIMasterState;
+
+
+#endif /* FSI_FSI_H */
hw/fsi/fsi-master.c new file +202 -0

202 lines not on this page · raw message

hw/fsi/fsi.c new file +25 -0
new file mode 100644
index 0000000000..180283746e
@@ -0,0 +1,25 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM Flexible Service Interface
+ */
+#include "qemu/osdep.h"
+
+#include "qapi/error.h"
+
+#include "hw/fsi/fsi.h"
+#include "hw/fsi/cfam.h"
+
+static const TypeInfo fsi_bus_info = {
+    .name = TYPE_FSI_BUS,
+    .parent = TYPE_BUS,
+    .instance_size = sizeof(FSIBus),
+};
+
+static void fsi_bus_register_types(void)
+{
+    type_register_static(&fsi_bus_info);
+}
+
+type_init(fsi_bus_register_types);
hw/fsi/meson.build +1 -1
index a9e7cd4099..f617943b4a 100644
@@ -1,4 +1,4 @@
 system_ss.add(when: 'CONFIG_FSI_LBUS', if_true: files('lbus.c'))
 system_ss.add(when: 'CONFIG_FSI_SCRATCHPAD', if_true: files('engine-scratchpad.c'))
 system_ss.add(when: 'CONFIG_FSI_CFAM', if_true: files('cfam.c'))
-system_ss.add(when: 'CONFIG_FSI', if_true: files('fsi-slave.c'))
+system_ss.add(when: 'CONFIG_FSI', if_true: files('fsi.c','fsi-master.c','fsi-slave.c'))
hw/fsi/trace-events +2 -0
index b57b2dcc86..89d8cd62c8 100644
@@ -7,3 +7,5 @@ fsi_cfam_unimplemented_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%"
 fsi_cfam_config_write_noaddr(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
 fsi_slave_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
 fsi_slave_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
+fsi_master_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_master_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
Signature · 1 line
2.39.2
#11 lore raw

PATCH v6 5/10 hw/fsi: IBM's On-chip Peripheral Bus

In reply to Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

The On-Chip Peripheral Bus (OPB): A low-speed bus typically found in POWER processors. This now makes an appearance in the ASPEED SoC due to tight integration of the FSI master IP with the OPB, mainly the existence of an MMIO-mapping of the CFAM address straight onto a sub-region of the OPB address space.

---
v2:
- Incorporated review comment by Joel.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric & Daniel

---
 include/hw/fsi/opb.h |  43 ++++++++++++
 hw/fsi/fsi-master.c  |   3 +-
 hw/fsi/opb.c         | 164 +++++++++++++++++++++++++++++++++++++++++++
 hw/fsi/Kconfig       |   4 ++
 hw/fsi/meson.build   |   1 +
 hw/fsi/trace-events  |   8 +++
 6 files changed, 221 insertions(+), 2 deletions(-)
 create mode 100644 include/hw/fsi/opb.h
 create mode 100644 hw/fsi/opb.c
include/hw/fsi/opb.h new file +43 -0
new file mode 100644
index 0000000000..9b143564ae
@@ -0,0 +1,43 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * IBM On-Chip Peripheral Bus
+ */
+#ifndef FSI_OPB_H
+#define FSI_OPB_H
+
+#include "exec/memory.h"
+#include "hw/fsi/fsi-master.h"
+
+#define TYPE_OP_BUS "opb"
+OBJECT_DECLARE_SIMPLE_TYPE(OPBus, OP_BUS)
+
+typedef struct OPBus {
+        /*< private >*/
+        BusState bus;
+
+        /*< public >*/
+        MemoryRegion mr;
+        AddressSpace as;
+
+        /* Model OPB as dumb enough just to provide an address-space */
+        /* TODO: Maybe don't store device state in the bus? */
+        FSIMasterState fsi;
+} OPBus;
+
+typedef struct OPBusClass {
+        BusClass parent_class;
+} OPBusClass;
+
+uint8_t fsi_opb_read8(OPBus *opb, hwaddr addr);
+uint16_t fsi_opb_read16(OPBus *opb, hwaddr addr);
+uint32_t fsi_opb_read32(OPBus *opb, hwaddr addr);
+void fsi_opb_write8(OPBus *opb, hwaddr addr, uint8_t data);
+void fsi_opb_write16(OPBus *opb, hwaddr addr, uint16_t data);
+void fsi_opb_write32(OPBus *opb, hwaddr addr, uint32_t data);
+
+void fsi_opb_fsi_master_address(OPBus *opb, hwaddr addr);
+void fsi_opb_opb2fsi_address(OPBus *opb, hwaddr addr);
+
+#endif /* FSI_OPB_H */
hw/fsi/fsi-master.c +1 -2
index fa23af3de2..91d1a57c4e 100644
@@ -11,8 +11,7 @@
 #include "trace.h"
 
 #include "hw/fsi/fsi-master.h"
-
-#define TYPE_OP_BUS "opb"
+#include "hw/fsi/opb.h"
 
 #define TO_REG(x)                               ((x) >> 2)
 
hw/fsi/opb.c new file +164 -0

164 lines not on this page · raw message

hw/fsi/Kconfig +4 -0
index 8d712e77ed..0f6e6d331a 100644
@@ -1,3 +1,7 @@
+config FSI_OPB
+    bool
+    select FSI_CFAM
+
 config FSI_CFAM
     bool
     select FSI
hw/fsi/meson.build +1 -0
index f617943b4a..407b8c2775 100644
@@ -2,3 +2,4 @@ system_ss.add(when: 'CONFIG_FSI_LBUS', if_true: files('lbus.c'))
 system_ss.add(when: 'CONFIG_FSI_SCRATCHPAD', if_true: files('engine-scratchpad.c'))
 system_ss.add(when: 'CONFIG_FSI_CFAM', if_true: files('cfam.c'))
 system_ss.add(when: 'CONFIG_FSI', if_true: files('fsi.c','fsi-master.c','fsi-slave.c'))
+system_ss.add(when: 'CONFIG_FSI_OPB', if_true: files('opb.c'))
hw/fsi/trace-events +8 -0
index 89d8cd62c8..13b211a815 100644
@@ -9,3 +9,11 @@ fsi_slave_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
 fsi_slave_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
 fsi_master_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
 fsi_master_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
+fsi_opb_read8(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_opb_read16(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_opb_read32(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_opb_write8(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_opb_write16(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_opb_write32(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_opb_unimplemented_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_opb_unimplemented_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
Signature · 1 line
2.39.2
#12 lore raw

PATCH v6 6/10 hw/fsi: Aspeed APB2OPB interface

In reply to Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

An APB-to-OPB bridge enabling access to the OPB from the ARM core in the AST2600. Hardware limitations prevent the OPB from being directly mapped into APB, so all accesses are indirect through the bridge.

---
v2:
- Incorporated review comments by Joel
v3:
- Incorporated review comments by Thomas Huth
v4:
- Compile FSI with ASPEED_SOC only.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric.

---
 include/hw/fsi/aspeed-apb2opb.h |  33 ++++
 hw/fsi/aspeed-apb2opb.c         | 280 ++++++++++++++++++++++++++++++++
 hw/arm/Kconfig                  |   1 +
 hw/fsi/Kconfig                  |   4 +
 hw/fsi/meson.build              |   1 +
 hw/fsi/trace-events             |   2 +
 6 files changed, 321 insertions(+)
 create mode 100644 include/hw/fsi/aspeed-apb2opb.h
 create mode 100644 hw/fsi/aspeed-apb2opb.c
include/hw/fsi/aspeed-apb2opb.h new file +33 -0
new file mode 100644
index 0000000000..a81ae67023
@@ -0,0 +1,33 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB2OPB Bridge
+ */
+#ifndef FSI_ASPEED_APB2OPB_H
+#define FSI_ASPEED_APB2OPB_H
+
+#include "hw/sysbus.h"
+#include "hw/fsi/opb.h"
+
+#define TYPE_ASPEED_APB2OPB "aspeed.apb2opb"
+OBJECT_DECLARE_SIMPLE_TYPE(AspeedAPB2OPBState, ASPEED_APB2OPB)
+
+#define ASPEED_APB2OPB_NR_REGS ((0xe8 >> 2) + 1)
+
+#define ASPEED_FSI_NUM 2
+
+typedef struct AspeedAPB2OPBState {
+    /*< private >*/
+    SysBusDevice parent_obj;
+
+    /*< public >*/
+    MemoryRegion iomem;
+
+    uint32_t regs[ASPEED_APB2OPB_NR_REGS];
+    qemu_irq irq;
+
+    OPBus opb[ASPEED_FSI_NUM];
+} AspeedAPB2OPBState;
+
+#endif /* FSI_ASPEED_APB2OPB_H */
hw/fsi/aspeed-apb2opb.c new file +280 -0

280 lines not on this page · raw message

hw/arm/Kconfig +1 -0
index 7e68348440..d963de74c9 100644
@@ -555,6 +555,7 @@ config ASPEED_SOC
     select LED
     select PMBUS
     select MAX31785
+    select FSI_APB2OPB_ASPEED
 
 config MPS2
     bool
hw/fsi/Kconfig +4 -0
index 0f6e6d331a..6bbcb8f6ca 100644
@@ -1,3 +1,7 @@
+config FSI_APB2OPB_ASPEED
+    bool
+    select FSI_OPB
+
 config FSI_OPB
     bool
     select FSI_CFAM
hw/fsi/meson.build +1 -0
index 407b8c2775..1bc6bb63cc 100644
@@ -3,3 +3,4 @@ system_ss.add(when: 'CONFIG_FSI_SCRATCHPAD', if_true: files('engine-scratchpad.c
 system_ss.add(when: 'CONFIG_FSI_CFAM', if_true: files('cfam.c'))
 system_ss.add(when: 'CONFIG_FSI', if_true: files('fsi.c','fsi-master.c','fsi-slave.c'))
 system_ss.add(when: 'CONFIG_FSI_OPB', if_true: files('opb.c'))
+system_ss.add(when: 'CONFIG_FSI_APB2OPB_ASPEED', if_true: files('aspeed-apb2opb.c'))
hw/fsi/trace-events +2 -0
index 13b211a815..9a45843eb6 100644
@@ -17,3 +17,5 @@ fsi_opb_write16(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
 fsi_opb_write32(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
 fsi_opb_unimplemented_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
 fsi_opb_unimplemented_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
+fsi_aspeed_apb2opb_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_aspeed_apb2opb_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
Signature · 1 line
2.39.2
#13 lore raw

Re: [PATCH v6 06/10] hw/fsi: Aspeed APB2OPB interface

Quoted 156 lines · Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

An APB-to-OPB bridge enabling access to the OPB from the ARM core in the AST2600. Hardware limitations prevent the OPB from being directly mapped into APB, so all accesses are indirect through the bridge.

---
v2:
- Incorporated review comments by Joel
v3:
- Incorporated review comments by Thomas Huth
v4:
- Compile FSI with ASPEED_SOC only.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric.

---
  include/hw/fsi/aspeed-apb2opb.h |  33 ++++
  hw/fsi/aspeed-apb2opb.c         | 280 ++++++++++++++++++++++++++++++++
  hw/arm/Kconfig                  |   1 +
  hw/fsi/Kconfig                  |   4 +
  hw/fsi/meson.build              |   1 +
  hw/fsi/trace-events             |   2 +
  6 files changed, 321 insertions(+)

create mode 100644 include/hw/fsi/aspeed-apb2opb.h create mode 100644 hw/fsi/aspeed-apb2opb.c

include/hw/fsi/aspeed-apb2opb.h new file +33 -0
new file mode 100644
index 0000000000..a81ae67023
@@ -0,0 +1,33 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB2OPB Bridge
+ */
+#ifndef FSI_ASPEED_APB2OPB_H
+#define FSI_ASPEED_APB2OPB_H
+
+#include "hw/sysbus.h"
+#include "hw/fsi/opb.h"
+
+#define TYPE_ASPEED_APB2OPB "aspeed.apb2opb"
+OBJECT_DECLARE_SIMPLE_TYPE(AspeedAPB2OPBState, ASPEED_APB2OPB)
+
+#define ASPEED_APB2OPB_NR_REGS ((0xe8 >> 2) + 1)
+
+#define ASPEED_FSI_NUM 2
+
+typedef struct AspeedAPB2OPBState {
+    /*< private >*/
+    SysBusDevice parent_obj;
+
+    /*< public >*/
+    MemoryRegion iomem;
+
+    uint32_t regs[ASPEED_APB2OPB_NR_REGS];
+    qemu_irq irq;
+
+    OPBus opb[ASPEED_FSI_NUM];
+} AspeedAPB2OPBState;
+
+#endif /* FSI_ASPEED_APB2OPB_H */
hw/fsi/aspeed-apb2opb.c new file +80 -0
new file mode 100644
index 0000000000..6f97a6bc7d
@@ -0,0 +1,280 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB-OPB FSI interface
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/log.h"
+#include "qom/object.h"
+#include "qapi/error.h"
+#include "trace.h"
+
+#include "hw/fsi/aspeed-apb2opb.h"
+#include "hw/qdev-core.h"
+
+#define TO_REG(x) (x >> 2)
+#define GENMASK(t, b) (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) - 1))
+
+#define APB2OPB_VERSION                    TO_REG(0x00)
+#define   APB2OPB_VERSION_VER              GENMASK(7, 0)
+
+#define APB2OPB_TRIGGER                    TO_REG(0x04)
+#define   APB2OPB_TRIGGER_EN               BIT(0)
+
+#define APB2OPB_CONTROL                    TO_REG(0x08)
+#define   APB2OPB_CONTROL_OFF              GENMASK(31, 13)
+
+#define APB2OPB_OPB2FSI                    TO_REG(0x0c)
+#define   APB2OPB_OPB2FSI_OFF              GENMASK(31, 22)
+
+#define APB2OPB_OPB0_SEL                   TO_REG(0x10)
+#define APB2OPB_OPB1_SEL                   TO_REG(0x28)
+#define   APB2OPB_OPB_SEL_EN               BIT(0)
+
+#define APB2OPB_OPB0_MODE                  TO_REG(0x14)
+#define APB2OPB_OPB1_MODE                  TO_REG(0x2c)
+#define   APB2OPB_OPB_MODE_RD              BIT(0)
+
+#define APB2OPB_OPB0_XFER                  TO_REG(0x18)
+#define APB2OPB_OPB1_XFER                  TO_REG(0x30)
+#define   APB2OPB_OPB_XFER_FULL            BIT(1)
+#define   APB2OPB_OPB_XFER_HALF            BIT(0)
+
+#define APB2OPB_OPB0_ADDR                  TO_REG(0x1c)
+#define APB2OPB_OPB0_WRITE_DATA            TO_REG(0x20)
+
+#define APB2OPB_OPB1_ADDR                  TO_REG(0x34)
+#define APB2OPB_OPB1_WRITE_DATA                  TO_REG(0x38)
+
+#define APB2OPB_IRQ_STS                    TO_REG(0x48)
+#define   APB2OPB_IRQ_STS_OPB1_TX_ACK      BIT(17)
+#define   APB2OPB_IRQ_STS_OPB0_TX_ACK      BIT(16)
+
+#define APB2OPB_OPB0_WRITE_WORD_ENDIAN     TO_REG(0x4c)
+#define   APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE 0x0011101b
+#define APB2OPB_OPB0_WRITE_BYTE_ENDIAN     TO_REG(0x50)
+#define   APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE 0x0c330f3f
+#define APB2OPB_OPB1_WRITE_WORD_ENDIAN     TO_REG(0x54)
+#define APB2OPB_OPB1_WRITE_BYTE_ENDIAN     TO_REG(0x58)
+#define APB2OPB_OPB0_READ_BYTE_ENDIAN      TO_REG(0x5c)
+#define APB2OPB_OPB1_READ_BYTE_ENDIAN      TO_REG(0x60)
+#define   APB2OPB_OPB0_READ_WORD_ENDIAN_BE  0x00030b1b
+
+#define APB2OPB_OPB0_READ_DATA         TO_REG(0x84)
+#define APB2OPB_OPB1_READ_DATA         TO_REG(0x90)
+
+/*
+ * The following magic values came from AST2600 data sheet
+ * The register values are defined under section "FSI controller"
+ * as initial values.
+ */
+static const uint32_t aspeed_apb2opb_reset[ASPEED_APB2OPB_NR_REGS] = {
+     [APB2OPB_VERSION]                = 0x000000a1,
+     [APB2OPB_OPB0_WRITE_WORD_ENDIAN] = 0x0044eee4,
+     [APB2OPB_OPB0_WRITE_BYTE_ENDIAN] = 0x0055aaff,
+     [APB2OPB_OPB1_WRITE_WORD_ENDIAN] = 0x00117717,
+     [APB2OPB_OPB1_WRITE_BYTE_ENDIAN] = 0xffaa5500,
+     [APB2OPB_OPB0_READ_BYTE_ENDIAN]  = 0x0044eee4,
+     [APB2OPB_OPB1_READ_BYTE_ENDIAN]  = 0x00117717

hw/fsi/aspeed-apb2opb.c line 90

Quoted 10 lines
hw/fsi/aspeed-apb2opb.c new file +10 -0
+};
+
+static uint64_t fsi_aspeed_apb2opb_read(void *opaque, hwaddr addr,
+                                        unsigned size)
+{
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);
+
+    trace_fsi_aspeed_apb2opb_read(addr, size);
+
+    if (addr + size > sizeof(s->regs)) {

hmm, the parameter 'size' is a memop transaction size not an address offset.

Quoted 7 lines
hw/fsi/aspeed-apb2opb.c new file +7 -0
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds read: 0x%"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return 0;
+    }
+
+    return s->regs[TO_REG(addr)];

hw/fsi/aspeed-apb2opb.c line 107

Quoted 10 lines
hw/fsi/aspeed-apb2opb.c new file +10 -0
+}
+
+static void fsi_aspeed_apb2opb_write(void *opaque, hwaddr addr, uint64_t data,
+                                     unsigned size)
+{
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);
+
+    trace_fsi_aspeed_apb2opb_write(addr, size, data);
+
+    if (addr + size > sizeof(s->regs)) {

same comment.

hw/fsi/aspeed-apb2opb.c line 116

Quoted 9 lines
hw/fsi/aspeed-apb2opb.c new file +9 -0
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds write: %"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return;
+    }
+
+    switch (TO_REG(addr)) {
+    case APB2OPB_CONTROL:
+        fsi_opb_fsi_master_address(&s->opb[0], data & APB2OPB_CONTROL_OFF);

fsi_opb_fsi_master_address() should statically defined in this file

hw/fsi/aspeed-apb2opb.c line 119

Quoted 3 lines
hw/fsi/aspeed-apb2opb.c new file +3 -0
+        break;
+    case APB2OPB_OPB2FSI:
+        fsi_opb_opb2fsi_address(&s->opb[0], data & APB2OPB_OPB2FSI_OFF);

same for fsi_opb_opb2fsi_address()

Quoted 41 lines
hw/fsi/aspeed-apb2opb.c new file +41 -0
+        break;
+    case APB2OPB_OPB0_WRITE_WORD_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_OPB0_WRITE_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_OPB0_READ_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_READ_WORD_ENDIAN_BE) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_READ_WORD_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_TRIGGER:
+    {
+        uint32_t opb, op_mode, op_size, op_addr, op_data;
+
+        assert((s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) ^
+               (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN));
+
+        if (s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 0;
+            op_mode = s->regs[APB2OPB_OPB0_MODE];
+            op_size = s->regs[APB2OPB_OPB0_XFER];
+            op_addr = s->regs[APB2OPB_OPB0_ADDR];
+            op_data = s->regs[APB2OPB_OPB0_WRITE_DATA];
+        } else if (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 1;
+            op_mode = s->regs[APB2OPB_OPB1_MODE];
+            op_size = s->regs[APB2OPB_OPB1_XFER];
+            op_addr = s->regs[APB2OPB_OPB1_ADDR];
+            op_data = s->regs[APB2OPB_OPB1_WRITE_DATA];

hw/fsi/aspeed-apb2opb.c line 170

Quoted 10 lines
hw/fsi/aspeed-apb2opb.c new file +10 -0
+        } else {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Invalid operation: 0x%"HWADDR_PRIx" for %u\n",
+                          __func__, addr, size);
+            return;
+        }
+
+        if (op_size & ~(APB2OPB_OPB_XFER_HALF | APB2OPB_OPB_XFER_FULL)) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "OPB transaction failed: Unrecognised access width: %d\n",

Unrecognized

Quoted 14 lines
hw/fsi/aspeed-apb2opb.c new file +14 -0
+                          op_size);
+            return;
+        }
+
+        op_size += 1;
+
+        if (op_mode & APB2OPB_OPB_MODE_RD) {
+            int index = opb ? APB2OPB_OPB1_READ_DATA
+                : APB2OPB_OPB0_READ_DATA;
+
+            switch (op_size) {
+            case 1:
+                s->regs[index] = fsi_opb_read8(&s->opb[opb], op_addr);
+                break;

hw/fsi/aspeed-apb2opb.c line 194

Quoted 10 lines
hw/fsi/aspeed-apb2opb.c new file +10 -0
+            case 2:
+                s->regs[index] = fsi_opb_read16(&s->opb[opb], op_addr);
+                break;
+            case 4:
+                s->regs[index] = fsi_opb_read32(&s->opb[opb], op_addr);
+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n",
+                              __func__, size);

this should use op_size and not size and seems redudant with the unrecognized test above.

Quoted 11 lines
hw/fsi/aspeed-apb2opb.c new file +11 -0
+                return;
+            }
+        } else {
+            /* FIXME: Endian swizzling */
+            switch (op_size) {
+            case 1:
+                fsi_opb_write8(&s->opb[opb], op_addr, op_data);
+                break;
+            case 2:
+                fsi_opb_write16(&s->opb[opb], op_addr, op_data);
+                break;

hw/fsi/aspeed-apb2opb.c line 215

Quoted 10 lines
hw/fsi/aspeed-apb2opb.c new file +10 -0
+            case 4:
+                fsi_opb_write32(&s->opb[opb], op_addr, op_data);
+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n",
+                              __func__, op_size);
+                return;
+            }
+        }

The above is equivalent to :

        MemTxResult result;
        bool is_write = !(op_mode & APB2OPB_OPB_MODE_RD);
        int index = opb ? APB2OPB_OPB1_READ_DATA : APB2OPB_OPB0_READ_DATA;
        AddressSpace *as = &s->opb[opb].as;

        result = address_space_rw(as, op_addr, MEMTXATTRS_UNSPECIFIED,
                                  &op_data, op_size, is_write);
        if (result != MEMTX_OK) {
            qemu_log_mask(LOG_GUEST_ERROR, "%s: OPB %s failed @%08x\n",
                          __func__, is_write ? "write" : "read", op_addr);
            return;
        }

        if (!is_write) {
            s->regs[index] = op_data;
        }

and the fsi_opb_* routines are useless to me.

Quoted 108 lines
hw/fsi/aspeed-apb2opb.c new file +65 -0
+        s->regs[APB2OPB_IRQ_STS] |= opb ? APB2OPB_IRQ_STS_OPB1_TX_ACK
+            : APB2OPB_IRQ_STS_OPB0_TX_ACK;
+        break;
+    }
+    }
+
+    s->regs[TO_REG(addr)] = data;
+}
+
+static const struct MemoryRegionOps aspeed_apb2opb_ops = {
+    .read = fsi_aspeed_apb2opb_read,
+    .write = fsi_aspeed_apb2opb_write,
+    .valid.max_access_size = 4,
+    .valid.min_access_size = 4,
+    .impl.max_access_size = 4,
+    .impl.min_access_size = 4,
+    .endianness = DEVICE_LITTLE_ENDIAN,
+};
+
+static void fsi_aspeed_apb2opb_realize(DeviceState *dev, Error **errp)
+{
+    SysBusDevice *sbd = SYS_BUS_DEVICE(dev);
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(dev);
+
+    qbus_init(&s->opb[0], sizeof(s->opb[0]), TYPE_OP_BUS,
+                        DEVICE(s), NULL);
+    qbus_init(&s->opb[1], sizeof(s->opb[1]), TYPE_OP_BUS,
+                        DEVICE(s), NULL);
+
+    sysbus_init_irq(sbd, &s->irq);
+
+    memory_region_init_io(&s->iomem, OBJECT(s), &aspeed_apb2opb_ops, s,
+                          TYPE_ASPEED_APB2OPB, 0x1000);
+    sysbus_init_mmio(sbd, &s->iomem);
+}
+
+static void fsi_aspeed_apb2opb_reset(DeviceState *dev)
+{
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(dev);
+
+    memcpy(s->regs, aspeed_apb2opb_reset, ASPEED_APB2OPB_NR_REGS);
+}
+
+static void fsi_aspeed_apb2opb_class_init(ObjectClass *klass, void *data)
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+
+    dc->desc = "ASPEED APB2OPB Bridge";
+    dc->realize = fsi_aspeed_apb2opb_realize;
+    dc->reset = fsi_aspeed_apb2opb_reset;
+}
+
+static const TypeInfo aspeed_apb2opb_info = {
+    .name = TYPE_ASPEED_APB2OPB,
+    .parent = TYPE_SYS_BUS_DEVICE,
+    .instance_size = sizeof(AspeedAPB2OPBState),
+    .class_init = fsi_aspeed_apb2opb_class_init,
+};
+
+static void aspeed_apb2opb_register_types(void)
+{
+    type_register_static(&aspeed_apb2opb_info);
+}
+
+type_init(aspeed_apb2opb_register_types);
hw/arm/Kconfig +1 -0
index 7e68348440..d963de74c9 100644
@@ -555,6 +555,7 @@ config ASPEED_SOC
      select LED
      select PMBUS
      select MAX31785
+    select FSI_APB2OPB_ASPEED
 
  config MPS2
      bool
hw/fsi/Kconfig +4 -0
index 0f6e6d331a..6bbcb8f6ca 100644
@@ -1,3 +1,7 @@
+config FSI_APB2OPB_ASPEED
+    bool
+    select FSI_OPB
+
  config FSI_OPB
      bool
      select FSI_CFAM
hw/fsi/meson.build +1 -0
index 407b8c2775..1bc6bb63cc 100644
@@ -3,3 +3,4 @@ system_ss.add(when: 'CONFIG_FSI_SCRATCHPAD', if_true: files('engine-scratchpad.c
  system_ss.add(when: 'CONFIG_FSI_CFAM', if_true: files('cfam.c'))
  system_ss.add(when: 'CONFIG_FSI', if_true: files('fsi.c','fsi-master.c','fsi-slave.c'))
  system_ss.add(when: 'CONFIG_FSI_OPB', if_true: files('opb.c'))
+system_ss.add(when: 'CONFIG_FSI_APB2OPB_ASPEED', if_true: files('aspeed-apb2opb.c'))
hw/fsi/trace-events +2 -0
index 13b211a815..9a45843eb6 100644
@@ -17,3 +17,5 @@ fsi_opb_write16(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
  fsi_opb_write32(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
  fsi_opb_unimplemented_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
  fsi_opb_unimplemented_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
+fsi_aspeed_apb2opb_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d"
+fsi_aspeed_apb2opb_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64
#14 lore raw

Re: [PATCH v6 06/10] hw/fsi: Aspeed APB2OPB interface

Hello Cedric,

Quoted 163 lines · Cédric Le Goater
Quoted 162 lines · Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

An APB-to-OPB bridge enabling access to the OPB from the ARM core in the AST2600. Hardware limitations prevent the OPB from being directly mapped into APB, so all accesses are indirect through the bridge.

---
v2:
- Incorporated review comments by Joel
v3:
- Incorporated review comments by Thomas Huth
v4:
   - Compile FSI with ASPEED_SOC only.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric.
---
  include/hw/fsi/aspeed-apb2opb.h |  33 ++++ hw/fsi/aspeed-apb2opb.c         | 280 ++++++++++++++++++++++++++++++++ hw/arm/Kconfig                  |   1 +
  hw/fsi/Kconfig                  |   4 +
  hw/fsi/meson.build              |   1 + hw/fsi/trace-events             |   2 +

  6 files changed, 321 insertions(+)

  create mode 100644 include/hw/fsi/aspeed-apb2opb.h create mode 100644 hw/fsi/aspeed-apb2opb.c

include/hw/fsi/aspeed-apb2opb.h new file +33 -0
new file mode 100644
index 0000000000..a81ae67023
@@ -0,0 +1,33 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB2OPB Bridge
+ */
+#ifndef FSI_ASPEED_APB2OPB_H
+#define FSI_ASPEED_APB2OPB_H
+
+#include "hw/sysbus.h"
+#include "hw/fsi/opb.h"
+
+#define TYPE_ASPEED_APB2OPB "aspeed.apb2opb"
+OBJECT_DECLARE_SIMPLE_TYPE(AspeedAPB2OPBState, ASPEED_APB2OPB)
+
+#define ASPEED_APB2OPB_NR_REGS ((0xe8 >> 2) + 1)
+
+#define ASPEED_FSI_NUM 2
+
+typedef struct AspeedAPB2OPBState {
+    /*< private >*/
+    SysBusDevice parent_obj;
+
+    /*< public >*/
+    MemoryRegion iomem;
+
+    uint32_t regs[ASPEED_APB2OPB_NR_REGS];
+    qemu_irq irq;
+
+    OPBus opb[ASPEED_FSI_NUM];
+} AspeedAPB2OPBState;
+
+#endif /* FSI_ASPEED_APB2OPB_H */
hw/fsi/aspeed-apb2opb.c new file +18 -0
new file mode 100644
index 0000000000..6f97a6bc7d
@@ -0,0 +1,280 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB-OPB FSI interface
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/log.h"
+#include "qom/object.h"
+#include "qapi/error.h"
+#include "trace.h"
+
+#include "hw/fsi/aspeed-apb2opb.h"
+#include "hw/qdev-core.h"
+
+#define TO_REG(x) (x >> 2)
+#define GENMASK(t, b) (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) - 

1))

hw/fsi/aspeed-apb2opb.c new file +67 -0
+
+#define APB2OPB_VERSION                    TO_REG(0x00)
+#define   APB2OPB_VERSION_VER              GENMASK(7, 0)
+
+#define APB2OPB_TRIGGER                    TO_REG(0x04)
+#define   APB2OPB_TRIGGER_EN               BIT(0)
+
+#define APB2OPB_CONTROL                    TO_REG(0x08)
+#define   APB2OPB_CONTROL_OFF              GENMASK(31, 13)
+
+#define APB2OPB_OPB2FSI                    TO_REG(0x0c)
+#define   APB2OPB_OPB2FSI_OFF              GENMASK(31, 22)
+
+#define APB2OPB_OPB0_SEL                   TO_REG(0x10)
+#define APB2OPB_OPB1_SEL                   TO_REG(0x28)
+#define   APB2OPB_OPB_SEL_EN               BIT(0)
+
+#define APB2OPB_OPB0_MODE                  TO_REG(0x14)
+#define APB2OPB_OPB1_MODE                  TO_REG(0x2c)
+#define   APB2OPB_OPB_MODE_RD              BIT(0)
+
+#define APB2OPB_OPB0_XFER                  TO_REG(0x18)
+#define APB2OPB_OPB1_XFER                  TO_REG(0x30)
+#define   APB2OPB_OPB_XFER_FULL            BIT(1)
+#define   APB2OPB_OPB_XFER_HALF            BIT(0)
+
+#define APB2OPB_OPB0_ADDR                  TO_REG(0x1c)
+#define APB2OPB_OPB0_WRITE_DATA            TO_REG(0x20)
+
+#define APB2OPB_OPB1_ADDR                  TO_REG(0x34)
+#define APB2OPB_OPB1_WRITE_DATA                  TO_REG(0x38)
+
+#define APB2OPB_IRQ_STS                    TO_REG(0x48)
+#define   APB2OPB_IRQ_STS_OPB1_TX_ACK      BIT(17)
+#define   APB2OPB_IRQ_STS_OPB0_TX_ACK      BIT(16)
+
+#define APB2OPB_OPB0_WRITE_WORD_ENDIAN     TO_REG(0x4c)
+#define   APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE 0x0011101b
+#define APB2OPB_OPB0_WRITE_BYTE_ENDIAN     TO_REG(0x50)
+#define   APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE 0x0c330f3f
+#define APB2OPB_OPB1_WRITE_WORD_ENDIAN     TO_REG(0x54)
+#define APB2OPB_OPB1_WRITE_BYTE_ENDIAN     TO_REG(0x58)
+#define APB2OPB_OPB0_READ_BYTE_ENDIAN      TO_REG(0x5c)
+#define APB2OPB_OPB1_READ_BYTE_ENDIAN      TO_REG(0x60)
+#define   APB2OPB_OPB0_READ_WORD_ENDIAN_BE  0x00030b1b
+
+#define APB2OPB_OPB0_READ_DATA         TO_REG(0x84)
+#define APB2OPB_OPB1_READ_DATA         TO_REG(0x90)
+
+/*
+ * The following magic values came from AST2600 data sheet
+ * The register values are defined under section "FSI controller"
+ * as initial values.
+ */
+static const uint32_t aspeed_apb2opb_reset[ASPEED_APB2OPB_NR_REGS] = {
+     [APB2OPB_VERSION]                = 0x000000a1,
+     [APB2OPB_OPB0_WRITE_WORD_ENDIAN] = 0x0044eee4,
+     [APB2OPB_OPB0_WRITE_BYTE_ENDIAN] = 0x0055aaff,
+     [APB2OPB_OPB1_WRITE_WORD_ENDIAN] = 0x00117717,
+     [APB2OPB_OPB1_WRITE_BYTE_ENDIAN] = 0xffaa5500,
+     [APB2OPB_OPB0_READ_BYTE_ENDIAN]  = 0x0044eee4,
+     [APB2OPB_OPB1_READ_BYTE_ENDIAN]  = 0x00117717
+};
+
+static uint64_t fsi_aspeed_apb2opb_read(void *opaque, hwaddr addr,
+                                        unsigned size)
+{
Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 90

Quoted 5 lines

+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);

hw/fsi/aspeed-apb2opb.c new file +4 -0
+
+    trace_fsi_aspeed_apb2opb_read(addr, size);
+
+    if (addr + size > sizeof(s->regs)) {

hmm, the parameter 'size' is a memop transaction size not an address offset.

OK, Changed it to validate the register (index) instead of addr + size.

Quoted 14 lines
Quoted 14 lines

+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds read: 0x%"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return 0;
+    }

hw/fsi/aspeed-apb2opb.c new file +5 -0
+
+    return s->regs[TO_REG(addr)];
+}
+
+static void fsi_aspeed_apb2opb_write(void *opaque, hwaddr addr, 

uint64_t data,
+                                     unsigned size)

Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 107

Quoted 6 lines
hw/fsi/aspeed-apb2opb.c new file +6 -0
+{
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);
+
+    trace_fsi_aspeed_apb2opb_write(addr, size, data);
+
+    if (addr + size > sizeof(s->regs)) {

same comment.

Fixed it same as above.

Quoted 14 lines
Quoted 13 lines

+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds write: %"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return;
+    }
+
+    switch (TO_REG(addr)) {
+    case APB2OPB_CONTROL:
+        fsi_opb_fsi_master_address(&s->opb[0], data & APB2OPB_CONTROL_OFF);

Quoted 1 line

fsi_opb_fsi_master_address() should statically defined in this file

We have separation of OPB bus implementation and APB2OPB interface. If we move this function here then we will be exposing OPB implementation here.

Quoted 8 lines
Quoted 5 lines

+        break;
+    case APB2OPB_OPB2FSI:
+        fsi_opb_opb2fsi_address(&s->opb[0], data & APB2OPB_OPB2FSI_OFF);

same for fsi_opb_opb2fsi_address()

Same as above.

Quoted 52 lines
Quoted 52 lines

+        break;
+    case APB2OPB_OPB0_WRITE_WORD_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE); +        }
+        break;
+    case APB2OPB_OPB0_WRITE_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE); +        }
+        break;
+    case APB2OPB_OPB0_READ_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_READ_WORD_ENDIAN_BE) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_READ_WORD_ENDIAN_BE); +        }
+        break;
+    case APB2OPB_TRIGGER:
+    {
+        uint32_t opb, op_mode, op_size, op_addr, op_data;

hw/fsi/aspeed-apb2opb.c new file +19 -0
+
+        assert((s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) ^
+               (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN));
+
+        if (s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 0;
+            op_mode = s->regs[APB2OPB_OPB0_MODE];
+            op_size = s->regs[APB2OPB_OPB0_XFER];
+            op_addr = s->regs[APB2OPB_OPB0_ADDR];
+            op_data = s->regs[APB2OPB_OPB0_WRITE_DATA];
+        } else if (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 1;
+            op_mode = s->regs[APB2OPB_OPB1_MODE];
+            op_size = s->regs[APB2OPB_OPB1_XFER];
+            op_addr = s->regs[APB2OPB_OPB1_ADDR];
+            op_data = s->regs[APB2OPB_OPB1_WRITE_DATA];
+        } else {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Invalid operation: 0x%"HWADDR_PRIx" 

for %u\n",
+                          __func__, addr, size);
+            return;
+        }

Quoted 8 lines
Quoted 6 lines

+
+        if (op_size & ~(APB2OPB_OPB_XFER_HALF | APB2OPB_OPB_XFER_FULL)) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "OPB transaction failed: Unrecognised access width: %d\n",

Unrecognized

Fixed

Quoted 20 lines
Quoted 20 lines

+                          op_size);
+            return;
+        }

hw/fsi/aspeed-apb2opb.c new file +16 -0
+
+        op_size += 1;
+
+        if (op_mode & APB2OPB_OPB_MODE_RD) {
+            int index = opb ? APB2OPB_OPB1_READ_DATA
+                : APB2OPB_OPB0_READ_DATA;
+
+            switch (op_size) {
+            case 1:
+                s->regs[index] = fsi_opb_read8(&s->opb[opb], op_addr);
+                break;
+            case 2:
+                s->regs[index] = fsi_opb_read16(&s->opb[opb], op_addr);
+                break;
+            case 4:
+                s->regs[index] = fsi_opb_read32(&s->opb[opb], op_addr);
Quoted 8 lines
Quoted 5 lines

+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n", +                              __func__, size);

this should use op_size and not size and seems redudant with the unrecognized test above.

true, Keeping it in case bits meaning change in future.

Quoted 40 lines
Quoted 15 lines

+                return;
+            }
+        } else {
+            /* FIXME: Endian swizzling */
+            switch (op_size) {
+            case 1:
+                fsi_opb_write8(&s->opb[opb], op_addr, op_data); +                break;
+            case 2:
+                fsi_opb_write16(&s->opb[opb], op_addr, op_data); +                break;
+            case 4:
+                fsi_opb_write32(&s->opb[opb], op_addr, op_data);

Quoted 8 lines

+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n", +                              __func__, op_size);
+                return;
+            }
+        }

The above is equivalent to :

        MemTxResult result;
        bool is_write = !(op_mode & APB2OPB_OPB_MODE_RD);
        int index = opb ? APB2OPB_OPB1_READ_DATA : APB2OPB_OPB0_READ_DATA; AddressSpace *as = &s->opb[opb].as;

        result = address_space_rw(as, op_addr, MEMTXATTRS_UNSPECIFIED, &op_data, op_size, is_write); if (result != MEMTX_OK) { qemu_log_mask(LOG_GUEST_ERROR, "%s: OPB %s failed @%08x\n", __func__, is_write ? "write" : "read", op_addr); return;
        }

Quoted 5 lines

        if (!is_write) {
            s->regs[index] = op_data;
        }

and the fsi_opb_* routines are useless to me.

We are trying to keep the separation between OPB implementation and interface hence we have all those fsi_opb_*. I feel that we should keep as it is so that future extensions will be easier. Please let me know.

Quoted 118 lines
Quoted 117 lines

+        s->regs[APB2OPB_IRQ_STS] |= opb ? APB2OPB_IRQ_STS_OPB1_TX_ACK +            : APB2OPB_IRQ_STS_OPB0_TX_ACK;
+        break;
+    }
+    }

hw/fsi/aspeed-apb2opb.c new file +39 -0
+
+    s->regs[TO_REG(addr)] = data;
+}
+
+static const struct MemoryRegionOps aspeed_apb2opb_ops = {
+    .read = fsi_aspeed_apb2opb_read,
+    .write = fsi_aspeed_apb2opb_write,
+    .valid.max_access_size = 4,
+    .valid.min_access_size = 4,
+    .impl.max_access_size = 4,
+    .impl.min_access_size = 4,
+    .endianness = DEVICE_LITTLE_ENDIAN,
+};
+
+static void fsi_aspeed_apb2opb_realize(DeviceState *dev, Error **errp)
+{
+    SysBusDevice *sbd = SYS_BUS_DEVICE(dev);
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(dev);
+
+    qbus_init(&s->opb[0], sizeof(s->opb[0]), TYPE_OP_BUS,
+                        DEVICE(s), NULL);
+    qbus_init(&s->opb[1], sizeof(s->opb[1]), TYPE_OP_BUS,
+                        DEVICE(s), NULL);
+
+    sysbus_init_irq(sbd, &s->irq);
+
+    memory_region_init_io(&s->iomem, OBJECT(s), &aspeed_apb2opb_ops, s,
+                          TYPE_ASPEED_APB2OPB, 0x1000);
+    sysbus_init_mmio(sbd, &s->iomem);
+}
+
+static void fsi_aspeed_apb2opb_reset(DeviceState *dev)
+{
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(dev);
+
+    memcpy(s->regs, aspeed_apb2opb_reset, ASPEED_APB2OPB_NR_REGS);
+}
+
+static void fsi_aspeed_apb2opb_class_init(ObjectClass *klass, void 

*data)

hw/fsi/aspeed-apb2opb.c new file +21 -0
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+
+    dc->desc = "ASPEED APB2OPB Bridge";
+    dc->realize = fsi_aspeed_apb2opb_realize;
+    dc->reset = fsi_aspeed_apb2opb_reset;
+}
+
+static const TypeInfo aspeed_apb2opb_info = {
+    .name = TYPE_ASPEED_APB2OPB,
+    .parent = TYPE_SYS_BUS_DEVICE,
+    .instance_size = sizeof(AspeedAPB2OPBState),
+    .class_init = fsi_aspeed_apb2opb_class_init,
+};
+
+static void aspeed_apb2opb_register_types(void)
+{
+    type_register_static(&aspeed_apb2opb_info);
+}
+
+type_init(aspeed_apb2opb_register_types);
hw/arm/Kconfig +0 -0
index 7e68348440..d963de74c9 100644
@@ -555,6 +555,7 @@ config ASPEED_SOC

      select LED
      select PMBUS
      select MAX31785
+    select FSI_APB2OPB_ASPEED
    config MPS2
      bool

hw/fsi/Kconfig +4 -0
index 0f6e6d331a..6bbcb8f6ca 100644
@@ -1,3 +1,7 @@
+config FSI_APB2OPB_ASPEED
+    bool
+    select FSI_OPB
+

  config FSI_OPB
      bool
      select FSI_CFAM

hw/fsi/meson.build +0 -0
index 407b8c2775..1bc6bb63cc 100644
@@ -3,3 +3,4 @@ system_ss.add(when: 'CONFIG_FSI_SCRATCHPAD', if_true: 

files('engine-scratchpad.c
  system_ss.add(when: 'CONFIG_FSI_CFAM', if_true: files('cfam.c')) system_ss.add(when: 'CONFIG_FSI', if_true: files('fsi.c','fsi-master.c','fsi-slave.c')) system_ss.add(when: 'CONFIG_FSI_OPB', if_true: files('opb.c')) +system_ss.add(when: 'CONFIG_FSI_APB2OPB_ASPEED', if_true:
files('aspeed-apb2opb.c'))

hw/fsi/trace-events +0 -0
index 13b211a815..9a45843eb6 100644
@@ -17,3 +17,5 @@ fsi_opb_write16(uint64_t addr, uint32_t size) 

"@0x%" PRIx64 " size=%d"
  fsi_opb_write32(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d" fsi_opb_unimplemented_read(uint64_t addr, uint32_t size) "@0x%" PRIx64 " size=%d" fsi_opb_unimplemented_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64 +fsi_aspeed_apb2opb_read(uint64_t addr, uint32_t size) "@0x%" PRIx64
" size=%d"
+fsi_aspeed_apb2opb_write(uint64_t addr, uint32_t size, uint64_t
data) "@0x%" PRIx64 " size=%d value=0x%"PRIx64

#15 lore raw

Re: [PATCH v6 06/10] hw/fsi: Aspeed APB2OPB interface

Quoted 341 lines · Ninad Palsule

Hello Cedric,

Quoted 162 lines · Cédric Le Goater
Quoted 161 lines · Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

An APB-to-OPB bridge enabling access to the OPB from the ARM core in the AST2600. Hardware limitations prevent the OPB from being directly mapped into APB, so all accesses are indirect through the bridge.

---
v2:
- Incorporated review comments by Joel
v3:
- Incorporated review comments by Thomas Huth
v4:
   - Compile FSI with ASPEED_SOC only.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric.
---
  include/hw/fsi/aspeed-apb2opb.h |  33 ++++ hw/fsi/aspeed-apb2opb.c         | 280 ++++++++++++++++++++++++++++++++ hw/arm/Kconfig                  |   1 +
  hw/fsi/Kconfig                  |   4 +
  hw/fsi/meson.build              |   1 + hw/fsi/trace-events             |   2 +

  6 files changed, 321 insertions(+)

  create mode 100644 include/hw/fsi/aspeed-apb2opb.h create mode 100644 hw/fsi/aspeed-apb2opb.c

include/hw/fsi/aspeed-apb2opb.h new file +33 -0
new file mode 100644
index 0000000000..a81ae67023
@@ -0,0 +1,33 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB2OPB Bridge
+ */
+#ifndef FSI_ASPEED_APB2OPB_H
+#define FSI_ASPEED_APB2OPB_H
+
+#include "hw/sysbus.h"
+#include "hw/fsi/opb.h"
+
+#define TYPE_ASPEED_APB2OPB "aspeed.apb2opb"
+OBJECT_DECLARE_SIMPLE_TYPE(AspeedAPB2OPBState, ASPEED_APB2OPB)
+
+#define ASPEED_APB2OPB_NR_REGS ((0xe8 >> 2) + 1)
+
+#define ASPEED_FSI_NUM 2
+
+typedef struct AspeedAPB2OPBState {
+    /*< private >*/
+    SysBusDevice parent_obj;
+
+    /*< public >*/
+    MemoryRegion iomem;
+
+    uint32_t regs[ASPEED_APB2OPB_NR_REGS];
+    qemu_irq irq;
+
+    OPBus opb[ASPEED_FSI_NUM];
+} AspeedAPB2OPBState;
+
+#endif /* FSI_ASPEED_APB2OPB_H */
hw/fsi/aspeed-apb2opb.c new file +85 -0
new file mode 100644
index 0000000000..6f97a6bc7d
@@ -0,0 +1,280 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB-OPB FSI interface
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/log.h"
+#include "qom/object.h"
+#include "qapi/error.h"
+#include "trace.h"
+
+#include "hw/fsi/aspeed-apb2opb.h"
+#include "hw/qdev-core.h"
+
+#define TO_REG(x) (x >> 2)
+#define GENMASK(t, b) (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) - 1))
+
+#define APB2OPB_VERSION                    TO_REG(0x00)
+#define   APB2OPB_VERSION_VER              GENMASK(7, 0)
+
+#define APB2OPB_TRIGGER                    TO_REG(0x04)
+#define   APB2OPB_TRIGGER_EN               BIT(0)
+
+#define APB2OPB_CONTROL                    TO_REG(0x08)
+#define   APB2OPB_CONTROL_OFF              GENMASK(31, 13)
+
+#define APB2OPB_OPB2FSI                    TO_REG(0x0c)
+#define   APB2OPB_OPB2FSI_OFF              GENMASK(31, 22)
+
+#define APB2OPB_OPB0_SEL                   TO_REG(0x10)
+#define APB2OPB_OPB1_SEL                   TO_REG(0x28)
+#define   APB2OPB_OPB_SEL_EN               BIT(0)
+
+#define APB2OPB_OPB0_MODE                  TO_REG(0x14)
+#define APB2OPB_OPB1_MODE                  TO_REG(0x2c)
+#define   APB2OPB_OPB_MODE_RD              BIT(0)
+
+#define APB2OPB_OPB0_XFER                  TO_REG(0x18)
+#define APB2OPB_OPB1_XFER                  TO_REG(0x30)
+#define   APB2OPB_OPB_XFER_FULL            BIT(1)
+#define   APB2OPB_OPB_XFER_HALF            BIT(0)
+
+#define APB2OPB_OPB0_ADDR                  TO_REG(0x1c)
+#define APB2OPB_OPB0_WRITE_DATA            TO_REG(0x20)
+
+#define APB2OPB_OPB1_ADDR                  TO_REG(0x34)
+#define APB2OPB_OPB1_WRITE_DATA                  TO_REG(0x38)
+
+#define APB2OPB_IRQ_STS                    TO_REG(0x48)
+#define   APB2OPB_IRQ_STS_OPB1_TX_ACK      BIT(17)
+#define   APB2OPB_IRQ_STS_OPB0_TX_ACK      BIT(16)
+
+#define APB2OPB_OPB0_WRITE_WORD_ENDIAN     TO_REG(0x4c)
+#define   APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE 0x0011101b
+#define APB2OPB_OPB0_WRITE_BYTE_ENDIAN     TO_REG(0x50)
+#define   APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE 0x0c330f3f
+#define APB2OPB_OPB1_WRITE_WORD_ENDIAN     TO_REG(0x54)
+#define APB2OPB_OPB1_WRITE_BYTE_ENDIAN     TO_REG(0x58)
+#define APB2OPB_OPB0_READ_BYTE_ENDIAN      TO_REG(0x5c)
+#define APB2OPB_OPB1_READ_BYTE_ENDIAN      TO_REG(0x60)
+#define   APB2OPB_OPB0_READ_WORD_ENDIAN_BE  0x00030b1b
+
+#define APB2OPB_OPB0_READ_DATA         TO_REG(0x84)
+#define APB2OPB_OPB1_READ_DATA         TO_REG(0x90)
+
+/*
+ * The following magic values came from AST2600 data sheet
+ * The register values are defined under section "FSI controller"
+ * as initial values.
+ */
+static const uint32_t aspeed_apb2opb_reset[ASPEED_APB2OPB_NR_REGS] = {
+     [APB2OPB_VERSION]                = 0x000000a1,
+     [APB2OPB_OPB0_WRITE_WORD_ENDIAN] = 0x0044eee4,
+     [APB2OPB_OPB0_WRITE_BYTE_ENDIAN] = 0x0055aaff,
+     [APB2OPB_OPB1_WRITE_WORD_ENDIAN] = 0x00117717,
+     [APB2OPB_OPB1_WRITE_BYTE_ENDIAN] = 0xffaa5500,
+     [APB2OPB_OPB0_READ_BYTE_ENDIAN]  = 0x0044eee4,
+     [APB2OPB_OPB1_READ_BYTE_ENDIAN]  = 0x00117717
+};
+
+static uint64_t fsi_aspeed_apb2opb_read(void *opaque, hwaddr addr,
+                                        unsigned size)
+{
Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 90

Quoted 5 lines
hw/fsi/aspeed-apb2opb.c new file +5 -0
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);
+
+    trace_fsi_aspeed_apb2opb_read(addr, size);
+
+    if (addr + size > sizeof(s->regs)) {

hmm, the parameter 'size' is a memop transaction size not an address offset.

OK, Changed it to validate the register (index) instead of addr + size.

Quoted 12 lines
Quoted 12 lines
hw/fsi/aspeed-apb2opb.c new file +11 -0
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds read: 0x%"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return 0;
+    }
+
+    return s->regs[TO_REG(addr)];
+}
+
+static void fsi_aspeed_apb2opb_write(void *opaque, hwaddr addr, uint64_t data,
+                                     unsigned size)
Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 107

Quoted 6 lines
hw/fsi/aspeed-apb2opb.c new file +6 -0
+{
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);
+
+    trace_fsi_aspeed_apb2opb_write(addr, size, data);
+
+    if (addr + size > sizeof(s->regs)) {

same comment.

Fixed it same as above.

Quoted 12 lines

hw/fsi/aspeed-apb2opb.c line 116

Quoted 11 lines
hw/fsi/aspeed-apb2opb.c new file +9 -0
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds write: %"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return;
+    }
+
+    switch (TO_REG(addr)) {
+    case APB2OPB_CONTROL:
+        fsi_opb_fsi_master_address(&s->opb[0], data & APB2OPB_CONTROL_OFF);
Quoted 1 line

fsi_opb_fsi_master_address() should statically defined in this file

We have separation of OPB bus implementation and APB2OPB interface. If we move this function here then we will be exposing OPB implementation here.

Quoted 7 lines

hw/fsi/aspeed-apb2opb.c line 119

Quoted 4 lines
hw/fsi/aspeed-apb2opb.c new file +3 -0
+        break;
+    case APB2OPB_OPB2FSI:
+        fsi_opb_opb2fsi_address(&s->opb[0], data & APB2OPB_OPB2FSI_OFF);

same for fsi_opb_opb2fsi_address()

Same as above.

Quoted 46 lines
Quoted 46 lines
hw/fsi/aspeed-apb2opb.c new file +45 -0
+        break;
+    case APB2OPB_OPB0_WRITE_WORD_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_OPB0_WRITE_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_OPB0_READ_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_READ_WORD_ENDIAN_BE) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_READ_WORD_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_TRIGGER:
+    {
+        uint32_t opb, op_mode, op_size, op_addr, op_data;
+
+        assert((s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) ^
+               (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN));
+
+        if (s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 0;
+            op_mode = s->regs[APB2OPB_OPB0_MODE];
+            op_size = s->regs[APB2OPB_OPB0_XFER];
+            op_addr = s->regs[APB2OPB_OPB0_ADDR];
+            op_data = s->regs[APB2OPB_OPB0_WRITE_DATA];
+        } else if (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 1;
+            op_mode = s->regs[APB2OPB_OPB1_MODE];
+            op_size = s->regs[APB2OPB_OPB1_XFER];
+            op_addr = s->regs[APB2OPB_OPB1_ADDR];
+            op_data = s->regs[APB2OPB_OPB1_WRITE_DATA];
+        } else {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Invalid operation: 0x%"HWADDR_PRIx" for %u\n",
+                          __func__, addr, size);
Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 170

Quoted 6 lines
hw/fsi/aspeed-apb2opb.c new file +6 -0
+            return;
+        }
+
+        if (op_size & ~(APB2OPB_OPB_XFER_HALF | APB2OPB_OPB_XFER_FULL)) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "OPB transaction failed: Unrecognised access width: %d\n",

Unrecognized

Fixed

Quoted 20 lines
Quoted 20 lines
hw/fsi/aspeed-apb2opb.c new file +19 -0
+                          op_size);
+            return;
+        }
+
+        op_size += 1;
+
+        if (op_mode & APB2OPB_OPB_MODE_RD) {
+            int index = opb ? APB2OPB_OPB1_READ_DATA
+                : APB2OPB_OPB0_READ_DATA;
+
+            switch (op_size) {
+            case 1:
+                s->regs[index] = fsi_opb_read8(&s->opb[opb], op_addr);
+                break;
+            case 2:
+                s->regs[index] = fsi_opb_read16(&s->opb[opb], op_addr);
+                break;
+            case 4:
+                s->regs[index] = fsi_opb_read32(&s->opb[opb], op_addr);
Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 194

Quoted 5 lines
hw/fsi/aspeed-apb2opb.c new file +5 -0
+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n",
+                              __func__, size);

this should use op_size and not size and seems redudant with the unrecognized test above.

true, Keeping it in case bits meaning change in future.

Quoted 40 lines
Quoted 13 lines
hw/fsi/aspeed-apb2opb.c new file +11 -0
+                return;
+            }
+        } else {
+            /* FIXME: Endian swizzling */
+            switch (op_size) {
+            case 1:
+                fsi_opb_write8(&s->opb[opb], op_addr, op_data);
+                break;
+            case 2:
+                fsi_opb_write16(&s->opb[opb], op_addr, op_data);
+                break;

hw/fsi/aspeed-apb2opb.c line 215

Quoted 10 lines
hw/fsi/aspeed-apb2opb.c new file +10 -0
+            case 4:
+                fsi_opb_write32(&s->opb[opb], op_addr, op_data);
+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n",
+                              __func__, op_size);
+                return;
+            }
+        }

The above is equivalent to :

        MemTxResult result;
        bool is_write = !(op_mode & APB2OPB_OPB_MODE_RD);
        int index = opb ? APB2OPB_OPB1_READ_DATA : APB2OPB_OPB0_READ_DATA; AddressSpace *as = &s->opb[opb].as;

        result = address_space_rw(as, op_addr, MEMTXATTRS_UNSPECIFIED, &op_data, op_size, is_write); if (result != MEMTX_OK) { qemu_log_mask(LOG_GUEST_ERROR, "%s: OPB %s failed @%08x\n", __func__, is_write ? "write" : "read", op_addr); return;
        }

Quoted 6 lines
Quoted 5 lines

        if (!is_write) {
            s->regs[index] = op_data;
        }

and the fsi_opb_* routines are useless to me.

We are trying to keep the separation between OPB implementation and interface hence we have all those fsi_opb_*. I feel that we should keep as it is so that future extensions will be easier. Please let me know.

Well, I can't really tell because I don't know enough about FSI :/

The models look fragile and I have spent already a lot of time trying to untangle what they are trying to do. Please ask your teammates or let's see in the next QEMU cycle.

Thanks,

C.

#16 lore raw

Re: [PATCH v6 06/10] hw/fsi: Aspeed APB2OPB interface

Hello Cedric & Andrew,

Quoted 366 lines · Cédric Le Goater
Quoted 365 lines · Ninad Palsule

Hello Cedric,

Quoted 165 lines · Cédric Le Goater
Quoted 164 lines · Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

An APB-to-OPB bridge enabling access to the OPB from the ARM core in the AST2600. Hardware limitations prevent the OPB from being directly mapped into APB, so all accesses are indirect through the bridge.

---
v2:
- Incorporated review comments by Joel
v3:
- Incorporated review comments by Thomas Huth
v4:
   - Compile FSI with ASPEED_SOC only.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric.
---
  include/hw/fsi/aspeed-apb2opb.h |  33 ++++
  hw/fsi/aspeed-apb2opb.c         | 280
++++++++++++++++++++++++++++++++
  hw/arm/Kconfig                  |   1 +
  hw/fsi/Kconfig                  |   4 +
  hw/fsi/meson.build              |   1 +
  hw/fsi/trace-events             |   2 +

  6 files changed, 321 insertions(+)

  create mode 100644 include/hw/fsi/aspeed-apb2opb.h create mode 100644 hw/fsi/aspeed-apb2opb.c

include/hw/fsi/aspeed-apb2opb.h new file +33 -0
new file mode 100644
index 0000000000..a81ae67023
@@ -0,0 +1,33 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB2OPB Bridge
+ */
+#ifndef FSI_ASPEED_APB2OPB_H
+#define FSI_ASPEED_APB2OPB_H
+
+#include "hw/sysbus.h"
+#include "hw/fsi/opb.h"
+
+#define TYPE_ASPEED_APB2OPB "aspeed.apb2opb"
+OBJECT_DECLARE_SIMPLE_TYPE(AspeedAPB2OPBState, ASPEED_APB2OPB)
+
+#define ASPEED_APB2OPB_NR_REGS ((0xe8 >> 2) + 1)
+
+#define ASPEED_FSI_NUM 2
+
+typedef struct AspeedAPB2OPBState {
+    /*< private >*/
+    SysBusDevice parent_obj;
+
+    /*< public >*/
+    MemoryRegion iomem;
+
+    uint32_t regs[ASPEED_APB2OPB_NR_REGS];
+    qemu_irq irq;
+
+    OPBus opb[ASPEED_FSI_NUM];
+} AspeedAPB2OPBState;
+
+#endif /* FSI_ASPEED_APB2OPB_H */
hw/fsi/aspeed-apb2opb.c new file +73 -1
new file mode 100644
index 0000000000..6f97a6bc7d
@@ -0,0 +1,280 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB-OPB FSI interface
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/log.h"
+#include "qom/object.h"
+#include "qapi/error.h"
+#include "trace.h"
+
+#include "hw/fsi/aspeed-apb2opb.h"
+#include "hw/qdev-core.h"
+
+#define TO_REG(x) (x >> 2)
+#define GENMASK(t, b) (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) 
- 1))
+
+#define APB2OPB_VERSION                    TO_REG(0x00)
+#define   APB2OPB_VERSION_VER              GENMASK(7, 0)
+
+#define APB2OPB_TRIGGER                    TO_REG(0x04)
+#define   APB2OPB_TRIGGER_EN               BIT(0)
+
+#define APB2OPB_CONTROL                    TO_REG(0x08)
+#define   APB2OPB_CONTROL_OFF              GENMASK(31, 13)
+
+#define APB2OPB_OPB2FSI                    TO_REG(0x0c)
+#define   APB2OPB_OPB2FSI_OFF              GENMASK(31, 22)
+
+#define APB2OPB_OPB0_SEL                   TO_REG(0x10)
+#define APB2OPB_OPB1_SEL                   TO_REG(0x28)
+#define   APB2OPB_OPB_SEL_EN               BIT(0)
+
+#define APB2OPB_OPB0_MODE                  TO_REG(0x14)
+#define APB2OPB_OPB1_MODE                  TO_REG(0x2c)
+#define   APB2OPB_OPB_MODE_RD              BIT(0)
+
+#define APB2OPB_OPB0_XFER                  TO_REG(0x18)
+#define APB2OPB_OPB1_XFER                  TO_REG(0x30)
+#define   APB2OPB_OPB_XFER_FULL            BIT(1)
+#define   APB2OPB_OPB_XFER_HALF            BIT(0)
+
+#define APB2OPB_OPB0_ADDR                  TO_REG(0x1c)
+#define APB2OPB_OPB0_WRITE_DATA            TO_REG(0x20)
+
+#define APB2OPB_OPB1_ADDR                  TO_REG(0x34)
+#define APB2OPB_OPB1_WRITE_DATA TO_REG(0x38)
+
+#define APB2OPB_IRQ_STS                    TO_REG(0x48)
+#define   APB2OPB_IRQ_STS_OPB1_TX_ACK      BIT(17)
+#define   APB2OPB_IRQ_STS_OPB0_TX_ACK      BIT(16)
+
+#define APB2OPB_OPB0_WRITE_WORD_ENDIAN     TO_REG(0x4c)
+#define   APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE 0x0011101b
+#define APB2OPB_OPB0_WRITE_BYTE_ENDIAN     TO_REG(0x50)
+#define   APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE 0x0c330f3f
+#define APB2OPB_OPB1_WRITE_WORD_ENDIAN     TO_REG(0x54)
+#define APB2OPB_OPB1_WRITE_BYTE_ENDIAN     TO_REG(0x58)
+#define APB2OPB_OPB0_READ_BYTE_ENDIAN      TO_REG(0x5c)
+#define APB2OPB_OPB1_READ_BYTE_ENDIAN      TO_REG(0x60)
+#define   APB2OPB_OPB0_READ_WORD_ENDIAN_BE  0x00030b1b
+
+#define APB2OPB_OPB0_READ_DATA         TO_REG(0x84)
+#define APB2OPB_OPB1_READ_DATA         TO_REG(0x90)
+
+/*
+ * The following magic values came from AST2600 data sheet
+ * The register values are defined under section "FSI controller"
+ * as initial values.
+ */
+static const uint32_t aspeed_apb2opb_reset[ASPEED_APB2OPB_NR_REGS] 

= {
+     [APB2OPB_VERSION]                = 0x000000a1, +     [APB2OPB_OPB0_WRITE_WORD_ENDIAN] = 0x0044eee4, +     [APB2OPB_OPB0_WRITE_BYTE_ENDIAN] = 0x0055aaff, +     [APB2OPB_OPB1_WRITE_WORD_ENDIAN] = 0x00117717, +     [APB2OPB_OPB1_WRITE_BYTE_ENDIAN] = 0xffaa5500, +     [APB2OPB_OPB0_READ_BYTE_ENDIAN]  = 0x0044eee4, +     [APB2OPB_OPB1_READ_BYTE_ENDIAN]  = 0x00117717

hw/fsi/aspeed-apb2opb.c new file +5 -0
+};
+
+static uint64_t fsi_aspeed_apb2opb_read(void *opaque, hwaddr addr,
+                                        unsigned size)
+{
Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 90

Quoted 5 lines

+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);

hw/fsi/aspeed-apb2opb.c new file +4 -0
+
+    trace_fsi_aspeed_apb2opb_read(addr, size);
+
+    if (addr + size > sizeof(s->regs)) {

hmm, the parameter 'size' is a memop transaction size not an address offset.

OK, Changed it to validate the register (index) instead of addr + size.

Quoted 14 lines
Quoted 14 lines

+ qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds read: 0x%"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return 0;
+    }

hw/fsi/aspeed-apb2opb.c new file +5 -0
+
+    return s->regs[TO_REG(addr)];
+}
+
+static void fsi_aspeed_apb2opb_write(void *opaque, hwaddr addr, 

uint64_t data,
+                                     unsigned size)

Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 107

Quoted 6 lines
hw/fsi/aspeed-apb2opb.c new file +6 -0
+{
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);
+
+    trace_fsi_aspeed_apb2opb_write(addr, size, data);
+
+    if (addr + size > sizeof(s->regs)) {

same comment.

Fixed it same as above.

Quoted 14 lines
Quoted 13 lines

+ qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds write: %"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return;
+    }
+
+    switch (TO_REG(addr)) {
+    case APB2OPB_CONTROL:
+        fsi_opb_fsi_master_address(&s->opb[0], data & APB2OPB_CONTROL_OFF);

Quoted 1 line

fsi_opb_fsi_master_address() should statically defined in this file

We have separation of OPB bus implementation and APB2OPB interface. If we move this function here then we will be exposing OPB implementation here.

Quoted 8 lines
Quoted 5 lines

+        break;
+    case APB2OPB_OPB2FSI:
+        fsi_opb_opb2fsi_address(&s->opb[0], data & APB2OPB_OPB2FSI_OFF);

same for fsi_opb_opb2fsi_address()

Same as above.

Quoted 55 lines
Quoted 55 lines

+        break;
+    case APB2OPB_OPB0_WRITE_WORD_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE) { +            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_OPB0_WRITE_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE) { +            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_OPB0_READ_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_READ_WORD_ENDIAN_BE) { +            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_READ_WORD_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_TRIGGER:
+    {
+        uint32_t opb, op_mode, op_size, op_addr, op_data;

hw/fsi/aspeed-apb2opb.c new file +19 -0
+
+        assert((s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) ^
+               (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN));
+
+        if (s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 0;
+            op_mode = s->regs[APB2OPB_OPB0_MODE];
+            op_size = s->regs[APB2OPB_OPB0_XFER];
+            op_addr = s->regs[APB2OPB_OPB0_ADDR];
+            op_data = s->regs[APB2OPB_OPB0_WRITE_DATA];
+        } else if (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 1;
+            op_mode = s->regs[APB2OPB_OPB1_MODE];
+            op_size = s->regs[APB2OPB_OPB1_XFER];
+            op_addr = s->regs[APB2OPB_OPB1_ADDR];
+            op_data = s->regs[APB2OPB_OPB1_WRITE_DATA];
+        } else {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Invalid operation: 0x%"HWADDR_PRIx" 

for %u\n",
+                          __func__, addr, size);
+            return;
+        }

Quoted 8 lines
Quoted 6 lines

+
+        if (op_size & ~(APB2OPB_OPB_XFER_HALF | APB2OPB_OPB_XFER_FULL)) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "OPB transaction failed: Unrecognised access width: %d\n",

Unrecognized

Fixed

Quoted 23 lines
Quoted 23 lines

+                          op_size);
+            return;
+        }

hw/fsi/aspeed-apb2opb.c new file +10 -0
+
+        op_size += 1;
+
+        if (op_mode & APB2OPB_OPB_MODE_RD) {
+            int index = opb ? APB2OPB_OPB1_READ_DATA
+                : APB2OPB_OPB0_READ_DATA;
+
+            switch (op_size) {
+            case 1:
+                s->regs[index] = fsi_opb_read8(&s->opb[opb], 

op_addr);
+                break;
+            case 2:
+                s->regs[index] = fsi_opb_read16(&s->opb[opb], op_addr);
+                break;
+            case 4:
+                s->regs[index] = fsi_opb_read32(&s->opb[opb], op_addr);

Quoted 8 lines
Quoted 5 lines

+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n", +                              __func__, size);

this should use op_size and not size and seems redudant with the unrecognized test above.

true, Keeping it in case bits meaning change in future.

Quoted 44 lines
Quoted 15 lines

+                return;
+            }
+        } else {
+            /* FIXME: Endian swizzling */
+            switch (op_size) {
+            case 1:
+                fsi_opb_write8(&s->opb[opb], op_addr, op_data); +                break;
+            case 2:
+                fsi_opb_write16(&s->opb[opb], op_addr, op_data); +                break;
+            case 4:
+                fsi_opb_write32(&s->opb[opb], op_addr, op_data);

Quoted 8 lines

+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n", +                              __func__, op_size);
+                return;
+            }
+        }

The above is equivalent to :

        MemTxResult result;
        bool is_write = !(op_mode & APB2OPB_OPB_MODE_RD);
        int index = opb ? APB2OPB_OPB1_READ_DATA : APB2OPB_OPB0_READ_DATA; AddressSpace *as = &s->opb[opb].as;

        result = address_space_rw(as, op_addr, MEMTXATTRS_UNSPECIFIED, &op_data, op_size, is_write); if (result != MEMTX_OK) { qemu_log_mask(LOG_GUEST_ERROR, "%s: OPB %s failed @%08x\n", __func__, is_write ? "write" : "read", op_addr); return;
        }

        if (!is_write) {
            s->regs[index] = op_data;
        }

Quoted 8 lines
Quoted 2 lines
Quoted 1 line

and the fsi_opb_* routines are useless to me.

We are trying to keep the separation between OPB implementation and interface hence we have all those fsi_opb_*. I feel that we should keep as it is so that future extensions will be easier. Please let me know.

Well, I can't really tell because I don't know enough about FSI :/

The models look fragile and I have spent already a lot of time trying to untangle what they are trying to do. Please ask your teammates or let's see in the next QEMU cycle.

Thanks for helping me to get in this patch.

On chip peripherals are connected to OPB bus for easier access. All FSI masters (including cascaded) will be attached to OPB bus but we may never implement it. So the current model is only instantiating bus and having wrapper around mmio. Are you suggesting to simplify it by not having this indirection as its not providing much? If that's the case then I can change as per your suggestion.

Andrew, Do you have any thoughts about this?

Quoted 6 lines

Thanks,

C.

#17 lore raw

Re: [PATCH v6 06/10] hw/fsi: Aspeed APB2OPB interface

In reply to Cédric Le Goater

Hello Cedric,

Quoted 215 lines · Cédric Le Goater
Quoted 214 lines · Ninad Palsule

Hello Cedric,

Quoted 165 lines · Cédric Le Goater
Quoted 164 lines · Ninad Palsule

This is a part of patchset where IBM's Flexible Service Interface is introduced.

An APB-to-OPB bridge enabling access to the OPB from the ARM core in the AST2600. Hardware limitations prevent the OPB from being directly mapped into APB, so all accesses are indirect through the bridge.

---
v2:
- Incorporated review comments by Joel
v3:
- Incorporated review comments by Thomas Huth
v4:
   - Compile FSI with ASPEED_SOC only.
v5:
- Incorporated review comments by Cedric.
v6:
- Incorporated review comments by Cedric.
---
  include/hw/fsi/aspeed-apb2opb.h |  33 ++++
  hw/fsi/aspeed-apb2opb.c         | 280
++++++++++++++++++++++++++++++++
  hw/arm/Kconfig                  |   1 +
  hw/fsi/Kconfig                  |   4 +
  hw/fsi/meson.build              |   1 +
  hw/fsi/trace-events             |   2 +

  6 files changed, 321 insertions(+)

  create mode 100644 include/hw/fsi/aspeed-apb2opb.h create mode 100644 hw/fsi/aspeed-apb2opb.c

include/hw/fsi/aspeed-apb2opb.h new file +33 -0
new file mode 100644
index 0000000000..a81ae67023
@@ -0,0 +1,33 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB2OPB Bridge
+ */
+#ifndef FSI_ASPEED_APB2OPB_H
+#define FSI_ASPEED_APB2OPB_H
+
+#include "hw/sysbus.h"
+#include "hw/fsi/opb.h"
+
+#define TYPE_ASPEED_APB2OPB "aspeed.apb2opb"
+OBJECT_DECLARE_SIMPLE_TYPE(AspeedAPB2OPBState, ASPEED_APB2OPB)
+
+#define ASPEED_APB2OPB_NR_REGS ((0xe8 >> 2) + 1)
+
+#define ASPEED_FSI_NUM 2
+
+typedef struct AspeedAPB2OPBState {
+    /*< private >*/
+    SysBusDevice parent_obj;
+
+    /*< public >*/
+    MemoryRegion iomem;
+
+    uint32_t regs[ASPEED_APB2OPB_NR_REGS];
+    qemu_irq irq;
+
+    OPBus opb[ASPEED_FSI_NUM];
+} AspeedAPB2OPBState;
+
+#endif /* FSI_ASPEED_APB2OPB_H */
hw/fsi/aspeed-apb2opb.c new file +73 -1
new file mode 100644
index 0000000000..6f97a6bc7d
@@ -0,0 +1,280 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Copyright (C) 2023 IBM Corp.
+ *
+ * ASPEED APB-OPB FSI interface
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/log.h"
+#include "qom/object.h"
+#include "qapi/error.h"
+#include "trace.h"
+
+#include "hw/fsi/aspeed-apb2opb.h"
+#include "hw/qdev-core.h"
+
+#define TO_REG(x) (x >> 2)
+#define GENMASK(t, b) (((1ULL << ((t) + 1)) - 1) & ~((1ULL << (b)) 
- 1))
+
+#define APB2OPB_VERSION                    TO_REG(0x00)
+#define   APB2OPB_VERSION_VER              GENMASK(7, 0)
+
+#define APB2OPB_TRIGGER                    TO_REG(0x04)
+#define   APB2OPB_TRIGGER_EN               BIT(0)
+
+#define APB2OPB_CONTROL                    TO_REG(0x08)
+#define   APB2OPB_CONTROL_OFF              GENMASK(31, 13)
+
+#define APB2OPB_OPB2FSI                    TO_REG(0x0c)
+#define   APB2OPB_OPB2FSI_OFF              GENMASK(31, 22)
+
+#define APB2OPB_OPB0_SEL                   TO_REG(0x10)
+#define APB2OPB_OPB1_SEL                   TO_REG(0x28)
+#define   APB2OPB_OPB_SEL_EN               BIT(0)
+
+#define APB2OPB_OPB0_MODE                  TO_REG(0x14)
+#define APB2OPB_OPB1_MODE                  TO_REG(0x2c)
+#define   APB2OPB_OPB_MODE_RD              BIT(0)
+
+#define APB2OPB_OPB0_XFER                  TO_REG(0x18)
+#define APB2OPB_OPB1_XFER                  TO_REG(0x30)
+#define   APB2OPB_OPB_XFER_FULL            BIT(1)
+#define   APB2OPB_OPB_XFER_HALF            BIT(0)
+
+#define APB2OPB_OPB0_ADDR                  TO_REG(0x1c)
+#define APB2OPB_OPB0_WRITE_DATA            TO_REG(0x20)
+
+#define APB2OPB_OPB1_ADDR                  TO_REG(0x34)
+#define APB2OPB_OPB1_WRITE_DATA TO_REG(0x38)
+
+#define APB2OPB_IRQ_STS                    TO_REG(0x48)
+#define   APB2OPB_IRQ_STS_OPB1_TX_ACK      BIT(17)
+#define   APB2OPB_IRQ_STS_OPB0_TX_ACK      BIT(16)
+
+#define APB2OPB_OPB0_WRITE_WORD_ENDIAN     TO_REG(0x4c)
+#define   APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE 0x0011101b
+#define APB2OPB_OPB0_WRITE_BYTE_ENDIAN     TO_REG(0x50)
+#define   APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE 0x0c330f3f
+#define APB2OPB_OPB1_WRITE_WORD_ENDIAN     TO_REG(0x54)
+#define APB2OPB_OPB1_WRITE_BYTE_ENDIAN     TO_REG(0x58)
+#define APB2OPB_OPB0_READ_BYTE_ENDIAN      TO_REG(0x5c)
+#define APB2OPB_OPB1_READ_BYTE_ENDIAN      TO_REG(0x60)
+#define   APB2OPB_OPB0_READ_WORD_ENDIAN_BE  0x00030b1b
+
+#define APB2OPB_OPB0_READ_DATA         TO_REG(0x84)
+#define APB2OPB_OPB1_READ_DATA         TO_REG(0x90)
+
+/*
+ * The following magic values came from AST2600 data sheet
+ * The register values are defined under section "FSI controller"
+ * as initial values.
+ */
+static const uint32_t aspeed_apb2opb_reset[ASPEED_APB2OPB_NR_REGS] 

= {
+     [APB2OPB_VERSION]                = 0x000000a1, +     [APB2OPB_OPB0_WRITE_WORD_ENDIAN] = 0x0044eee4, +     [APB2OPB_OPB0_WRITE_BYTE_ENDIAN] = 0x0055aaff, +     [APB2OPB_OPB1_WRITE_WORD_ENDIAN] = 0x00117717, +     [APB2OPB_OPB1_WRITE_BYTE_ENDIAN] = 0xffaa5500, +     [APB2OPB_OPB0_READ_BYTE_ENDIAN]  = 0x0044eee4, +     [APB2OPB_OPB1_READ_BYTE_ENDIAN]  = 0x00117717

hw/fsi/aspeed-apb2opb.c new file +5 -0
+};
+
+static uint64_t fsi_aspeed_apb2opb_read(void *opaque, hwaddr addr,
+                                        unsigned size)
+{
Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 90

Quoted 5 lines

+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);

hw/fsi/aspeed-apb2opb.c new file +4 -0
+
+    trace_fsi_aspeed_apb2opb_read(addr, size);
+
+    if (addr + size > sizeof(s->regs)) {

hmm, the parameter 'size' is a memop transaction size not an address offset.

OK, Changed it to validate the register (index) instead of addr + size.

Quoted 14 lines
Quoted 14 lines

+ qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds read: 0x%"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return 0;
+    }

hw/fsi/aspeed-apb2opb.c new file +5 -0
+
+    return s->regs[TO_REG(addr)];
+}
+
+static void fsi_aspeed_apb2opb_write(void *opaque, hwaddr addr, 

uint64_t data,
+                                     unsigned size)

Quoted 8 lines

hw/fsi/aspeed-apb2opb.c line 107

Quoted 6 lines
hw/fsi/aspeed-apb2opb.c new file +6 -0
+{
+    AspeedAPB2OPBState *s = ASPEED_APB2OPB(opaque);
+
+    trace_fsi_aspeed_apb2opb_write(addr, size, data);
+
+    if (addr + size > sizeof(s->regs)) {

same comment.

Fixed it same as above.

Quoted 14 lines
Quoted 13 lines

+ qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Out of bounds write: %"HWADDR_PRIx" for %u\n",
+                      __func__, addr, size);
+        return;
+    }
+
+    switch (TO_REG(addr)) {
+    case APB2OPB_CONTROL:
+        fsi_opb_fsi_master_address(&s->opb[0], data & APB2OPB_CONTROL_OFF);

Quoted 2 lines
Quoted 2 lines
Quoted 1 line

fsi_opb_fsi_master_address() should statically defined in this file

We have separation of OPB bus implementation and APB2OPB interface. If we move this function here then we will be exposing OPB implementation here.

Defined a static function and removed from opb.c

Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
Quoted 5 lines

+        break;
+    case APB2OPB_OPB2FSI:
+        fsi_opb_opb2fsi_address(&s->opb[0], data & APB2OPB_OPB2FSI_OFF);

same for fsi_opb_opb2fsi_address()

Same as above.

Same as above.

Quoted 140 lines
Quoted 140 lines
Quoted 55 lines
Quoted 55 lines

+        break;
+    case APB2OPB_OPB0_WRITE_WORD_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE) { +            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_WORD_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_OPB0_WRITE_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE) { +            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_WRITE_BYTE_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_OPB0_READ_BYTE_ENDIAN:
+        if (data != APB2OPB_OPB0_READ_WORD_ENDIAN_BE) { +            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Bridge needs to be driven as BE (0x%x)\n",
+                          __func__, APB2OPB_OPB0_READ_WORD_ENDIAN_BE);
+        }
+        break;
+    case APB2OPB_TRIGGER:
+    {
+        uint32_t opb, op_mode, op_size, op_addr, op_data;

hw/fsi/aspeed-apb2opb.c new file +19 -0
+
+        assert((s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) ^
+               (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN));
+
+        if (s->regs[APB2OPB_OPB0_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 0;
+            op_mode = s->regs[APB2OPB_OPB0_MODE];
+            op_size = s->regs[APB2OPB_OPB0_XFER];
+            op_addr = s->regs[APB2OPB_OPB0_ADDR];
+            op_data = s->regs[APB2OPB_OPB0_WRITE_DATA];
+        } else if (s->regs[APB2OPB_OPB1_SEL] & APB2OPB_OPB_SEL_EN) {
+            opb = 1;
+            op_mode = s->regs[APB2OPB_OPB1_MODE];
+            op_size = s->regs[APB2OPB_OPB1_XFER];
+            op_addr = s->regs[APB2OPB_OPB1_ADDR];
+            op_data = s->regs[APB2OPB_OPB1_WRITE_DATA];
+        } else {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: Invalid operation: 0x%"HWADDR_PRIx" 

for %u\n",
+                          __func__, addr, size);
+            return;
+        }

Quoted 8 lines
Quoted 6 lines

+
+        if (op_size & ~(APB2OPB_OPB_XFER_HALF | APB2OPB_OPB_XFER_FULL)) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "OPB transaction failed: Unrecognised access width: %d\n",

Unrecognized

Fixed

Quoted 23 lines
Quoted 23 lines

+                          op_size);
+            return;
+        }

hw/fsi/aspeed-apb2opb.c new file +10 -0
+
+        op_size += 1;
+
+        if (op_mode & APB2OPB_OPB_MODE_RD) {
+            int index = opb ? APB2OPB_OPB1_READ_DATA
+                : APB2OPB_OPB0_READ_DATA;
+
+            switch (op_size) {
+            case 1:
+                s->regs[index] = fsi_opb_read8(&s->opb[opb], 

op_addr);
+                break;
+            case 2:
+                s->regs[index] = fsi_opb_read16(&s->opb[opb], op_addr);
+                break;
+            case 4:
+                s->regs[index] = fsi_opb_read32(&s->opb[opb], op_addr);

Quoted 8 lines
Quoted 5 lines

+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n", +                              __func__, size);

this should use op_size and not size and seems redudant with the unrecognized test above.

true, Keeping it in case bits meaning change in future.

Quoted 44 lines
Quoted 15 lines

+                return;
+            }
+        } else {
+            /* FIXME: Endian swizzling */
+            switch (op_size) {
+            case 1:
+                fsi_opb_write8(&s->opb[opb], op_addr, op_data); +                break;
+            case 2:
+                fsi_opb_write16(&s->opb[opb], op_addr, op_data); +                break;
+            case 4:
+                fsi_opb_write32(&s->opb[opb], op_addr, op_data);

Quoted 8 lines

+                break;
+            default:
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "%s: Size not supported: %u\n", +                              __func__, op_size);
+                return;
+            }
+        }

The above is equivalent to :

        MemTxResult result;
        bool is_write = !(op_mode & APB2OPB_OPB_MODE_RD);
        int index = opb ? APB2OPB_OPB1_READ_DATA : APB2OPB_OPB0_READ_DATA; AddressSpace *as = &s->opb[opb].as;

        result = address_space_rw(as, op_addr, MEMTXATTRS_UNSPECIFIED, &op_data, op_size, is_write); if (result != MEMTX_OK) { qemu_log_mask(LOG_GUEST_ERROR, "%s: OPB %s failed @%08x\n", __func__, is_write ? "write" : "read", op_addr); return;
        }

        if (!is_write) {
            s->regs[index] = op_data;
        }

Quoted 8 lines
Quoted 2 lines
Quoted 1 line

and the fsi_opb_* routines are useless to me.

We are trying to keep the separation between OPB implementation and interface hence we have all those fsi_opb_*. I feel that we should keep as it is so that future extensions will be easier. Please let me know.

Well, I can't really tell because I don't know enough about FSI :/

The models look fragile and I have spent already a lot of time trying to untangle what they are trying to do. Please ask your teammates or let's see in the next QEMU cycle.

I have decided to go with the approach you suggested and it looks much better. Fixed it.

Thanks for the review.

Regards,

Ninad

Quoted 6 lines

Thanks,

C.

maintainer
#18 lore raw

Re: [PATCH v6 06/10] hw/fsi: Aspeed APB2OPB interface

Quoted 14 lines · Ninad Palsule

Hello Cedric,

Quoted 10 lines · Cédric Le Goater
Quoted 8 lines · Ninad Palsule

Hello Cedric,

Quoted 1 line · Cédric Le Goater

and the fsi_opb_* routines are useless to me.

We are trying to keep the separation between OPB implementation and interface hence we have all those fsi_opb_*. I feel that we should keep as it is so that future extensions will be easier. Please let me know.

Quoted 8 lines
Quoted 5 lines

Well, I can't really tell because I don't know enough about FSI :/

The models look fragile and I have spent already a lot of time trying to untangle what they are trying to do. Please ask your teammates or let's see in the next QEMU cycle.

I have decided to go with the approach you suggested and it looks much better. Fixed it.

I intended to reply to this before Ninad sent out v7, but life intervened.

If we can't justify it with the code we have now I think it's right to pull it out. Add the code to support the things we're trying to do when we need to do them. As long as we don't do anything that precludes us from adding that code later (and I can't really imagine how we'd corner ourselves like that).

We should bear in mind I wrote the initial models several years ago in the space of about a week while I was trying to learn FSI (and more deeply about the QEMU bus and address space modelling). I think I was doing that to unblock some CI due to the introduction of the kernel driver for the Aspeed FSI hardware. The models were pretty rough - prior to all this review the code reflected my hazy understanding of the problems. I didn't get time to remove the complexities introduced by my misunderstandings, and now it's been so long that I'm not much help with fixing them.

Andrew

#19 lore raw

PATCH v6 7/10 hw/arm: Hook up FSI module in AST2600

In reply to Ninad Palsule

This patchset introduces IBM's Flexible Service Interface(FSI).

Time for some fun with inter-processor buses. FSI allows a service processor access to the internal buses of a host POWER processor to perform configuration or debugging.

FSI has long existed in POWER processes and so comes with some baggage, including how it has been integrated into the ASPEED SoC.

Working backwards from the POWER processor, the fundamental pieces of interest for the implementation are:

1. The Common FRU Access Macro (CFAM), an address space containing various "engines" that drive accesses on buses internal and external to the POWER chip. Examples include the SBEFIFO and I2C masters. The engines hang off of an internal Local Bus (LBUS) which is described by the CFAM configuration block.

2. The FSI slave: The slave is the terminal point of the FSI bus for FSI symbols addressed to it. Slaves can be cascaded off of one another. The slave's configuration registers appear in address space of the CFAM to which it is attached.

3. The FSI master: A controller in the platform service processor (e.g. BMC) driving CFAM engine accesses into the POWER chip. At the hardware level FSI is a bit-based protocol supporting synchronous and DMA-driven accesses of engines in a CFAM.

4. The On-Chip Peripheral Bus (OPB): A low-speed bus typically found in POWER processors. This now makes an appearance in the ASPEED SoC due to tight integration of the FSI master IP with the OPB, mainly the existence of an MMIO-mapping of the CFAM address straight onto a sub-region of the OPB address space.

5. An APB-to-OPB bridge enabling access to the OPB from the ARM core in the AST2600. Hardware limitations prevent the OPB from being directly mapped into APB, so all accesses are indirect through the bridge.

The implementation appears as following in the qemu device tree:

(qemu) info qtree
bus: main-system-bus
type System
...
dev: aspeed.apb2opb, id ""
gpio-out "sysbus-irq" 1
mmio 000000001e79b000/0000000000001000
bus: opb.1
type opb
dev: fsi.master, id ""
bus: fsi.bus.1
type fsi.bus
dev: cfam.config, id ""
dev: cfam, id ""
bus: fsi.lbus.1
type lbus
dev: scratchpad, id ""
address = 0 (0x0)
bus: opb.0
type opb
dev: fsi.master, id ""
bus: fsi.bus.0
type fsi.bus
dev: cfam.config, id ""
dev: cfam, id ""
bus: fsi.lbus.0
type lbus
dev: scratchpad, id ""
address = 0 (0x0)

The LBUS is modelled to maintain the qdev bus hierarchy and to take advantage of the object model to automatically generate the CFAM configuration block. The configuration block presents engines in the order they are attached to the CFAM's LBUS. Engine implementations should subclass the LBusDevice and set the 'config' member of LBusDeviceClass to match the engine's type.

CFAM designs offer a lot of flexibility, for instance it is possible for a CFAM to be simultaneously driven from multiple FSI links. The modeling is not so complete; it's assumed that each CFAM is attached to a single FSI slave (as a consequence the CFAM subclasses the FSI slave).

As for FSI, its symbols and wire-protocol are not modelled at all. This is not necessary to get FSI off the ground thanks to the mapping of the CFAM address space onto the OPB address space - the models follow this directly and map the CFAM memory region into the OPB's memory region. Future work includes supporting more advanced accesses that drive the FSI master directly rather than indirectly via the CFAM mapping, which will require implementing the FSI state machine and methods for each of the FSI symbols on the slave. Further down the track we can also look at supporting the bitbanged SoftFSI drivers in Linux by extending the FSI slave model to resolve sequences of GPIO IRQs into FSI symbols, and calling the associated symbol method on the slave to map the access onto the CFAM.

Testing:

    Tested by reading cfam config address 0 on rainier machine type.

    root@p10bmc:~# pdbg -a getcfam 0x0
    p0: 0x0 = 0xc0022d15
---
 include/hw/arm/aspeed_soc.h |  4 ++++
 hw/arm/aspeed_ast2600.c     | 19 +++++++++++++++++++
 2 files changed, 23 insertions(+)
include/hw/arm/aspeed_soc.h +4 -0
index 8adff70072..db3ba3abc7 100644
@@ -36,6 +36,7 @@
 #include "hw/misc/aspeed_lpc.h"
 #include "hw/misc/unimp.h"
 #include "hw/misc/aspeed_peci.h"
+#include "hw/fsi/aspeed-apb2opb.h"
 #include "hw/char/serial.h"
 
 #define ASPEED_SPIS_NUM  2
@@ -96,6 +97,7 @@ struct AspeedSoCState {
     UnimplementedDeviceState udc;
     UnimplementedDeviceState sgpiom;
     UnimplementedDeviceState jtag[ASPEED_JTAG_NUM];
+    AspeedAPB2OPBState fsi[2];
 };
 
 #define TYPE_ASPEED_SOC "aspeed-soc"
@@ -191,6 +193,8 @@ enum {
     ASPEED_DEV_SGPIOM,
     ASPEED_DEV_JTAG0,
     ASPEED_DEV_JTAG1,
+    ASPEED_DEV_FSI1,
+    ASPEED_DEV_FSI2,
 };
 
 #define ASPEED_SOC_SPI_BOOT_ADDR 0x0
hw/arm/aspeed_ast2600.c +19 -0
index e122e1c32d..2ab6260955 100644
@@ -75,6 +75,8 @@ static const hwaddr aspeed_soc_ast2600_memmap[] = {
     [ASPEED_DEV_UART12]    = 0x1E790600,
     [ASPEED_DEV_UART13]    = 0x1E790700,
     [ASPEED_DEV_VUART]     = 0x1E787000,
+    [ASPEED_DEV_FSI1]      = 0x1E79B000,
+    [ASPEED_DEV_FSI2]      = 0x1E79B100,
     [ASPEED_DEV_I3C]       = 0x1E7A0000,
     [ASPEED_DEV_SDRAM]     = 0x80000000,
 };
@@ -132,6 +134,8 @@ static const int aspeed_soc_ast2600_irqmap[] = {
     [ASPEED_DEV_ETH4]      = 33,
     [ASPEED_DEV_KCS]       = 138,   /* 138 -> 142 */
     [ASPEED_DEV_DP]        = 62,
+    [ASPEED_DEV_FSI1]      = 100,
+    [ASPEED_DEV_FSI2]      = 101,
     [ASPEED_DEV_I3C]       = 102,   /* 102 -> 107 */
 };
 
@@ -262,6 +266,10 @@ static void aspeed_soc_ast2600_init(Object *obj)
     object_initialize_child(obj, "emmc-boot-controller",
                             &s->emmc_boot_controller,
                             TYPE_UNIMPLEMENTED_DEVICE);
+
+    for (i = 0; i < ASPEED_FSI_NUM; i++) {
+        object_initialize_child(obj, "fsi[*]", &s->fsi[i], TYPE_ASPEED_APB2OPB);
+    }
 }
 
 /*
@@ -622,6 +630,17 @@ static void aspeed_soc_ast2600_realize(DeviceState *dev, Error **errp)
         return;
     }
     aspeed_mmio_map(s, SYS_BUS_DEVICE(&s->sbc), 0, sc->memmap[ASPEED_DEV_SBC]);
+
+    /* FSI */
+    for (i = 0; i < ASPEED_FSI_NUM; i++) {
+        if (!sysbus_realize(SYS_BUS_DEVICE(&s->fsi[i]), errp)) {
+            return;
+        }
+        aspeed_mmio_map(s, SYS_BUS_DEVICE(&s->fsi[i]), 0,
+                        sc->memmap[ASPEED_DEV_FSI1 + i]);
+        sysbus_connect_irq(SYS_BUS_DEVICE(&s->fsi[i]), 0,
+                           aspeed_soc_get_irq(s, ASPEED_DEV_FSI1 + i));
+    }
 }
 
 static void aspeed_soc_ast2600_class_init(ObjectClass *oc, void *data)
Signature · 1 line
2.39.2
Reviewed-by on v6 7/10
#20 lore raw

Re: [PATCH v6 07/10] hw/arm: Hook up FSI module in AST2600

Quoted 101 lines · Ninad Palsule

This patchset introduces IBM's Flexible Service Interface(FSI).

Time for some fun with inter-processor buses. FSI allows a service processor access to the internal buses of a host POWER processor to perform configuration or debugging.

FSI has long existed in POWER processes and so comes with some baggage, including how it has been integrated into the ASPEED SoC.

Working backwards from the POWER processor, the fundamental pieces of interest for the implementation are:

1. The Common FRU Access Macro (CFAM), an address space containing various "engines" that drive accesses on buses internal and external to the POWER chip. Examples include the SBEFIFO and I2C masters. The engines hang off of an internal Local Bus (LBUS) which is described by the CFAM configuration block.

2. The FSI slave: The slave is the terminal point of the FSI bus for FSI symbols addressed to it. Slaves can be cascaded off of one another. The slave's configuration registers appear in address space of the CFAM to which it is attached.

3. The FSI master: A controller in the platform service processor (e.g. BMC) driving CFAM engine accesses into the POWER chip. At the hardware level FSI is a bit-based protocol supporting synchronous and DMA-driven accesses of engines in a CFAM.

4. The On-Chip Peripheral Bus (OPB): A low-speed bus typically found in POWER processors. This now makes an appearance in the ASPEED SoC due to tight integration of the FSI master IP with the OPB, mainly the existence of an MMIO-mapping of the CFAM address straight onto a sub-region of the OPB address space.

5. An APB-to-OPB bridge enabling access to the OPB from the ARM core in the AST2600. Hardware limitations prevent the OPB from being directly mapped into APB, so all accesses are indirect through the bridge.

The implementation appears as following in the qemu device tree:

(qemu) info qtree
bus: main-system-bus
type System
...
dev: aspeed.apb2opb, id ""
gpio-out "sysbus-irq" 1
mmio 000000001e79b000/0000000000001000
bus: opb.1
type opb
dev: fsi.master, id ""
bus: fsi.bus.1
type fsi.bus
dev: cfam.config, id ""
dev: cfam, id ""
bus: fsi.lbus.1
type lbus
dev: scratchpad, id ""
address = 0 (0x0)
bus: opb.0
type opb
dev: fsi.master, id ""
bus: fsi.bus.0
type fsi.bus
dev: cfam.config, id ""
dev: cfam, id ""
bus: fsi.lbus.0
type lbus
dev: scratchpad, id ""
address = 0 (0x0)

The LBUS is modelled to maintain the qdev bus hierarchy and to take advantage of the object model to automatically generate the CFAM configuration block. The configuration block presents engines in the order they are attached to the CFAM's LBUS. Engine implementations should subclass the LBusDevice and set the 'config' member of LBusDeviceClass to match the engine's type.

CFAM designs offer a lot of flexibility, for instance it is possible for a CFAM to be simultaneously driven from multiple FSI links. The modeling is not so complete; it's assumed that each CFAM is attached to a single FSI slave (as a consequence the CFAM subclasses the FSI slave).

As for FSI, its symbols and wire-protocol are not modelled at all. This is not necessary to get FSI off the ground thanks to the mapping of the CFAM address space onto the OPB address space - the models follow this directly and map the CFAM memory region into the OPB's memory region. Future work includes supporting more advanced accesses that drive the FSI master directly rather than indirectly via the CFAM mapping, which will require implementing the FSI state machine and methods for each of the FSI symbols on the slave. Further down the track we can also look at supporting the bitbanged SoftFSI drivers in Linux by extending the FSI slave model to resolve sequences of GPIO IRQs into FSI symbols, and calling the associated symbol method on the slave to map the access onto the CFAM.

Testing:

     Tested by reading cfam config address 0 on rainier machine type.

     root@p10bmc:~# pdbg -a getcfam 0x0
     p0: 0x0 = 0xc0022d15
Quoted 6 lines
---
  include/hw/arm/aspeed_soc.h |  4 ++++
  hw/arm/aspeed_ast2600.c     | 19 +++++++++++++++++++
  2 files changed, 23 insertions(+)
#21 lore raw

PATCH v6 8/10 hw/fsi: Added qtest

In reply to Ninad Palsule

Added basic qtests for FSI model.

---
v3:
- Added new qtest as per Cedric's comment.
V4:
- Remove MAINTAINER and documentation changes from this commit
v6:
- Incorporated review comments by Thomas Huth.

---
 tests/qtest/fsi-test.c  | 207 ++++++++++++++++++++++++++++++++++++++++
 tests/qtest/meson.build |   1 +
 2 files changed, 208 insertions(+)
 create mode 100644 tests/qtest/fsi-test.c
tests/qtest/fsi-test.c new file +207 -0
new file mode 100644
index 0000000000..01a0739092
@@ -0,0 +1,207 @@
+/*
+ * QTest testcases for IBM's Flexible Service Interface (FSI)
+ *
+ * Copyright (c) 2023 IBM Corporation
+ *
+ * Authors:
+ *   Ninad Palsule <[email protected]>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ */
+
+#include "qemu/osdep.h"
+#include <glib/gstdio.h>
+
+#include "qemu/module.h"
+#include "libqtest-single.h"
+
+/* Registers from ast2600 specifications */
+#define ASPEED_FSI_ENGINER_TRIGGER   0x04
+#define ASPEED_FSI_OPB0_BUS_SELECT   0x10
+#define ASPEED_FSI_OPB1_BUS_SELECT   0x28
+#define ASPEED_FSI_OPB0_RW_DIRECTION 0x14
+#define ASPEED_FSI_OPB1_RW_DIRECTION 0x2c
+#define ASPEED_FSI_OPB0_XFER_SIZE    0x18
+#define ASPEED_FSI_OPB1_XFER_SIZE    0x30
+#define ASPEED_FSI_OPB0_BUS_ADDR     0x1c
+#define ASPEED_FSI_OPB1_BUS_ADDR     0x34
+#define ASPEED_FSI_INTRRUPT_CLEAR    0x40
+#define ASPEED_FSI_INTRRUPT_STATUS   0x48
+#define ASPEED_FSI_OPB0_BUS_STATUS   0x80
+#define ASPEED_FSI_OPB1_BUS_STATUS   0x8c
+#define ASPEED_FSI_OPB0_READ_DATA    0x84
+#define ASPEED_FSI_OPB1_READ_DATA    0x90
+
+/*
+ * FSI Base addresses from the ast2600 specifications.
+ */
+#define AST2600_OPB_FSI0_BASE_ADDR 0x1e79b000
+#define AST2600_OPB_FSI1_BASE_ADDR 0x1e79b100
+
+static uint32_t aspeed_fsi_base_addr;
+
+static uint32_t aspeed_fsi_readl(QTestState *s, uint32_t reg)
+{
+    return qtest_readl(s, aspeed_fsi_base_addr + reg);
+}
+
+static void aspeed_fsi_writel(QTestState *s, uint32_t reg, uint32_t val)
+{
+    qtest_writel(s, aspeed_fsi_base_addr + reg, val);
+}
+
+/* Setup base address and select register */
+static void test_fsi_setup(QTestState *s, uint32_t base_addr)
+{
+    uint32_t curval;
+
+    /* Set the base select register */
+    if (base_addr == AST2600_OPB_FSI0_BASE_ADDR) {
+        aspeed_fsi_base_addr = base_addr;
+
+        /* Unselect FSI1 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x0);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x0);
+
+        /* Select FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x1);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x1);
+    } else if (base_addr == AST2600_OPB_FSI1_BASE_ADDR) {
+        aspeed_fsi_base_addr = base_addr;
+
+        /* Unselect FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x0);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x0);
+
+        /* Select FSI1 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x1);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x1);
+    } else {
+        g_assert_not_reached();
+    }
+}
+
+static void test_fsi_reg_change(QTestState *s, uint32_t reg, uint32_t newval)
+{
+    uint32_t base;
+    uint32_t curval;
+
+    base = aspeed_fsi_readl(s, reg);
+    aspeed_fsi_writel(s, reg, newval);
+    curval = aspeed_fsi_readl(s, reg);
+    g_assert_cmpuint(curval, ==, newval);
+    aspeed_fsi_writel(s, reg, base);
+    curval = aspeed_fsi_readl(s, reg);
+    g_assert_cmpuint(curval, ==, base);
+}
+
+static void test_fsi0_master_regs(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+
+    test_fsi_setup(s, AST2600_OPB_FSI0_BASE_ADDR);
+
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_RW_DIRECTION, 0xF3F4F514);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_XFER_SIZE, 0xF3F4F518);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_BUS_ADDR, 0xF3F4F51c);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_CLEAR, 0xF3F4F540);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_STATUS, 0xF3F4F548);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_BUS_STATUS, 0xF3F4F580);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_READ_DATA, 0xF3F4F584);
+}
+
+static void test_fsi1_master_regs(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+
+    test_fsi_setup(s, AST2600_OPB_FSI1_BASE_ADDR);
+
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_RW_DIRECTION, 0xF3F4F514);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_XFER_SIZE, 0xF3F4F518);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_BUS_ADDR, 0xF3F4F51c);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_CLEAR, 0xF3F4F540);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_STATUS, 0xF3F4F548);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_BUS_STATUS, 0xF3F4F580);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_READ_DATA, 0xF3F4F584);
+}
+
+static void test_fsi0_getcfam_addr0(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+    uint32_t curval;
+
+    test_fsi_setup(s, AST2600_OPB_FSI0_BASE_ADDR);
+
+    /* Master access direction read */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB0_RW_DIRECTION, 0x1);
+    /* word */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB0_XFER_SIZE, 0x3);
+    /* Address */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_ADDR, 0xa0000000);
+    aspeed_fsi_writel(s, ASPEED_FSI_INTRRUPT_CLEAR, 0x1);
+    aspeed_fsi_writel(s, ASPEED_FSI_ENGINER_TRIGGER, 0x1);
+
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_INTRRUPT_STATUS);
+    g_assert_cmpuint(curval, ==, 0x10000);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_STATUS);
+    g_assert_cmpuint(curval, ==, 0x0);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_READ_DATA);
+    g_assert_cmpuint(curval, ==, 0x152d02c0);
+}
+
+static void test_fsi1_getcfam_addr0(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+    uint32_t curval;
+
+    test_fsi_setup(s, AST2600_OPB_FSI1_BASE_ADDR);
+
+    /* Master access direction read */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB1_RW_DIRECTION, 0x1);
+
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB1_XFER_SIZE, 0x3);
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_ADDR, 0xa0000000);
+    aspeed_fsi_writel(s, ASPEED_FSI_INTRRUPT_CLEAR, 0x1);
+    aspeed_fsi_writel(s, ASPEED_FSI_ENGINER_TRIGGER, 0x1);
+
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_INTRRUPT_STATUS);
+    g_assert_cmpuint(curval, ==, 0x20000);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_STATUS);
+    g_assert_cmpuint(curval, ==, 0x0);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_READ_DATA);
+    g_assert_cmpuint(curval, ==, 0x152d02c0);
+}
+
+int main(int argc, char **argv)
+{
+    int ret = -1;
+    QTestState *s;
+
+    g_test_init(&argc, &argv, NULL);
+
+    s = qtest_init("-machine ast2600-evb ");
+
+    /* Tests for OPB/FSI0 */
+    qtest_add_data_func("/fsi-test/test_fsi0_master_regs", s,
+                        test_fsi0_master_regs);
+
+    qtest_add_data_func("/fsi-test/test_fsi0_getcfam_addr0", s,
+                        test_fsi0_getcfam_addr0);
+
+    /* Tests for OPB/FSI1 */
+    qtest_add_data_func("/fsi-test/test_fsi1_master_regs", s,
+                        test_fsi1_master_regs);
+
+    qtest_add_data_func("/fsi-test/test_fsi1_getcfam_addr0", s,
+                        test_fsi1_getcfam_addr0);
+
+    ret = g_test_run();
+    qtest_quit(s);
+
+    return ret;
+}
tests/qtest/meson.build +1 -0
index d6022ebd64..74228a4aed 100644
@@ -207,6 +207,7 @@ qtests_arm = \
   (config_all_devices.has_key('CONFIG_TPM_TIS_I2C') ? ['tpm-tis-i2c-test'] : []) + \
   (config_all_devices.has_key('CONFIG_VEXPRESS') ? ['test-arm-mptimer'] : []) + \
   (config_all_devices.has_key('CONFIG_MICROBIT') ? ['microbit-test'] : []) + \
+  (config_all_devices.has_key('CONFIG_FSI_APB2OPB_ASPEED') ? ['fsi-test'] : []) + \
   ['arm-cpu-features',
    'boot-serial-test']
 
Signature · 1 line
2.39.2
Acked-by on v6 8/10
#22 lore raw

Re: [PATCH v6 08/10] hw/fsi: Added qtest

Quoted 85 lines · Ninad Palsule

Added basic qtests for FSI model.

---
v3:
- Added new qtest as per Cedric's comment.
V4:
- Remove MAINTAINER and documentation changes from this commit
v6:
- Incorporated review comments by Thomas Huth.

---
  tests/qtest/fsi-test.c  | 207 ++++++++++++++++++++++++++++++++++++++++
  tests/qtest/meson.build |   1 +
  2 files changed, 208 insertions(+)

create mode 100644 tests/qtest/fsi-test.c

tests/qtest/fsi-test.c new file +63 -0
new file mode 100644
index 0000000000..01a0739092
@@ -0,0 +1,207 @@
+/*
+ * QTest testcases for IBM's Flexible Service Interface (FSI)
+ *
+ * Copyright (c) 2023 IBM Corporation
+ *
+ * Authors:
+ *   Ninad Palsule <[email protected]>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ */
+
+#include "qemu/osdep.h"
+#include <glib/gstdio.h>
+
+#include "qemu/module.h"
+#include "libqtest-single.h"
+
+/* Registers from ast2600 specifications */
+#define ASPEED_FSI_ENGINER_TRIGGER   0x04
+#define ASPEED_FSI_OPB0_BUS_SELECT   0x10
+#define ASPEED_FSI_OPB1_BUS_SELECT   0x28
+#define ASPEED_FSI_OPB0_RW_DIRECTION 0x14
+#define ASPEED_FSI_OPB1_RW_DIRECTION 0x2c
+#define ASPEED_FSI_OPB0_XFER_SIZE    0x18
+#define ASPEED_FSI_OPB1_XFER_SIZE    0x30
+#define ASPEED_FSI_OPB0_BUS_ADDR     0x1c
+#define ASPEED_FSI_OPB1_BUS_ADDR     0x34
+#define ASPEED_FSI_INTRRUPT_CLEAR    0x40
+#define ASPEED_FSI_INTRRUPT_STATUS   0x48
+#define ASPEED_FSI_OPB0_BUS_STATUS   0x80
+#define ASPEED_FSI_OPB1_BUS_STATUS   0x8c
+#define ASPEED_FSI_OPB0_READ_DATA    0x84
+#define ASPEED_FSI_OPB1_READ_DATA    0x90
+
+/*
+ * FSI Base addresses from the ast2600 specifications.
+ */
+#define AST2600_OPB_FSI0_BASE_ADDR 0x1e79b000
+#define AST2600_OPB_FSI1_BASE_ADDR 0x1e79b100
+
+static uint32_t aspeed_fsi_base_addr;
+
+static uint32_t aspeed_fsi_readl(QTestState *s, uint32_t reg)
+{
+    return qtest_readl(s, aspeed_fsi_base_addr + reg);
+}
+
+static void aspeed_fsi_writel(QTestState *s, uint32_t reg, uint32_t val)
+{
+    qtest_writel(s, aspeed_fsi_base_addr + reg, val);
+}
+
+/* Setup base address and select register */
+static void test_fsi_setup(QTestState *s, uint32_t base_addr)
+{
+    uint32_t curval;
+
+    /* Set the base select register */
+    if (base_addr == AST2600_OPB_FSI0_BASE_ADDR) {
+        aspeed_fsi_base_addr = base_addr;
+
+        /* Unselect FSI1 */

tests/qtest/fsi-test.c line 73

Quoted 10 lines
tests/qtest/fsi-test.c new file +10 -0
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x0);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x0);
+
+        /* Select FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x1);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x1);
+    } else if (base_addr == AST2600_OPB_FSI1_BASE_ADDR) {
+        aspeed_fsi_base_addr = base_addr;

You could move "aspeed_fsi_base_addr = base_addr" before the if-statement, so that you don't have to repeat it in both cases.

tests/qtest/fsi-test.c line 87

Quoted 13 lines
tests/qtest/fsi-test.c new file +13 -0
+        /* Unselect FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x0);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x0);
+
+        /* Select FSI1 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x1);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x1);
+    } else {
+        g_assert_not_reached();
+    }
+}

Anyway:

#23 lore raw

Re: [PATCH v6 08/10] hw/fsi: Added qtest

Hello Thomas,

Quoted 92 lines · Thomas Huth
Quoted 91 lines · Ninad Palsule

Added basic qtests for FSI model.

---
v3:
  - Added new qtest as per Cedric's comment.
V4:
  - Remove MAINTAINER and documentation changes from this commit v6:
  - Incorporated review comments by Thomas Huth.
---
  tests/qtest/fsi-test.c  | 207 ++++++++++++++++++++++++++++++++++++++++ tests/qtest/meson.build |   1 +

  2 files changed, 208 insertions(+)

  create mode 100644 tests/qtest/fsi-test.c

tests/qtest/fsi-test.c new file +9 -0
new file mode 100644
index 0000000000..01a0739092
@@ -0,0 +1,207 @@
+/*
+ * QTest testcases for IBM's Flexible Service Interface (FSI)
+ *
+ * Copyright (c) 2023 IBM Corporation
+ *
+ * Authors:
+ *   Ninad Palsule <[email protected]>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 

or later.
+ * See the COPYING file in the top-level directory.
+ */

tests/qtest/fsi-test.c new file +38 -0
+
+#include "qemu/osdep.h"
+#include <glib/gstdio.h>
+
+#include "qemu/module.h"
+#include "libqtest-single.h"
+
+/* Registers from ast2600 specifications */
+#define ASPEED_FSI_ENGINER_TRIGGER   0x04
+#define ASPEED_FSI_OPB0_BUS_SELECT   0x10
+#define ASPEED_FSI_OPB1_BUS_SELECT   0x28
+#define ASPEED_FSI_OPB0_RW_DIRECTION 0x14
+#define ASPEED_FSI_OPB1_RW_DIRECTION 0x2c
+#define ASPEED_FSI_OPB0_XFER_SIZE    0x18
+#define ASPEED_FSI_OPB1_XFER_SIZE    0x30
+#define ASPEED_FSI_OPB0_BUS_ADDR     0x1c
+#define ASPEED_FSI_OPB1_BUS_ADDR     0x34
+#define ASPEED_FSI_INTRRUPT_CLEAR    0x40
+#define ASPEED_FSI_INTRRUPT_STATUS   0x48
+#define ASPEED_FSI_OPB0_BUS_STATUS   0x80
+#define ASPEED_FSI_OPB1_BUS_STATUS   0x8c
+#define ASPEED_FSI_OPB0_READ_DATA    0x84
+#define ASPEED_FSI_OPB1_READ_DATA    0x90
+
+/*
+ * FSI Base addresses from the ast2600 specifications.
+ */
+#define AST2600_OPB_FSI0_BASE_ADDR 0x1e79b000
+#define AST2600_OPB_FSI1_BASE_ADDR 0x1e79b100
+
+static uint32_t aspeed_fsi_base_addr;
+
+static uint32_t aspeed_fsi_readl(QTestState *s, uint32_t reg)
+{
+    return qtest_readl(s, aspeed_fsi_base_addr + reg);
+}
+
+static void aspeed_fsi_writel(QTestState *s, uint32_t reg, uint32_t 

val)

tests/qtest/fsi-test.c new file +18 -0
+{
+    qtest_writel(s, aspeed_fsi_base_addr + reg, val);
+}
+
+/* Setup base address and select register */
+static void test_fsi_setup(QTestState *s, uint32_t base_addr)
+{
+    uint32_t curval;
+
+    /* Set the base select register */
+    if (base_addr == AST2600_OPB_FSI0_BASE_ADDR) {
+        aspeed_fsi_base_addr = base_addr;
+
+        /* Unselect FSI1 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x0);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x0);
+
Quoted 8 lines
Quoted 6 lines

+        /* Select FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x1); +        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT); +        g_assert_cmpuint(curval, ==, 0x1);
+    } else if (base_addr == AST2600_OPB_FSI1_BASE_ADDR) { +        aspeed_fsi_base_addr = base_addr;

You could move "aspeed_fsi_base_addr = base_addr" before the if-statement, so that you don't have to repeat it in both cases.

Fixed it. I will update it in version 7.

Thanks for the review.

Regards,

Ninad

Quoted 18 lines
Quoted 5 lines

+        /* Unselect FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x0); +        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT); +        g_assert_cmpuint(curval, ==, 0x0);

tests/qtest/fsi-test.c line 87

Quoted 9 lines
tests/qtest/fsi-test.c new file +9 -0
+
+        /* Select FSI1 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x1);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x1);
+    } else {
+        g_assert_not_reached();
+    }
+}

Anyway:

#24 lore raw

Re: [PATCH v6 08/10] hw/fsi: Added qtest

In reply to Ninad Palsule
Quoted 10 lines · Ninad Palsule

Added basic qtests for FSI model.

---
v3:
- Added new qtest as per Cedric's comment.
V4:
- Remove MAINTAINER and documentation changes from this commit
v6:
- Incorporated review comments by Thomas Huth.

Quoted 2 lines
---
  tests/qtest/fsi-test.c  | 207 ++++++++++++++++++++++++++++++++++++++++

please rename the file to aspeed-fsi-test.c

Thanks,

C.

Quoted 229 lines
  tests/qtest/meson.build |   1 +
  2 files changed, 208 insertions(+)

create mode 100644 tests/qtest/fsi-test.c

tests/qtest/fsi-test.c new file +207 -0
new file mode 100644
index 0000000000..01a0739092
@@ -0,0 +1,207 @@
+/*
+ * QTest testcases for IBM's Flexible Service Interface (FSI)
+ *
+ * Copyright (c) 2023 IBM Corporation
+ *
+ * Authors:
+ *   Ninad Palsule <[email protected]>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ */
+
+#include "qemu/osdep.h"
+#include <glib/gstdio.h>
+
+#include "qemu/module.h"
+#include "libqtest-single.h"
+
+/* Registers from ast2600 specifications */
+#define ASPEED_FSI_ENGINER_TRIGGER   0x04
+#define ASPEED_FSI_OPB0_BUS_SELECT   0x10
+#define ASPEED_FSI_OPB1_BUS_SELECT   0x28
+#define ASPEED_FSI_OPB0_RW_DIRECTION 0x14
+#define ASPEED_FSI_OPB1_RW_DIRECTION 0x2c
+#define ASPEED_FSI_OPB0_XFER_SIZE    0x18
+#define ASPEED_FSI_OPB1_XFER_SIZE    0x30
+#define ASPEED_FSI_OPB0_BUS_ADDR     0x1c
+#define ASPEED_FSI_OPB1_BUS_ADDR     0x34
+#define ASPEED_FSI_INTRRUPT_CLEAR    0x40
+#define ASPEED_FSI_INTRRUPT_STATUS   0x48
+#define ASPEED_FSI_OPB0_BUS_STATUS   0x80
+#define ASPEED_FSI_OPB1_BUS_STATUS   0x8c
+#define ASPEED_FSI_OPB0_READ_DATA    0x84
+#define ASPEED_FSI_OPB1_READ_DATA    0x90
+
+/*
+ * FSI Base addresses from the ast2600 specifications.
+ */
+#define AST2600_OPB_FSI0_BASE_ADDR 0x1e79b000
+#define AST2600_OPB_FSI1_BASE_ADDR 0x1e79b100
+
+static uint32_t aspeed_fsi_base_addr;
+
+static uint32_t aspeed_fsi_readl(QTestState *s, uint32_t reg)
+{
+    return qtest_readl(s, aspeed_fsi_base_addr + reg);
+}
+
+static void aspeed_fsi_writel(QTestState *s, uint32_t reg, uint32_t val)
+{
+    qtest_writel(s, aspeed_fsi_base_addr + reg, val);
+}
+
+/* Setup base address and select register */
+static void test_fsi_setup(QTestState *s, uint32_t base_addr)
+{
+    uint32_t curval;
+
+    /* Set the base select register */
+    if (base_addr == AST2600_OPB_FSI0_BASE_ADDR) {
+        aspeed_fsi_base_addr = base_addr;
+
+        /* Unselect FSI1 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x0);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x0);
+
+        /* Select FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x1);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x1);
+    } else if (base_addr == AST2600_OPB_FSI1_BASE_ADDR) {
+        aspeed_fsi_base_addr = base_addr;
+
+        /* Unselect FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x0);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x0);
+
+        /* Select FSI1 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x1);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x1);
+    } else {
+        g_assert_not_reached();
+    }
+}
+
+static void test_fsi_reg_change(QTestState *s, uint32_t reg, uint32_t newval)
+{
+    uint32_t base;
+    uint32_t curval;
+
+    base = aspeed_fsi_readl(s, reg);
+    aspeed_fsi_writel(s, reg, newval);
+    curval = aspeed_fsi_readl(s, reg);
+    g_assert_cmpuint(curval, ==, newval);
+    aspeed_fsi_writel(s, reg, base);
+    curval = aspeed_fsi_readl(s, reg);
+    g_assert_cmpuint(curval, ==, base);
+}
+
+static void test_fsi0_master_regs(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+
+    test_fsi_setup(s, AST2600_OPB_FSI0_BASE_ADDR);
+
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_RW_DIRECTION, 0xF3F4F514);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_XFER_SIZE, 0xF3F4F518);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_BUS_ADDR, 0xF3F4F51c);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_CLEAR, 0xF3F4F540);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_STATUS, 0xF3F4F548);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_BUS_STATUS, 0xF3F4F580);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_READ_DATA, 0xF3F4F584);
+}
+
+static void test_fsi1_master_regs(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+
+    test_fsi_setup(s, AST2600_OPB_FSI1_BASE_ADDR);
+
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_RW_DIRECTION, 0xF3F4F514);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_XFER_SIZE, 0xF3F4F518);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_BUS_ADDR, 0xF3F4F51c);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_CLEAR, 0xF3F4F540);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_STATUS, 0xF3F4F548);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_BUS_STATUS, 0xF3F4F580);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_READ_DATA, 0xF3F4F584);
+}
+
+static void test_fsi0_getcfam_addr0(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+    uint32_t curval;
+
+    test_fsi_setup(s, AST2600_OPB_FSI0_BASE_ADDR);
+
+    /* Master access direction read */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB0_RW_DIRECTION, 0x1);
+    /* word */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB0_XFER_SIZE, 0x3);
+    /* Address */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_ADDR, 0xa0000000);
+    aspeed_fsi_writel(s, ASPEED_FSI_INTRRUPT_CLEAR, 0x1);
+    aspeed_fsi_writel(s, ASPEED_FSI_ENGINER_TRIGGER, 0x1);
+
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_INTRRUPT_STATUS);
+    g_assert_cmpuint(curval, ==, 0x10000);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_STATUS);
+    g_assert_cmpuint(curval, ==, 0x0);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_READ_DATA);
+    g_assert_cmpuint(curval, ==, 0x152d02c0);
+}
+
+static void test_fsi1_getcfam_addr0(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+    uint32_t curval;
+
+    test_fsi_setup(s, AST2600_OPB_FSI1_BASE_ADDR);
+
+    /* Master access direction read */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB1_RW_DIRECTION, 0x1);
+
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB1_XFER_SIZE, 0x3);
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_ADDR, 0xa0000000);
+    aspeed_fsi_writel(s, ASPEED_FSI_INTRRUPT_CLEAR, 0x1);
+    aspeed_fsi_writel(s, ASPEED_FSI_ENGINER_TRIGGER, 0x1);
+
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_INTRRUPT_STATUS);
+    g_assert_cmpuint(curval, ==, 0x20000);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_STATUS);
+    g_assert_cmpuint(curval, ==, 0x0);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_READ_DATA);
+    g_assert_cmpuint(curval, ==, 0x152d02c0);
+}
+
+int main(int argc, char **argv)
+{
+    int ret = -1;
+    QTestState *s;
+
+    g_test_init(&argc, &argv, NULL);
+
+    s = qtest_init("-machine ast2600-evb ");
+
+    /* Tests for OPB/FSI0 */
+    qtest_add_data_func("/fsi-test/test_fsi0_master_regs", s,
+                        test_fsi0_master_regs);
+
+    qtest_add_data_func("/fsi-test/test_fsi0_getcfam_addr0", s,
+                        test_fsi0_getcfam_addr0);
+
+    /* Tests for OPB/FSI1 */
+    qtest_add_data_func("/fsi-test/test_fsi1_master_regs", s,
+                        test_fsi1_master_regs);
+
+    qtest_add_data_func("/fsi-test/test_fsi1_getcfam_addr0", s,
+                        test_fsi1_getcfam_addr0);
+
+    ret = g_test_run();
+    qtest_quit(s);
+
+    return ret;
+}
tests/qtest/meson.build +1 -0
index d6022ebd64..74228a4aed 100644
@@ -207,6 +207,7 @@ qtests_arm = \
    (config_all_devices.has_key('CONFIG_TPM_TIS_I2C') ? ['tpm-tis-i2c-test'] : []) + \
    (config_all_devices.has_key('CONFIG_VEXPRESS') ? ['test-arm-mptimer'] : []) + \
    (config_all_devices.has_key('CONFIG_MICROBIT') ? ['microbit-test'] : []) + \
+  (config_all_devices.has_key('CONFIG_FSI_APB2OPB_ASPEED') ? ['fsi-test'] : []) + \
    ['arm-cpu-features',
     'boot-serial-test']
 
#25 lore raw

Re: [PATCH v6 08/10] hw/fsi: Added qtest

Hello Cedric,

Quoted 14 lines · Cédric Le Goater
Quoted 12 lines · Ninad Palsule

Added basic qtests for FSI model.

---
v3:
  - Added new qtest as per Cedric's comment.
V4:
  - Remove MAINTAINER and documentation changes from this commit v6:
  - Incorporated review comments by Thomas Huth.
---
  tests/qtest/fsi-test.c  | 207 ++++++++++++++++++++++++++++++++++++++++

Quoted 1 line

please rename the file to aspeed-fsi-test.c

Renamed it.

Thanks for the review.

Regards,

Ninad

Quoted 239 lines

Thanks,

C.

Quoted 232 lines

  tests/qtest/meson.build |   1 +

  2 files changed, 208 insertions(+)

  create mode 100644 tests/qtest/fsi-test.c

tests/qtest/fsi-test.c new file +9 -0
new file mode 100644
index 0000000000..01a0739092
@@ -0,0 +1,207 @@
+/*
+ * QTest testcases for IBM's Flexible Service Interface (FSI)
+ *
+ * Copyright (c) 2023 IBM Corporation
+ *
+ * Authors:
+ *   Ninad Palsule <[email protected]>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 

or later.
+ * See the COPYING file in the top-level directory.
+ */

tests/qtest/fsi-test.c new file +38 -0
+
+#include "qemu/osdep.h"
+#include <glib/gstdio.h>
+
+#include "qemu/module.h"
+#include "libqtest-single.h"
+
+/* Registers from ast2600 specifications */
+#define ASPEED_FSI_ENGINER_TRIGGER   0x04
+#define ASPEED_FSI_OPB0_BUS_SELECT   0x10
+#define ASPEED_FSI_OPB1_BUS_SELECT   0x28
+#define ASPEED_FSI_OPB0_RW_DIRECTION 0x14
+#define ASPEED_FSI_OPB1_RW_DIRECTION 0x2c
+#define ASPEED_FSI_OPB0_XFER_SIZE    0x18
+#define ASPEED_FSI_OPB1_XFER_SIZE    0x30
+#define ASPEED_FSI_OPB0_BUS_ADDR     0x1c
+#define ASPEED_FSI_OPB1_BUS_ADDR     0x34
+#define ASPEED_FSI_INTRRUPT_CLEAR    0x40
+#define ASPEED_FSI_INTRRUPT_STATUS   0x48
+#define ASPEED_FSI_OPB0_BUS_STATUS   0x80
+#define ASPEED_FSI_OPB1_BUS_STATUS   0x8c
+#define ASPEED_FSI_OPB0_READ_DATA    0x84
+#define ASPEED_FSI_OPB1_READ_DATA    0x90
+
+/*
+ * FSI Base addresses from the ast2600 specifications.
+ */
+#define AST2600_OPB_FSI0_BASE_ADDR 0x1e79b000
+#define AST2600_OPB_FSI1_BASE_ADDR 0x1e79b100
+
+static uint32_t aspeed_fsi_base_addr;
+
+static uint32_t aspeed_fsi_readl(QTestState *s, uint32_t reg)
+{
+    return qtest_readl(s, aspeed_fsi_base_addr + reg);
+}
+
+static void aspeed_fsi_writel(QTestState *s, uint32_t reg, uint32_t 

val)

tests/qtest/fsi-test.c new file +40 -0
+{
+    qtest_writel(s, aspeed_fsi_base_addr + reg, val);
+}
+
+/* Setup base address and select register */
+static void test_fsi_setup(QTestState *s, uint32_t base_addr)
+{
+    uint32_t curval;
+
+    /* Set the base select register */
+    if (base_addr == AST2600_OPB_FSI0_BASE_ADDR) {
+        aspeed_fsi_base_addr = base_addr;
+
+        /* Unselect FSI1 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x0);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x0);
+
+        /* Select FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x1);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x1);
+    } else if (base_addr == AST2600_OPB_FSI1_BASE_ADDR) {
+        aspeed_fsi_base_addr = base_addr;
+
+        /* Unselect FSI0 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_SELECT, 0x0);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x0);
+
+        /* Select FSI1 */
+        aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_SELECT, 0x1);
+        curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_SELECT);
+        g_assert_cmpuint(curval, ==, 0x1);
+    } else {
+        g_assert_not_reached();
+    }
+}
+
+static void test_fsi_reg_change(QTestState *s, uint32_t reg, 

uint32_t newval)

tests/qtest/fsi-test.c new file +118 -0
+{
+    uint32_t base;
+    uint32_t curval;
+
+    base = aspeed_fsi_readl(s, reg);
+    aspeed_fsi_writel(s, reg, newval);
+    curval = aspeed_fsi_readl(s, reg);
+    g_assert_cmpuint(curval, ==, newval);
+    aspeed_fsi_writel(s, reg, base);
+    curval = aspeed_fsi_readl(s, reg);
+    g_assert_cmpuint(curval, ==, base);
+}
+
+static void test_fsi0_master_regs(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+
+    test_fsi_setup(s, AST2600_OPB_FSI0_BASE_ADDR);
+
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_RW_DIRECTION, 0xF3F4F514);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_XFER_SIZE, 0xF3F4F518);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_BUS_ADDR, 0xF3F4F51c);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_CLEAR, 0xF3F4F540);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_STATUS, 0xF3F4F548);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_BUS_STATUS, 0xF3F4F580);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB0_READ_DATA, 0xF3F4F584);
+}
+
+static void test_fsi1_master_regs(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+
+    test_fsi_setup(s, AST2600_OPB_FSI1_BASE_ADDR);
+
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_RW_DIRECTION, 0xF3F4F514);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_XFER_SIZE, 0xF3F4F518);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_BUS_ADDR, 0xF3F4F51c);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_CLEAR, 0xF3F4F540);
+    test_fsi_reg_change(s, ASPEED_FSI_INTRRUPT_STATUS, 0xF3F4F548);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_BUS_STATUS, 0xF3F4F580);
+    test_fsi_reg_change(s, ASPEED_FSI_OPB1_READ_DATA, 0xF3F4F584);
+}
+
+static void test_fsi0_getcfam_addr0(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+    uint32_t curval;
+
+    test_fsi_setup(s, AST2600_OPB_FSI0_BASE_ADDR);
+
+    /* Master access direction read */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB0_RW_DIRECTION, 0x1);
+    /* word */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB0_XFER_SIZE, 0x3);
+    /* Address */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB0_BUS_ADDR, 0xa0000000);
+    aspeed_fsi_writel(s, ASPEED_FSI_INTRRUPT_CLEAR, 0x1);
+    aspeed_fsi_writel(s, ASPEED_FSI_ENGINER_TRIGGER, 0x1);
+
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_INTRRUPT_STATUS);
+    g_assert_cmpuint(curval, ==, 0x10000);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_BUS_STATUS);
+    g_assert_cmpuint(curval, ==, 0x0);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB0_READ_DATA);
+    g_assert_cmpuint(curval, ==, 0x152d02c0);
+}
+
+static void test_fsi1_getcfam_addr0(const void *data)
+{
+    QTestState *s = (QTestState *)data;
+    uint32_t curval;
+
+    test_fsi_setup(s, AST2600_OPB_FSI1_BASE_ADDR);
+
+    /* Master access direction read */
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB1_RW_DIRECTION, 0x1);
+
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB1_XFER_SIZE, 0x3);
+    aspeed_fsi_writel(s, ASPEED_FSI_OPB1_BUS_ADDR, 0xa0000000);
+    aspeed_fsi_writel(s, ASPEED_FSI_INTRRUPT_CLEAR, 0x1);
+    aspeed_fsi_writel(s, ASPEED_FSI_ENGINER_TRIGGER, 0x1);
+
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_INTRRUPT_STATUS);
+    g_assert_cmpuint(curval, ==, 0x20000);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_BUS_STATUS);
+    g_assert_cmpuint(curval, ==, 0x0);
+    curval = aspeed_fsi_readl(s, ASPEED_FSI_OPB1_READ_DATA);
+    g_assert_cmpuint(curval, ==, 0x152d02c0);
+}
+
+int main(int argc, char **argv)
+{
+    int ret = -1;
+    QTestState *s;
+
+    g_test_init(&argc, &argv, NULL);
+
+    s = qtest_init("-machine ast2600-evb ");
+
+    /* Tests for OPB/FSI0 */
+    qtest_add_data_func("/fsi-test/test_fsi0_master_regs", s,
+                        test_fsi0_master_regs);
+
+    qtest_add_data_func("/fsi-test/test_fsi0_getcfam_addr0", s,
+                        test_fsi0_getcfam_addr0);
+
+    /* Tests for OPB/FSI1 */
+    qtest_add_data_func("/fsi-test/test_fsi1_master_regs", s,
+                        test_fsi1_master_regs);
+
+    qtest_add_data_func("/fsi-test/test_fsi1_getcfam_addr0", s,
+                        test_fsi1_getcfam_addr0);
+
+    ret = g_test_run();
+    qtest_quit(s);
+
+    return ret;
+}
tests/qtest/meson.build +1 -0
index d6022ebd64..74228a4aed 100644
@@ -207,6 +207,7 @@ qtests_arm = \
     (config_all_devices.has_key('CONFIG_TPM_TIS_I2C') ? ['tpm-tis-i2c-test'] : []) + \
     (config_all_devices.has_key('CONFIG_VEXPRESS') ? ['test-arm-mptimer'] : []) + \
     (config_all_devices.has_key('CONFIG_MICROBIT') ? ['microbit-test'] : []) + \
+  (config_all_devices.has_key('CONFIG_FSI_APB2OPB_ASPEED') ? 
 ['fsi-test'] : []) + \
     ['arm-cpu-features',

     'boot-serial-test']

#26 lore raw

PATCH v6 9/10 hw/fsi: Added FSI documentation

In reply to Ninad Palsule

Documentation for IBM FSI model.

---
v4:
- Added separate commit for documentation

---
 docs/specs/fsi.rst | 141 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 141 insertions(+)
 create mode 100644 docs/specs/fsi.rst
docs/specs/fsi.rst new file +141 -0
new file mode 100644
index 0000000000..73b082afe1
@@ -0,0 +1,141 @@
+======================================
+IBM's Flexible Service Interface (FSI)
+======================================
+
+The QEMU FSI emulation implements hardware interfaces between ASPEED SOC, FSI
+master/slave and the end engine.
+
+FSI is a point-to-point two wire interface which is capable of supporting
+distances of up to 4 meters. FSI interfaces have been used successfully for
+many years in IBM servers to attach IBM Flexible Support Processors(FSP) to
+CPUs and IBM ASICs.
+
+FSI allows a service processor access to the internal buses of a host POWER
+processor to perform configuration or debugging. FSI has long existed in POWER
+processes and so comes with some baggage, including how it has been integrated
+into the ASPEED SoC.
+
+Working backwards from the POWER processor, the fundamental pieces of interest
+for the implementation are:
+
+1. The Common FRU Access Macro (CFAM), an address space containing various
+   "engines" that drive accesses on buses internal and external to the POWER
+   chip. Examples include the SBEFIFO and I2C masters. The engines hang off of
+   an internal Local Bus (LBUS) which is described by the CFAM configuration
+   block.
+
+2. The FSI slave: The slave is the terminal point of the FSI bus for FSI
+   symbols addressed to it. Slaves can be cascaded off of one another. The
+   slave's configuration registers appear in address space of the CFAM to
+   which it is attached.
+
+3. The FSI master: A controller in the platform service processor (e.g. BMC)
+   driving CFAM engine accesses into the POWER chip. At the hardware level
+   FSI is a bit-based protocol supporting synchronous and DMA-driven accesses
+   of engines in a CFAM.
+
+4. The On-Chip Peripheral Bus (OPB): A low-speed bus typically found in POWER
+   processors. This now makes an appearance in the ASPEED SoC due to tight
+   integration of the FSI master IP with the OPB, mainly the existence of an
+   MMIO-mapping of the CFAM address straight onto a sub-region of the OPB
+   address space.
+
+5. An APB-to-OPB bridge enabling access to the OPB from the ARM core in the
+   AST2600. Hardware limitations prevent the OPB from being directly mapped
+   into APB, so all accesses are indirect through the bridge.
+
+The LBUS is modelled to maintain the qdev bus hierarchy and to take advantages
+of the object model to automatically generate the CFAM configuration block.
+The configuration block presents engines in the order they are attached to the
+CFAM's LBUS. Engine implementations should subclass the LBusDevice and set the
+'config' member of LBusDeviceClass to match the engine's type.
+
+CFAM designs offer a lot of flexibility, for instance it is possible for a
+CFAM to be simultaneously driven from multiple FSI links. The modeling is not
+so complete; it's assumed that each CFAM is attached to a single FSI slave (as
+a consequence the CFAM subclasses the FSI slave).
+
+As for FSI, its symbols and wire-protocol are not modelled at all. This is not
+necessary to get FSI off the ground thanks to the mapping of the CFAM address
+space onto the OPB address space - the models follow this directly and map the
+CFAM memory region into the OPB's memory region.
+
+QEMU files related to FSI interface:
+ - ``hw/fsi/aspeed-apb2opb.c``
+ - ``include/hw/fsi/aspeed-apb2opb.h``
+ - ``hw/fsi/opb.c``
+ - ``include/hw/fsi/opb.h``
+ - ``hw/fsi/fsi.c``
+ - ``include/hw/fsi/fsi.h``
+ - ``hw/fsi/fsi-master.c``
+ - ``include/hw/fsi/fsi-master.h``
+ - ``hw/fsi/fsi-slave.c``
+ - ``include/hw/fsi/fsi-slave.h``
+ - ``hw/fsi/cfam.c``
+ - ``include/hw/fsi/cfam.h``
+ - ``hw/fsi/engine-scratchpad.c``
+ - ``include/hw/fsi/engine-scratchpad.h``
+ - ``include/hw/fsi/lbus.h``
+
+The following commands start the rainier machine with built-in FSI model.
+There are no model specific arguments.
+
+.. code-block:: console
+
+  qemu-system-arm -M rainier-bmc -nographic \
+  -kernel fitImage-linux.bin \
+  -dtb aspeed-bmc-ibm-rainier.dtb \
+  -initrd obmc-phosphor-initramfs.rootfs.cpio.xz \
+  -drive file=obmc-phosphor-image.rootfs.wic.qcow2,if=sd,index=2 \
+  -append "rootwait console=ttyS4,115200n8 root=PARTLABEL=rofs-a"
+
+The implementation appears as following in the qemu device tree:
+
+.. code-block:: console
+
+  (qemu) info qtree
+  bus: main-system-bus
+    type System
+    ...
+    dev: aspeed.apb2opb, id ""
+      gpio-out "sysbus-irq" 1
+      mmio 000000001e79b000/0000000000001000
+      bus: opb.1
+        type opb
+        dev: fsi.master, id ""
+          bus: fsi.bus.1
+            type fsi.bus
+            dev: cfam.config, id ""
+            dev: cfam, id ""
+              bus: lbus.1
+                type lbus
+                dev: scratchpad, id ""
+                  address = 0 (0x0)
+      bus: opb.0
+        type opb
+        dev: fsi.master, id ""
+          bus: fsi.bus.0
+            type fsi.bus
+            dev: cfam.config, id ""
+            dev: cfam, id ""
+              bus: lbus.0
+                type lbus
+                dev: scratchpad, id ""
+                  address = 0 (0x0)
+
+pdbg is a simple application to allow debugging of the host POWER processors
+from the BMC. (see the `pdbg source repository` for more details)
+
+.. code-block:: console
+
+  root@p10bmc:~# pdbg -a getcfam 0x0
+  p0: 0x0 = 0xc0022d15
+
+Refer following documents for more details.
+
+.. _FSI specification:
+   https://openpowerfoundation.org/specifications/fsi/
+   https://wiki.raptorcs.com/w/images/9/97/OpenFSI-spec-20161212.pdf
+
+.. _pdbg source repository:
+   https://github.com/open-power/pdbg
Signature · 1 line
2.39.2
#27 lore raw

Re: [PATCH v6 09/10] hw/fsi: Added FSI documentation

Quoted 10 lines · Ninad Palsule

Documentation for IBM FSI model.

---
v4:
- Added separate commit for documentation

---
  docs/specs/fsi.rst | 141 +++++++++++++++++++++++++++++++++++++++++++++
  1 file changed, 141 insertions(+)

create mode 100644 docs/specs/fsi.rst

Documentation build is broken.

a 'fsi" entry should be added in docs/specs/index.rst. More below.

Quoted 123 lines
docs/specs/fsi.rst new file +117 -0
new file mode 100644
index 0000000000..73b082afe1
@@ -0,0 +1,141 @@
+======================================
+IBM's Flexible Service Interface (FSI)
+======================================
+
+The QEMU FSI emulation implements hardware interfaces between ASPEED SOC, FSI
+master/slave and the end engine.
+
+FSI is a point-to-point two wire interface which is capable of supporting
+distances of up to 4 meters. FSI interfaces have been used successfully for
+many years in IBM servers to attach IBM Flexible Support Processors(FSP) to
+CPUs and IBM ASICs.
+
+FSI allows a service processor access to the internal buses of a host POWER
+processor to perform configuration or debugging. FSI has long existed in POWER
+processes and so comes with some baggage, including how it has been integrated
+into the ASPEED SoC.
+
+Working backwards from the POWER processor, the fundamental pieces of interest
+for the implementation are:
+
+1. The Common FRU Access Macro (CFAM), an address space containing various
+   "engines" that drive accesses on buses internal and external to the POWER
+   chip. Examples include the SBEFIFO and I2C masters. The engines hang off of
+   an internal Local Bus (LBUS) which is described by the CFAM configuration
+   block.
+
+2. The FSI slave: The slave is the terminal point of the FSI bus for FSI
+   symbols addressed to it. Slaves can be cascaded off of one another. The
+   slave's configuration registers appear in address space of the CFAM to
+   which it is attached.
+
+3. The FSI master: A controller in the platform service processor (e.g. BMC)
+   driving CFAM engine accesses into the POWER chip. At the hardware level
+   FSI is a bit-based protocol supporting synchronous and DMA-driven accesses
+   of engines in a CFAM.
+
+4. The On-Chip Peripheral Bus (OPB): A low-speed bus typically found in POWER
+   processors. This now makes an appearance in the ASPEED SoC due to tight
+   integration of the FSI master IP with the OPB, mainly the existence of an
+   MMIO-mapping of the CFAM address straight onto a sub-region of the OPB
+   address space.
+
+5. An APB-to-OPB bridge enabling access to the OPB from the ARM core in the
+   AST2600. Hardware limitations prevent the OPB from being directly mapped
+   into APB, so all accesses are indirect through the bridge.
+
+The LBUS is modelled to maintain the qdev bus hierarchy and to take advantages
+of the object model to automatically generate the CFAM configuration block.
+The configuration block presents engines in the order they are attached to the
+CFAM's LBUS. Engine implementations should subclass the LBusDevice and set the
+'config' member of LBusDeviceClass to match the engine's type.
+
+CFAM designs offer a lot of flexibility, for instance it is possible for a
+CFAM to be simultaneously driven from multiple FSI links. The modeling is not
+so complete; it's assumed that each CFAM is attached to a single FSI slave (as
+a consequence the CFAM subclasses the FSI slave).
+
+As for FSI, its symbols and wire-protocol are not modelled at all. This is not
+necessary to get FSI off the ground thanks to the mapping of the CFAM address
+space onto the OPB address space - the models follow this directly and map the
+CFAM memory region into the OPB's memory region.
+
+QEMU files related to FSI interface:
+ - ``hw/fsi/aspeed-apb2opb.c``
+ - ``include/hw/fsi/aspeed-apb2opb.h``
+ - ``hw/fsi/opb.c``
+ - ``include/hw/fsi/opb.h``
+ - ``hw/fsi/fsi.c``
+ - ``include/hw/fsi/fsi.h``
+ - ``hw/fsi/fsi-master.c``
+ - ``include/hw/fsi/fsi-master.h``
+ - ``hw/fsi/fsi-slave.c``
+ - ``include/hw/fsi/fsi-slave.h``
+ - ``hw/fsi/cfam.c``
+ - ``include/hw/fsi/cfam.h``
+ - ``hw/fsi/engine-scratchpad.c``
+ - ``include/hw/fsi/engine-scratchpad.h``
+ - ``include/hw/fsi/lbus.h``
+
+The following commands start the rainier machine with built-in FSI model.
+There are no model specific arguments.
+
+.. code-block:: console
+
+  qemu-system-arm -M rainier-bmc -nographic \
+  -kernel fitImage-linux.bin \
+  -dtb aspeed-bmc-ibm-rainier.dtb \
+  -initrd obmc-phosphor-initramfs.rootfs.cpio.xz \
+  -drive file=obmc-phosphor-image.rootfs.wic.qcow2,if=sd,index=2 \
+  -append "rootwait console=ttyS4,115200n8 root=PARTLABEL=rofs-a"
+
+The implementation appears as following in the qemu device tree:
+
+.. code-block:: console
+
+  (qemu) info qtree
+  bus: main-system-bus
+    type System
+    ...
+    dev: aspeed.apb2opb, id ""
+      gpio-out "sysbus-irq" 1
+      mmio 000000001e79b000/0000000000001000
+      bus: opb.1
+        type opb
+        dev: fsi.master, id ""
+          bus: fsi.bus.1
+            type fsi.bus
+            dev: cfam.config, id ""
+            dev: cfam, id ""
+              bus: lbus.1
+                type lbus
+                dev: scratchpad, id ""
+                  address = 0 (0x0)
+      bus: opb.0
+        type opb
+        dev: fsi.master, id ""
+          bus: fsi.bus.0

docs/specs/fsi.rst line 127

Quoted 10 lines
docs/specs/fsi.rst new file +10 -0
+            type fsi.bus
+            dev: cfam.config, id ""
+            dev: cfam, id ""
+              bus: lbus.0
+                type lbus
+                dev: scratchpad, id ""
+                  address = 0 (0x0)
+
+pdbg is a simple application to allow debugging of the host POWER processors
+from the BMC. (see the `pdbg source repository` for more details)

+ from the BMC. (see the ``pdbg source repository`` for more details)

Please check before sending.

Thanks,

C.

Quoted 14 lines
docs/specs/fsi.rst new file +14 -0
+
+.. code-block:: console
+
+  root@p10bmc:~# pdbg -a getcfam 0x0
+  p0: 0x0 = 0xc0022d15
+
+Refer following documents for more details.
+
+.. _FSI specification:
+   https://openpowerfoundation.org/specifications/fsi/
+   https://wiki.raptorcs.com/w/images/9/97/OpenFSI-spec-20161212.pdf
+
+.. _pdbg source repository:
+   https://github.com/open-power/pdbg
#28 lore raw

Re: [PATCH v6 09/10] hw/fsi: Added FSI documentation

Hello Cedric,

Quoted 13 lines · Cédric Le Goater
Quoted 10 lines · Ninad Palsule

Documentation for IBM FSI model.

---
v4:
   - Added separate commit for documentation
---
  docs/specs/fsi.rst | 141 +++++++++++++++++++++++++++++++++++++++++++++

  1 file changed, 141 insertions(+)

  create mode 100644 docs/specs/fsi.rst

Quoted 3 lines

Documentation build is broken.

a 'fsi" entry should be added in docs/specs/index.rst. More below.

Sorry about that. Added entry in the index.rst

Quoted 175 lines
Quoted 170 lines
docs/specs/fsi.rst new file +5 -0
new file mode 100644
index 0000000000..73b082afe1
@@ -0,0 +1,141 @@
+======================================
+IBM's Flexible Service Interface (FSI)
+======================================
+
+The QEMU FSI emulation implements hardware interfaces between ASPEED 

SOC, FSI

docs/specs/fsi.rst new file +3 -0
+master/slave and the end engine.
+
+FSI is a point-to-point two wire interface which is capable of 

supporting
+distances of up to 4 meters. FSI interfaces have been used successfully for
+many years in IBM servers to attach IBM Flexible Support Processors(FSP) to

docs/specs/fsi.rst new file +3 -0
+CPUs and IBM ASICs.
+
+FSI allows a service processor access to the internal buses of a 

host POWER
+processor to perform configuration or debugging. FSI has long existed in POWER
+processes and so comes with some baggage, including how it has been integrated

docs/specs/fsi.rst new file +3 -0
+into the ASPEED SoC.
+
+Working backwards from the POWER processor, the fundamental pieces 

of interest

docs/specs/fsi.rst new file +3 -0
+for the implementation are:
+
+1. The Common FRU Access Macro (CFAM), an address space containing 

various
+   "engines" that drive accesses on buses internal and external to the POWER
+   chip. Examples include the SBEFIFO and I2C masters. The engines hang off of
+   an internal Local Bus (LBUS) which is described by the CFAM configuration
+   block.
+
+2. The FSI slave: The slave is the terminal point of the FSI bus for FSI
+   symbols addressed to it. Slaves can be cascaded off of one another. The
+   slave's configuration registers appear in address space of the CFAM to
+   which it is attached.
+
+3. The FSI master: A controller in the platform service processor (e.g. BMC)
+   driving CFAM engine accesses into the POWER chip. At the hardware level
+   FSI is a bit-based protocol supporting synchronous and DMA-driven accesses
+   of engines in a CFAM.
+
+4. The On-Chip Peripheral Bus (OPB): A low-speed bus typically found in POWER
+   processors. This now makes an appearance in the ASPEED SoC due to tight
+   integration of the FSI master IP with the OPB, mainly the existence of an
+   MMIO-mapping of the CFAM address straight onto a sub-region of the OPB
+   address space.
+
+5. An APB-to-OPB bridge enabling access to the OPB from the ARM core in the
+   AST2600. Hardware limitations prevent the OPB from being directly mapped
+   into APB, so all accesses are indirect through the bridge.

docs/specs/fsi.rst new file +2 -0
+
+The LBUS is modelled to maintain the qdev bus hierarchy and to take 

advantages
+of the object model to automatically generate the CFAM configuration block.
+The configuration block presents engines in the order they are attached to the
+CFAM's LBUS. Engine implementations should subclass the LBusDevice and set the
+'config' member of LBusDeviceClass to match the engine's type.

docs/specs/fsi.rst new file +2 -0
+
+CFAM designs offer a lot of flexibility, for instance it is possible 

for a
+CFAM to be simultaneously driven from multiple FSI links. The modeling is not
+so complete; it's assumed that each CFAM is attached to a single FSI slave (as

docs/specs/fsi.rst new file +3 -0
+a consequence the CFAM subclasses the FSI slave).
+
+As for FSI, its symbols and wire-protocol are not modelled at all. 

This is not
+necessary to get FSI off the ground thanks to the mapping of the CFAM address
+space onto the OPB address space - the models follow this directly and map the

docs/specs/fsi.rst new file +20 -0
+CFAM memory region into the OPB's memory region.
+
+QEMU files related to FSI interface:
+ - ``hw/fsi/aspeed-apb2opb.c``
+ - ``include/hw/fsi/aspeed-apb2opb.h``
+ - ``hw/fsi/opb.c``
+ - ``include/hw/fsi/opb.h``
+ - ``hw/fsi/fsi.c``
+ - ``include/hw/fsi/fsi.h``
+ - ``hw/fsi/fsi-master.c``
+ - ``include/hw/fsi/fsi-master.h``
+ - ``hw/fsi/fsi-slave.c``
+ - ``include/hw/fsi/fsi-slave.h``
+ - ``hw/fsi/cfam.c``
+ - ``include/hw/fsi/cfam.h``
+ - ``hw/fsi/engine-scratchpad.c``
+ - ``include/hw/fsi/engine-scratchpad.h``
+ - ``include/hw/fsi/lbus.h``
+
+The following commands start the rainier machine with built-in FSI 

model.

docs/specs/fsi.rst new file +46 -0
+There are no model specific arguments.
+
+.. code-block:: console
+
+  qemu-system-arm -M rainier-bmc -nographic \
+  -kernel fitImage-linux.bin \
+  -dtb aspeed-bmc-ibm-rainier.dtb \
+  -initrd obmc-phosphor-initramfs.rootfs.cpio.xz \
+  -drive file=obmc-phosphor-image.rootfs.wic.qcow2,if=sd,index=2 \
+  -append "rootwait console=ttyS4,115200n8 root=PARTLABEL=rofs-a"
+
+The implementation appears as following in the qemu device tree:
+
+.. code-block:: console
+
+  (qemu) info qtree
+  bus: main-system-bus
+    type System
+    ...
+    dev: aspeed.apb2opb, id ""
+      gpio-out "sysbus-irq" 1
+      mmio 000000001e79b000/0000000000001000
+      bus: opb.1
+        type opb
+        dev: fsi.master, id ""
+          bus: fsi.bus.1
+            type fsi.bus
+            dev: cfam.config, id ""
+            dev: cfam, id ""
+              bus: lbus.1
+                type lbus
+                dev: scratchpad, id ""
+                  address = 0 (0x0)
+      bus: opb.0
+        type opb
+        dev: fsi.master, id ""
+          bus: fsi.bus.0
+            type fsi.bus
+            dev: cfam.config, id ""
+            dev: cfam, id ""
+              bus: lbus.0
+                type lbus
+                dev: scratchpad, id ""
+                  address = 0 (0x0)
+
+pdbg is a simple application to allow debugging of the host POWER 
Quoted 2 lines

processors
+from the BMC. (see the pdbg source repository for more details)

+ from the BMC. (see the  ``pdbg source repository`` for more details)

Quoted 1 line

Please check before sending.

Sorry about that. Fixed it. Checked it using "rstcheck" tool and also through web browser.

Thanks for the review.

Regards,

Ninad

Quoted 21 lines

Thanks,

C.

Quoted 14 lines
docs/specs/fsi.rst new file +14 -0
+
+.. code-block:: console
+
+  root@p10bmc:~# pdbg -a getcfam 0x0
+  p0: 0x0 = 0xc0022d15
+
+Refer following documents for more details.
+
+.. _FSI specification:
+   https://openpowerfoundation.org/specifications/fsi/
+ https://wiki.raptorcs.com/w/images/9/97/OpenFSI-spec-20161212.pdf
+
+.. _pdbg source repository:
+   https://github.com/open-power/pdbg
#29 lore raw

PATCH v6 10/10 hw/fsi: Update MAINTAINER list

In reply to Ninad Palsule

Added maintainer for IBM FSI model

---
V4:
- Added separate commit for MAINTAINER change.

V5:
- Use * instead of listing all files in dir

---
 MAINTAINERS | 8 ++++++++
 1 file changed, 8 insertions(+)
MAINTAINERS +8 -0
index 7f9912baa0..381923c937 100644
@@ -3476,6 +3476,14 @@ F: tests/qtest/adm1272-test.c
 F: tests/qtest/max34451-test.c
 F: tests/qtest/isl_pmbus_vr-test.c
 
+FSI
+M: Ninad Palsule <[email protected]>
+S: Maintained
+F: hw/fsi/*
+F: include/hw/fsi/*
+F: docs/specs/fsi.rst
+F: tests/qtest/fsi-test.c
+
 Firmware schema specifications
 M: Philippe Mathieu-Daudé <[email protected]>
 R: Daniel P. Berrange <[email protected]>
Signature · 1 line
2.39.2