Skip to contentSkip to conversation

RISC-V Add the OpenTitan Machine

PATCH Superseded Alistair Francis 10 participants 35 messages May 28, 2020 – May 19, 2023 RISC-V

v5 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

v5 cover RISC-V Add the OpenTitan Machine

OpenTitan is an open source silicon Root of Trust (RoT) project. This series adds initial support for the OpenTitan machine to QEMU.

This series add the Ibex CPU to the QEMU RISC-V target. It then adds the OpenTitan machine, the Ibex UART and the Ibex PLIC.

The UART has been tested sending and receiving data.

With this series QEMU can boot the OpenTitan ROM, Tock OS and a Tock userspace app.

The Ibex PLIC is similar to the RISC-V PLIC (and is based on the QEMU implementation) with some differences. The hope is that the Ibex PLIC will converge to follow the RISC-V spec. As that happens I want to update the QEMU Ibex PLIC and hopefully eventually replace the current PLIC as the implementation is a little overlay complex.

For more details on OpenTitan, see here: https://docs.opentitan.org/

v5:
- Add some of the missing unimplemented devices
- Don't set PMP feature in init() function
v4:
- Don't set the reset vector in realise
- Fix a bug where the MMU is always enabled
- Fixup the PMP/MMU size logic
v3:
- Small fixes pointed out in review
v2:
- Rebase on master
- Get uart receive working

Alistair Francis (11):
riscv/boot: Add a missing header include target/riscv: Don't overwrite the reset vector target/riscv: Disable the MMU correctly target/riscv: Don't set PMP feature in the cpu init target/riscv: Add the lowRISC Ibex CPU
riscv: Initial commit of OpenTitan machine hw/char: Initial commit of Ibex UART hw/intc: Initial commit of lowRISC Ibex PLIC riscv/opentitan: Connect the PLIC device riscv/opentitan: Connect the UART device target/riscv: Use a smaller guess size for no-MMU PMP

 default-configs/riscv32-softmmu.mak |   1 +
 default-configs/riscv64-softmmu.mak |  11 +-
 include/hw/char/ibex_uart.h         | 110 +++++++
 include/hw/intc/ibex_plic.h         |  63 ++++
 include/hw/riscv/boot.h             |   1 +
 include/hw/riscv/opentitan.h        |  84 +++++
 target/riscv/cpu.h                  |   1 +
 hw/char/ibex_uart.c                 | 492 ++++++++++++++++++++++++++++
 hw/intc/ibex_plic.c                 | 261 +++++++++++++++
 hw/riscv/opentitan.c                | 219 +++++++++++++
 target/riscv/cpu.c                  |  27 +-
 target/riscv/pmp.c                  |  14 +-
 MAINTAINERS                         |  13 +
 hw/char/Makefile.objs               |   1 +
 hw/intc/Makefile.objs               |   1 +
 hw/riscv/Kconfig                    |   9 +
 hw/riscv/Makefile.objs              |   1 +
 17 files changed, 1291 insertions(+), 18 deletions(-)
 create mode 100644 include/hw/char/ibex_uart.h
 create mode 100644 include/hw/intc/ibex_plic.h
 create mode 100644 include/hw/riscv/opentitan.h
 create mode 100644 hw/char/ibex_uart.c
 create mode 100644 hw/intc/ibex_plic.c
 create mode 100644 hw/riscv/opentitan.c
Signature · 1 line
2.26.2
#1 lore raw

PATCH v5 1/11 riscv/boot: Add a missing header include

As the functions declared in this header use the symbol_fn_t typedef itself declared in "hw/loader.h", we need to include it here to make the header file self-contained.

---
 include/hw/riscv/boot.h | 1 +
 1 file changed, 1 insertion(+)
include/hw/riscv/boot.h +1 -0
index 474a940ad5..9daa98da08 100644
@@ -21,6 +21,7 @@
 #define RISCV_BOOT_H
 
 #include "exec/cpu-defs.h"
+#include "hw/loader.h"
 
 void riscv_find_and_load_firmware(MachineState *machine,
                                   const char *default_machine_firmware,
Signature · 1 line
2.26.2
#2 lore raw

PATCH v5 2/11 target/riscv: Don't overwrite the reset vector

In reply to Alistair Francis

The reset vector is set in the init function don't set it again in realize.

---
 target/riscv/cpu.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
target/riscv/cpu.c +2 -1
index 059d71f2c7..5eb3c02735 100644
@@ -133,6 +133,7 @@ static void riscv_base32_cpu_init(Object *obj)
     CPURISCVState *env = &RISCV_CPU(obj)->env;
     /* We set this in the realise function */
     set_misa(env, 0);
+    set_resetvec(env, DEFAULT_RSTVEC);
 }
 
 static void rv32gcsu_priv1_09_1_cpu_init(Object *obj)
@@ -180,6 +181,7 @@ static void riscv_base64_cpu_init(Object *obj)
     CPURISCVState *env = &RISCV_CPU(obj)->env;
     /* We set this in the realise function */
     set_misa(env, 0);
+    set_resetvec(env, DEFAULT_RSTVEC);
 }
 
 static void rv64gcsu_priv1_09_1_cpu_init(Object *obj)
@@ -399,7 +401,6 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)
     }
 
     set_priv_version(env, priv_version);
-    set_resetvec(env, DEFAULT_RSTVEC);
 
     if (cpu->cfg.mmu) {
         set_feature(env, RISCV_FEATURE_MMU);
Signature · 1 line
2.26.2
#3 lore raw

PATCH v5 3/11 target/riscv: Disable the MMU correctly

In reply to Alistair Francis

Previously if we didn't enable the MMU it would be enabled in the realize() function anyway. Let's ensure that if we don't want the MMU we disable it. We also don't need to enable the MMU as it will be enalbed in realize() by default.

---
 target/riscv/cpu.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
target/riscv/cpu.c +3 -4
index 5eb3c02735..8deba3d16d 100644
@@ -142,7 +142,6 @@ static void rv32gcsu_priv1_09_1_cpu_init(Object *obj)
     set_misa(env, RV32 | RVI | RVM | RVA | RVF | RVD | RVC | RVS | RVU);
     set_priv_version(env, PRIV_VERSION_1_09_1);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_MMU);
     set_feature(env, RISCV_FEATURE_PMP);
 }
 
@@ -152,7 +151,6 @@ static void rv32gcsu_priv1_10_0_cpu_init(Object *obj)
     set_misa(env, RV32 | RVI | RVM | RVA | RVF | RVD | RVC | RVS | RVU);
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_MMU);
     set_feature(env, RISCV_FEATURE_PMP);
 }
 
@@ -163,6 +161,7 @@ static void rv32imacu_nommu_cpu_init(Object *obj)
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
     set_feature(env, RISCV_FEATURE_PMP);
+    qdev_prop_set_bit(DEVICE(obj), "mmu", false);
 }
 
 static void rv32imafcu_nommu_cpu_init(Object *obj)
@@ -172,6 +171,7 @@ static void rv32imafcu_nommu_cpu_init(Object *obj)
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
     set_feature(env, RISCV_FEATURE_PMP);
+    qdev_prop_set_bit(DEVICE(obj), "mmu", false);
 }
 
 #elif defined(TARGET_RISCV64)
@@ -190,7 +190,6 @@ static void rv64gcsu_priv1_09_1_cpu_init(Object *obj)
     set_misa(env, RV64 | RVI | RVM | RVA | RVF | RVD | RVC | RVS | RVU);
     set_priv_version(env, PRIV_VERSION_1_09_1);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_MMU);
     set_feature(env, RISCV_FEATURE_PMP);
 }
 
@@ -200,7 +199,6 @@ static void rv64gcsu_priv1_10_0_cpu_init(Object *obj)
     set_misa(env, RV64 | RVI | RVM | RVA | RVF | RVD | RVC | RVS | RVU);
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_MMU);
     set_feature(env, RISCV_FEATURE_PMP);
 }
 
@@ -211,6 +209,7 @@ static void rv64imacu_nommu_cpu_init(Object *obj)
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
     set_feature(env, RISCV_FEATURE_PMP);
+    qdev_prop_set_bit(DEVICE(obj), "mmu", false);
 }
 
 #endif
Signature · 1 line
2.26.2
Reviewed-by on v5 3/11
#4 lore raw

Re: [PATCH v5 03/11] target/riscv: Disable the MMU correctly

Quoted 4 lines · Alistair Francis

Previously if we didn't enable the MMU it would be enabled in the realize() function anyway. Let's ensure that if we don't want the MMU we disable it. We also don't need to enable the MMU as it will be enalbed

typo: enabled

Quoted 7 lines

in realize() by default.

---
 target/riscv/cpu.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
#5 lore raw

PATCH v5 4/11 target/riscv: Don't set PMP feature in the cpu init

In reply to Alistair Francis

The PMP is enabled by default via the "pmp" property so there is no need for us to set it in the init function. As all CPUs have PMP support just remove the set_feature() call in the CPU init functions.

---
 target/riscv/cpu.c | 7 -------
 1 file changed, 7 deletions(-)
target/riscv/cpu.c +0 -7
index 8deba3d16d..406e8f37d7 100644
@@ -142,7 +142,6 @@ static void rv32gcsu_priv1_09_1_cpu_init(Object *obj)
     set_misa(env, RV32 | RVI | RVM | RVA | RVF | RVD | RVC | RVS | RVU);
     set_priv_version(env, PRIV_VERSION_1_09_1);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_PMP);
 }
 
 static void rv32gcsu_priv1_10_0_cpu_init(Object *obj)
@@ -151,7 +150,6 @@ static void rv32gcsu_priv1_10_0_cpu_init(Object *obj)
     set_misa(env, RV32 | RVI | RVM | RVA | RVF | RVD | RVC | RVS | RVU);
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_PMP);
 }
 
 static void rv32imacu_nommu_cpu_init(Object *obj)
@@ -160,7 +158,6 @@ static void rv32imacu_nommu_cpu_init(Object *obj)
     set_misa(env, RV32 | RVI | RVM | RVA | RVC | RVU);
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_PMP);
     qdev_prop_set_bit(DEVICE(obj), "mmu", false);
 }
 
@@ -170,7 +167,6 @@ static void rv32imafcu_nommu_cpu_init(Object *obj)
     set_misa(env, RV32 | RVI | RVM | RVA | RVF | RVC | RVU);
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_PMP);
     qdev_prop_set_bit(DEVICE(obj), "mmu", false);
 }
 
@@ -190,7 +186,6 @@ static void rv64gcsu_priv1_09_1_cpu_init(Object *obj)
     set_misa(env, RV64 | RVI | RVM | RVA | RVF | RVD | RVC | RVS | RVU);
     set_priv_version(env, PRIV_VERSION_1_09_1);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_PMP);
 }
 
 static void rv64gcsu_priv1_10_0_cpu_init(Object *obj)
@@ -199,7 +194,6 @@ static void rv64gcsu_priv1_10_0_cpu_init(Object *obj)
     set_misa(env, RV64 | RVI | RVM | RVA | RVF | RVD | RVC | RVS | RVU);
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_PMP);
 }
 
 static void rv64imacu_nommu_cpu_init(Object *obj)
@@ -208,7 +202,6 @@ static void rv64imacu_nommu_cpu_init(Object *obj)
     set_misa(env, RV64 | RVI | RVM | RVA | RVC | RVU);
     set_priv_version(env, PRIV_VERSION_1_10_0);
     set_resetvec(env, DEFAULT_RSTVEC);
-    set_feature(env, RISCV_FEATURE_PMP);
     qdev_prop_set_bit(DEVICE(obj), "mmu", false);
 }
 
Signature · 1 line
2.26.2
Reviewed-by on v5 4/11
#6 lore raw

Re: [PATCH v5 04/11] target/riscv: Don't set PMP feature in the cpu init

Quoted 10 lines · Alistair Francis

The PMP is enabled by default via the "pmp" property so there is no need for us to set it in the init function. As all CPUs have PMP support just remove the set_feature() call in the CPU init functions.

---
 target/riscv/cpu.c | 7 -------
 1 file changed, 7 deletions(-)
#7 lore raw

PATCH v5 5/11 target/riscv: Add the lowRISC Ibex CPU

In reply to Alistair Francis

Ibex is a small and efficient, 32-bit, in-order RISC-V core with a 2-stage pipeline that implements the RV32IMC instruction set architecture.

For more details on lowRISC see here:
https://github.com/lowRISC/ibex

---
 target/riscv/cpu.h |  1 +
 target/riscv/cpu.c | 10 ++++++++++
 2 files changed, 11 insertions(+)
target/riscv/cpu.h +1 -0
index d0e7f5b9c5..8733d7467f 100644
@@ -35,6 +35,7 @@
 #define TYPE_RISCV_CPU_ANY              RISCV_CPU_TYPE_NAME("any")
 #define TYPE_RISCV_CPU_BASE32           RISCV_CPU_TYPE_NAME("rv32")
 #define TYPE_RISCV_CPU_BASE64           RISCV_CPU_TYPE_NAME("rv64")
+#define TYPE_RISCV_CPU_IBEX             RISCV_CPU_TYPE_NAME("lowrisc-ibex")
 #define TYPE_RISCV_CPU_SIFIVE_E31       RISCV_CPU_TYPE_NAME("sifive-e31")
 #define TYPE_RISCV_CPU_SIFIVE_E34       RISCV_CPU_TYPE_NAME("sifive-e34")
 #define TYPE_RISCV_CPU_SIFIVE_E51       RISCV_CPU_TYPE_NAME("sifive-e51")
target/riscv/cpu.c +10 -0
index 406e8f37d7..6e0d4d1dda 100644
@@ -152,6 +152,15 @@ static void rv32gcsu_priv1_10_0_cpu_init(Object *obj)
     set_resetvec(env, DEFAULT_RSTVEC);
 }
 
+static void rv32imcu_nommu_cpu_init(Object *obj)
+{
+    CPURISCVState *env = &RISCV_CPU(obj)->env;
+    set_misa(env, RV32 | RVI | RVM | RVC | RVU);
+    set_priv_version(env, PRIV_VERSION_1_10_0);
+    set_resetvec(env, 0x8090);
+    qdev_prop_set_bit(DEVICE(obj), "mmu", false);
+}
+
 static void rv32imacu_nommu_cpu_init(Object *obj)
 {
     CPURISCVState *env = &RISCV_CPU(obj)->env;
@@ -611,6 +620,7 @@ static const TypeInfo riscv_cpu_type_infos[] = {
     DEFINE_CPU(TYPE_RISCV_CPU_ANY,              riscv_any_cpu_init),
 #if defined(TARGET_RISCV32)
     DEFINE_CPU(TYPE_RISCV_CPU_BASE32,           riscv_base32_cpu_init),
+    DEFINE_CPU(TYPE_RISCV_CPU_IBEX,             rv32imcu_nommu_cpu_init),
     DEFINE_CPU(TYPE_RISCV_CPU_SIFIVE_E31,       rv32imacu_nommu_cpu_init),
     DEFINE_CPU(TYPE_RISCV_CPU_SIFIVE_E34,       rv32imafcu_nommu_cpu_init),
     DEFINE_CPU(TYPE_RISCV_CPU_SIFIVE_U34,       rv32gcsu_priv1_10_0_cpu_init),
Signature · 1 line
2.26.2
#8 lore raw

PATCH v5 6/11 riscv: Initial commit of OpenTitan machine

In reply to Alistair Francis

This adds a barebone OpenTitan machine to QEMU.

---
 default-configs/riscv32-softmmu.mak |   1 +
 default-configs/riscv64-softmmu.mak |  11 +-
 include/hw/riscv/opentitan.h        |  68 ++++++++++
 hw/riscv/opentitan.c                | 184 ++++++++++++++++++++++++++++
 MAINTAINERS                         |   9 ++
 hw/riscv/Kconfig                    |   5 +
 hw/riscv/Makefile.objs              |   1 +
 7 files changed, 278 insertions(+), 1 deletion(-)
 create mode 100644 include/hw/riscv/opentitan.h
 create mode 100644 hw/riscv/opentitan.c
default-configs/riscv32-softmmu.mak +1 -0
index 1ae077ed87..94a236c9c2 100644
@@ -10,3 +10,4 @@ CONFIG_SPIKE=y
 CONFIG_SIFIVE_E=y
 CONFIG_SIFIVE_U=y
 CONFIG_RISCV_VIRT=y
+CONFIG_OPENTITAN=y
default-configs/riscv64-softmmu.mak +10 -1
index 235c6f473f..aaf6d735bb 100644
@@ -1,3 +1,12 @@
 # Default configuration for riscv64-softmmu
 
-include riscv32-softmmu.mak
+# Uncomment the following lines to disable these optional devices:
+#
+#CONFIG_PCI_DEVICES=n
+
+# Boards:
+#
+CONFIG_SPIKE=y
+CONFIG_SIFIVE_E=y
+CONFIG_SIFIVE_U=y
+CONFIG_RISCV_VIRT=y
include/hw/riscv/opentitan.h new file +68 -0

68 lines not on this page · raw message

hw/riscv/opentitan.c new file +184 -0

184 lines not on this page · raw message

MAINTAINERS +9 -0
index 0944d9c731..3e7d9cb0a5 100644
@@ -1238,6 +1238,15 @@ F: pc-bios/canyonlands.dt[sb]
 F: pc-bios/u-boot-sam460ex-20100605.bin
 F: roms/u-boot-sam460ex
 
+RISC-V Machines
+---------------
+OpenTitan
+M: Alistair Francis <[email protected]>
+L: [email protected]
+S: Supported
+F: hw/riscv/opentitan.c
+F: include/hw/riscv/opentitan.h
+
 SH4 Machines
 ------------
 R2D
hw/riscv/Kconfig +5 -0
index ff9fbe958a..94d19571f7 100644
@@ -27,6 +27,11 @@ config SPIKE
     select HTIF
     select SIFIVE
 
+config OPENTITAN
+    bool
+    select HART
+    select UNIMP
+
 config RISCV_VIRT
     bool
     imply PCI_DEVICES
hw/riscv/Makefile.objs +1 -0
index fc3c6dd7c8..57cc708f5d 100644
@@ -1,6 +1,7 @@
 obj-y += boot.o
 obj-$(CONFIG_SPIKE) += riscv_htif.o
 obj-$(CONFIG_HART) += riscv_hart.o
+obj-$(CONFIG_OPENTITAN) += opentitan.o
 obj-$(CONFIG_SIFIVE_E) += sifive_e.o
 obj-$(CONFIG_SIFIVE_E) += sifive_e_prci.o
 obj-$(CONFIG_SIFIVE) += sifive_clint.o
Signature · 1 line
2.26.2
#9 lore raw

Re: [PATCH v5 06/11] riscv: Initial commit of OpenTitan machine

Hi Alistair,

Quoted 12 lines · Alistair Francis

This adds a barebone OpenTitan machine to QEMU.

---

include/hw/riscv/opentitan.h new file +0 -0
new file mode 100644
index 0000000000..a4b6499444
@@ -0,0 +1,68 @@

[...]

include/hw/riscv/opentitan.h

Quoted 6 lines
include/hw/riscv/opentitan.h new file +6 -0
+
+static const struct MemmapEntry {
+    hwaddr base;
+    hwaddr size;
+} ibex_memmap[] = {
+    [IBEX_ROM] =            {  0x00008000,   0xc000 },

Shouldn't the ROM size be 0x4000 (which make the end of rom at 0xc000) ?

Not sure if that's exactly this platform you are modeling but the following doc says the ROM size is 16kB (0x4000):
https://github.com/lowRISC/opentitan/blob/master/hw/top_earlgrey/doc/_index.md

--
Damien

#10 lore raw

Re: [PATCH v5 06/11] riscv: Initial commit of OpenTitan machine

Quoted 28 lines · Damien Hedde

Hi Alistair,

Quoted 12 lines · Alistair Francis

This adds a barebone OpenTitan machine to QEMU.

---

include/hw/riscv/opentitan.h new file +0 -0
new file mode 100644
index 0000000000..a4b6499444
@@ -0,0 +1,68 @@

[...]

include/hw/riscv/opentitan.h

Quoted 6 lines
include/hw/riscv/opentitan.h new file +6 -0
+
+static const struct MemmapEntry {
+    hwaddr base;
+    hwaddr size;
+} ibex_memmap[] = {
+    [IBEX_ROM] =            {  0x00008000,   0xc000 },

Shouldn't the ROM size be 0x4000 (which make the end of rom at 0xc000) ?

Quoted 3 lines

Not sure if that's exactly this platform you are modeling but the following doc says the ROM size is 16kB (0x4000):
https://github.com/lowRISC/opentitan/blob/master/hw/top_earlgrey/doc/_index.md

Good catch. This is why I prefer the IEC notation:

    [IBEX_ROM] =            {  0x00008000,   16 * KiB },

You can then verify the mapping running 'info mtree' in the monitor.

Quoted 5 lines

--
Damien

#11 lore raw

Re: [PATCH v5 06/11] riscv: Initial commit of OpenTitan machine

Quoted 30 lines · Philippe Mathieu-Daudé
Quoted 28 lines · Damien Hedde

Hi Alistair,

Quoted 12 lines · Alistair Francis

This adds a barebone OpenTitan machine to QEMU.

---

include/hw/riscv/opentitan.h new file +0 -0
new file mode 100644
index 0000000000..a4b6499444
@@ -0,0 +1,68 @@

[...]

include/hw/riscv/opentitan.h

Quoted 6 lines
include/hw/riscv/opentitan.h new file +6 -0
+
+static const struct MemmapEntry {
+    hwaddr base;
+    hwaddr size;
+} ibex_memmap[] = {
+    [IBEX_ROM] =            {  0x00008000,   0xc000 },

Shouldn't the ROM size be 0x4000 (which make the end of rom at 0xc000) ?

Quoted 3 lines
Quoted 3 lines

Not sure if that's exactly this platform you are modeling but the following doc says the ROM size is 16kB (0x4000):
https://github.com/lowRISC/opentitan/blob/master/hw/top_earlgrey/doc/_index.md

Yep you are right, I have fixed this.

Quoted 6 lines

Good catch. This is why I prefer the IEC notation:

    [IBEX_ROM] =            {  0x00008000,   16 * KiB },

You can then verify the mapping running 'info mtree' in the monitor.

Thanks!

Alistair

Quoted 7 lines
Quoted 6 lines

--
Damien

#12 lore raw

Re: [PATCH v5 06/11] riscv: Initial commit of OpenTitan machine

In reply to Alistair Francis
Quoted 8 lines · Alistair Francis

This adds a barebone OpenTitan machine to QEMU.

---
 default-configs/riscv32-softmmu.mak |   1 +
 default-configs/riscv64-softmmu.mak |  11 +-

Just noticed this, but:

Quoted 17 lines
default-configs/riscv32-softmmu.mak +1 -0
index 1ae077ed87..94a236c9c2 100644
@@ -10,3 +10,4 @@ CONFIG_SPIKE=y
 CONFIG_SIFIVE_E=y
 CONFIG_SIFIVE_U=y
 CONFIG_RISCV_VIRT=y
+CONFIG_OPENTITAN=y
default-configs/riscv64-softmmu.mak +0 -1
index 235c6f473f..aaf6d735bb 100644
@@ -1,3 +1,12 @@
 # Default configuration for riscv64-softmmu
 
-include riscv32-softmmu.mak

default-configs/riscv64-softmmu.mak line 12

Quoted 10 lines
default-configs/riscv64-softmmu.mak +10 -0
+# Uncomment the following lines to disable these optional devices:
+#
+#CONFIG_PCI_DEVICES=n
+
+# Boards:
+#
+CONFIG_SPIKE=y
+CONFIG_SIFIVE_E=y
+CONFIG_SIFIVE_U=y
+CONFIG_RISCV_VIRT=y

...shouldn't the riscv64-softmmu config have CONFIG_OPENTITAN too? The usual principle is that the 64-bit executable can run the 32-bit boards too.

thanks
-- PMM

#13 lore raw

Re: [PATCH v5 06/11] riscv: Initial commit of OpenTitan machine

Quoted 29 lines · Peter Maydell
Quoted 8 lines · Alistair Francis

This adds a barebone OpenTitan machine to QEMU.

---
 default-configs/riscv32-softmmu.mak |   1 +
 default-configs/riscv64-softmmu.mak |  11 +-

Just noticed this, but:

Quoted 16 lines
default-configs/riscv32-softmmu.mak +1 -0
index 1ae077ed87..94a236c9c2 100644
@@ -10,3 +10,4 @@ CONFIG_SPIKE=y
 CONFIG_SIFIVE_E=y
 CONFIG_SIFIVE_U=y
 CONFIG_RISCV_VIRT=y
+CONFIG_OPENTITAN=y
default-configs/riscv64-softmmu.mak +0 -0
index 235c6f473f..aaf6d735bb 100644
@@ -1,3 +1,12 @@
 # Default configuration for riscv64-softmmu
 
Quoted 15 lines

default-configs/riscv64-softmmu.mak line 12

Quoted 11 lines
default-configs/riscv64-softmmu.mak +10 -1
-include riscv32-softmmu.mak
+# Uncomment the following lines to disable these optional devices:
+#
+#CONFIG_PCI_DEVICES=n
+
+# Boards:
+#
+CONFIG_SPIKE=y
+CONFIG_SIFIVE_E=y
+CONFIG_SIFIVE_U=y
+CONFIG_RISCV_VIRT=y

...shouldn't the riscv64-softmmu config have CONFIG_OPENTITAN too? The usual principle is that the 64-bit executable can run the 32-bit boards too.

I didn't know that was the general case. I'll send a patch to enable this.

Alistair

Quoted 3 lines

thanks
-- PMM

#14 lore raw

Re: [PATCH v5 06/11] riscv: Initial commit of OpenTitan machine

Quoted 7 lines · Alistair Francis
Quoted 3 lines · Peter Maydell

...shouldn't the riscv64-softmmu config have CONFIG_OPENTITAN too? The usual principle is that the 64-bit executable can run the 32-bit boards too.

I didn't know that was the general case. I'll send a patch to enable this.

Somebody on IRC suggested that the riscv code currently assumes that #ifdef TARGET_RISCV64 implies a 64-bit CPU, ie that the 32-bit CPUs don't actually behave correctly if built into the qemu-system-riscv64 process, so you might want to check that things seem to work when you enable it...

thanks
-- pMM

maintainer
#15 lore raw

Re: [PATCH v5 06/11] riscv: Initial commit of OpenTitan machine

Quoted 7 lines · Peter Maydell
Quoted 6 lines · Alistair Francis
Quoted 3 lines · Peter Maydell

...shouldn't the riscv64-softmmu config have CONFIG_OPENTITAN too? The usual principle is that the 64-bit executable can run the 32-bit boards too.

Quoted 7 lines
Quoted 1 line

I didn't know that was the general case. I'll send a patch to enable this.

Somebody on IRC suggested that the riscv code currently assumes that #ifdef TARGET_RISCV64 implies a 64-bit CPU, ie that the 32-bit CPUs don't actually behave correctly if built into the qemu-system-riscv64 process, so you might want to check that things seem to work when you enable it...

IIRC there's a whole bunch of that floating around the RISC-V port, so just tossing a 32-bit CPU into qemu-system-riscv64 is going to result in garbage. At the time we originally did the port we though we weren't goin to have mixed-XLEN systems, but the ISA now allows for it so in theory we should go fix all that.

I don't think anyone has looked into doing it, though, as there isn't much interest in mixed-XLEN systems (though oddly enough there was a glibc post yesterday about them).

#16 lore raw

Re: [PATCH v5 06/11] riscv: Initial commit of OpenTitan machine

Quoted 9 lines · Palmer Dabbelt
Quoted 7 lines · Peter Maydell
Quoted 6 lines · Alistair Francis
Quoted 3 lines · Peter Maydell

...shouldn't the riscv64-softmmu config have CONFIG_OPENTITAN too? The usual principle is that the 64-bit executable can run the 32-bit boards too.

Quoted 7 lines
Quoted 7 lines
Quoted 1 line

I didn't know that was the general case. I'll send a patch to enable this.

Somebody on IRC suggested that the riscv code currently assumes that #ifdef TARGET_RISCV64 implies a 64-bit CPU, ie that the 32-bit CPUs don't actually behave correctly if built into the qemu-system-riscv64 process, so you might want to check that things seem to work when you enable it...

Yep, we don't even build the 32-bit CPUs for 64-bit.

I'm going to leave this config disabled and we can add a TODO of fixing up riscv64-softmmu to support 32-bit and 64-bit.

Quoted 6 lines

IIRC there's a whole bunch of that floating around the RISC-V port, so just tossing a 32-bit CPU into qemu-system-riscv64 is going to result in garbage. At the time we originally did the port we though we weren't goin to have mixed-XLEN systems, but the ISA now allows for it so in theory we should go fix all that.

Dynamically changing XLEN is then another can of worms. Although there isn't much interest, it becomes more useful with the Hypervisor extensions supporting 32-bit guests on 64-bit systems.

Alistair

Quoted 4 lines

I don't think anyone has looked into doing it, though, as there isn't much interest in mixed-XLEN systems (though oddly enough there was a glibc post yesterday about them).

#17 lore raw

Re: [PATCH v5 6/11] riscv: Initial commit of OpenTitan machine

In reply to Alistair Francis

Hi Alistair,

[Old patch]

Quoted 53 lines · Alistair Francis

This adds a barebone OpenTitan machine to QEMU.

---
  default-configs/riscv32-softmmu.mak |   1 +
  default-configs/riscv64-softmmu.mak |  11 +-
  include/hw/riscv/opentitan.h        |  68 ++++++++++
  hw/riscv/opentitan.c                | 184 ++++++++++++++++++++++++++++
  MAINTAINERS                         |   9 ++
  hw/riscv/Kconfig                    |   5 +
  hw/riscv/Makefile.objs              |   1 +
  7 files changed, 278 insertions(+), 1 deletion(-)

create mode 100644 include/hw/riscv/opentitan.h
create mode 100644 hw/riscv/opentitan.c

include/hw/riscv/opentitan.h new file +30 -0
new file mode 100644
index 0000000000..a4b6499444
@@ -0,0 +1,68 @@
+/*
+ * QEMU RISC-V Board Compatible with OpenTitan FPGA platform
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef HW_OPENTITAN_H
+#define HW_OPENTITAN_H
+
+#include "hw/riscv/riscv_hart.h"
+
+#define TYPE_RISCV_IBEX_SOC "riscv.lowrisc.ibex.soc"
+#define RISCV_IBEX_SOC(obj) \
+    OBJECT_CHECK(LowRISCIbexSoCState, (obj), TYPE_RISCV_IBEX_SOC)
+
+typedef struct LowRISCIbexSoCState {
+    /*< private >*/
+    SysBusDevice parent_obj;

include/hw/riscv/opentitan.h line 40

Quoted 10 lines
include/hw/riscv/opentitan.h new file +10 -0
+
+    /*< public >*/
+    RISCVHartArrayState cpus;
+    MemoryRegion flash_mem;
+    MemoryRegion rom;
+} LowRISCIbexSoCState;
+
+typedef struct OpenTitanState {
+    /*< private >*/
+    SysBusDevice parent_obj;

Shouldn't this object inheritate from MachineState ...?

Quoted 58 lines
include/hw/riscv/opentitan.h new file +3 -0
+    /*< public >*/
+    LowRISCIbexSoCState soc;
+} OpenTitanState;
hw/riscv/opentitan.c new file +47 -0
new file mode 100644
index 0000000000..b4fb836466
@@ -0,0 +1,184 @@
+/*
+ * QEMU RISC-V Board Compatible with OpenTitan FPGA platform
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * Provides a board compatible with the OpenTitan FPGA platform:
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/riscv/opentitan.h"
+#include "qapi/error.h"
+#include "hw/boards.h"
+#include "hw/misc/unimp.h"
+#include "hw/riscv/boot.h"
+#include "exec/address-spaces.h"
+
+static const struct MemmapEntry {
+    hwaddr base;
+    hwaddr size;
+} ibex_memmap[] = {
+    [IBEX_ROM] =            {  0x00008000,   0xc000 },
+    [IBEX_RAM] =            {  0x10000000,  0x10000 },
+    [IBEX_FLASH] =          {  0x20000000,  0x80000 },
+    [IBEX_UART] =           {  0x40000000,  0x10000 },
+    [IBEX_GPIO] =           {  0x40010000,  0x10000 },
+    [IBEX_SPI] =            {  0x40020000,  0x10000 },
+    [IBEX_FLASH_CTRL] =     {  0x40030000,  0x10000 },
+    [IBEX_PINMUX] =         {  0x40070000,  0x10000 },
+    [IBEX_RV_TIMER] =       {  0x40080000,  0x10000 },
+    [IBEX_PLIC] =           {  0x40090000,  0x10000 },
+    [IBEX_PWRMGR] =         {  0x400A0000,  0x10000 },
+    [IBEX_RSTMGR] =         {  0x400B0000,  0x10000 },
+    [IBEX_CLKMGR] =         {  0x400C0000,  0x10000 },
+    [IBEX_AES] =            {  0x40110000,  0x10000 },
+    [IBEX_HMAC] =           {  0x40120000,  0x10000 },

hw/riscv/opentitan.c line 57

Quoted 10 lines
hw/riscv/opentitan.c new file +10 -0
+    [IBEX_ALERT_HANDLER] =  {  0x40130000,  0x10000 },
+    [IBEX_NMI_GEN] =        {  0x40140000,  0x10000 },
+    [IBEX_USBDEV] =         {  0x40150000,  0x10000 },
+    [IBEX_PADCTRL] =        {  0x40160000,  0x10000 }
+};
+
+static void riscv_opentitan_init(MachineState *machine)
+{
+    const struct MemmapEntry *memmap = ibex_memmap;
+    OpenTitanState *s = g_new0(OpenTitanState, 1);

... because looking at QOM design issue, this line looks dubious.

Quoted 24 lines
hw/riscv/opentitan.c new file +24 -0
+    MemoryRegion *sys_mem = get_system_memory();
+    MemoryRegion *main_mem = g_new(MemoryRegion, 1);
+
+    /* Initialize SoC */
+    object_initialize_child(OBJECT(machine), "soc", &s->soc,
+                            sizeof(s->soc), TYPE_RISCV_IBEX_SOC,
+                            &error_abort, NULL);
+    object_property_set_bool(OBJECT(&s->soc), true, "realized",
+                            &error_abort);
+
+    memory_region_init_ram(main_mem, NULL, "riscv.lowrisc.ibex.ram",
+        memmap[IBEX_RAM].size, &error_fatal);
+    memory_region_add_subregion(sys_mem,
+        memmap[IBEX_RAM].base, main_mem);
+
+
+    if (machine->firmware) {
+        riscv_load_firmware(machine->firmware, memmap[IBEX_RAM].base, NULL);
+    }
+
+    if (machine->kernel_filename) {
+        riscv_load_kernel(machine->kernel_filename, NULL);
+    }
+}
#18 lore raw

PATCH v5 7/11 hw/char: Initial commit of Ibex UART

In reply to Alistair Francis

This is the initial commit of the Ibex UART device. Serial TX is working, while RX has been implemeneted but untested.

This is based on the documentation from:
https://docs.opentitan.org/hw/ip/uart/doc/

---
 include/hw/char/ibex_uart.h | 110 ++++++++
 hw/char/ibex_uart.c         | 492 ++++++++++++++++++++++++++++++++++++
 MAINTAINERS                 |   2 +
 hw/char/Makefile.objs       |   1 +
 hw/riscv/Kconfig            |   4 +
 5 files changed, 609 insertions(+)
 create mode 100644 include/hw/char/ibex_uart.h
 create mode 100644 hw/char/ibex_uart.c
include/hw/char/ibex_uart.h new file +110 -0
new file mode 100644
index 0000000000..2bec772615
@@ -0,0 +1,110 @@
+/*
+ * QEMU lowRISC Ibex UART device
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef HW_IBEX_UART_H
+#define HW_IBEX_UART_H
+
+#include "hw/sysbus.h"
+#include "chardev/char-fe.h"
+#include "qemu/timer.h"
+
+#define IBEX_UART_INTR_STATE   0x00
+    #define INTR_STATE_TX_WATERMARK (1 << 0)
+    #define INTR_STATE_RX_WATERMARK (1 << 1)
+    #define INTR_STATE_TX_EMPTY     (1 << 2)
+    #define INTR_STATE_RX_OVERFLOW  (1 << 3)
+#define IBEX_UART_INTR_ENABLE  0x04
+#define IBEX_UART_INTR_TEST    0x08
+
+#define IBEX_UART_CTRL         0x0c
+    #define UART_CTRL_TX_ENABLE     (1 << 0)
+    #define UART_CTRL_RX_ENABLE     (1 << 1)
+    #define UART_CTRL_NF            (1 << 2)
+    #define UART_CTRL_SLPBK         (1 << 4)
+    #define UART_CTRL_LLPBK         (1 << 5)
+    #define UART_CTRL_PARITY_EN     (1 << 6)
+    #define UART_CTRL_PARITY_ODD    (1 << 7)
+    #define UART_CTRL_RXBLVL        (3 << 8)
+    #define UART_CTRL_NCO           (0xFFFF << 16)
+
+#define IBEX_UART_STATUS       0x10
+    #define UART_STATUS_TXFULL  (1 << 0)
+    #define UART_STATUS_RXFULL  (1 << 1)
+    #define UART_STATUS_TXEMPTY (1 << 2)
+    #define UART_STATUS_RXIDLE  (1 << 4)
+    #define UART_STATUS_RXEMPTY (1 << 5)
+
+#define IBEX_UART_RDATA        0x14
+#define IBEX_UART_WDATA        0x18
+
+#define IBEX_UART_FIFO_CTRL    0x1c
+    #define FIFO_CTRL_RXRST          (1 << 0)
+    #define FIFO_CTRL_TXRST          (1 << 1)
+    #define FIFO_CTRL_RXILVL         (7 << 2)
+    #define FIFO_CTRL_RXILVL_SHIFT   (2)
+    #define FIFO_CTRL_TXILVL         (3 << 5)
+    #define FIFO_CTRL_TXILVL_SHIFT   (5)
+
+#define IBEX_UART_FIFO_STATUS  0x20
+#define IBEX_UART_OVRD         0x24
+#define IBEX_UART_VAL          0x28
+#define IBEX_UART_TIMEOUT_CTRL 0x2c
+
+#define IBEX_UART_TX_FIFO_SIZE 16
+
+#define TYPE_IBEX_UART "ibex-uart"
+#define IBEX_UART(obj) \
+    OBJECT_CHECK(IbexUartState, (obj), TYPE_IBEX_UART)
+
+typedef struct {
+    /* <private> */
+    SysBusDevice parent_obj;
+
+    /* <public> */
+    MemoryRegion mmio;
+
+    uint8_t tx_fifo[IBEX_UART_TX_FIFO_SIZE];
+    uint32_t tx_level;
+
+    QEMUTimer *fifo_trigger_handle;
+    uint64_t char_tx_time;
+
+    uint32_t uart_intr_state;
+    uint32_t uart_intr_enable;
+    uint32_t uart_ctrl;
+    uint32_t uart_status;
+    uint32_t uart_rdata;
+    uint32_t uart_fifo_ctrl;
+    uint32_t uart_fifo_status;
+    uint32_t uart_ovrd;
+    uint32_t uart_val;
+    uint32_t uart_timeout_ctrl;
+
+    CharBackend chr;
+    qemu_irq tx_watermark;
+    qemu_irq rx_watermark;
+    qemu_irq tx_empty;
+    qemu_irq rx_overflow;
+} IbexUartState;
+#endif /* HW_IBEX_UART_H */
hw/char/ibex_uart.c new file +492 -0

492 lines not on this page · raw message

MAINTAINERS +2 -0
index 3e7d9cb0a5..a1ce186a7e 100644
@@ -1245,7 +1245,9 @@ M: Alistair Francis <[email protected]>
 L: [email protected]
 S: Supported
 F: hw/riscv/opentitan.c
+F: hw/char/ibex_uart.c
 F: include/hw/riscv/opentitan.h
+F: include/hw/char/ibex_uart.h
 
 SH4 Machines
 ------------
hw/char/Makefile.objs +1 -0
index 9e9a6c1aff..633996be5b 100644
@@ -12,6 +12,7 @@ common-obj-$(CONFIG_VIRTIO_SERIAL) += virtio-console.o
 common-obj-$(CONFIG_XILINX) += xilinx_uartlite.o
 common-obj-$(CONFIG_XEN) += xen_console.o
 common-obj-$(CONFIG_CADENCE) += cadence_uart.o
+common-obj-$(CONFIG_IBEX) += ibex_uart.o
 
 common-obj-$(CONFIG_EXYNOS4) += exynos4210_uart.o
 common-obj-$(CONFIG_COLDFIRE) += mcf_uart.o
hw/riscv/Kconfig +4 -0
index 94d19571f7..28947ef3e0 100644
@@ -4,6 +4,9 @@ config HTIF
 config HART
     bool
 
+config IBEX
+    bool
+
 config SIFIVE
     bool
     select MSI_NONBROKEN
@@ -29,6 +32,7 @@ config SPIKE
 
 config OPENTITAN
     bool
+    select IBEX
     select HART
     select UNIMP
 
Signature · 1 line
2.26.2
#19 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

Quoted 8 lines · Alistair Francis

This is the initial commit of the Ibex UART device. Serial TX is working, while RX has been implemeneted but untested.

This is based on the documentation from:
https://docs.opentitan.org/hw/ip/uart/doc/

Ping! This is the last patch not reviewed.

Alistair

Quoted 675 lines
---
 include/hw/char/ibex_uart.h | 110 ++++++++
 hw/char/ibex_uart.c         | 492 ++++++++++++++++++++++++++++++++++++
 MAINTAINERS                 |   2 +
 hw/char/Makefile.objs       |   1 +
 hw/riscv/Kconfig            |   4 +
 5 files changed, 609 insertions(+)
 create mode 100644 include/hw/char/ibex_uart.h
 create mode 100644 hw/char/ibex_uart.c
include/hw/char/ibex_uart.h new file +110 -0
new file mode 100644
index 0000000000..2bec772615
@@ -0,0 +1,110 @@
+/*
+ * QEMU lowRISC Ibex UART device
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef HW_IBEX_UART_H
+#define HW_IBEX_UART_H
+
+#include "hw/sysbus.h"
+#include "chardev/char-fe.h"
+#include "qemu/timer.h"
+
+#define IBEX_UART_INTR_STATE   0x00
+    #define INTR_STATE_TX_WATERMARK (1 << 0)
+    #define INTR_STATE_RX_WATERMARK (1 << 1)
+    #define INTR_STATE_TX_EMPTY     (1 << 2)
+    #define INTR_STATE_RX_OVERFLOW  (1 << 3)
+#define IBEX_UART_INTR_ENABLE  0x04
+#define IBEX_UART_INTR_TEST    0x08
+
+#define IBEX_UART_CTRL         0x0c
+    #define UART_CTRL_TX_ENABLE     (1 << 0)
+    #define UART_CTRL_RX_ENABLE     (1 << 1)
+    #define UART_CTRL_NF            (1 << 2)
+    #define UART_CTRL_SLPBK         (1 << 4)
+    #define UART_CTRL_LLPBK         (1 << 5)
+    #define UART_CTRL_PARITY_EN     (1 << 6)
+    #define UART_CTRL_PARITY_ODD    (1 << 7)
+    #define UART_CTRL_RXBLVL        (3 << 8)
+    #define UART_CTRL_NCO           (0xFFFF << 16)
+
+#define IBEX_UART_STATUS       0x10
+    #define UART_STATUS_TXFULL  (1 << 0)
+    #define UART_STATUS_RXFULL  (1 << 1)
+    #define UART_STATUS_TXEMPTY (1 << 2)
+    #define UART_STATUS_RXIDLE  (1 << 4)
+    #define UART_STATUS_RXEMPTY (1 << 5)
+
+#define IBEX_UART_RDATA        0x14
+#define IBEX_UART_WDATA        0x18
+
+#define IBEX_UART_FIFO_CTRL    0x1c
+    #define FIFO_CTRL_RXRST          (1 << 0)
+    #define FIFO_CTRL_TXRST          (1 << 1)
+    #define FIFO_CTRL_RXILVL         (7 << 2)
+    #define FIFO_CTRL_RXILVL_SHIFT   (2)
+    #define FIFO_CTRL_TXILVL         (3 << 5)
+    #define FIFO_CTRL_TXILVL_SHIFT   (5)
+
+#define IBEX_UART_FIFO_STATUS  0x20
+#define IBEX_UART_OVRD         0x24
+#define IBEX_UART_VAL          0x28
+#define IBEX_UART_TIMEOUT_CTRL 0x2c
+
+#define IBEX_UART_TX_FIFO_SIZE 16
+
+#define TYPE_IBEX_UART "ibex-uart"
+#define IBEX_UART(obj) \
+    OBJECT_CHECK(IbexUartState, (obj), TYPE_IBEX_UART)
+
+typedef struct {
+    /* <private> */
+    SysBusDevice parent_obj;
+
+    /* <public> */
+    MemoryRegion mmio;
+
+    uint8_t tx_fifo[IBEX_UART_TX_FIFO_SIZE];
+    uint32_t tx_level;
+
+    QEMUTimer *fifo_trigger_handle;
+    uint64_t char_tx_time;
+
+    uint32_t uart_intr_state;
+    uint32_t uart_intr_enable;
+    uint32_t uart_ctrl;
+    uint32_t uart_status;
+    uint32_t uart_rdata;
+    uint32_t uart_fifo_ctrl;
+    uint32_t uart_fifo_status;
+    uint32_t uart_ovrd;
+    uint32_t uart_val;
+    uint32_t uart_timeout_ctrl;
+
+    CharBackend chr;
+    qemu_irq tx_watermark;
+    qemu_irq rx_watermark;
+    qemu_irq tx_empty;
+    qemu_irq rx_overflow;
+} IbexUartState;
+#endif /* HW_IBEX_UART_H */
hw/char/ibex_uart.c new file +492 -0
new file mode 100644
index 0000000000..c416325d73
@@ -0,0 +1,492 @@
+/*
+ * QEMU lowRISC Ibex UART device
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * For details check the documentation here:
+ *    https://docs.opentitan.org/hw/ip/uart/doc/
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/char/ibex_uart.h"
+#include "hw/irq.h"
+#include "hw/qdev-properties.h"
+#include "migration/vmstate.h"
+#include "qemu/log.h"
+#include "qemu/module.h"
+
+static void ibex_uart_update_irqs(IbexUartState *s)
+{
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_TX_WATERMARK) {
+        qemu_set_irq(s->tx_watermark, 1);
+    } else {
+        qemu_set_irq(s->tx_watermark, 0);
+    }
+
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_RX_WATERMARK) {
+        qemu_set_irq(s->rx_watermark, 1);
+    } else {
+        qemu_set_irq(s->rx_watermark, 0);
+    }
+
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_TX_EMPTY) {
+        qemu_set_irq(s->tx_empty, 1);
+    } else {
+        qemu_set_irq(s->tx_empty, 0);
+    }
+
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_RX_OVERFLOW) {
+        qemu_set_irq(s->rx_overflow, 1);
+    } else {
+        qemu_set_irq(s->rx_overflow, 0);
+    }
+}
+
+static int ibex_uart_can_receive(void *opaque)
+{
+    IbexUartState *s = opaque;
+
+    if (s->uart_ctrl & UART_CTRL_RX_ENABLE) {
+        return 1;
+    }
+
+    return 0;
+}
+
+static void ibex_uart_receive(void *opaque, const uint8_t *buf, int size)
+{
+    IbexUartState *s = opaque;
+    uint8_t rx_fifo_level = (s->uart_fifo_ctrl & FIFO_CTRL_RXILVL)
+                            >> FIFO_CTRL_RXILVL_SHIFT;
+
+    s->uart_rdata = *buf;
+
+    s->uart_status &= ~UART_STATUS_RXIDLE;
+    s->uart_status &= ~UART_STATUS_RXEMPTY;
+
+    if (size > rx_fifo_level) {
+        s->uart_intr_state |= INTR_STATE_RX_WATERMARK;
+    }
+
+    ibex_uart_update_irqs(s);
+}
+
+static gboolean ibex_uart_xmit(GIOChannel *chan, GIOCondition cond,
+                               void *opaque)
+{
+    IbexUartState *s = opaque;
+    uint8_t tx_fifo_level = (s->uart_fifo_ctrl & FIFO_CTRL_TXILVL)
+                            >> FIFO_CTRL_TXILVL_SHIFT;
+    int ret;
+
+    /* instant drain the fifo when there's no back-end */
+    if (!qemu_chr_fe_backend_connected(&s->chr)) {
+        s->tx_level = 0;
+        return FALSE;
+    }
+
+    if (!s->tx_level) {
+        s->uart_status &= UART_STATUS_TXFULL;
+        s->uart_status |= UART_STATUS_TXEMPTY;
+        s->uart_intr_state |= INTR_STATE_TX_EMPTY;
+        s->uart_intr_state &= ~INTR_STATE_TX_WATERMARK;
+        ibex_uart_update_irqs(s);
+        return FALSE;
+    }
+
+    ret = qemu_chr_fe_write(&s->chr, s->tx_fifo, s->tx_level);
+
+    if (ret >= 0) {
+        s->tx_level -= ret;
+        memmove(s->tx_fifo, s->tx_fifo + ret, s->tx_level);
+    }
+
+    if (s->tx_level) {
+        guint r = qemu_chr_fe_add_watch(&s->chr, G_IO_OUT | G_IO_HUP,
+                                        ibex_uart_xmit, s);
+        if (!r) {
+            s->tx_level = 0;
+            return FALSE;
+        }
+    }
+
+    /* Clear the TX Full bit */
+    if (s->tx_level != IBEX_UART_TX_FIFO_SIZE) {
+        s->uart_status &= ~UART_STATUS_TXFULL;
+    }
+
+    /* Disable the TX_WATERMARK IRQ */
+    if (s->tx_level < tx_fifo_level) {
+        s->uart_intr_state &= ~INTR_STATE_TX_WATERMARK;
+    }
+
+    /* Set TX empty */
+    if (s->tx_level == 0) {
+        s->uart_status |= UART_STATUS_TXEMPTY;
+        s->uart_intr_state |= INTR_STATE_TX_EMPTY;
+    }
+
+    ibex_uart_update_irqs(s);
+    return FALSE;
+}
+
+static void uart_write_tx_fifo(IbexUartState *s, const uint8_t *buf,
+                               int size)
+{
+    uint64_t current_time = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
+    uint8_t tx_fifo_level = (s->uart_fifo_ctrl & FIFO_CTRL_TXILVL)
+                            >> FIFO_CTRL_TXILVL_SHIFT;
+
+    if (size > IBEX_UART_TX_FIFO_SIZE - s->tx_level) {
+        size = IBEX_UART_TX_FIFO_SIZE - s->tx_level;
+        qemu_log_mask(LOG_GUEST_ERROR, "ibex_uart: TX FIFO overflow");
+    }
+
+    memcpy(s->tx_fifo + s->tx_level, buf, size);
+    s->tx_level += size;
+
+    if (s->tx_level > 0) {
+        s->uart_status &= ~UART_STATUS_TXEMPTY;
+    }
+
+    if (s->tx_level >= tx_fifo_level) {
+        s->uart_intr_state |= INTR_STATE_TX_WATERMARK;
+        ibex_uart_update_irqs(s);
+    }
+
+    if (s->tx_level == IBEX_UART_TX_FIFO_SIZE) {
+        s->uart_status |= UART_STATUS_TXFULL;
+    }
+
+    timer_mod(s->fifo_trigger_handle, current_time +
+              (s->char_tx_time * 4));
+}
+
+static void ibex_uart_reset(DeviceState *dev)
+{
+    IbexUartState *s = IBEX_UART(dev);
+
+    s->uart_intr_state = 0x00000000;
+    s->uart_intr_state = 0x00000000;
+    s->uart_intr_enable = 0x00000000;
+    s->uart_ctrl = 0x00000000;
+    s->uart_status = 0x0000003c;
+    s->uart_rdata = 0x00000000;
+    s->uart_fifo_ctrl = 0x00000000;
+    s->uart_fifo_status = 0x00000000;
+    s->uart_ovrd = 0x00000000;
+    s->uart_val = 0x00000000;
+    s->uart_timeout_ctrl = 0x00000000;
+
+    s->tx_level = 0;
+
+    s->char_tx_time = (NANOSECONDS_PER_SECOND / 230400) * 10;
+
+    ibex_uart_update_irqs(s);
+}
+
+static uint64_t ibex_uart_read(void *opaque, hwaddr addr,
+                                       unsigned int size)
+{
+    IbexUartState *s = opaque;
+    uint64_t retvalue = 0;
+
+    switch (addr) {
+    case IBEX_UART_INTR_STATE:
+        retvalue = s->uart_intr_state;
+        break;
+    case IBEX_UART_INTR_ENABLE:
+        retvalue = s->uart_intr_enable;
+        break;
+    case IBEX_UART_INTR_TEST:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: wdata is write only\n", __func__);
+        break;
+
+    case IBEX_UART_CTRL:
+        retvalue = s->uart_ctrl;
+        break;
+    case IBEX_UART_STATUS:
+        retvalue = s->uart_status;
+        break;
+
+    case IBEX_UART_RDATA:
+        retvalue = s->uart_rdata;
+        if (s->uart_ctrl & UART_CTRL_RX_ENABLE) {
+            qemu_chr_fe_accept_input(&s->chr);
+
+            s->uart_status |= UART_STATUS_RXIDLE;
+            s->uart_status |= UART_STATUS_RXEMPTY;
+        }
+        break;
+    case IBEX_UART_WDATA:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: wdata is write only\n", __func__);
+        break;
+
+    case IBEX_UART_FIFO_CTRL:
+        retvalue = s->uart_fifo_ctrl;
+        break;
+    case IBEX_UART_FIFO_STATUS:
+        retvalue = s->uart_fifo_status;
+
+        retvalue |= s->tx_level & 0x1F;
+
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: RX fifos are not supported\n", __func__);
+        break;
+
+    case IBEX_UART_OVRD:
+        retvalue = s->uart_ovrd;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: ovrd is not supported\n", __func__);
+        break;
+    case IBEX_UART_VAL:
+        retvalue = s->uart_val;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: val is not supported\n", __func__);
+        break;
+    case IBEX_UART_TIMEOUT_CTRL:
+        retvalue = s->uart_timeout_ctrl;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: timeout_ctrl is not supported\n", __func__);
+        break;
+    default:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Bad offset 0x%"HWADDR_PRIx"\n", __func__, addr);
+        return 0;
+    }
+
+    return retvalue;
+}
+
+static void ibex_uart_write(void *opaque, hwaddr addr,
+                                  uint64_t val64, unsigned int size)
+{
+    IbexUartState *s = opaque;
+    uint32_t value = val64;
+
+    switch (addr) {
+    case IBEX_UART_INTR_STATE:
+        /* Write 1 clear */
+        s->uart_intr_state &= ~value;
+        ibex_uart_update_irqs(s);
+        break;
+    case IBEX_UART_INTR_ENABLE:
+        s->uart_intr_enable = value;
+        ibex_uart_update_irqs(s);
+        break;
+    case IBEX_UART_INTR_TEST:
+        s->uart_intr_state |= value;
+        ibex_uart_update_irqs(s);
+        break;
+
+    case IBEX_UART_CTRL:
+        s->uart_ctrl = value;
+
+        if (value & UART_CTRL_NF) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_NF is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_SLPBK) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_SLPBK is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_LLPBK) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_LLPBK is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_PARITY_EN) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_PARITY_EN is not supported\n",
+                          __func__);
+        }
+        if (value & UART_CTRL_PARITY_ODD) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_PARITY_ODD is not supported\n",
+                          __func__);
+        }
+        if (value & UART_CTRL_RXBLVL) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_RXBLVL is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_NCO) {
+            uint64_t baud = ((value & UART_CTRL_NCO) >> 16);
+            baud *= 1000;
+            baud /= 2 ^ 20;
+
+            s->char_tx_time = (NANOSECONDS_PER_SECOND / baud) * 10;
+        }
+        break;
+    case IBEX_UART_STATUS:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: status is read only\n", __func__);
+        break;
+
+    case IBEX_UART_RDATA:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: rdata is read only\n", __func__);
+        break;
+    case IBEX_UART_WDATA:
+        uart_write_tx_fifo(s, (uint8_t *) &value, 1);
+        break;
+
+    case IBEX_UART_FIFO_CTRL:
+        s->uart_fifo_ctrl = value;
+
+        if (value & FIFO_CTRL_RXRST) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: RX fifos are not supported\n", __func__);
+        }
+        if (value & FIFO_CTRL_TXRST) {
+            s->tx_level = 0;
+        }
+        break;
+    case IBEX_UART_FIFO_STATUS:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: fifo_status is read only\n", __func__);
+        break;
+
+    case IBEX_UART_OVRD:
+        s->uart_ovrd = value;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: ovrd is not supported\n", __func__);
+        break;
+    case IBEX_UART_VAL:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: val is read only\n", __func__);
+        break;
+    case IBEX_UART_TIMEOUT_CTRL:
+        s->uart_timeout_ctrl = value;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: timeout_ctrl is not supported\n", __func__);
+        break;
+    default:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Bad offset 0x%"HWADDR_PRIx"\n", __func__, addr);
+    }
+}
+
+static void fifo_trigger_update(void *opaque)
+{
+    IbexUartState *s = opaque;
+
+    if (s->uart_ctrl & UART_CTRL_TX_ENABLE) {
+        ibex_uart_xmit(NULL, G_IO_OUT, s);
+    }
+}
+
+static const MemoryRegionOps ibex_uart_ops = {
+    .read = ibex_uart_read,
+    .write = ibex_uart_write,
+    .endianness = DEVICE_NATIVE_ENDIAN,
+    .impl.min_access_size = 4,
+    .impl.max_access_size = 4,
+};
+
+static int cadence_uart_post_load(void *opaque, int version_id)
+{
+    IbexUartState *s = opaque;
+
+    ibex_uart_update_irqs(s);
+    return 0;
+}
+
+static const VMStateDescription vmstate_ibex_uart = {
+    .name = TYPE_IBEX_UART,
+    .version_id = 1,
+    .minimum_version_id = 1,
+    .post_load = cadence_uart_post_load,
+    .fields = (VMStateField[]) {
+        VMSTATE_UINT8_ARRAY(tx_fifo, IbexUartState,
+                            IBEX_UART_TX_FIFO_SIZE),
+        VMSTATE_UINT32(tx_level, IbexUartState),
+        VMSTATE_UINT64(char_tx_time, IbexUartState),
+        VMSTATE_TIMER_PTR(fifo_trigger_handle, IbexUartState),
+        VMSTATE_UINT32(uart_intr_state, IbexUartState),
+        VMSTATE_UINT32(uart_intr_enable, IbexUartState),
+        VMSTATE_UINT32(uart_ctrl, IbexUartState),
+        VMSTATE_UINT32(uart_status, IbexUartState),
+        VMSTATE_UINT32(uart_rdata, IbexUartState),
+        VMSTATE_UINT32(uart_fifo_ctrl, IbexUartState),
+        VMSTATE_UINT32(uart_fifo_status, IbexUartState),
+        VMSTATE_UINT32(uart_ovrd, IbexUartState),
+        VMSTATE_UINT32(uart_val, IbexUartState),
+        VMSTATE_UINT32(uart_timeout_ctrl, IbexUartState),
+        VMSTATE_END_OF_LIST()
+    }
+};
+
+static Property ibex_uart_properties[] = {
+    DEFINE_PROP_CHR("chardev", IbexUartState, chr),
+    DEFINE_PROP_END_OF_LIST(),
+};
+
+static void ibex_uart_init(Object *obj)
+{
+    IbexUartState *s = IBEX_UART(obj);
+
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->tx_watermark);
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->rx_watermark);
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->tx_empty);
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->rx_overflow);
+
+    memory_region_init_io(&s->mmio, obj, &ibex_uart_ops, s,
+                          TYPE_IBEX_UART, 0x400);
+    sysbus_init_mmio(SYS_BUS_DEVICE(obj), &s->mmio);
+}
+
+static void ibex_uart_realize(DeviceState *dev, Error **errp)
+{
+    IbexUartState *s = IBEX_UART(dev);
+
+    s->fifo_trigger_handle = timer_new_ns(QEMU_CLOCK_VIRTUAL,
+                                          fifo_trigger_update, s);
+
+    qemu_chr_fe_set_handlers(&s->chr, ibex_uart_can_receive,
+                             ibex_uart_receive, NULL, NULL,
+                             s, NULL, true);
+}
+
+static void ibex_uart_class_init(ObjectClass *klass, void *data)
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+
+    dc->reset = ibex_uart_reset;
+    dc->realize = ibex_uart_realize;
+    dc->vmsd = &vmstate_ibex_uart;
+    device_class_set_props(dc, ibex_uart_properties);
+}
+
+static const TypeInfo ibex_uart_info = {
+    .name          = TYPE_IBEX_UART,
+    .parent        = TYPE_SYS_BUS_DEVICE,
+    .instance_size = sizeof(IbexUartState),
+    .instance_init = ibex_uart_init,
+    .class_init    = ibex_uart_class_init,
+};
+
+static void ibex_uart_register_types(void)
+{
+    type_register_static(&ibex_uart_info);
+}
+
+type_init(ibex_uart_register_types)
MAINTAINERS +2 -0
index 3e7d9cb0a5..a1ce186a7e 100644
@@ -1245,7 +1245,9 @@ M: Alistair Francis <[email protected]>
 L: [email protected]
 S: Supported
 F: hw/riscv/opentitan.c
+F: hw/char/ibex_uart.c
 F: include/hw/riscv/opentitan.h
+F: include/hw/char/ibex_uart.h
 
 SH4 Machines
 ------------
hw/char/Makefile.objs +1 -0
index 9e9a6c1aff..633996be5b 100644
@@ -12,6 +12,7 @@ common-obj-$(CONFIG_VIRTIO_SERIAL) += virtio-console.o
 common-obj-$(CONFIG_XILINX) += xilinx_uartlite.o
 common-obj-$(CONFIG_XEN) += xen_console.o
 common-obj-$(CONFIG_CADENCE) += cadence_uart.o
+common-obj-$(CONFIG_IBEX) += ibex_uart.o
 
 common-obj-$(CONFIG_EXYNOS4) += exynos4210_uart.o
 common-obj-$(CONFIG_COLDFIRE) += mcf_uart.o
hw/riscv/Kconfig +4 -1
index 94d19571f7..28947ef3e0 100644
@@ -4,6 +4,9 @@ config HTIF
 config HART
     bool
 
+config IBEX
+    bool
+
 config SIFIVE
     bool
     select MSI_NONBROKEN
@@ -29,6 +32,7 @@ config SPIKE
 
 config OPENTITAN
     bool
+    select IBEX
     select HART
     select UNIMP
 
--

2.26.2

#20 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

In reply to Alistair Francis
Quoted 236 lines · Alistair Francis

This is the initial commit of the Ibex UART device. Serial TX is working, while RX has been implemeneted but untested.

This is based on the documentation from:
https://docs.opentitan.org/hw/ip/uart/doc/

---
  include/hw/char/ibex_uart.h | 110 ++++++++
  hw/char/ibex_uart.c         | 492 ++++++++++++++++++++++++++++++++++++
  MAINTAINERS                 |   2 +
  hw/char/Makefile.objs       |   1 +
  hw/riscv/Kconfig            |   4 +
  5 files changed, 609 insertions(+)

create mode 100644 include/hw/char/ibex_uart.h
create mode 100644 hw/char/ibex_uart.c

include/hw/char/ibex_uart.h new file +110 -0
new file mode 100644
index 0000000000..2bec772615
@@ -0,0 +1,110 @@
+/*
+ * QEMU lowRISC Ibex UART device
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef HW_IBEX_UART_H
+#define HW_IBEX_UART_H
+
+#include "hw/sysbus.h"
+#include "chardev/char-fe.h"
+#include "qemu/timer.h"
+
+#define IBEX_UART_INTR_STATE   0x00
+    #define INTR_STATE_TX_WATERMARK (1 << 0)
+    #define INTR_STATE_RX_WATERMARK (1 << 1)
+    #define INTR_STATE_TX_EMPTY     (1 << 2)
+    #define INTR_STATE_RX_OVERFLOW  (1 << 3)
+#define IBEX_UART_INTR_ENABLE  0x04
+#define IBEX_UART_INTR_TEST    0x08
+
+#define IBEX_UART_CTRL         0x0c
+    #define UART_CTRL_TX_ENABLE     (1 << 0)
+    #define UART_CTRL_RX_ENABLE     (1 << 1)
+    #define UART_CTRL_NF            (1 << 2)
+    #define UART_CTRL_SLPBK         (1 << 4)
+    #define UART_CTRL_LLPBK         (1 << 5)
+    #define UART_CTRL_PARITY_EN     (1 << 6)
+    #define UART_CTRL_PARITY_ODD    (1 << 7)
+    #define UART_CTRL_RXBLVL        (3 << 8)
+    #define UART_CTRL_NCO           (0xFFFF << 16)
+
+#define IBEX_UART_STATUS       0x10
+    #define UART_STATUS_TXFULL  (1 << 0)
+    #define UART_STATUS_RXFULL  (1 << 1)
+    #define UART_STATUS_TXEMPTY (1 << 2)
+    #define UART_STATUS_RXIDLE  (1 << 4)
+    #define UART_STATUS_RXEMPTY (1 << 5)
+
+#define IBEX_UART_RDATA        0x14
+#define IBEX_UART_WDATA        0x18
+
+#define IBEX_UART_FIFO_CTRL    0x1c
+    #define FIFO_CTRL_RXRST          (1 << 0)
+    #define FIFO_CTRL_TXRST          (1 << 1)
+    #define FIFO_CTRL_RXILVL         (7 << 2)
+    #define FIFO_CTRL_RXILVL_SHIFT   (2)
+    #define FIFO_CTRL_TXILVL         (3 << 5)
+    #define FIFO_CTRL_TXILVL_SHIFT   (5)
+
+#define IBEX_UART_FIFO_STATUS  0x20
+#define IBEX_UART_OVRD         0x24
+#define IBEX_UART_VAL          0x28
+#define IBEX_UART_TIMEOUT_CTRL 0x2c
+
+#define IBEX_UART_TX_FIFO_SIZE 16
+
+#define TYPE_IBEX_UART "ibex-uart"
+#define IBEX_UART(obj) \
+    OBJECT_CHECK(IbexUartState, (obj), TYPE_IBEX_UART)
+
+typedef struct {
+    /* <private> */
+    SysBusDevice parent_obj;
+
+    /* <public> */
+    MemoryRegion mmio;
+
+    uint8_t tx_fifo[IBEX_UART_TX_FIFO_SIZE];
+    uint32_t tx_level;
+
+    QEMUTimer *fifo_trigger_handle;
+    uint64_t char_tx_time;
+
+    uint32_t uart_intr_state;
+    uint32_t uart_intr_enable;
+    uint32_t uart_ctrl;
+    uint32_t uart_status;
+    uint32_t uart_rdata;
+    uint32_t uart_fifo_ctrl;
+    uint32_t uart_fifo_status;
+    uint32_t uart_ovrd;
+    uint32_t uart_val;
+    uint32_t uart_timeout_ctrl;
+
+    CharBackend chr;
+    qemu_irq tx_watermark;
+    qemu_irq rx_watermark;
+    qemu_irq tx_empty;
+    qemu_irq rx_overflow;
+} IbexUartState;
+#endif /* HW_IBEX_UART_H */
hw/char/ibex_uart.c new file +97 -0
new file mode 100644
index 0000000000..c416325d73
@@ -0,0 +1,492 @@
+/*
+ * QEMU lowRISC Ibex UART device
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * For details check the documentation here:
+ *    https://docs.opentitan.org/hw/ip/uart/doc/
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/char/ibex_uart.h"
+#include "hw/irq.h"
+#include "hw/qdev-properties.h"
+#include "migration/vmstate.h"
+#include "qemu/log.h"
+#include "qemu/module.h"
+
+static void ibex_uart_update_irqs(IbexUartState *s)
+{
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_TX_WATERMARK) {
+        qemu_set_irq(s->tx_watermark, 1);
+    } else {
+        qemu_set_irq(s->tx_watermark, 0);
+    }
+
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_RX_WATERMARK) {
+        qemu_set_irq(s->rx_watermark, 1);
+    } else {
+        qemu_set_irq(s->rx_watermark, 0);
+    }
+
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_TX_EMPTY) {
+        qemu_set_irq(s->tx_empty, 1);
+    } else {
+        qemu_set_irq(s->tx_empty, 0);
+    }
+
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_RX_OVERFLOW) {
+        qemu_set_irq(s->rx_overflow, 1);
+    } else {
+        qemu_set_irq(s->rx_overflow, 0);
+    }
+}
+
+static int ibex_uart_can_receive(void *opaque)
+{
+    IbexUartState *s = opaque;
+
+    if (s->uart_ctrl & UART_CTRL_RX_ENABLE) {
+        return 1;
+    }
+
+    return 0;
+}
+
+static void ibex_uart_receive(void *opaque, const uint8_t *buf, int size)
+{
+    IbexUartState *s = opaque;
+    uint8_t rx_fifo_level = (s->uart_fifo_ctrl & FIFO_CTRL_RXILVL)
+                            >> FIFO_CTRL_RXILVL_SHIFT;
+
+    s->uart_rdata = *buf;
+
+    s->uart_status &= ~UART_STATUS_RXIDLE;
+    s->uart_status &= ~UART_STATUS_RXEMPTY;
+
+    if (size > rx_fifo_level) {
+        s->uart_intr_state |= INTR_STATE_RX_WATERMARK;
+    }
+
+    ibex_uart_update_irqs(s);
+}
+
+static gboolean ibex_uart_xmit(GIOChannel *chan, GIOCondition cond,
+                               void *opaque)
+{
+    IbexUartState *s = opaque;
+    uint8_t tx_fifo_level = (s->uart_fifo_ctrl & FIFO_CTRL_TXILVL)
+                            >> FIFO_CTRL_TXILVL_SHIFT;

hw/char/ibex_uart.c line 107

Quoted 10 lines
hw/char/ibex_uart.c new file +10 -0
+    int ret;
+
+    /* instant drain the fifo when there's no back-end */
+    if (!qemu_chr_fe_backend_connected(&s->chr)) {
+        s->tx_level = 0;
+        return FALSE;
+    }
+
+    if (!s->tx_level) {
+        s->uart_status &= UART_STATUS_TXFULL;

Maybe  s->uart_status &= ~UART_STATUS_TXFULL;

Quoted 288 lines
hw/char/ibex_uart.c new file +288 -0
+        s->uart_status |= UART_STATUS_TXEMPTY;
+        s->uart_intr_state |= INTR_STATE_TX_EMPTY;
+        s->uart_intr_state &= ~INTR_STATE_TX_WATERMARK;
+        ibex_uart_update_irqs(s);
+        return FALSE;
+    }
+
+    ret = qemu_chr_fe_write(&s->chr, s->tx_fifo, s->tx_level);
+
+    if (ret >= 0) {
+        s->tx_level -= ret;
+        memmove(s->tx_fifo, s->tx_fifo + ret, s->tx_level);
+    }
+
+    if (s->tx_level) {
+        guint r = qemu_chr_fe_add_watch(&s->chr, G_IO_OUT | G_IO_HUP,
+                                        ibex_uart_xmit, s);
+        if (!r) {
+            s->tx_level = 0;
+            return FALSE;
+        }
+    }
+
+    /* Clear the TX Full bit */
+    if (s->tx_level != IBEX_UART_TX_FIFO_SIZE) {
+        s->uart_status &= ~UART_STATUS_TXFULL;
+    }
+
+    /* Disable the TX_WATERMARK IRQ */
+    if (s->tx_level < tx_fifo_level) {
+        s->uart_intr_state &= ~INTR_STATE_TX_WATERMARK;
+    }
+
+    /* Set TX empty */
+    if (s->tx_level == 0) {
+        s->uart_status |= UART_STATUS_TXEMPTY;
+        s->uart_intr_state |= INTR_STATE_TX_EMPTY;
+    }
+
+    ibex_uart_update_irqs(s);
+    return FALSE;
+}
+
+static void uart_write_tx_fifo(IbexUartState *s, const uint8_t *buf,
+                               int size)
+{
+    uint64_t current_time = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
+    uint8_t tx_fifo_level = (s->uart_fifo_ctrl & FIFO_CTRL_TXILVL)
+                            >> FIFO_CTRL_TXILVL_SHIFT;
+
+    if (size > IBEX_UART_TX_FIFO_SIZE - s->tx_level) {
+        size = IBEX_UART_TX_FIFO_SIZE - s->tx_level;
+        qemu_log_mask(LOG_GUEST_ERROR, "ibex_uart: TX FIFO overflow");
+    }
+
+    memcpy(s->tx_fifo + s->tx_level, buf, size);
+    s->tx_level += size;
+
+    if (s->tx_level > 0) {
+        s->uart_status &= ~UART_STATUS_TXEMPTY;
+    }
+
+    if (s->tx_level >= tx_fifo_level) {
+        s->uart_intr_state |= INTR_STATE_TX_WATERMARK;
+        ibex_uart_update_irqs(s);
+    }
+
+    if (s->tx_level == IBEX_UART_TX_FIFO_SIZE) {
+        s->uart_status |= UART_STATUS_TXFULL;
+    }
+
+    timer_mod(s->fifo_trigger_handle, current_time +
+              (s->char_tx_time * 4));
+}
+
+static void ibex_uart_reset(DeviceState *dev)
+{
+    IbexUartState *s = IBEX_UART(dev);
+
+    s->uart_intr_state = 0x00000000;
+    s->uart_intr_state = 0x00000000;
+    s->uart_intr_enable = 0x00000000;
+    s->uart_ctrl = 0x00000000;
+    s->uart_status = 0x0000003c;
+    s->uart_rdata = 0x00000000;
+    s->uart_fifo_ctrl = 0x00000000;
+    s->uart_fifo_status = 0x00000000;
+    s->uart_ovrd = 0x00000000;
+    s->uart_val = 0x00000000;
+    s->uart_timeout_ctrl = 0x00000000;
+
+    s->tx_level = 0;
+
+    s->char_tx_time = (NANOSECONDS_PER_SECOND / 230400) * 10;
+
+    ibex_uart_update_irqs(s);
+}
+
+static uint64_t ibex_uart_read(void *opaque, hwaddr addr,
+                                       unsigned int size)
+{
+    IbexUartState *s = opaque;
+    uint64_t retvalue = 0;
+
+    switch (addr) {
+    case IBEX_UART_INTR_STATE:
+        retvalue = s->uart_intr_state;
+        break;
+    case IBEX_UART_INTR_ENABLE:
+        retvalue = s->uart_intr_enable;
+        break;
+    case IBEX_UART_INTR_TEST:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: wdata is write only\n", __func__);
+        break;
+
+    case IBEX_UART_CTRL:
+        retvalue = s->uart_ctrl;
+        break;
+    case IBEX_UART_STATUS:
+        retvalue = s->uart_status;
+        break;
+
+    case IBEX_UART_RDATA:
+        retvalue = s->uart_rdata;
+        if (s->uart_ctrl & UART_CTRL_RX_ENABLE) {
+            qemu_chr_fe_accept_input(&s->chr);
+
+            s->uart_status |= UART_STATUS_RXIDLE;
+            s->uart_status |= UART_STATUS_RXEMPTY;
+        }
+        break;
+    case IBEX_UART_WDATA:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: wdata is write only\n", __func__);
+        break;
+
+    case IBEX_UART_FIFO_CTRL:
+        retvalue = s->uart_fifo_ctrl;
+        break;
+    case IBEX_UART_FIFO_STATUS:
+        retvalue = s->uart_fifo_status;
+
+        retvalue |= s->tx_level & 0x1F;
+
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: RX fifos are not supported\n", __func__);
+        break;
+
+    case IBEX_UART_OVRD:
+        retvalue = s->uart_ovrd;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: ovrd is not supported\n", __func__);
+        break;
+    case IBEX_UART_VAL:
+        retvalue = s->uart_val;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: val is not supported\n", __func__);
+        break;
+    case IBEX_UART_TIMEOUT_CTRL:
+        retvalue = s->uart_timeout_ctrl;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: timeout_ctrl is not supported\n", __func__);
+        break;
+    default:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Bad offset 0x%"HWADDR_PRIx"\n", __func__, addr);
+        return 0;
+    }
+
+    return retvalue;
+}
+
+static void ibex_uart_write(void *opaque, hwaddr addr,
+                                  uint64_t val64, unsigned int size)
+{
+    IbexUartState *s = opaque;
+    uint32_t value = val64;
+
+    switch (addr) {
+    case IBEX_UART_INTR_STATE:
+        /* Write 1 clear */
+        s->uart_intr_state &= ~value;
+        ibex_uart_update_irqs(s);
+        break;
+    case IBEX_UART_INTR_ENABLE:
+        s->uart_intr_enable = value;
+        ibex_uart_update_irqs(s);
+        break;
+    case IBEX_UART_INTR_TEST:
+        s->uart_intr_state |= value;
+        ibex_uart_update_irqs(s);
+        break;
+
+    case IBEX_UART_CTRL:
+        s->uart_ctrl = value;
+
+        if (value & UART_CTRL_NF) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_NF is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_SLPBK) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_SLPBK is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_LLPBK) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_LLPBK is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_PARITY_EN) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_PARITY_EN is not supported\n",
+                          __func__);
+        }
+        if (value & UART_CTRL_PARITY_ODD) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_PARITY_ODD is not supported\n",
+                          __func__);
+        }
+        if (value & UART_CTRL_RXBLVL) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_RXBLVL is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_NCO) {
+            uint64_t baud = ((value & UART_CTRL_NCO) >> 16);
+            baud *= 1000;
+            baud /= 2 ^ 20;
+
+            s->char_tx_time = (NANOSECONDS_PER_SECOND / baud) * 10;
+        }
+        break;
+    case IBEX_UART_STATUS:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: status is read only\n", __func__);
+        break;
+
+    case IBEX_UART_RDATA:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: rdata is read only\n", __func__);
+        break;
+    case IBEX_UART_WDATA:
+        uart_write_tx_fifo(s, (uint8_t *) &value, 1);
+        break;
+
+    case IBEX_UART_FIFO_CTRL:
+        s->uart_fifo_ctrl = value;
+
+        if (value & FIFO_CTRL_RXRST) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: RX fifos are not supported\n", __func__);
+        }
+        if (value & FIFO_CTRL_TXRST) {
+            s->tx_level = 0;
+        }
+        break;
+    case IBEX_UART_FIFO_STATUS:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: fifo_status is read only\n", __func__);
+        break;
+
+    case IBEX_UART_OVRD:
+        s->uart_ovrd = value;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: ovrd is not supported\n", __func__);
+        break;
+    case IBEX_UART_VAL:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: val is read only\n", __func__);
+        break;
+    case IBEX_UART_TIMEOUT_CTRL:
+        s->uart_timeout_ctrl = value;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: timeout_ctrl is not supported\n", __func__);
+        break;
+    default:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Bad offset 0x%"HWADDR_PRIx"\n", __func__, addr);
+    }
+}
+
+static void fifo_trigger_update(void *opaque)
+{
+    IbexUartState *s = opaque;
+
+    if (s->uart_ctrl & UART_CTRL_TX_ENABLE) {
+        ibex_uart_xmit(NULL, G_IO_OUT, s);
+    }
+}

hw/char/ibex_uart.c line 405

Quoted 10 lines
hw/char/ibex_uart.c new file +10 -0
+
+static const MemoryRegionOps ibex_uart_ops = {
+    .read = ibex_uart_read,
+    .write = ibex_uart_write,
+    .endianness = DEVICE_NATIVE_ENDIAN,
+    .impl.min_access_size = 4,
+    .impl.max_access_size = 4,
+};
+
+static int cadence_uart_post_load(void *opaque, int version_id)

Not a good name.

Quoted 135 lines
hw/char/ibex_uart.c new file +87 -0
+{
+    IbexUartState *s = opaque;
+
+    ibex_uart_update_irqs(s);
+    return 0;
+}
+
+static const VMStateDescription vmstate_ibex_uart = {
+    .name = TYPE_IBEX_UART,
+    .version_id = 1,
+    .minimum_version_id = 1,
+    .post_load = cadence_uart_post_load,
+    .fields = (VMStateField[]) {
+        VMSTATE_UINT8_ARRAY(tx_fifo, IbexUartState,
+                            IBEX_UART_TX_FIFO_SIZE),
+        VMSTATE_UINT32(tx_level, IbexUartState),
+        VMSTATE_UINT64(char_tx_time, IbexUartState),
+        VMSTATE_TIMER_PTR(fifo_trigger_handle, IbexUartState),
+        VMSTATE_UINT32(uart_intr_state, IbexUartState),
+        VMSTATE_UINT32(uart_intr_enable, IbexUartState),
+        VMSTATE_UINT32(uart_ctrl, IbexUartState),
+        VMSTATE_UINT32(uart_status, IbexUartState),
+        VMSTATE_UINT32(uart_rdata, IbexUartState),
+        VMSTATE_UINT32(uart_fifo_ctrl, IbexUartState),
+        VMSTATE_UINT32(uart_fifo_status, IbexUartState),
+        VMSTATE_UINT32(uart_ovrd, IbexUartState),
+        VMSTATE_UINT32(uart_val, IbexUartState),
+        VMSTATE_UINT32(uart_timeout_ctrl, IbexUartState),
+        VMSTATE_END_OF_LIST()
+    }
+};
+
+static Property ibex_uart_properties[] = {
+    DEFINE_PROP_CHR("chardev", IbexUartState, chr),
+    DEFINE_PROP_END_OF_LIST(),
+};
+
+static void ibex_uart_init(Object *obj)
+{
+    IbexUartState *s = IBEX_UART(obj);
+
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->tx_watermark);
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->rx_watermark);
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->tx_empty);
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->rx_overflow);
+
+    memory_region_init_io(&s->mmio, obj, &ibex_uart_ops, s,
+                          TYPE_IBEX_UART, 0x400);
+    sysbus_init_mmio(SYS_BUS_DEVICE(obj), &s->mmio);
+}
+
+static void ibex_uart_realize(DeviceState *dev, Error **errp)
+{
+    IbexUartState *s = IBEX_UART(dev);
+
+    s->fifo_trigger_handle = timer_new_ns(QEMU_CLOCK_VIRTUAL,
+                                          fifo_trigger_update, s);
+
+    qemu_chr_fe_set_handlers(&s->chr, ibex_uart_can_receive,
+                             ibex_uart_receive, NULL, NULL,
+                             s, NULL, true);
+}
+
+static void ibex_uart_class_init(ObjectClass *klass, void *data)
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+
+    dc->reset = ibex_uart_reset;
+    dc->realize = ibex_uart_realize;
+    dc->vmsd = &vmstate_ibex_uart;
+    device_class_set_props(dc, ibex_uart_properties);
+}
+
+static const TypeInfo ibex_uart_info = {
+    .name          = TYPE_IBEX_UART,
+    .parent        = TYPE_SYS_BUS_DEVICE,
+    .instance_size = sizeof(IbexUartState),
+    .instance_init = ibex_uart_init,
+    .class_init    = ibex_uart_class_init,
+};
+
+static void ibex_uart_register_types(void)
+{
+    type_register_static(&ibex_uart_info);
+}
+
+type_init(ibex_uart_register_types)
MAINTAINERS +2 -0
index 3e7d9cb0a5..a1ce186a7e 100644
@@ -1245,7 +1245,9 @@ M: Alistair Francis <[email protected]>
  L: [email protected]
  S: Supported
  F: hw/riscv/opentitan.c
+F: hw/char/ibex_uart.c
  F: include/hw/riscv/opentitan.h
+F: include/hw/char/ibex_uart.h
 
  SH4 Machines
  ------------
hw/char/Makefile.objs +1 -0
index 9e9a6c1aff..633996be5b 100644
@@ -12,6 +12,7 @@ common-obj-$(CONFIG_VIRTIO_SERIAL) += virtio-console.o
  common-obj-$(CONFIG_XILINX) += xilinx_uartlite.o
  common-obj-$(CONFIG_XEN) += xen_console.o
  common-obj-$(CONFIG_CADENCE) += cadence_uart.o
+common-obj-$(CONFIG_IBEX) += ibex_uart.o
 
  common-obj-$(CONFIG_EXYNOS4) += exynos4210_uart.o
  common-obj-$(CONFIG_COLDFIRE) += mcf_uart.o
hw/riscv/Kconfig +4 -0
index 94d19571f7..28947ef3e0 100644
@@ -4,6 +4,9 @@ config HTIF
  config HART
      bool
 
+config IBEX
+    bool
+
  config SIFIVE
      bool
      select MSI_NONBROKEN
@@ -29,6 +32,7 @@ config SPIKE
 
  config OPENTITAN
      bool
+    select IBEX
      select HART
      select UNIMP
 
#21 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

Hi Alistair,

There are still some questions I don't understand.

1. Is the baud rate  or fifo a necessary feature to simulate?
As you can see, qemu_chr_fe_write will send the byte as soon as possible. When you want to transmit a byte through WDATA,  you can call qemu_chr_fe_write directly.

2.  The baud rate calculation method is not strictly right.
I think when a byte write to FIFO,  char_tx_time * 8 is the correct time to send the byte instead of char_tx_time * 4.

3.  Why add a watch here?

Quoted 13 lines

+        s->uart_status |= UART_STATUS_TXEMPTY; +        s->uart_intr_state |= INTR_STATE_TX_EMPTY; +        s->uart_intr_state &= ~INTR_STATE_TX_WATERMARK; +        ibex_uart_update_irqs(s);
+        return FALSE;
+    }

hw/char/ibex_uart.c new file +7 -0
+
+    ret = qemu_chr_fe_write(&s->chr, s->tx_fifo, s->tx_level);
+
+    if (ret >= 0) {
+        s->tx_level -= ret;
+        memmove(s->tx_fifo, s->tx_fifo + ret, s->tx_level);
+    }

hw/char/ibex_uart.c line 130

Quoted 10 lines
hw/char/ibex_uart.c new file +10 -0
+
+    if (s->tx_level) {
+        guint r = qemu_chr_fe_add_watch(&s->chr, G_IO_OUT | G_IO_HUP,
+                                        ibex_uart_xmit, s);
+        if (!r) {
+            s->tx_level = 0;
+            return FALSE;
+        }
+    }
+ 

Zhiwei

#22 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

Quoted 9 lines · LIU Zhiwei

Hi Alistair,

There are still some questions I don't understand.

1. Is the baud rate or fifo a necessary feature to simulate? As you can see, qemu_chr_fe_write will send the byte as soon as possible. When you want to transmit a byte through WDATA, you can call qemu_chr_fe_write directly.

So qemu_chr_fe_write() will send the data straight away. This doesn't match what teh hardware does though. So by modelling a FIFO and a delay in sending we can better match the hardware.

Quoted 5 lines

2. The baud rate calculation method is not strictly right. I think when a byte write to FIFO, char_tx_time * 8 is the correct time to send the byte instead of
char_tx_time * 4.

Do you mind explaining why 8 is correct instead of 4?

Quoted 2 lines

3. Why add a watch here?

This is based on the Cadence UART implementation in QEMU (which does the same thing). This will trigger a callback when we can write more data or when the backend has hung up.

Alistair

Quoted 27 lines
Quoted 13 lines
hw/char/ibex_uart.c new file +13 -0
+        s->uart_status |= UART_STATUS_TXEMPTY;
+        s->uart_intr_state |= INTR_STATE_TX_EMPTY;
+        s->uart_intr_state &= ~INTR_STATE_TX_WATERMARK;
+        ibex_uart_update_irqs(s);
+        return FALSE;
+    }
+
+    ret = qemu_chr_fe_write(&s->chr, s->tx_fifo, s->tx_level);
+
+    if (ret >= 0) {
+        s->tx_level -= ret;
+        memmove(s->tx_fifo, s->tx_fifo + ret, s->tx_level);
+    }

hw/char/ibex_uart.c line 130

Quoted 10 lines
hw/char/ibex_uart.c new file +10 -0
+
+    if (s->tx_level) {
+        guint r = qemu_chr_fe_add_watch(&s->chr, G_IO_OUT | G_IO_HUP,
+                                        ibex_uart_xmit, s);
+        if (!r) {
+            s->tx_level = 0;
+            return FALSE;
+        }
+    }
+

Zhiwei

#23 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

Quoted 5 lines · Alistair Francis
Quoted 4 lines · LIU Zhiwei

Hi Alistair,

There are still some questions I don't understand.

Quoted 7 lines
Quoted 4 lines

1. Is the baud rate or fifo a necessary feature to simulate? As you can see, qemu_chr_fe_write will send the byte as soon as possible. When you want to transmit a byte through WDATA, you can call qemu_chr_fe_write directly.

So qemu_chr_fe_write() will send the data straight away. This doesn't match what teh hardware does though. So by modelling a FIFO and a delay in sending we can better match the hardware.

I see many UARTs have similar features. Does the software really care about these features? Usually I just want to print something to the terminal through UART. Most simulation in QEMU is for running software, not exactly the details of hardware. For example, we will not simulate the 16x oversamples in this UART.

There is no error here. Personally I  think it is necessary to simulate the FIFO and baud rate, maybe for supporting some backends.

Can someone give a reasonable answer for this question?

Quoted 5 lines
Quoted 4 lines

2. The baud rate calculation method is not strictly right. I think when a byte write to FIFO, char_tx_time * 8 is the correct time to send the byte instead of
char_tx_time * 4.

Do you mind explaining why 8 is correct instead of 4?

Usually write a byte to WDATA will trigger a uart_write_tx_fifo. Translate a bit will take char_tx_time. So it will take char_tx_time * 8 to transmit a byte.

Quoted 4 lines
Quoted 1 line

3. Why add a watch here?

This is based on the Cadence UART implementation in QEMU (which does the same thing). This will trigger a callback when we can write more data or when the backend has hung up.

Many other serials do the same thing, like virtio-console and serial. So it may be a common interface here. I will try to understand it(Not yet).

Zhiwei

Quoted 28 lines

Alistair

Quoted 26 lines
Quoted 13 lines
hw/char/ibex_uart.c new file +13 -0
+        s->uart_status |= UART_STATUS_TXEMPTY;
+        s->uart_intr_state |= INTR_STATE_TX_EMPTY;
+        s->uart_intr_state &= ~INTR_STATE_TX_WATERMARK;
+        ibex_uart_update_irqs(s);
+        return FALSE;
+    }
+
+    ret = qemu_chr_fe_write(&s->chr, s->tx_fifo, s->tx_level);
+
+    if (ret >= 0) {
+        s->tx_level -= ret;
+        memmove(s->tx_fifo, s->tx_fifo + ret, s->tx_level);
+    }

hw/char/ibex_uart.c line 130

Quoted 10 lines
hw/char/ibex_uart.c new file +10 -0
+
+    if (s->tx_level) {
+        guint r = qemu_chr_fe_add_watch(&s->chr, G_IO_OUT | G_IO_HUP,
+                                        ibex_uart_xmit, s);
+        if (!r) {
+            s->tx_level = 0;
+            return FALSE;
+        }
+    }
+

Zhiwei

#24 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

Quoted 11 lines · LIU Zhiwei
Quoted 9 lines · Alistair Francis
Quoted 8 lines · LIU Zhiwei

Hi Alistair,

There are still some questions I don't understand.

1. Is the baud rate or fifo a necessary feature to simulate? As you can see, qemu_chr_fe_write will send the byte as soon as possible. When you want to transmit a byte through WDATA, you can call qemu_chr_fe_write directly.

Quoted 6 lines
Quoted 3 lines

So qemu_chr_fe_write() will send the data straight away. This doesn't match what teh hardware does though. So by modelling a FIFO and a delay in sending we can better match the hardware.

I see many UARTs have similar features. Does the software really care about these features? Usually I just want to print something to the terminal through UART.

In this case Tock (which is the OS used for OpenTitan) does car about these features as it relies on interrupts generated by the HW to complete the serial send task. It also just makes the QEMU model more accurate.

Quoted 3 lines

Most simulation in QEMU is for running software, not exactly the details of hardware.
For example, we will not simulate the 16x oversamples in this UART.

Agreed. Lots of UARTs don't bother modelling the delay from the hardware as generally it doesn't matter. In this case it does make a difference for the software and it makes the QEMU model more accurate, which is always a good thing.

Quoted 4 lines

There is no error here. Personally I think it is necessary to simulate the FIFO and baud rate,
maybe for supporting some backends.

So baud rate doesn't need to be modelled as we aren't actually sending UART data, just pretending and then printing it.

Quoted 2 lines

Can someone give a reasonable answer for this question?

Which question?

Quoted 8 lines
Quoted 5 lines
Quoted 4 lines

2. The baud rate calculation method is not strictly right. I think when a byte write to FIFO, char_tx_time * 8 is the correct time to send the byte instead of
char_tx_time * 4.

Do you mind explaining why 8 is correct instead of 4?

Usually write a byte to WDATA will trigger a uart_write_tx_fifo. Translate a bit will take
char_tx_time. So it will take char_tx_time * 8 to transmit a byte.

I see your point. I just used the 4 as that is what the Cadence one does. I don't think it matters too much as it's just the delay for a timer (that isn't used as an accurate timer).

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

3. Why add a watch here?

This is based on the Cadence UART implementation in QEMU (which does the same thing). This will trigger a callback when we can write more data or when the backend has hung up.

Many other serials do the same thing, like virtio-console and serial. So it may be a common
interface here. I will try to understand it(Not yet).

Yep, it's just a more complete model of that the HW does.

Alistair

#25 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

Quoted 14 lines · Alistair Francis
Quoted 13 lines · LIU Zhiwei
Quoted 12 lines · Alistair Francis
Quoted 8 lines · LIU Zhiwei

Hi Alistair,

There are still some questions I don't understand.

1. Is the baud rate or fifo a necessary feature to simulate? As you can see, qemu_chr_fe_write will send the byte as soon as possible. When you want to transmit a byte through WDATA, you can call qemu_chr_fe_write directly.

So qemu_chr_fe_write() will send the data straight away. This doesn't match what teh hardware does though. So by modelling a FIFO and a delay in sending we can better match the hardware.

Quoted 7 lines
Quoted 3 lines

I see many UARTs have similar features. Does the software really care about these features? Usually I just want to print something to the terminal through UART.

In this case Tock (which is the OS used for OpenTitan) does car about these features as it relies on interrupts generated by the HW to complete the serial send task. It also just makes the QEMU model more accurate.

Fair enough. I see the "tx_watermark" interrupt, which needs the FIFO. At least, it can verify the ISP.

Quoted 8 lines
Quoted 3 lines

Most simulation in QEMU is for running software, not exactly the details of hardware.
For example, we will not simulate the 16x oversamples in this UART.

Agreed. Lots of UARTs don't bother modelling the delay from the hardware as generally it doesn't matter. In this case it does make a difference for the software and it makes the QEMU model more accurate, which is always a good thing.

Quoted 8 lines
Quoted 3 lines

There is no error here. Personally I think it is necessary to simulate the FIFO and baud rate,
maybe for supporting some backends.

So baud rate doesn't need to be modelled as we aren't actually sending UART data, just pretending and then printing it.

Quoted 1 line

Can someone give a reasonable answer for this question?

Which question?

I see  the UART can work with many  different backends,  such as pty , file, socket and so on. I wonder if this a backend, which has some requirements on the baud rate.  You can ignore it, as it doesn't matter.

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

2. The baud rate calculation method is not strictly right. I think when a byte write to FIFO, char_tx_time * 8 is the correct time to send the byte instead of
char_tx_time * 4.

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

Do you mind explaining why 8 is correct instead of 4?

Usually write a byte to WDATA will trigger a uart_write_tx_fifo. Translate a bit will take
char_tx_time. So it will take char_tx_time * 8 to transmit a byte.

I see your point. I just used the 4 as that is what the Cadence one does. I don't think it matters too much as it's just the delay for a timer (that isn't used as an accurate timer).

Got it. Just a way to send the bytes at sometime later.

Quoted 8 lines
Quoted 7 lines
Quoted 4 lines
Quoted 1 line

3. Why add a watch here?

This is based on the Cadence UART implementation in QEMU (which does the same thing). This will trigger a callback when we can write more data or when the backend has hung up.

Many other serials do the same thing, like virtio-console and serial. So it may be a common
interface here. I will try to understand it(Not yet).

Yep, it's just a more complete model of that the HW does.

I try to boot a RISC-V Linux, and set a breakpoint  to a watch callback function. The breakpoint did't match.

I just wonder if there is a case really need the callback function.

Zhiwei

Quoted 2 lines

Alistair

#26 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

Quoted 21 lines · LIU Zhiwei
Quoted 17 lines · Alistair Francis
Quoted 16 lines · LIU Zhiwei
Quoted 5 lines · Alistair Francis
Quoted 4 lines · LIU Zhiwei

Hi Alistair,

There are still some questions I don't understand.

Quoted 7 lines
Quoted 4 lines

1. Is the baud rate or fifo a necessary feature to simulate? As you can see, qemu_chr_fe_write will send the byte as soon as possible. When you want to transmit a byte through WDATA, you can call qemu_chr_fe_write directly.

So qemu_chr_fe_write() will send the data straight away. This doesn't match what teh hardware does though. So by modelling a FIFO and a delay in sending we can better match the hardware.

I see many UARTs have similar features. Does the software really care about these features? Usually I just want to print something to the terminal through UART.

Quoted 8 lines
Quoted 4 lines

In this case Tock (which is the OS used for OpenTitan) does car about these features as it relies on interrupts generated by the HW to complete the serial send task. It also just makes the QEMU model more accurate.

Fair enough. I see the "tx_watermark" interrupt, which needs the FIFO. At least,
it can verify the ISP.

Exactly :)

Quoted 11 lines
Quoted 11 lines
Quoted 3 lines

Most simulation in QEMU is for running software, not exactly the details of hardware.
For example, we will not simulate the 16x oversamples in this UART.

Agreed. Lots of UARTs don't bother modelling the delay from the hardware as generally it doesn't matter. In this case it does make a difference for the software and it makes the QEMU model more accurate, which is always a good thing.

Quoted 3 lines

There is no error here. Personally I think it is necessary to simulate the FIFO and baud rate,
maybe for supporting some backends.

Quoted 8 lines
Quoted 5 lines

So baud rate doesn't need to be modelled as we aren't actually sending UART data, just pretending and then printing it.

Quoted 1 line

Can someone give a reasonable answer for this question?

Which question?

I see the UART can work with many different backends, such as pty , file, socket and so on.
I wonder if this a backend, which has some requirements on the baud

The backend should be independent so it doesn't matter what baud rate we choose here.

Quoted 22 lines

rate. You can ignore it,
as it doesn't matter.

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

2. The baud rate calculation method is not strictly right. I think when a byte write to FIFO, char_tx_time * 8 is the correct time to send the byte instead of
char_tx_time * 4.

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

Do you mind explaining why 8 is correct instead of 4?

Usually write a byte to WDATA will trigger a uart_write_tx_fifo. Translate a bit will take
char_tx_time. So it will take char_tx_time * 8 to transmit a byte.

I see your point. I just used the 4 as that is what the Cadence one does. I don't think it matters too much as it's just the delay for a timer (that isn't used as an accurate timer).

Got it. Just a way to send the bytes at sometime later.

Quoted 7 lines
Quoted 7 lines
Quoted 4 lines
Quoted 1 line

3. Why add a watch here?

This is based on the Cadence UART implementation in QEMU (which does the same thing). This will trigger a callback when we can write more data or when the backend has hung up.

Many other serials do the same thing, like virtio-console and serial. So it may be a common
interface here. I will try to understand it(Not yet).

Quoted 6 lines
Quoted 1 line

Yep, it's just a more complete model of that the HW does.

I try to boot a RISC-V Linux, and set a breakpoint to a watch callback function.
The breakpoint did't match.

I just wonder if there is a case really need the callback function.

AFAIK Linux doesn't support the Ibex UART (or Ibex at all) so it shouldn't be triggered.

Alistair

Quoted 5 lines

Zhiwei

Quoted 2 lines

Alistair

Reviewed-by
#27 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

Quoted 73 lines · Alistair Francis
Quoted 20 lines · LIU Zhiwei
Quoted 17 lines · Alistair Francis
Quoted 16 lines · LIU Zhiwei
Quoted 5 lines · Alistair Francis
Quoted 4 lines · LIU Zhiwei

Hi Alistair,

There are still some questions I don't understand.

Quoted 7 lines
Quoted 4 lines

1. Is the baud rate or fifo a necessary feature to simulate? As you can see, qemu_chr_fe_write will send the byte as soon as possible. When you want to transmit a byte through WDATA, you can call qemu_chr_fe_write directly.

So qemu_chr_fe_write() will send the data straight away. This doesn't match what teh hardware does though. So by modelling a FIFO and a delay in sending we can better match the hardware.

I see many UARTs have similar features. Does the software really care about these features? Usually I just want to print something to the terminal through UART.

Quoted 7 lines
Quoted 4 lines

In this case Tock (which is the OS used for OpenTitan) does car about these features as it relies on interrupts generated by the HW to complete the serial send task. It also just makes the QEMU model more accurate.

Fair enough. I see the "tx_watermark" interrupt, which needs the FIFO. At least,
it can verify the ISP.

Exactly :)

Quoted 11 lines
Quoted 11 lines
Quoted 3 lines

Most simulation in QEMU is for running software, not exactly the details of hardware.
For example, we will not simulate the 16x oversamples in this UART.

Agreed. Lots of UARTs don't bother modelling the delay from the hardware as generally it doesn't matter. In this case it does make a difference for the software and it makes the QEMU model more accurate, which is always a good thing.

Quoted 3 lines

There is no error here. Personally I think it is necessary to simulate the FIFO and baud rate,
maybe for supporting some backends.

Quoted 8 lines
Quoted 5 lines

So baud rate doesn't need to be modelled as we aren't actually sending UART data, just pretending and then printing it.

Quoted 1 line

Can someone give a reasonable answer for this question?

Which question?

I see the UART can work with many different backends, such as pty , file, socket and so on.
I wonder if this a backend, which has some requirements on the baud

The backend should be independent so it doesn't matter what baud rate we choose here.

Quoted 21 lines

rate. You can ignore it,
as it doesn't matter.

Quoted 11 lines
Quoted 8 lines
Quoted 5 lines
Quoted 4 lines

2. The baud rate calculation method is not strictly right. I think when a byte write to FIFO, char_tx_time * 8 is the correct time to send the byte instead of
char_tx_time * 4.

Do you mind explaining why 8 is correct instead of 4?

Usually write a byte to WDATA will trigger a uart_write_tx_fifo. Translate a bit will take
char_tx_time. So it will take char_tx_time * 8 to transmit a byte.

I see your point. I just used the 4 as that is what the Cadence one does. I don't think it matters too much as it's just the delay for a timer (that isn't used as an accurate timer).

Got it. Just a way to send the bytes at sometime later.

Quoted 7 lines
Quoted 7 lines
Quoted 4 lines
Quoted 1 line

3. Why add a watch here?

This is based on the Cadence UART implementation in QEMU (which does the same thing). This will trigger a callback when we can write more data or when the backend has hung up.

Many other serials do the same thing, like virtio-console and serial. So it may be a common
interface here. I will try to understand it(Not yet).

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

Yep, it's just a more complete model of that the HW does.

I try to boot a RISC-V Linux, and set a breakpoint to a watch callback function.
The breakpoint did't match.

I just wonder if there is a case really need the callback function.

AFAIK Linux doesn't support the Ibex UART (or Ibex at all) so it shouldn't be triggered.

I tried with the UART in the virt board.  It also registers the watch callback on G_IO_HUP and G_IO_OUT.

However I will through the question alone in another mail.

After the two points addressed,

Zhiwei

Quoted 5 lines

Alistair

Quoted 2 lines

Zhiwei

Quoted 1 line

Alistair

#28 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

Quoted 82 lines · LIU Zhiwei
Quoted 78 lines · Alistair Francis
Quoted 20 lines · LIU Zhiwei
Quoted 17 lines · Alistair Francis
Quoted 16 lines · LIU Zhiwei
Quoted 5 lines · Alistair Francis
Quoted 4 lines · LIU Zhiwei

Hi Alistair,

There are still some questions I don't understand.

Quoted 7 lines
Quoted 4 lines

1. Is the baud rate or fifo a necessary feature to simulate? As you can see, qemu_chr_fe_write will send the byte as soon as possible. When you want to transmit a byte through WDATA, you can call qemu_chr_fe_write directly.

So qemu_chr_fe_write() will send the data straight away. This doesn't match what teh hardware does though. So by modelling a FIFO and a delay in sending we can better match the hardware.

I see many UARTs have similar features. Does the software really care about these features? Usually I just want to print something to the terminal through UART.

Quoted 7 lines
Quoted 4 lines

In this case Tock (which is the OS used for OpenTitan) does car about these features as it relies on interrupts generated by the HW to complete the serial send task. It also just makes the QEMU model more accurate.

Fair enough. I see the "tx_watermark" interrupt, which needs the FIFO. At least,
it can verify the ISP.

Exactly :)

Quoted 11 lines
Quoted 11 lines
Quoted 3 lines

Most simulation in QEMU is for running software, not exactly the details of hardware.
For example, we will not simulate the 16x oversamples in this UART.

Agreed. Lots of UARTs don't bother modelling the delay from the hardware as generally it doesn't matter. In this case it does make a difference for the software and it makes the QEMU model more accurate, which is always a good thing.

Quoted 3 lines

There is no error here. Personally I think it is necessary to simulate the FIFO and baud rate,
maybe for supporting some backends.

Quoted 8 lines
Quoted 5 lines

So baud rate doesn't need to be modelled as we aren't actually sending UART data, just pretending and then printing it.

Quoted 1 line

Can someone give a reasonable answer for this question?

Which question?

I see the UART can work with many different backends, such as pty , file, socket and so on.
I wonder if this a backend, which has some requirements on the baud

The backend should be independent so it doesn't matter what baud rate we choose here.

Quoted 26 lines

rate. You can ignore it,
as it doesn't matter.

Quoted 11 lines
Quoted 8 lines
Quoted 5 lines
Quoted 4 lines

2. The baud rate calculation method is not strictly right. I think when a byte write to FIFO, char_tx_time * 8 is the correct time to send the byte instead of
char_tx_time * 4.

Do you mind explaining why 8 is correct instead of 4?

Usually write a byte to WDATA will trigger a uart_write_tx_fifo. Translate a bit will take
char_tx_time. So it will take char_tx_time * 8 to transmit a byte.

I see your point. I just used the 4 as that is what the Cadence one does. I don't think it matters too much as it's just the delay for a timer (that isn't used as an accurate timer).

Got it. Just a way to send the bytes at sometime later.

Quoted 8 lines
Quoted 7 lines
Quoted 4 lines
Quoted 1 line

3. Why add a watch here?

This is based on the Cadence UART implementation in QEMU (which does the same thing). This will trigger a callback when we can write more data or when the backend has hung up.

Many other serials do the same thing, like virtio-console and serial. So it may be a common
interface here. I will try to understand it(Not yet).

Yep, it's just a more complete model of that the HW does.

I try to boot a RISC-V Linux, and set a breakpoint to a watch callback function.
The breakpoint did't match.

Quoted 8 lines
Quoted 3 lines
Quoted 1 line

I just wonder if there is a case really need the callback function.

AFAIK Linux doesn't support the Ibex UART (or Ibex at all) so it shouldn't be triggered.

I tried with the UART in the virt board. It also registers the watch callback on
G_IO_HUP and G_IO_OUT.

However I will through the question alone in another mail.

Ah, sorry I misunderstood what you meant.

I haven't looked at it, it's possible it's not enabled by Linux.

Quoted 4 lines

After the two points addressed,

Thanks!

Alistair

Quoted 10 lines

Zhiwei

Quoted 5 lines

Alistair

Quoted 2 lines

Zhiwei

Quoted 1 line

Alistair

Reviewed-by
#29 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

In reply to Alistair Francis
Quoted 73 lines · Alistair Francis
Quoted 20 lines · LIU Zhiwei
Quoted 17 lines · Alistair Francis
Quoted 16 lines · LIU Zhiwei
Quoted 5 lines · Alistair Francis
Quoted 4 lines · LIU Zhiwei

Hi Alistair,

There are still some questions I don't understand.

Quoted 7 lines
Quoted 4 lines

1. Is the baud rate or fifo a necessary feature to simulate? As you can see, qemu_chr_fe_write will send the byte as soon as possible. When you want to transmit a byte through WDATA, you can call qemu_chr_fe_write directly.

So qemu_chr_fe_write() will send the data straight away. This doesn't match what teh hardware does though. So by modelling a FIFO and a delay in sending we can better match the hardware.

I see many UARTs have similar features. Does the software really care about these features? Usually I just want to print something to the terminal through UART.

Quoted 7 lines
Quoted 4 lines

In this case Tock (which is the OS used for OpenTitan) does car about these features as it relies on interrupts generated by the HW to complete the serial send task. It also just makes the QEMU model more accurate.

Fair enough. I see the "tx_watermark" interrupt, which needs the FIFO. At least,
it can verify the ISP.

Exactly :)

Quoted 11 lines
Quoted 11 lines
Quoted 3 lines

Most simulation in QEMU is for running software, not exactly the details of hardware.
For example, we will not simulate the 16x oversamples in this UART.

Agreed. Lots of UARTs don't bother modelling the delay from the hardware as generally it doesn't matter. In this case it does make a difference for the software and it makes the QEMU model more accurate, which is always a good thing.

Quoted 3 lines

There is no error here. Personally I think it is necessary to simulate the FIFO and baud rate,
maybe for supporting some backends.

Quoted 8 lines
Quoted 5 lines

So baud rate doesn't need to be modelled as we aren't actually sending UART data, just pretending and then printing it.

Quoted 1 line

Can someone give a reasonable answer for this question?

Which question?

I see the UART can work with many different backends, such as pty , file, socket and so on.
I wonder if this a backend, which has some requirements on the baud

The backend should be independent so it doesn't matter what baud rate we choose here.

Quoted 21 lines

rate. You can ignore it,
as it doesn't matter.

Quoted 11 lines
Quoted 8 lines
Quoted 5 lines
Quoted 4 lines

2. The baud rate calculation method is not strictly right. I think when a byte write to FIFO, char_tx_time * 8 is the correct time to send the byte instead of
char_tx_time * 4.

Do you mind explaining why 8 is correct instead of 4?

Usually write a byte to WDATA will trigger a uart_write_tx_fifo. Translate a bit will take
char_tx_time. So it will take char_tx_time * 8 to transmit a byte.

I see your point. I just used the 4 as that is what the Cadence one does. I don't think it matters too much as it's just the delay for a timer (that isn't used as an accurate timer).

Got it. Just a way to send the bytes at sometime later.

Quoted 7 lines
Quoted 7 lines
Quoted 4 lines
Quoted 1 line

3. Why add a watch here?

This is based on the Cadence UART implementation in QEMU (which does the same thing). This will trigger a callback when we can write more data or when the backend has hung up.

Many other serials do the same thing, like virtio-console and serial. So it may be a common
interface here. I will try to understand it(Not yet).

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

Yep, it's just a more complete model of that the HW does.

I try to boot a RISC-V Linux, and set a breakpoint to a watch callback function.
The breakpoint did't match.

I just wonder if there is a case really need the callback function.

AFAIK Linux doesn't support the Ibex UART (or Ibex at all) so it shouldn't be triggered.

I tried with the UART in the virt board.  It also registers the watch callback on G_IO_HUP and G_IO_OUT.

However I will throw the question alone in another thread.

After the two comments pointed at another email addressed,

Zhiwei

Quoted 5 lines

Alistair

Quoted 2 lines

Zhiwei

Quoted 1 line

Alistair

#30 lore raw

Re: [PATCH v5 07/11] hw/char: Initial commit of Ibex UART

In reply to LIU Zhiwei
Quoted 243 lines · LIU Zhiwei
Quoted 239 lines · Alistair Francis

This is the initial commit of the Ibex UART device. Serial TX is working, while RX has been implemeneted but untested.

This is based on the documentation from:
https://docs.opentitan.org/hw/ip/uart/doc/

---
  include/hw/char/ibex_uart.h | 110 ++++++++
  hw/char/ibex_uart.c         | 492 ++++++++++++++++++++++++++++++++++++
  MAINTAINERS                 |   2 +
  hw/char/Makefile.objs       |   1 +
  hw/riscv/Kconfig            |   4 +
  5 files changed, 609 insertions(+)

create mode 100644 include/hw/char/ibex_uart.h
create mode 100644 hw/char/ibex_uart.c

include/hw/char/ibex_uart.h new file +110 -0
new file mode 100644
index 0000000000..2bec772615
@@ -0,0 +1,110 @@
+/*
+ * QEMU lowRISC Ibex UART device
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef HW_IBEX_UART_H
+#define HW_IBEX_UART_H
+
+#include "hw/sysbus.h"
+#include "chardev/char-fe.h"
+#include "qemu/timer.h"
+
+#define IBEX_UART_INTR_STATE   0x00
+    #define INTR_STATE_TX_WATERMARK (1 << 0)
+    #define INTR_STATE_RX_WATERMARK (1 << 1)
+    #define INTR_STATE_TX_EMPTY     (1 << 2)
+    #define INTR_STATE_RX_OVERFLOW  (1 << 3)
+#define IBEX_UART_INTR_ENABLE  0x04
+#define IBEX_UART_INTR_TEST    0x08
+
+#define IBEX_UART_CTRL         0x0c
+    #define UART_CTRL_TX_ENABLE     (1 << 0)
+    #define UART_CTRL_RX_ENABLE     (1 << 1)
+    #define UART_CTRL_NF            (1 << 2)
+    #define UART_CTRL_SLPBK         (1 << 4)
+    #define UART_CTRL_LLPBK         (1 << 5)
+    #define UART_CTRL_PARITY_EN     (1 << 6)
+    #define UART_CTRL_PARITY_ODD    (1 << 7)
+    #define UART_CTRL_RXBLVL        (3 << 8)
+    #define UART_CTRL_NCO           (0xFFFF << 16)
+
+#define IBEX_UART_STATUS       0x10
+    #define UART_STATUS_TXFULL  (1 << 0)
+    #define UART_STATUS_RXFULL  (1 << 1)
+    #define UART_STATUS_TXEMPTY (1 << 2)
+    #define UART_STATUS_RXIDLE  (1 << 4)
+    #define UART_STATUS_RXEMPTY (1 << 5)
+
+#define IBEX_UART_RDATA        0x14
+#define IBEX_UART_WDATA        0x18
+
+#define IBEX_UART_FIFO_CTRL    0x1c
+    #define FIFO_CTRL_RXRST          (1 << 0)
+    #define FIFO_CTRL_TXRST          (1 << 1)
+    #define FIFO_CTRL_RXILVL         (7 << 2)
+    #define FIFO_CTRL_RXILVL_SHIFT   (2)
+    #define FIFO_CTRL_TXILVL         (3 << 5)
+    #define FIFO_CTRL_TXILVL_SHIFT   (5)
+
+#define IBEX_UART_FIFO_STATUS  0x20
+#define IBEX_UART_OVRD         0x24
+#define IBEX_UART_VAL          0x28
+#define IBEX_UART_TIMEOUT_CTRL 0x2c
+
+#define IBEX_UART_TX_FIFO_SIZE 16
+
+#define TYPE_IBEX_UART "ibex-uart"
+#define IBEX_UART(obj) \
+    OBJECT_CHECK(IbexUartState, (obj), TYPE_IBEX_UART)
+
+typedef struct {
+    /* <private> */
+    SysBusDevice parent_obj;
+
+    /* <public> */
+    MemoryRegion mmio;
+
+    uint8_t tx_fifo[IBEX_UART_TX_FIFO_SIZE];
+    uint32_t tx_level;
+
+    QEMUTimer *fifo_trigger_handle;
+    uint64_t char_tx_time;
+
+    uint32_t uart_intr_state;
+    uint32_t uart_intr_enable;
+    uint32_t uart_ctrl;
+    uint32_t uart_status;
+    uint32_t uart_rdata;
+    uint32_t uart_fifo_ctrl;
+    uint32_t uart_fifo_status;
+    uint32_t uart_ovrd;
+    uint32_t uart_val;
+    uint32_t uart_timeout_ctrl;
+
+    CharBackend chr;
+    qemu_irq tx_watermark;
+    qemu_irq rx_watermark;
+    qemu_irq tx_empty;
+    qemu_irq rx_overflow;
+} IbexUartState;
+#endif /* HW_IBEX_UART_H */
hw/char/ibex_uart.c new file +100 -0
new file mode 100644
index 0000000000..c416325d73
@@ -0,0 +1,492 @@
+/*
+ * QEMU lowRISC Ibex UART device
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * For details check the documentation here:
+ *    https://docs.opentitan.org/hw/ip/uart/doc/
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/char/ibex_uart.h"
+#include "hw/irq.h"
+#include "hw/qdev-properties.h"
+#include "migration/vmstate.h"
+#include "qemu/log.h"
+#include "qemu/module.h"
+
+static void ibex_uart_update_irqs(IbexUartState *s)
+{
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_TX_WATERMARK) {
+        qemu_set_irq(s->tx_watermark, 1);
+    } else {
+        qemu_set_irq(s->tx_watermark, 0);
+    }
+
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_RX_WATERMARK) {
+        qemu_set_irq(s->rx_watermark, 1);
+    } else {
+        qemu_set_irq(s->rx_watermark, 0);
+    }
+
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_TX_EMPTY) {
+        qemu_set_irq(s->tx_empty, 1);
+    } else {
+        qemu_set_irq(s->tx_empty, 0);
+    }
+
+    if (s->uart_intr_state & s->uart_intr_enable & INTR_STATE_RX_OVERFLOW) {
+        qemu_set_irq(s->rx_overflow, 1);
+    } else {
+        qemu_set_irq(s->rx_overflow, 0);
+    }
+}
+
+static int ibex_uart_can_receive(void *opaque)
+{
+    IbexUartState *s = opaque;
+
+    if (s->uart_ctrl & UART_CTRL_RX_ENABLE) {
+        return 1;
+    }
+
+    return 0;
+}
+
+static void ibex_uart_receive(void *opaque, const uint8_t *buf, int size)
+{
+    IbexUartState *s = opaque;
+    uint8_t rx_fifo_level = (s->uart_fifo_ctrl & FIFO_CTRL_RXILVL)
+                            >> FIFO_CTRL_RXILVL_SHIFT;
+
+    s->uart_rdata = *buf;
+
+    s->uart_status &= ~UART_STATUS_RXIDLE;
+    s->uart_status &= ~UART_STATUS_RXEMPTY;
+
+    if (size > rx_fifo_level) {
+        s->uart_intr_state |= INTR_STATE_RX_WATERMARK;
+    }
+
+    ibex_uart_update_irqs(s);
+}
+
+static gboolean ibex_uart_xmit(GIOChannel *chan, GIOCondition cond,
+                               void *opaque)
+{
+    IbexUartState *s = opaque;
+    uint8_t tx_fifo_level = (s->uart_fifo_ctrl & FIFO_CTRL_TXILVL)
+                            >> FIFO_CTRL_TXILVL_SHIFT;
+    int ret;
+
+    /* instant drain the fifo when there's no back-end */
Quoted 8 lines

hw/char/ibex_uart.c line 107

Quoted 7 lines
hw/char/ibex_uart.c new file +7 -0
+    if (!qemu_chr_fe_backend_connected(&s->chr)) {
+        s->tx_level = 0;
+        return FALSE;
+    }
+
+    if (!s->tx_level) {
+        s->uart_status &= UART_STATUS_TXFULL;

Maybe s->uart_status &= ~UART_STATUS_TXFULL;

Good spot, fixed.

Quoted 291 lines
Quoted 291 lines
hw/char/ibex_uart.c new file +291 -0
+        s->uart_status |= UART_STATUS_TXEMPTY;
+        s->uart_intr_state |= INTR_STATE_TX_EMPTY;
+        s->uart_intr_state &= ~INTR_STATE_TX_WATERMARK;
+        ibex_uart_update_irqs(s);
+        return FALSE;
+    }
+
+    ret = qemu_chr_fe_write(&s->chr, s->tx_fifo, s->tx_level);
+
+    if (ret >= 0) {
+        s->tx_level -= ret;
+        memmove(s->tx_fifo, s->tx_fifo + ret, s->tx_level);
+    }
+
+    if (s->tx_level) {
+        guint r = qemu_chr_fe_add_watch(&s->chr, G_IO_OUT | G_IO_HUP,
+                                        ibex_uart_xmit, s);
+        if (!r) {
+            s->tx_level = 0;
+            return FALSE;
+        }
+    }
+
+    /* Clear the TX Full bit */
+    if (s->tx_level != IBEX_UART_TX_FIFO_SIZE) {
+        s->uart_status &= ~UART_STATUS_TXFULL;
+    }
+
+    /* Disable the TX_WATERMARK IRQ */
+    if (s->tx_level < tx_fifo_level) {
+        s->uart_intr_state &= ~INTR_STATE_TX_WATERMARK;
+    }
+
+    /* Set TX empty */
+    if (s->tx_level == 0) {
+        s->uart_status |= UART_STATUS_TXEMPTY;
+        s->uart_intr_state |= INTR_STATE_TX_EMPTY;
+    }
+
+    ibex_uart_update_irqs(s);
+    return FALSE;
+}
+
+static void uart_write_tx_fifo(IbexUartState *s, const uint8_t *buf,
+                               int size)
+{
+    uint64_t current_time = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
+    uint8_t tx_fifo_level = (s->uart_fifo_ctrl & FIFO_CTRL_TXILVL)
+                            >> FIFO_CTRL_TXILVL_SHIFT;
+
+    if (size > IBEX_UART_TX_FIFO_SIZE - s->tx_level) {
+        size = IBEX_UART_TX_FIFO_SIZE - s->tx_level;
+        qemu_log_mask(LOG_GUEST_ERROR, "ibex_uart: TX FIFO overflow");
+    }
+
+    memcpy(s->tx_fifo + s->tx_level, buf, size);
+    s->tx_level += size;
+
+    if (s->tx_level > 0) {
+        s->uart_status &= ~UART_STATUS_TXEMPTY;
+    }
+
+    if (s->tx_level >= tx_fifo_level) {
+        s->uart_intr_state |= INTR_STATE_TX_WATERMARK;
+        ibex_uart_update_irqs(s);
+    }
+
+    if (s->tx_level == IBEX_UART_TX_FIFO_SIZE) {
+        s->uart_status |= UART_STATUS_TXFULL;
+    }
+
+    timer_mod(s->fifo_trigger_handle, current_time +
+              (s->char_tx_time * 4));
+}
+
+static void ibex_uart_reset(DeviceState *dev)
+{
+    IbexUartState *s = IBEX_UART(dev);
+
+    s->uart_intr_state = 0x00000000;
+    s->uart_intr_state = 0x00000000;
+    s->uart_intr_enable = 0x00000000;
+    s->uart_ctrl = 0x00000000;
+    s->uart_status = 0x0000003c;
+    s->uart_rdata = 0x00000000;
+    s->uart_fifo_ctrl = 0x00000000;
+    s->uart_fifo_status = 0x00000000;
+    s->uart_ovrd = 0x00000000;
+    s->uart_val = 0x00000000;
+    s->uart_timeout_ctrl = 0x00000000;
+
+    s->tx_level = 0;
+
+    s->char_tx_time = (NANOSECONDS_PER_SECOND / 230400) * 10;
+
+    ibex_uart_update_irqs(s);
+}
+
+static uint64_t ibex_uart_read(void *opaque, hwaddr addr,
+                                       unsigned int size)
+{
+    IbexUartState *s = opaque;
+    uint64_t retvalue = 0;
+
+    switch (addr) {
+    case IBEX_UART_INTR_STATE:
+        retvalue = s->uart_intr_state;
+        break;
+    case IBEX_UART_INTR_ENABLE:
+        retvalue = s->uart_intr_enable;
+        break;
+    case IBEX_UART_INTR_TEST:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: wdata is write only\n", __func__);
+        break;
+
+    case IBEX_UART_CTRL:
+        retvalue = s->uart_ctrl;
+        break;
+    case IBEX_UART_STATUS:
+        retvalue = s->uart_status;
+        break;
+
+    case IBEX_UART_RDATA:
+        retvalue = s->uart_rdata;
+        if (s->uart_ctrl & UART_CTRL_RX_ENABLE) {
+            qemu_chr_fe_accept_input(&s->chr);
+
+            s->uart_status |= UART_STATUS_RXIDLE;
+            s->uart_status |= UART_STATUS_RXEMPTY;
+        }
+        break;
+    case IBEX_UART_WDATA:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: wdata is write only\n", __func__);
+        break;
+
+    case IBEX_UART_FIFO_CTRL:
+        retvalue = s->uart_fifo_ctrl;
+        break;
+    case IBEX_UART_FIFO_STATUS:
+        retvalue = s->uart_fifo_status;
+
+        retvalue |= s->tx_level & 0x1F;
+
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: RX fifos are not supported\n", __func__);
+        break;
+
+    case IBEX_UART_OVRD:
+        retvalue = s->uart_ovrd;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: ovrd is not supported\n", __func__);
+        break;
+    case IBEX_UART_VAL:
+        retvalue = s->uart_val;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: val is not supported\n", __func__);
+        break;
+    case IBEX_UART_TIMEOUT_CTRL:
+        retvalue = s->uart_timeout_ctrl;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: timeout_ctrl is not supported\n", __func__);
+        break;
+    default:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Bad offset 0x%"HWADDR_PRIx"\n", __func__, addr);
+        return 0;
+    }
+
+    return retvalue;
+}
+
+static void ibex_uart_write(void *opaque, hwaddr addr,
+                                  uint64_t val64, unsigned int size)
+{
+    IbexUartState *s = opaque;
+    uint32_t value = val64;
+
+    switch (addr) {
+    case IBEX_UART_INTR_STATE:
+        /* Write 1 clear */
+        s->uart_intr_state &= ~value;
+        ibex_uart_update_irqs(s);
+        break;
+    case IBEX_UART_INTR_ENABLE:
+        s->uart_intr_enable = value;
+        ibex_uart_update_irqs(s);
+        break;
+    case IBEX_UART_INTR_TEST:
+        s->uart_intr_state |= value;
+        ibex_uart_update_irqs(s);
+        break;
+
+    case IBEX_UART_CTRL:
+        s->uart_ctrl = value;
+
+        if (value & UART_CTRL_NF) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_NF is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_SLPBK) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_SLPBK is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_LLPBK) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_LLPBK is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_PARITY_EN) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_PARITY_EN is not supported\n",
+                          __func__);
+        }
+        if (value & UART_CTRL_PARITY_ODD) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_PARITY_ODD is not supported\n",
+                          __func__);
+        }
+        if (value & UART_CTRL_RXBLVL) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: UART_CTRL_RXBLVL is not supported\n", __func__);
+        }
+        if (value & UART_CTRL_NCO) {
+            uint64_t baud = ((value & UART_CTRL_NCO) >> 16);
+            baud *= 1000;
+            baud /= 2 ^ 20;
+
+            s->char_tx_time = (NANOSECONDS_PER_SECOND / baud) * 10;
+        }
+        break;
+    case IBEX_UART_STATUS:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: status is read only\n", __func__);
+        break;
+
+    case IBEX_UART_RDATA:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: rdata is read only\n", __func__);
+        break;
+    case IBEX_UART_WDATA:
+        uart_write_tx_fifo(s, (uint8_t *) &value, 1);
+        break;
+
+    case IBEX_UART_FIFO_CTRL:
+        s->uart_fifo_ctrl = value;
+
+        if (value & FIFO_CTRL_RXRST) {
+            qemu_log_mask(LOG_UNIMP,
+                          "%s: RX fifos are not supported\n", __func__);
+        }
+        if (value & FIFO_CTRL_TXRST) {
+            s->tx_level = 0;
+        }
+        break;
+    case IBEX_UART_FIFO_STATUS:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: fifo_status is read only\n", __func__);
+        break;
+
+    case IBEX_UART_OVRD:
+        s->uart_ovrd = value;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: ovrd is not supported\n", __func__);
+        break;
+    case IBEX_UART_VAL:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: val is read only\n", __func__);
+        break;
+    case IBEX_UART_TIMEOUT_CTRL:
+        s->uart_timeout_ctrl = value;
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: timeout_ctrl is not supported\n", __func__);
+        break;
+    default:
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: Bad offset 0x%"HWADDR_PRIx"\n", __func__, addr);
+    }
+}
+
+static void fifo_trigger_update(void *opaque)
+{
+    IbexUartState *s = opaque;
+
+    if (s->uart_ctrl & UART_CTRL_TX_ENABLE) {
+        ibex_uart_xmit(NULL, G_IO_OUT, s);
+    }
+}
+
+static const MemoryRegionOps ibex_uart_ops = {
+    .read = ibex_uart_read,
Quoted 8 lines

hw/char/ibex_uart.c line 405

Quoted 7 lines
hw/char/ibex_uart.c new file +7 -0
+    .write = ibex_uart_write,
+    .endianness = DEVICE_NATIVE_ENDIAN,
+    .impl.min_access_size = 4,
+    .impl.max_access_size = 4,
+};
+
+static int cadence_uart_post_load(void *opaque, int version_id)

Not a good name.

Also fixed.

Alistair

Quoted 137 lines
Quoted 135 lines
hw/char/ibex_uart.c new file +87 -0
+{
+    IbexUartState *s = opaque;
+
+    ibex_uart_update_irqs(s);
+    return 0;
+}
+
+static const VMStateDescription vmstate_ibex_uart = {
+    .name = TYPE_IBEX_UART,
+    .version_id = 1,
+    .minimum_version_id = 1,
+    .post_load = cadence_uart_post_load,
+    .fields = (VMStateField[]) {
+        VMSTATE_UINT8_ARRAY(tx_fifo, IbexUartState,
+                            IBEX_UART_TX_FIFO_SIZE),
+        VMSTATE_UINT32(tx_level, IbexUartState),
+        VMSTATE_UINT64(char_tx_time, IbexUartState),
+        VMSTATE_TIMER_PTR(fifo_trigger_handle, IbexUartState),
+        VMSTATE_UINT32(uart_intr_state, IbexUartState),
+        VMSTATE_UINT32(uart_intr_enable, IbexUartState),
+        VMSTATE_UINT32(uart_ctrl, IbexUartState),
+        VMSTATE_UINT32(uart_status, IbexUartState),
+        VMSTATE_UINT32(uart_rdata, IbexUartState),
+        VMSTATE_UINT32(uart_fifo_ctrl, IbexUartState),
+        VMSTATE_UINT32(uart_fifo_status, IbexUartState),
+        VMSTATE_UINT32(uart_ovrd, IbexUartState),
+        VMSTATE_UINT32(uart_val, IbexUartState),
+        VMSTATE_UINT32(uart_timeout_ctrl, IbexUartState),
+        VMSTATE_END_OF_LIST()
+    }
+};
+
+static Property ibex_uart_properties[] = {
+    DEFINE_PROP_CHR("chardev", IbexUartState, chr),
+    DEFINE_PROP_END_OF_LIST(),
+};
+
+static void ibex_uart_init(Object *obj)
+{
+    IbexUartState *s = IBEX_UART(obj);
+
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->tx_watermark);
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->rx_watermark);
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->tx_empty);
+    sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->rx_overflow);
+
+    memory_region_init_io(&s->mmio, obj, &ibex_uart_ops, s,
+                          TYPE_IBEX_UART, 0x400);
+    sysbus_init_mmio(SYS_BUS_DEVICE(obj), &s->mmio);
+}
+
+static void ibex_uart_realize(DeviceState *dev, Error **errp)
+{
+    IbexUartState *s = IBEX_UART(dev);
+
+    s->fifo_trigger_handle = timer_new_ns(QEMU_CLOCK_VIRTUAL,
+                                          fifo_trigger_update, s);
+
+    qemu_chr_fe_set_handlers(&s->chr, ibex_uart_can_receive,
+                             ibex_uart_receive, NULL, NULL,
+                             s, NULL, true);
+}
+
+static void ibex_uart_class_init(ObjectClass *klass, void *data)
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+
+    dc->reset = ibex_uart_reset;
+    dc->realize = ibex_uart_realize;
+    dc->vmsd = &vmstate_ibex_uart;
+    device_class_set_props(dc, ibex_uart_properties);
+}
+
+static const TypeInfo ibex_uart_info = {
+    .name          = TYPE_IBEX_UART,
+    .parent        = TYPE_SYS_BUS_DEVICE,
+    .instance_size = sizeof(IbexUartState),
+    .instance_init = ibex_uart_init,
+    .class_init    = ibex_uart_class_init,
+};
+
+static void ibex_uart_register_types(void)
+{
+    type_register_static(&ibex_uart_info);
+}
+
+type_init(ibex_uart_register_types)
MAINTAINERS +2 -0
index 3e7d9cb0a5..a1ce186a7e 100644
@@ -1245,7 +1245,9 @@ M: Alistair Francis <[email protected]>
  L: [email protected]
  S: Supported
  F: hw/riscv/opentitan.c
+F: hw/char/ibex_uart.c
  F: include/hw/riscv/opentitan.h
+F: include/hw/char/ibex_uart.h
 
  SH4 Machines
  ------------
hw/char/Makefile.objs +1 -0
index 9e9a6c1aff..633996be5b 100644
@@ -12,6 +12,7 @@ common-obj-$(CONFIG_VIRTIO_SERIAL) += virtio-console.o
  common-obj-$(CONFIG_XILINX) += xilinx_uartlite.o
  common-obj-$(CONFIG_XEN) += xen_console.o
  common-obj-$(CONFIG_CADENCE) += cadence_uart.o
+common-obj-$(CONFIG_IBEX) += ibex_uart.o
 
  common-obj-$(CONFIG_EXYNOS4) += exynos4210_uart.o
  common-obj-$(CONFIG_COLDFIRE) += mcf_uart.o
hw/riscv/Kconfig +4 -0
index 94d19571f7..28947ef3e0 100644
@@ -4,6 +4,9 @@ config HTIF
  config HART
      bool
 
+config IBEX
+    bool
+
  config SIFIVE
      bool
      select MSI_NONBROKEN
@@ -29,6 +32,7 @@ config SPIKE
 
  config OPENTITAN
      bool
+    select IBEX
      select HART
      select UNIMP
 
#31 lore raw

PATCH v5 8/11 hw/intc: Initial commit of lowRISC Ibex PLIC

In reply to Alistair Francis

The Ibex core contains a PLIC that although similar to the RISC-V spec is not RISC-V spec compliant.

This patch implements a Ibex PLIC in a somewhat generic way.

As the current RISC-V PLIC needs tidying up, my hope is that as the Ibex PLIC move towards spec compliance this PLIC implementation can be updated until it can replace the current PLIC.

---
 include/hw/intc/ibex_plic.h |  63 +++++++++
 hw/intc/ibex_plic.c         | 261 ++++++++++++++++++++++++++++++++++++
 MAINTAINERS                 |   2 +
 hw/intc/Makefile.objs       |   1 +
 4 files changed, 327 insertions(+)
 create mode 100644 include/hw/intc/ibex_plic.h
 create mode 100644 hw/intc/ibex_plic.c
include/hw/intc/ibex_plic.h new file +63 -0
new file mode 100644
index 0000000000..ddc7909903
@@ -0,0 +1,63 @@
+/*
+ * QEMU RISC-V lowRISC Ibex PLIC
+ *
+ * Copyright (c) 2020 Western Digital
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef HW_IBEX_PLIC_H
+#define HW_IBEX_PLIC_H
+
+#include "hw/sysbus.h"
+
+#define TYPE_IBEX_PLIC "ibex-plic"
+#define IBEX_PLIC(obj) \
+    OBJECT_CHECK(IbexPlicState, (obj), TYPE_IBEX_PLIC)
+
+typedef struct IbexPlicState {
+    /*< private >*/
+    SysBusDevice parent_obj;
+
+    /*< public >*/
+    MemoryRegion mmio;
+
+    uint32_t *pending;
+    uint32_t *source;
+    uint32_t *priority;
+    uint32_t *enable;
+    uint32_t threshold;
+    uint32_t claim;
+
+    /* config */
+    uint32_t num_cpus;
+    uint32_t num_sources;
+
+    uint32_t pending_base;
+    uint32_t pending_num;
+
+    uint32_t source_base;
+    uint32_t source_num;
+
+    uint32_t priority_base;
+    uint32_t priority_num;
+
+    uint32_t enable_base;
+    uint32_t enable_num;
+
+    uint32_t threshold_base;
+
+    uint32_t claim_base;
+} IbexPlicState;
+
+#endif /* HW_IBEX_PLIC_H */
hw/intc/ibex_plic.c new file +261 -0

261 lines not on this page · raw message

MAINTAINERS +2 -0
index a1ce186a7e..d6b2819e44 100644
@@ -1246,8 +1246,10 @@ L: [email protected]
 S: Supported
 F: hw/riscv/opentitan.c
 F: hw/char/ibex_uart.c
+F: hw/intc/ibex_plic.c
 F: include/hw/riscv/opentitan.h
 F: include/hw/char/ibex_uart.h
+F: include/hw/intc/ibex_plic.h
 
 SH4 Machines
 ------------
hw/intc/Makefile.objs +1 -0
index f726d87532..a61e6728fe 100644
@@ -49,3 +49,4 @@ obj-$(CONFIG_ARM_GIC) += arm_gicv3_cpuif.o
 obj-$(CONFIG_MIPS_CPS) += mips_gic.o
 obj-$(CONFIG_NIOS2) += nios2_iic.o
 obj-$(CONFIG_OMPIC) += ompic.o
+obj-$(CONFIG_IBEX) += ibex_plic.o
Signature · 1 line
2.26.2
#32 lore raw

PATCH v5 9/11 riscv/opentitan: Connect the PLIC device

In reply to Alistair Francis
---
 include/hw/riscv/opentitan.h |  3 +++
 hw/riscv/opentitan.c         | 19 +++++++++++++++++--
 2 files changed, 20 insertions(+), 2 deletions(-)
include/hw/riscv/opentitan.h +3 -0
index a4b6499444..76f72905a8 100644
@@ -20,6 +20,7 @@
 #define HW_OPENTITAN_H
 
 #include "hw/riscv/riscv_hart.h"
+#include "hw/intc/ibex_plic.h"
 
 #define TYPE_RISCV_IBEX_SOC "riscv.lowrisc.ibex.soc"
 #define RISCV_IBEX_SOC(obj) \
@@ -31,6 +32,8 @@ typedef struct LowRISCIbexSoCState {
 
     /*< public >*/
     RISCVHartArrayState cpus;
+    IbexPlicState plic;
+
     MemoryRegion flash_mem;
     MemoryRegion rom;
 } LowRISCIbexSoCState;
hw/riscv/opentitan.c +17 -2
index b4fb836466..46a3a93c5e 100644
@@ -25,6 +25,7 @@
 #include "hw/misc/unimp.h"
 #include "hw/riscv/boot.h"
 #include "exec/address-spaces.h"
+#include "sysemu/sysemu.h"
 
 static const struct MemmapEntry {
     hwaddr base;
@@ -97,6 +98,9 @@ static void riscv_lowrisc_ibex_soc_init(Object *obj)
     object_initialize_child(obj, "cpus", &s->cpus,
                             sizeof(s->cpus), TYPE_RISCV_HART_ARRAY,
                             &error_abort, NULL);
+
+    sysbus_init_child_obj(obj, "plic", &s->plic,
+                          sizeof(s->plic), TYPE_IBEX_PLIC);
 }
 
 static void riscv_lowrisc_ibex_soc_realize(DeviceState *dev_soc, Error **errp)
@@ -105,6 +109,9 @@ static void riscv_lowrisc_ibex_soc_realize(DeviceState *dev_soc, Error **errp)
     MachineState *ms = MACHINE(qdev_get_machine());
     LowRISCIbexSoCState *s = RISCV_IBEX_SOC(dev_soc);
     MemoryRegion *sys_mem = get_system_memory();
+    DeviceState *dev;
+    SysBusDevice *busdev;
+    Error *err = NULL;
 
     object_property_set_str(OBJECT(&s->cpus), ms->cpu_type, "cpu-type",
                             &error_abort);
@@ -125,6 +132,16 @@ static void riscv_lowrisc_ibex_soc_realize(DeviceState *dev_soc, Error **errp)
     memory_region_add_subregion(sys_mem, memmap[IBEX_FLASH].base,
                                 &s->flash_mem);
 
+    /* PLIC */
+    dev = DEVICE(&s->plic);
+    object_property_set_bool(OBJECT(&s->plic), true, "realized", &err);
+    if (err != NULL) {
+        error_propagate(errp, err);
+        return;
+    }
+    busdev = SYS_BUS_DEVICE(dev);
+    sysbus_mmio_map(busdev, 0, memmap[IBEX_PLIC].base);
+
     create_unimplemented_device("riscv.lowrisc.ibex.uart",
         memmap[IBEX_UART].base, memmap[IBEX_UART].size);
     create_unimplemented_device("riscv.lowrisc.ibex.gpio",
@@ -145,8 +162,6 @@ static void riscv_lowrisc_ibex_soc_realize(DeviceState *dev_soc, Error **errp)
         memmap[IBEX_AES].base, memmap[IBEX_AES].size);
     create_unimplemented_device("riscv.lowrisc.ibex.hmac",
         memmap[IBEX_HMAC].base, memmap[IBEX_HMAC].size);
-    create_unimplemented_device("riscv.lowrisc.ibex.plic",
-        memmap[IBEX_PLIC].base, memmap[IBEX_PLIC].size);
     create_unimplemented_device("riscv.lowrisc.ibex.pinmux",
         memmap[IBEX_PINMUX].base, memmap[IBEX_PINMUX].size);
     create_unimplemented_device("riscv.lowrisc.ibex.alert_handler",
Signature · 1 line
2.26.2
#33 lore raw

PATCH v5 10/11 riscv/opentitan: Connect the UART device

In reply to Alistair Francis
---
 include/hw/riscv/opentitan.h | 13 +++++++++++++
 hw/riscv/opentitan.c         | 24 ++++++++++++++++++++++--
 2 files changed, 35 insertions(+), 2 deletions(-)
include/hw/riscv/opentitan.h +13 -0
index 76f72905a8..8f29b9cbbf 100644
@@ -21,6 +21,7 @@
 
 #include "hw/riscv/riscv_hart.h"
 #include "hw/intc/ibex_plic.h"
+#include "hw/char/ibex_uart.h"
 
 #define TYPE_RISCV_IBEX_SOC "riscv.lowrisc.ibex.soc"
 #define RISCV_IBEX_SOC(obj) \
@@ -33,6 +34,7 @@ typedef struct LowRISCIbexSoCState {
     /*< public >*/
     RISCVHartArrayState cpus;
     IbexPlicState plic;
+    IbexUartState uart;
 
     MemoryRegion flash_mem;
     MemoryRegion rom;
@@ -68,4 +70,15 @@ enum {
     IBEX_PADCTRL,
 };
 
+enum {
+    IBEX_UART_RX_PARITY_ERR_IRQ = 0x28,
+    IBEX_UART_RX_TIMEOUT_IRQ = 0x27,
+    IBEX_UART_RX_BREAK_ERR_IRQ = 0x26,
+    IBEX_UART_RX_FRAME_ERR_IRQ = 0x25,
+    IBEX_UART_RX_OVERFLOW_IRQ = 0x24,
+    IBEX_UART_TX_EMPTY_IRQ = 0x23,
+    IBEX_UART_RX_WATERMARK_IRQ = 0x22,
+    IBEX_UART_TX_WATERMARK_IRQ = 0x21,
+};
+
 #endif
hw/riscv/opentitan.c +22 -2
index 46a3a93c5e..a8844a870b 100644
@@ -101,6 +101,9 @@ static void riscv_lowrisc_ibex_soc_init(Object *obj)
 
     sysbus_init_child_obj(obj, "plic", &s->plic,
                           sizeof(s->plic), TYPE_IBEX_PLIC);
+
+    sysbus_init_child_obj(obj, "uart", &s->uart,
+                          sizeof(s->uart), TYPE_IBEX_UART);
 }
 
 static void riscv_lowrisc_ibex_soc_realize(DeviceState *dev_soc, Error **errp)
@@ -142,8 +145,25 @@ static void riscv_lowrisc_ibex_soc_realize(DeviceState *dev_soc, Error **errp)
     busdev = SYS_BUS_DEVICE(dev);
     sysbus_mmio_map(busdev, 0, memmap[IBEX_PLIC].base);
 
-    create_unimplemented_device("riscv.lowrisc.ibex.uart",
-        memmap[IBEX_UART].base, memmap[IBEX_UART].size);
+    /* UART */
+    dev = DEVICE(&(s->uart));
+    qdev_prop_set_chr(dev, "chardev", serial_hd(0));
+    object_property_set_bool(OBJECT(&s->uart), true, "realized", &err);
+    if (err != NULL) {
+        error_propagate(errp, err);
+        return;
+    }
+    busdev = SYS_BUS_DEVICE(dev);
+    sysbus_mmio_map(busdev, 0, memmap[IBEX_UART].base);
+    sysbus_connect_irq(busdev, 0, qdev_get_gpio_in(DEVICE(&s->plic),
+                       IBEX_UART_TX_WATERMARK_IRQ));
+    sysbus_connect_irq(busdev, 1, qdev_get_gpio_in(DEVICE(&s->plic),
+                       IBEX_UART_RX_WATERMARK_IRQ));
+    sysbus_connect_irq(busdev, 2, qdev_get_gpio_in(DEVICE(&s->plic),
+                       IBEX_UART_TX_EMPTY_IRQ));
+    sysbus_connect_irq(busdev, 3, qdev_get_gpio_in(DEVICE(&s->plic),
+                       IBEX_UART_RX_OVERFLOW_IRQ));
+
     create_unimplemented_device("riscv.lowrisc.ibex.gpio",
         memmap[IBEX_GPIO].base, memmap[IBEX_GPIO].size);
     create_unimplemented_device("riscv.lowrisc.ibex.spi",
Signature · 1 line
2.26.2
#34 lore raw

PATCH v5 11/11 target/riscv: Use a smaller guess size for no-MMU PMP

In reply to Alistair Francis
---
 target/riscv/pmp.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)
target/riscv/pmp.c +9 -5
index 0e6b640fbd..9418660f1b 100644
@@ -233,12 +233,16 @@ bool pmp_hart_has_privs(CPURISCVState *env, target_ulong addr,
         return true;
     }
 
-    /*
-     * if size is unknown (0), assume that all bytes
-     * from addr to the end of the page will be accessed.
-     */
     if (size == 0) {
-        pmp_size = -(addr | TARGET_PAGE_MASK);
+        if (riscv_feature(env, RISCV_FEATURE_MMU)) {
+            /*
+             * If size is unknown (0), assume that all bytes
+             * from addr to the end of the page will be accessed.
+             */
+            pmp_size = -(addr | TARGET_PAGE_MASK);
+        } else {
+            pmp_size = sizeof(target_ulong);
+        }
     } else {
         pmp_size = size;
     }
Signature · 1 line
2.26.2