hw/riscv: Add Microchip PolarFire SoC Icicle Kit board support
Subsystems (2)
cover hw/riscv: Add Microchip PolarFire SoC Icicle Kit board support
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
For more details about Microchip PolarFire Soc, please see:
https://www.microsemi.com/product-directory/soc-fpgas/5498-polarfire-soc-fpga
The Icicle Kit board information can be found here:
https://www.microsemi.com/existing-parts/parts/152514
Unlike SiFive FU540, the RISC-V core resect vector is at 0x20220000. The RISC-V CPU and HART codes has been updated to set the core's reset vector based on a configurable property from machine codes.
The following perepherals are created as an unimplemented device:
- Bus Error Uint 0/1/2/3/4
- L2 cache controller
- SYSREG
- MPUCFG
- IOSCBCFG
- GPIO
The following perepherals are emulated:
- SiFive CLINT
- SiFive PLIC
- PolarFire SoC Multi-Mode UART
- PolarFire SoC DMA
- Cadence eMMC/SDHCI controller
- Cadence Gigabit Ethernet MAC
Some bugs in the SD card codes are fixed during the development.
The BIOS image used by this machine is hss.bin, aka Hart Software Services, which can be built from:
https://github.com/polarfire-soc/hart-software-services
To launch this machine:
$ qemu-system-riscv64 -M microchip-icicle-kit -smp 5 \
-bios path/to/hss.bin -sd path/to/sdcard.img \
-nic tap,ifname=tap,script=no,model=cadence_gem \
-display none -serial stdio \
-chardev socket,id=serial1,path=serial1.sock,server,wait \
-serial chardev:serial1The memory is set to 1 GiB by default to match the hardware. A sanity check on ram size is performed in the machine init routine to prompt user to increase the RAM size to > 1 GiB when less than 1 GiB ram is detected.
HSS output is on the first serial port (stdio) and U-Boot/Linux outputs on the 2nd serial port. OpenSBI outputs on a random serial port due to the lottery mechanism used during the multi-core boot.
Bin Meng (18):
target/riscv: cpu: Add a new 'resetvec' property
hw/riscv: hart: Add a new 'resetvec' property target/riscv: cpu: Set reset vector based on the configured property
value
hw/riscv: Initial support for Microchip PolarFire SoC Icicle Kit board hw/char: Add Microchip PolarFire SoC MMUART emulation hw/riscv: microchip_pfsoc: Connect 5 MMUARTs
hw/sd: sd: Fix incorrect populated function switch status data
structure
hw/sd: sd: Correctly set the high capacity bit
hw/sd: sdhci: Make sdhci_poweron_reset() internal visible hw/sd: Add Cadence SDHCI emulation
hw/riscv: microchip_pfsoc: Connect a Cadence SDHCI controller and an
SD card
hw/dma: Add Microchip PolarFire Soc DMA controller emulation hw/riscv: microchip_pfsoc: Connect a DMA controller hw/net: cadence_gem: Add a new 'phy-addr' property hw/riscv: microchip_pfsoc: Connect 2 Cadence GEMs hw/riscv: microchip_pfsoc: Hook GPIO controllers
hw/riscv: clint: Avoid using hard-coded timebase frequency hw/riscv: microchip_pfsoc: Document the software used for testing
MAINTAINERS | 11 + default-configs/riscv64-softmmu.mak | 1 + hw/char/Kconfig | 3 + hw/char/Makefile.objs | 1 + hw/char/mchp_pfsoc_mmuart.c | 82 +++++++ hw/dma/Kconfig | 3 + hw/dma/Makefile.objs | 1 + hw/dma/mchp_pfsoc_dma.c | 322 +++++++++++++++++++++++++ hw/net/cadence_gem.c | 7 +- hw/riscv/Kconfig | 9 + hw/riscv/Makefile.objs | 1 + hw/riscv/microchip_pfsoc.c | 456 ++++++++++++++++++++++++++++++++++++ hw/riscv/opentitan.c | 1 + hw/riscv/riscv_hart.c | 3 + hw/riscv/sifive_clint.c | 25 +- hw/riscv/sifive_e.c | 4 +- hw/riscv/sifive_u.c | 5 +- hw/riscv/spike.c | 2 +- hw/riscv/virt.c | 3 +- hw/sd/Kconfig | 4 + hw/sd/Makefile.objs | 1 + hw/sd/cadence_sdhci.c | 162 +++++++++++++ hw/sd/sd.c | 8 +- hw/sd/sdhci-internal.h | 1 + hw/sd/sdhci.c | 2 +- include/hw/char/mchp_pfsoc_mmuart.h | 61 +++++ include/hw/dma/mchp_pfsoc_dma.h | 57 +++++ include/hw/net/cadence_gem.h | 2 + include/hw/riscv/microchip_pfsoc.h | 125 ++++++++++ include/hw/riscv/riscv_hart.h | 1 + include/hw/riscv/sifive_clint.h | 3 +- include/hw/sd/cadence_sdhci.h | 65 +++++ target/riscv/cpu.c | 8 +- target/riscv/cpu.h | 7 +- target/riscv/cpu_helper.c | 4 +- target/riscv/csr.c | 4 +- 36 files changed, 1424 insertions(+), 31 deletions(-) create mode 100644 hw/char/mchp_pfsoc_mmuart.c create mode 100644 hw/dma/mchp_pfsoc_dma.c create mode 100644 hw/riscv/microchip_pfsoc.c create mode 100644 hw/sd/cadence_sdhci.c create mode 100644 include/hw/char/mchp_pfsoc_mmuart.h create mode 100644 include/hw/dma/mchp_pfsoc_dma.h create mode 100644 include/hw/riscv/microchip_pfsoc.h create mode 100644 include/hw/sd/cadence_sdhci.h
Signature · 1 line
2.7.4
PATCH 1/18 target/riscv: cpu: Add a new 'resetvec' property
From: Bin Meng <[email protected]>
Currently the reset vector address is hard-coded in a RISC-V CPU's instance_init() routine. In a real world we can have 2 exact same CPUs except for the reset vector address, which is pretty common in the RISC-V core IP licensing business.
Normally reset vector address is a configurable parameter. Let's create a 64-bit property to store the reset vector address which covers both 32-bit and 64-bit CPUs.
---
target/riscv/cpu.c | 1 + target/riscv/cpu.h | 1 + 2 files changed, 2 insertions(+)
target/riscv/cpu.c +1 -0
@@ -518,6 +518,7 @@ static Property riscv_cpu_properties[] = {DEFINE_PROP_UINT16 ("elen" ,RISCVCPU , cfg.elen,64 ),DEFINE_PROP_BOOL ("mmu" ,RISCVCPU , cfg.mmu, true),DEFINE_PROP_BOOL ("pmp" ,RISCVCPU , cfg.pmp, true), +DEFINE_PROP_UINT64 ("resetvec" ,RISCVCPU , cfg.resetvec,DEFAULT_RSTVEC ), DEFINE_PROP_END_OF_LIST(), };
target/riscv/cpu.h +1 -0
@@ -291,6 +291,7 @@ typedef struct RISCVCPU {uint16_t elen;bool mmu;bool pmp; +uint64_t resetvec; } cfg; }RISCVCPU ;
Signature · 1 line
2.7.4
Re: [PATCH 01/18] target/riscv: cpu: Add a new 'resetvec' property
Quoted 8 lines · Bin Meng
From: Bin Meng <[email protected]>
Currently the reset vector address is hard-coded in a RISC-V CPU's
instance_init()routine. In a real world we can have 2 exact same CPUs except for the reset vector address, which is pretty common in the RISC-V core IP licensing business.
Quoted 5 lines
Normally reset vector address is a configurable parameter. Let's create a 64-bit property to store the reset vector address which covers both 32-bit and 64-bit CPUs.
Signed-off-by Bin Meng <[email protected]>
Alistair
Quoted 34 lines
---
target/riscv/cpu.c | 1 + target/riscv/cpu.h | 1 + 2 files changed, 2 insertions(+)target/riscv/cpu.c +1
-0@@ -518,6 +518,7 @@ static Property riscv_cpu_properties[] = {DEFINE_PROP_UINT16 ("elen" ,RISCVCPU , cfg.elen,64 ),DEFINE_PROP_BOOL ("mmu" ,RISCVCPU , cfg.mmu, true),DEFINE_PROP_BOOL ("pmp" ,RISCVCPU , cfg.pmp, true), +DEFINE_PROP_UINT64 ("resetvec" ,RISCVCPU , cfg.resetvec,DEFAULT_RSTVEC ), DEFINE_PROP_END_OF_LIST(), };target/riscv/cpu.h +1
-1@@ -291,6 +291,7 @@ typedef struct RISCVCPU {uint16_t elen;bool mmu;bool pmp; +uint64_t resetvec; } cfg; }RISCVCPU ; --2.7.4
PATCH 2/18 hw/riscv: hart: Add a new 'resetvec' property
In reply to Bin MengFrom: Bin Meng <[email protected]>
RISC-V machines do not instantiate RISC-V CPUs directly, instead they do that via the hart array. Add a new property for the reset vector address to allow the value to be passed to the CPU, before CPU is realized.
---
hw/riscv/riscv_hart.c | 3 +++ include/hw/riscv/riscv_hart.h | 1 + 2 files changed, 4 insertions(+)
hw/riscv/riscv_hart.c +3 -0
@@ -31,6 +31,8 @@ static Property riscv_harts_props[] = {DEFINE_PROP_UINT32 ("num-harts" , RISCVHartArrayState, num_harts,1 ),DEFINE_PROP_UINT32 ("hartid-base" , RISCVHartArrayState, hartid_base,0 ),DEFINE_PROP_STRING ("cpu-type" , RISCVHartArrayState, cpu_type), +DEFINE_PROP_UINT64 ("resetvec" , RISCVHartArrayState, resetvec, +DEFAULT_RSTVEC ), DEFINE_PROP_END_OF_LIST(), };
@@ -44,6 +46,7 @@ static bool riscv_hart_realize(RISCVHartArrayState *s, int idx, char *cpu_type, Error **errp) { object_initialize_child(OBJECT(s), "harts[*]", &s->harts[idx], cpu_type); +qdev_prop_set_uint64 (DEVICE (&s->harts[idx]),"resetvec" , s->resetvec); s->harts[idx].env.mhartid = s->hartid_base + idx;qemu_register_reset (riscv_harts_cpu_reset, &s->harts[idx]);return qdev_realize (DEVICE (&s->harts[idx]),NULL , errp);
include/hw/riscv/riscv_hart.h +1 -0
@@ -37,6 +37,7 @@ typedef struct RISCVHartArrayState {uint32_t num_harts;uint32_t hartid_base;char *cpu_type; +uint64_t resetvec;RISCVCPU *harts; } RISCVHartArrayState;
Signature · 1 line
2.7.4
Re: [PATCH 02/18] hw/riscv: hart: Add a new 'resetvec' property
Quoted 9 lines · Bin Meng
From: Bin Meng <[email protected]>
RISC-V machines do not instantiate RISC-V CPUs directly, instead they do that via the hart array. Add a new property for the reset vector address to allow the value to be passed to the CPU, before CPU is realized.
Signed-off-by Bin Meng <[email protected]>
Alistair
Quoted 43 lines
---
hw/riscv/riscv_hart.c | 3 +++ include/hw/riscv/riscv_hart.h | 1 + 2 files changed, 4 insertions(+)hw/riscv/riscv_hart.c +3
-0@@ -31,6 +31,8 @@ static Property riscv_harts_props[] = {DEFINE_PROP_UINT32 ("num-harts" , RISCVHartArrayState, num_harts,1 ),DEFINE_PROP_UINT32 ("hartid-base" , RISCVHartArrayState, hartid_base,0 ),DEFINE_PROP_STRING ("cpu-type" , RISCVHartArrayState, cpu_type), +DEFINE_PROP_UINT64 ("resetvec" , RISCVHartArrayState, resetvec, +DEFAULT_RSTVEC ), DEFINE_PROP_END_OF_LIST(), };@@ -44,6 +46,7 @@ static bool riscv_hart_realize(RISCVHartArrayState *s, int idx, char *cpu_type, Error **errp) { object_initialize_child(OBJECT(s), "harts[*]", &s->harts[idx], cpu_type); +qdev_prop_set_uint64 (DEVICE (&s->harts[idx]),"resetvec" , s->resetvec); s->harts[idx].env.mhartid = s->hartid_base + idx;qemu_register_reset (riscv_harts_cpu_reset, &s->harts[idx]);return qdev_realize (DEVICE (&s->harts[idx]),NULL , errp);include/hw/riscv/riscv_hart.h +1
-1@@ -37,6 +37,7 @@ typedef struct RISCVHartArrayState {uint32_t num_harts;uint32_t hartid_base;char *cpu_type; +uint64_t resetvec;RISCVCPU *harts; } RISCVHartArrayState; --2.7.4
PATCH 3/18 target/riscv: cpu: Set reset vector based on the configured property value
In reply to Bin MengFrom: Bin Meng <[email protected]>
Now that we have the newly introduced 'resetvec' property in the RISC-V CPU and HART, instead of hard-coding the reset vector addr in the CPU's instance_init(), move that to riscv_cpu_realize() based on the configured property value from the RISC-V machines.
---
hw/riscv/opentitan.c | 1 + hw/riscv/sifive_e.c | 1 + hw/riscv/sifive_u.c | 2 ++ target/riscv/cpu.c | 7 ++----- 4 files changed, 6 insertions(+), 5 deletions(-)
hw/riscv/opentitan.c +1 -0
@@ -111,6 +111,7 @@ static void lowrisc_ibex_soc_realize(DeviceState *dev_soc, Error **errp) &error_abort);object_property_set_int (OBJECT (&s->cpus),"num-harts" , ms->smp.cpus, &error_abort); +object_property_set_int (OBJECT (&s->cpus),"resetvec" ,0x8090 , &error_abort);sysbus_realize (SYS_BUS_DEVICE (&s->cpus), &error_abort);/* Boot ROM */
hw/riscv/sifive_e.c +1 -0
@@ -177,6 +177,7 @@ static void sifive_e_soc_init(Object *obj)object_initialize_child (obj,"cpus" , &s->cpus,TYPE_RISCV_HART_ARRAY );object_property_set_int (OBJECT (&s->cpus),"num-harts" , ms->smp.cpus, &error_abort); +object_property_set_int (OBJECT (&s->cpus),"resetvec" ,0x1004 , &error_abort);object_initialize_child (obj,"riscv.sifive.e.gpio0" , &s->gpio,TYPE_SIFIVE_GPIO ); }
hw/riscv/sifive_u.c +2 -0
@@ -611,6 +611,7 @@ static void sifive_u_soc_instance_init(Object *obj)qdev_prop_set_uint32 (DEVICE (&s->e_cpus),"num-harts" ,1 );qdev_prop_set_uint32 (DEVICE (&s->e_cpus),"hartid-base" ,0 );qdev_prop_set_string (DEVICE (&s->e_cpus),"cpu-type" ,SIFIVE_E_CPU ); +qdev_prop_set_uint64 (DEVICE (&s->e_cpus),"resetvec" ,0x1004 );object_initialize_child (obj,"u-cluster" , &s->u_cluster,TYPE_CPU_CLUSTER );qdev_prop_set_uint32 (DEVICE (&s->u_cluster),"cluster-id" ,1 );
@@ -620,6 +621,7 @@ static void sifive_u_soc_instance_init(Object *obj)qdev_prop_set_uint32 (DEVICE (&s->u_cpus),"num-harts" , ms->smp.cpus -1 );qdev_prop_set_uint32 (DEVICE (&s->u_cpus),"hartid-base" ,1 );qdev_prop_set_string (DEVICE (&s->u_cpus),"cpu-type" ,SIFIVE_U_CPU ); +qdev_prop_set_uint64 (DEVICE (&s->u_cpus),"resetvec" ,0x1004 );object_initialize_child (obj,"prci" , &s->prci,TYPE_SIFIVE_U_PRCI );object_initialize_child (obj,"otp" , &s->otp,TYPE_SIFIVE_U_OTP );
target/riscv/cpu.c +2 -5
@@ -128,7 +128,6 @@ static void riscv_any_cpu_init(Object *obj)CPURISCVState *env = &RISCV_CPU (obj)->env;set_misa (env,RVXLEN |RVI |RVM |RVA |RVF |RVD |RVC |RVU );set_priv_version (env,PRIV_VERSION_1_11_0 ); -set_resetvec (env,DEFAULT_RSTVEC ); }static void riscv_base_cpu_init (Object *obj)
@@ -136,7 +135,6 @@ static void riscv_base_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 rvxx_sifive_u_cpu_init (Object *obj)
@@ -144,7 +142,6 @@ static void rvxx_sifive_u_cpu_init(Object *obj)CPURISCVState *env = &RISCV_CPU (obj)->env;set_misa (env,RVXLEN |RVI |RVM |RVA |RVF |RVD |RVC |RVS |RVU );set_priv_version (env,PRIV_VERSION_1_10_0 ); -set_resetvec (env,0x1004 ); }static void rvxx_sifive_e_cpu_init (Object *obj)
@@ -152,7 +149,6 @@ static void rvxx_sifive_e_cpu_init(Object *obj)CPURISCVState *env = &RISCV_CPU (obj)->env;set_misa (env,RVXLEN |RVI |RVM |RVA |RVC |RVU );set_priv_version (env,PRIV_VERSION_1_10_0 ); -set_resetvec (env,0x1004 );qdev_prop_set_bit (DEVICE (obj),"mmu" , false); }
@@ -163,7 +159,6 @@ static void rv32_ibex_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); }
@@ -373,6 +368,8 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)set_feature (env,RISCV_FEATURE_PMP ); } +set_resetvec (env, cpu->cfg.resetvec); +/* If misa isn't set (rv32 and rv64 machines) set it here */ if (!env->misa) {/* Do some ISA extension error checking */
Signature · 1 line
2.7.4
Re: [PATCH 03/18] target/riscv: cpu: Set reset vector based on the configured property value
Quoted 9 lines · Bin Meng
From: Bin Meng <[email protected]>
Now that we have the newly introduced 'resetvec' property in the RISC-V CPU and HART, instead of hard-coding the reset vector addr in the CPU's
instance_init(), move that toriscv_cpu_realize()based on the configured property value from the RISC-V machines.Signed-off-by Bin Meng <[email protected]>
Alistair
Quoted 109 lines
---
hw/riscv/opentitan.c | 1 + hw/riscv/sifive_e.c | 1 + hw/riscv/sifive_u.c | 2 ++ target/riscv/cpu.c | 7 ++----- 4 files changed, 6 insertions(+), 5 deletions(-)hw/riscv/opentitan.c +1
-0@@ -111,6 +111,7 @@ static void lowrisc_ibex_soc_realize(DeviceState *dev_soc, Error **errp) &error_abort);object_property_set_int (OBJECT (&s->cpus),"num-harts" , ms->smp.cpus, &error_abort); +object_property_set_int (OBJECT (&s->cpus),"resetvec" ,0x8090 , &error_abort);sysbus_realize (SYS_BUS_DEVICE (&s->cpus), &error_abort);/* Boot ROM */ hw/riscv/sifive_e.c +1
-0@@ -177,6 +177,7 @@ static void sifive_e_soc_init(Object *obj)object_initialize_child (obj,"cpus" , &s->cpus,TYPE_RISCV_HART_ARRAY );object_property_set_int (OBJECT (&s->cpus),"num-harts" , ms->smp.cpus, &error_abort); +object_property_set_int (OBJECT (&s->cpus),"resetvec" ,0x1004 , &error_abort);object_initialize_child (obj,"riscv.sifive.e.gpio0" , &s->gpio,TYPE_SIFIVE_GPIO ); }hw/riscv/sifive_u.c +2
-0@@ -611,6 +611,7 @@ static void sifive_u_soc_instance_init(Object *obj)qdev_prop_set_uint32 (DEVICE (&s->e_cpus),"num-harts" ,1 );qdev_prop_set_uint32 (DEVICE (&s->e_cpus),"hartid-base" ,0 );qdev_prop_set_string (DEVICE (&s->e_cpus),"cpu-type" ,SIFIVE_E_CPU ); +qdev_prop_set_uint64 (DEVICE (&s->e_cpus),"resetvec" ,0x1004 );object_initialize_child (obj,"u-cluster" , &s->u_cluster,TYPE_CPU_CLUSTER );qdev_prop_set_uint32 (DEVICE (&s->u_cluster),"cluster-id" ,1 );@@ -620,6 +621,7 @@ static void sifive_u_soc_instance_init(Object *obj)qdev_prop_set_uint32 (DEVICE (&s->u_cpus),"num-harts" , ms->smp.cpus -1 );qdev_prop_set_uint32 (DEVICE (&s->u_cpus),"hartid-base" ,1 );qdev_prop_set_string (DEVICE (&s->u_cpus),"cpu-type" ,SIFIVE_U_CPU ); +qdev_prop_set_uint64 (DEVICE (&s->u_cpus),"resetvec" ,0x1004 );object_initialize_child (obj,"prci" , &s->prci,TYPE_SIFIVE_U_PRCI );object_initialize_child (obj,"otp" , &s->otp,TYPE_SIFIVE_U_OTP );target/riscv/cpu.c +2
-6@@ -128,7 +128,6 @@ static void riscv_any_cpu_init(Object *obj)CPURISCVState *env = &RISCV_CPU (obj)->env;set_misa (env,RVXLEN |RVI |RVM |RVA |RVF |RVD |RVC |RVU );set_priv_version (env,PRIV_VERSION_1_11_0 ); -set_resetvec (env,DEFAULT_RSTVEC ); }static void riscv_base_cpu_init (Object *obj)@@ -136,7 +135,6 @@ static void riscv_base_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 rvxx_sifive_u_cpu_init (Object *obj)@@ -144,7 +142,6 @@ static void rvxx_sifive_u_cpu_init(Object *obj)CPURISCVState *env = &RISCV_CPU (obj)->env;set_misa (env,RVXLEN |RVI |RVM |RVA |RVF |RVD |RVC |RVS |RVU );set_priv_version (env,PRIV_VERSION_1_10_0 ); -set_resetvec (env,0x1004 ); }static void rvxx_sifive_e_cpu_init (Object *obj)@@ -152,7 +149,6 @@ static void rvxx_sifive_e_cpu_init(Object *obj)CPURISCVState *env = &RISCV_CPU (obj)->env;set_misa (env,RVXLEN |RVI |RVM |RVA |RVC |RVU );set_priv_version (env,PRIV_VERSION_1_10_0 ); -set_resetvec (env,0x1004 );qdev_prop_set_bit (DEVICE (obj),"mmu" , false); }@@ -163,7 +159,6 @@ static void rv32_ibex_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); }@@ -373,6 +368,8 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)set_feature (env,RISCV_FEATURE_PMP ); } +set_resetvec (env, cpu->cfg.resetvec); +/* If misa isn't set (rv32 and rv64 machines) set it here */ if (!env->misa) {/* Do some ISA extension error checking */ --2.7.4
PATCH 4/18 hw/riscv: Initial support for Microchip PolarFire SoC Icicle Kit board
In reply to Bin MengFrom: Bin Meng <[email protected]>
This is an initial support for Microchip PolarFire SoC Icicle Kit. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
For more details about Microchip PolarFire Soc, please see:
https://www.microsemi.com/product-directory/soc-fpgas/5498-polarfire-soc-fpga
Unlike SiFive FU540, the RISC-V core resect vector is at 0x20220000. The following perepherals are created as an unimplemented device:
- Bus Error Uint 0/1/2/3/4
- L2 cache controller
- SYSREG
- MPUCFG
- IOSCBCFG
More devices will be added later.
The BIOS image used by this machine is hss.bin, aka Hart Software Services, which can be built from:
https://github.com/polarfire-soc/hart-software-services
To launch this machine:
$ qemu-system-riscv64 -nographic -M microchip-icicle-kit
The memory is set to 1 GiB by default to match the hardware. A sanity check on ram size is performed in the machine init routine to prompt user to increase the RAM size to > 1 GiB when less than 1 GiB ram is detected.
---
MAINTAINERS | 7 + default-configs/riscv64-softmmu.mak | 1 + hw/riscv/Kconfig | 6 + hw/riscv/Makefile.objs | 1 + hw/riscv/microchip_pfsoc.c | 312 ++++++++++++++++++++++++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 88 ++++++++++ 6 files changed, 415 insertions(+) create mode 100644 hw/riscv/microchip_pfsoc.c create mode 100644 include/hw/riscv/microchip_pfsoc.h
MAINTAINERS +7 -0
@@ -1314,6 +1314,13 @@ F: include/hw/riscv/opentitan.h F: include/hw/char/ibex_uart.h F: include/hw/intc/ibex_plic.h +Microchip PolarFire SoC Icicle Kit +M: Bin Meng <[email protected]> +L: [email protected] +S: Supported +F: hw/riscv/microchip_pfsoc.c +F: include/hw/riscv/microchip_pfsoc.h + RX Machines ----------- rx-gdbsim
default-configs/riscv64-softmmu.mak +1 -0
@@ -10,3 +10,4 @@ CONFIG_SPIKE=y CONFIG_SIFIVE_E=y CONFIG_SIFIVE_U=y CONFIG_RISCV_VIRT=y +CONFIG_MICROCHIP_PFSOC=y
hw/riscv/Kconfig +6 -0
@@ -48,3 +48,9 @@ config RISCV_VIRT select PCI_EXPRESS_GENERIC_BRIDGE select PFLASH_CFI01 select SIFIVE + +config MICROCHIP_PFSOC + bool + select HART + select SIFIVE + select UNIMP
hw/riscv/Makefile.objs +1 -0
@@ -14,3 +14,4 @@ obj-$(CONFIG_SIFIVE_U) += sifive_u_prci.o obj-$(CONFIG_SIFIVE) += sifive_uart.o obj-$(CONFIG_SPIKE) += spike.o obj-$(CONFIG_RISCV_VIRT) += virt.o +obj-$(CONFIG_MICROCHIP_PFSOC) += microchip_pfsoc.o
hw/riscv/microchip_pfsoc.c new file +312 -0
312 lines not on this page · raw message
include/hw/riscv/microchip_pfsoc.h new file +88 -0
88 lines not on this page · raw message
Signature · 1 line
2.7.4
Re: [PATCH 04/18] hw/riscv: Initial support for Microchip PolarFire SoC Icicle Kit board
Quoted 26 lines · Bin Meng
From: Bin Meng <[email protected]>
This is an initial support for Microchip PolarFire SoC Icicle Kit. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
For more details about Microchip PolarFire Soc, please see:
https://www.microsemi.com/product-directory/soc-fpgas/5498-polarfire-soc-fpgaUnlike SiFive FU540, the RISC-V core resect vector is at
0x20220000. The following perepherals are created as an unimplemented device:- Bus Error Uint 0/1/2/3/4
- L2 cache controller
- SYSREG
- MPUCFG
- IOSCBCFGMore devices will be added later.
The BIOS image used by this machine is hss.bin, aka Hart Software Services, which can be built from:
https://github.com/polarfire-soc/hart-software-servicesTo launch this machine:
Quoted 8 lines
$ qemu-system-riscv64 -nographic -M microchip-icicle-kitThe memory is set to 1 GiB by default to match the hardware. A sanity check on ram size is performed in the machine init routine to prompt user to increase the RAM size to > 1 GiB when less than 1 GiB ram is detected.
Signed-off-by Bin Meng <[email protected]>
Looks good. I didn't check the addresses and PLIC config data with the datasheet, but I'm assuming it's correct.
Alistair
Quoted 478 lines
---
MAINTAINERS | 7 + default-configs/riscv64-softmmu.mak | 1 + hw/riscv/Kconfig | 6 + hw/riscv/Makefile.objs | 1 + hw/riscv/microchip_pfsoc.c | 312 ++++++++++++++++++++++++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 88 ++++++++++ 6 files changed, 415 insertions(+) create mode 100644 hw/riscv/microchip_pfsoc.c create mode 100644 include/hw/riscv/microchip_pfsoc.hMAINTAINERS +7
-0@@ -1314,6 +1314,13 @@ F: include/hw/riscv/opentitan.h F: include/hw/char/ibex_uart.h F: include/hw/intc/ibex_plic.h +Microchip PolarFire SoC Icicle Kit +M: Bin Meng <[email protected]> +L: [email protected] +S: Supported +F: hw/riscv/microchip_pfsoc.c +F: include/hw/riscv/microchip_pfsoc.h + RX Machines ----------- rx-gdbsimdefault-configs/riscv64-softmmu.mak +1
-0@@ -10,3 +10,4 @@ CONFIG_SPIKE=y CONFIG_SIFIVE_E=y CONFIG_SIFIVE_U=y CONFIG_RISCV_VIRT=y +CONFIG_MICROCHIP_PFSOC=yhw/riscv/Kconfig +6
-0@@ -48,3 +48,9 @@ config RISCV_VIRT select PCI_EXPRESS_GENERIC_BRIDGE select PFLASH_CFI01 select SIFIVE + +config MICROCHIP_PFSOC + bool + select HART + select SIFIVE + select UNIMPhw/riscv/Makefile.objs +1
-0@@ -14,3 +14,4 @@ obj-$(CONFIG_SIFIVE_U) += sifive_u_prci.o obj-$(CONFIG_SIFIVE) += sifive_uart.o obj-$(CONFIG_SPIKE) += spike.o obj-$(CONFIG_RISCV_VIRT) += virt.o +obj-$(CONFIG_MICROCHIP_PFSOC) += microchip_pfsoc.ohw/riscv/microchip_pfsoc.c new file +312
-0@@ -0,0 +1,312 @@ +/* +* QEMU RISC-V Board Compatible with Microchip PolarFire SoC Icicle Kit +* +* Copyright (c) 2020 Wind River Systems, Inc. +* +* Author: +* Bin Meng <[email protected]> +* +* Provides a board compatible with the Microchip PolarFire SoC Icicle Kit +* +* 0) CLINT (Core Level Interruptor) +* 1) PLIC (Platform Level Interrupt Controller) +* 2) eNVM (Embedded Non-Volatile Memory) +* +* This board currently generates devicetree dynamically that indicates at least +* two harts and up to five harts. +* +* 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 "qemu/error-report.h" +#include "qemu/log.h" +#include "qemu/units.h" +#include "qemu/cutils.h" +#include "qapi/error.h" +#include "hw/boards.h" +#include "hw/irq.h" +#include "hw/loader.h" +#include "hw/sysbus.h" +#include "hw/cpu/cluster.h" +#include "target/riscv/cpu.h" +#include "hw/misc/unimp.h" +#include "hw/riscv/boot.h" +#include "hw/riscv/riscv_hart.h" +#include "hw/riscv/sifive_clint.h" +#include "hw/riscv/sifive_plic.h" +#include "hw/riscv/microchip_pfsoc.h" + +/* +* The BIOS image used by this machine is called Hart Software Services (HSS). +* See https://github.com/polarfire-soc/hart-software-services +*/ +#define BIOS_FILENAME "hss.bin" +#define RESET_VECTOR 0x20220000 + +static const struct MemmapEntry { +hwaddr base; +hwaddr size; +} microchip_pfsoc_memmap[] = { + [MICROCHIP_PFSOC_DEBUG ] = {0x0 ,0x1000 }, + [MICROCHIP_PFSOC_E51_DTIM ] = {0x1000000 ,0x2000 }, + [MICROCHIP_PFSOC_BUSERR_UNIT0 ] = {0x1700000 ,0x1000 }, + [MICROCHIP_PFSOC_BUSERR_UNIT1 ] = {0x1701000 ,0x1000 }, + [MICROCHIP_PFSOC_BUSERR_UNIT2 ] = {0x1702000 ,0x1000 }, + [MICROCHIP_PFSOC_BUSERR_UNIT3 ] = {0x1703000 ,0x1000 }, + [MICROCHIP_PFSOC_BUSERR_UNIT4 ] = {0x1704000 ,0x1000 }, + [MICROCHIP_PFSOC_CLINT ] = {0x2000000 ,0x10000 }, + [MICROCHIP_PFSOC_L2CC ] = {0x2010000 ,0x1000 }, + [MICROCHIP_PFSOC_L2LIM ] = {0x8000000 ,0x2000000 }, + [MICROCHIP_PFSOC_PLIC ] = {0xc000000 ,0x4000000 }, + [MICROCHIP_PFSOC_SYSREG ] = {0x20002000 ,0x2000 }, + [MICROCHIP_PFSOC_MPUCFG ] = {0x20005000 ,0x1000 }, + [MICROCHIP_PFSOC_ENVM_CFG ] = {0x20200000 ,0x1000 }, + [MICROCHIP_PFSOC_ENVM_DATA ] = {0x20220000 ,0x20000 }, + [MICROCHIP_PFSOC_IOSCB_CFG ] = {0x37080000 ,0x1000 }, + [MICROCHIP_PFSOC_DRAM ] = {0x80000000 ,0x0 }, +}; + +static void microchip_pfsoc_soc_instance_init (Object *obj) +{ +MachineState *ms =MACHINE (qdev_get_machine ()); +MicrochipPFSoCState *s =MICROCHIP_PFSOC (obj); + +object_initialize_child (obj,"e-cluster" , &s->e_cluster,TYPE_CPU_CLUSTER ); +qdev_prop_set_uint32 (DEVICE (&s->e_cluster),"cluster-id" ,0 ); + +object_initialize_child (OBJECT (&s->e_cluster),"e-cpus" , &s->e_cpus, +TYPE_RISCV_HART_ARRAY ); +qdev_prop_set_uint32 (DEVICE (&s->e_cpus),"num-harts" ,1 ); +qdev_prop_set_uint32 (DEVICE (&s->e_cpus),"hartid-base" ,0 ); +qdev_prop_set_string (DEVICE (&s->e_cpus),"cpu-type" , +TYPE_RISCV_CPU_SIFIVE_E51 ); +qdev_prop_set_uint64 (DEVICE (&s->e_cpus),"resetvec" ,RESET_VECTOR ); + +object_initialize_child (obj,"u-cluster" , &s->u_cluster,TYPE_CPU_CLUSTER ); +qdev_prop_set_uint32 (DEVICE (&s->u_cluster),"cluster-id" ,1 ); + +object_initialize_child (OBJECT (&s->u_cluster),"u-cpus" , &s->u_cpus, +TYPE_RISCV_HART_ARRAY ); +qdev_prop_set_uint32 (DEVICE (&s->u_cpus),"num-harts" , ms->smp.cpus -1 ); +qdev_prop_set_uint32 (DEVICE (&s->u_cpus),"hartid-base" ,1 ); +qdev_prop_set_string (DEVICE (&s->u_cpus),"cpu-type" , +TYPE_RISCV_CPU_SIFIVE_U54 ); +qdev_prop_set_uint64 (DEVICE (&s->u_cpus),"resetvec" ,RESET_VECTOR ); +} + +static void microchip_pfsoc_soc_realize (DeviceState *dev,Error **errp) +{ +MachineState *ms =MACHINE (qdev_get_machine ()); +MicrochipPFSoCState *s =MICROCHIP_PFSOC (dev); +const struct MemmapEntry *memmap = microchip_pfsoc_memmap; +MemoryRegion *system_memory =get_system_memory (); +MemoryRegion *e51_dtim_mem =g_new (MemoryRegion,1 ); +MemoryRegion *l2lim_mem =g_new (MemoryRegion,1 ); +MemoryRegion *envm_data =g_new (MemoryRegion,1 ); +char *plic_hart_config; +size_t plic_hart_config_len; +int i; + +sysbus_realize (SYS_BUS_DEVICE (&s->e_cpus), &error_abort); +sysbus_realize (SYS_BUS_DEVICE (&s->u_cpus), &error_abort); +/* +* The cluster must be realized after the RISC-V hart array container, +* as the container's CPU object is only created on realize, and the +* CPU must exist and have been parented into the cluster before the +* cluster is realized. +*/ +qdev_realize (DEVICE (&s->e_cluster),NULL , &error_abort); +qdev_realize (DEVICE (&s->u_cluster),NULL , &error_abort); + +/* E51 DTIM */ +memory_region_init_ram (e51_dtim_mem,NULL ,"microchip.pfsoc.e51_dtim_mem" , + memmap[MICROCHIP_PFSOC_E51_DTIM ].size, &error_fatal); +memory_region_add_subregion (system_memory, + memmap[MICROCHIP_PFSOC_E51_DTIM ].base, + e51_dtim_mem); + +/* Bus Error Units */ +create_unimplemented_device ("microchip.pfsoc.buserr_unit0_mem" , + memmap[MICROCHIP_PFSOC_BUSERR_UNIT0 ].base, + memmap[MICROCHIP_PFSOC_BUSERR_UNIT0 ].size); +create_unimplemented_device ("microchip.pfsoc.buserr_unit1_mem" , + memmap[MICROCHIP_PFSOC_BUSERR_UNIT1 ].base, + memmap[MICROCHIP_PFSOC_BUSERR_UNIT1 ].size); +create_unimplemented_device ("microchip.pfsoc.buserr_unit2_mem" , + memmap[MICROCHIP_PFSOC_BUSERR_UNIT2 ].base, + memmap[MICROCHIP_PFSOC_BUSERR_UNIT2 ].size); +create_unimplemented_device ("microchip.pfsoc.buserr_unit3_mem" , + memmap[MICROCHIP_PFSOC_BUSERR_UNIT3 ].base, + memmap[MICROCHIP_PFSOC_BUSERR_UNIT3 ].size); +create_unimplemented_device ("microchip.pfsoc.buserr_unit4_mem" , + memmap[MICROCHIP_PFSOC_BUSERR_UNIT4 ].base, + memmap[MICROCHIP_PFSOC_BUSERR_UNIT4 ].size); + +/* CLINT */ +sifive_clint_create (memmap[MICROCHIP_PFSOC_CLINT ].base, + memmap[MICROCHIP_PFSOC_CLINT ].size, ms->smp.cpus, +SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , false); + +/* L2 cache controller */ +create_unimplemented_device ("microchip.pfsoc.l2cc" , + memmap[MICROCHIP_PFSOC_L2CC ].base, memmap[MICROCHIP_PFSOC_L2CC ].size); + +/* +* Add L2-LIM at reset size. +* This should be reduced in size as the L2 Cache Controller WayEnable +* register is incremented. Unfortunately I don't see a nice (or any) way +* to handle reducing or blocking out the L2 LIM while still allowing it +* be re returned to all enabled after a reset. For the time being, just +* leave it enabled all the time. This won't break anything, but will be +* too generous to misbehaving guests. +*/ +memory_region_init_ram (l2lim_mem,NULL ,"microchip.pfsoc.l2lim" , + memmap[MICROCHIP_PFSOC_L2LIM ].size, &error_fatal); +memory_region_add_subregion (system_memory, + memmap[MICROCHIP_PFSOC_L2LIM ].base, + l2lim_mem); + +/* create PLIC hart topology configuration string */ + plic_hart_config_len = (strlen (MICROCHIP_PFSOC_PLIC_HART_CONFIG ) +1 ) * + ms->smp.cpus; + plic_hart_config =g_malloc0 (plic_hart_config_len); +for (i =0 ; i < ms->smp.cpus; i++) { +if (i !=0 ) { +strncat (plic_hart_config,"," MICROCHIP_PFSOC_PLIC_HART_CONFIG , + plic_hart_config_len); + }else { +strncat (plic_hart_config,"M" , plic_hart_config_len); + } + plic_hart_config_len -= (strlen (MICROCHIP_PFSOC_PLIC_HART_CONFIG ) +1 ); + } + +/* PLIC */ + s->plic =sifive_plic_create (memmap[MICROCHIP_PFSOC_PLIC ].base, + plic_hart_config, +MICROCHIP_PFSOC_PLIC_NUM_SOURCES , +MICROCHIP_PFSOC_PLIC_NUM_PRIORITIES , +MICROCHIP_PFSOC_PLIC_PRIORITY_BASE , +MICROCHIP_PFSOC_PLIC_PENDING_BASE , +MICROCHIP_PFSOC_PLIC_ENABLE_BASE , +MICROCHIP_PFSOC_PLIC_ENABLE_STRIDE , +MICROCHIP_PFSOC_PLIC_CONTEXT_BASE , +MICROCHIP_PFSOC_PLIC_CONTEXT_STRIDE , + memmap[MICROCHIP_PFSOC_PLIC ].size); +g_free (plic_hart_config); + +/* SYSREG */ +create_unimplemented_device ("microchip.pfsoc.sysreg" , + memmap[MICROCHIP_PFSOC_SYSREG ].base, + memmap[MICROCHIP_PFSOC_SYSREG ].size); + +/* MPUCFG */ +create_unimplemented_device ("microchip.pfsoc.mpucfg" , + memmap[MICROCHIP_PFSOC_MPUCFG ].base, + memmap[MICROCHIP_PFSOC_MPUCFG ].size); + +/* eNVM */ +memory_region_init_rom (envm_data,OBJECT (dev),"microchip.pfsoc.envm.data" , + memmap[MICROCHIP_PFSOC_ENVM_DATA ].size, + &error_fatal); +memory_region_add_subregion (system_memory, + memmap[MICROCHIP_PFSOC_ENVM_DATA ].base, + envm_data); + +/* IOSCBCFG */ +create_unimplemented_device ("microchip.pfsoc.ioscb.cfg" , + memmap[MICROCHIP_PFSOC_IOSCB_CFG ].base, + memmap[MICROCHIP_PFSOC_IOSCB_CFG ].size); +} + +static void microchip_pfsoc_soc_class_init (ObjectClass *oc,void *data) +{ +DeviceClass *dc =DEVICE_CLASS (oc); + + dc->realize = microchip_pfsoc_soc_realize; +/* Reason: Uses serial_hds in realize function, thus can't be used twice */ + dc->user_creatable = false; +} + +static const TypeInfo microchip_pfsoc_soc_type_info = { + .name =TYPE_MICROCHIP_PFSOC , + .parent =TYPE_DEVICE , + .instance_size =sizeof (MicrochipPFSoCState), + .instance_init = microchip_pfsoc_soc_instance_init, + .class_init = microchip_pfsoc_soc_class_init, +}; + +static void microchip_pfsoc_soc_register_types (void ) +{ +type_register_static (µchip_pfsoc_soc_type_info); +} + +type_init (microchip_pfsoc_soc_register_types) + +static void microchip_icicle_kit_machine_init(MachineState *machine) +{ +MachineClass *mc =MACHINE_GET_CLASS (machine); +const struct MemmapEntry *memmap = microchip_pfsoc_memmap; +MicrochipIcicleKitState *s =MICROCHIP_ICICLE_KIT_MACHINE (machine); +MemoryRegion *system_memory =get_system_memory (); +MemoryRegion *main_mem =g_new (MemoryRegion,1 ); + +/* Sanity check on RAM size */ +if (machine->ram_size < mc->default_ram_size) { +char *sz =size_to_str (mc->default_ram_size); +error_report ("Invalid RAM size, should be bigger than %s" , sz); +g_free (sz); +exit (EXIT_FAILURE ); + } + +/* Initialize SoC */ +object_initialize_child (OBJECT (machine),"soc" , &s->soc, +TYPE_MICROCHIP_PFSOC ); +qdev_realize (DEVICE (&s->soc),NULL , &error_abort); + +/* Register RAM */ +memory_region_init_ram (main_mem,NULL ,"microchip.icicle.kit.ram" , + machine->ram_size, &error_fatal); +memory_region_add_subregion (system_memory, + memmap[MICROCHIP_PFSOC_DRAM ].base, main_mem); + +/* Load the firmware */ +riscv_find_and_load_firmware (machine,BIOS_FILENAME ,RESET_VECTOR ,NULL ); +} + +static void microchip_icicle_kit_machine_class_init (ObjectClass *oc,void *data) +{ +MachineClass *mc =MACHINE_CLASS (oc); + + mc->desc ="Microchip PolarFire SoC Icicle Kit" ; + mc->init = microchip_icicle_kit_machine_init; + mc->max_cpus =MICROCHIP_PFSOC_MANAGEMENT_CPU_COUNT + +MICROCHIP_PFSOC_COMPUTE_CPU_COUNT ; + mc->min_cpus =MICROCHIP_PFSOC_MANAGEMENT_CPU_COUNT +1 ; + mc->default_cpus = mc->min_cpus; + mc->default_ram_size =1 * GiB; +} + +static const TypeInfo microchip_icicle_kit_machine_typeinfo = { + .name =MACHINE_TYPE_NAME ("microchip-icicle-kit" ), + .parent =TYPE_MACHINE , + .class_init = microchip_icicle_kit_machine_class_init, + .instance_size =sizeof (MicrochipIcicleKitState), +}; + +static void microchip_icicle_kit_machine_init_register_types (void ) +{ +type_register_static (µchip_icicle_kit_machine_typeinfo); +} + +type_init (microchip_icicle_kit_machine_init_register_types)include/hw/riscv/microchip_pfsoc.h new file +88
-1@@ -0,0 +1,88 @@ +/* +* Microchip PolarFire SoC machine interface +* +* Copyright (c) 2020 Wind River Systems, Inc. +* +* Author: +* Bin Meng <[email protected]> +* +* 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_MICROCHIP_PFSOC_H +#define HW_MICROCHIP_PFSOC_H + +typedef struct MicrochipPFSoCState { +/*< private >*/ +DeviceState parent_obj; + +/*< public >*/ +CPUClusterState e_cluster; +CPUClusterState u_cluster; +RISCVHartArrayState e_cpus; +RISCVHartArrayState u_cpus; +DeviceState *plic; +}MicrochipPFSoCState ; + +#define TYPE_MICROCHIP_PFSOC "microchip.pfsoc" +#define MICROCHIP_PFSOC (obj) \ + OBJECT_CHECK(MicrochipPFSoCState, (obj), TYPE_MICROCHIP_PFSOC) + +typedef struct MicrochipIcicleKitState { +/*< private >*/ +MachineState parent_obj; + +/*< public >*/ +MicrochipPFSoCState soc; +}MicrochipIcicleKitState ; + +#define TYPE_MICROCHIP_ICICLE_KIT_MACHINE \ + MACHINE_TYPE_NAME("microchip-icicle-kit") +#define MICROCHIP_ICICLE_KIT_MACHINE (obj) \ + OBJECT_CHECK(MicrochipIcicleKitState, (obj), \ + TYPE_MICROCHIP_ICICLE_KIT_MACHINE) + +enum { +MICROCHIP_PFSOC_DEBUG , +MICROCHIP_PFSOC_E51_DTIM , +MICROCHIP_PFSOC_BUSERR_UNIT0 , +MICROCHIP_PFSOC_BUSERR_UNIT1 , +MICROCHIP_PFSOC_BUSERR_UNIT2 , +MICROCHIP_PFSOC_BUSERR_UNIT3 , +MICROCHIP_PFSOC_BUSERR_UNIT4 , +MICROCHIP_PFSOC_CLINT , +MICROCHIP_PFSOC_L2CC , +MICROCHIP_PFSOC_L2LIM , +MICROCHIP_PFSOC_PLIC , +MICROCHIP_PFSOC_SYSREG , +MICROCHIP_PFSOC_MPUCFG , +MICROCHIP_PFSOC_ENVM_CFG , +MICROCHIP_PFSOC_ENVM_DATA , +MICROCHIP_PFSOC_IOSCB_CFG , +MICROCHIP_PFSOC_DRAM , +}; + +#define MICROCHIP_PFSOC_MANAGEMENT_CPU_COUNT 1 +#define MICROCHIP_PFSOC_COMPUTE_CPU_COUNT 4 + +#define MICROCHIP_PFSOC_PLIC_HART_CONFIG "MS" +#define MICROCHIP_PFSOC_PLIC_NUM_SOURCES 185 +#define MICROCHIP_PFSOC_PLIC_NUM_PRIORITIES 7 +#define MICROCHIP_PFSOC_PLIC_PRIORITY_BASE 0x04 +#define MICROCHIP_PFSOC_PLIC_PENDING_BASE 0x1000 +#define MICROCHIP_PFSOC_PLIC_ENABLE_BASE 0x2000 +#define MICROCHIP_PFSOC_PLIC_ENABLE_STRIDE 0x80 +#define MICROCHIP_PFSOC_PLIC_CONTEXT_BASE 0x200000 +#define MICROCHIP_PFSOC_PLIC_CONTEXT_STRIDE 0x1000 + +#endif /* HW_MICROCHIP_PFSOC_H */ --2.7.4
PATCH 5/18 hw/char: Add Microchip PolarFire SoC MMUART emulation
In reply to Bin MengFrom: Bin Meng <[email protected]>
Microchip PolarFire SoC MMUART is ns16550 compatible, with some additional registers. Create a simple MMUART model built on top of the existing ns16550 model.
---
MAINTAINERS | 2 + hw/char/Kconfig | 3 ++ hw/char/Makefile.objs | 1 + hw/char/mchp_pfsoc_mmuart.c | 82 +++++++++++++++++++++++++++++++++++++ include/hw/char/mchp_pfsoc_mmuart.h | 61 +++++++++++++++++++++++++++ 5 files changed, 149 insertions(+) create mode 100644 hw/char/mchp_pfsoc_mmuart.c create mode 100644 include/hw/char/mchp_pfsoc_mmuart.h
MAINTAINERS +2 -0
@@ -1319,7 +1319,9 @@ M: Bin Meng <[email protected]> L: [email protected] S: Supported F: hw/riscv/microchip_pfsoc.c +F: hw/char/mchp_pfsoc_mmuart.c F: include/hw/riscv/microchip_pfsoc.h +F: include/hw/char/mchp_pfsoc_mmuart.h RX Machines -----------
hw/char/Kconfig +3 -0
@@ -52,3 +52,6 @@ config RENESAS_SCI config AVR_USART bool + +config MCHP_PFSOC_MMUART + bool
hw/char/Makefile.objs +1 -0
@@ -33,6 +33,7 @@ common-obj-$(CONFIG_LM32) += lm32_juart.o common-obj-$(CONFIG_LM32) += lm32_uart.o common-obj-$(CONFIG_MILKYMIST) += milkymist-uart.o common-obj-$(CONFIG_SCLPCONSOLE) += sclpconsole.o sclpconsole-lm.o +common-obj-$(CONFIG_MCHP_PFSOC_MMUART) += mchp_pfsoc_mmuart.o obj-$(CONFIG_VIRTIO) += virtio-serial-bus.o obj-$(CONFIG_PSERIES) += spapr_vty.o
hw/char/mchp_pfsoc_mmuart.c new file +82 -0
82 lines not on this page · raw message
include/hw/char/mchp_pfsoc_mmuart.h new file +61 -0
61 lines not on this page · raw message
Signature · 1 line
2.7.4
Re: [PATCH 05/18] hw/char: Add Microchip PolarFire SoC MMUART emulation
Quoted 85 lines · Bin Meng
From: Bin Meng <[email protected]>
Microchip PolarFire SoC MMUART is ns16550 compatible, with some additional registers. Create a simple MMUART model built on top of the existing ns16550 model.
Signed-off-by Bin Meng <[email protected]>---
MAINTAINERS | 2 + hw/char/Kconfig | 3 ++ hw/char/Makefile.objs | 1 + hw/char/mchp_pfsoc_mmuart.c | 82 +++++++++++++++++++++++++++++++++++++ include/hw/char/mchp_pfsoc_mmuart.h | 61 +++++++++++++++++++++++++++ 5 files changed, 149 insertions(+) create mode 100644 hw/char/mchp_pfsoc_mmuart.c create mode 100644 include/hw/char/mchp_pfsoc_mmuart.hMAINTAINERS +2
-0@@ -1319,7 +1319,9 @@ M: Bin Meng <[email protected]> L: [email protected] S: Supported F: hw/riscv/microchip_pfsoc.c +F: hw/char/mchp_pfsoc_mmuart.c F: include/hw/riscv/microchip_pfsoc.h +F: include/hw/char/mchp_pfsoc_mmuart.h RX Machines -----------hw/char/Kconfig +3
-0@@ -52,3 +52,6 @@ config RENESAS_SCI config AVR_USART bool + +config MCHP_PFSOC_MMUART + boolhw/char/Makefile.objs +1
-0@@ -33,6 +33,7 @@ common-obj-$(CONFIG_LM32) += lm32_juart.o common-obj-$(CONFIG_LM32) += lm32_uart.o common-obj-$(CONFIG_MILKYMIST) += milkymist-uart.o common-obj-$(CONFIG_SCLPCONSOLE) += sclpconsole.o sclpconsole-lm.o +common-obj-$(CONFIG_MCHP_PFSOC_MMUART) += mchp_pfsoc_mmuart.o obj-$(CONFIG_VIRTIO) += virtio-serial-bus.o obj-$(CONFIG_PSERIES) += spapr_vty.ohw/char/mchp_pfsoc_mmuart.c new file +23
-0@@ -0,0 +1,82 @@ +/* +* Microchip PolarFire SoC MMUART emulation +* +* Copyright (c) 2020 Wind River Systems, Inc. +* +* Author: +* Bin Meng <[email protected]> +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU General Public License as +* published by the Free Software Foundation; either version 2 or +* (at your option) version 3 of the License. +* +* This program is distributed in the hope that it will be useful, +* but WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +* GNU General Public License for more details. +* +* You should have received a copy of the GNU General Public License along +* with this program; if not, see <http://www.gnu.org/licenses/>. +*/ + +#include "qemu/osdep.h"
hw/char/mchp_pfsoc_mmuart.c line 33
Quoted 10 lines
hw/char/mchp_pfsoc_mmuart.c new file +10
-0+#include "qemu/log.h" +#include "chardev/char.h" +#include "exec/address-spaces.h" +#include "hw/char/mchp_pfsoc_mmuart.h" + +static uint64_t mchp_pfsoc_mmuart_read (void *opaque,hwaddr addr,unsigned size) +{ +MchpPfSoCMMUartState *s = opaque; + +if ((addr %sizeof (uint32_t )) || (addr >=0x34 )) {
You can specify impl.min_access_size and impl.max_access_size as part of MemoryRegionOps so the allignment check isn't required here.
Quoted 5 lines
hw/char/mchp_pfsoc_mmuart.c new file +5
-0+qemu_log_mask (LOG_GUEST_ERROR ,"%s: read: addr=0x%" HWADDR_PRIx"\n" , + __func__, addr); +return 0 ; + } +
hw/char/mchp_pfsoc_mmuart.c line 48
Quoted 10 lines
hw/char/mchp_pfsoc_mmuart.c new file +10
-0+return s->reg[addr /sizeof (uint32_t )]; +} + +static void mchp_pfsoc_mmuart_write (void *opaque,hwaddr addr, +uint64_t value,unsigned size) +{ +MchpPfSoCMMUartState *s = opaque; +uint32_t val32 = (uint32_t )value; + +if ((addr %sizeof (uint32_t )) || (addr >=0x34 )) {
Same comment here about allignment.
Alistair
Quoted 105 lines
hw/char/mchp_pfsoc_mmuart.c new file +34
-0+qemu_log_mask (LOG_GUEST_ERROR ,"%s: bad write: addr=0x%" HWADDR_PRIx +" v=0x%x\n" , __func__, addr, val32); +return ; + } + + s->reg[addr /sizeof (uint32_t )] = val32; +} + +static const MemoryRegionOps mchp_pfsoc_mmuart_ops = { + .read = mchp_pfsoc_mmuart_read, + .write = mchp_pfsoc_mmuart_write, + .endianness =DEVICE_LITTLE_ENDIAN , +}; + +MchpPfSoCMMUartState *mchp_pfsoc_mmuart_create (MemoryRegion *sysmem, +hwaddr base,qemu_irq irq,Chardev *chr) +{ +MchpPfSoCMMUartState *s; + + s =g_new0 (MchpPfSoCMMUartState,1 ); + +memory_region_init_io (&s->iomem,NULL , &mchp_pfsoc_mmuart_ops, s, +"mchp.pfsoc.mmuart" ,0x1000 ); + + s->base = base; + s->irq = irq; + + s->serial =serial_mm_init (sysmem, base,2 , irq,399193 , chr, +DEVICE_LITTLE_ENDIAN ); + +memory_region_add_subregion (sysmem, base +0x20 , &s->iomem); + +return s; +}include/hw/char/mchp_pfsoc_mmuart.h new file +61
-1@@ -0,0 +1,61 @@ +/* +* Microchip PolarFire SoC MMUART emulation +* +* Copyright (c) 2020 Wind River Systems, Inc. +* +* Author: +* Bin Meng <[email protected]> +* +* 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_MCHP_PFSOC_MMUART_H +#define HW_MCHP_PFSOC_MMUART_H + +#include "hw/char/serial.h" + +#define MCHP_PFSOC_MMUART_REG_SIZE 52 + +typedef struct MchpPfSoCMMUartState { +MemoryRegion iomem; +hwaddr base; +qemu_irq irq; + +SerialMM *serial; + +uint32_t reg[MCHP_PFSOC_MMUART_REG_SIZE /sizeof (uint32_t )]; +}MchpPfSoCMMUartState ; + +/** +* mchp_pfsoc_mmuart_create - Create a Microchip PolarFire SoC MMUART +* +* This is a helper routine for board to create a MMUART device that is +* compatible with Microchip PolarFire SoC. +* +* @sysmem: system memory region to map +* @base: base address of the MMUART registers +* @irq: IRQ number of the MMUART device +* @chr: character device to associate to +* +* @return: a pointer to the device specific control structure +*/ +MchpPfSoCMMUartState *mchp_pfsoc_mmuart_create (MemoryRegion *sysmem, +hwaddr base,qemu_irq irq,Chardev *chr); + +#endif /* HW_MCHP_PFSOC_MMUART_H */ --2.7.4
PATCH 6/18 hw/riscv: microchip_pfsoc: Connect 5 MMUARTs
In reply to Bin MengFrom: Bin Meng <[email protected]>
Microchip PolarFire SoC has 5 MMUARTs, and the Icicle Kit board wires 4 of them out. Let's connect all 5 MMUARTs.
---
hw/riscv/Kconfig | 1 + hw/riscv/microchip_pfsoc.c | 30 ++++++++++++++++++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 20 ++++++++++++++++++++ 3 files changed, 51 insertions(+)
hw/riscv/Kconfig +1 -0
@@ -54,3 +54,4 @@ config MICROCHIP_PFSOC select HART select SIFIVE select UNIMP + select MCHP_PFSOC_MMUART
hw/riscv/microchip_pfsoc.c +30 -0
@@ -11,6 +11,7 @@* 0) CLINT (Core Level Interruptor) * 1) PLIC (Platform Level Interrupt Controller) * 2) eNVM (Embedded Non-Volatile Memory) +* 3) MMUARTs (Multi-Mode UART) * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts.
@@ -38,6 +39,7 @@#include "hw/irq.h" #include "hw/loader.h" #include "hw/sysbus.h" +#include "chardev/char.h" #include "hw/cpu/cluster.h" #include "target/riscv/cpu.h" #include "hw/misc/unimp.h"
@@ -46,6 +48,7 @@#include "hw/riscv/sifive_clint.h" #include "hw/riscv/sifive_plic.h" #include "hw/riscv/microchip_pfsoc.h" +#include "sysemu/sysemu.h" /* * The BIOS image used by this machine is called Hart Software Services (HSS).
@@ -69,8 +72,13 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_L2CC] = { 0x2010000, 0x1000 }, [MICROCHIP_PFSOC_L2LIM] = { 0x8000000, 0x2000000 }, [MICROCHIP_PFSOC_PLIC] = { 0xc000000, 0x4000000 }, + [MICROCHIP_PFSOC_MMUART0] = { 0x20000000, 0x1000 }, [MICROCHIP_PFSOC_SYSREG] = { 0x20002000, 0x2000 }, [MICROCHIP_PFSOC_MPUCFG] = { 0x20005000, 0x1000 }, + [MICROCHIP_PFSOC_MMUART1] = { 0x20100000, 0x1000 }, + [MICROCHIP_PFSOC_MMUART2] = { 0x20102000, 0x1000 }, + [MICROCHIP_PFSOC_MMUART3] = { 0x20104000, 0x1000 }, + [MICROCHIP_PFSOC_MMUART4] = { 0x20106000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_CFG] = { 0x20200000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_DATA] = { 0x20220000, 0x20000 }, [MICROCHIP_PFSOC_IOSCB_CFG] = { 0x37080000, 0x1000 },
@@ -215,6 +223,28 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp) memmap[MICROCHIP_PFSOC_MPUCFG ].base, memmap[MICROCHIP_PFSOC_MPUCFG].size); +/* MMUARTs */ + s->serial0 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART0 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART0_IRQ ), +serial_hd (0 )); + s->serial1 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART1 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART1_IRQ ), +serial_hd (1 )); + s->serial2 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART2 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART2_IRQ ), +serial_hd (2 )); + s->serial3 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART3 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART3_IRQ ), +serial_hd (3 )); + s->serial4 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART4 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART4_IRQ ), +serial_hd (4 )); +/* eNVM */ memory_region_init_rom(envm_data, OBJECT(dev), "microchip.pfsoc.envm.data", memmap[MICROCHIP_PFSOC_ENVM_DATA].size,
include/hw/riscv/microchip_pfsoc.h +20 -0
@@ -22,6 +22,8 @@ #ifndef HW_MICROCHIP_PFSOC_H #define HW_MICROCHIP_PFSOC_H +#include "hw/char/mchp_pfsoc_mmuart.h" + typedef struct MicrochipPFSoCState { /*< private >*/ DeviceState parent_obj;
@@ -32,6 +34,11 @@ typedef struct MicrochipPFSoCState {RISCVHartArrayState e_cpus;RISCVHartArrayState u_cpus;DeviceState *plic; +MchpPfSoCMMUartState *serial0; +MchpPfSoCMMUartState *serial1; +MchpPfSoCMMUartState *serial2; +MchpPfSoCMMUartState *serial3; +MchpPfSoCMMUartState *serial4; } MicrochipPFSoCState;#define TYPE_MICROCHIP_PFSOC "microchip.pfsoc"
@@ -64,14 +71,27 @@ enum {MICROCHIP_PFSOC_L2CC ,MICROCHIP_PFSOC_L2LIM ,MICROCHIP_PFSOC_PLIC , +MICROCHIP_PFSOC_MMUART0 ,MICROCHIP_PFSOC_SYSREG ,MICROCHIP_PFSOC_MPUCFG , +MICROCHIP_PFSOC_MMUART1 , +MICROCHIP_PFSOC_MMUART2 , +MICROCHIP_PFSOC_MMUART3 , +MICROCHIP_PFSOC_MMUART4 ,MICROCHIP_PFSOC_ENVM_CFG ,MICROCHIP_PFSOC_ENVM_DATA ,MICROCHIP_PFSOC_IOSCB_CFG , MICROCHIP_PFSOC_DRAM, }; +enum { +MICROCHIP_PFSOC_MMUART0_IRQ =90 , +MICROCHIP_PFSOC_MMUART1_IRQ =91 , +MICROCHIP_PFSOC_MMUART2_IRQ =92 , +MICROCHIP_PFSOC_MMUART3_IRQ =93 , +MICROCHIP_PFSOC_MMUART4_IRQ =94 , +}; +#define MICROCHIP_PFSOC_MANAGEMENT_CPU_COUNT 1#define MICROCHIP_PFSOC_COMPUTE_CPU_COUNT 4
Signature · 1 line
2.7.4
Re: [PATCH 06/18] hw/riscv: microchip_pfsoc: Connect 5 MMUARTs
Quoted 7 lines · Bin Meng
From: Bin Meng <[email protected]>
Microchip PolarFire SoC has 5 MMUARTs, and the Icicle Kit board wires 4 of them out. Let's connect all 5 MMUARTs.
Signed-off-by Bin Meng <[email protected]>
Alistair
Quoted 144 lines
---
hw/riscv/Kconfig | 1 + hw/riscv/microchip_pfsoc.c | 30 ++++++++++++++++++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 20 ++++++++++++++++++++ 3 files changed, 51 insertions(+)hw/riscv/Kconfig +1
-0@@ -54,3 +54,4 @@ config MICROCHIP_PFSOC select HART select SIFIVE select UNIMP + select MCHP_PFSOC_MMUARThw/riscv/microchip_pfsoc.c +30
-0@@ -11,6 +11,7 @@* 0) CLINT (Core Level Interruptor) * 1) PLIC (Platform Level Interrupt Controller) * 2) eNVM (Embedded Non-Volatile Memory) +* 3) MMUARTs (Multi-Mode UART) * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts. @@ -38,6 +39,7 @@#include "hw/irq.h" #include "hw/loader.h" #include "hw/sysbus.h" +#include "chardev/char.h" #include "hw/cpu/cluster.h" #include "target/riscv/cpu.h" #include "hw/misc/unimp.h" @@ -46,6 +48,7 @@#include "hw/riscv/sifive_clint.h" #include "hw/riscv/sifive_plic.h" #include "hw/riscv/microchip_pfsoc.h" +#include "sysemu/sysemu.h" /* * The BIOS image used by this machine is called Hart Software Services (HSS). @@ -69,8 +72,13 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_L2CC] = { 0x2010000, 0x1000 }, [MICROCHIP_PFSOC_L2LIM] = { 0x8000000, 0x2000000 }, [MICROCHIP_PFSOC_PLIC] = { 0xc000000, 0x4000000 }, + [MICROCHIP_PFSOC_MMUART0] = { 0x20000000, 0x1000 }, [MICROCHIP_PFSOC_SYSREG] = { 0x20002000, 0x2000 }, [MICROCHIP_PFSOC_MPUCFG] = { 0x20005000, 0x1000 }, + [MICROCHIP_PFSOC_MMUART1] = { 0x20100000, 0x1000 }, + [MICROCHIP_PFSOC_MMUART2] = { 0x20102000, 0x1000 }, + [MICROCHIP_PFSOC_MMUART3] = { 0x20104000, 0x1000 }, + [MICROCHIP_PFSOC_MMUART4] = { 0x20106000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_CFG] = { 0x20200000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_DATA] = { 0x20220000, 0x20000 }, [MICROCHIP_PFSOC_IOSCB_CFG] = { 0x37080000, 0x1000 },@@ -215,6 +223,28 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp) memmap[MICROCHIP_PFSOC_MPUCFG ].base, memmap[MICROCHIP_PFSOC_MPUCFG].size); +/* MMUARTs */ + s->serial0 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART0 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART0_IRQ ), +serial_hd (0 )); + s->serial1 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART1 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART1_IRQ ), +serial_hd (1 )); + s->serial2 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART2 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART2_IRQ ), +serial_hd (2 )); + s->serial3 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART3 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART3_IRQ ), +serial_hd (3 )); + s->serial4 =mchp_pfsoc_mmuart_create (system_memory, + memmap[MICROCHIP_PFSOC_MMUART4 ].base, +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART4_IRQ ), +serial_hd (4 )); +/* eNVM */ memory_region_init_rom(envm_data, OBJECT(dev), "microchip.pfsoc.envm.data", memmap[MICROCHIP_PFSOC_ENVM_DATA].size,include/hw/riscv/microchip_pfsoc.h +20
-1@@ -22,6 +22,8 @@ #ifndef HW_MICROCHIP_PFSOC_H #define HW_MICROCHIP_PFSOC_H +#include "hw/char/mchp_pfsoc_mmuart.h" + typedef struct MicrochipPFSoCState { /*< private >*/ DeviceState parent_obj;@@ -32,6 +34,11 @@ typedef struct MicrochipPFSoCState {RISCVHartArrayState e_cpus;RISCVHartArrayState u_cpus;DeviceState *plic; +MchpPfSoCMMUartState *serial0; +MchpPfSoCMMUartState *serial1; +MchpPfSoCMMUartState *serial2; +MchpPfSoCMMUartState *serial3; +MchpPfSoCMMUartState *serial4; } MicrochipPFSoCState;#define TYPE_MICROCHIP_PFSOC "microchip.pfsoc"@@ -64,14 +71,27 @@ enum {MICROCHIP_PFSOC_L2CC ,MICROCHIP_PFSOC_L2LIM ,MICROCHIP_PFSOC_PLIC , +MICROCHIP_PFSOC_MMUART0 ,MICROCHIP_PFSOC_SYSREG ,MICROCHIP_PFSOC_MPUCFG , +MICROCHIP_PFSOC_MMUART1 , +MICROCHIP_PFSOC_MMUART2 , +MICROCHIP_PFSOC_MMUART3 , +MICROCHIP_PFSOC_MMUART4 ,MICROCHIP_PFSOC_ENVM_CFG ,MICROCHIP_PFSOC_ENVM_DATA ,MICROCHIP_PFSOC_IOSCB_CFG , MICROCHIP_PFSOC_DRAM, }; +enum { +MICROCHIP_PFSOC_MMUART0_IRQ =90 , +MICROCHIP_PFSOC_MMUART1_IRQ =91 , +MICROCHIP_PFSOC_MMUART2_IRQ =92 , +MICROCHIP_PFSOC_MMUART3_IRQ =93 , +MICROCHIP_PFSOC_MMUART4_IRQ =94 , +}; +#define MICROCHIP_PFSOC_MANAGEMENT_CPU_COUNT 1#define MICROCHIP_PFSOC_COMPUTE_CPU_COUNT 4 --2.7.4
PATCH 7/18 hw/sd: sd: Fix incorrect populated function switch status data structure
In reply to Bin MengFrom: Bin Meng <[email protected]>
At present the function switch status data structure bit [399:376] are wrongly pupulated. These 3 bytes encode function switch status for the 6 function groups, with 4 bits per group, starting from function group 6 at bit 399, then followed by function group 5 at bit 395, and so on.
However the codes mistakenly fills in the function group 1 status at bit 399. This fixes the code logic.
---
hw/sd/sd.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
hw/sd/sd.c +5 -1
@@ -806,11 +806,15 @@ static void sd_function_switch(SDState *sd, uint32_t arg) sd->data[11 ] =0x43 ; sd->data[12 ] =0x80 ;/* Supported group 1 functions */ sd->data[13 ] =0x03 ; + + sd->data[14 ] =0 ; + sd->data[15 ] =0 ; + sd->data[16 ] =0 ;for (i =0 ; i <6 ; i ++) { new_func = (arg >> (i *4 )) &0x0f ;if (mode && new_func !=0x0f ) sd->function_group[i] = new_func; - sd->data[14 + (i >>1 )] = new_func << ((i *4 ) &4 ); + sd->data[16 - (i >>1 )] |= new_func << ((i %2 ) *4 ); }memset (&sd->data[17 ],0 ,47 );stw_be_p (sd->data +64 ,sd_crc16 (sd->data,64 ));
Signature · 1 line
2.7.4
Re: [PATCH 07/18] hw/sd: sd: Fix incorrect populated function switch status data structure
+Sai Pavan
Quoted 11 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the function switch status data structure bit [399:376] are wrongly pupulated. These 3 bytes encode function switch status for the 6 function groups, with 4 bits per group, starting from function group 6 at bit 399, then followed by function group 5 at bit 395, and so on.
However the codes mistakenly fills in the function group 1 status at bit 399. This fixes the code logic.
Quoted 11 lines
Signed-off-by Bin Meng <[email protected]>---
hw/sd/sd.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)hw/sd/sd.c +0
-0
hw/sd/sd.c static void sd_function_switch(SDState *sd, uint32_t arg) · line 812
Quoted 7 lines
hw/sd/sd.c +4
-0@@ -806,11 +806,15 @@ static void sd_function_switch(SDState *sd, uint32_t arg) sd->data[11 ] =0x43 ; sd->data[12 ] =0x80 ;/* Supported group 1 functions */ sd->data[13 ] =0x03 ; + + sd->data[14 ] =0 ; + sd->data[15 ] =0 ; + sd->data[16 ] =0 ;
Pointless zero initialization.
I'll wait until next week in case Sai Pavan wants to test this patch (I don't have access to the Xilinx images anymore) then I'll queue this via my sd-next tree.
Regards,
Phil.
Quoted 10 lines
hw/sd/sd.c +1
-1for (i =0 ; i <6 ; i ++) { new_func = (arg >> (i *4 )) &0x0f ;if (mode && new_func !=0x0f ) sd->function_group[i] = new_func; - sd->data[14 + (i >>1 )] = new_func << ((i *4 ) &4 ); + sd->data[16 - (i >>1 )] |= new_func << ((i %2 ) *4 ); }memset (&sd->data[17 ],0 ,47 );stw_be_p (sd->data +64 ,sd_crc16 (sd->data,64 ));
RE: [PATCH 07/18] hw/sd: sd: Fix incorrect populated function switch status data structure
Thanks Philippe,
I would test this and get back.
Regards,
Sai Pavan
Quoted 70 lines
-----Original Message-----
From: Philippe Mathieu-Daudé <[email protected]> On Behalf Of Philippe Mathieu-Daudé
Sent: Saturday, August 15, 2020 1:29 PM
To: Bin Meng <[email protected]>; Alistair Francis <[email protected]>; Bastian Koppelmann <[email protected] paderborn.de>; Palmer Dabbelt <[email protected]>; Sagar Karandikar <[email protected]>; [email protected]; qemu- [email protected]; Sai Pavan Boddu <[email protected]>
Cc: Bin Meng <[email protected]>; [email protected]
Subject: Re: [PATCH 07/18] hw/sd: sd: Fix incorrect populated function switch status data structure+Sai Pavan
Quoted 10 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the function switch status data structure bit [399:376] are wrongly pupulated. These 3 bytes encode function switch status for the 6 function groups, with 4 bits per group, starting from function group 6 at bit 399, then followed by function group 5 at bit 395, and so on.
However the codes mistakenly fills in the function group 1 status at bit 399. This fixes the code logic.
Fixes a1bb27b1e9 ("SD card emulation (initial implementation)")hw/sd/sd.c static void sd_function_switch(SDState *sd,
Quoted 11 lines
Signed-off-by Bin Meng <[email protected]>---
hw/sd/sd.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)hw/sd/sd.c +0
-0@@ -806,11 +806,15 @@ static void sd_function_switch(SDState *sd,
uint32_targ)hw/sd/sd.c static void sd_function_switch(SDState *sd, uint32_t arg) · line 812
Quoted 7 lines
hw/sd/sd.c +4
-0sd->data[11 ] =0x43 ; sd->data[12 ] =0x80 ;/* Supported group 1 functions */ sd->data[13 ] =0x03 ; + + sd->data[14 ] =0 ; + sd->data[15 ] =0 ; + sd->data[16 ] =0 ;Pointless zero initialization.
Reviewed-by Philippe Mathieu-Daudé <[email protected]>I'll wait until next week in case Sai Pavan wants to test this patch (I don't have access to the Xilinx images anymore) then I'll queue this via my sd-next tree.
Regards,
Phil.
Quoted 10 lines
hw/sd/sd.c +1
-1for (i =0 ; i <6 ; i ++) { new_func = (arg >> (i *4 )) &0x0f ;if (mode && new_func !=0x0f ) sd->function_group[i] = new_func; - sd->data[14 + (i >>1 )] = new_func << ((i *4 ) &4 ); + sd->data[16 - (i >>1 )] |= new_func << ((i %2 ) *4 ); }memset (&sd->data[17 ],0 ,47 );stw_be_p (sd->data +64 ,sd_crc16 (sd->data,64 ));
Re: [PATCH 07/18] hw/sd: sd: Fix incorrect populated function switch status data structure
Hi Philippe,
First two patch of SD look good. Tested them over zynqmp and versal platforms.
Thanks,
Sai Pavan
Quoted 78 lines · Sai Pavan Boddu
Thanks Philippe,
I would test this and get back.
Regards,
Sai PavanQuoted 70 lines
-----Original Message-----
From: Philippe Mathieu-Daudé <[email protected]> On Behalf Of Philippe Mathieu-Daudé
Sent: Saturday, August 15, 2020 1:29 PM
To: Bin Meng <[email protected]>; Alistair Francis <[email protected]>; Bastian Koppelmann <[email protected] paderborn.de>; Palmer Dabbelt <[email protected]>; Sagar Karandikar <[email protected]>; [email protected]; qemu- [email protected]; Sai Pavan Boddu <[email protected]>
Cc: Bin Meng <[email protected]>; [email protected]
Subject: Re: [PATCH 07/18] hw/sd: sd: Fix incorrect populated function switch status data structure+Sai Pavan
Quoted 10 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the function switch status data structure bit [399:376] are wrongly pupulated. These 3 bytes encode function switch status for the 6 function groups, with 4 bits per group, starting from function group 6 at bit 399, then followed by function group 5 at bit 395, and so on.
However the codes mistakenly fills in the function group 1 status at bit 399. This fixes the code logic.
Fixes a1bb27b1e9 ("SD card emulation (initial implementation)")hw/sd/sd.c static void sd_function_switch(SDState *sd,
Quoted 11 lines
Signed-off-by Bin Meng <[email protected]>---
hw/sd/sd.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)hw/sd/sd.c +0
-0@@ -806,11 +806,15 @@ static void sd_function_switch(SDState *sd,
uint32_targ)hw/sd/sd.c static void sd_function_switch(SDState *sd, uint32_t arg) · line 812
Quoted 7 lines
hw/sd/sd.c +4
-0sd->data[11 ] =0x43 ; sd->data[12 ] =0x80 ;/* Supported group 1 functions */ sd->data[13 ] =0x03 ; + + sd->data[14 ] =0 ; + sd->data[15 ] =0 ; + sd->data[16 ] =0 ;Pointless zero initialization.
Reviewed-by Philippe Mathieu-Daudé <[email protected]>I'll wait until next week in case Sai Pavan wants to test this patch (I don't have access to the Xilinx images anymore) then I'll queue this via my sd-next tree.
Regards,
Phil.
Quoted 10 lines
hw/sd/sd.c +1
-1for (i =0 ; i <6 ; i ++) { new_func = (arg >> (i *4 )) &0x0f ;if (mode && new_func !=0x0f ) sd->function_group[i] = new_func; - sd->data[14 + (i >>1 )] = new_func << ((i *4 ) &4 ); + sd->data[16 - (i >>1 )] |= new_func << ((i %2 ) *4 ); }memset (&sd->data[17 ],0 ,47 );stw_be_p (sd->data +64 ,sd_crc16 (sd->data,64 ));
Re: [PATCH 07/18] hw/sd: sd: Fix incorrect populated function switch status data structure
Hi Sai,
Quoted 6 lines · Sai Pavan Boddu
Hi Philippe,
First two patch of SD look good. Tested them over zynqmp and versal platforms.
Thanks for testing.
Can I add your Tested-by tag?
Regards,
Bin
RE: [PATCH 07/18] hw/sd: sd: Fix incorrect populated function switch status data structure
Hi Bin,
Quoted 20 lines
-----Original Message-----
From: Bin Meng <[email protected]>
Sent: Friday, August 21, 2020 3:38 PM
To: Sai Pavan Boddu <[email protected]>
Cc: Philippe Mathieu-Daudé <[email protected]>; Alistair Francis <[email protected]>; Bastian Koppelmann <[email protected] paderborn.de>; Palmer Dabbelt <[email protected]>; Sagar Karandikar <[email protected]>; [email protected]; qemu- [email protected]; Bin Meng <[email protected]>; qemu- [email protected]; Sai Pavan Boddu <[email protected]>
Subject: Re: [PATCH 07/18] hw/sd: sd: Fix incorrect populated function switch status data structureHi Sai,
Quoted 3 lines · Sai Pavan Boddu
Hi Philippe,
Quoted 7 lines
Quoted 3 lines
First two patch of SD look good. Tested them over zynqmp and versal platforms.
Thanks for testing.
Can I add your Tested-by tag?
[Sai Pavan Boddu] Sure. Thanks.
Regards,
Sai Pavan
Quoted 3 lines
Regards,
Bin
PATCH 8/18 hw/sd: sd: Correctly set the high capacity bit
In reply to Bin MengFrom: Bin Meng <[email protected]>
Per the SD spec, Standard Capacity SD Memory Card (SDSC) supports capacity up to and including 2 GiB.
---
hw/sd/sd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
hw/sd/sd.c +1 -1
@@ -313,7 +313,7 @@ static void sd_ocr_powerup(void *opaque)/* card power-up OK */ sd->ocr =FIELD_DP32 (sd->ocr,OCR ,CARD_POWER_UP ,1 ); -if (sd->size >1 * GiB) { +if (sd->size >2 * GiB) { sd->ocr =FIELD_DP32 (sd->ocr,OCR ,CARD_CAPACITY ,1 ); } }
Signature · 1 line
2.7.4
Re: [PATCH 08/18] hw/sd: sd: Correctly set the high capacity bit
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
Per the SD spec, Standard Capacity SD Memory Card (SDSC) supports capacity up to and including 2 GiB.
Quoted 11 lines
Signed-off-by Bin Meng <[email protected]>---
hw/sd/sd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)hw/sd/sd.c +0
-0
hw/sd/sd.c static void sd_ocr_powerup(void *opaque) · line 316
Quoted 5 lines
hw/sd/sd.c +1
-1@@ -313,7 +313,7 @@ static void sd_ocr_powerup(void *opaque)/* card power-up OK */ sd->ocr =FIELD_DP32 (sd->ocr,OCR ,CARD_POWER_UP ,1 ); -if (sd->size >1 * GiB) { +if (sd->size >2 * GiB) {
But you need to fix sd_set_csd() too, else this is incomplete.
Thanks,
Phil.
Quoted 4 lines
hw/sd/sd.c +0
-0sd->ocr =FIELD_DP32 (sd->ocr,OCR ,CARD_CAPACITY ,1 ); } }
Re: [PATCH 08/18] hw/sd: sd: Correctly set the high capacity bit
Hi Philippe,
Quoted 24 lines · Philippe Mathieu-Daudé
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
Per the SD spec, Standard Capacity SD Memory Card (SDSC) supports capacity up to and including 2 GiB.
Fixes 2d7adea4fe ("hw/sd: Support SDHC size cards")Quoted 14 lines
Signed-off-by Bin Meng <[email protected]>---
hw/sd/sd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)hw/sd/sd.c +0
-0@@ -313,7 +313,7 @@ static void sd_ocr_powerup(void *opaque)/* card power-up OK */ sd->ocr =FIELD_DP32 (sd->ocr,OCR ,CARD_POWER_UP ,1 );
Quoted 5 lines
hw/sd/sd.c static void sd_ocr_powerup(void *opaque) · line 316
Quoted 2 lines
hw/sd/sd.c +1
-1-if (sd->size >1 * GiB) { +if (sd->size >2 * GiB) {But you need to fix
sd_set_csd()too, else this is incomplete.
Ah yes, I missed that. Will fix in v2. Thanks for the review.
Regards,
Bin
PATCH 9/18 hw/sd: sdhci: Make sdhci_poweron_reset() internal visible
In reply to Bin MengFrom: Bin Meng <[email protected]>
sdhci_poweron_reset() might be needed for any SDHCI compatible device that is built on top of the generic SDHCI device.
---
hw/sd/sdhci-internal.h | 1 + hw/sd/sdhci.c | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-)
hw/sd/sdhci-internal.h +1 -0
@@ -342,5 +342,6 @@ void sdhci_uninitfn(SDHCIState *s);void sdhci_common_realize (SDHCIState *s,Error **errp);void sdhci_common_unrealize (SDHCIState *s);void sdhci_common_class_init (ObjectClass *klass,void *data); +void sdhci_poweron_reset (DeviceState *dev);#endif
hw/sd/sdhci.c +1 -1
@@ -294,7 +294,7 @@ static void sdhci_reset(SDHCIState *s) s->pending_insert_state = false; } -static void sdhci_poweron_reset (DeviceState *dev) +void sdhci_poweron_reset (DeviceState *dev) {/* QOM (ie power-on) reset. This is identical to reset * commanded via device register apart from handling of the
Signature · 1 line
2.7.4
Re: [PATCH 09/18] hw/sd: sdhci: Make sdhci_poweron_reset() internal visible
Quoted 4 lines · Bin Meng
From: Bin Meng <[email protected]>
sdhci_poweron_reset()might be needed for any SDHCI compatible device that is built on top of the generic SDHCI device.
NAck. Please use device_legacy_reset() instead.
In next patch:
device_legacy_reset(DEVICE(&sdhci->slot));
Thanks,
Phil.
Quoted 33 lines
Signed-off-by Bin Meng <[email protected]>---
hw/sd/sdhci-internal.h | 1 + hw/sd/sdhci.c | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-)hw/sd/sdhci-internal.h +1
-0@@ -342,5 +342,6 @@ void sdhci_uninitfn(SDHCIState *s);void sdhci_common_realize (SDHCIState *s,Error **errp);void sdhci_common_unrealize (SDHCIState *s);void sdhci_common_class_init (ObjectClass *klass,void *data); +void sdhci_poweron_reset (DeviceState *dev);#endif hw/sd/sdhci.c +1
-1@@ -294,7 +294,7 @@ static void sdhci_reset(SDHCIState *s) s->pending_insert_state = false; } -static void sdhci_poweron_reset (DeviceState *dev) +void sdhci_poweron_reset (DeviceState *dev) {/* QOM (ie power-on) reset. This is identical to reset * commanded via device register apart from handling of the
Re: [PATCH 09/18] hw/sd: sdhci: Make sdhci_poweron_reset() internal visible
Hi Philippe,
Quoted 4 lines · Philippe Mathieu-Daudé
Quoted 2 lines · Bin Meng
From: Bin Meng <[email protected]>
Quoted 9 lines
Quoted 2 lines
sdhci_poweron_reset()might be needed for any SDHCI compatible device that is built on top of the generic SDHCI device.NAck. Please use
device_legacy_reset()instead.In next patch:
device_legacy_reset(DEVICE(&sdhci->slot));
The function comments say this API is deprecated.
/**
* device_legacy_reset:
*
* Reset a single device (by calling the reset method).
* Note: This function is deprecated and will be removed when it becomes unused.
* Please use device_cold_reset() now.
*/
Should we use device_cold_reset() here?
Regards,
Bin
Re: [PATCH 09/18] hw/sd: sdhci: Make sdhci_poweron_reset() internal visible
Quoted 27 lines · Bin Meng
Hi Philippe,
Quoted 4 lines · Philippe Mathieu-Daudé
Quoted 2 lines · Bin Meng
From: Bin Meng <[email protected]>
Quoted 9 lines
Quoted 2 lines
sdhci_poweron_reset()might be needed for any SDHCI compatible device that is built on top of the generic SDHCI device.NAck. Please use
device_legacy_reset()instead.In next patch:
device_legacy_reset(DEVICE(&sdhci->slot));The function comments say this API is deprecated.
/**
*device_legacy_reset:
*
* Reset a single device (by calling the reset method).
* Note: This function is deprecated and will be removed when it becomes unused.
* Please usedevice_cold_reset()now.
*/
Quoted 1 line
Should we use
device_cold_reset()here?
Oops yes, certainly.
Quoted 4 lines
Regards,
Bin
PATCH 10/18 hw/sd: Add Cadence SDHCI emulation
In reply to Bin MengFrom: Bin Meng <[email protected]>
Cadence SD/SDIO/eMMC Host Controller (SD4HC) is an SDHCI compatible controller. The SDHCI compatible registers start from offset 0x200, which are called Slot Register Set (SRS) in its datasheet.
This creates a Cadence SDHCI model built on top of the existing generic SDHCI model. Cadence specific Host Register Set (HRS) is implemented to make guest software happy.
---
hw/sd/Kconfig | 4 ++ hw/sd/Makefile.objs | 1 + hw/sd/cadence_sdhci.c | 162 ++++++++++++++++++++++++++++++++++++++++++ include/hw/sd/cadence_sdhci.h | 65 +++++++++++++++++ 4 files changed, 232 insertions(+) create mode 100644 hw/sd/cadence_sdhci.c create mode 100644 include/hw/sd/cadence_sdhci.h
hw/sd/Kconfig +4 -0
@@ -19,3 +19,7 @@ config SDHCI_PCI default y if PCI_DEVICES depends on PCI select SDHCI + +config CADENCE_SDHCI + bool + select SDHCI
hw/sd/Makefile.objs +1 -0
@@ -10,3 +10,4 @@ common-obj-$(CONFIG_OMAP) += omap_mmc.o common-obj-$(CONFIG_PXA2XX) += pxa2xx_mmci.o common-obj-$(CONFIG_RASPI) += bcm2835_sdhost.o common-obj-$(CONFIG_ASPEED_SOC) += aspeed_sdhci.o +common-obj-$(CONFIG_CADENCE_SDHCI) += cadence_sdhci.o
hw/sd/cadence_sdhci.c new file +162 -0
162 lines not on this page · raw message
include/hw/sd/cadence_sdhci.h new file +65 -0
65 lines not on this page · raw message
Signature · 1 line
2.7.4
Re: [PATCH 10/18] hw/sd: Add Cadence SDHCI emulation
Quoted 13 lines · Bin Meng
From: Bin Meng <[email protected]>
Cadence SD/SDIO/eMMC Host Controller (SD4HC) is an SDHCI compatible controller. The SDHCI compatible registers start from offset
0x200, which are called Slot Register Set (SRS) in its datasheet.This creates a Cadence SDHCI model built on top of the existing generic SDHCI model. Cadence specific Host Register Set (HRS) is implemented to make guest software happy.
Signed-off-by Bin Meng <[email protected]>---
Quoted 7 lines
hw/sd/Kconfig | 4 ++ hw/sd/Makefile.objs | 1 + hw/sd/cadence_sdhci.c | 162 ++++++++++++++++++++++++++++++++++++++++++ include/hw/sd/cadence_sdhci.h | 65 +++++++++++++++++ 4 files changed, 232 insertions(+) create mode 100644 hw/sd/cadence_sdhci.c create mode 100644 include/hw/sd/cadence_sdhci.h
Please consider using scripts/git.orderfile.
Quoted 54 lines
hw/sd/Kconfig +4
-0@@ -19,3 +19,7 @@ config SDHCI_PCI default y if PCI_DEVICES depends on PCI select SDHCI + +config CADENCE_SDHCI + bool + select SDHCIhw/sd/Makefile.objs +1
-0@@ -10,3 +10,4 @@ common-obj-$(CONFIG_OMAP) += omap_mmc.o common-obj-$(CONFIG_PXA2XX) += pxa2xx_mmci.o common-obj-$(CONFIG_RASPI) += bcm2835_sdhost.o common-obj-$(CONFIG_ASPEED_SOC) += aspeed_sdhci.o +common-obj-$(CONFIG_CADENCE_SDHCI) += cadence_sdhci.ohw/sd/cadence_sdhci.c new file +26
-0@@ -0,0 +1,162 @@ +/* +* Cadence SDHCI emulation +* +* Copyright (c) 2020 Wind River Systems, Inc. +* +* Author: +* Bin Meng <[email protected]> +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU General Public License as +* published by the Free Software Foundation; either version 2 or +* (at your option) version 3 of the License. +* +* This program is distributed in the hope that it will be useful, +* but WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +* GNU General Public License for more details. +* +* You should have received a copy of the GNU General Public License along +* with this program; if not, see <http://www.gnu.org/licenses/>. +*/ + +#include "qemu/osdep.h" +#include "qemu/error-report.h" +#include "qemu/log.h" +#include "qapi/error.h"
hw/sd/cadence_sdhci.c line 36
Quoted 10 lines
hw/sd/cadence_sdhci.c new file +10
-0+#include "migration/vmstate.h" +#include "hw/irq.h" +#include "hw/sd/cadence_sdhci.h" +#include "sdhci-internal.h" + +#define TO_REG (addr) ((addr) / sizeof(uint32_t)) + +static void cadence_sdhci_reset (DeviceState *dev) +{ +CadenceSDHCIState *sdhci =CADENCE_SDHCI (dev);
We usually use 's' for the model state.
hw/sd/cadence_sdhci.c line 39
Quoted 3 lines
hw/sd/cadence_sdhci.c new file +3
-0+ +memset (sdhci->regs,0 ,CADENCE_SDHCI_REG_SIZE ); + sdhci->regs[TO_REG (SDHCI_CDNS_HRS00 )] =SDHCI_CDNS_HRS00_POR_VAL ;
As commented in previous patch, here goes:
device_legacy_reset(DEVICE(&sdhci->slot));
Quoted 23 lines
hw/sd/cadence_sdhci.c new file +23
-0+} + +static uint64_t cadence_sdhci_read (void *opaque,hwaddr addr,unsigned int size) +{ +uint32_t val =0 ; +CadenceSDHCIState *sdhci = opaque; + +if (addr <CADENCE_SDHCI_REG_SIZE ) { + val = sdhci->regs[TO_REG (addr)]; + }else { +qemu_log_mask (LOG_GUEST_ERROR , +"%s: Out-of-bounds read at 0x%" HWADDR_PRIx"\n" , + __func__, addr); + } + +return (uint64_t )val; +} + +static void cadence_sdhci_write (void *opaque,hwaddr addr,uint64_t val, +unsigned int size) +{ +CadenceSDHCIState *sdhci = opaque; +uint32_t val32 = (uint32_t )val;
hw/sd/cadence_sdhci.c line 72
Quoted 10 lines
hw/sd/cadence_sdhci.c new file +10
-0+ +switch (addr) { +case SDHCI_CDNS_HRS00 : +/* +* The only writable bit is SWR (software reset) and it automatically +* clears to zero, so essentially this register remains unchanged. +*/ +if (val32 &SDHCI_CDNS_HRS00_SWR ) { +cadence_sdhci_reset (DEVICE (sdhci)); +sdhci_poweron_reset (DEVICE (&sdhci->slot));
The 'slot' reset has to go in cadence_sdhci_reset().
Quoted 28 lines
hw/sd/cadence_sdhci.c new file +28
-0+ } + +break ; +case SDHCI_CDNS_HRS04 : +/* +* Only emulate the ACK bit behavior when read or write transaction +* are requested. +*/ +if (val32 & (SDHCI_CDNS_HRS04_WR |SDHCI_CDNS_HRS04_RD )) { + val32 |=SDHCI_CDNS_HRS04_ACK ; + }else { + val32 &= ~SDHCI_CDNS_HRS04_ACK ; + } + + sdhci->regs[TO_REG (addr)] = val32; +break ; +case SDHCI_CDNS_HRS06 : +if (val32 &SDHCI_CDNS_HRS06_TUNE_UP ) { + val32 &= ~SDHCI_CDNS_HRS06_TUNE_UP ; + } + + sdhci->regs[TO_REG (addr)] = val32; +break ; +default : +if (addr <CADENCE_SDHCI_REG_SIZE ) { + sdhci->regs[TO_REG (addr)] = val32; + }else { + qemu_log_mask(LOG_GUEST_ERROR,
hw/sd/cadence_sdhci.c line 110
Quoted 10 lines
hw/sd/cadence_sdhci.c new file +10
-0+"%s: Out-of-bounds write at 0x%" HWADDR_PRIx"\n" , + __func__, addr); + } + } +} + +static const MemoryRegionOps cadence_sdhci_ops = { + .read = cadence_sdhci_read, + .write = cadence_sdhci_write, + .endianness =DEVICE_NATIVE_ENDIAN ,
To make it obvious, please also add:
.impl = {
.min_access_size = 4,
.max_access_size = 4,
},hw/sd/cadence_sdhci.c line 121
Quoted 11 lines
hw/sd/cadence_sdhci.c new file +11
-0+ .valid = { + .min_access_size = 4, + .max_access_size = 4, + } +}; + +static void cadence_sdhci_realize (DeviceState *dev,Error **errp) +{ +SysBusDevice *sbd =SYS_BUS_DEVICE (dev); +CadenceSDHCIState *sdhci =CADENCE_SDHCI (dev); +SysBusDevice *sbd_slot =SYS_BUS_DEVICE (&sdhci->slot);
Where is sdhci->slot initialized??
hw/sd/cadence_sdhci.c line 129
Quoted 8 lines
hw/sd/cadence_sdhci.c new file +8
-0+ +memory_region_init_io (&sdhci->iomem,OBJECT (sdhci), &cadence_sdhci_ops, + sdhci,TYPE_CADENCE_SDHCI ,0x1000 ); +sysbus_init_mmio (sbd, &sdhci->iomem); + +sysbus_realize (sbd_slot, errp); +memory_region_add_subregion (&sdhci->iomem,SDHCI_CDNS_SRS_BASE , + &sdhci->slot.iomem);
You should not access private field that way, you have to use the public interface:
memory_region_add_subregion(&sdhci->iomem, SDHCI_CDNS_SRS_BASE,
sysbus_mmio_get_region(SYS_BUS_DEVICE(&sdhci->slot), 0));To have this model useful, you need to propagate at least the 'sd-bus' and 'dma' properties from the 'slot', else how the TYPE_CADENCE_SDHCI user is supposed to set them?
Quoted 73 lines
hw/sd/cadence_sdhci.c new file +33
-0+} + +static const VMStateDescription vmstate_cadence_sdhci = { + .name =TYPE_CADENCE_SDHCI , + .version_id =1 , + .fields = (VMStateField []) { +VMSTATE_UINT32_ARRAY (regs, CadenceSDHCIState,CADENCE_SDHCI_NUM_REGS ), +VMSTATE_END_OF_LIST (), + }, +}; + +static void cadence_sdhci_class_init (ObjectClass *classp,void *data) +{ +DeviceClass *dc =DEVICE_CLASS (classp); + + dc->realize = cadence_sdhci_realize; + dc->reset = cadence_sdhci_reset; + dc->vmsd = &vmstate_cadence_sdhci; +} + +static TypeInfo cadence_sdhci_info = { + .name =TYPE_CADENCE_SDHCI , + .parent =TYPE_SYS_BUS_DEVICE , + .instance_size =sizeof (CadenceSDHCIState), + .class_init = cadence_sdhci_class_init, +}; + +static void cadence_sdhci_register_types (void ) +{ +type_register_static (&cadence_sdhci_info); +} + +type_init (cadence_sdhci_register_types)include/hw/sd/cadence_sdhci.h new file +34
-0@@ -0,0 +1,65 @@ +/* +* Cadence SDHCI emulation +* +* Copyright (c) 2020 Wind River Systems, Inc. +* +* Author: +* Bin Meng <[email protected]> +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU General Public License as +* published by the Free Software Foundation; either version 2 or +* (at your option) version 3 of the License. +* +* This program is distributed in the hope that it will be useful, +* but WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +* GNU General Public License for more details. +* +* You should have received a copy of the GNU General Public License along +* with this program; if not, see <http://www.gnu.org/licenses/>. +*/ + +#ifndef CADENCE_SDHCI_H +#define CADENCE_SDHCI_H + +#include "qemu/bitops.h" +#include "hw/sd/sdhci.h" + +/* HRS - Host Register Set (specific to Cadence) */ + +#define SDHCI_CDNS_HRS00 0x00/* general information */ +#define SDHCI_CDNS_HRS00_SWR BIT(0) +#define SDHCI_CDNS_HRS00_POR_VAL 0x00010000 +
include/hw/sd/cadence_sdhci.h line 44
Quoted 10 lines
include/hw/sd/cadence_sdhci.h new file +10
-0+#define SDHCI_CDNS_HRS04 0x10/* PHY access port */ +#define SDHCI_CDNS_HRS04_WR BIT(24) +#define SDHCI_CDNS_HRS04_RD BIT(25) +#define SDHCI_CDNS_HRS04_ACK BIT(26) + +#define SDHCI_CDNS_HRS06 0x18/* eMMC control */ +#define SDHCI_CDNS_HRS06_TUNE_UP BIT(15) + +/* SRS - Slot Register Set (SDHCI-compatible) */ +#define SDHCI_CDNS_SRS_BASE 0x200
Please keep the previous definitions local to cadence_sdhci.c.
Quoted 4 lines
include/hw/sd/cadence_sdhci.h new file +4
-0+ +#define CADENCE_SDHCI_CAPABILITIES 0x01E80080 +#define CADENCE_SDHCI_REG_SIZE 0x100 +#define CADENCE_SDHCI_NUM_REGS (CADENCE_SDHCI_REG_SIZE / sizeof(uint32_t))
include/hw/sd/cadence_sdhci.h line 58
Quoted 10 lines
include/hw/sd/cadence_sdhci.h new file +10
-0+ +typedef struct CadenceSDHCIState { + SysBusDevice parent; + + MemoryRegion iomem; + qemu_irq irq; + + uint32_t regs[CADENCE_SDHCI_NUM_REGS]; + + SDHCIState slot;
Why name this 'slot' and not 'sdhci' or 'generic_sdhci'?
Thanks,
Phil.
Quoted 8 lines
include/hw/sd/cadence_sdhci.h new file +7
-0+} CadenceSDHCIState; + +#define TYPE_CADENCE_SDHCI "cadence.sdhci" +#define CADENCE_SDHCI (obj) OBJECT_CHECK(CadenceSDHCIState, (obj), \ + TYPE_CADENCE_SDHCI) + +#endif /* CADENCE_SDHCI_H */
PATCH 11/18 hw/riscv: microchip_pfsoc: Connect a Cadence SDHCI controller and an SD card
In reply to Bin MengFrom: Bin Meng <[email protected]>
Microchip PolarFire SoC integrates one Cadence SDHCI controller. On the Icicle Kit board, one eMMC chip and an external SD card connect to this controller depending on different configuration.
As QEMU does not support eMMC yet, we just emulate the SD card configuration. To test this, the Hart Software Services (HSS) should choose the SD card configuration:
$ cp boards/icicle-kit-es/def_config.sdcard .config $ make BOARD=icicle-kit-es
The SD card image can be built from the Yocto BSP at:
https://github.com/polarfire-soc/meta-polarfire-soc-yocto-bsp
Note the generated SD card image should be resized before use:
$ qemu-img resize /path/to/sdcard.img 4G
Launch QEMU with the following command:
$ qemu-system-riscv64 -nographic -M microchip-icicle-kit -sd sdcard.img
---
hw/riscv/Kconfig | 1 + hw/riscv/microchip_pfsoc.c | 26 ++++++++++++++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 4 ++++ 3 files changed, 31 insertions(+)
hw/riscv/Kconfig +1 -0
@@ -55,3 +55,4 @@ config MICROCHIP_PFSOC select SIFIVE select UNIMP select MCHP_PFSOC_MMUART + select CADENCE_SDHCI
hw/riscv/microchip_pfsoc.c +26 -0
@@ -12,6 +12,7 @@* 1) PLIC (Platform Level Interrupt Controller) * 2) eNVM (Embedded Non-Volatile Memory) * 3) MMUARTs (Multi-Mode UART) +* 4) Cadence eMMC/SDHC controller and an SD card connected to it * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts.
@@ -75,6 +76,7 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_MMUART0] = { 0x20000000, 0x1000 }, [MICROCHIP_PFSOC_SYSREG] = { 0x20002000, 0x2000 }, [MICROCHIP_PFSOC_MPUCFG] = { 0x20005000, 0x1000 }, + [MICROCHIP_PFSOC_EMMC_SD] = { 0x20008000, 0x1000 }, [MICROCHIP_PFSOC_MMUART1] = { 0x20100000, 0x1000 }, [MICROCHIP_PFSOC_MMUART2] = { 0x20102000, 0x1000 }, [MICROCHIP_PFSOC_MMUART3] = { 0x20104000, 0x1000 },
@@ -111,6 +113,11 @@ static void microchip_pfsoc_soc_instance_init(Object *obj)qdev_prop_set_string (DEVICE (&s->u_cpus),"cpu-type" ,TYPE_RISCV_CPU_SIFIVE_U54 );qdev_prop_set_uint64 (DEVICE (&s->u_cpus),"resetvec" ,RESET_VECTOR ); + +object_initialize_child (obj,"sd-controller" , &s->sdhci, +TYPE_CADENCE_SDHCI ); +object_initialize_child (OBJECT (&s->sdhci),"sd-controller.sdhci" , + &s->sdhci.slot,TYPE_SYSBUS_SDHCI ); }static void microchip_pfsoc_soc_realize (DeviceState *dev,Error **errp)
@@ -223,6 +230,13 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp) memmap[MICROCHIP_PFSOC_MPUCFG ].base, memmap[MICROCHIP_PFSOC_MPUCFG].size); +/* SDHCI */ +sysbus_realize (SYS_BUS_DEVICE (&s->sdhci), errp); +sysbus_mmio_map (SYS_BUS_DEVICE (&s->sdhci),0 , + memmap[MICROCHIP_PFSOC_EMMC_SD ].base); +sysbus_connect_irq (SYS_BUS_DEVICE (&s->sdhci.slot),0 , +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_EMMC_SD_IRQ )); +/* MMUARTs */ s->serial0 = mchp_pfsoc_mmuart_create(system_memory, memmap[MICROCHIP_PFSOC_MMUART0].base,
@@ -290,6 +304,7 @@ static void microchip_icicle_kit_machine_init(MachineState *machine)MicrochipIcicleKitState *s =MICROCHIP_ICICLE_KIT_MACHINE (machine);MemoryRegion *system_memory =get_system_memory ();MemoryRegion *main_mem =g_new (MemoryRegion,1 ); +DriveInfo *dinfo =drive_get_next (IF_SD );/* Sanity check on RAM size */ if (machine->ram_size < mc->default_ram_size) {
@@ -312,6 +327,17 @@ static void microchip_icicle_kit_machine_init(MachineState *machine)/* Load the firmware */ riscv_find_and_load_firmware (machine,BIOS_FILENAME ,RESET_VECTOR ,NULL ); + +/* Attach an SD card */ +if (dinfo) { +SDHCIState *sdhci = &(s->soc.sdhci.slot); +DeviceState *card =qdev_new (TYPE_SD_CARD ); +BusState *bus =qdev_get_child_bus (DEVICE (sdhci),"sd-bus" ); + +qdev_prop_set_drive_err (card,"drive" ,blk_by_legacy_dinfo (dinfo), + &error_fatal); +qdev_realize_and_unref (card, bus, &error_fatal); + } }static void microchip_icicle_kit_machine_class_init (ObjectClass *oc,void *data)
include/hw/riscv/microchip_pfsoc.h +4 -0
@@ -23,6 +23,7 @@#define HW_MICROCHIP_PFSOC_H #include "hw/char/mchp_pfsoc_mmuart.h" +#include "hw/sd/cadence_sdhci.h" typedef struct MicrochipPFSoCState { /*< private >*/
@@ -39,6 +40,7 @@ typedef struct MicrochipPFSoCState {MchpPfSoCMMUartState *serial2;MchpPfSoCMMUartState *serial3;MchpPfSoCMMUartState *serial4; +CadenceSDHCIState sdhci; } MicrochipPFSoCState;#define TYPE_MICROCHIP_PFSOC "microchip.pfsoc"
@@ -74,6 +76,7 @@ enum {MICROCHIP_PFSOC_MMUART0 ,MICROCHIP_PFSOC_SYSREG ,MICROCHIP_PFSOC_MPUCFG , +MICROCHIP_PFSOC_EMMC_SD ,MICROCHIP_PFSOC_MMUART1 ,MICROCHIP_PFSOC_MMUART2 , MICROCHIP_PFSOC_MMUART3,
@@ -85,6 +88,7 @@ enum { };enum { +MICROCHIP_PFSOC_EMMC_SD_IRQ =88 ,MICROCHIP_PFSOC_MMUART0_IRQ =90 ,MICROCHIP_PFSOC_MMUART1_IRQ =91 ,MICROCHIP_PFSOC_MMUART2_IRQ =92 ,
Signature · 1 line
2.7.4
Re: [PATCH 11/18] hw/riscv: microchip_pfsoc: Connect a Cadence SDHCI controller and an SD card
Quoted 60 lines · Bin Meng
From: Bin Meng <[email protected]>
Microchip PolarFire SoC integrates one Cadence SDHCI controller. On the Icicle Kit board, one eMMC chip and an external SD card connect to this controller depending on different configuration.
As QEMU does not support eMMC yet, we just emulate the SD card configuration. To test this, the Hart Software Services (HSS) should choose the SD card configuration:
$ cp boards/icicle-kit-es/def_config.sdcard .config $ make BOARD=icicle-kit-esThe SD card image can be built from the Yocto BSP at:
https://github.com/polarfire-soc/meta-polarfire-soc-yocto-bspNote the generated SD card image should be resized before use:
$ qemu-img resize /path/to/sdcard.img 4GLaunch QEMU with the following command:
$ qemu-system-riscv64 -nographic -M microchip-icicle-kit -sd sdcard.imgSigned-off-by Bin Meng <[email protected]>---
hw/riscv/Kconfig | 1 + hw/riscv/microchip_pfsoc.c | 26 ++++++++++++++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 4 ++++ 3 files changed, 31 insertions(+)hw/riscv/Kconfig +1
-0@@ -55,3 +55,4 @@ config MICROCHIP_PFSOC select SIFIVE select UNIMP select MCHP_PFSOC_MMUART + select CADENCE_SDHCIhw/riscv/microchip_pfsoc.c +2
-0@@ -12,6 +12,7 @@* 1) PLIC (Platform Level Interrupt Controller) * 2) eNVM (Embedded Non-Volatile Memory) * 3) MMUARTs (Multi-Mode UART) +* 4) Cadence eMMC/SDHC controller and an SD card connected to it * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts. @@ -75,6 +76,7 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_MMUART0] = { 0x20000000, 0x1000 }, [MICROCHIP_PFSOC_SYSREG] = { 0x20002000, 0x2000 }, [MICROCHIP_PFSOC_MPUCFG] = { 0x20005000, 0x1000 }, + [MICROCHIP_PFSOC_EMMC_SD] = { 0x20008000, 0x1000 }, [MICROCHIP_PFSOC_MMUART1] = { 0x20100000, 0x1000 }, [MICROCHIP_PFSOC_MMUART2] = { 0x20102000, 0x1000 }, [MICROCHIP_PFSOC_MMUART3] = { 0x20104000, 0x1000 },
hw/riscv/microchip_pfsoc.c static void microchip_pfsoc_soc_instance_init(Object *obj) · line 120
Quoted 8 lines
hw/riscv/microchip_pfsoc.c +5
-0@@ -111,6 +113,11 @@ static void microchip_pfsoc_soc_instance_init(Object *obj)qdev_prop_set_string (DEVICE (&s->u_cpus),"cpu-type" ,TYPE_RISCV_CPU_SIFIVE_U54 );qdev_prop_set_uint64 (DEVICE (&s->u_cpus),"resetvec" ,RESET_VECTOR ); + +object_initialize_child (obj,"sd-controller" , &s->sdhci, +TYPE_CADENCE_SDHCI ); +object_initialize_child (OBJECT (&s->sdhci),"sd-controller.sdhci" , + &s->sdhci.slot,TYPE_SYSBUS_SDHCI );
OK now I see this patch, so NAck. TYPE_SYSBUS_SDHCI has to be initialized in TYPE_CADENCE_SDHCI, not in all TYPE_CADENCE_SDHCI users.
hw/riscv/microchip_pfsoc.c static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp) · line 233
Quoted 8 lines
hw/riscv/microchip_pfsoc.c +1
-0}static void microchip_pfsoc_soc_realize (DeviceState *dev,Error **errp)@@ -223,6 +230,13 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp) memmap[MICROCHIP_PFSOC_MPUCFG ].base, memmap[MICROCHIP_PFSOC_MPUCFG].size); +/* SDHCI */
Consider setting the SDHCI DMA MR here.
Quoted 18 lines
hw/riscv/microchip_pfsoc.c +7
-0+sysbus_realize (SYS_BUS_DEVICE (&s->sdhci), errp); +sysbus_mmio_map (SYS_BUS_DEVICE (&s->sdhci),0 , + memmap[MICROCHIP_PFSOC_EMMC_SD ].base); +sysbus_connect_irq (SYS_BUS_DEVICE (&s->sdhci.slot),0 , +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_EMMC_SD_IRQ )); +/* MMUARTs */ s->serial0 = mchp_pfsoc_mmuart_create(system_memory, memmap[MICROCHIP_PFSOC_MMUART0].base,@@ -290,6 +304,7 @@ static void microchip_icicle_kit_machine_init(MachineState *machine)MicrochipIcicleKitState *s =MICROCHIP_ICICLE_KIT_MACHINE (machine);MemoryRegion *system_memory =get_system_memory ();MemoryRegion *main_mem =g_new (MemoryRegion,1 ); +DriveInfo *dinfo =drive_get_next (IF_SD );/* Sanity check on RAM size */ if (machine->ram_size < mc->default_ram_size) {
hw/riscv/microchip_pfsoc.c static void microchip_icicle_kit_machine_init(MachineState *machine) · line 333
Quoted 7 lines
hw/riscv/microchip_pfsoc.c +4
-0@@ -312,6 +327,17 @@ static void microchip_icicle_kit_machine_init(MachineState *machine)/* Load the firmware */ riscv_find_and_load_firmware (machine,BIOS_FILENAME ,RESET_VECTOR ,NULL ); + +/* Attach an SD card */ +if (dinfo) { +SDHCIState *sdhci = &(s->soc.sdhci.slot);
NAck again, you are not supposed to access this field here. This is why in the previous patch I asked why the "sd-bus" property was not propagated from child to parent.
Thanks,
Phil.
Quoted 47 lines
hw/riscv/microchip_pfsoc.c +7
-0+DeviceState *card =qdev_new (TYPE_SD_CARD ); +BusState *bus =qdev_get_child_bus (DEVICE (sdhci),"sd-bus" ); + +qdev_prop_set_drive_err (card,"drive" ,blk_by_legacy_dinfo (dinfo), + &error_fatal); +qdev_realize_and_unref (card, bus, &error_fatal); + } }static void microchip_icicle_kit_machine_class_init (ObjectClass *oc,void *data)include/hw/riscv/microchip_pfsoc.h +4
-0@@ -23,6 +23,7 @@#define HW_MICROCHIP_PFSOC_H #include "hw/char/mchp_pfsoc_mmuart.h" +#include "hw/sd/cadence_sdhci.h" typedef struct MicrochipPFSoCState { /*< private >*/@@ -39,6 +40,7 @@ typedef struct MicrochipPFSoCState {MchpPfSoCMMUartState *serial2;MchpPfSoCMMUartState *serial3;MchpPfSoCMMUartState *serial4; +CadenceSDHCIState sdhci; } MicrochipPFSoCState;#define TYPE_MICROCHIP_PFSOC "microchip.pfsoc"@@ -74,6 +76,7 @@ enum {MICROCHIP_PFSOC_MMUART0 ,MICROCHIP_PFSOC_SYSREG ,MICROCHIP_PFSOC_MPUCFG , +MICROCHIP_PFSOC_EMMC_SD ,MICROCHIP_PFSOC_MMUART1 ,MICROCHIP_PFSOC_MMUART2 , MICROCHIP_PFSOC_MMUART3,@@ -85,6 +88,7 @@ enum { };enum { +MICROCHIP_PFSOC_EMMC_SD_IRQ =88 ,MICROCHIP_PFSOC_MMUART0_IRQ =90 ,MICROCHIP_PFSOC_MMUART1_IRQ =91 ,MICROCHIP_PFSOC_MMUART2_IRQ =92 ,
PATCH 12/18 hw/dma: Add Microchip PolarFire Soc DMA controller emulation
In reply to Bin MengFrom: Bin Meng <[email protected]>
Microchip PolarFire SoC integrates a DMA engine that supports:
* Independent concurrent DMA transfers using 4 DMA channels
* Generation of interrupts on various conditions during execution
This creates a simple model to support polling mode which is enough for firmware usage. While there are codes for interrupts handling, please note the interrupt path has not been validated due to missing kernel driver for testing as of now.
---
MAINTAINERS | 2 + hw/dma/Kconfig | 3 + hw/dma/Makefile.objs | 1 + hw/dma/mchp_pfsoc_dma.c | 322 ++++++++++++++++++++++++++++++++++++++++ include/hw/dma/mchp_pfsoc_dma.h | 57 +++++++ 5 files changed, 385 insertions(+) create mode 100644 hw/dma/mchp_pfsoc_dma.c create mode 100644 include/hw/dma/mchp_pfsoc_dma.h
MAINTAINERS +2 -0
@@ -1320,8 +1320,10 @@ L: [email protected] S: Supported F: hw/riscv/microchip_pfsoc.c F: hw/char/mchp_pfsoc_mmuart.c +F: hw/dma/mchp_pfsoc_dma.c F: include/hw/riscv/microchip_pfsoc.h F: include/hw/char/mchp_pfsoc_mmuart.h +F: include/hw/dma/mchp_pfsoc_dma.h RX Machines -----------
hw/dma/Kconfig +3 -0
@@ -20,3 +20,6 @@ config ZYNQ_DEVCFG config STP2000 bool + +config MCHP_PFSOC_DMA + bool
hw/dma/Makefile.objs +1 -0
@@ -14,3 +14,4 @@ common-obj-$(CONFIG_XLNX_ZYNQMP_ARM) += xlnx-zdma.o common-obj-$(CONFIG_OMAP) += omap_dma.o soc_dma.o common-obj-$(CONFIG_PXA2XX) += pxa2xx_dma.o common-obj-$(CONFIG_RASPI) += bcm2835_dma.o +common-obj-$(CONFIG_MCHP_PFSOC_DMA) += mchp_pfsoc_dma.o
hw/dma/mchp_pfsoc_dma.c new file +322 -0
322 lines not on this page · raw message
include/hw/dma/mchp_pfsoc_dma.h new file +57 -0
57 lines not on this page · raw message
Signature · 1 line
2.7.4
PATCH 13/18 hw/riscv: microchip_pfsoc: Connect a DMA controller
In reply to Bin MengFrom: Bin Meng <[email protected]>
Connect a DMA controller to Microchip PolarFire SoC. Note interrupt has not been connected due to missing information in the manual how interrupts are routed to PLIC.
On the Icicle Kit board, the HSS firmware utilizes the on-chip DMA controller to move the 2nd stage bootloader in the system memory.
---
hw/riscv/Kconfig | 1 + hw/riscv/microchip_pfsoc.c | 10 ++++++++++ include/hw/riscv/microchip_pfsoc.h | 3 +++ 3 files changed, 14 insertions(+)
hw/riscv/Kconfig +1 -0
@@ -55,4 +55,5 @@ config MICROCHIP_PFSOC select SIFIVE select UNIMP select MCHP_PFSOC_MMUART + select MCHP_PFSOC_DMA select CADENCE_SDHCI
hw/riscv/microchip_pfsoc.c +10 -0
@@ -13,6 +13,7 @@* 2) eNVM (Embedded Non-Volatile Memory) * 3) MMUARTs (Multi-Mode UART) * 4) Cadence eMMC/SDHC controller and an SD card connected to it +* 5) DMA (Direct Memory Access Controller) * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts.
@@ -71,6 +72,7 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_BUSERR_UNIT4] = { 0x1704000, 0x1000 }, [MICROCHIP_PFSOC_CLINT] = { 0x2000000, 0x10000 }, [MICROCHIP_PFSOC_L2CC] = { 0x2010000, 0x1000 }, + [MICROCHIP_PFSOC_DMA] = { 0x3000000, 0x100000 }, [MICROCHIP_PFSOC_L2LIM] = { 0x8000000, 0x2000000 }, [MICROCHIP_PFSOC_PLIC] = { 0xc000000, 0x4000000 }, [MICROCHIP_PFSOC_MMUART0] = { 0x20000000, 0x1000 },
@@ -114,6 +116,9 @@ static void microchip_pfsoc_soc_instance_init(Object *obj) TYPE_RISCV_CPU_SIFIVE_U54);qdev_prop_set_uint64 (DEVICE (&s->u_cpus),"resetvec" ,RESET_VECTOR ); +object_initialize_child (obj,"dma-controller" , &s->dma, +TYPE_MCHP_PFSOC_DMA ); +object_initialize_child (obj,"sd-controller" , &s->sdhci,TYPE_CADENCE_SDHCI ); object_initialize_child(OBJECT(&s->sdhci), "sd-controller.sdhci",
@@ -220,6 +225,11 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp) memmap[MICROCHIP_PFSOC_PLIC].size);g_free (plic_hart_config); +/* DMA */ +sysbus_realize (SYS_BUS_DEVICE (&s->dma), errp); +sysbus_mmio_map (SYS_BUS_DEVICE (&s->dma),0 , + memmap[MICROCHIP_PFSOC_DMA ].base); +/* SYSREG */ create_unimplemented_device("microchip.pfsoc.sysreg", memmap[MICROCHIP_PFSOC_SYSREG].base,
include/hw/riscv/microchip_pfsoc.h +3 -0
@@ -23,6 +23,7 @@#define HW_MICROCHIP_PFSOC_H #include "hw/char/mchp_pfsoc_mmuart.h" +#include "hw/dma/mchp_pfsoc_dma.h" #include "hw/sd/cadence_sdhci.h" typedef struct MicrochipPFSoCState {
@@ -40,6 +41,7 @@ typedef struct MicrochipPFSoCState {MchpPfSoCMMUartState *serial2;MchpPfSoCMMUartState *serial3;MchpPfSoCMMUartState *serial4; +MchpPfSoCDMAState dma;CadenceSDHCIState sdhci; } MicrochipPFSoCState;
@@ -71,6 +73,7 @@ enum {MICROCHIP_PFSOC_BUSERR_UNIT4 ,MICROCHIP_PFSOC_CLINT ,MICROCHIP_PFSOC_L2CC , +MICROCHIP_PFSOC_DMA ,MICROCHIP_PFSOC_L2LIM ,MICROCHIP_PFSOC_PLIC , MICROCHIP_PFSOC_MMUART0,
Signature · 1 line
2.7.4
Re: [PATCH 13/18] hw/riscv: microchip_pfsoc: Connect a DMA controller
Quoted 48 lines · Bin Meng
From: Bin Meng <[email protected]>
Connect a DMA controller to Microchip PolarFire SoC. Note interrupt has not been connected due to missing information in the manual how interrupts are routed to PLIC.
On the Icicle Kit board, the HSS firmware utilizes the on-chip DMA controller to move the 2nd stage bootloader in the system memory.
Signed-off-by Bin Meng <[email protected]>---
hw/riscv/Kconfig | 1 + hw/riscv/microchip_pfsoc.c | 10 ++++++++++ include/hw/riscv/microchip_pfsoc.h | 3 +++ 3 files changed, 14 insertions(+)hw/riscv/Kconfig +1
-0@@ -55,4 +55,5 @@ config MICROCHIP_PFSOC select SIFIVE select UNIMP select MCHP_PFSOC_MMUART + select MCHP_PFSOC_DMA select CADENCE_SDHCIhw/riscv/microchip_pfsoc.c +2
-0@@ -13,6 +13,7 @@* 2) eNVM (Embedded Non-Volatile Memory) * 3) MMUARTs (Multi-Mode UART) * 4) Cadence eMMC/SDHC controller and an SD card connected to it +* 5) DMA (Direct Memory Access Controller) * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts. @@ -71,6 +72,7 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_BUSERR_UNIT4] = { 0x1704000, 0x1000 }, [MICROCHIP_PFSOC_CLINT] = { 0x2000000, 0x10000 }, [MICROCHIP_PFSOC_L2CC] = { 0x2010000, 0x1000 }, + [MICROCHIP_PFSOC_DMA] = { 0x3000000, 0x100000 }, [MICROCHIP_PFSOC_L2LIM] = { 0x8000000, 0x2000000 }, [MICROCHIP_PFSOC_PLIC] = { 0xc000000, 0x4000000 }, [MICROCHIP_PFSOC_MMUART0] = { 0x20000000, 0x1000 },
hw/riscv/microchip_pfsoc.c static void microchip_pfsoc_soc_instance_init(Object *obj) · line 123
Quoted 8 lines
hw/riscv/microchip_pfsoc.c +3
-0@@ -114,6 +116,9 @@ static void microchip_pfsoc_soc_instance_init(Object *obj) TYPE_RISCV_CPU_SIFIVE_U54);qdev_prop_set_uint64 (DEVICE (&s->u_cpus),"resetvec" ,RESET_VECTOR ); +object_initialize_child (obj,"dma-controller" , &s->dma, +TYPE_MCHP_PFSOC_DMA ); +object_initialize_child (obj,"sd-controller" , &s->sdhci,TYPE_CADENCE_SDHCI );
I haven't looked at the chip specs, but maybe you can add the SD controller after the DMA controller so so you can directly link a DMA address space to it.
Quoted 42 lines
hw/riscv/microchip_pfsoc.c +5
-0object_initialize_child(OBJECT(&s->sdhci), "sd-controller.sdhci",@@ -220,6 +225,11 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp) memmap[MICROCHIP_PFSOC_PLIC].size);g_free (plic_hart_config); +/* DMA */ +sysbus_realize (SYS_BUS_DEVICE (&s->dma), errp); +sysbus_mmio_map (SYS_BUS_DEVICE (&s->dma),0 , + memmap[MICROCHIP_PFSOC_DMA ].base); +/* SYSREG */ create_unimplemented_device("microchip.pfsoc.sysreg", memmap[MICROCHIP_PFSOC_SYSREG].base,include/hw/riscv/microchip_pfsoc.h +3
-0@@ -23,6 +23,7 @@#define HW_MICROCHIP_PFSOC_H #include "hw/char/mchp_pfsoc_mmuart.h" +#include "hw/dma/mchp_pfsoc_dma.h" #include "hw/sd/cadence_sdhci.h" typedef struct MicrochipPFSoCState {@@ -40,6 +41,7 @@ typedef struct MicrochipPFSoCState {MchpPfSoCMMUartState *serial2;MchpPfSoCMMUartState *serial3;MchpPfSoCMMUartState *serial4; +MchpPfSoCDMAState dma;CadenceSDHCIState sdhci; } MicrochipPFSoCState;@@ -71,6 +73,7 @@ enum {MICROCHIP_PFSOC_BUSERR_UNIT4 ,MICROCHIP_PFSOC_CLINT ,MICROCHIP_PFSOC_L2CC , +MICROCHIP_PFSOC_DMA ,MICROCHIP_PFSOC_L2LIM ,MICROCHIP_PFSOC_PLIC , MICROCHIP_PFSOC_MMUART0,
Re: [PATCH 13/18] hw/riscv: microchip_pfsoc: Connect a DMA controller
Hi Philippe,
Quoted 53 lines · Philippe Mathieu-Daudé
Quoted 51 lines · Bin Meng
From: Bin Meng <[email protected]>
Connect a DMA controller to Microchip PolarFire SoC. Note interrupt has not been connected due to missing information in the manual how interrupts are routed to PLIC.
On the Icicle Kit board, the HSS firmware utilizes the on-chip DMA controller to move the 2nd stage bootloader in the system memory.
Signed-off-by Bin Meng <[email protected]>---
hw/riscv/Kconfig | 1 + hw/riscv/microchip_pfsoc.c | 10 ++++++++++ include/hw/riscv/microchip_pfsoc.h | 3 +++ 3 files changed, 14 insertions(+)hw/riscv/Kconfig +1
-0@@ -55,4 +55,5 @@ config MICROCHIP_PFSOC select SIFIVE select UNIMP select MCHP_PFSOC_MMUART + select MCHP_PFSOC_DMA select CADENCE_SDHCIhw/riscv/microchip_pfsoc.c +2
-0@@ -13,6 +13,7 @@* 2) eNVM (Embedded Non-Volatile Memory) * 3) MMUARTs (Multi-Mode UART) * 4) Cadence eMMC/SDHC controller and an SD card connected to it +* 5) DMA (Direct Memory Access Controller) * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts. @@ -71,6 +72,7 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_BUSERR_UNIT4] = { 0x1704000, 0x1000 }, [MICROCHIP_PFSOC_CLINT] = { 0x2000000, 0x10000 }, [MICROCHIP_PFSOC_L2CC] = { 0x2010000, 0x1000 }, + [MICROCHIP_PFSOC_DMA] = { 0x3000000, 0x100000 }, [MICROCHIP_PFSOC_L2LIM] = { 0x8000000, 0x2000000 }, [MICROCHIP_PFSOC_PLIC] = { 0xc000000, 0x4000000 }, [MICROCHIP_PFSOC_MMUART0] = { 0x20000000, 0x1000 },@@ -114,6 +116,9 @@ static void microchip_pfsoc_soc_instance_init(Object *obj) TYPE_RISCV_CPU_SIFIVE_U54);qdev_prop_set_uint64 (DEVICE (&s->u_cpus),"resetvec" ,RESET_VECTOR );
Quoted 10 lines
hw/riscv/microchip_pfsoc.c static void microchip_pfsoc_soc_instance_init(Object *obj) · line 123
Quoted 5 lines
hw/riscv/microchip_pfsoc.c +3
-0+object_initialize_child (obj,"dma-controller" , &s->dma, +TYPE_MCHP_PFSOC_DMA ); +object_initialize_child (obj,"sd-controller" , &s->sdhci,TYPE_CADENCE_SDHCI );I haven't looked at the chip specs, but maybe you can add the SD controller after the DMA controller so so you can directly link a DMA address space to it.
I am not sure I understand what you meant about adding the SD controller after the DMA controller. The Cadence SD controller has its own built-in DMA and does not depend on this DMA controller.
Regards,
Bin
Re: [PATCH 13/18] hw/riscv: microchip_pfsoc: Connect a DMA controller
Quoted 67 lines · Bin Meng
Hi Philippe,
Quoted 53 lines · Philippe Mathieu-Daudé
Quoted 51 lines · Bin Meng
From: Bin Meng <[email protected]>
Connect a DMA controller to Microchip PolarFire SoC. Note interrupt has not been connected due to missing information in the manual how interrupts are routed to PLIC.
On the Icicle Kit board, the HSS firmware utilizes the on-chip DMA controller to move the 2nd stage bootloader in the system memory.
Signed-off-by Bin Meng <[email protected]>---
hw/riscv/Kconfig | 1 + hw/riscv/microchip_pfsoc.c | 10 ++++++++++ include/hw/riscv/microchip_pfsoc.h | 3 +++ 3 files changed, 14 insertions(+)hw/riscv/Kconfig +1
-0@@ -55,4 +55,5 @@ config MICROCHIP_PFSOC select SIFIVE select UNIMP select MCHP_PFSOC_MMUART + select MCHP_PFSOC_DMA select CADENCE_SDHCIhw/riscv/microchip_pfsoc.c +2
-0@@ -13,6 +13,7 @@* 2) eNVM (Embedded Non-Volatile Memory) * 3) MMUARTs (Multi-Mode UART) * 4) Cadence eMMC/SDHC controller and an SD card connected to it +* 5) DMA (Direct Memory Access Controller) * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts. @@ -71,6 +72,7 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_BUSERR_UNIT4] = { 0x1704000, 0x1000 }, [MICROCHIP_PFSOC_CLINT] = { 0x2000000, 0x10000 }, [MICROCHIP_PFSOC_L2CC] = { 0x2010000, 0x1000 }, + [MICROCHIP_PFSOC_DMA] = { 0x3000000, 0x100000 }, [MICROCHIP_PFSOC_L2LIM] = { 0x8000000, 0x2000000 }, [MICROCHIP_PFSOC_PLIC] = { 0xc000000, 0x4000000 }, [MICROCHIP_PFSOC_MMUART0] = { 0x20000000, 0x1000 },@@ -114,6 +116,9 @@ static void microchip_pfsoc_soc_instance_init(Object *obj) TYPE_RISCV_CPU_SIFIVE_U54);qdev_prop_set_uint64 (DEVICE (&s->u_cpus),"resetvec" ,RESET_VECTOR );Quoted 10 lines
hw/riscv/microchip_pfsoc.c static void microchip_pfsoc_soc_instance_init(Object *obj) · line 123
Quoted 5 lines
hw/riscv/microchip_pfsoc.c +3
-0+object_initialize_child (obj,"dma-controller" , &s->dma, +TYPE_MCHP_PFSOC_DMA ); +object_initialize_child (obj,"sd-controller" , &s->sdhci,TYPE_CADENCE_SDHCI );I haven't looked at the chip specs, but maybe you can add the SD controller after the DMA controller so so you can directly link a DMA address space to it.
Quoted 3 lines
I am not sure I understand what you meant about adding the SD controller after the DMA controller. The Cadence SD controller has its own built-in DMA and does not depend on this DMA controller.
Ah OK. What I'm concerned about is the SD controller do its DMA access in a proper DMA address space.
Quoted 4 lines
Regards,
Bin
PATCH 14/18 hw/net: cadence_gem: Add a new 'phy-addr' property
In reply to Bin MengFrom: Bin Meng <[email protected]>
At present the PHY address of the PHY connected to GEM is hard-coded to either 23 (BOARD_PHY_ADDRESS) or 0. This might not be the case for all boards. Add a new 'phy-addr' property so that board can specify the PHY address for each GEM instance.
---
hw/net/cadence_gem.c | 7 +++++-- include/hw/net/cadence_gem.h | 2 ++ 2 files changed, 7 insertions(+), 2 deletions(-)
hw/net/cadence_gem.c +5 -2
@@ -1446,7 +1446,8 @@ static uint64_t gem_read(void *opaque, hwaddr offset, unsigned size)uint32_t phy_addr, reg_num; phy_addr = (retval &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (retval &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ; retval &=0xFFFF0000 ; retval |=gem_phy_read (s, reg_num);
@@ -1569,7 +1570,8 @@ static void gem_write(void *opaque, hwaddr offset, uint64_t val,uint32_t phy_addr, reg_num; phy_addr = (val &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (val &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ;gem_phy_write (s, reg_num, val); }
@@ -1682,6 +1684,7 @@ static Property gem_properties[] = {DEFINE_NIC_PROPERTIES (CadenceGEMState, conf),DEFINE_PROP_UINT32 ("revision" , CadenceGEMState, revision,GEM_MODID_VALUE ), +DEFINE_PROP_UINT8 ("phy-addr" , CadenceGEMState, phy_addr,0 ),DEFINE_PROP_UINT8 ("num-priority-queues" , CadenceGEMState, num_priority_queues,1 ), DEFINE_PROP_UINT8("num-type1-screeners", CadenceGEMState,
include/hw/net/cadence_gem.h +2 -0
@@ -73,6 +73,8 @@ typedef struct CadenceGEMState {/* Mask of register bits which are write 1 to clear */ uint32_t regs_w1c[CADENCE_GEM_MAXREG ]; +/* PHY address */ +uint8_t phy_addr;/* PHY registers backing store */ uint16_t phy_regs[32 ];
Signature · 1 line
2.7.4
Re: [PATCH 14/18] hw/net: cadence_gem: Add a new 'phy-addr' property
Quoted 39 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the PHY address of the PHY connected to GEM is hard-coded to either 23 (
BOARD_PHY_ADDRESS) or 0. This might not be the case for all boards. Add a new 'phy-addr' property so that board can specify the PHY address for each GEM instance.Signed-off-by Bin Meng <[email protected]>---
hw/net/cadence_gem.c | 7 +++++-- include/hw/net/cadence_gem.h | 2 ++ 2 files changed, 7 insertions(+), 2 deletions(-)hw/net/cadence_gem.c +4
-2@@ -1446,7 +1446,8 @@ static uint64_t gem_read(void *opaque, hwaddr offset, unsigned size)uint32_t phy_addr, reg_num; phy_addr = (retval &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (retval &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ; retval &=0xFFFF0000 ; retval |=gem_phy_read (s, reg_num);@@ -1569,7 +1570,8 @@ static void gem_write(void *opaque, hwaddr offset, uint64_t val,uint32_t phy_addr, reg_num; phy_addr = (val &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (val &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ;gem_phy_write (s, reg_num, val); }
hw/net/cadence_gem.c static Property gem_properties[] = { · line 1687
Quoted 4 lines
hw/net/cadence_gem.c +1
-0@@ -1682,6 +1684,7 @@ static Property gem_properties[] = {DEFINE_NIC_PROPERTIES (CadenceGEMState, conf),DEFINE_PROP_UINT32 ("revision" , CadenceGEMState, revision,GEM_MODID_VALUE ), + DEFINE_PROP_UINT8("phy-addr", CadenceGEMState, phy_addr, 0),
This patch would be complete by moving the BOARD_PHY_ADDRESS definition to each board using this NIC, and setting the "phy-addr" property to this value.
Quoted 17 lines
hw/net/cadence_gem.c +0
-0DEFINE_PROP_UINT8 ("num-priority-queues" , CadenceGEMState, num_priority_queues,1 ), DEFINE_PROP_UINT8("num-type1-screeners", CadenceGEMState,include/hw/net/cadence_gem.h +2
-0@@ -73,6 +73,8 @@ typedef struct CadenceGEMState {/* Mask of register bits which are write 1 to clear */ uint32_t regs_w1c[CADENCE_GEM_MAXREG ]; +/* PHY address */ +uint8_t phy_addr;/* PHY registers backing store */ uint16_t phy_regs[32 ];
Re: [PATCH 14/18] hw/net: cadence_gem: Add a new 'phy-addr' property
Quoted 33 lines · Philippe Mathieu-Daudé
Quoted 31 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the PHY address of the PHY connected to GEM is hard-coded to either 23 (
BOARD_PHY_ADDRESS) or 0. This might not be the case for all boards. Add a new 'phy-addr' property so that board can specify the PHY address for each GEM instance.Signed-off-by Bin Meng <[email protected]>---
hw/net/cadence_gem.c | 7 +++++-- include/hw/net/cadence_gem.h | 2 ++ 2 files changed, 7 insertions(+), 2 deletions(-)hw/net/cadence_gem.c +2
-1@@ -1446,7 +1446,8 @@ static uint64_t gem_read(void *opaque, hwaddr offset, unsigned size)uint32_t phy_addr, reg_num; phy_addr = (retval &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (retval &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ; retval &=0xFFFF0000 ; retval |=gem_phy_read (s, reg_num);@@ -1569,7 +1570,8 @@ static void gem_write(void *opaque, hwaddr offset, uint64_t val,uint32_t phy_addr, reg_num;
Quoted 16 lines
hw/net/cadence_gem.c static Property gem_properties[] = { · line 1687
Quoted 12 lines
hw/net/cadence_gem.c +3
-1phy_addr = (val &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (val &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ;gem_phy_write (s, reg_num, val); }@@ -1682,6 +1684,7 @@ static Property gem_properties[] = {DEFINE_NIC_PROPERTIES (CadenceGEMState, conf),DEFINE_PROP_UINT32 ("revision" , CadenceGEMState, revision,GEM_MODID_VALUE ), + DEFINE_PROP_UINT8("phy-addr", CadenceGEMState, phy_addr, 0),This patch would be complete by moving the
BOARD_PHY_ADDRESSdefinition to each board using this NIC, and setting the "phy-addr" property to this value.
I actually have no idea which board in QEMU is using this special PHY address instead of default 0.
It looks BOARD_PHY_ADDRESS has been there since the initial version of the cadence_gem model.
commit e9f186e514a70557d695cadd2c2287ef97737023
Author: Peter A. G. Crosthwaite <[email protected]> Date: Mon Mar 5 14:39:12 2012 +1000
cadence_gem: initial version of device model
Device model for cadence gem ethernet controller.
Later PHY address 0 was added via the following commit:
commit 553893736885e4f2dda424bff3e2200e1b6482a5
Author: Peter Crosthwaite <[email protected]> Date: Thu Apr 3 23:55:19 2014 -0700
net: cadence_gem: Make phy respond to broadcast
Phys must respond to address 0 by specification. Implement.
Message-id:
6f4d53b04ddbfb19895bfb61a595e69f1c08859a.1396594056.git.peter.crosthwaite@xilinx.com
I doubt the commit message said that PHYs must respond to address 0. I am not aware of such specs. The issue was probably that whatever board 2nd commit was tested against did not have a PHY at address BOARD_PHY_ADDRESS.
+ a couple of Xilinx folks to comment.
Quoted 12 lines
Quoted 12 lines
hw/net/cadence_gem.c +0
-0DEFINE_PROP_UINT8 ("num-priority-queues" , CadenceGEMState, num_priority_queues,1 ), DEFINE_PROP_UINT8("num-type1-screeners", CadenceGEMState,include/hw/net/cadence_gem.h +0
-0@@ -73,6 +73,8 @@ typedef struct CadenceGEMState {/* Mask of register bits which are write 1 to clear */ uint32_t regs_w1c[CADENCE_GEM_MAXREG ];
Quoted 4 lines
include/hw/net/cadence_gem.h typedef struct CadenceGEMState { · line 79
Quoted 4 lines
include/hw/net/cadence_gem.h +2
-0+/* PHY address */ +uint8_t phy_addr;/* PHY registers backing store */ uint16_t phy_regs[32 ];
Regards,
Bin
Re: [PATCH 14/18] hw/net: cadence_gem: Add a new 'phy-addr' property
Quoted 51 lines · Bin Meng
Quoted 33 lines · Philippe Mathieu-Daudé
Quoted 31 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the PHY address of the PHY connected to GEM is hard-coded to either 23 (
BOARD_PHY_ADDRESS) or 0. This might not be the case for all boards. Add a new 'phy-addr' property so that board can specify the PHY address for each GEM instance.Signed-off-by Bin Meng <[email protected]>---
hw/net/cadence_gem.c | 7 +++++-- include/hw/net/cadence_gem.h | 2 ++ 2 files changed, 7 insertions(+), 2 deletions(-)hw/net/cadence_gem.c +2
-1@@ -1446,7 +1446,8 @@ static uint64_t gem_read(void *opaque, hwaddr offset, unsigned size)uint32_t phy_addr, reg_num; phy_addr = (retval &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (retval &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ; retval &=0xFFFF0000 ; retval |=gem_phy_read (s, reg_num);@@ -1569,7 +1570,8 @@ static void gem_write(void *opaque, hwaddr offset, uint64_t val,uint32_t phy_addr, reg_num;Quoted 16 lines
hw/net/cadence_gem.c static Property gem_properties[] = { · line 1687
Quoted 12 lines
hw/net/cadence_gem.c +3
-1phy_addr = (val &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (val &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ;gem_phy_write (s, reg_num, val); }@@ -1682,6 +1684,7 @@ static Property gem_properties[] = {DEFINE_NIC_PROPERTIES (CadenceGEMState, conf),DEFINE_PROP_UINT32 ("revision" , CadenceGEMState, revision,GEM_MODID_VALUE ), + DEFINE_PROP_UINT8("phy-addr", CadenceGEMState, phy_addr, 0),This patch would be complete by moving the
BOARD_PHY_ADDRESSdefinition to each board using this NIC, and setting the "phy-addr" property to this value.
Quoted 2 lines
I actually have no idea which board in QEMU is using this special PHY address instead of default 0.
It seems safe to assume all do use address 23.
Anyway you can simply get ride of BOARD_PHY_ADDRESS in the read/write using:
DEFINE_PROP_UINT8("phy-addr", CadenceGEMState,
phy_addr, BOARD_PHY_ADDRESS),
Quoted 61 lines
It looks
BOARD_PHY_ADDRESShas been there since the initial version of thecadence_gemmodel.commit
e9f186e514a70557d695cadd2c2287ef97737023Author: Peter A. G. Crosthwaite <[email protected]> Date: Mon Mar 5 14:39:12 2012 +1000
cadence_gem: initial version of device modelDevice model for cadence gem ethernet controller.
Signed-off-by Peter A. G. Crosthwaite <[email protected]>Signed-off-by John Linn <[email protected]>Signed-off-by Edgar E. Iglesias <[email protected]>Later PHY address 0 was added via the following commit:
commit
553893736885e4f2dda424bff3e2200e1b6482a5Author: Peter Crosthwaite <[email protected]> Date: Thu Apr 3 23:55:19 2014 -0700net:
cadence_gem: Make phy respond to broadcastPhys must respond to address 0 by specification. Implement.
Signed-off-by Nathan Rossi <[email protected]>Signed-off-by Peter Crosthwaite <[email protected]>Message-id:
6f4d53b04ddbfb19895bfb61a595e69f1c08859a.1396594056.git.peter.crosthwaite@xilinx.comReviewed-by Beniamino Galvani <[email protected]>Signed-off-by Peter Maydell <[email protected]>I doubt the commit message said that PHYs must respond to address 0. I am not aware of such specs. The issue was probably that whatever board 2nd commit was tested against did not have a PHY at address
BOARD_PHY_ADDRESS.+ a couple of Xilinx folks to comment.
Quoted 12 lines
Quoted 12 lines
hw/net/cadence_gem.c +0
-0DEFINE_PROP_UINT8 ("num-priority-queues" , CadenceGEMState, num_priority_queues,1 ), DEFINE_PROP_UINT8("num-type1-screeners", CadenceGEMState,include/hw/net/cadence_gem.h +0
-0@@ -73,6 +73,8 @@ typedef struct CadenceGEMState {/* Mask of register bits which are write 1 to clear */ uint32_t regs_w1c[CADENCE_GEM_MAXREG ];Quoted 4 lines
include/hw/net/cadence_gem.h typedef struct CadenceGEMState { · line 79
Quoted 4 lines
include/hw/net/cadence_gem.h +2
-0+/* PHY address */ +uint8_t phy_addr;/* PHY registers backing store */ uint16_t phy_regs[32 ];Regards,
Bin
Re: [PATCH 14/18] hw/net: cadence_gem: Add a new 'phy-addr' property
In reply to Bin MengQuoted 52 lines · Bin Meng
Quoted 33 lines · Philippe Mathieu-Daudé
Quoted 31 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the PHY address of the PHY connected to GEM is hard-coded to either 23 (
BOARD_PHY_ADDRESS) or 0. This might not be the case for all boards. Add a new 'phy-addr' property so that board can specify the PHY address for each GEM instance.Signed-off-by Bin Meng <[email protected]>---
hw/net/cadence_gem.c | 7 +++++-- include/hw/net/cadence_gem.h | 2 ++ 2 files changed, 7 insertions(+), 2 deletions(-)hw/net/cadence_gem.c +2
-1@@ -1446,7 +1446,8 @@ static uint64_t gem_read(void *opaque, hwaddr offset, unsigned size)uint32_t phy_addr, reg_num; phy_addr = (retval &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (retval &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ; retval &=0xFFFF0000 ; retval |=gem_phy_read (s, reg_num);@@ -1569,7 +1570,8 @@ static void gem_write(void *opaque, hwaddr offset, uint64_t val,uint32_t phy_addr, reg_num;Quoted 16 lines
hw/net/cadence_gem.c static Property gem_properties[] = { · line 1687
Quoted 12 lines
hw/net/cadence_gem.c +3
-1phy_addr = (val &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (val &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ;gem_phy_write (s, reg_num, val); }@@ -1682,6 +1684,7 @@ static Property gem_properties[] = {DEFINE_NIC_PROPERTIES (CadenceGEMState, conf),DEFINE_PROP_UINT32 ("revision" , CadenceGEMState, revision,GEM_MODID_VALUE ), + DEFINE_PROP_UINT8("phy-addr", CadenceGEMState, phy_addr, 0),This patch would be complete by moving the
BOARD_PHY_ADDRESSdefinition to each board using this NIC, and setting the "phy-addr" property to this value.
Quoted 2 lines
I actually have no idea which board in QEMU is using this special PHY address instead of default 0.
Given Xilinx's QEMU fork has not used this value for quite some time, I suspect it was only used to match an early chip emulation platform/board.
https://github.com/Xilinx/qemu/blame/master/hw/net/cadence_gem.c#L248
Quoted 31 lines
It looks
BOARD_PHY_ADDRESShas been there since the initial version of thecadence_gemmodel.commit
e9f186e514a70557d695cadd2c2287ef97737023Author: Peter A. G. Crosthwaite <[email protected]> Date: Mon Mar 5 14:39:12 2012 +1000
cadence_gem: initial version of device modelDevice model for cadence gem ethernet controller.
Signed-off-by Peter A. G. Crosthwaite <[email protected]>Signed-off-by John Linn <[email protected]>Signed-off-by Edgar E. Iglesias <[email protected]>Later PHY address 0 was added via the following commit:
commit
553893736885e4f2dda424bff3e2200e1b6482a5Author: Peter Crosthwaite <[email protected]> Date: Thu Apr 3 23:55:19 2014 -0700net:
cadence_gem: Make phy respond to broadcastPhys must respond to address 0 by specification. Implement.
Signed-off-by Nathan Rossi <[email protected]>Signed-off-by Peter Crosthwaite <[email protected]>Message-id:
6f4d53b04ddbfb19895bfb61a595e69f1c08859a.1396594056.git.peter.crosthwaite@xilinx.com
Quoted 7 lines
Reviewed-by Beniamino Galvani <[email protected]>Signed-off-by Peter Maydell <[email protected]>I doubt the commit message said that PHYs must respond to address 0. I am not aware of such specs. The issue was probably that whatever board 2nd commit was tested against did not have a PHY at address
BOARD_PHY_ADDRESS.
It is common for phy devices to support it as a broadcast address. It is also commonly written in Xilinx documentation that it is treated as a broadcast address. e.g. the axi ethernet core
(https://www.xilinx.com/support/documentation/ip_documentation/axi_ethernet/v7_0/pg138-axi-ethernet.pdf
page 45). But 802.3 does not require it, instead address 0 is only reserved.
With this commit the "must" refers to the device specification, in that QEMU is emulating a real phy (though more specifically the phy(s) that were being emulated for Zynq boards) that does respond to address 0 as a broadcast. This change was a trade off in order to make QEMU behave as it would on hardware, such that software using address 0 as broadcast would work correctly.
Regards,
Nathan
Quoted 22 lines
+ a couple of Xilinx folks to comment.
Quoted 12 lines
Quoted 12 lines
hw/net/cadence_gem.c +0
-0DEFINE_PROP_UINT8 ("num-priority-queues" , CadenceGEMState, num_priority_queues,1 ), DEFINE_PROP_UINT8("num-type1-screeners", CadenceGEMState,include/hw/net/cadence_gem.h +0
-0@@ -73,6 +73,8 @@ typedef struct CadenceGEMState {/* Mask of register bits which are write 1 to clear */ uint32_t regs_w1c[CADENCE_GEM_MAXREG ];Quoted 4 lines
include/hw/net/cadence_gem.h typedef struct CadenceGEMState { · line 79
Quoted 4 lines
include/hw/net/cadence_gem.h +2
-0+/* PHY address */ +uint8_t phy_addr;/* PHY registers backing store */ uint16_t phy_regs[32 ];Regards,
Bin
Re: [PATCH 14/18] hw/net: cadence_gem: Add a new 'phy-addr' property
Quoted 109 lines · Nathan Rossi
Quoted 52 lines · Bin Meng
Quoted 33 lines · Philippe Mathieu-Daudé
Quoted 31 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the PHY address of the PHY connected to GEM is hard-coded to either 23 (
BOARD_PHY_ADDRESS) or 0. This might not be the case for all boards. Add a new 'phy-addr' property so that board can specify the PHY address for each GEM instance.Signed-off-by Bin Meng <[email protected]>---
hw/net/cadence_gem.c | 7 +++++-- include/hw/net/cadence_gem.h | 2 ++ 2 files changed, 7 insertions(+), 2 deletions(-)hw/net/cadence_gem.c +2
-1@@ -1446,7 +1446,8 @@ static uint64_t gem_read(void *opaque, hwaddr offset, unsigned size)uint32_t phy_addr, reg_num; phy_addr = (retval &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (retval &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ; retval &=0xFFFF0000 ; retval |=gem_phy_read (s, reg_num);@@ -1569,7 +1570,8 @@ static void gem_write(void *opaque, hwaddr offset, uint64_t val,uint32_t phy_addr, reg_num;Quoted 16 lines
hw/net/cadence_gem.c static Property gem_properties[] = { · line 1687
Quoted 12 lines
hw/net/cadence_gem.c +3
-1phy_addr = (val &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (val &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ;gem_phy_write (s, reg_num, val); }@@ -1682,6 +1684,7 @@ static Property gem_properties[] = {DEFINE_NIC_PROPERTIES (CadenceGEMState, conf),DEFINE_PROP_UINT32 ("revision" , CadenceGEMState, revision,GEM_MODID_VALUE ), + DEFINE_PROP_UINT8("phy-addr", CadenceGEMState, phy_addr, 0),This patch would be complete by moving the
BOARD_PHY_ADDRESSdefinition to each board using this NIC, and setting the "phy-addr" property to this value.Quoted 2 lines
I actually have no idea which board in QEMU is using this special PHY address instead of default 0.
Given Xilinx's QEMU fork has not used this value for quite some time, I suspect it was only used to match an early chip emulation platform/board.
https://github.com/Xilinx/qemu/blame/master/hw/net/cadence_gem.c#L248
Quoted 31 lines
It looks
BOARD_PHY_ADDRESShas been there since the initial version of thecadence_gemmodel.commit
e9f186e514a70557d695cadd2c2287ef97737023Author: Peter A. G. Crosthwaite <[email protected]> Date: Mon Mar 5 14:39:12 2012 +1000
cadence_gem: initial version of device modelDevice model for cadence gem ethernet controller.
Signed-off-by Peter A. G. Crosthwaite <[email protected]>Signed-off-by John Linn <[email protected]>Signed-off-by Edgar E. Iglesias <[email protected]>Later PHY address 0 was added via the following commit:
commit
553893736885e4f2dda424bff3e2200e1b6482a5Author: Peter Crosthwaite <[email protected]> Date: Thu Apr 3 23:55:19 2014 -0700net:
cadence_gem: Make phy respond to broadcastPhys must respond to address 0 by specification. Implement.
Signed-off-by Nathan Rossi <[email protected]>Signed-off-by Peter Crosthwaite <[email protected]>Message-id:
6f4d53b04ddbfb19895bfb61a595e69f1c08859a.1396594056.git.peter.crosthwaite@xilinx.comQuoted 7 lines
Reviewed-by Beniamino Galvani <[email protected]>Signed-off-by Peter Maydell <[email protected]>I doubt the commit message said that PHYs must respond to address 0. I am not aware of such specs. The issue was probably that whatever board 2nd commit was tested against did not have a PHY at address
BOARD_PHY_ADDRESS.It is common for phy devices to support it as a broadcast address. It is also commonly written in Xilinx documentation that it is treated as a broadcast address. e.g. the axi ethernet core
(https://www.xilinx.com/support/documentation/ip_documentation/axi_ethernet/v7_0/pg138-axi-ethernet.pdf
page 45). But 802.3 does not require it, instead address 0 is only reserved.
Quoted 7 lines
With this commit the "must" refers to the device specification, in that QEMU is emulating a real phy (though more specifically the phy(s) that were being emulated for Zynq boards) that does respond to address 0 as a broadcast. This change was a trade off in order to make QEMU behave as it would on hardware, such that software using address 0 as broadcast would work correctly.
Thanks Nathan. So is it safe to just remove BOARD_PHY_ADDRESS and set "phy-addr" property default value to 0?
Regards,
Bin
Re: [PATCH 14/18] hw/net: cadence_gem: Add a new 'phy-addr' property
Quoted 118 lines · Bin Meng
Quoted 109 lines · Nathan Rossi
Quoted 52 lines · Bin Meng
Quoted 33 lines · Philippe Mathieu-Daudé
Quoted 31 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the PHY address of the PHY connected to GEM is hard-coded to either 23 (
BOARD_PHY_ADDRESS) or 0. This might not be the case for all boards. Add a new 'phy-addr' property so that board can specify the PHY address for each GEM instance.Signed-off-by Bin Meng <[email protected]>---
hw/net/cadence_gem.c | 7 +++++-- include/hw/net/cadence_gem.h | 2 ++ 2 files changed, 7 insertions(+), 2 deletions(-)hw/net/cadence_gem.c +2
-1@@ -1446,7 +1446,8 @@ static uint64_t gem_read(void *opaque, hwaddr offset, unsigned size)uint32_t phy_addr, reg_num; phy_addr = (retval &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (retval &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ; retval &=0xFFFF0000 ; retval |=gem_phy_read (s, reg_num);@@ -1569,7 +1570,8 @@ static void gem_write(void *opaque, hwaddr offset, uint64_t val,uint32_t phy_addr, reg_num;Quoted 16 lines
hw/net/cadence_gem.c static Property gem_properties[] = { · line 1687
Quoted 12 lines
hw/net/cadence_gem.c +3
-1phy_addr = (val &GEM_PHYMNTNC_ADDR ) >>GEM_PHYMNTNC_ADDR_SHFT ; -if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 ) { +if (phy_addr ==BOARD_PHY_ADDRESS || phy_addr ==0 || + phy_addr == s->phy_addr) { reg_num = (val &GEM_PHYMNTNC_REG ) >>GEM_PHYMNTNC_REG_SHIFT ;gem_phy_write (s, reg_num, val); }@@ -1682,6 +1684,7 @@ static Property gem_properties[] = {DEFINE_NIC_PROPERTIES (CadenceGEMState, conf),DEFINE_PROP_UINT32 ("revision" , CadenceGEMState, revision,GEM_MODID_VALUE ), + DEFINE_PROP_UINT8("phy-addr", CadenceGEMState, phy_addr, 0),This patch would be complete by moving the
BOARD_PHY_ADDRESSdefinition to each board using this NIC, and setting the "phy-addr" property to this value.Quoted 2 lines
I actually have no idea which board in QEMU is using this special PHY address instead of default 0.
Given Xilinx's QEMU fork has not used this value for quite some time, I suspect it was only used to match an early chip emulation platform/board.
https://github.com/Xilinx/qemu/blame/master/hw/net/cadence_gem.c#L248
Quoted 31 lines
It looks
BOARD_PHY_ADDRESShas been there since the initial version of thecadence_gemmodel.commit
e9f186e514a70557d695cadd2c2287ef97737023Author: Peter A. G. Crosthwaite <[email protected]> Date: Mon Mar 5 14:39:12 2012 +1000
cadence_gem: initial version of device modelDevice model for cadence gem ethernet controller.
Signed-off-by Peter A. G. Crosthwaite <[email protected]>Signed-off-by John Linn <[email protected]>Signed-off-by Edgar E. Iglesias <[email protected]>Later PHY address 0 was added via the following commit:
commit
553893736885e4f2dda424bff3e2200e1b6482a5Author: Peter Crosthwaite <[email protected]> Date: Thu Apr 3 23:55:19 2014 -0700net:
cadence_gem: Make phy respond to broadcastPhys must respond to address 0 by specification. Implement.
Signed-off-by Nathan Rossi <[email protected]>Signed-off-by Peter Crosthwaite <[email protected]>Message-id:
6f4d53b04ddbfb19895bfb61a595e69f1c08859a.1396594056.git.peter.crosthwaite@xilinx.comQuoted 7 lines
Reviewed-by Beniamino Galvani <[email protected]>Signed-off-by Peter Maydell <[email protected]>I doubt the commit message said that PHYs must respond to address 0. I am not aware of such specs. The issue was probably that whatever board 2nd commit was tested against did not have a PHY at address
BOARD_PHY_ADDRESS.It is common for phy devices to support it as a broadcast address. It is also commonly written in Xilinx documentation that it is treated as a broadcast address. e.g. the axi ethernet core
(https://www.xilinx.com/support/documentation/ip_documentation/axi_ethernet/v7_0/pg138-axi-ethernet.pdf
page 45). But 802.3 does not require it, instead address 0 is only reserved.Quoted 7 lines
With this commit the "must" refers to the device specification, in that QEMU is emulating a real phy (though more specifically the phy(s) that were being emulated for Zynq boards) that does respond to address 0 as a broadcast. This change was a trade off in order to make QEMU behave as it would on hardware, such that software using address 0 as broadcast would work correctly.
Quoted 2 lines
Thanks Nathan. So is it safe to just remove
BOARD_PHY_ADDRESSand set "phy-addr" property default value to 0?
I'd do as following:
First patch, introduce "phy-addr" property (default to BOARD_PHY_ADDRESS) and remove BOARD_PHY_ADDRESS in code:
DEFINE_PROP_UINT8("phy-addr", CadenceGEMState,
phy_addr, BOARD_PHY_ADDRESS),
Second patch set "phy-addr" to BOARD_PHY_ADDRESS in all current boards using this PHY and set the default to 0.
Thanks,
Phil.
Quoted 4 lines
Regards,
Bin
PATCH 15/18 hw/riscv: microchip_pfsoc: Connect 2 Cadence GEMs
In reply to Bin MengFrom: Bin Meng <[email protected]>
Microchip PolarFire SoC integrates 2 Candence GEMs to provide IEEE 802.3 standard-compliant 10/100/1000 Mbps ethernet interface.
On the Icicle Kit board, GEM0 connects to a PHY at address 8 while GEM1 connects to a PHY at address 9.
The 2nd stage bootloader (U-Boot) is using GEM1 by default, so we must specify 2 '-nic' options from the command line in order to get a working ethernet.
---
hw/riscv/microchip_pfsoc.c | 39 ++++++++++++++++++++++++++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 7 +++++++ 2 files changed, 46 insertions(+)
hw/riscv/microchip_pfsoc.c +39 -0
@@ -14,6 +14,7 @@* 3) MMUARTs (Multi-Mode UART) * 4) Cadence eMMC/SDHC controller and an SD card connected to it * 5) DMA (Direct Memory Access Controller) +* 6) GEM (Gigabit Ethernet MAC Controller) * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts.
@@ -59,6 +60,9 @@#define BIOS_FILENAME "hss.bin"#define RESET_VECTOR 0x20220000 +/* GEM version */ +#define GEM_REVISION 0x0107010c + static const struct MemmapEntry { hwaddr base; hwaddr size;
@@ -83,6 +87,8 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_MMUART2] = { 0x20102000, 0x1000 }, [MICROCHIP_PFSOC_MMUART3] = { 0x20104000, 0x1000 }, [MICROCHIP_PFSOC_MMUART4] = { 0x20106000, 0x1000 }, + [MICROCHIP_PFSOC_GEM0] = { 0x20110000, 0x2000 }, + [MICROCHIP_PFSOC_GEM1] = { 0x20112000, 0x2000 }, [MICROCHIP_PFSOC_ENVM_CFG] = { 0x20200000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_DATA] = { 0x20220000, 0x20000 }, [MICROCHIP_PFSOC_IOSCB_CFG] = { 0x37080000, 0x1000 },
@@ -119,6 +125,9 @@ static void microchip_pfsoc_soc_instance_init(Object *obj)object_initialize_child (obj,"dma-controller" , &s->dma,TYPE_MCHP_PFSOC_DMA ); +object_initialize_child (obj,"gem0" , &s->gem0,TYPE_CADENCE_GEM ); +object_initialize_child (obj,"gem1" , &s->gem1,TYPE_CADENCE_GEM ); +object_initialize_child (obj,"sd-controller" , &s->sdhci,TYPE_CADENCE_SDHCI ); object_initialize_child(OBJECT(&s->sdhci), "sd-controller.sdhci",
@@ -136,6 +145,7 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp)MemoryRegion *envm_data =g_new (MemoryRegion,1 );char *plic_hart_config;size_t plic_hart_config_len; +NICInfo *nd;int i;sysbus_realize (SYS_BUS_DEVICE (&s->e_cpus), &error_abort);
@@ -269,6 +279,35 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp)qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART4_IRQ ), serial_hd(4)); +/* GEMs */ + + nd = &nd_table[0 ]; +if (nd->used) { +qemu_check_nic_model (nd,TYPE_CADENCE_GEM ); +qdev_set_nic_properties (DEVICE (&s->gem0), nd); + } + nd = &nd_table[1 ]; +if (nd->used) { +qemu_check_nic_model (nd,TYPE_CADENCE_GEM ); +qdev_set_nic_properties (DEVICE (&s->gem1), nd); + } + +object_property_set_int (OBJECT (&s->gem0),"revision" ,GEM_REVISION , errp); +object_property_set_int (OBJECT (&s->gem0),"phy-addr" ,8 , errp); +sysbus_realize (SYS_BUS_DEVICE (&s->gem0), errp); +sysbus_mmio_map (SYS_BUS_DEVICE (&s->gem0),0 , + memmap[MICROCHIP_PFSOC_GEM0 ].base); +sysbus_connect_irq (SYS_BUS_DEVICE (&s->gem0),0 , +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_GEM0_IRQ )); + +object_property_set_int (OBJECT (&s->gem1),"revision" ,GEM_REVISION , errp); +object_property_set_int (OBJECT (&s->gem1),"phy-addr" ,9 , errp); +sysbus_realize (SYS_BUS_DEVICE (&s->gem1), errp); +sysbus_mmio_map (SYS_BUS_DEVICE (&s->gem1),0 , + memmap[MICROCHIP_PFSOC_GEM1 ].base); +sysbus_connect_irq (SYS_BUS_DEVICE (&s->gem1),0 , +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_GEM1_IRQ )); +/* eNVM */ memory_region_init_rom(envm_data, OBJECT(dev), "microchip.pfsoc.envm.data", memmap[MICROCHIP_PFSOC_ENVM_DATA].size,
include/hw/riscv/microchip_pfsoc.h +7 -0
@@ -24,6 +24,7 @@#include "hw/char/mchp_pfsoc_mmuart.h" #include "hw/dma/mchp_pfsoc_dma.h" +#include "hw/net/cadence_gem.h" #include "hw/sd/cadence_sdhci.h" typedef struct MicrochipPFSoCState {
@@ -42,6 +43,8 @@ typedef struct MicrochipPFSoCState {MchpPfSoCMMUartState *serial3;MchpPfSoCMMUartState *serial4;MchpPfSoCDMAState dma; +CadenceGEMState gem0; +CadenceGEMState gem1;CadenceSDHCIState sdhci; } MicrochipPFSoCState;
@@ -84,6 +87,8 @@ enum {MICROCHIP_PFSOC_MMUART2 ,MICROCHIP_PFSOC_MMUART3 ,MICROCHIP_PFSOC_MMUART4 , +MICROCHIP_PFSOC_GEM0 , +MICROCHIP_PFSOC_GEM1 ,MICROCHIP_PFSOC_ENVM_CFG ,MICROCHIP_PFSOC_ENVM_DATA , MICROCHIP_PFSOC_IOSCB_CFG,
@@ -91,6 +96,8 @@ enum { };enum { +MICROCHIP_PFSOC_GEM0_IRQ =64 , +MICROCHIP_PFSOC_GEM1_IRQ =70 ,MICROCHIP_PFSOC_EMMC_SD_IRQ =88 ,MICROCHIP_PFSOC_MMUART0_IRQ =90 ,MICROCHIP_PFSOC_MMUART1_IRQ =91 ,
Signature · 1 line
2.7.4
Re: [PATCH 15/18] hw/riscv: microchip_pfsoc: Connect 2 Cadence GEMs
Quoted 6 lines · Bin Meng
From: Bin Meng <[email protected]>
Microchip PolarFire SoC integrates 2 Candence GEMs to provide IEEE 802.3 standard-compliant 10/100/1000 Mbps ethernet interface.
Quoted 8 lines
On the Icicle Kit board, GEM0 connects to a PHY at address 8 while GEM1 connects to a PHY at address 9.
The 2nd stage bootloader (U-Boot) is using GEM1 by default, so we must specify 2 '-nic' options from the command line in order to get a working ethernet.
Signed-off-by Bin Meng <[email protected]>
Alistair
Quoted 134 lines
---
hw/riscv/microchip_pfsoc.c | 39 ++++++++++++++++++++++++++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 7 +++++++ 2 files changed, 46 insertions(+)hw/riscv/microchip_pfsoc.c +39
-0@@ -14,6 +14,7 @@* 3) MMUARTs (Multi-Mode UART) * 4) Cadence eMMC/SDHC controller and an SD card connected to it * 5) DMA (Direct Memory Access Controller) +* 6) GEM (Gigabit Ethernet MAC Controller) * * This board currently generates devicetree dynamically that indicates at least * two harts and up to five harts. @@ -59,6 +60,9 @@#define BIOS_FILENAME "hss.bin"#define RESET_VECTOR 0x20220000 +/* GEM version */ +#define GEM_REVISION 0x0107010c + static const struct MemmapEntry { hwaddr base; hwaddr size;@@ -83,6 +87,8 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_MMUART2] = { 0x20102000, 0x1000 }, [MICROCHIP_PFSOC_MMUART3] = { 0x20104000, 0x1000 }, [MICROCHIP_PFSOC_MMUART4] = { 0x20106000, 0x1000 }, + [MICROCHIP_PFSOC_GEM0] = { 0x20110000, 0x2000 }, + [MICROCHIP_PFSOC_GEM1] = { 0x20112000, 0x2000 }, [MICROCHIP_PFSOC_ENVM_CFG] = { 0x20200000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_DATA] = { 0x20220000, 0x20000 }, [MICROCHIP_PFSOC_IOSCB_CFG] = { 0x37080000, 0x1000 },@@ -119,6 +125,9 @@ static void microchip_pfsoc_soc_instance_init(Object *obj)object_initialize_child (obj,"dma-controller" , &s->dma,TYPE_MCHP_PFSOC_DMA ); +object_initialize_child (obj,"gem0" , &s->gem0,TYPE_CADENCE_GEM ); +object_initialize_child (obj,"gem1" , &s->gem1,TYPE_CADENCE_GEM ); +object_initialize_child (obj,"sd-controller" , &s->sdhci,TYPE_CADENCE_SDHCI ); object_initialize_child(OBJECT(&s->sdhci), "sd-controller.sdhci",@@ -136,6 +145,7 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp)MemoryRegion *envm_data =g_new (MemoryRegion,1 );char *plic_hart_config;size_t plic_hart_config_len; +NICInfo *nd;int i;sysbus_realize (SYS_BUS_DEVICE (&s->e_cpus), &error_abort);@@ -269,6 +279,35 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp)qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_MMUART4_IRQ ), serial_hd(4)); +/* GEMs */ + + nd = &nd_table[0 ]; +if (nd->used) { +qemu_check_nic_model (nd,TYPE_CADENCE_GEM ); +qdev_set_nic_properties (DEVICE (&s->gem0), nd); + } + nd = &nd_table[1 ]; +if (nd->used) { +qemu_check_nic_model (nd,TYPE_CADENCE_GEM ); +qdev_set_nic_properties (DEVICE (&s->gem1), nd); + } + +object_property_set_int (OBJECT (&s->gem0),"revision" ,GEM_REVISION , errp); +object_property_set_int (OBJECT (&s->gem0),"phy-addr" ,8 , errp); +sysbus_realize (SYS_BUS_DEVICE (&s->gem0), errp); +sysbus_mmio_map (SYS_BUS_DEVICE (&s->gem0),0 , + memmap[MICROCHIP_PFSOC_GEM0 ].base); +sysbus_connect_irq (SYS_BUS_DEVICE (&s->gem0),0 , +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_GEM0_IRQ )); + +object_property_set_int (OBJECT (&s->gem1),"revision" ,GEM_REVISION , errp); +object_property_set_int (OBJECT (&s->gem1),"phy-addr" ,9 , errp); +sysbus_realize (SYS_BUS_DEVICE (&s->gem1), errp); +sysbus_mmio_map (SYS_BUS_DEVICE (&s->gem1),0 , + memmap[MICROCHIP_PFSOC_GEM1 ].base); +sysbus_connect_irq (SYS_BUS_DEVICE (&s->gem1),0 , +qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_GEM1_IRQ )); +/* eNVM */ memory_region_init_rom(envm_data, OBJECT(dev), "microchip.pfsoc.envm.data", memmap[MICROCHIP_PFSOC_ENVM_DATA].size,include/hw/riscv/microchip_pfsoc.h +7
-1@@ -24,6 +24,7 @@#include "hw/char/mchp_pfsoc_mmuart.h" #include "hw/dma/mchp_pfsoc_dma.h" +#include "hw/net/cadence_gem.h" #include "hw/sd/cadence_sdhci.h" typedef struct MicrochipPFSoCState {@@ -42,6 +43,8 @@ typedef struct MicrochipPFSoCState {MchpPfSoCMMUartState *serial3;MchpPfSoCMMUartState *serial4;MchpPfSoCDMAState dma; +CadenceGEMState gem0; +CadenceGEMState gem1;CadenceSDHCIState sdhci; } MicrochipPFSoCState;@@ -84,6 +87,8 @@ enum {MICROCHIP_PFSOC_MMUART2 ,MICROCHIP_PFSOC_MMUART3 ,MICROCHIP_PFSOC_MMUART4 , +MICROCHIP_PFSOC_GEM0 , +MICROCHIP_PFSOC_GEM1 ,MICROCHIP_PFSOC_ENVM_CFG ,MICROCHIP_PFSOC_ENVM_DATA , MICROCHIP_PFSOC_IOSCB_CFG,@@ -91,6 +96,8 @@ enum { };enum { +MICROCHIP_PFSOC_GEM0_IRQ =64 , +MICROCHIP_PFSOC_GEM1_IRQ =70 ,MICROCHIP_PFSOC_EMMC_SD_IRQ =88 ,MICROCHIP_PFSOC_MMUART0_IRQ =90 ,MICROCHIP_PFSOC_MMUART1_IRQ =91 , --2.7.4
PATCH 16/18 hw/riscv: microchip_pfsoc: Hook GPIO controllers
In reply to Bin MengFrom: Bin Meng <[email protected]>
Microchip PolarFire SoC integrates 3 GPIOs controllers. It seems enough to create unimplemented devices to cover their register spaces at this point.
With this commit, QEMU can boot to U-Boot (2nd stage bootloader) all the way to the Linux shell login prompt, with a modified HSS (1st stage bootloader).
---
hw/riscv/microchip_pfsoc.c | 14 ++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 3 +++ 2 files changed, 17 insertions(+)
hw/riscv/microchip_pfsoc.c +14 -0
@@ -89,6 +89,9 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_MMUART4] = { 0x20106000, 0x1000 }, [MICROCHIP_PFSOC_GEM0] = { 0x20110000, 0x2000 }, [MICROCHIP_PFSOC_GEM1] = { 0x20112000, 0x2000 }, + [MICROCHIP_PFSOC_GPIO0] = { 0x20120000, 0x1000 }, + [MICROCHIP_PFSOC_GPIO1] = { 0x20121000, 0x1000 }, + [MICROCHIP_PFSOC_GPIO2] = { 0x20122000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_CFG] = { 0x20200000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_DATA] = { 0x20220000, 0x20000 }, [MICROCHIP_PFSOC_IOSCB_CFG] = { 0x37080000, 0x1000 },
@@ -308,6 +311,17 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp)sysbus_connect_irq (SYS_BUS_DEVICE (&s->gem1),0 ,qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_GEM1_IRQ )); +/* GPIOs */ +create_unimplemented_device ("microchip.pfsoc.gpio0" , + memmap[MICROCHIP_PFSOC_GPIO0 ].base, + memmap[MICROCHIP_PFSOC_GPIO0 ].size); +create_unimplemented_device ("microchip.pfsoc.gpio1" , + memmap[MICROCHIP_PFSOC_GPIO1 ].base, + memmap[MICROCHIP_PFSOC_GPIO1 ].size); +create_unimplemented_device ("microchip.pfsoc.gpio2" , + memmap[MICROCHIP_PFSOC_GPIO2 ].base, + memmap[MICROCHIP_PFSOC_GPIO2 ].size); +/* eNVM */ memory_region_init_rom(envm_data, OBJECT(dev), "microchip.pfsoc.envm.data", memmap[MICROCHIP_PFSOC_ENVM_DATA].size,
include/hw/riscv/microchip_pfsoc.h +3 -0
@@ -89,6 +89,9 @@ enum {MICROCHIP_PFSOC_MMUART4 ,MICROCHIP_PFSOC_GEM0 ,MICROCHIP_PFSOC_GEM1 , +MICROCHIP_PFSOC_GPIO0 , +MICROCHIP_PFSOC_GPIO1 , +MICROCHIP_PFSOC_GPIO2 ,MICROCHIP_PFSOC_ENVM_CFG ,MICROCHIP_PFSOC_ENVM_DATA , MICROCHIP_PFSOC_IOSCB_CFG,
Signature · 1 line
2.7.4
Re: [PATCH 16/18] hw/riscv: microchip_pfsoc: Hook GPIO controllers
Quoted 7 lines · Bin Meng
From: Bin Meng <[email protected]>
Microchip PolarFire SoC integrates 3 GPIOs controllers. It seems enough to create unimplemented devices to cover their register spaces at this point.
Quoted 5 lines
With this commit, QEMU can boot to U-Boot (2nd stage bootloader) all the way to the Linux shell login prompt, with a modified HSS (1st stage bootloader).
Signed-off-by Bin Meng <[email protected]>
Alistair
Quoted 56 lines
---
hw/riscv/microchip_pfsoc.c | 14 ++++++++++++++ include/hw/riscv/microchip_pfsoc.h | 3 +++ 2 files changed, 17 insertions(+)hw/riscv/microchip_pfsoc.c +14
-0@@ -89,6 +89,9 @@ static const struct MemmapEntry { [MICROCHIP_PFSOC_MMUART4] = { 0x20106000, 0x1000 }, [MICROCHIP_PFSOC_GEM0] = { 0x20110000, 0x2000 }, [MICROCHIP_PFSOC_GEM1] = { 0x20112000, 0x2000 }, + [MICROCHIP_PFSOC_GPIO0] = { 0x20120000, 0x1000 }, + [MICROCHIP_PFSOC_GPIO1] = { 0x20121000, 0x1000 }, + [MICROCHIP_PFSOC_GPIO2] = { 0x20122000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_CFG] = { 0x20200000, 0x1000 }, [MICROCHIP_PFSOC_ENVM_DATA] = { 0x20220000, 0x20000 }, [MICROCHIP_PFSOC_IOSCB_CFG] = { 0x37080000, 0x1000 },@@ -308,6 +311,17 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp)sysbus_connect_irq (SYS_BUS_DEVICE (&s->gem1),0 ,qdev_get_gpio_in (DEVICE (s->plic),MICROCHIP_PFSOC_GEM1_IRQ )); +/* GPIOs */ +create_unimplemented_device ("microchip.pfsoc.gpio0" , + memmap[MICROCHIP_PFSOC_GPIO0 ].base, + memmap[MICROCHIP_PFSOC_GPIO0 ].size); +create_unimplemented_device ("microchip.pfsoc.gpio1" , + memmap[MICROCHIP_PFSOC_GPIO1 ].base, + memmap[MICROCHIP_PFSOC_GPIO1 ].size); +create_unimplemented_device ("microchip.pfsoc.gpio2" , + memmap[MICROCHIP_PFSOC_GPIO2 ].base, + memmap[MICROCHIP_PFSOC_GPIO2 ].size); +/* eNVM */ memory_region_init_rom(envm_data, OBJECT(dev), "microchip.pfsoc.envm.data", memmap[MICROCHIP_PFSOC_ENVM_DATA].size,include/hw/riscv/microchip_pfsoc.h +3
-1@@ -89,6 +89,9 @@ enum {MICROCHIP_PFSOC_MMUART4 ,MICROCHIP_PFSOC_GEM0 ,MICROCHIP_PFSOC_GEM1 , +MICROCHIP_PFSOC_GPIO0 , +MICROCHIP_PFSOC_GPIO1 , +MICROCHIP_PFSOC_GPIO2 ,MICROCHIP_PFSOC_ENVM_CFG ,MICROCHIP_PFSOC_ENVM_DATA , MICROCHIP_PFSOC_IOSCB_CFG, --2.7.4
PATCH 17/18 hw/riscv: clint: Avoid using hard-coded timebase frequency
In reply to Bin MengFrom: Bin Meng <[email protected]>
At present the CLINT timestamp is using a hard-coded timebase frequency value SIFIVE_CLINT_TIMEBASE_FREQ. This might not be true for all boards.
Add a new 'timebase-freq' property to the CLINT device, and update various functions to accept this as a parameter.
---
hw/riscv/microchip_pfsoc.c | 6 +++++- hw/riscv/sifive_clint.c | 25 ++++++++++++++----------- hw/riscv/sifive_e.c | 3 ++- hw/riscv/sifive_u.c | 3 ++- hw/riscv/spike.c | 2 +- hw/riscv/virt.c | 3 ++- include/hw/riscv/sifive_clint.h | 3 ++- target/riscv/cpu.h | 6 ++++-- target/riscv/cpu_helper.c | 4 +++- target/riscv/csr.c | 4 ++-- 10 files changed, 37 insertions(+), 22 deletions(-)
hw/riscv/microchip_pfsoc.c +5 -1
@@ -60,6 +60,9 @@#define BIOS_FILENAME "hss.bin"#define RESET_VECTOR 0x20220000 +/* CLINT timebase frequency */ +#define CLINT_TIMEBASE_FREQ 1000000 +/* GEM version */ #define GEM_REVISION 0x0107010c
@@ -189,7 +192,8 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp)/* CLINT */ sifive_clint_create (memmap[MICROCHIP_PFSOC_CLINT ].base, memmap[MICROCHIP_PFSOC_CLINT ].size, ms->smp.cpus, -SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , false); +SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , +CLINT_TIMEBASE_FREQ , false);/* L2 cache controller */ create_unimplemented_device("microchip.pfsoc.l2cc",
hw/riscv/sifive_clint.c +14 -11
@@ -29,22 +29,23 @@#include "hw/riscv/sifive_clint.h" #include "qemu/timer.h" -static uint64_t cpu_riscv_read_rtc (void ) +static uint64_t cpu_riscv_read_rtc (uint32_t timebase_freq) {return muldiv64 (qemu_clock_get_ns (QEMU_CLOCK_VIRTUAL ), -SIFIVE_CLINT_TIMEBASE_FREQ ,NANOSECONDS_PER_SECOND ); + timebase_freq,NANOSECONDS_PER_SECOND ); }/* * Called when timecmp is written to update the QEMU timer or immediately * trigger timer interrupt if mtimecmp <= current timer value. */ -static void sifive_clint_write_timecmp (RISCVCPU *cpu,uint64_t value) +static void sifive_clint_write_timecmp (RISCVCPU *cpu,uint64_t value, +uint32_t timebase_freq) {uint64_t next;uint64_t diff; -uint64_t rtc_r =cpu_riscv_read_rtc (); +uint64_t rtc_r =cpu_riscv_read_rtc (timebase_freq); cpu->env.timecmp = value;if (cpu->env.timecmp <= rtc_r) {
@@ -59,7 +60,7 @@ static void sifive_clint_write_timecmp(RISCVCPU *cpu, uint64_t value) diff = cpu->env.timecmp - rtc_r;/* back to ns (note args switched in muldiv64) */ next =qemu_clock_get_ns (QEMU_CLOCK_VIRTUAL ) + -muldiv64 (diff,NANOSECONDS_PER_SECOND ,SIFIVE_CLINT_TIMEBASE_FREQ ); +muldiv64 (diff,NANOSECONDS_PER_SECOND , timebase_freq);timer_mod (cpu->env.timer, next); }
@@ -111,10 +112,10 @@ static uint64_t sifive_clint_read(void *opaque, hwaddr addr, unsigned size) } }else if (addr == clint->time_base) {/* time_lo */ -return cpu_riscv_read_rtc () &0xFFFFFFFF ; +return cpu_riscv_read_rtc (clint->timebase_freq) &0xFFFFFFFF ; }else if (addr == clint->time_base +4 ) {/* time_hi */ -return (cpu_riscv_read_rtc () >>32 ) &0xFFFFFFFF ; +return (cpu_riscv_read_rtc (clint->timebase_freq) >>32 ) &0xFFFFFFFF ; }error_report ("clint: invalid read: %08x" , (uint32_t )addr);
@@ -151,13 +152,13 @@ static void sifive_clint_write(void *opaque, hwaddr addr, uint64_t value,/* timecmp_lo */ uint64_t timecmp_hi = env->timecmp >>32 ;sifive_clint_write_timecmp (RISCV_CPU (cpu), - timecmp_hi <<32 | (value &0xFFFFFFFF )); + timecmp_hi <<32 | (value &0xFFFFFFFF ), clint->timebase_freq);return ; }else if ((addr &0x7 ) ==4 ) {/* timecmp_hi */ uint64_t timecmp_lo = env->timecmp;sifive_clint_write_timecmp (RISCV_CPU (cpu), - value <<32 | (timecmp_lo &0xFFFFFFFF )); + value <<32 | (timecmp_lo &0xFFFFFFFF ), clint->timebase_freq); }else {error_report ("clint: invalid timecmp write: %08x" , (uint32_t )addr); }
@@ -191,6 +192,7 @@ static Property sifive_clint_properties[] = {DEFINE_PROP_UINT32 ("timecmp-base" , SiFiveCLINTState, timecmp_base,0 ),DEFINE_PROP_UINT32 ("time-base" , SiFiveCLINTState, time_base,0 ),DEFINE_PROP_UINT32 ("aperture-size" , SiFiveCLINTState, aperture_size,0 ), +DEFINE_PROP_UINT32 ("timebase-freq" , SiFiveCLINTState, timebase_freq,0 ), DEFINE_PROP_END_OF_LIST(), };
@@ -229,7 +231,7 @@ type_init(sifive_clint_register_types)*/ DeviceState *sifive_clint_create (hwaddr addr,hwaddr size,uint32_t num_harts,uint32_t sip_base,uint32_t timecmp_base,uint32_t time_base, -bool provide_rdtime) +uint32_t timebase_freq,bool provide_rdtime) {int i;for (i =0 ; i < num_harts; i++) {
@@ -239,7 +241,7 @@ DeviceState *sifive_clint_create(hwaddr addr, hwaddr size, uint32_t num_harts,continue ; }if (provide_rdtime) { -riscv_cpu_set_rdtime_fn (env, cpu_riscv_read_rtc); +riscv_cpu_set_rdtime_fn (env, cpu_riscv_read_rtc, timebase_freq); } env->timer =timer_new_ns (QEMU_CLOCK_VIRTUAL , &sifive_clint_timer_cb, cpu);
@@ -252,6 +254,7 @@ DeviceState *sifive_clint_create(hwaddr addr, hwaddr size, uint32_t num_harts,qdev_prop_set_uint32 (dev,"timecmp-base" , timecmp_base);qdev_prop_set_uint32 (dev,"time-base" , time_base);qdev_prop_set_uint32 (dev,"aperture-size" , size); +qdev_prop_set_uint32 (dev,"timebase-freq" , timebase_freq);sysbus_realize_and_unref (SYS_BUS_DEVICE (dev), &error_fatal);sysbus_mmio_map (SYS_BUS_DEVICE (dev),0 , addr);return dev;
hw/riscv/sifive_e.c +2 -1
@@ -213,7 +213,8 @@ static void sifive_e_soc_realize(DeviceState *dev, Error **errp) memmap[SIFIVE_E_PLIC].size);sifive_clint_create (memmap[SIFIVE_E_CLINT ].base, memmap[SIFIVE_E_CLINT ].size, ms->smp.cpus, -SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , false); +SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , +SIFIVE_CLINT_TIMEBASE_FREQ , false);create_unimplemented_device ("riscv.sifive.e.aon" , memmap[SIFIVE_E_AON ].base, memmap[SIFIVE_E_AON ].size);sifive_e_prci_create (memmap[SIFIVE_E_PRCI ].base);
hw/riscv/sifive_u.c +2 -1
@@ -706,7 +706,8 @@ static void sifive_u_soc_realize(DeviceState *dev, Error **errp) serial_hd(1), qdev_get_gpio_in(DEVICE(s->plic), SIFIVE_U_UART1_IRQ));sifive_clint_create (memmap[SIFIVE_U_CLINT ].base, memmap[SIFIVE_U_CLINT ].size, ms->smp.cpus, -SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , false); +SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , +SIFIVE_CLINT_TIMEBASE_FREQ , false);if (!sysbus_realize (SYS_BUS_DEVICE (&s->prci), errp)) {return ;
hw/riscv/spike.c +1 -1
@@ -235,7 +235,7 @@ static void spike_board_init(MachineState *machine)/* Core Local Interruptor (timer and IPI) */ sifive_clint_create (memmap[SPIKE_CLINT ].base, memmap[SPIKE_CLINT ].size, smp_cpus,SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , - false); +SIFIVE_CLINT_TIMEBASE_FREQ , false); }static void spike_machine_init (MachineClass *mc)
hw/riscv/virt.c +2 -1
@@ -572,7 +572,8 @@ static void virt_machine_init(MachineState *machine) memmap[VIRT_PLIC].size);sifive_clint_create (memmap[VIRT_CLINT ].base, memmap[VIRT_CLINT ].size, smp_cpus, -SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , true); +SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , +SIFIVE_CLINT_TIMEBASE_FREQ , true);sifive_test_create (memmap[VIRT_TEST ].base);for (i =0 ; i <VIRTIO_COUNT ; i++) {
include/hw/riscv/sifive_clint.h +2 -1
@@ -38,11 +38,12 @@ typedef struct SiFiveCLINTState {uint32_t timecmp_base;uint32_t time_base;uint32_t aperture_size; +uint32_t timebase_freq; } SiFiveCLINTState;DeviceState *sifive_clint_create (hwaddr addr,hwaddr size,uint32_t num_harts,uint32_t sip_base,uint32_t timecmp_base,uint32_t time_base, -bool provide_rdtime); +uint32_t timebase_freq,bool provide_rdtime);enum {SIFIVE_SIP_BASE =0x0 ,
target/riscv/cpu.h +4 -2
@@ -219,7 +219,8 @@ struct CPURISCVState {pmp_table_t pmp_state;/* machine specific rdtime callback */ -uint64_t (*rdtime_fn)(void ); +uint64_t (*rdtime_fn)(uint32_t ); +uint32_t rdtime_fn_arg;/* True if in debugger mode. */ bool debugger;
@@ -347,7 +348,8 @@ void riscv_cpu_swap_hypervisor_regs(CPURISCVState *env);int riscv_cpu_claim_interrupts (RISCVCPU *cpu,uint32_t interrupts);uint32_t riscv_cpu_update_mip (RISCVCPU *cpu,uint32_t mask,uint32_t value);#define BOOL_TO_MASK (x) (-!!(x))/* helper for riscv_cpu_update_mip value */ -void riscv_cpu_set_rdtime_fn (CPURISCVState *env,uint64_t (*fn)(void )); +void riscv_cpu_set_rdtime_fn (CPURISCVState *env,uint64_t (*fn)(uint32_t ), +uint32_t arg);#endif void riscv_cpu_set_mode (CPURISCVState *env,target_ulong newpriv);
target/riscv/cpu_helper.c +3 -1
@@ -258,9 +258,11 @@ uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t value)return old; } -void riscv_cpu_set_rdtime_fn (CPURISCVState *env,uint64_t (*fn)(void )) +void riscv_cpu_set_rdtime_fn (CPURISCVState *env,uint64_t (*fn)(uint32_t ), +uint32_t arg) { env->rdtime_fn = fn; + env->rdtime_fn_arg = arg; }void riscv_cpu_set_mode (CPURISCVState *env,target_ulong newpriv)
target/riscv/csr.c +2 -2
@@ -294,7 +294,7 @@ static int read_time(CPURISCVState *env, int csrno, target_ulong *val)return -1 ; } - *val = env->rdtime_fn () + delta; + *val = env->rdtime_fn (env->rdtime_fn_arg) + delta;return 0 ; }
@@ -307,7 +307,7 @@ static int read_timeh(CPURISCVState *env, int csrno, target_ulong *val) return -1; } - *val = (env->rdtime_fn () + delta) >>32 ; + *val = (env->rdtime_fn (env->rdtime_fn_arg) + delta) >>32 ;return 0 ; }#endif
Signature · 1 line
2.7.4
Re: [PATCH 17/18] hw/riscv: clint: Avoid using hard-coded timebase frequency
Quoted 11 lines · Bin Meng
From: Bin Meng <[email protected]>
At present the CLINT timestamp is using a hard-coded timebase frequency value
SIFIVE_CLINT_TIMEBASE_FREQ. This might not be true for all boards.Add a new 'timebase-freq' property to the CLINT device, and update various functions to accept this as a parameter.
Signed-off-by Bin Meng <[email protected]>
Alistair
Quoted 282 lines
---
hw/riscv/microchip_pfsoc.c | 6 +++++- hw/riscv/sifive_clint.c | 25 ++++++++++++++----------- hw/riscv/sifive_e.c | 3 ++- hw/riscv/sifive_u.c | 3 ++- hw/riscv/spike.c | 2 +- hw/riscv/virt.c | 3 ++- include/hw/riscv/sifive_clint.h | 3 ++- target/riscv/cpu.h | 6 ++++-- target/riscv/cpu_helper.c | 4 +++- target/riscv/csr.c | 4 ++-- 10 files changed, 37 insertions(+), 22 deletions(-)hw/riscv/microchip_pfsoc.c +5
-1@@ -60,6 +60,9 @@#define BIOS_FILENAME "hss.bin"#define RESET_VECTOR 0x20220000 +/* CLINT timebase frequency */ +#define CLINT_TIMEBASE_FREQ 1000000 +/* GEM version */ #define GEM_REVISION 0x0107010c@@ -189,7 +192,8 @@ static void microchip_pfsoc_soc_realize(DeviceState *dev, Error **errp)/* CLINT */ sifive_clint_create (memmap[MICROCHIP_PFSOC_CLINT ].base, memmap[MICROCHIP_PFSOC_CLINT ].size, ms->smp.cpus, -SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , false); +SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , +CLINT_TIMEBASE_FREQ , false);/* L2 cache controller */ create_unimplemented_device("microchip.pfsoc.l2cc",hw/riscv/sifive_clint.c +14
-11@@ -29,22 +29,23 @@#include "hw/riscv/sifive_clint.h" #include "qemu/timer.h" -static uint64_t cpu_riscv_read_rtc (void ) +static uint64_t cpu_riscv_read_rtc (uint32_t timebase_freq) {return muldiv64 (qemu_clock_get_ns (QEMU_CLOCK_VIRTUAL ), -SIFIVE_CLINT_TIMEBASE_FREQ ,NANOSECONDS_PER_SECOND ); + timebase_freq,NANOSECONDS_PER_SECOND ); }/* * Called when timecmp is written to update the QEMU timer or immediately * trigger timer interrupt if mtimecmp <= current timer value. */ -static void sifive_clint_write_timecmp (RISCVCPU *cpu,uint64_t value) +static void sifive_clint_write_timecmp (RISCVCPU *cpu,uint64_t value, +uint32_t timebase_freq) {uint64_t next;uint64_t diff; -uint64_t rtc_r =cpu_riscv_read_rtc (); +uint64_t rtc_r =cpu_riscv_read_rtc (timebase_freq); cpu->env.timecmp = value;if (cpu->env.timecmp <= rtc_r) {@@ -59,7 +60,7 @@ static void sifive_clint_write_timecmp(RISCVCPU *cpu, uint64_t value) diff = cpu->env.timecmp - rtc_r;/* back to ns (note args switched in muldiv64) */ next =qemu_clock_get_ns (QEMU_CLOCK_VIRTUAL ) + -muldiv64 (diff,NANOSECONDS_PER_SECOND ,SIFIVE_CLINT_TIMEBASE_FREQ ); +muldiv64 (diff,NANOSECONDS_PER_SECOND , timebase_freq);timer_mod (cpu->env.timer, next); }@@ -111,10 +112,10 @@ static uint64_t sifive_clint_read(void *opaque, hwaddr addr, unsigned size) } }else if (addr == clint->time_base) {/* time_lo */ -return cpu_riscv_read_rtc () &0xFFFFFFFF ; +return cpu_riscv_read_rtc (clint->timebase_freq) &0xFFFFFFFF ; }else if (addr == clint->time_base +4 ) {/* time_hi */ -return (cpu_riscv_read_rtc () >>32 ) &0xFFFFFFFF ; +return (cpu_riscv_read_rtc (clint->timebase_freq) >>32 ) &0xFFFFFFFF ; }error_report ("clint: invalid read: %08x" , (uint32_t )addr);@@ -151,13 +152,13 @@ static void sifive_clint_write(void *opaque, hwaddr addr, uint64_t value,/* timecmp_lo */ uint64_t timecmp_hi = env->timecmp >>32 ;sifive_clint_write_timecmp (RISCV_CPU (cpu), - timecmp_hi <<32 | (value &0xFFFFFFFF )); + timecmp_hi <<32 | (value &0xFFFFFFFF ), clint->timebase_freq);return ; }else if ((addr &0x7 ) ==4 ) {/* timecmp_hi */ uint64_t timecmp_lo = env->timecmp;sifive_clint_write_timecmp (RISCV_CPU (cpu), - value <<32 | (timecmp_lo &0xFFFFFFFF )); + value <<32 | (timecmp_lo &0xFFFFFFFF ), clint->timebase_freq); }else {error_report ("clint: invalid timecmp write: %08x" , (uint32_t )addr); }@@ -191,6 +192,7 @@ static Property sifive_clint_properties[] = {DEFINE_PROP_UINT32 ("timecmp-base" , SiFiveCLINTState, timecmp_base,0 ),DEFINE_PROP_UINT32 ("time-base" , SiFiveCLINTState, time_base,0 ),DEFINE_PROP_UINT32 ("aperture-size" , SiFiveCLINTState, aperture_size,0 ), +DEFINE_PROP_UINT32 ("timebase-freq" , SiFiveCLINTState, timebase_freq,0 ), DEFINE_PROP_END_OF_LIST(), };@@ -229,7 +231,7 @@ type_init(sifive_clint_register_types)*/ DeviceState *sifive_clint_create (hwaddr addr,hwaddr size,uint32_t num_harts,uint32_t sip_base,uint32_t timecmp_base,uint32_t time_base, -bool provide_rdtime) +uint32_t timebase_freq,bool provide_rdtime) {int i;for (i =0 ; i < num_harts; i++) {@@ -239,7 +241,7 @@ DeviceState *sifive_clint_create(hwaddr addr, hwaddr size, uint32_t num_harts,continue ; }if (provide_rdtime) { -riscv_cpu_set_rdtime_fn (env, cpu_riscv_read_rtc); +riscv_cpu_set_rdtime_fn (env, cpu_riscv_read_rtc, timebase_freq); } env->timer =timer_new_ns (QEMU_CLOCK_VIRTUAL , &sifive_clint_timer_cb, cpu);@@ -252,6 +254,7 @@ DeviceState *sifive_clint_create(hwaddr addr, hwaddr size, uint32_t num_harts,qdev_prop_set_uint32 (dev,"timecmp-base" , timecmp_base);qdev_prop_set_uint32 (dev,"time-base" , time_base);qdev_prop_set_uint32 (dev,"aperture-size" , size); +qdev_prop_set_uint32 (dev,"timebase-freq" , timebase_freq);sysbus_realize_and_unref (SYS_BUS_DEVICE (dev), &error_fatal);sysbus_mmio_map (SYS_BUS_DEVICE (dev),0 , addr);return dev;hw/riscv/sifive_e.c +2
-1@@ -213,7 +213,8 @@ static void sifive_e_soc_realize(DeviceState *dev, Error **errp) memmap[SIFIVE_E_PLIC].size);sifive_clint_create (memmap[SIFIVE_E_CLINT ].base, memmap[SIFIVE_E_CLINT ].size, ms->smp.cpus, -SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , false); +SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , +SIFIVE_CLINT_TIMEBASE_FREQ , false);create_unimplemented_device ("riscv.sifive.e.aon" , memmap[SIFIVE_E_AON ].base, memmap[SIFIVE_E_AON ].size);sifive_e_prci_create (memmap[SIFIVE_E_PRCI ].base);hw/riscv/sifive_u.c +2
-1@@ -706,7 +706,8 @@ static void sifive_u_soc_realize(DeviceState *dev, Error **errp) serial_hd(1), qdev_get_gpio_in(DEVICE(s->plic), SIFIVE_U_UART1_IRQ));sifive_clint_create (memmap[SIFIVE_U_CLINT ].base, memmap[SIFIVE_U_CLINT ].size, ms->smp.cpus, -SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , false); +SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , +SIFIVE_CLINT_TIMEBASE_FREQ , false);if (!sysbus_realize (SYS_BUS_DEVICE (&s->prci), errp)) {return ;hw/riscv/spike.c +1
-1@@ -235,7 +235,7 @@ static void spike_board_init(MachineState *machine)/* Core Local Interruptor (timer and IPI) */ sifive_clint_create (memmap[SPIKE_CLINT ].base, memmap[SPIKE_CLINT ].size, smp_cpus,SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , - false); +SIFIVE_CLINT_TIMEBASE_FREQ , false); }static void spike_machine_init (MachineClass *mc)hw/riscv/virt.c +2
-1@@ -572,7 +572,8 @@ static void virt_machine_init(MachineState *machine) memmap[VIRT_PLIC].size);sifive_clint_create (memmap[VIRT_CLINT ].base, memmap[VIRT_CLINT ].size, smp_cpus, -SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , true); +SIFIVE_SIP_BASE ,SIFIVE_TIMECMP_BASE ,SIFIVE_TIME_BASE , +SIFIVE_CLINT_TIMEBASE_FREQ , true);sifive_test_create (memmap[VIRT_TEST ].base);for (i =0 ; i <VIRTIO_COUNT ; i++) {include/hw/riscv/sifive_clint.h +2
-1@@ -38,11 +38,12 @@ typedef struct SiFiveCLINTState {uint32_t timecmp_base;uint32_t time_base;uint32_t aperture_size; +uint32_t timebase_freq; } SiFiveCLINTState;DeviceState *sifive_clint_create (hwaddr addr,hwaddr size,uint32_t num_harts,uint32_t sip_base,uint32_t timecmp_base,uint32_t time_base, -bool provide_rdtime); +uint32_t timebase_freq,bool provide_rdtime);enum {SIFIVE_SIP_BASE =0x0 ,target/riscv/cpu.h +4
-2@@ -219,7 +219,8 @@ struct CPURISCVState {pmp_table_t pmp_state;/* machine specific rdtime callback */ -uint64_t (*rdtime_fn)(void ); +uint64_t (*rdtime_fn)(uint32_t ); +uint32_t rdtime_fn_arg;/* True if in debugger mode. */ bool debugger;@@ -347,7 +348,8 @@ void riscv_cpu_swap_hypervisor_regs(CPURISCVState *env);int riscv_cpu_claim_interrupts (RISCVCPU *cpu,uint32_t interrupts);uint32_t riscv_cpu_update_mip (RISCVCPU *cpu,uint32_t mask,uint32_t value);#define BOOL_TO_MASK (x) (-!!(x))/* helper for riscv_cpu_update_mip value */ -void riscv_cpu_set_rdtime_fn (CPURISCVState *env,uint64_t (*fn)(void )); +void riscv_cpu_set_rdtime_fn (CPURISCVState *env,uint64_t (*fn)(uint32_t ), +uint32_t arg);#endif void riscv_cpu_set_mode (CPURISCVState *env,target_ulong newpriv);target/riscv/cpu_helper.c +3
-1@@ -258,9 +258,11 @@ uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t value)return old; } -void riscv_cpu_set_rdtime_fn (CPURISCVState *env,uint64_t (*fn)(void )) +void riscv_cpu_set_rdtime_fn (CPURISCVState *env,uint64_t (*fn)(uint32_t ), +uint32_t arg) { env->rdtime_fn = fn; + env->rdtime_fn_arg = arg; }void riscv_cpu_set_mode (CPURISCVState *env,target_ulong newpriv)target/riscv/csr.c +2
-3@@ -294,7 +294,7 @@ static int read_time(CPURISCVState *env, int csrno, target_ulong *val)return -1 ; } - *val = env->rdtime_fn () + delta; + *val = env->rdtime_fn (env->rdtime_fn_arg) + delta;return 0 ; }@@ -307,7 +307,7 @@ static int read_timeh(CPURISCVState *env, int csrno, target_ulong *val) return -1; } - *val = (env->rdtime_fn () + delta) >>32 ; + *val = (env->rdtime_fn (env->rdtime_fn_arg) + delta) >>32 ;return 0 ; }#endif --2.7.4
PATCH 18/18 hw/riscv: microchip_pfsoc: Document the software used for testing
In reply to Bin MengFrom: Bin Meng <[email protected]>
Add some useful comments to document the software used for testing. including how to patch HSS to bypass the DDR memory initialization, HSS and Yocto BSP build instructions, etc.
To launch this machine for testing:
$ qemu-system-riscv64 -M microchip-icicle-kit -smp 5 \
-bios path/to/hss.bin -sd path/to/sdcard.img \
-nic user,model=cadence_gem \
-nic tap,ifname=tap,model=cadence_gem \
-display none -serial stdio \
-chardev socket,id=serial1,path=serial1.sock,server,wait \
-serial chardev:serial1---
hw/riscv/microchip_pfsoc.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+)
hw/riscv/microchip_pfsoc.c +21 -0
@@ -56,6 +56,27 @@/* * The BIOS image used by this machine is called Hart Software Services (HSS). * See https://github.com/polarfire-soc/hart-software-services +* +* As of now the DDR memory controller in the Microchip PolarFire SoC has not +* been modeled. Simply creating unimplemented devices does not make HSS happy. +* Emulating the DDR memory controller is tedious, so a patched HSS should be +* used as the BIOS for this machine. +* +* To patch HSS, open boards/icicle-kit-es/hss_board_init.c in the HSS source +* tree, find the boardInitFunctions[] array that contains the initialization +* routines for this board, and remove the line that contains 'HSS_DDRInit'. +* +* QEMU does not support eMMC hence the SD configuration shall be used in the +* HSS and Yocto BSP build. The eMMC configuration is not supported. +* +* Instructions to build HSS: +* $ cp boards/icicle-kit-es/def_config.sdcard .config +* $ make BOARD=icicle-kit-es +* +* For Yocto build, "MACHINE=icicle-kit-es-sd" should be specified, otherwise +* when booting Linux kernel the rootfs cannot be mounted. The generated image +* is something like: mpfs-dev-cli-icicle-kit-es-sd.rootfs.wic. Resize the file +* with 'qemu-image' to a power of 2 before passing to QEMU '-sd' command line. */ #define BIOS_FILENAME "hss.bin"#define RESET_VECTOR 0x20220000
Signature · 1 line
2.7.4
Re: [PATCH 18/18] hw/riscv: microchip_pfsoc: Document the software used for testing
Quoted 42 lines · Bin Meng
From: Bin Meng <[email protected]>
Add some useful comments to document the software used for testing. including how to patch HSS to bypass the DDR memory initialization, HSS and Yocto BSP build instructions, etc.
To launch this machine for testing:
$ qemu-system-riscv64 -M microchip-icicle-kit -smp 5 \ -bios path/to/hss.bin -sd path/to/sdcard.img \ -nic user,model=cadence_gem \ -nic tap,ifname=tap,model=cadence_gem \ -display none -serial stdio \ -chardev socket,id=serial1,path=serial1.sock,server,wait \ -serial chardev:serial1Signed-off-by Bin Meng <[email protected]>---
hw/riscv/microchip_pfsoc.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+)hw/riscv/microchip_pfsoc.c +11
-0@@ -56,6 +56,27 @@/* * The BIOS image used by this machine is called Hart Software Services (HSS). * See https://github.com/polarfire-soc/hart-software-services +* +* As of now the DDR memory controller in the Microchip PolarFire SoC has not +* been modeled. Simply creating unimplemented devices does not make HSS happy. +* Emulating the DDR memory controller is tedious, so a patched HSS should be +* used as the BIOS for this machine. +* +* To patch HSS, open boards/icicle-kit-es/hss_board_init.c in the HSS source +* tree, find the boardInitFunctions[] array that contains the initialization +* routines for this board, and remove the line that contains 'HSS_DDRInit'. +* +* QEMU does not support eMMC hence the SD configuration shall be used in the
hw/riscv/microchip_pfsoc.c line 79
Quoted 10 lines
hw/riscv/microchip_pfsoc.c +10
-0+* HSS and Yocto BSP build. The eMMC configuration is not supported. +* +* Instructions to build HSS: +* $ cp boards/icicle-kit-es/def_config.sdcard .config +* $ make BOARD=icicle-kit-es +* +* For Yocto build, "MACHINE=icicle-kit-es-sd" should be specified, otherwise +* when booting Linux kernel the rootfs cannot be mounted. The generated image +* is something like: mpfs-dev-cli-icicle-kit-es-sd.rootfs.wic. Resize the file +* with 'qemu-image' to a power of 2 before passing to QEMU '-sd' command line.
Instead of adding comments in the code (where it's hard for users to find and prone to getting out of date) can you drop this patch and instead update the RISC-V QEMU wiki?
https://wiki.qemu.org/Documentation/Platforms/RISCV
Just add an Icicle section.
Alistair
Quoted 7 lines
hw/riscv/microchip_pfsoc.c +0
-1*/ #define BIOS_FILENAME "hss.bin"#define RESET_VECTOR 0x20220000 --2.7.4
Quoted 6 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Quoted 34 lines
For more details about Microchip PolarFire Soc, please see:
https://www.microsemi.com/product-directory/soc-fpgas/5498-polarfire-soc-fpgaThe Icicle Kit board information can be found here:
https://www.microsemi.com/existing-parts/parts/152514Unlike SiFive FU540, the RISC-V core resect vector is at
0x20220000. The RISC-V CPU and HART codes has been updated to set the core's reset vector based on a configurable property from machine codes.The following perepherals are created as an unimplemented device:
- Bus Error Uint 0/1/2/3/4
- L2 cache controller
- SYSREG
- MPUCFG
- IOSCBCFG
- GPIOThe following perepherals are emulated:
- SiFive CLINT
- SiFive PLIC
- PolarFire SoC Multi-Mode UART
- PolarFire SoC DMA
- Cadence eMMC/SDHCI controller
- Cadence Gigabit Ethernet MACSome bugs in the SD card codes are fixed during the development.
The BIOS image used by this machine is hss.bin, aka Hart Software Services, which can be built from:
https://github.com/polarfire-soc/hart-software-services
Quoted 7 lines
To launch this machine:
$ qemu-system-riscv64 -M microchip-icicle-kit -smp 5 \ -bios path/to/hss.bin -sd path/to/sdcard.img \ -nic tap,ifname=tap,script=no,model=cadence_gem \ -display none -serial stdio \ -chardev socket,id=serial1,path=serial1.sock,server,wait \ -serial chardev:serial1
Currently, it is fine to use HSS (with OpenSBI v0.6 as a library) but this is not aligned with the existing booting flow of many RISC-V systems.
It will be nice to have standard U-Boot RISC-V boot-flow working on Microchip PolarFire SoC:
U-Boot SPL (BIOS) => FW_DYNAMIC (Generic) => U-Boot S-mode
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE support
I think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI FW_DYNAMIC.
Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Do you have plans for adding U-Boot SPL support for this board ??
Regards,
Anup
Quoted 84 lines
The memory is set to 1 GiB by default to match the hardware. A sanity check on ram size is performed in the machine init routine to prompt user to increase the RAM size to > 1 GiB when less than 1 GiB ram is detected.
HSS output is on the first serial port (stdio) and U-Boot/Linux outputs on the 2nd serial port. OpenSBI outputs on a random serial port due to the lottery mechanism used during the multi-core boot.
Bin Meng (18):
target/riscv: cpu: Add a new 'resetvec' property
hw/riscv: hart: Add a new 'resetvec' property target/riscv: cpu: Set reset vector based on the configured property
value
hw/riscv: Initial support for Microchip PolarFire SoC Icicle Kit board hw/char: Add Microchip PolarFire SoC MMUART emulation hw/riscv:microchip_pfsoc: Connect 5 MMUARTs
hw/sd: sd: Fix incorrect populated function switch status data
structure
hw/sd: sd: Correctly set the high capacity bit
hw/sd: sdhci: Makesdhci_poweron_reset()internal visible hw/sd: Add Cadence SDHCI emulation
hw/riscv:microchip_pfsoc: Connect a Cadence SDHCI controller and an
SD card
hw/dma: Add Microchip PolarFire Soc DMA controller emulation hw/riscv:microchip_pfsoc: Connect a DMA controller hw/net:cadence_gem: Add a new 'phy-addr' property hw/riscv:microchip_pfsoc: Connect 2 Cadence GEMs hw/riscv:microchip_pfsoc: Hook GPIO controllers
hw/riscv: clint: Avoid using hard-coded timebase frequency hw/riscv:microchip_pfsoc: Document the software used for testingMAINTAINERS | 11 + default-configs/riscv64-softmmu.mak | 1 + hw/char/Kconfig | 3 + hw/char/Makefile.objs | 1 + hw/char/mchp_pfsoc_mmuart.c | 82 +++++++ hw/dma/Kconfig | 3 + hw/dma/Makefile.objs | 1 + hw/dma/mchp_pfsoc_dma.c | 322 +++++++++++++++++++++++++ hw/net/cadence_gem.c | 7 +- hw/riscv/Kconfig | 9 + hw/riscv/Makefile.objs | 1 + hw/riscv/microchip_pfsoc.c | 456 ++++++++++++++++++++++++++++++++++++ hw/riscv/opentitan.c | 1 + hw/riscv/riscv_hart.c | 3 + hw/riscv/sifive_clint.c | 25 +- hw/riscv/sifive_e.c | 4 +- hw/riscv/sifive_u.c | 5 +- hw/riscv/spike.c | 2 +- hw/riscv/virt.c | 3 +- hw/sd/Kconfig | 4 + hw/sd/Makefile.objs | 1 + hw/sd/cadence_sdhci.c | 162 +++++++++++++ hw/sd/sd.c | 8 +- hw/sd/sdhci-internal.h | 1 + hw/sd/sdhci.c | 2 +- include/hw/char/mchp_pfsoc_mmuart.h | 61 +++++ include/hw/dma/mchp_pfsoc_dma.h | 57 +++++ include/hw/net/cadence_gem.h | 2 + include/hw/riscv/microchip_pfsoc.h | 125 ++++++++++ include/hw/riscv/riscv_hart.h | 1 + include/hw/riscv/sifive_clint.h | 3 +- include/hw/sd/cadence_sdhci.h | 65 +++++ target/riscv/cpu.c | 8 +- target/riscv/cpu.h | 7 +- target/riscv/cpu_helper.c | 4 +- target/riscv/csr.c | 4 +- 36 files changed, 1424 insertions(+), 31 deletions(-) create mode 100644 hw/char/mchp_pfsoc_mmuart.c create mode 100644 hw/dma/mchp_pfsoc_dma.c create mode 100644 hw/riscv/microchip_pfsoc.c create mode 100644 hw/sd/cadence_sdhci.c create mode 100644 include/hw/char/mchp_pfsoc_mmuart.h create mode 100644 include/hw/dma/mchp_pfsoc_dma.h create mode 100644 include/hw/riscv/microchip_pfsoc.h create mode 100644 include/hw/sd/cadence_sdhci.h--
2.7.4
Hi Anup,
Quoted 10 lines · Anup Patel
Quoted 6 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 43 lines
Quoted 35 lines
For more details about Microchip PolarFire Soc, please see:
https://www.microsemi.com/product-directory/soc-fpgas/5498-polarfire-soc-fpgaThe Icicle Kit board information can be found here:
https://www.microsemi.com/existing-parts/parts/152514Unlike SiFive FU540, the RISC-V core resect vector is at
0x20220000. The RISC-V CPU and HART codes has been updated to set the core's reset vector based on a configurable property from machine codes.The following perepherals are created as an unimplemented device:
- Bus Error Uint 0/1/2/3/4
- L2 cache controller
- SYSREG
- MPUCFG
- IOSCBCFG
- GPIOThe following perepherals are emulated:
- SiFive CLINT
- SiFive PLIC
- PolarFire SoC Multi-Mode UART
- PolarFire SoC DMA
- Cadence eMMC/SDHCI controller
- Cadence Gigabit Ethernet MACSome bugs in the SD card codes are fixed during the development.
The BIOS image used by this machine is hss.bin, aka Hart Software Services, which can be built from:
https://github.com/polarfire-soc/hart-software-servicesQuoted 7 lines
To launch this machine:
$ qemu-system-riscv64 -M microchip-icicle-kit -smp 5 \ -bios path/to/hss.bin -sd path/to/sdcard.img \ -nic tap,ifname=tap,script=no,model=cadence_gem \ -display none -serial stdio \ -chardev socket,id=serial1,path=serial1.sock,server,wait \ -serial chardev:serial1
Quoted 3 lines
Currently, it is fine to use HSS (with OpenSBI v0.6 as a library) but this is not aligned with the existing booting flow of many RISC-V systems.
Yep, unfortunately this is the case currently.
Quoted 5 lines
It will be nice to have standard U-Boot RISC-V boot-flow working on Microchip PolarFire SoC:
U-Boot SPL (BIOS) =>FW_DYNAMIC(Generic) => U-Boot S-mode
Agreed.
Quoted 9 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.
Quoted 6 lines
Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
Regards,
Bin
Hi Anup,
Quoted 28 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Hi Anup,
Quoted 7 lines · Anup Patel
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 14 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Quoted 5 lines
Quoted 1 line
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
I currently do not have a plan to support U-Boot SPL. The idea of the HSS is to contain all the silicon specific initialization and configuration code within the HSS before jumping to U-Boot S-mode. I would rather keep all this within the HSS for the time being. I would wait until we reach production silicon before attempting to move this to U-Boot SPL as the HSS is likely to contain some opaque silicon related changes for another while.
Regards,
Cyril.
Quoted 37 lines · via
Hi Anup,
Quoted 28 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Hi Anup,
Quoted 7 lines · Anup Patel
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 14 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Quoted 5 lines
Quoted 1 line
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
Quoted 7 lines
I currently do not have a plan to support U-Boot SPL. The idea of the HSS is to contain all the silicon specific initialization and configuration code within the HSS before jumping to U-Boot S-mode. I would rather keep all this within the HSS for the time being. I would wait until we reach production silicon before attempting to move this to U-Boot SPL as the HSS is likely to contain some opaque silicon related changes for another while.
That is unfortunate, a lot of work has gone into making the boot flow simple and easy to use.
QEMU now includes OpenSBI by default to make it easy for users to boot Linux. The Icicle Kit board is now the most difficult QEMU board to boot Linux on. Not to mention it makes it hard to impossible to support it in standard tool flows such as meta-riscv.
Alistair
Quoted 6 lines
Regards,
Cyril.
Quoted 49 lines · Alistair Francis
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 36 lines · via
Hi Anup,
Quoted 28 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Hi Anup,
Quoted 7 lines · Anup Patel
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 14 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Quoted 5 lines
Quoted 1 line
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
Quoted 7 lines
I currently do not have a plan to support U-Boot SPL. The idea of the HSS is to contain all the silicon specific initialization and configuration code within the HSS before jumping to U-Boot S-mode. I would rather keep all this within the HSS for the time being. I would wait until we reach production silicon before attempting to move this to U-Boot SPL as the HSS is likely to contain some opaque silicon related changes for another while.
That is unfortunate, a lot of work has gone into making the boot flow simple and easy to use.
Quoted 6 lines
QEMU now includes OpenSBI by default to make it easy for users to boot Linux. The Icicle Kit board is now the most difficult QEMU board to boot Linux on. Not to mention it makes it hard to impossible to support it in standard tool flows such as meta-riscv.
Alistair
If it is such a problem we can add a U-Boot SPL stage and the HSS can be treated as standard SoC ROM code.
Cyril.
Quoted 56 lines · On Tue, Aug 18, 2020 at 1:23 AM <[email protected]> wrote:
Quoted 54 lines · Alistair Francis
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 36 lines · via
Hi Anup,
Quoted 28 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Hi Anup,
Quoted 7 lines · Anup Patel
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 14 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Quoted 5 lines
Quoted 1 line
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
Quoted 7 lines
I currently do not have a plan to support U-Boot SPL. The idea of the HSS is to contain all the silicon specific initialization and configuration code within the HSS before jumping to U-Boot S-mode. I would rather keep all this within the HSS for the time being. I would wait until we reach production silicon before attempting to move this to U-Boot SPL as the HSS is likely to contain some opaque silicon related changes for another while.
That is unfortunate, a lot of work has gone into making the boot flow simple and easy to use.
QEMU now includes OpenSBI by default to make it easy for users to boot Linux. The Icicle Kit board is now the most difficult QEMU board to boot Linux on. Not to mention it makes it hard to impossible to support it in standard tool flows such as meta-riscv.
Quoted 4 lines
Quoted 1 line
Alistair
If it is such a problem we can add a U-Boot SPL stage and the HSS can be treated as standard SoC ROM code.
It's not mandatory for U-Boot SPL to have stable DRAM calibration settings from the start itself. The initial U-Boot SPL support for most platforms (accross
architectures) usually include basic working DRAM calibration settings which is later on updated with separate patches. Also, we don't need all U-Boot drivers to be upstreamed in one-go as this can be done in phases.
The advantage we have with PolarFire SoC Icicle board is that we already have a U-Boot S-mode. and we believe the OpenSBI generic platform will work fine for PolarFire SoC Icicle board so the only thing missing right now is the U-Boot SPL support for OpenSource boot-flow.
It will certainly accelerate open-source development if we have boot-flow U-Boot_SPL => OpenSBI (FW_DYNAMIC) => U-Boot_S-mode working on PolarFire SoC Icicle board. Initially, we just need DRAM, SD/eMMC, and Serial port support for U-Boot SPL and U-Boot S-mode. Later on, more patches can add ethernet and other booting device drivers to U-Boot.
Regarding security services of HSS, we are working on a OpenSBI feature which will allow HSS security services to run as independent binary in M-mode (not linked to OpenSBI) and OpenSBI FW_DYNAMIC will be a separate binary acting as a secure monitor.
Regards,
Anup
Quoted 79 lines · Anup Patel
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 50 lines · On Tue, Aug 18, 2020 at 1:23 AM <[email protected]> wrote:
Quoted 49 lines · Alistair Francis
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 36 lines · via
Hi Anup,
Quoted 28 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Hi Anup,
Quoted 7 lines · Anup Patel
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 14 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Quoted 5 lines
Quoted 1 line
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
Quoted 7 lines
I currently do not have a plan to support U-Boot SPL. The idea of the HSS is to contain all the silicon specific initialization and configuration code within the HSS before jumping to U-Boot S-mode. I would rather keep all this within the HSS for the time being. I would wait until we reach production silicon before attempting to move this to U-Boot SPL as the HSS is likely to contain some opaque silicon related changes for another while.
That is unfortunate, a lot of work has gone into making the boot flow simple and easy to use.
Quoted 8 lines
Quoted 6 lines
QEMU now includes OpenSBI by default to make it easy for users to boot Linux. The Icicle Kit board is now the most difficult QEMU board to boot Linux on. Not to mention it makes it hard to impossible to support it in standard tool flows such as meta-riscv.
Alistair
If it is such a problem we can add a U-Boot SPL stage and the HSS can be treated as standard SoC ROM code.
It's not mandatory for U-Boot SPL to have stable DRAM calibration settings from the start itself. The initial U-Boot SPL support for most platforms (accross
architectures) usually include basic working DRAM calibration settings which is later on updated with separate patches. Also, we don't need all U-Boot drivers to be upstreamed in one-go as this can be done in phases.The advantage we have with PolarFire SoC Icicle board is that we already have a U-Boot S-mode. and we believe the OpenSBI generic platform will work fine for PolarFire SoC Icicle board so the only thing missing right now is the U-Boot SPL support for OpenSource boot-flow.
It will certainly accelerate open-source development if we have boot-flow U-Boot_SPL => OpenSBI (
FW_DYNAMIC) => U-Boot_S-mode working on PolarFire SoC Icicle board. Initially, we just need DRAM, SD/eMMC, and Serial port support for U-Boot SPL and U-Boot S-mode. Later on, more patches can add ethernet and other booting device drivers to U-Boot.
Quoted 7 lines
Regarding security services of HSS, we are working on a OpenSBI feature which will allow HSS security services to run as independent binary in M-mode (not linked to OpenSBI) and OpenSBI
FW_DYNAMICwill be a separate binary acting as a secure monitor.Regards,
Anup
What I have in mind is that the external memory will be up and running by the time we get to U-Boot SPL. In the case of PolarFire SoC the ROM code equivalent brings up the DDR memory interface so we do not need to worry about this as part of U-Boot.
Sounds to me the component that needs to be upstreamed next is the eMMC support so that it can be used as part of U-Boot SPL. Correct?
Regards,
Cyril.
Quoted 86 lines · On Tue, Aug 18, 2020 at 6:39 PM <[email protected]> wrote:
Quoted 84 lines · Anup Patel
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 50 lines · On Tue, Aug 18, 2020 at 1:23 AM <[email protected]> wrote:
Quoted 49 lines · Alistair Francis
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 36 lines · via
Hi Anup,
Quoted 28 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Hi Anup,
Quoted 7 lines · Anup Patel
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 14 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Quoted 5 lines
Quoted 1 line
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
Quoted 7 lines
I currently do not have a plan to support U-Boot SPL. The idea of the HSS is to contain all the silicon specific initialization and configuration code within the HSS before jumping to U-Boot S-mode. I would rather keep all this within the HSS for the time being. I would wait until we reach production silicon before attempting to move this to U-Boot SPL as the HSS is likely to contain some opaque silicon related changes for another while.
That is unfortunate, a lot of work has gone into making the boot flow simple and easy to use.
Quoted 8 lines
Quoted 6 lines
QEMU now includes OpenSBI by default to make it easy for users to boot Linux. The Icicle Kit board is now the most difficult QEMU board to boot Linux on. Not to mention it makes it hard to impossible to support it in standard tool flows such as meta-riscv.
Alistair
If it is such a problem we can add a U-Boot SPL stage and the HSS can be treated as standard SoC ROM code.
It's not mandatory for U-Boot SPL to have stable DRAM calibration settings from the start itself. The initial U-Boot SPL support for most platforms (accross
architectures) usually include basic working DRAM calibration settings which is later on updated with separate patches. Also, we don't need all U-Boot drivers to be upstreamed in one-go as this can be done in phases.The advantage we have with PolarFire SoC Icicle board is that we already have a U-Boot S-mode. and we believe the OpenSBI generic platform will work fine for PolarFire SoC Icicle board so the only thing missing right now is the U-Boot SPL support for OpenSource boot-flow.
It will certainly accelerate open-source development if we have boot-flow U-Boot_SPL => OpenSBI (
FW_DYNAMIC) => U-Boot_S-mode working on PolarFire SoC Icicle board. Initially, we just need DRAM, SD/eMMC, and Serial port support for U-Boot SPL and U-Boot S-mode. Later on, more patches can add ethernet and other booting device drivers to U-Boot.Regarding security services of HSS, we are working on a OpenSBI feature which will allow HSS security services to run as independent binary in M-mode (not linked to OpenSBI) and OpenSBI
FW_DYNAMICwill be a separate binary acting as a secure monitor.
Quoted 7 lines
Quoted 2 lines
Regards,
AnupWhat I have in mind is that the external memory will be up and running by the time we get to U-Boot SPL. In the case of PolarFire SoC the ROM code equivalent brings up the DDR memory interface so we do not need to worry about this as part of U-Boot.
Keeping DRAM configuration as part of a separate ROM booting stage prior to the U-Boot SPL sounds good to me. This will lead to following boot-flow:
ROM/HSS (M-mode) => U-Boot SPL (M-mode) => OpenSBI (M-mode) => U-Boot S-mode
Quoted 3 lines
Sounds to me the component that needs to be upstreamed next is the eMMC support so that it can be used as part of U-Boot SPL. Correct?
Maybe as a PHASE1 patch series for PolarFire U-Boot support can target the following:
1. Minimal U-Boot board support for PolarFire IcIcle Board with single defconfig to build both U-Boot SPL and U-Boot S-mode
2. Serial port driver (probably re-use existing driver)
3. SD and eMMC driver
Supporting SD booting is highly desirable for PHASE1. At least, U-Boot SPL, OpenSBI, U-Boot S-mode, and Linux should come as separate images from SD card. The ROM/HSS can reside and boot from on-board flash/eMMC.
Above is my suggestion based on experience with SiFive Unleashed.
The general idea behind OpenSource boot-flow is to have each booting stage as a separate binary so that users can selectively upgrade a particular booting stage without re-compiling/re-flashing other booting stages.
Various distros and yocto already support building most of the above booting stages (U-Boot, OpenSBI, Linux, etc) so we can reuse a lot of existing work. Eventually, I am confident we will endup moving more stuff from ROM/HSS to U-Boot SPL for ease in maintenance.
Regards,
Anup
Quoted 96 lines · Anup Patel
Quoted 86 lines · On Tue, Aug 18, 2020 at 6:39 PM <[email protected]> wrote:
Quoted 84 lines · Anup Patel
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 50 lines · On Tue, Aug 18, 2020 at 1:23 AM <[email protected]> wrote:
Quoted 49 lines · Alistair Francis
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 36 lines · via
Hi Anup,
Quoted 28 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Hi Anup,
Quoted 7 lines · Anup Patel
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 14 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Quoted 5 lines
Quoted 1 line
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
Quoted 7 lines
I currently do not have a plan to support U-Boot SPL. The idea of the HSS is to contain all the silicon specific initialization and configuration code within the HSS before jumping to U-Boot S-mode. I would rather keep all this within the HSS for the time being. I would wait until we reach production silicon before attempting to move this to U-Boot SPL as the HSS is likely to contain some opaque silicon related changes for another while.
That is unfortunate, a lot of work has gone into making the boot flow simple and easy to use.
Quoted 8 lines
Quoted 6 lines
QEMU now includes OpenSBI by default to make it easy for users to boot Linux. The Icicle Kit board is now the most difficult QEMU board to boot Linux on. Not to mention it makes it hard to impossible to support it in standard tool flows such as meta-riscv.
Alistair
If it is such a problem we can add a U-Boot SPL stage and the HSS can be treated as standard SoC ROM code.
It's not mandatory for U-Boot SPL to have stable DRAM calibration settings from the start itself. The initial U-Boot SPL support for most platforms (accross
architectures) usually include basic working DRAM calibration settings which is later on updated with separate patches. Also, we don't need all U-Boot drivers to be upstreamed in one-go as this can be done in phases.The advantage we have with PolarFire SoC Icicle board is that we already have a U-Boot S-mode. and we believe the OpenSBI generic platform will work fine for PolarFire SoC Icicle board so the only thing missing right now is the U-Boot SPL support for OpenSource boot-flow.
It will certainly accelerate open-source development if we have boot-flow U-Boot_SPL => OpenSBI (
FW_DYNAMIC) => U-Boot_S-mode working on PolarFire SoC Icicle board. Initially, we just need DRAM, SD/eMMC, and Serial port support for U-Boot SPL and U-Boot S-mode. Later on, more patches can add ethernet and other booting device drivers to U-Boot.Regarding security services of HSS, we are working on a OpenSBI feature which will allow HSS security services to run as independent binary in M-mode (not linked to OpenSBI) and OpenSBI
FW_DYNAMICwill be a separate binary acting as a secure monitor.Quoted 7 lines
Quoted 2 lines
Regards,
AnupWhat I have in mind is that the external memory will be up and running by the time we get to U-Boot SPL. In the case of PolarFire SoC the ROM code equivalent brings up the DDR memory interface so we do not need to worry about this as part of U-Boot.
Quoted 4 lines
Keeping DRAM configuration as part of a separate ROM booting stage prior to the U-Boot SPL sounds good to me. This will lead to following boot-flow:
ROM/HSS (M-mode) => U-Boot SPL (M-mode) => OpenSBI (M-mode) => U-Boot S-mode
If we want to keep the HSS, meaning that DDR memory is already initialized, then there is no need to create an additional step of U-Boot SPL phase, because there is no size limitation any more.
Right now, the boot workflow on PolarFire is:
HSS + OpenSBI (as a library) (M-mode) => U-Boot S-mode
It's just not the canonical way because of the HSS and OpenSBI as a library.
A canonical way should be:
U-Boot SPL (M-mode) => OpenSBI dynamic (M-mode) => U-Boot S-mode
So we should be aiming to replace HSS with U-Boot SPL.
Quoted 24 lines
Quoted 4 lines
Sounds to me the component that needs to be upstreamed next is the eMMC support so that it can be used as part of U-Boot SPL. Correct?
Maybe as a PHASE1 patch series for PolarFire U-Boot support can target the following:
1. Minimal U-Boot board support for PolarFire IcIcle Board with single defconfig to build both U-Boot SPL and U-Boot S-mode
2. Serial port driver (probably re-use existing driver)
3. SD and eMMC driverSupporting SD booting is highly desirable for PHASE1. At least, U-Boot SPL, OpenSBI, U-Boot S-mode, and Linux should come as separate images from SD card. The ROM/HSS can reside and boot from on-board flash/eMMC.
Above is my suggestion based on experience with SiFive Unleashed.
The general idea behind OpenSource boot-flow is to have each booting stage as a separate binary so that users can selectively upgrade a particular booting stage without re-compiling/re-flashing other booting stages.
Quoted 4 lines
Various distros and yocto already support building most of the above booting stages (U-Boot, OpenSBI, Linux, etc) so we can reuse a lot of existing work. Eventually, I am confident we will endup moving more stuff from ROM/HSS to U-Boot SPL for ease in maintenance.
Regards,
Bin
Quoted 136 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 88 lines · Anup Patel
Quoted 87 lines · On Tue, Aug 18, 2020 at 6:39 PM <[email protected]> wrote:
Quoted 86 lines · Anup Patel
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 50 lines · On Tue, Aug 18, 2020 at 1:23 AM <[email protected]> wrote:
Quoted 49 lines · Alistair Francis
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 36 lines · via
Hi Anup,
Quoted 28 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Hi Anup,
Quoted 7 lines · Anup Patel
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 14 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Quoted 5 lines
Quoted 1 line
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
Quoted 7 lines
I currently do not have a plan to support U-Boot SPL. The idea of the HSS is to contain all the silicon specific initialization and configuration code within the HSS before jumping to U-Boot S-mode. I would rather keep all this within the HSS for the time being. I would wait until we reach production silicon before attempting to move this to U-Boot SPL as the HSS is likely to contain some opaque silicon related changes for another while.
That is unfortunate, a lot of work has gone into making the boot flow simple and easy to use.
Quoted 8 lines
Quoted 6 lines
QEMU now includes OpenSBI by default to make it easy for users to boot Linux. The Icicle Kit board is now the most difficult QEMU board to boot Linux on. Not to mention it makes it hard to impossible to support it in standard tool flows such as meta-riscv.
Alistair
If it is such a problem we can add a U-Boot SPL stage and the HSS can be treated as standard SoC ROM code.
It's not mandatory for U-Boot SPL to have stable DRAM calibration settings from the start itself. The initial U-Boot SPL support for most platforms (accross
architectures) usually include basic working DRAM calibration settings which is later on updated with separate patches. Also, we don't need all U-Boot drivers to be upstreamed in one-go as this can be done in phases.The advantage we have with PolarFire SoC Icicle board is that we already have a U-Boot S-mode. and we believe the OpenSBI generic platform will work fine for PolarFire SoC Icicle board so the only thing missing right now is the U-Boot SPL support for OpenSource boot-flow.
It will certainly accelerate open-source development if we have boot-flow U-Boot_SPL => OpenSBI (
FW_DYNAMIC) => U-Boot_S-mode working on PolarFire SoC Icicle board. Initially, we just need DRAM, SD/eMMC, and Serial port support for U-Boot SPL and U-Boot S-mode. Later on, more patches can add ethernet and other booting device drivers to U-Boot.Regarding security services of HSS, we are working on a OpenSBI feature which will allow HSS security services to run as independent binary in M-mode (not linked to OpenSBI) and OpenSBI
FW_DYNAMICwill be a separate binary acting as a secure monitor.Regards,
AnupQuoted 8 lines
Quoted 4 lines
What I have in mind is that the external memory will be up and running by the time we get to U-Boot SPL. In the case of PolarFire SoC the ROM code equivalent brings up the DDR memory interface so we do not need to worry about this as part of U-Boot.
Keeping DRAM configuration as part of a separate ROM booting stage prior to the U-Boot SPL sounds good to me. This will lead to following boot-flow:
ROM/HSS (M-mode) => U-Boot SPL (M-mode) => OpenSBI (M-mode) => U-Boot S-mode
If we want to keep the HSS, meaning that DDR memory is already initialized, then there is no need to create an additional step of U-Boot SPL phase, because there is no size limitation any more.
Right now, the boot workflow on PolarFire is:
HSS + OpenSBI (as a library) (M-mode) => U-Boot S-mode
It's just not the canonical way because of the HSS and OpenSBI as a library.
A canonical way should be:
U-Boot SPL (M-mode) => OpenSBI dynamic (M-mode) => U-Boot S-mode
So we should be aiming to replace HSS with U-Boot SPL.
Quoted 21 lines
Quoted 2 lines
Sounds to me the component that needs to be upstreamed next is the eMMC support so that it can be used as part of U-Boot SPL. Correct?
Maybe as a PHASE1 patch series for PolarFire U-Boot support can target the following:
1. Minimal U-Boot board support for PolarFire IcIcle Board with single defconfig to build both U-Boot SPL and U-Boot S-mode
2. Serial port driver (probably re-use existing driver)
3. SD and eMMC driverSupporting SD booting is highly desirable for PHASE1. At least, U-Boot SPL, OpenSBI, U-Boot S-mode, and Linux should come as separate images from SD card. The ROM/HSS can reside and boot from on-board flash/eMMC.
Above is my suggestion based on experience with SiFive Unleashed.
The general idea behind OpenSource boot-flow is to have each booting stage as a separate binary so that users can selectively upgrade a particular booting stage without re-compiling/re-flashing other booting stages.
Quoted 6 lines
Quoted 4 lines
Various distros and yocto already support building most of the above booting stages (U-Boot, OpenSBI, Linux, etc) so we can reuse a lot of existing work. Eventually, I am confident we will endup moving more stuff from ROM/HSS to U-Boot SPL for ease in maintenance.
Regards,
Bin
I am OK with supporting U-Boot SPL as a community boot flow but the HSS is not going to disappear completely for many PolarFire SoC deployments. We need to have the HSS running on the monitor core after the system has booted for various reasons. Some of the current HSS functionalities can be migrated at a later time.
The HSS was designed to easily enable and disable features through Kconfig which should allow to assign features to different boot stages depending on the overall system deployment. Keeping the HSS as ROM code capable of starting various early boot stages in either M mode or S mode allows us (Microchip engineering and non pure Linux users) to mix different operating systems on the same platform and more easily debug hardware bring-up.
What I am trying to get at is that we need to balance the QEMU ease of use with the ability to bring-up and debug actual hardware. My preferred approach is to keep all the hardware bring up in one place, the HSS, so that you do not have to worry about obscure silicon idiosyncrasies once you reach the open source boot flow. Silicon/board bring up steps/features can be brought out of the HSS into the open source boot stages once these steps are mature.
PolarFire SoC is the third family of SoC FPGA I have worked on. So I have a fair idea of the amount of pain in front of me to bring the entire device family to market :-) We decided early on in the architecture of PolarFire SoC that we would make all code executing on the RISC-V processors open source and user modifiable. This includes all code typically hidden away in your typical SoC ROM code. This is why PolarFire SoC has on-chip flash memory used to store the HSS and why the HSS should be thought as including ROM code. The current Icicle Kit uses engineering samples from the first device in the PolarFire SoC family. So we are really at the beginning of the process of validating very early boot steps on a significant number of devices. I do not want to expose the open source community to the pain of bringing up a family of devices because of the decision to open source all code. So I am asking you to not rush in trying to migrate components from the HSS to a later boot stage just yet. It can be done but based on experience I think that now is not the right time yet.
Regards,
Cyril.
Quoted 151 lines · On Wed, Aug 19, 2020 at 3:13 AM <[email protected]> wrote:
Quoted 136 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 88 lines · Anup Patel
Quoted 87 lines · On Tue, Aug 18, 2020 at 6:39 PM <[email protected]> wrote:
Quoted 86 lines · Anup Patel
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 50 lines · On Tue, Aug 18, 2020 at 1:23 AM <[email protected]> wrote:
Quoted 49 lines · Alistair Francis
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 36 lines · via
Hi Anup,
Quoted 28 lines · Bin Meng
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Hi Anup,
Quoted 7 lines · Anup Patel
Quoted 5 lines · Bin Meng
From: Bin Meng <[email protected]>
This adds support for Microchip PolarFire SoC Icicle Kit board. The Icicle Kit board integrates a PolarFire SoC, with one SiFive's E51 plus four U54 cores and many on-chip peripherals and an FPGA.
Nice Work !!! This is very helpful.
Thanks!
Quoted 14 lines
The Microchip HSS is quite convoluted. It has:
1. DDR Init
2. Boot device support
3. SBI support using OpenSBI as library
4. Simple TEE supportI think point 1) and 2) above should be part of U-Boot SPL. The point 3) can be OpenSBI
FW_DYNAMIC.Lastly,for point 4), we are working on a new OpenSBI feature using which we can run independent Secure OS and Non-Secure OS using U-Boot_SPL+OpenSBI (for both SiFive Unleashed and Microchip PolarFire).
Quoted 5 lines
Quoted 1 line
Do you have plans for adding U-Boot SPL support for this board ??
+ Cyril Jean from Microchip
I will have to leave this question to Cyril to comment.
Quoted 7 lines
I currently do not have a plan to support U-Boot SPL. The idea of the HSS is to contain all the silicon specific initialization and configuration code within the HSS before jumping to U-Boot S-mode. I would rather keep all this within the HSS for the time being. I would wait until we reach production silicon before attempting to move this to U-Boot SPL as the HSS is likely to contain some opaque silicon related changes for another while.
That is unfortunate, a lot of work has gone into making the boot flow simple and easy to use.
Quoted 8 lines
Quoted 6 lines
QEMU now includes OpenSBI by default to make it easy for users to boot Linux. The Icicle Kit board is now the most difficult QEMU board to boot Linux on. Not to mention it makes it hard to impossible to support it in standard tool flows such as meta-riscv.
Alistair
If it is such a problem we can add a U-Boot SPL stage and the HSS can be treated as standard SoC ROM code.
It's not mandatory for U-Boot SPL to have stable DRAM calibration settings from the start itself. The initial U-Boot SPL support for most platforms (accross
architectures) usually include basic working DRAM calibration settings which is later on updated with separate patches. Also, we don't need all U-Boot drivers to be upstreamed in one-go as this can be done in phases.The advantage we have with PolarFire SoC Icicle board is that we already have a U-Boot S-mode. and we believe the OpenSBI generic platform will work fine for PolarFire SoC Icicle board so the only thing missing right now is the U-Boot SPL support for OpenSource boot-flow.
It will certainly accelerate open-source development if we have boot-flow U-Boot_SPL => OpenSBI (
FW_DYNAMIC) => U-Boot_S-mode working on PolarFire SoC Icicle board. Initially, we just need DRAM, SD/eMMC, and Serial port support for U-Boot SPL and U-Boot S-mode. Later on, more patches can add ethernet and other booting device drivers to U-Boot.Regarding security services of HSS, we are working on a OpenSBI feature which will allow HSS security services to run as independent binary in M-mode (not linked to OpenSBI) and OpenSBI
FW_DYNAMICwill be a separate binary acting as a secure monitor.Regards,
AnupQuoted 8 lines
Quoted 4 lines
What I have in mind is that the external memory will be up and running by the time we get to U-Boot SPL. In the case of PolarFire SoC the ROM code equivalent brings up the DDR memory interface so we do not need to worry about this as part of U-Boot.
Keeping DRAM configuration as part of a separate ROM booting stage prior to the U-Boot SPL sounds good to me. This will lead to following boot-flow:
ROM/HSS (M-mode) => U-Boot SPL (M-mode) => OpenSBI (M-mode) => U-Boot S-mode
If we want to keep the HSS, meaning that DDR memory is already initialized, then there is no need to create an additional step of U-Boot SPL phase, because there is no size limitation any more.
Right now, the boot workflow on PolarFire is:
HSS + OpenSBI (as a library) (M-mode) => U-Boot S-mode
It's just not the canonical way because of the HSS and OpenSBI as a library.
A canonical way should be:
U-Boot SPL (M-mode) => OpenSBI dynamic (M-mode) => U-Boot S-mode
So we should be aiming to replace HSS with U-Boot SPL.
Quoted 21 lines
Quoted 2 lines
Sounds to me the component that needs to be upstreamed next is the eMMC support so that it can be used as part of U-Boot SPL. Correct?
Maybe as a PHASE1 patch series for PolarFire U-Boot support can target the following:
1. Minimal U-Boot board support for PolarFire IcIcle Board with single defconfig to build both U-Boot SPL and U-Boot S-mode
2. Serial port driver (probably re-use existing driver)
3. SD and eMMC driverSupporting SD booting is highly desirable for PHASE1. At least, U-Boot SPL, OpenSBI, U-Boot S-mode, and Linux should come as separate images from SD card. The ROM/HSS can reside and boot from on-board flash/eMMC.
Above is my suggestion based on experience with SiFive Unleashed.
The general idea behind OpenSource boot-flow is to have each booting stage as a separate binary so that users can selectively upgrade a particular booting stage without re-compiling/re-flashing other booting stages.
Quoted 6 lines
Quoted 4 lines
Various distros and yocto already support building most of the above booting stages (U-Boot, OpenSBI, Linux, etc) so we can reuse a lot of existing work. Eventually, I am confident we will endup moving more stuff from ROM/HSS to U-Boot SPL for ease in maintenance.
Regards,
BinI am OK with supporting U-Boot SPL as a community boot flow but the HSS is not going to disappear completely for many PolarFire SoC deployments. We need to have the HSS running on the monitor core after the system has booted for various reasons. Some of the current HSS functionalities can be migrated at a later time.
Quoted 7 lines
The HSS was designed to easily enable and disable features through
Kconfigwhich should allow to assign features to different boot stages depending on the overall system deployment. Keeping the HSS as ROM code capable of starting various early boot stages in either M mode or S mode allows us (Microchip engineering and non pure Linux users) to mix different operating systems on the same platform and more easily debug hardware bring-up.
Thanks for explaining this.
I have changed the CC list a bit to be a little more targeted.
Quoted 3 lines
What I am trying to get at is that we need to balance the QEMU ease of use with the ability to bring-up and debug actual hardware. My preferred
Agreed. I don't think anyone has been saying that these changes should be for QEMU only. All of the changes talked about apply to both the QEMU model and the hardware.
Quoted 3 lines
approach is to keep all the hardware bring up in one place, the HSS, so that you do not have to worry about obscure silicon idiosyncrasies once you reach the open source boot flow. Silicon/board bring up
That is a noble goal, but doesn't HSS currently pull in the OpenSBI library, so it's doing more than just dealing with silicon bugs. As well as that HSS is not supported by distros, so distros can't build or ship it. For example OE won't build it in the meta-riscv layer, meaning that the Icle board just isn't supported. This means the board misses out on core work, for example it won't be able to utilise the RISC-V AGL work that has started.
Also, how are users expected to update HSS as bugs and vulnerabilities are found if it isn't shipped by distros?
Quoted 8 lines
steps/features can be brought out of the HSS into the open source boot stages once these steps are mature.
PolarFire SoC is the third family of SoC FPGA I have worked on. So I have a fair idea of the amount of pain in front of me to bring the entire device family to market :-) We decided early on in the architecture of PolarFire SoC that we would make all code executing on the RISC-V processors open source and user modifiable. This includes all
Awesome!!! That is so great to hear.
Quoted 1 line
code typically hidden away in your typical SoC ROM code. This is why
Again, this is just awesome!
Quoted 2 lines
PolarFire SoC has on-chip flash memory used to store the HSS and why the HSS should be thought as including ROM code. The current Icicle Kit uses
If we could change it to be more like a traditional ROM in terms of features then that would be really helpful. In that case we wouldn't need it for QEMU and only use it for hardware. It would only handle basic hardware boot up and then hand it over to U-Boot SPL/OpenSBI.
Quoted 5 lines
engineering samples from the first device in the PolarFire SoC family. So we are really at the beginning of the process of validating very early boot steps on a significant number of devices. I do not want to expose the open source community to the pain of bringing up a family of devices because of the decision to open source all code. So I am asking
I understand what you mean. In saying that it is usually easiest to start with the goal of upstreaming and then work towards that. That way all of the architectures decisions are made in that context. Maybe at first you have a fork while everything is in flux, but then it's really helpful to have the work upstreamed.
I think we have all seen too many "open source" projects that just fork everything, dump their code and never contribute back. Unfortunately when that happens we end up with duplicate work and frustrated engineers. As well as that the original project misses out on all the advantages of open source as they aren't benefiting from others work as well.
Quoted 3 lines
you to not rush in trying to migrate components from the HSS to a later boot stage just yet. It can be done but based on experience I think that now is not the right time yet.
The open source focus is great to hear. I think we are all really glad that MicroChip is doing this. It's a great first step, but it doesn't end there.
Alistair
Quoted 5 lines
Regards,
Cyril.
Patchew URL: https://patchew.org/QEMU/[email protected]/
Patchew URL: https://patchew.org/QEMU/[email protected]/
Hi,
This series failed the docker-quick@centos7 build test. Please find the testing commands and their output below. If you have Docker installed, you can probably reproduce it locally.
=== TEST SCRIPT BEGIN ===
#!/bin/bash
make docker-image-centos7 V=1 NETWORK=1
time make docker-test-quick@centos7 SHOW_ENV=1 J=14 NETWORK=1 === TEST SCRIPT END ===
TEST check-unit: tests/test-char
Unexpected error in object_property_try_add() at /tmp/qemu-test/src/qom/object.c:1181:
attempt to add duplicate property 'serial-id' to object (type 'container')
ERROR test-char - too few tests run (expected 38, got 9)
make: *** [check-unit] Error 1
make: *** Waiting for unfinished jobs....
TEST check-qtest-x86_64: tests/qtest/hd-geo-test
qemu-system-aarch64: -accel kvm: invalid accelerator kvm
---
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['sudo', '-n', 'docker', 'run', '--label', 'com.qemu.instance.uuid=5d87ed4b32104d3fa300f89248e59809', '-u', '1003', '--security-opt', 'seccomp=unconfined', '--rm', '-e', 'TARGET_LIST=', '-e', 'EXTRA_CONFIGURE_OPTS=', '-e', 'V=', '-e', 'J=14', '-e', 'DEBUG=', '-e', 'SHOW_ENV=1', '-e', 'CCACHE_DIR=/var/tmp/ccache', '-v', '/home/patchew2/.cache/qemu-docker-ccache:/var/tmp/ccache:z', '-v', '/var/tmp/patchew-tester-tmp-okrr9kbl/src/docker-src.2020-08-14-13.56.55.12251:/var/tmp/qemu:z,ro', 'qemu/centos7', '/var/tmp/qemu/run', 'test-quick']' returned non-zero exit status 2. filter=--filter=label=com.qemu.instance.uuid=5d87ed4b32104d3fa300f89248e59809
make[1]: *** [docker-run] Error 1
make[1]: Leaving directory `/var/tmp/patchew-tester-tmp-okrr9kbl/src'
make: *** [docker-run-test-quick@centos7] Error 2
real 13m35.300s
user 0m8.266s
The full log is available at
http://patchew.org/logs/[email protected]/testing.docker-quick@centos7/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]