lib: utils: Implement "64bit-mmio" property parsing
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
From: Guo Ren <[email protected]>
Figure out CLINT has_64bit_mmio from DT node and using antonym for compatibility.
--- lib/utils/fdt/fdt_helper.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
lib/utils/fdt/fdt_helper.c +2 -1
@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */ clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , &count)) + clint->has_64bit_mmio = FALSE;return 0 ; }
Signature · 1 line
2.7.4
PATCH 2/2 lib: utils: reset: Add T-HEAD sample platform reset driver
From: Guo Ren <[email protected]>
This driver is for T-HEAD test chip, fpga. It could work with all T-HEAD riscv processors: C9xx series.
example1: (Using io-regs for reset)
reset: reset-sample {
compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; entry-reg = <0xff 0xff019050>; entry-cnt = <4>; control-reg = <0xff 0xff015004>; control-val = <0x1c>; csr-copy = <0x7f3 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc>; };
example2: (Using csr-regs for reset)
reset: reset-sample {
compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = <0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0>; };
example3: (Only delegate plic enable to S-mode)
reset: reset-sample {
compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; };
After this patch, all T-HEAD c9xx would use platform/generic with fw_dynamic as default:
CROSS_COMPILE=riscv64-linux-gnu- PLATFORM=generic FW_PIC=y /usr/bin/make
dts full example:
cpus {
#address-cells = <1>;
#size-cells = <0>;
timebase-frequency = <0x2dc6c0>;
cpu at 0 {
device_type = "cpu";
reg = <0>;
status = "okay";
compatible = "riscv";
riscv,isa = "rv64imafdcsu";
mmu-type = "riscv,sv39";
cpu0_intc: interrupt-controller {
#interrupt-cells = <1>;
compatible = "riscv,cpu-intc";
interrupt-controller;
};
};
cpu at 1 {
device_type = "cpu";
reg = <1>;
status = "fail";
compatible = "riscv";
riscv,isa = "rv64imafdcsu";
mmu-type = "riscv,sv39";
cpu1_intc: interrupt-controller {
#interrupt-cells = <1>;
compatible = "riscv,cpu-intc";
interrupt-controller;
};
};
cpu at 2 {
device_type = "cpu";
reg = <2>;
status = "fail";
compatible = "riscv";
riscv,isa = "rv64imafdcsu";
mmu-type = "riscv,sv39";
cpu2_intc: interrupt-controller {
#interrupt-cells = <1>;
compatible = "riscv,cpu-intc";
interrupt-controller;
};
};
cpu at 3 {
device_type = "cpu";
reg = <3>;
status = "fail";
compatible = "riscv";
riscv,isa = "rv64imafdcsu";
mmu-type = "riscv,sv39";
cpu3_intc: interrupt-controller {
#interrupt-cells = <1>;
compatible = "riscv,cpu-intc";
interrupt-controller;
};
};
};
soc {
#address-cells = <2>;
#size-cells = <2>;
compatible = "simple-bus";
ranges;
reset: reset-sample {
compatible = "thead,reset-sample";
plic-delegate = <0xff 0xd81ffffc>;
using-csr-reset;
csr-copy = <
0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc
0x3b0 0x3b1 0x3b2 0x3b3
0x3b4 0x3b5 0x3b6 0x3b7
0x3a0
>;
};
clint0: clint at ffdc000000 {
compatible = "riscv,clint0";
interrupts-extended = <
&cpu0_intc 3 &cpu0_intc 7
&cpu1_intc 3 &cpu1_intc 7
&cpu2_intc 3 &cpu2_intc 7
&cpu3_intc 3 &cpu3_intc 7
&cpu4_intc 3 &cpu4_intc 7
>;
reg = <0xff 0xdc000000 0x0 0x04000000>;
clint,has-no-64bit-mmio;
};
intc: interrupt-controller at ffd8000000 {
#interrupt-cells = <1>;
compatible = "riscv,plic0";
interrupt-controller;
interrupts-extended = <
&cpu0_intc 0xffffffff &cpu0_intc 9
&cpu1_intc 0xffffffff &cpu1_intc 9
&cpu2_intc 0xffffffff &cpu2_intc 9
&cpu3_intc 0xffffffff &cpu3_intc 9
>;
reg = <0xff 0xd8000000 0x0 0x04000000>;
reg-names = "control";
riscv,max-priority = <7>;
riscv,ndev = <80>;
};
}The platform/thead will be deprecated.
--- lib/utils/reset/fdt_reset.c | 2 + lib/utils/reset/fdt_reset_thead.c | 117 ++++++++++++++++++++++++++++++++++ lib/utils/reset/fdt_reset_thead.h | 25 ++++++++ lib/utils/reset/fdt_reset_thead_asm.S | 47 ++++++++++++++ lib/utils/reset/objects.mk | 2 + 5 files changed, 193 insertions(+) create mode 100644 lib/utils/reset/fdt_reset_thead.c create mode 100644 lib/utils/reset/fdt_reset_thead.h create mode 100644 lib/utils/reset/fdt_reset_thead_asm.S
lib/utils/reset/fdt_reset.c +2 -0
@@ -13,10 +13,12 @@extern struct fdt_reset fdt_reset_sifive;extern struct fdt_reset fdt_reset_htif; +extern struct fdt_reset fdt_reset_thead;static struct fdt_reset *reset_drivers[] = { &fdt_reset_sifive, &fdt_reset_htif, + &fdt_reset_thead, };static struct fdt_reset *current_driver =NULL ;
lib/utils/reset/fdt_reset_thead.c new file +117 -0
117 lines not on this page · raw message
lib/utils/reset/fdt_reset_thead.h new file +25 -0
@@ -0,0 +1,25 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause +*/ + +#define MAX_CUSTOM_CSR 32 + +#ifndef __ASSEMBLER__ +#include <libfdt.h> +#include <sbi/riscv_io.h> +#include <sbi/sbi_bitops.h> +#include <sbi/sbi_hart.h> +#include <sbi/sbi_scratch.h> +#include <sbi_utils/fdt/fdt_helper.h> +#include <sbi_utils/reset/fdt_reset.h> + +struct custom_csr { +unsigned long index; +unsigned long value; +}; + +u64 __fdt_reset_thead_csrr (void ); + +extern struct custom_csr custom_csr[MAX_CUSTOM_CSR ]; +extern u32 __reset_thead_csr_stub[]; +#endif
lib/utils/reset/fdt_reset_thead_asm.S new file +47 -0
47 lines not on this page · raw message
lib/utils/reset/objects.mk +2 -0
@@ -10,3 +10,5 @@ libsbiutils-objs-y += reset/fdt_reset.o libsbiutils-objs-y += reset/fdt_reset_htif.o libsbiutils-objs-y += reset/fdt_reset_sifive.o +libsbiutils-objs-y += reset/fdt_reset_thead.o +libsbiutils-objs-y += reset/fdt_reset_thead_asm.o
Signature · 1 line
2.7.4
PATCH 2/2 lib: utils: reset: Add T-HEAD sample platform reset driver
ping kindly ... Any problems with the patch, please let me know.
Quoted 401 lines · On Tue, Apr 13, 2021 at 11:45 AM <[email protected]> wrote:
From: Guo Ren <[email protected]>
This driver is for T-HEAD test chip, fpga. It could work with all T-HEAD riscv processors: C9xx series.
example1: (Using io-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; entry-reg = <0xff 0xff019050>; entry-cnt = <4>; control-reg = <0xff 0xff015004>; control-val = <0x1c>; csr-copy = <0x7f3 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc>; };example2: (Using csr-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = <0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0>; };example3: (Only delegate plic enable to S-mode)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; };After this patch, all T-HEAD c9xx would use platform/generic with
fw_dynamicas default:CROSS_COMPILE=riscv64-linux-gnu- PLATFORM=generic FW_PIC=y
/usr/bin/makedts full example:
cpus { #address-cells = <1>; #size-cells = <0>; timebase-frequency = <0x2dc6c0>; cpu at 0 { device_type = "cpu"; reg = <0>; status = "okay"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu0_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 1 { device_type = "cpu"; reg = <1>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu1_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 2 { device_type = "cpu"; reg = <2>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu2_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 3 { device_type = "cpu"; reg = <3>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu3_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; }; soc { #address-cells = <2>; #size-cells = <2>; compatible = "simple-bus"; ranges; reset: reset-sample { compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = < 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0 >; }; clint0: clint at ffdc000000 { compatible = "riscv,clint0"; interrupts-extended = < &cpu0_intc 3 &cpu0_intc 7 &cpu1_intc 3 &cpu1_intc 7 &cpu2_intc 3 &cpu2_intc 7 &cpu3_intc 3 &cpu3_intc 7 &cpu4_intc 3 &cpu4_intc 7 >; reg = <0xff 0xdc000000 0x0 0x04000000>; clint,has-no-64bit-mmio; }; intc: interrupt-controller at ffd8000000 { #interrupt-cells = <1>; compatible = "riscv,plic0"; interrupt-controller; interrupts-extended = < &cpu0_intc 0xffffffff &cpu0_intc 9 &cpu1_intc 0xffffffff &cpu1_intc 9 &cpu2_intc 0xffffffff &cpu2_intc 9 &cpu3_intc 0xffffffff &cpu3_intc 9 >; reg = <0xff 0xd8000000 0x0 0x04000000>; reg-names = "control"; riscv,max-priority = <7>; riscv,ndev = <80>; }; }The platform/thead will be deprecated.
Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Atish Patra <[email protected]>Cc Xiang W <[email protected]>Cc Bin Meng <[email protected]>--- lib/utils/reset/fdt_reset.c | 2 + lib/utils/reset/fdt_reset_thead.c | 117 ++++++++++++++++++++++++++++++++++ lib/utils/reset/fdt_reset_thead.h | 25 ++++++++ lib/utils/reset/fdt_reset_thead_asm.S | 47 ++++++++++++++ lib/utils/reset/objects.mk | 2 + 5 files changed, 193 insertions(+) create mode 100644 lib/utils/reset/fdt_reset_thead.c create mode 100644 lib/utils/reset/fdt_reset_thead.h create mode 100644 lib/utils/reset/fdt_reset_thead_asm.Slib/utils/reset/fdt_reset.c +2
-0@@ -13,10 +13,12 @@extern struct fdt_reset fdt_reset_sifive;extern struct fdt_reset fdt_reset_htif; +extern struct fdt_reset fdt_reset_thead;static struct fdt_reset *reset_drivers[] = { &fdt_reset_sifive, &fdt_reset_htif, + &fdt_reset_thead, };static struct fdt_reset *current_driver =NULL ;lib/utils/reset/fdt_reset_thead.c new file +117
-0@@ -0,0 +1,117 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause +*/ + +#include "fdt_reset_thead.h" + +struct custom_csr custom_csr[MAX_CUSTOM_CSR ]; + +#define CSR_OPCODE 0x39073 +static void clone_csrs (int cnt) +{ +unsigned long i; + +for (i =0 ; i < cnt; i++) { +/* Write csr BIT[31 - 20] to stub */ + __reset_thead_csr_stub[3 *i +1 ] = +CSR_OPCODE | (custom_csr[i].index <<20 ); + +/* Mask csr BIT[31 - 20] */ + *(u32 *)&__fdt_reset_thead_csrr &=BIT (20 ) -1 ; +smp_mb (); + +/* Write csr BIT[31 - 20] to __fdt_reset_thead_csrr */ + *(u32 *)&__fdt_reset_thead_csrr |= custom_csr[i].index <<20 ; +smp_mb (); + + __asm__ __volatile__("fence.i\n" :::"memory" ); + + custom_csr[i].value =__fdt_reset_thead_csrr (); + } +} + +extern void __thead_pre_start_warm (void ); +static int thead_reset_init (void *fdt,int nodeoff, +const struct fdt_match *match) +{ +void *p; +const fdt64_t *val; +const fdt32_t *val_w; +int len, i, cnt; +u32 tmp =0 ; + +/* Prepare clone csrs */ + val_w =fdt_getprop (fdt, nodeoff,"csr-copy" , &len); +if (len >0 && val_w) { + cnt = len /sizeof (fdt32_t); + +if (cnt >MAX_CUSTOM_CSR ) +sbi_hart_hang (); + +for (i =0 ; i < cnt; i++) { + custom_csr[i].index =fdt32_to_cpu (val_w[i]); + } + } + +if (cnt) +clone_csrs (cnt); + +/* Delegate plic enable regs for S-mode */ + val =fdt_getprop (fdt, nodeoff,"plic-delegate" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); +writel (BIT (0 ), p); + } + +/* Old reset method for secondary harts */ +if (fdt_getprop (fdt, nodeoff,"using-csr-reset" , &len)) { +csr_write (0x7c7 , (u64)&__thead_pre_start_warm); +csr_write (0x7c6 ,-1 ); + } + +/* Custom reset method for secondary harts */ + val =fdt_getprop (fdt, nodeoff,"entry-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w =fdt_getprop (fdt, nodeoff,"entry-cnt" , &len); +if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + +for (i =0 ; i < tmp; i++) { +writel ((u64)(&__thead_pre_start_warm), p +8 *i); +writel ((u64)(&__thead_pre_start_warm) >>32 , p +8 *i +4 ); + } + } + + val =fdt_getprop (fdt, nodeoff,"control-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w =fdt_getprop (fdt, nodeoff,"control-val" , &len); +if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + tmp |=readl (p); +writel (tmp, p); + } + } + } + +return 0 ; +} + +void thead_system_reset (u32 type,u32 reason) +{ + __asm__ __volatile__("ebreak\n" ); +} + +static const struct fdt_match thead_reset_match[] = { + { .compatible ="thead,reset-sample" }, + { }, +}; + +struct fdt_reset fdt_reset_thead = { + .match_table = thead_reset_match, + .init = thead_reset_init, + .system_reset = thead_system_reset +};lib/utils/reset/fdt_reset_thead.h new file +25
-0@@ -0,0 +1,25 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause +*/ + +#define MAX_CUSTOM_CSR 32 + +#ifndef __ASSEMBLER__ +#include <libfdt.h> +#include <sbi/riscv_io.h> +#include <sbi/sbi_bitops.h> +#include <sbi/sbi_hart.h> +#include <sbi/sbi_scratch.h> +#include <sbi_utils/fdt/fdt_helper.h> +#include <sbi_utils/reset/fdt_reset.h> + +struct custom_csr { +unsigned long index; +unsigned long value; +}; + +u64 __fdt_reset_thead_csrr (void ); + +extern struct custom_csr custom_csr[MAX_CUSTOM_CSR ]; +extern u32 __reset_thead_csr_stub[]; +#endif lib/utils/reset/fdt_reset_thead_asm.S new file +47
-0@@ -0,0 +1,47 @@ +/* + * SPDX-License-Identifier: BSD-2-Clause + */ + +#include <sbi/riscv_asm.h> +#include "fdt_reset_thead.h" + +/* + * csrrs rd, csr, rs1 + * |31 20|19 15|14 12|11 7|6 0| + * csr rs1 010 rd 1110011 + */ +#define CSR_STUB addi x0, x0, 0 + + .option norvc + .align 3 + .global __fdt_reset_thead_csrr +__fdt_reset_thead_csrr: + csrrs a0, 0, x0 + ret + + .align 3 + .global __thead_pre_start_warm +__thead_pre_start_warm: + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + + lla t1, custom_csr + + .global __reset_thead_csr_stub +__reset_thead_csr_stub: +.rept MAX_CUSTOM_CSR + REG_L t2, 8(t1) + CSR_STUB + addi t1, t1, 16 +.endr + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + j _start_warmlib/utils/reset/objects.mk +2
-1@@ -10,3 +10,5 @@ libsbiutils-objs-y += reset/fdt_reset.o libsbiutils-objs-y += reset/fdt_reset_htif.o libsbiutils-objs-y += reset/fdt_reset_sifive.o +libsbiutils-objs-y += reset/fdt_reset_thead.o +libsbiutils-objs-y += reset/fdt_reset_thead_asm.o --2.7.4
Signature · 4 lines
Best Regards Guo Ren ML: https://lore.kernel.org/linux-csky/
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
In reply to [email protected]ping kindly ... Any problems with the patch, please let me know.
Quoted 31 lines · On Tue, Apr 13, 2021 at 11:44 AM <[email protected]> wrote:
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>--- lib/utils/fdt/fdt_helper.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)lib/utils/fdt/fdt_helper.c +2
-2@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */ clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , &count)) + clint->has_64bit_mmio = FALSE;return 0 ; } --2.7.4
Signature · 4 lines
Best Regards Guo Ren ML: https://lore.kernel.org/linux-csky/
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
In reply to [email protected]Quoted 12 lines
-----Original Message-----
From: opensbi <[email protected]> On Behalf Of guoren at kernel.org
Sent: 13 April 2021 09:14
To: guoren at kernel.org; Alistair Francis <[email protected]>; anup at brainfault.org
Cc: opensbi at lists.infradead.org; Guo Ren <[email protected]>; Anup Patel <[email protected]>; Xiang W <[email protected]>
Subject: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsing
Quoted 8 lines
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>
Looks good to me.
Regards,
Anup
Quoted 29 lines
--- lib/utils/fdt/fdt_helper.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)lib/utils/fdt/fdt_helper.c index renamed from lib/utils/fdt/fdt_helper.c +0
-0bf19ff9..909de01 100644
lib/utils/fdt/fdt_helper.c +2
-1@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, int nodeoffset,bool for_timer, if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */ clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , &count)) + clint->has_64bit_mmio = FALSE;return 0;
}
--
2.7.4--
opensbi mailing list
opensbi at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/opensbi
PATCH 2/2 lib: utils: reset: Add T-HEAD sample platform reset driver
In reply to [email protected]Quoted 200 lines
-----Original Message-----
From: guoren at kernel.org <[email protected]>
Sent: 13 April 2021 09:14
To: guoren at kernel.org; Alistair Francis <[email protected]>; anup at brainfault.org
Cc: opensbi at lists.infradead.org; Guo Ren <[email protected]>; Anup Patel <[email protected]>; Atish Patra <[email protected]>; Xiang W <[email protected]>; Bin Meng <[email protected]>
Subject: [PATCH 2/2] lib: utils: reset: Add T-HEAD sample platform reset driverFrom: Guo Ren <[email protected]>
This driver is for T-HEAD test chip, fpga. It could work with all T-HEAD riscv
processors: C9xx series.example1: (Using io-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; entry-reg = <0xff 0xff019050>; entry-cnt = <4>; control-reg = <0xff 0xff015004>; control-val = <0x1c>; csr-copy = <0x7f3 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc>; };example2: (Using csr-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = <0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0>; };example3: (Only delegate plic enable to S-mode)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; };After this patch, all T-HEAD c9xx would use platform/generic with
fw_dynamicas default:CROSS_COMPILE=riscv64-linux-gnu- PLATFORM=generic FW_PIC=y
/usr/bin/makedts full example:
cpus { #address-cells = <1>; #size-cells = <0>; timebase-frequency = <0x2dc6c0>; cpu at 0 { device_type = "cpu"; reg = <0>; status = "okay"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu0_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 1 { device_type = "cpu"; reg = <1>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu1_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 2 { device_type = "cpu"; reg = <2>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu2_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 3 { device_type = "cpu"; reg = <3>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu3_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; }; soc { #address-cells = <2>; #size-cells = <2>; compatible = "simple-bus"; ranges; reset: reset-sample { compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = < 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0 >; }; clint0: clint at ffdc000000 { compatible = "riscv,clint0"; interrupts-extended = < &cpu0_intc 3 &cpu0_intc 7 &cpu1_intc 3 &cpu1_intc 7 &cpu2_intc 3 &cpu2_intc 7 &cpu3_intc 3 &cpu3_intc 7 &cpu4_intc 3 &cpu4_intc 7 >; reg = <0xff 0xdc000000 0x0 0x04000000>; clint,has-no-64bit-mmio; }; intc: interrupt-controller at ffd8000000 { #interrupt-cells = <1>; compatible = "riscv,plic0"; interrupt-controller; interrupts-extended = < &cpu0_intc 0xffffffff &cpu0_intc 9 &cpu1_intc 0xffffffff &cpu1_intc 9 &cpu2_intc 0xffffffff &cpu2_intc 9 &cpu3_intc 0xffffffff &cpu3_intc 9 >; reg = <0xff 0xd8000000 0x0 0x04000000>; reg-names = "control"; riscv,max-priority = <7>; riscv,ndev = <80>; }; }The platform/thead will be deprecated.
Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Atish Patra <[email protected]>Cc Xiang W <[email protected]>Cc Bin Meng <[email protected]>--- lib/utils/reset/fdt_reset.c | 2 + lib/utils/reset/fdt_reset_thead.c | 117 ++++++++++++++++++++++++++++++++++ lib/utils/reset/fdt_reset_thead.h | 25 ++++++++ lib/utils/reset/fdt_reset_thead_asm.S | 47 ++++++++++++++ lib/utils/reset/objects.mk | 2 + 5 files changed, 193 insertions(+) create mode 100644 lib/utils/reset/fdt_reset_thead.c create mode 100644
lib/utils/reset/fdt_reset_thead.hcreate mode 100644lib/utils/reset/fdt_reset_thead_asm.Slib/utils/reset/fdt_reset.c index renamed from lib/utils/reset/fdt_reset.c +0
-0dead8a3..82532c2 100644
lib/utils/reset/fdt_reset.c +2
-0@@ -13,10 +13,12 @@extern struct fdt_reset fdt_reset_sifive;extern struct fdt_reset fdt_reset_htif; +extern struct fdt_reset fdt_reset_thead;static struct fdt_reset *reset_drivers[] = { &fdt_reset_sifive, &fdt_reset_htif, + &fdt_reset_thead, }; static struct fdt_reset *current_driver = NULL; diff --git
a/lib/utils/reset/fdt_reset_thead.cb/lib/utils/reset/fdt_reset_thead.cnew file mode 100644
index 0000000..12eb1fc
lib/utils/reset/fdt_reset_thead.c line 3
Quoted 3 lines
lib/utils/reset/fdt_reset_thead.c new file +3
-0@@ -0,0 +1,117 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause */ +
Move most of the includes from fdt_reset_thead.h to here.
Keep includes in fdt_reset_thead.h to minimum.
Quoted 21 lines
lib/utils/reset/fdt_reset_thead.c new file +20
-0+#include "fdt_reset_thead.h" + +struct custom_csr custom_csr[MAX_CUSTOM_CSR ]; + +#define CSR_OPCODE 0x39073 +static void clone_csrs (int cnt) +{ +unsigned long i; + +for (i =0 ; i < cnt; i++) { +/* Write csr BIT[31 - 20] to stub */ + __reset_thead_csr_stub[3 *i +1 ] = +CSR_OPCODE | (custom_csr[i].index <<20 ); + +/* Mask csr BIT[31 - 20] */ + *(u32 *)&__fdt_reset_thead_csrr &=BIT (20 ) -1 ; +smp_mb (); + +/* Write csr BIT[31 - 20] to __fdt_reset_thead_csrr */ + *(u32 *)&__fdt_reset_thead_csrr |= custom_csr[i].index <<20;
lib/utils/reset/fdt_reset_thead.c line 27
Quoted 3 lines
lib/utils/reset/fdt_reset_thead.c new file +3
-0+smp_mb (); + + __asm__ __volatile__("fence.i\n" :::"memory" );
Please add macro in sbi/riscv_barrier.h for this and use it here.
Quoted 70 lines
lib/utils/reset/fdt_reset_thead.c new file +55
-0+ + custom_csr[i].value =__fdt_reset_thead_csrr (); + } +} + +extern void __thead_pre_start_warm (void );static int +thead_reset_init (void *fdt,int nodeoff, +const struct fdt_match *match) +{ +void *p; +const fdt64_t *val; +const fdt32_t *val_w; +int len, i, cnt; +u32 tmp =0 ; + +/* Prepare clone csrs */ + val_w =fdt_getprop (fdt, nodeoff,"csr-copy" , &len); +if (len >0 && val_w) { + cnt = len /sizeof (fdt32_t); + +if (cnt >MAX_CUSTOM_CSR ) +sbi_hart_hang (); + +for (i =0 ; i < cnt; i++) { + custom_csr[i].index =fdt32_to_cpu (val_w[i]); + } + } + +if (cnt) +clone_csrs (cnt); + +/* Delegate plic enable regs for S-mode */ + val =fdt_getprop (fdt, nodeoff,"plic-delegate" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); +writel (BIT (0 ), p); + } + +/* Old reset method for secondary harts */ +if (fdt_getprop (fdt, nodeoff,"using-csr-reset" , &len)) { +csr_write (0x7c7 , (u64)&__thead_pre_start_warm); +csr_write (0x7c6 ,-1 ); + } + +/* Custom reset method for secondary harts */ + val =fdt_getprop (fdt, nodeoff,"entry-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w =fdt_getprop (fdt, nodeoff,"entry-cnt" , &len); +if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + +for (i =0 ; i < tmp; i++) { + writel((u64)(&__thead_pre_start_warm), p +8*i);
lib/utils/reset/fdt_reset_thead.c new file +1
-0+ writel((u64)(&__thead_pre_start_warm) >>32, p + 8*i + 4);
lib/utils/reset/fdt_reset_thead.c new file +8
-0+ } + } + + val =fdt_getprop (fdt, nodeoff,"control-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w = fdt_getprop(fdt, nodeoff, "control-val",&len);
lib/utils/reset/fdt_reset_thead.c new file +3
-0+if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + tmp |=readl (p);
lib/utils/reset/fdt_reset_thead.c line 105
Quoted 10 lines
lib/utils/reset/fdt_reset_thead.c new file +10
-0+writel (tmp, p); + } + } + } + +return 0 ; +} + +void thead_system_reset (u32 type,u32 reason) { + __asm__ __volatile__("ebreak\n" );
Add a macro in sbi/riscv_asm.h for "ebreak" and use it here.
lib/utils/reset/fdt_reset_thead.c line 116
Quoted 11 lines
lib/utils/reset/fdt_reset_thead.c new file +11
-0+} + +static const struct fdt_match thead_reset_match[] = { + { .compatible ="thead,reset-sample" }, + { }, +}; + +struct fdt_reset fdt_reset_thead = { + .match_table = thead_reset_match, + .init = thead_reset_init, + .system_reset = thead_system_reset
We need system_reset_check() callback to be implemented here otherwise your system_reset() callback will not be called.
lib/utils/reset/fdt_reset_thead.h line 2
Quoted 10 lines
lib/utils/reset/fdt_reset_thead.c new file +1
-0+};lib/utils/reset/fdt_reset_thead.h new file +2
-0@@ -0,0 +1,25 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause */
Need header protection defines here.
#ifndef __FDT_RESET_THEAD__
#define __FDT_RESET_THEAD__
And one extra "#endif" at end of the file.
Quoted 84 lines
lib/utils/reset/fdt_reset_thead.h new file +21
-0+ +#define MAX_CUSTOM_CSR 32 + +#ifndef __ASSEMBLER__ +#include <libfdt.h> +#include <sbi/riscv_io.h> +#include <sbi/sbi_bitops.h> +#include <sbi/sbi_hart.h> +#include <sbi/sbi_scratch.h> +#include <sbi_utils/fdt/fdt_helper.h> +#include <sbi_utils/reset/fdt_reset.h> + +struct custom_csr { +unsigned long index; +unsigned long value; +}; + +u64 __fdt_reset_thead_csrr (void ); + +extern struct custom_csr custom_csr[MAX_CUSTOM_CSR ];extern u32 +__reset_thead_csr_stub[];#endif lib/utils/reset/fdt_reset_thead_asm.S new file +46
-0@@ -0,0 +1,47 @@ +/* + * SPDX-License-Identifier: BSD-2-Clause */ + +#include <sbi/riscv_asm.h> +#include "fdt_reset_thead.h" + +/* + * csrrs rd, csr, rs1 + * |31 20|19 15|14 12|11 7|6 0| + * csr rs1 010 rd 1110011 + */ +#define CSR_STUB addi x0, x0, 0 + + .option norvc + .align 3 + .global __fdt_reset_thead_csrr +__fdt_reset_thead_csrr: + csrrs a0, 0, x0 + ret + + .align 3 + .global __thead_pre_start_warm +__thead_pre_start_warm: + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + + lla t1, custom_csr + + .global __reset_thead_csr_stub +__reset_thead_csr_stub: +.rept MAX_CUSTOM_CSR + REG_L t2, 8(t1) + CSR_STUB + addi t1, t1, 16 +.endr + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + j _start_warmlib/utils/reset/objects.mk index renamed from lib/utils/reset/objects.mk +0
-0b447261..b6619f4 100644
lib/utils/reset/objects.mk +0
-0@@ -10,3 +10,5 @@ libsbiutils-objs-y += reset/fdt_reset.o libsbiutils-objs-y +=
reset/fdt_reset_htif.olibsbiutils-objs-y +=reset/fdt_reset_sifive.olib/utils/reset/objects.mk +2
-1+libsbiutils-objs-y += reset/fdt_reset_thead.o libsbiutils-objs-y += +reset/fdt_reset_thead_asm.o --
Quoted 1 line
2.7.4
Regards,
Anup
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
In reply to [email protected]Quoted 18 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)lib/utils/fdt/fdt_helper.c +0
-0
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer, · line 446
Quoted 6 lines
lib/utils/fdt/fdt_helper.c +1
-1@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */ clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , &count))
I do not see this specified in clint.yaml.
Jess
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
Quoted 23 lines · Jessica Clarke
Quoted 21 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,if (clint->hart_count < count) clint->hart_count = count;
Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer, · line 446
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +1
-1-/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */ clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , &count))I do not see this specified in
clint.yaml.
Also, the RISC-V privileged is very clear:
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
Jess
PATCH 2/2 lib: utils: reset: Add T-HEAD sample platform reset driver
In reply to [email protected]Quoted 146 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
This driver is for T-HEAD test chip, fpga. It could work with all T-HEAD riscv processors: C9xx series.
example1: (Using io-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; entry-reg = <0xff 0xff019050>; entry-cnt = <4>; control-reg = <0xff 0xff015004>; control-val = <0x1c>; csr-copy = <0x7f3 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc>; };example2: (Using csr-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = <0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0>; };example3: (Only delegate plic enable to S-mode)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; };After this patch, all T-HEAD c9xx would use platform/generic with
fw_dynamicas default:CROSS_COMPILE=riscv64-linux-gnu- PLATFORM=generic FW_PIC=y
/usr/bin/makedts full example:
cpus { #address-cells = <1>; #size-cells = <0>; timebase-frequency = <0x2dc6c0>; cpu at 0 { device_type = "cpu"; reg = <0>; status = "okay"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu0_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 1 { device_type = "cpu"; reg = <1>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu1_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 2 { device_type = "cpu"; reg = <2>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu2_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 3 { device_type = "cpu"; reg = <3>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu3_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; }; soc { #address-cells = <2>; #size-cells = <2>; compatible = "simple-bus"; ranges; reset: reset-sample { compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = < 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0 >; }; clint0: clint at ffdc000000 { compatible = "riscv,clint0"; interrupts-extended = < &cpu0_intc 3 &cpu0_intc 7 &cpu1_intc 3 &cpu1_intc 7 &cpu2_intc 3 &cpu2_intc 7 &cpu3_intc 3 &cpu3_intc 7 &cpu4_intc 3 &cpu4_intc 7 >; reg = <0xff 0xdc000000 0x0 0x04000000>; clint,has-no-64bit-mmio; }; intc: interrupt-controller at ffd8000000 { #interrupt-cells = <1>; compatible = "riscv,plic0"; interrupt-controller; interrupts-extended = < &cpu0_intc 0xffffffff &cpu0_intc 9 &cpu1_intc 0xffffffff &cpu1_intc 9 &cpu2_intc 0xffffffff &cpu2_intc 9 &cpu3_intc 0xffffffff &cpu3_intc 9 >; reg = <0xff 0xd8000000 0x0 0x04000000>; reg-names = "control"; riscv,max-priority = <7>; riscv,ndev = <80>; }; }
Quoted 1 line
The platform/thead will be deprecated.
This seems extremely complicated for just doing a reset, yet the reset itself is just an ebreak? Is there any documentation for how this is meant to work rather than just the above device tree spew, because right now it doesn?t make any sense to me?
Jess
Quoted 258 lines
Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Atish Patra <[email protected]>Cc Xiang W <[email protected]>Cc Bin Meng <[email protected]>---
lib/utils/reset/fdt_reset.c | 2 + lib/utils/reset/fdt_reset_thead.c | 117 ++++++++++++++++++++++++++++++++++ lib/utils/reset/fdt_reset_thead.h | 25 ++++++++ lib/utils/reset/fdt_reset_thead_asm.S | 47 ++++++++++++++ lib/utils/reset/objects.mk | 2 + 5 files changed, 193 insertions(+)create mode 100644
lib/utils/reset/fdt_reset_thead.ccreate mode 100644lib/utils/reset/fdt_reset_thead.hcreate mode 100644lib/utils/reset/fdt_reset_thead_asm.Slib/utils/reset/fdt_reset.c +2
-0@@ -13,10 +13,12 @@extern struct fdt_reset fdt_reset_sifive;extern struct fdt_reset fdt_reset_htif; +extern struct fdt_reset fdt_reset_thead;static struct fdt_reset *reset_drivers[] = { &fdt_reset_sifive, &fdt_reset_htif, + &fdt_reset_thead, };static struct fdt_reset *current_driver =NULL ;lib/utils/reset/fdt_reset_thead.c new file +117
-0@@ -0,0 +1,117 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause +*/ + +#include "fdt_reset_thead.h" + +struct custom_csr custom_csr[MAX_CUSTOM_CSR ]; + +#define CSR_OPCODE 0x39073 +static void clone_csrs (int cnt) +{ +unsigned long i; + +for (i =0 ; i < cnt; i++) { +/* Write csr BIT[31 - 20] to stub */ + __reset_thead_csr_stub[3 *i +1 ] = +CSR_OPCODE | (custom_csr[i].index <<20 ); + +/* Mask csr BIT[31 - 20] */ + *(u32 *)&__fdt_reset_thead_csrr &=BIT (20 ) -1 ; +smp_mb (); + +/* Write csr BIT[31 - 20] to __fdt_reset_thead_csrr */ + *(u32 *)&__fdt_reset_thead_csrr |= custom_csr[i].index <<20 ; +smp_mb (); + + __asm__ __volatile__("fence.i\n" :::"memory" ); + + custom_csr[i].value =__fdt_reset_thead_csrr (); + } +} + +extern void __thead_pre_start_warm (void ); +static int thead_reset_init (void *fdt,int nodeoff, +const struct fdt_match *match) +{ +void *p; +const fdt64_t *val; +const fdt32_t *val_w; +int len, i, cnt; +u32 tmp =0 ; + +/* Prepare clone csrs */ + val_w =fdt_getprop (fdt, nodeoff,"csr-copy" , &len); +if (len >0 && val_w) { + cnt = len /sizeof (fdt32_t); + +if (cnt >MAX_CUSTOM_CSR ) +sbi_hart_hang (); + +for (i =0 ; i < cnt; i++) { + custom_csr[i].index =fdt32_to_cpu (val_w[i]); + } + } + +if (cnt) +clone_csrs (cnt); + +/* Delegate plic enable regs for S-mode */ + val =fdt_getprop (fdt, nodeoff,"plic-delegate" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); +writel (BIT (0 ), p); + } + +/* Old reset method for secondary harts */ +if (fdt_getprop (fdt, nodeoff,"using-csr-reset" , &len)) { +csr_write (0x7c7 , (u64)&__thead_pre_start_warm); +csr_write (0x7c6 ,-1 ); + } + +/* Custom reset method for secondary harts */ + val =fdt_getprop (fdt, nodeoff,"entry-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w =fdt_getprop (fdt, nodeoff,"entry-cnt" , &len); +if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + +for (i =0 ; i < tmp; i++) { +writel ((u64)(&__thead_pre_start_warm), p +8 *i); +writel ((u64)(&__thead_pre_start_warm) >>32 , p +8 *i +4 ); + } + } + + val =fdt_getprop (fdt, nodeoff,"control-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w =fdt_getprop (fdt, nodeoff,"control-val" , &len); +if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + tmp |=readl (p); +writel (tmp, p); + } + } + } + +return 0 ; +} + +void thead_system_reset (u32 type,u32 reason) +{ + __asm__ __volatile__("ebreak\n" ); +} + +static const struct fdt_match thead_reset_match[] = { + { .compatible ="thead,reset-sample" }, + { }, +}; + +struct fdt_reset fdt_reset_thead = { + .match_table = thead_reset_match, + .init = thead_reset_init, + .system_reset = thead_system_reset +};lib/utils/reset/fdt_reset_thead.h new file +25
-0@@ -0,0 +1,25 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause +*/ + +#define MAX_CUSTOM_CSR 32 + +#ifndef __ASSEMBLER__ +#include <libfdt.h> +#include <sbi/riscv_io.h> +#include <sbi/sbi_bitops.h> +#include <sbi/sbi_hart.h> +#include <sbi/sbi_scratch.h> +#include <sbi_utils/fdt/fdt_helper.h> +#include <sbi_utils/reset/fdt_reset.h> + +struct custom_csr { +unsigned long index; +unsigned long value; +}; + +u64 __fdt_reset_thead_csrr (void ); + +extern struct custom_csr custom_csr[MAX_CUSTOM_CSR ]; +extern u32 __reset_thead_csr_stub[]; +#endif lib/utils/reset/fdt_reset_thead_asm.S new file +47
-0@@ -0,0 +1,47 @@ +/* + * SPDX-License-Identifier: BSD-2-Clause + */ + +#include <sbi/riscv_asm.h> +#include "fdt_reset_thead.h" + +/* + * csrrs rd, csr, rs1 + * |31 20|19 15|14 12|11 7|6 0| + * csr rs1 010 rd 1110011 + */ +#define CSR_STUB addi x0, x0, 0 + + .option norvc + .align 3 + .global __fdt_reset_thead_csrr +__fdt_reset_thead_csrr: + csrrs a0, 0, x0 + ret + + .align 3 + .global __thead_pre_start_warm +__thead_pre_start_warm: + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + + lla t1, custom_csr + + .global __reset_thead_csr_stub +__reset_thead_csr_stub: +.rept MAX_CUSTOM_CSR + REG_L t2, 8(t1) + CSR_STUB + addi t1, t1, 16 +.endr + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + j _start_warmlib/utils/reset/objects.mk +0
-0@@ -10,3 +10,5 @@libsbiutils-objs-y +=
reset/fdt_reset.o
libsbiutils-objs-y +=reset/fdt_reset_htif.o
libsbiutils-objs-y +=reset/fdt_reset_sifive.olib/utils/reset/objects.mk +2
-0+libsbiutils-objs-y += reset/fdt_reset_thead.o +libsbiutils-objs-y += reset/fdt_reset_thead_asm.oSignature · 7 lines
2.7.4 -- opensbi mailing list opensbi at lists.infradead.org http://lists.infradead.org/mailman/listinfo/opensbi
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
In reply to Jessica ClarkeQuoted 35 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 17:11
To: guoren at kernel.org
Cc: Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Anup Patel <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsinglib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset,
Quoted 18 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, int nodeoffset,bool
for_timer,Quoted 3 lines
lib/utils/fdt/fdt_helper.c +0
-0if (clint->hart_count < count) clint->hart_count = count;
Quoted 6 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, · line 445
Quoted 1 line
lib/utils/fdt/fdt_helper.c +0
-1-/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer, · line 446
Quoted 2 lines
lib/utils/fdt/fdt_helper.c +1
-0clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , &count))I do not see this specified in
clint.yaml.
Yes, this is a new property should be added in Linux documentation as well.
Thanks for pointing.
Regards,
Anup
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
In reply to Jessica ClarkeQuoted 45 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 17:16
To: guoren at kernel.org
Cc: Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Anup Patel <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 15 lines · Jessica Clarke
Quoted 13 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int
Quoted 5 lines
lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, intnodeoffset, bool
for_timer,Quoted 4 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,
Quoted 4 lines
lib/utils/fdt/fdt_helper.c +0
-1if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */
Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 447
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +2
-0clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , +&count))I do not see this specified in
clint.yaml.
Quoted 7 lines
Also, the RISC-V privileged is very clear:
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
There is no CLINT spec maintained by RISC-V international.
All implementation have SiFive compatible CLINT or some variant of it.
Regards,
Anup
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
Quoted 51 lines · Anup Patel
Quoted 51 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 17:16
To: guoren at kernel.org
Cc: Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Anup Patel <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 15 lines · Jessica Clarke
Quoted 13 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int
Quoted 5 lines
lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, intnodeoffset, bool
for_timer,Quoted 4 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,
Quoted 4 lines
lib/utils/fdt/fdt_helper.c +0
-1if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */
Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 447
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +2
-0clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , +&count))I do not see this specified in
clint.yaml.Also, the RISC-V privileged is very clear:
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
Quoted 6 lines
Quoted 2 lines
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
There is no CLINT spec maintained by RISC-V international.
All implementation have SiFive compatible CLINT or some variant of it.
That was a quote from the ratified privileged spec. The memory mapping may not be specified by RISC-V Intl, but the fact that mtime(cmp) are 64-bit registers is very much specified.
Jess
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
Quoted 69 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 18:47
To: Anup Patel <[email protected]>
Cc: guoren at kernel.org; Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 9 lines · Anup Patel
Quoted 9 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 17:16
To: guoren at kernel.org
Cc: Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; GuoRen
Quoted 1 line
Quoted 1 line
<[email protected]>; Anup Patel <[email protected]>; Xiang
W
Quoted 45 lines
Quoted 44 lines
<[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 15 lines · Jessica Clarke
Quoted 13 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int
Quoted 5 lines
(quoted diff) +0
-0
b/lib/utils/fdt/fdt_helper.cindex bf19ff9..909de01 100644lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, intnodeoffset, bool
for_timer,Quoted 4 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,
Quoted 4 lines
lib/utils/fdt/fdt_helper.c +0
-1if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */
Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 447
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +2
-0clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , +&count))I do not see this specified in
clint.yaml.Also, the RISC-V privileged is very clear:
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
Quoted 7 lines
Quoted 3 lines
There is no CLINT spec maintained by RISC-V international.
All implementation have SiFive compatible CLINT or some variant of it.
That was a quote from the ratified privileged spec. The memory mapping may not be specified by RISC-V Intl, but the fact that mtime(cmp) are 64-bit registers is very much specified.
Yes, mtime(cmp) has to be a 64bit wide as-per ratified spec but the spec does not require 64bit MMIO to be supported for these MMIO registers.
In fact, for RV32 64bit MMIO is not possible at all because machine word is 32bits wide.
The T-Head implementation does not violate the mtime(cmp) requirements of ratified spec but it is certainly bit strange because they have a RV64 system but their custom CLINT only supports 32bit MMIO accesses.
Regards,
Anup
Quoted 2 lines
Jess
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
Quoted 81 lines · Anup Patel
Quoted 70 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 18:47
To: Anup Patel <[email protected]>
Cc: guoren at kernel.org; Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 9 lines · Anup Patel
Quoted 9 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 17:16
To: guoren at kernel.org
Cc: Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; GuoRen
Quoted 1 line
Quoted 1 line
<[email protected]>; Anup Patel <[email protected]>; Xiang
W
Quoted 45 lines
Quoted 44 lines
<[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 15 lines · Jessica Clarke
Quoted 13 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int
Quoted 5 lines
(quoted diff) +0
-0
b/lib/utils/fdt/fdt_helper.cindex bf19ff9..909de01 100644lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, intnodeoffset, bool
for_timer,Quoted 4 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,
Quoted 4 lines
lib/utils/fdt/fdt_helper.c +0
-1if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */
Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 447
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +2
-0clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , +&count))I do not see this specified in
clint.yaml.Also, the RISC-V privileged is very clear:
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
Quoted 7 lines
Quoted 3 lines
There is no CLINT spec maintained by RISC-V international.
All implementation have SiFive compatible CLINT or some variant of it.
That was a quote from the ratified privileged spec. The memory mapping may not be specified by RISC-V Intl, but the fact that mtime(cmp) are 64-bit registers is very much specified.
Yes, mtime(cmp) has to be a 64bit wide as-per ratified spec but the spec does not require 64bit MMIO to be supported for these MMIO registers.
Quoted 6 lines
In fact, for RV32 64bit MMIO is not possible at all because machine word is 32bits wide.
The T-Head implementation does not violate the mtime(cmp) requirements of ratified spec but it is certainly bit strange because they have a RV64 system but their custom CLINT only supports 32bit MMIO accesses.
The quoted part of the spec says that, on RV64, aligned 64-bit accesses are atomic. There is the clear implication there that 64-bit accesses are allowed. This is entirely a hardware bug that violates the spec.
Jess
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
Quoted 101 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 18:58
To: Anup Patel <[email protected]>
Cc: guoren at kernel.org; Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 18 lines · Anup Patel
Quoted 18 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 18:47
To: Anup Patel <[email protected]>
Cc: guoren at kernel.org; Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 3 lines · Anup Patel
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
Quoted 6 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 17:16
To: guoren at kernel.org
Cc: Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; GuoRen
Quoted 1 line
Quoted 1 line
<[email protected]>; Anup Patel <[email protected]>;
Xiang
Quoted 27 lines
Quoted 27 lines
W
Quoted 26 lines
Quoted 26 lines
<[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 15 lines · Jessica Clarke
Quoted 13 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int
Quoted 5 lines
(quoted diff) +0
-0
b/lib/utils/fdt/fdt_helper.cindex bf19ff9..909de01 100644lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, intnodeoffset, bool
for_timer,Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 445
Quoted 4 lines
lib/utils/fdt/fdt_helper.c +0
-1if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node
Quoted 17 lines
Quoted 17 lines
Quoted 17 lines
Quoted 7 lines
*/
Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 447
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +2
-0clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , +&count))I do not see this specified in
clint.yaml.Quoted 7 lines
Also, the RISC-V privileged is very clear:
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
There is no CLINT spec maintained by RISC-V international.
Quoted 8 lines
Quoted 5 lines
Quoted 1 line
All implementation have SiFive compatible CLINT or some variant of it.
That was a quote from the ratified privileged spec. The memory mapping may not be specified by RISC-V Intl, but the fact that mtime(cmp) are 64-bit registers is very much specified.
Yes, mtime(cmp) has to be a 64bit wide as-per ratified spec but the spec does not require 64bit MMIO to be supported for these MMIO
registers.
Quoted 4 lines
In fact, for RV32 64bit MMIO is not possible at all because machine word is 32bits wide.
Quoted 8 lines
Quoted 1 line
The T-Head implementation does not violate the mtime(cmp)
requirements
Quoted 2 lines
of ratified spec but it is certainly bit strange because they have a RV64 system but their custom CLINT only supports 32bit MMIO accesses.
The quoted part of the spec says that, on RV64, aligned 64-bit accesses are atomic. There is the clear implication there that 64-bit accesses are allowed. This is entirely a hardware bug that violates the spec.
It does not say that aligned 64bit accesses have to be supported. There is no mandate in the statement.
It only says if 64bit accesses are supported then they have to be atomic.
Regards,
Anup
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
Quoted 107 lines · Anup Patel
Quoted 107 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 18:58
To: Anup Patel <[email protected]>
Cc: guoren at kernel.org; Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 18 lines · Anup Patel
Quoted 18 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 18:47
To: Anup Patel <[email protected]>
Cc: guoren at kernel.org; Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 3 lines · Anup Patel
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
Quoted 6 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 17:16
To: guoren at kernel.org
Cc: Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; GuoRen
Quoted 1 line
Quoted 1 line
<[email protected]>; Anup Patel <[email protected]>;
Xiang
Quoted 27 lines
Quoted 27 lines
W
Quoted 26 lines
Quoted 26 lines
<[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 15 lines · Jessica Clarke
Quoted 13 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int
Quoted 5 lines
(quoted diff) +0
-0
b/lib/utils/fdt/fdt_helper.cindex bf19ff9..909de01 100644lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, intnodeoffset, bool
for_timer,Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 445
Quoted 4 lines
lib/utils/fdt/fdt_helper.c +0
-1if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node
Quoted 17 lines
Quoted 17 lines
Quoted 17 lines
Quoted 7 lines
*/
Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 447
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +2
-0clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , +&count))I do not see this specified in
clint.yaml.Quoted 7 lines
Also, the RISC-V privileged is very clear:
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
There is no CLINT spec maintained by RISC-V international.
Quoted 8 lines
Quoted 5 lines
Quoted 1 line
All implementation have SiFive compatible CLINT or some variant of it.
That was a quote from the ratified privileged spec. The memory mapping may not be specified by RISC-V Intl, but the fact that mtime(cmp) are 64-bit registers is very much specified.
Yes, mtime(cmp) has to be a 64bit wide as-per ratified spec but the spec does not require 64bit MMIO to be supported for these MMIO
registers.
Quoted 5 lines
In fact, for RV32 64bit MMIO is not possible at all because machine word is 32bits wide.
The T-Head implementation does not violate the mtime(cmp)
requirements
Quoted 2 lines
of ratified spec but it is certainly bit strange because they have a RV64 system but their custom CLINT only supports 32bit MMIO accesses.
Quoted 8 lines
Quoted 3 lines
The quoted part of the spec says that, on RV64, aligned 64-bit accesses are atomic. There is the clear implication there that 64-bit accesses are allowed. This is entirely a hardware bug that violates the spec.
It does not say that aligned 64bit accesses have to be supported. There is no mandate in the statement.
It only says if 64bit accesses are supported then they have to be atomic.
There is no ?if? in the spec, just that they *are*.
Jess
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
Quoted 126 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 19:04
To: Anup Patel <[email protected]>
Cc: guoren at kernel.org; Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 33 lines · Anup Patel
Quoted 33 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 18:58
To: Anup Patel <[email protected]>
Cc: guoren at kernel.org; Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 18 lines · Anup Patel
Quoted 18 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 18:47
To: Anup Patel <[email protected]>
Cc: guoren at kernel.org; Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>; Guo Ren <[email protected]>; Xiang W <[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 3 lines · Anup Patel
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
-----Original Message-----
From: Jessica Clarke <[email protected]>
Sent: 17 April 2021 17:16
To: guoren at kernel.org
Cc: Alistair Francis <[email protected]>; Anup Patel <[email protected]>; OpenSBI <[email protected]>;Guo
Quoted 8 lines
Quoted 8 lines
Quoted 2 lines
Quoted 2 lines
Ren
Quoted 1 line
Quoted 1 line
<[email protected]>; Anup Patel <[email protected]>;
Xiang
Quoted 5 lines
Quoted 5 lines
W
Quoted 4 lines
Quoted 4 lines
<[email protected]>
Subject: Re: [PATCH 1/2] lib: utils: Implement "64bit-mmio" property parsingQuoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines · Jessica Clarke
Quoted 6 lines · guoren at kernel.org
Quoted 4 lines
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and usingantonym
Quoted 50 lines
Quoted 50 lines
Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 9 lines
Quoted 9 lines
for compatibility.
Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int
Quoted 5 lines
(quoted diff) +0
-0
b/lib/utils/fdt/fdt_helper.cindex bf19ff9..909de01 100644lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, intnodeoffset, bool
for_timer,Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 445
Quoted 4 lines
lib/utils/fdt/fdt_helper.c +0
-1if (clint->hart_count < count) clint->hart_count = count; -/* TODO: We should figure-out CLINT has_64bit_mmio from DT node
Quoted 17 lines
Quoted 17 lines
Quoted 17 lines
Quoted 7 lines
*/
Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int · line 447
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +2
-0clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , +&count))I do not see this specified in
clint.yaml.Quoted 7 lines
Also, the RISC-V privileged is very clear:
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
There is no CLINT spec maintained by RISC-V international.
Quoted 8 lines
Quoted 5 lines
Quoted 1 line
All implementation have SiFive compatible CLINT or some variant of it.
That was a quote from the ratified privileged spec. The memory mapping may not be specified by RISC-V Intl, but the fact that mtime(cmp) are 64-bit registers is very much specified.
Yes, mtime(cmp) has to be a 64bit wide as-per ratified spec but the spec does not require 64bit MMIO to be supported for these MMIO
registers.
Quoted 4 lines
In fact, for RV32 64bit MMIO is not possible at all because machine word is 32bits wide.
Quoted 7 lines
Quoted 7 lines
Quoted 1 line
The T-Head implementation does not violate the mtime(cmp)
requirements
Quoted 2 lines
of ratified spec but it is certainly bit strange because they have a RV64 system but their custom CLINT only supports 32bit MMIO accesses.
The quoted part of the spec says that, on RV64, aligned 64-bit accesses are atomic. There is the clear implication there that 64-bit
accesses are allowed.
Quoted 8 lines
Quoted 6 lines
Quoted 1 line
This is entirely a hardware bug that violates the spec.
It does not say that aligned 64bit accesses have to be supported. There is no mandate in the statement.
It only says if 64bit accesses are supported then they have to be atomic.
There is no ?if? in the spec, just that they *are*.
The statement does not say that on RV64 we cannot use 32bit MMIO to access for mtime(cmp) which means that 64bit MMIO is is not mandatory for mtime(cmp).
Further, the statement only tells of 64bit MMIO accesses on RV64 are atomic. It tells nothing about 32bit MMIO accesses on RV64.
Regards,
Anup
PATCH 2/2 lib: utils: reset: Add T-HEAD sample platform reset driver
In reply to Jessica ClarkeQuoted 148 lines · Jessica Clarke
Quoted 146 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
This driver is for T-HEAD test chip, fpga. It could work with all T-HEAD riscv processors: C9xx series.
example1: (Using io-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; entry-reg = <0xff 0xff019050>; entry-cnt = <4>; control-reg = <0xff 0xff015004>; control-val = <0x1c>; csr-copy = <0x7f3 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc>; };example2: (Using csr-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = <0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0>; };example3: (Only delegate plic enable to S-mode)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; };After this patch, all T-HEAD c9xx would use platform/generic with
fw_dynamicas default:CROSS_COMPILE=riscv64-linux-gnu- PLATFORM=generic FW_PIC=y
/usr/bin/makedts full example:
cpus { #address-cells = <1>; #size-cells = <0>; timebase-frequency = <0x2dc6c0>; cpu at 0 { device_type = "cpu"; reg = <0>; status = "okay"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu0_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 1 { device_type = "cpu"; reg = <1>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu1_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 2 { device_type = "cpu"; reg = <2>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu2_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 3 { device_type = "cpu"; reg = <3>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu3_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; }; soc { #address-cells = <2>; #size-cells = <2>; compatible = "simple-bus"; ranges; reset: reset-sample { compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = < 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0 >; }; clint0: clint at ffdc000000 { compatible = "riscv,clint0"; interrupts-extended = < &cpu0_intc 3 &cpu0_intc 7 &cpu1_intc 3 &cpu1_intc 7 &cpu2_intc 3 &cpu2_intc 7 &cpu3_intc 3 &cpu3_intc 7 &cpu4_intc 3 &cpu4_intc 7 >; reg = <0xff 0xdc000000 0x0 0x04000000>; clint,has-no-64bit-mmio; }; intc: interrupt-controller at ffd8000000 { #interrupt-cells = <1>; compatible = "riscv,plic0"; interrupt-controller; interrupts-extended = < &cpu0_intc 0xffffffff &cpu0_intc 9 &cpu1_intc 0xffffffff &cpu1_intc 9 &cpu2_intc 0xffffffff &cpu2_intc 9 &cpu3_intc 0xffffffff &cpu3_intc 9 >; reg = <0xff 0xd8000000 0x0 0x04000000>; reg-names = "control"; riscv,max-priority = <7>; riscv,ndev = <80>; }; }
Quoted 4 lines
Quoted 1 line
The platform/thead will be deprecated.
This seems extremely complicated for just doing a reset, yet the reset itself is just an ebreak?
The ebreak is used to break down the gdb test, Jtag downloads the vmlinux, gdb continues..., when the test finished, we'll call poweroff -> sbi_reset -> ebreak in Linux shell.
Test script would come out :)
Quoted 3 lines
Is there any documentation for how this is meant to work
rather than just the above device tree spew, because right now it doesn?t make any sense to me?
We are a CPU vendor, not a soc vendor. So this is a sample for our customers to help them bring up basic system in a short time.
Firstly, we give them a simple reset address, reset control address in SOC integration. So they needn't care about multi-harts boot code at first and all are controlled by the driver. For this we also design a fdt_reset_thead_asm.S for warm-init.
If customers want to bring up basic Linux with our CPU in FPGA, zebu, Veloce, any emulators ... even any tapouted chips ... They just need devicetree + gdbinit.script + fw_dynamic.bin + Image without writing any bootloader codes.
And fw_dynamic.bin + Image also could compatible with qemu + platform generic boards (sifive board). A union opensbi & Linux is good idea for riscv Linux eco-systrem developement.
That's the reason why we change into platform/generic.
Quoted 262 lines
Jess
Quoted 258 lines
Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Atish Patra <[email protected]>Cc Xiang W <[email protected]>Cc Bin Meng <[email protected]>---
lib/utils/reset/fdt_reset.c | 2 + lib/utils/reset/fdt_reset_thead.c | 117 ++++++++++++++++++++++++++++++++++ lib/utils/reset/fdt_reset_thead.h | 25 ++++++++ lib/utils/reset/fdt_reset_thead_asm.S | 47 ++++++++++++++ lib/utils/reset/objects.mk | 2 + 5 files changed, 193 insertions(+)create mode 100644
lib/utils/reset/fdt_reset_thead.ccreate mode 100644lib/utils/reset/fdt_reset_thead.hcreate mode 100644lib/utils/reset/fdt_reset_thead_asm.Slib/utils/reset/fdt_reset.c +2
-0@@ -13,10 +13,12 @@extern struct fdt_reset fdt_reset_sifive;extern struct fdt_reset fdt_reset_htif; +extern struct fdt_reset fdt_reset_thead;static struct fdt_reset *reset_drivers[] = { &fdt_reset_sifive, &fdt_reset_htif, + &fdt_reset_thead, };static struct fdt_reset *current_driver =NULL ;lib/utils/reset/fdt_reset_thead.c new file +117
-0@@ -0,0 +1,117 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause +*/ + +#include "fdt_reset_thead.h" + +struct custom_csr custom_csr[MAX_CUSTOM_CSR ]; + +#define CSR_OPCODE 0x39073 +static void clone_csrs (int cnt) +{ +unsigned long i; + +for (i =0 ; i < cnt; i++) { +/* Write csr BIT[31 - 20] to stub */ + __reset_thead_csr_stub[3 *i +1 ] = +CSR_OPCODE | (custom_csr[i].index <<20 ); + +/* Mask csr BIT[31 - 20] */ + *(u32 *)&__fdt_reset_thead_csrr &=BIT (20 ) -1 ; +smp_mb (); + +/* Write csr BIT[31 - 20] to __fdt_reset_thead_csrr */ + *(u32 *)&__fdt_reset_thead_csrr |= custom_csr[i].index <<20 ; +smp_mb (); + + __asm__ __volatile__("fence.i\n" :::"memory" ); + + custom_csr[i].value =__fdt_reset_thead_csrr (); + } +} + +extern void __thead_pre_start_warm (void ); +static int thead_reset_init (void *fdt,int nodeoff, +const struct fdt_match *match) +{ +void *p; +const fdt64_t *val; +const fdt32_t *val_w; +int len, i, cnt; +u32 tmp =0 ; + +/* Prepare clone csrs */ + val_w =fdt_getprop (fdt, nodeoff,"csr-copy" , &len); +if (len >0 && val_w) { + cnt = len /sizeof (fdt32_t); + +if (cnt >MAX_CUSTOM_CSR ) +sbi_hart_hang (); + +for (i =0 ; i < cnt; i++) { + custom_csr[i].index =fdt32_to_cpu (val_w[i]); + } + } + +if (cnt) +clone_csrs (cnt); + +/* Delegate plic enable regs for S-mode */ + val =fdt_getprop (fdt, nodeoff,"plic-delegate" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); +writel (BIT (0 ), p); + } + +/* Old reset method for secondary harts */ +if (fdt_getprop (fdt, nodeoff,"using-csr-reset" , &len)) { +csr_write (0x7c7 , (u64)&__thead_pre_start_warm); +csr_write (0x7c6 ,-1 ); + } + +/* Custom reset method for secondary harts */ + val =fdt_getprop (fdt, nodeoff,"entry-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w =fdt_getprop (fdt, nodeoff,"entry-cnt" , &len); +if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + +for (i =0 ; i < tmp; i++) { +writel ((u64)(&__thead_pre_start_warm), p +8 *i); +writel ((u64)(&__thead_pre_start_warm) >>32 , p +8 *i +4 ); + } + } + + val =fdt_getprop (fdt, nodeoff,"control-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w =fdt_getprop (fdt, nodeoff,"control-val" , &len); +if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + tmp |=readl (p); +writel (tmp, p); + } + } + } + +return 0 ; +} + +void thead_system_reset (u32 type,u32 reason) +{ + __asm__ __volatile__("ebreak\n" ); +} + +static const struct fdt_match thead_reset_match[] = { + { .compatible ="thead,reset-sample" }, + { }, +}; + +struct fdt_reset fdt_reset_thead = { + .match_table = thead_reset_match, + .init = thead_reset_init, + .system_reset = thead_system_reset +};lib/utils/reset/fdt_reset_thead.h new file +25
-0@@ -0,0 +1,25 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause +*/ + +#define MAX_CUSTOM_CSR 32 + +#ifndef __ASSEMBLER__ +#include <libfdt.h> +#include <sbi/riscv_io.h> +#include <sbi/sbi_bitops.h> +#include <sbi/sbi_hart.h> +#include <sbi/sbi_scratch.h> +#include <sbi_utils/fdt/fdt_helper.h> +#include <sbi_utils/reset/fdt_reset.h> + +struct custom_csr { +unsigned long index; +unsigned long value; +}; + +u64 __fdt_reset_thead_csrr (void ); + +extern struct custom_csr custom_csr[MAX_CUSTOM_CSR ]; +extern u32 __reset_thead_csr_stub[]; +#endif lib/utils/reset/fdt_reset_thead_asm.S new file +47
-0@@ -0,0 +1,47 @@ +/* + * SPDX-License-Identifier: BSD-2-Clause + */ + +#include <sbi/riscv_asm.h> +#include "fdt_reset_thead.h" + +/* + * csrrs rd, csr, rs1 + * |31 20|19 15|14 12|11 7|6 0| + * csr rs1 010 rd 1110011 + */ +#define CSR_STUB addi x0, x0, 0 + + .option norvc + .align 3 + .global __fdt_reset_thead_csrr +__fdt_reset_thead_csrr: + csrrs a0, 0, x0 + ret + + .align 3 + .global __thead_pre_start_warm +__thead_pre_start_warm: + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + + lla t1, custom_csr + + .global __reset_thead_csr_stub +__reset_thead_csr_stub: +.rept MAX_CUSTOM_CSR + REG_L t2, 8(t1) + CSR_STUB + addi t1, t1, 16 +.endr + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + j _start_warmlib/utils/reset/objects.mk +0
-0@@ -10,3 +10,5 @@libsbiutils-objs-y +=
reset/fdt_reset.o
libsbiutils-objs-y +=reset/fdt_reset_htif.o
libsbiutils-objs-y +=reset/fdt_reset_sifive.olib/utils/reset/objects.mk +2
-1+libsbiutils-objs-y += reset/fdt_reset_thead.o +libsbiutils-objs-y += reset/fdt_reset_thead_asm.o --2.7.4
--
opensbi mailing list
opensbi at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/opensbi
Signature · 4 lines
Best Regards Guo Ren ML: https://lore.kernel.org/linux-csky/
PATCH 2/2 lib: utils: reset: Add T-HEAD sample platform reset driver
In reply to Anup PatelThx Anup,
Quoted 207 lines · Anup Patel
Quoted 201 lines
-----Original Message-----
From: guoren at kernel.org <[email protected]>
Sent: 13 April 2021 09:14
To: guoren at kernel.org; Alistair Francis <[email protected]>; anup at brainfault.org
Cc: opensbi at lists.infradead.org; Guo Ren <[email protected]>; Anup Patel <[email protected]>; Atish Patra <[email protected]>; Xiang W <[email protected]>; Bin Meng <[email protected]>
Subject: [PATCH 2/2] lib: utils: reset: Add T-HEAD sample platform reset driverFrom: Guo Ren <[email protected]>
This driver is for T-HEAD test chip, fpga. It could work with all T-HEAD riscv
processors: C9xx series.example1: (Using io-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; entry-reg = <0xff 0xff019050>; entry-cnt = <4>; control-reg = <0xff 0xff015004>; control-val = <0x1c>; csr-copy = <0x7f3 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc>; };example2: (Using csr-regs for reset)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = <0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0>; };example3: (Only delegate plic enable to S-mode)
reset: reset-sample {compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; };After this patch, all T-HEAD c9xx would use platform/generic with
fw_dynamicas default:CROSS_COMPILE=riscv64-linux-gnu- PLATFORM=generic FW_PIC=y
/usr/bin/makedts full example:
cpus { #address-cells = <1>; #size-cells = <0>; timebase-frequency = <0x2dc6c0>; cpu at 0 { device_type = "cpu"; reg = <0>; status = "okay"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu0_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 1 { device_type = "cpu"; reg = <1>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu1_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 2 { device_type = "cpu"; reg = <2>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu2_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; cpu at 3 { device_type = "cpu"; reg = <3>; status = "fail"; compatible = "riscv"; riscv,isa = "rv64imafdcsu"; mmu-type = "riscv,sv39"; cpu3_intc: interrupt-controller { #interrupt-cells = <1>; compatible = "riscv,cpu-intc"; interrupt-controller; }; }; }; soc { #address-cells = <2>; #size-cells = <2>; compatible = "simple-bus"; ranges; reset: reset-sample { compatible = "thead,reset-sample"; plic-delegate = <0xff 0xd81ffffc>; using-csr-reset; csr-copy = < 0x7c0 0x7c1 0x7c2 0x7c3 0x7c5 0x7cc 0x3b0 0x3b1 0x3b2 0x3b3 0x3b4 0x3b5 0x3b6 0x3b7 0x3a0 >; }; clint0: clint at ffdc000000 { compatible = "riscv,clint0"; interrupts-extended = < &cpu0_intc 3 &cpu0_intc 7 &cpu1_intc 3 &cpu1_intc 7 &cpu2_intc 3 &cpu2_intc 7 &cpu3_intc 3 &cpu3_intc 7 &cpu4_intc 3 &cpu4_intc 7 >; reg = <0xff 0xdc000000 0x0 0x04000000>; clint,has-no-64bit-mmio; }; intc: interrupt-controller at ffd8000000 { #interrupt-cells = <1>; compatible = "riscv,plic0"; interrupt-controller; interrupts-extended = < &cpu0_intc 0xffffffff &cpu0_intc 9 &cpu1_intc 0xffffffff &cpu1_intc 9 &cpu2_intc 0xffffffff &cpu2_intc 9 &cpu3_intc 0xffffffff &cpu3_intc 9 >; reg = <0xff 0xd8000000 0x0 0x04000000>; reg-names = "control"; riscv,max-priority = <7>; riscv,ndev = <80>; }; }The platform/thead will be deprecated.
Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Atish Patra <[email protected]>Cc Xiang W <[email protected]>Cc Bin Meng <[email protected]>--- lib/utils/reset/fdt_reset.c | 2 + lib/utils/reset/fdt_reset_thead.c | 117 ++++++++++++++++++++++++++++++++++ lib/utils/reset/fdt_reset_thead.h | 25 ++++++++ lib/utils/reset/fdt_reset_thead_asm.S | 47 ++++++++++++++ lib/utils/reset/objects.mk | 2 + 5 files changed, 193 insertions(+) create mode 100644 lib/utils/reset/fdt_reset_thead.c create mode 100644
lib/utils/reset/fdt_reset_thead.hcreate mode 100644lib/utils/reset/fdt_reset_thead_asm.Slib/utils/reset/fdt_reset.c index renamed from lib/utils/reset/fdt_reset.c +0
-0dead8a3..82532c2 100644
lib/utils/reset/fdt_reset.c +2
-0@@ -13,10 +13,12 @@extern struct fdt_reset fdt_reset_sifive;extern struct fdt_reset fdt_reset_htif; +extern struct fdt_reset fdt_reset_thead;static struct fdt_reset *reset_drivers[] = { &fdt_reset_sifive, &fdt_reset_htif, + &fdt_reset_thead, }; static struct fdt_reset *current_driver = NULL; diff --git
a/lib/utils/reset/fdt_reset_thead.cb/lib/utils/reset/fdt_reset_thead.cnew file mode 100644
index 0000000..12eb1fclib/utils/reset/fdt_reset_thead.c line 3
Quoted 3 lines
lib/utils/reset/fdt_reset_thead.c new file +3
-0@@ -0,0 +1,117 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause */ +Move most of the includes from
fdt_reset_thead.hto here.
Quoted 1 line
Keep includes in
fdt_reset_thead.hto minimum.
Okay
Quoted 19 lines
Quoted 19 lines
lib/utils/reset/fdt_reset_thead.c new file +18
-0+#include "fdt_reset_thead.h" + +struct custom_csr custom_csr[MAX_CUSTOM_CSR ]; + +#define CSR_OPCODE 0x39073 +static void clone_csrs (int cnt) +{ +unsigned long i; + +for (i =0 ; i < cnt; i++) { +/* Write csr BIT[31 - 20] to stub */ + __reset_thead_csr_stub[3 *i +1 ] = +CSR_OPCODE | (custom_csr[i].index <<20 ); + +/* Mask csr BIT[31 - 20] */ + *(u32 *)&__fdt_reset_thead_csrr &=BIT (20 ) -1 ; +smp_mb (); +
Quoted 8 lines
lib/utils/reset/fdt_reset_thead.c line 27
Quoted 6 lines
lib/utils/reset/fdt_reset_thead.c new file +2
-0+/* Write csr BIT[31 - 20] to __fdt_reset_thead_csrr */ + *(u32 *)&__fdt_reset_thead_csrr |= custom_csr[i].index <<20;
lib/utils/reset/fdt_reset_thead.c new file +3
-0+smp_mb (); + + __asm__ __volatile__("fence.i\n" :::"memory" );Please add macro in
sbi/riscv_barrier.hfor this and use it here.
Okay
Quoted 75 lines
Quoted 75 lines
lib/utils/reset/fdt_reset_thead.c new file +55
-0+ + custom_csr[i].value =__fdt_reset_thead_csrr (); + } +} + +extern void __thead_pre_start_warm (void );static int +thead_reset_init (void *fdt,int nodeoff, +const struct fdt_match *match) +{ +void *p; +const fdt64_t *val; +const fdt32_t *val_w; +int len, i, cnt; +u32 tmp =0 ; + +/* Prepare clone csrs */ + val_w =fdt_getprop (fdt, nodeoff,"csr-copy" , &len); +if (len >0 && val_w) { + cnt = len /sizeof (fdt32_t); + +if (cnt >MAX_CUSTOM_CSR ) +sbi_hart_hang (); + +for (i =0 ; i < cnt; i++) { + custom_csr[i].index =fdt32_to_cpu (val_w[i]); + } + } + +if (cnt) +clone_csrs (cnt); + +/* Delegate plic enable regs for S-mode */ + val =fdt_getprop (fdt, nodeoff,"plic-delegate" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); +writel (BIT (0 ), p); + } + +/* Old reset method for secondary harts */ +if (fdt_getprop (fdt, nodeoff,"using-csr-reset" , &len)) { +csr_write (0x7c7 , (u64)&__thead_pre_start_warm); +csr_write (0x7c6 ,-1 ); + } + +/* Custom reset method for secondary harts */ + val =fdt_getprop (fdt, nodeoff,"entry-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w =fdt_getprop (fdt, nodeoff,"entry-cnt" , &len); +if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + +for (i =0 ; i < tmp; i++) { + writel((u64)(&__thead_pre_start_warm), p +8*i);
+ writel((u64)(&__thead_pre_start_warm) >>
32, p + 8*i + 4);lib/utils/reset/fdt_reset_thead.c new file +8
-0+ } + } + + val =fdt_getprop (fdt, nodeoff,"control-reg" , &len); +if (len >0 && val) { + p = (void *)fdt64_to_cpu (*val); + + val_w = fdt_getprop(fdt, nodeoff, "control-val",&len);
lib/utils/reset/fdt_reset_thead.c new file +7
-0+if (len >0 && val_w) { + tmp =fdt32_to_cpu (*val_w); + tmp |=readl (p); +writel (tmp, p); + } + } + }
Quoted 8 lines
lib/utils/reset/fdt_reset_thead.c line 105
Quoted 6 lines
lib/utils/reset/fdt_reset_thead.c new file +6
-0+ +return 0 ; +} + +void thead_system_reset (u32 type,u32 reason) { + __asm__ __volatile__("ebreak\n" );Add a macro in
sbi/riscv_asm.hfor "ebreak" and use it here.
Okay
Quoted 29 lines
lib/utils/reset/fdt_reset_thead.c line 116
Quoted 12 lines
lib/utils/reset/fdt_reset_thead.c new file +11
-0+} + +static const struct fdt_match thead_reset_match[] = { + { .compatible ="thead,reset-sample" }, + { }, +}; + +struct fdt_reset fdt_reset_thead = { + .match_table = thead_reset_match, + .init = thead_reset_init, + .system_reset = thead_system_resetWe need
system_reset_check()callback to be implemented here otherwise yoursystem_reset()callback will not be called.lib/utils/reset/fdt_reset_thead.h line 2
Quoted 10 lines
lib/utils/reset/fdt_reset_thead.c new file +1
-0+};lib/utils/reset/fdt_reset_thead.h new file +2
-0@@ -0,0 +1,25 @@ +/* +* SPDX-License-Identifier: BSD-2-Clause */ Need header protection defines here.
Quoted 4 lines
#ifndef
__FDT_RESET_THEAD__
#define__FDT_RESET_THEAD__And one extra "#endif" at end of the file.
Yes, I'll correct it.
Quoted 90 lines
Quoted 85 lines
lib/utils/reset/fdt_reset_thead.h new file +21
-0+ +#define MAX_CUSTOM_CSR 32 + +#ifndef __ASSEMBLER__ +#include <libfdt.h> +#include <sbi/riscv_io.h> +#include <sbi/sbi_bitops.h> +#include <sbi/sbi_hart.h> +#include <sbi/sbi_scratch.h> +#include <sbi_utils/fdt/fdt_helper.h> +#include <sbi_utils/reset/fdt_reset.h> + +struct custom_csr { +unsigned long index; +unsigned long value; +}; + +u64 __fdt_reset_thead_csrr (void ); + +extern struct custom_csr custom_csr[MAX_CUSTOM_CSR ];extern u32 +__reset_thead_csr_stub[];#endif lib/utils/reset/fdt_reset_thead_asm.S new file +46
-0@@ -0,0 +1,47 @@ +/* + * SPDX-License-Identifier: BSD-2-Clause */ + +#include <sbi/riscv_asm.h> +#include "fdt_reset_thead.h" + +/* + * csrrs rd, csr, rs1 + * |31 20|19 15|14 12|11 7|6 0| + * csr rs1 010 rd 1110011 + */ +#define CSR_STUB addi x0, x0, 0 + + .option norvc + .align 3 + .global __fdt_reset_thead_csrr +__fdt_reset_thead_csrr: + csrrs a0, 0, x0 + ret + + .align 3 + .global __thead_pre_start_warm +__thead_pre_start_warm: + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + + lla t1, custom_csr + + .global __reset_thead_csr_stub +__reset_thead_csr_stub: +.rept MAX_CUSTOM_CSR + REG_L t2, 8(t1) + CSR_STUB + addi t1, t1, 16 +.endr + /* + * Clear L1 cache & BTB & BHT ... + */ + li t1, 0x70013 + csrw 0x7c2, t1 + fence rw,rw + j _start_warmlib/utils/reset/objects.mk index renamed from lib/utils/reset/objects.mk +0
-0b447261..b6619f4 100644
lib/utils/reset/objects.mk +0
-0@@ -10,3 +10,5 @@ libsbiutils-objs-y += reset/fdt_reset.o libsbiutils-objs-y +=
reset/fdt_reset_htif.olibsbiutils-objs-y +=reset/fdt_reset_sifive.olib/utils/reset/objects.mk +2
-1+libsbiutils-objs-y += reset/fdt_reset_thead.o libsbiutils-objs-y += +reset/fdt_reset_thead_asm.o --Quoted 1 line
2.7.4
Regards,
Anup
Signature · 4 lines
Best Regards Guo Ren ML: https://lore.kernel.org/linux-csky/
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
In reply to Jessica ClarkeQuoted 33 lines · Jessica Clarke
Quoted 23 lines · Jessica Clarke
Quoted 21 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,if (clint->hart_count < count) clint->hart_count = count;Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer, · line 446
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +1
-1-/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */ clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , &count))I do not see this specified in
clint.yaml.Also, the RISC-V privileged is very clear:
Quoted 8 lines
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
Jess
Normal Linux use csr_read(CSR_TIME), and C9xx wouldn't support M-mode&MMU-less Linux. So the only sbi_set_timer would be affected.
Because mmio-32bit is perfect with RV32, it wouldn't cause problems in the RV64 linux mmu s-mode system.
--
Best Regards
Guo Ren
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
Quoted 41 lines · Guo Ren
Quoted 39 lines · Jessica Clarke
Quoted 23 lines · Jessica Clarke
Quoted 21 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,if (clint->hart_count < count) clint->hart_count = count;Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer, · line 446
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +1
-1-/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */ clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , &count))I do not see this specified in
clint.yaml.Also, the RISC-V privileged is very clear:
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
Quoted 8 lines
Quoted 2 lines
Jess
Normal Linux use
csr_read(CSR_TIME), and C9xx wouldn't support M-mode&MMU-less Linux. So the onlysbi_set_timerwould be affected.Because mmio-32bit is perfect with RV32, it wouldn't cause problems in the RV64 linux mmu s-mode system.
That doesn?t change the fact that it violates the spec for RV64.
Jess
PATCH 1/2 lib: utils: Implement "64bit-mmio" property parsing
Quoted 49 lines · Jessica Clarke
Quoted 47 lines · Guo Ren
Quoted 33 lines · Jessica Clarke
Quoted 23 lines · Jessica Clarke
Quoted 21 lines · guoren at kernel.org
From: Guo Ren <[email protected]>
Figure out CLINT
has_64bit_mmiofrom DT node and using antonym for compatibility.Signed-off-by Guo Ren <[email protected]>Cc Anup Patel <[email protected]>Cc Xiang W <[email protected]>---
lib/utils/fdt/fdt_helper.c| 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)lib/utils/fdt/fdt_helper.c +0
-0@@ -442,8 +442,9 @@ int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer,if (clint->hart_count < count) clint->hart_count = count;Quoted 5 lines
lib/utils/fdt/fdt_helper.c int fdt_parse_clint_node(void *fdt, int nodeoffset, bool for_timer, · line 446
Quoted 3 lines
lib/utils/fdt/fdt_helper.c +1
-1-/* TODO: We should figure-out CLINT has_64bit_mmio from DT node */ clint->has_64bit_mmio = TRUE; +if (fdt_getprop (fdt, nodeoffset,"clint,has-no-64bit-mmio" , &count))I do not see this specified in
clint.yaml.Also, the RISC-V privileged is very clear:
Quoted 8 lines
Quoted 2 lines
For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic.
So I would say your CLINT violates the spec and thus should not claim to be a RISC-V or SiFive-compatible CLINT.
Jess
Normal Linux use
csr_read(CSR_TIME), and C9xx wouldn't support M-mode&MMU-less Linux. So the onlysbi_set_timerwould be affected.
Quoted 7 lines
Quoted 2 lines
Because mmio-32bit is perfect with RV32, it wouldn't cause problems in the RV64 linux mmu s-mode system.
That doesn?t change the fact that it violates the spec for RV64.
Jess
"For RV64, naturally aligned 64-bit memory accesses to the mtime and mtimecmp registers are atomic."
The sentence didn't enforce RV64 must use 64-bit memory accesses to mtime* regs. It says if you are using RV64 64-bit memory accesses, then it would be atomic and you needn't do it like in RV32.
"Platforms provide a real-time counter, exposed as a memory-mapped machine-mode read-write register, time.".
That means the soc vendor determines the data width of the IO interconnects. You know a lot of them just have a 32-bit APB on hand and easily reuse it.
Conclude my views here:
- RV64 using 32-bit memory accesses to mtime* doesn't violate spec.
- Using 32-bit/64-bit IO access depends on platform integration, not
CPU vendor. (Of cause, we could recommend them using 64-bit IO access)
- Because 64-bit IO access always contains 32-bit IO access
capability, that why our sample using 32-bit IO access as default.
Signature · 4 lines
Best Regards Guo Ren ML: https://lore.kernel.org/linux-csky/