riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
// SPDX-License-Identifier: GPL-2.0-only
/*
* alternative runtime patching
* inspired by the ARM64 and x86 version
*
* Copyright ( C ) 2021 Sifive .
*/
# include <linux/init.h>
2022-05-11 21:29:12 +02:00
# include <linux/module.h>
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
# include <linux/cpu.h>
# include <linux/uaccess.h>
# include <asm/alternative.h>
2023-01-29 01:28:53 +08:00
# include <asm/module.h>
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
# include <asm/sections.h>
2023-01-29 01:28:53 +08:00
# include <asm/vdso.h>
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
# include <asm/vendorid_list.h>
# include <asm/sbi.h>
# include <asm/csr.h>
2022-12-23 23:13:32 +01:00
# include <asm/insn.h>
# include <asm/patch.h>
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
2022-05-11 21:29:20 +02:00
struct cpu_manufacturer_info_t {
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
unsigned long vendor_id ;
unsigned long arch_id ;
unsigned long imp_id ;
2022-05-26 22:56:44 +02:00
void ( * patch_func ) ( struct alt_entry * begin , struct alt_entry * end ,
2022-05-11 21:29:20 +02:00
unsigned long archid , unsigned long impid ,
unsigned int stage ) ;
} ;
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
2023-04-20 12:49:34 -07:00
static void riscv_fill_cpu_mfr_info ( struct cpu_manufacturer_info_t * cpu_mfr_info )
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
{
# ifdef CONFIG_RISCV_M_MODE
2022-05-11 21:29:20 +02:00
cpu_mfr_info - > vendor_id = csr_read ( CSR_MVENDORID ) ;
cpu_mfr_info - > arch_id = csr_read ( CSR_MARCHID ) ;
cpu_mfr_info - > imp_id = csr_read ( CSR_MIMPID ) ;
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
# else
2022-05-11 21:29:20 +02:00
cpu_mfr_info - > vendor_id = sbi_get_mvendorid ( ) ;
cpu_mfr_info - > arch_id = sbi_get_marchid ( ) ;
cpu_mfr_info - > imp_id = sbi_get_mimpid ( ) ;
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
# endif
2022-05-11 21:29:20 +02:00
switch ( cpu_mfr_info - > vendor_id ) {
2023-08-18 14:57:19 +01:00
# ifdef CONFIG_ERRATA_ANDES
case ANDESTECH_VENDOR_ID :
cpu_mfr_info - > patch_func = andes_errata_patch_func ;
break ;
# endif
2021-03-22 22:26:04 +08:00
# ifdef CONFIG_ERRATA_SIFIVE
case SIFIVE_VENDOR_ID :
2022-05-26 22:56:44 +02:00
cpu_mfr_info - > patch_func = sifive_errata_patch_func ;
2021-03-22 22:26:04 +08:00
break ;
2022-05-11 21:29:21 +02:00
# endif
# ifdef CONFIG_ERRATA_THEAD
case THEAD_VENDOR_ID :
2022-05-26 22:56:44 +02:00
cpu_mfr_info - > patch_func = thead_errata_patch_func ;
2022-05-11 21:29:21 +02:00
break ;
2021-03-22 22:26:04 +08:00
# endif
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
default :
2022-05-26 22:56:44 +02:00
cpu_mfr_info - > patch_func = NULL ;
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
}
}
2022-12-23 23:13:32 +01:00
static u32 riscv_instruction_at ( void * p )
{
u16 * parcel = p ;
return ( u32 ) parcel [ 0 ] | ( u32 ) parcel [ 1 ] < < 16 ;
}
static void riscv_alternative_fix_auipc_jalr ( void * ptr , u32 auipc_insn ,
u32 jalr_insn , int patch_offset )
{
u32 call [ 2 ] = { auipc_insn , jalr_insn } ;
s32 imm ;
/* get and adjust new target address */
imm = riscv_insn_extract_utype_itype_imm ( auipc_insn , jalr_insn ) ;
imm - = patch_offset ;
/* update instructions */
riscv_insn_insert_utype_itype_imm ( & call [ 0 ] , & call [ 1 ] , imm ) ;
/* patch the call place again */
patch_text_nosync ( ptr , call , sizeof ( u32 ) * 2 ) ;
}
2023-01-13 22:22:05 +01:00
static void riscv_alternative_fix_jal ( void * ptr , u32 jal_insn , int patch_offset )
{
s32 imm ;
/* get and adjust new target address */
imm = riscv_insn_extract_jtype_imm ( jal_insn ) ;
imm - = patch_offset ;
/* update instruction */
riscv_insn_insert_jtype_imm ( & jal_insn , imm ) ;
/* patch the call place again */
patch_text_nosync ( ptr , & jal_insn , sizeof ( u32 ) ) ;
}
2022-12-23 23:13:32 +01:00
void riscv_alternative_fix_offsets ( void * alt_ptr , unsigned int len ,
int patch_offset )
{
int num_insn = len / sizeof ( u32 ) ;
int i ;
for ( i = 0 ; i < num_insn ; i + + ) {
u32 insn = riscv_instruction_at ( alt_ptr + i * sizeof ( u32 ) ) ;
/*
* May be the start of an auipc + jalr pair
* Needs to check that at least one more instruction
* is in the list .
*/
if ( riscv_insn_is_auipc ( insn ) & & i < num_insn - 1 ) {
u32 insn2 = riscv_instruction_at ( alt_ptr + ( i + 1 ) * sizeof ( u32 ) ) ;
if ( ! riscv_insn_is_jalr ( insn2 ) )
continue ;
/* if instruction pair is a call, it will use the ra register */
if ( RV_EXTRACT_RD_REG ( insn ) ! = 1 )
continue ;
riscv_alternative_fix_auipc_jalr ( alt_ptr + i * sizeof ( u32 ) ,
insn , insn2 , patch_offset ) ;
2023-01-16 00:28:11 +08:00
i + + ;
2022-12-23 23:13:32 +01:00
}
2023-01-13 22:22:05 +01:00
if ( riscv_insn_is_jal ( insn ) ) {
s32 imm = riscv_insn_extract_jtype_imm ( insn ) ;
/* Don't modify jumps inside the alternative block */
if ( ( alt_ptr + i * sizeof ( u32 ) + imm ) > = alt_ptr & &
( alt_ptr + i * sizeof ( u32 ) + imm ) < ( alt_ptr + len ) )
continue ;
riscv_alternative_fix_jal ( alt_ptr + i * sizeof ( u32 ) ,
insn , patch_offset ) ;
}
2022-12-23 23:13:32 +01:00
}
}
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
/*
* This is called very early in the boot process ( directly after we run
* a feature detect on the boot CPU ) . No need to worry about other CPUs
* here .
*/
2022-05-11 21:29:12 +02:00
static void __init_or_module _apply_alternatives ( struct alt_entry * begin ,
struct alt_entry * end ,
unsigned int stage )
2022-05-11 21:29:11 +02:00
{
2022-05-11 21:29:20 +02:00
struct cpu_manufacturer_info_t cpu_mfr_info ;
riscv_fill_cpu_mfr_info ( & cpu_mfr_info ) ;
2022-05-11 21:29:18 +02:00
riscv_cpufeature_patch_func ( begin , end , stage ) ;
2022-05-26 22:56:44 +02:00
if ( ! cpu_mfr_info . patch_func )
2022-05-11 21:29:11 +02:00
return ;
2022-05-26 22:56:44 +02:00
cpu_mfr_info . patch_func ( begin , end ,
cpu_mfr_info . arch_id ,
cpu_mfr_info . imp_id ,
stage ) ;
2022-05-11 21:29:11 +02:00
}
2023-01-29 01:28:53 +08:00
# ifdef CONFIG_MMU
static void __init apply_vdso_alternatives ( void )
{
const Elf_Ehdr * hdr ;
const Elf_Shdr * shdr ;
const Elf_Shdr * alt ;
struct alt_entry * begin , * end ;
hdr = ( Elf_Ehdr * ) vdso_start ;
shdr = ( void * ) hdr + hdr - > e_shoff ;
alt = find_section ( hdr , shdr , " .alternative " ) ;
if ( ! alt )
return ;
begin = ( void * ) hdr + alt - > sh_offset ,
end = ( void * ) hdr + alt - > sh_offset + alt - > sh_size ,
_apply_alternatives ( ( struct alt_entry * ) begin ,
( struct alt_entry * ) end ,
RISCV_ALTERNATIVES_BOOT ) ;
}
# else
static void __init apply_vdso_alternatives ( void ) { }
# endif
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
void __init apply_boot_alternatives ( void )
{
/* If called on non-boot cpu things could go wrong */
WARN_ON ( smp_processor_id ( ) ! = 0 ) ;
2022-05-11 21:29:11 +02:00
_apply_alternatives ( ( struct alt_entry * ) __alt_start ,
( struct alt_entry * ) __alt_end ,
RISCV_ALTERNATIVES_BOOT ) ;
2023-01-29 01:28:53 +08:00
apply_vdso_alternatives ( ) ;
riscv: Introduce alternative mechanism to apply errata solution
Introduce the "alternative" mechanism from ARM64 and x86 to apply the CPU
vendors' errata solution at runtime. The main purpose of this patch is
to provide a framework. Therefore, the implementation is quite basic for
now so that some scenarios could not use this schemei, such as patching
code to a module, relocating the patching code and heterogeneous CPU
topology.
Users could use the macro ALTERNATIVE to apply an errata to the existing
code flow. In the macro ALTERNATIVE, users need to specify the manufacturer
information(vendorid, archid, and impid) for this errata. Therefore, kernel
will know this errata is suitable for which CPU core. During the booting
procedure, kernel will select the errata required by the CPU core and then
patch it. It means that the kernel only applies the errata to the specified
CPU core. In this case, the vendor's errata does not affect each other at
runtime. The above patching procedure only occurs during the booting phase,
so we only take the overhead of the "alternative" mechanism once.
This "alternative" mechanism is enabled by default to ensure that all
required errata will be applied. However, users can disable this feature by
the Kconfig "CONFIG_RISCV_ERRATA_ALTERNATIVE".
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
2021-03-22 22:26:03 +08:00
}
2022-05-11 21:29:12 +02:00
2022-05-11 21:29:21 +02:00
/*
* apply_early_boot_alternatives ( ) is called from setup_vm ( ) with MMU - off .
*
* Following requirements should be honoured for it to work correctly :
* 1 ) It should use PC - relative addressing for accessing kernel symbols .
* To achieve this we always use GCC cmodel = medany .
* 2 ) The compiler instrumentation for FTRACE will not work for setup_vm ( )
* so disable compiler instrumentation when FTRACE is enabled .
*
* Currently , the above requirements are honoured by using custom CFLAGS
* for alternative . o in kernel / Makefile .
*/
void __init apply_early_boot_alternatives ( void )
{
# ifdef CONFIG_RISCV_ALTERNATIVE_EARLY
_apply_alternatives ( ( struct alt_entry * ) __alt_start ,
( struct alt_entry * ) __alt_end ,
RISCV_ALTERNATIVES_EARLY_BOOT ) ;
# endif
}
2022-05-11 21:29:12 +02:00
# ifdef CONFIG_MODULES
void apply_module_alternatives ( void * start , size_t length )
{
_apply_alternatives ( ( struct alt_entry * ) start ,
( struct alt_entry * ) ( start + length ) ,
RISCV_ALTERNATIVES_MODULE ) ;
}
# endif