RISC-V: Hypervisor prep work part 2
Subsystems (2)
cover RISC-V: Hypervisor prep work part 2
The first four patches are ones that I have pulled out of my original Hypervisor series at an attempt to reduce the number of patches in the series.
These four patches all make sense without the Hypervisor series so can be merged seperatley and will reduce the review burden of the next version of the patches.
The fifth patch is a prep patch for the new v0.4 Hypervisor spec.
The final patch is unreleated to Hypervisor that I'm just slipping in here because it seems easier then sending it by itself.
Alistair Francis (5):
target/riscv: Don't set write permissions on dirty PTEs target/riscv: Remove strict perm checking for CSR R/W
riscv: plic: Remove unused interrupt functions target/riscv: Create function to test if FP is enabled target/riscv: Update the Hypervisor CSRs to v0.4
Atish Patra (1):
target/riscv: Fix Floating Point register names
hw/riscv/sifive_plic.c | 12 ------------ include/hw/riscv/sifive_plic.h | 3 --- target/riscv/cpu.c | 8 ++++---- target/riscv/cpu.h | 6 +++++- target/riscv/cpu_bits.h | 35 +++++++++++++++++----------------- target/riscv/cpu_helper.c | 16 ++++++++++++---- target/riscv/csr.c | 22 ++++++++++----------- 7 files changed, 50 insertions(+), 52 deletions(-)
Signature · 1 line
2.22.0
PATCH 1/6 target/riscv: Don't set write permissions on dirty PTEs
Setting write permission on dirty PTEs results in userspace inside a Hypervisor guest (VU) becoming corrupted. This appears to be becuase it ends up with write permission in the second stage translation in cases where we aren't doing a store.
--- target/riscv/cpu_helper.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)
target/riscv/cpu_helper.c +2 -4
@@ -340,10 +340,8 @@ restart:if ((pte &PTE_X )) { *prot |=PAGE_EXEC ; } -/* add write permission on stores or if the page is already dirty, -so that we TLB miss on later writes to update the dirty bit */ -if ((pte &PTE_W ) && - (access_type ==MMU_DATA_STORE || (pte &PTE_D ))) { +/* add write permission on stores */ +if ((pte &PTE_W ) && (access_type ==MMU_DATA_STORE )) { *prot |=PAGE_WRITE ; }return TRANSLATE_SUCCESS ;
Signature · 1 line
2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 1/6] target/riscv: Don't set write permissions on dirty PTEs
Quoted 2 lines · Alistair Francis
Setting write permission on dirty PTEs results in userspace inside a Hypervisor guest (VU) becoming corrupted. This appears to be becuase it
"because"
Quoted 26 lines
ends up with write permission in the second stage translation in cases where we aren't doing a store.
Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu_helper.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)target/riscv/cpu_helper.c +2
-4@@ -340,10 +340,8 @@ restart:if ((pte &PTE_X )) { *prot |=PAGE_EXEC ; } -/* add write permission on stores or if the page is already dirty, -so that we TLB miss on later writes to update the dirty bit */ -if ((pte &PTE_W ) && - (access_type ==MMU_DATA_STORE || (pte &PTE_D ))) { +/* add write permission on stores */ +if ((pte &PTE_W ) && (access_type ==MMU_DATA_STORE )) { *prot |=PAGE_WRITE ; }return TRANSLATE_SUCCESS ;
PATCH 2/6 target/riscv: Remove strict perm checking for CSR R/W
In reply to Alistair FrancisThe privledge check based on the CSR address mask 0x300 doesn't work when using Hypervisor extensions so remove the check
--- target/riscv/csr.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
target/riscv/csr.c +1 -2
@@ -797,9 +797,8 @@ int riscv_csrrw(CPURISCVState *env, int csrno, target_ulong *ret_value,/* check privileges and return -1 if check fails */ #if !defined(CONFIG_USER_ONLY ) -int csr_priv =get_field (csrno,0x300 );int read_only =get_field (csrno,0xC00 ) ==3 ; -if ((write_mask && read_only) || (env->priv < csr_priv)) { +if (write_mask && read_only) {return -1 ; }#endif
Signature · 1 line
2.22.0
Re: [Qemu-devel] [Qemu-riscv] [PATCH-4.2 v1 2/6] target/riscv: Remove strict perm checking for CSR R/W
Unless I'm missing something, this is the only place that QEMU checks the privilege level for read and writes to CSRs. The exact computation used here won't work with the hypervisor extension, but we also can't just get rid of privilege checking entirely...
Jonathan
Quoted 29 lines · Alistair Francis
The privledge check based on the CSR address mask
0x300doesn't work when using Hypervisor extensions so remove the checkSigned-off-by Alistair Francis <[email protected]>--- target/riscv/csr.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)target/riscv/csr.c +1
-2@@ -797,9 +797,8 @@ int riscv_csrrw(CPURISCVState *env, int csrno, target_ulong *ret_value,/* check privileges and return -1 if check fails */ #if !defined(CONFIG_USER_ONLY ) -int csr_priv =get_field (csrno,0x300 );int read_only =get_field (csrno,0xC00 ) ==3 ; -if ((write_mask && read_only) || (env->priv < csr_priv)) { +if (write_mask && read_only) {return -1 ; }#endif
--
2.22.0
Re: [Qemu-devel] [Qemu-riscv] [PATCH-4.2 v1 2/6] target/riscv: Remove strict perm checking for CSR R/W
Quoted 2 lines · Jonathan Behrens
Unless I'm missing something, this is the only place that QEMU checks the privilege level for read and writes to CSRs. The exact computation used here won't work with the hypervisor extension, but we also can't just get rid of privilege checking entirely...
The csr_ops struct contains a checker function, so there are still some checks occurring. I haven't done negative testing on this patch, but the current check doesn't seem to make any sense so it should be removed. We can separately discuss adding more checks but this current way base don CSR address just seems strange.
Alistair
Quoted 32 lines
Jonathan
Quoted 28 lines · Alistair Francis
The privledge check based on the CSR address mask
0x300doesn't work when using Hypervisor extensions so remove the checkSigned-off-by Alistair Francis <[email protected]>--- target/riscv/csr.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)target/riscv/csr.c +1
-3@@ -797,9 +797,8 @@ int riscv_csrrw(CPURISCVState *env, int csrno, target_ulong *ret_value,/* check privileges and return -1 if check fails */ #if !defined(CONFIG_USER_ONLY ) -int csr_priv =get_field (csrno,0x300 );int read_only =get_field (csrno,0xC00 ) ==3 ; -if ((write_mask && read_only) || (env->priv < csr_priv)) { +if (write_mask && read_only) {return -1 ; }#endif --2.22.0
Re: [Qemu-devel] [Qemu-riscv] [PATCH-4.2 v1 2/6] target/riscv: Remove strict perm checking for CSR R/W
The remaining checks are not sufficient. If you look at the bottom of csr.c, you'll see that for most of the M-mode CSRs the predicate is set to "any" which unconditionally allows access regardless of privilege mode. The S-mode CSR predicates similarly only check that supervisor mode exists, but not that the processor is current running in that mode. I do agree with you that using the register address to determine the required privilege mode is a little strange, but RISC-V is designed so that bits 8 and 9 of the CSR address encode that information: 0=U-mode, 1=S-mode, 2=HS-mode, 3=M-mode.
I also tested by booting RVirt <https://github.com/mit-pdos/RVirt> with a Linux guest and found that this patch caused it to instantly crash because guest CSR accesses weren't intercepted.
Jonathan
Quoted 51 lines · Alistair Francis
Quoted 2 lines · Jonathan Behrens
Unless I'm missing something, this is the only place that QEMU checks
the privilege level for read and writes to CSRs. The exact computation used here won't work with the hypervisor extension, but we also can't just get rid of privilege checking entirely...
The
csr_opsstruct contains a checker function, so there are still some checks occurring. I haven't done negative testing on this patch, but the current check doesn't seem to make any sense so it should be removed. We can separately discuss adding more checks but this current way base don CSR address just seems strange.Alistair
Quoted 4 lines
Jonathan
On Thu, Jul 25, 2019 at 2:56 PM Alistair Francis <
Quoted 9 lines · [email protected]> wrote:
Quoted 9 lines
The privledge check based on the CSR address mask
0x300doesn't work when using Hypervisor extensions so remove the checkSigned-off-by Alistair Francis <[email protected]>--- target/riscv/csr.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)Quoted 5 lines
target/riscv/csr.c int riscv_csrrw(CPURISCVState *env, int csrno,
Quoted 5 lines
target/riscv/csr.c +0
-0@@ -797,9 +797,8 @@ int riscv_csrrw(CPURISCVState *env, int csrno,
target_ulong*ret_value,Quoted 14 lines
Quoted 14 lines
target/riscv/csr.c +1
-3/* check privileges and return -1 if check fails */ #if !defined(CONFIG_USER_ONLY ) -int csr_priv =get_field (csrno,0x300 );int read_only =get_field (csrno,0xC00 ) ==3 ; -if ((write_mask && read_only) || (env->priv < csr_priv)) { +if (write_mask && read_only) {return -1 ; }#endif --2.22.0
Re: [Qemu-devel] [Qemu-riscv] [PATCH-4.2 v1 2/6] target/riscv: Remove strict perm checking for CSR R/W
Quoted 2 lines · Jonathan Behrens
The remaining checks are not sufficient. If you look at the bottom of
csr.c, you'll see that for most of the M-mode CSRs the predicate is set to "any" which unconditionally allows access regardless of privilege mode. The S-mode CSR predicates similarly only check that supervisor mode exists, but not that the processor is current running in that mode. I do agree with you that using the register address to determine the required privilege mode is a little strange, but RISC-V is designed so that bits 8 and 9 of the CSR address encode that information: 0=U-mode, 1=S-mode, 2=HS-mode, 3=M-mode.
Ah, I didn't realise that was guaranteed. I will drop this patch from this series and update it in my Hypervisor series.
Alistair
Quoted 51 lines
I also tested by booting RVirt with a Linux guest and found that this patch caused it to instantly crash because guest CSR accesses weren't intercepted.
Jonathan
Quoted 45 lines · Alistair Francis
Quoted 2 lines · Jonathan Behrens
Unless I'm missing something, this is the only place that QEMU checks the privilege level for read and writes to CSRs. The exact computation used here won't work with the hypervisor extension, but we also can't just get rid of privilege checking entirely...
The
csr_opsstruct contains a checker function, so there are still some checks occurring. I haven't done negative testing on this patch, but the current check doesn't seem to make any sense so it should be removed. We can separately discuss adding more checks but this current way base don CSR address just seems strange.Alistair
Quoted 32 lines
Jonathan
Quoted 28 lines · Alistair Francis
The privledge check based on the CSR address mask
0x300doesn't work when using Hypervisor extensions so remove the checkSigned-off-by Alistair Francis <[email protected]>--- target/riscv/csr.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)target/riscv/csr.c +1
-3@@ -797,9 +797,8 @@ int riscv_csrrw(CPURISCVState *env, int csrno, target_ulong *ret_value,/* check privileges and return -1 if check fails */ #if !defined(CONFIG_USER_ONLY ) -int csr_priv =get_field (csrno,0x300 );int read_only =get_field (csrno,0xC00 ) ==3 ; -if ((write_mask && read_only) || (env->priv < csr_priv)) { +if (write_mask && read_only) {return -1 ; }#endif --2.22.0
PATCH 3/6 riscv: plic: Remove unused interrupt functions
In reply to Alistair Francis--- hw/riscv/sifive_plic.c | 12 ------------ include/hw/riscv/sifive_plic.h | 3 --- 2 files changed, 15 deletions(-)
hw/riscv/sifive_plic.c +0 -12
@@ -161,18 +161,6 @@ static void sifive_plic_update(SiFivePLICState *plic) } } -void sifive_plic_raise_irq (SiFivePLICState *plic,uint32_t irq) -{ -sifive_plic_set_pending (plic, irq, true); -sifive_plic_update (plic); -} - -void sifive_plic_lower_irq (SiFivePLICState *plic,uint32_t irq) -{ -sifive_plic_set_pending (plic, irq, false); -sifive_plic_update (plic); -} -static uint32_t sifive_plic_claim (SiFivePLICState *plic,uint32_t addrid) {int i, j;
include/hw/riscv/sifive_plic.h +0 -3
@@ -69,9 +69,6 @@ typedef struct SiFivePLICState {uint32_t aperture_size; } SiFivePLICState; -void sifive_plic_raise_irq (SiFivePLICState *plic,uint32_t irq); -void sifive_plic_lower_irq (SiFivePLICState *plic,uint32_t irq); - DeviceState *sifive_plic_create(hwaddr addr, char *hart_config, uint32_t num_sources, uint32_t num_priorities, uint32_t priority_base, uint32_t pending_base,
Signature · 1 line
2.22.0
Re: [Qemu-devel] [Qemu-riscv] [PATCH-4.2 v1 3/6] riscv: plic: Remove unused interrupt functions
Quoted 49 lines · Alistair Francis
Signed-off-by Alistair Francis <[email protected]>--- hw/riscv/sifive_plic.c | 12 ------------ include/hw/riscv/sifive_plic.h | 3 --- 2 files changed, 15 deletions(-)hw/riscv/sifive_plic.c +0
-12@@ -161,18 +161,6 @@ static void sifive_plic_update(SiFivePLICState *plic) } } -void sifive_plic_raise_irq (SiFivePLICState *plic,uint32_t irq) -{ -sifive_plic_set_pending (plic, irq, true); -sifive_plic_update (plic); -} - -void sifive_plic_lower_irq (SiFivePLICState *plic,uint32_t irq) -{ -sifive_plic_set_pending (plic, irq, false); -sifive_plic_update (plic); -} -static uint32_t sifive_plic_claim (SiFivePLICState *plic,uint32_t addrid) {int i, j;include/hw/riscv/sifive_plic.h +0
-4@@ -69,9 +69,6 @@ typedef struct SiFivePLICState {uint32_t aperture_size; } SiFivePLICState; -void sifive_plic_raise_irq (SiFivePLICState *plic,uint32_t irq); -void sifive_plic_lower_irq (SiFivePLICState *plic,uint32_t irq); - DeviceState *sifive_plic_create(hwaddr addr, char *hart_config, uint32_t num_sources, uint32_t num_priorities, uint32_t priority_base, uint32_t pending_base, --2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 3/6] riscv: plic: Remove unused interrupt functions
In reply to Alistair FrancisQuoted 35 lines · Alistair Francis
Signed-off-by Alistair Francis <[email protected]>--- hw/riscv/sifive_plic.c | 12 ------------ include/hw/riscv/sifive_plic.h | 3 --- 2 files changed, 15 deletions(-)hw/riscv/sifive_plic.c +0
-12@@ -161,18 +161,6 @@ static void sifive_plic_update(SiFivePLICState *plic) } } -void sifive_plic_raise_irq (SiFivePLICState *plic,uint32_t irq) -{ -sifive_plic_set_pending (plic, irq, true); -sifive_plic_update (plic); -} - -void sifive_plic_lower_irq (SiFivePLICState *plic,uint32_t irq) -{ -sifive_plic_set_pending (plic, irq, false); -sifive_plic_update (plic); -} -static uint32_t sifive_plic_claim (SiFivePLICState *plic,uint32_t addrid) {int i, j;include/hw/riscv/sifive_plic.h +0
-0
include/hw/riscv/sifive_plic.h typedef struct SiFivePLICState { · line 74
Quoted 9 lines
include/hw/riscv/sifive_plic.h +0
-3@@ -69,9 +69,6 @@ typedef struct SiFivePLICState {uint32_t aperture_size; } SiFivePLICState; -void sifive_plic_raise_irq (SiFivePLICState *plic,uint32_t irq); -void sifive_plic_lower_irq (SiFivePLICState *plic,uint32_t irq); - DeviceState *sifive_plic_create(hwaddr addr, char *hart_config, uint32_t num_sources, uint32_t num_priorities, uint32_t priority_base, uint32_t pending_base,
Re: [Qemu-devel] [PATCH-4.2 v1 3/6] riscv: plic: Remove unused interrupt functions
In reply to Alistair FrancisQuoted 45 lines · Alistair Francis
Signed-off-by Alistair Francis <[email protected]>--- hw/riscv/sifive_plic.c | 12 ------------ include/hw/riscv/sifive_plic.h | 3 --- 2 files changed, 15 deletions(-)hw/riscv/sifive_plic.c +0
-12@@ -161,18 +161,6 @@ static void sifive_plic_update(SiFivePLICState *plic) } } -void sifive_plic_raise_irq (SiFivePLICState *plic,uint32_t irq) -{ -sifive_plic_set_pending (plic, irq, true); -sifive_plic_update (plic); -} - -void sifive_plic_lower_irq (SiFivePLICState *plic,uint32_t irq) -{ -sifive_plic_set_pending (plic, irq, false); -sifive_plic_update (plic); -} -static uint32_t sifive_plic_claim (SiFivePLICState *plic,uint32_t addrid) {int i, j;include/hw/riscv/sifive_plic.h +0
-4@@ -69,9 +69,6 @@ typedef struct SiFivePLICState {uint32_t aperture_size; } SiFivePLICState; -void sifive_plic_raise_irq (SiFivePLICState *plic,uint32_t irq); -void sifive_plic_lower_irq (SiFivePLICState *plic,uint32_t irq); - DeviceState *sifive_plic_create(hwaddr addr, char *hart_config, uint32_t num_sources, uint32_t num_priorities, uint32_t priority_base, uint32_t pending_base, --
Quoted 3 lines
2.22.0
PATCH 4/6 target/riscv: Create function to test if FP is enabled
In reply to Alistair FrancisLet's creaate a function that tests if floating point support is enabled. We can then protect all floating point operations based on if they are enabled.
This patch so far doesn't change anything, it's just preparing for the Hypervisor support for floating point operations.
--- target/riscv/cpu.h | 6 +++++- target/riscv/cpu_helper.c | 10 ++++++++++ target/riscv/csr.c | 19 ++++++++++--------- 3 files changed, 25 insertions(+), 10 deletions(-)
target/riscv/cpu.h +5 -1
@@ -255,6 +255,7 @@ void riscv_cpu_do_interrupt(CPUState *cpu);int riscv_cpu_gdb_read_register (CPUState *cpu,uint8_t *buf,int reg);int riscv_cpu_gdb_write_register (CPUState *cpu,uint8_t *buf,int reg);bool riscv_cpu_exec_interrupt (CPUState *cs,int interrupt_request); +bool riscv_cpu_fp_enabled (CPURISCVState *env);int riscv_cpu_mmu_index (CPURISCVState *env,bool ifetch);hwaddr riscv_cpu_get_phys_page_debug (CPUState *cpu,vaddr addr); void riscv_cpu_do_unaligned_access(CPUState *cs, vaddr addr,
@@ -298,7 +299,10 @@ static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,#ifdef CONFIG_USER_ONLY *flags =TB_FLAGS_MSTATUS_FS ;#else - *flags =cpu_mmu_index (env,0 ) | (env->mstatus &MSTATUS_FS ); + *flags =cpu_mmu_index (env,0 ); +if (riscv_cpu_fp_enabled (env)) { + *flags |= env->mstatus &MSTATUS_FS ; + }#endif }
target/riscv/cpu_helper.c +10 -0
@@ -71,6 +71,16 @@ bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request)#if !defined(CONFIG_USER_ONLY ) +/* Return true is floating point support is currently enabled */ +bool riscv_cpu_fp_enabled (CPURISCVState *env) +{ +if (env->mstatus &MSTATUS_FS ) { +return true; + } + +return false; +} +int riscv_cpu_claim_interrupts (RISCVCPU *cpu,uint32_t interrupts) {CPURISCVState *env = &cpu->env;
target/riscv/csr.c +10 -9
@@ -46,7 +46,7 @@ void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops)static int fs (CPURISCVState *env,int csrno) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif
@@ -108,7 +108,7 @@ static int pmp(CPURISCVState *env, int csrno)static int read_fflags (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif
@@ -119,7 +119,7 @@ static int read_fflags(CPURISCVState *env, int csrno, target_ulong *val)static int write_fflags (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;
@@ -131,7 +131,7 @@ static int write_fflags(CPURISCVState *env, int csrno, target_ulong val)static int read_frm (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif
@@ -142,7 +142,7 @@ static int read_frm(CPURISCVState *env, int csrno, target_ulong *val)static int write_frm (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;
@@ -154,7 +154,7 @@ static int write_frm(CPURISCVState *env, int csrno, target_ulong val)static int read_fcsr (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif
@@ -166,7 +166,7 @@ static int read_fcsr(CPURISCVState *env, int csrno, target_ulong *val)static int write_fcsr (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;
@@ -307,6 +307,7 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) {target_ulong mstatus = env->mstatus;target_ulong mask =0 ; +int dirty;/* flush tlb on mstatus fields that affect VM */ if (env->priv_ver <=PRIV_VERSION_1_09_1 ) {
@@ -340,8 +341,8 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) mstatus = (mstatus & ~mask) | (val & mask); -int dirty = ((mstatus &MSTATUS_FS ) ==MSTATUS_FS ) | - ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); + dirty =riscv_cpu_fp_enabled (env) | + ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); mstatus =set_field (mstatus,MSTATUS_SD , dirty); env->mstatus = mstatus;
Signature · 1 line
2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 4/6] target/riscv: Create function to test if FP is enabled
Quoted 1 line · Alistair Francis
Let's creaate a function that tests if floating point support is
"create"
Quoted 7 lines
enabled. We can then protect all floating point operations based on if they are enabled.
This patch so far doesn't change anything, it's just preparing for the Hypervisor support for floating point operations.
Signed-off-by Alistair Francis <[email protected]>
Quoted 138 lines
--- target/riscv/cpu.h | 6 +++++- target/riscv/cpu_helper.c | 10 ++++++++++ target/riscv/csr.c | 19 ++++++++++--------- 3 files changed, 25 insertions(+), 10 deletions(-)target/riscv/cpu.h +5
-1@@ -255,6 +255,7 @@ void riscv_cpu_do_interrupt(CPUState *cpu);int riscv_cpu_gdb_read_register (CPUState *cpu,uint8_t *buf,int reg);int riscv_cpu_gdb_write_register (CPUState *cpu,uint8_t *buf,int reg);bool riscv_cpu_exec_interrupt (CPUState *cs,int interrupt_request); +bool riscv_cpu_fp_enabled (CPURISCVState *env);int riscv_cpu_mmu_index (CPURISCVState *env,bool ifetch);hwaddr riscv_cpu_get_phys_page_debug (CPUState *cpu,vaddr addr); void riscv_cpu_do_unaligned_access(CPUState *cs, vaddr addr,@@ -298,7 +299,10 @@ static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,#ifdef CONFIG_USER_ONLY *flags =TB_FLAGS_MSTATUS_FS ;#else - *flags =cpu_mmu_index (env,0 ) | (env->mstatus &MSTATUS_FS ); + *flags =cpu_mmu_index (env,0 ); +if (riscv_cpu_fp_enabled (env)) { + *flags |= env->mstatus &MSTATUS_FS ; + }#endif }target/riscv/cpu_helper.c +10
-0@@ -71,6 +71,16 @@ bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request)#if !defined(CONFIG_USER_ONLY ) +/* Return true is floating point support is currently enabled */ +bool riscv_cpu_fp_enabled (CPURISCVState *env) +{ +if (env->mstatus &MSTATUS_FS ) { +return true; + } + +return false; +} +int riscv_cpu_claim_interrupts (RISCVCPU *cpu,uint32_t interrupts) {CPURISCVState *env = &cpu->env;target/riscv/csr.c +10
-9@@ -46,7 +46,7 @@ void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops)static int fs (CPURISCVState *env,int csrno) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -108,7 +108,7 @@ static int pmp(CPURISCVState *env, int csrno)static int read_fflags (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -119,7 +119,7 @@ static int read_fflags(CPURISCVState *env, int csrno, target_ulong *val)static int write_fflags (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -131,7 +131,7 @@ static int write_fflags(CPURISCVState *env, int csrno, target_ulong val)static int read_frm (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -142,7 +142,7 @@ static int read_frm(CPURISCVState *env, int csrno, target_ulong *val)static int write_frm (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -154,7 +154,7 @@ static int write_frm(CPURISCVState *env, int csrno, target_ulong val)static int read_fcsr (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -166,7 +166,7 @@ static int read_fcsr(CPURISCVState *env, int csrno, target_ulong *val)static int write_fcsr (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -307,6 +307,7 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) {target_ulong mstatus = env->mstatus;target_ulong mask =0 ; +int dirty;/* flush tlb on mstatus fields that affect VM */ if (env->priv_ver <=PRIV_VERSION_1_09_1 ) {@@ -340,8 +341,8 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) mstatus = (mstatus & ~mask) | (val & mask); -int dirty = ((mstatus &MSTATUS_FS ) ==MSTATUS_FS ) | - ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); + dirty =riscv_cpu_fp_enabled (env) | + ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); mstatus =set_field (mstatus,MSTATUS_SD , dirty); env->mstatus = mstatus;
Re: [Qemu-devel] [Qemu-riscv] [PATCH-4.2 v1 4/6] target/riscv: Create function to test if FP is enabled
In reply to Alistair FrancisQuoted 144 lines · Alistair Francis
Let's creaate a function that tests if floating point support is enabled. We can then protect all floating point operations based on if they are enabled.
This patch so far doesn't change anything, it's just preparing for the Hypervisor support for floating point operations.
Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu.h | 6 +++++- target/riscv/cpu_helper.c | 10 ++++++++++ target/riscv/csr.c | 19 ++++++++++--------- 3 files changed, 25 insertions(+), 10 deletions(-)target/riscv/cpu.h +5
-1@@ -255,6 +255,7 @@ void riscv_cpu_do_interrupt(CPUState *cpu);int riscv_cpu_gdb_read_register (CPUState *cpu,uint8_t *buf,int reg);int riscv_cpu_gdb_write_register (CPUState *cpu,uint8_t *buf,int reg);bool riscv_cpu_exec_interrupt (CPUState *cs,int interrupt_request); +bool riscv_cpu_fp_enabled (CPURISCVState *env);int riscv_cpu_mmu_index (CPURISCVState *env,bool ifetch);hwaddr riscv_cpu_get_phys_page_debug (CPUState *cpu,vaddr addr); void riscv_cpu_do_unaligned_access(CPUState *cs, vaddr addr,@@ -298,7 +299,10 @@ static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc, #ifdef CONFIG_USER_ONLY *flags =TB_FLAGS_MSTATUS_FS ;#else - *flags =cpu_mmu_index (env,0 ) | (env->mstatus &MSTATUS_FS ); + *flags =cpu_mmu_index (env,0 ); +if (riscv_cpu_fp_enabled (env)) { + *flags |= env->mstatus &MSTATUS_FS ; + }#endif }target/riscv/cpu_helper.c +10
-0@@ -71,6 +71,16 @@ bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request)#if !defined(CONFIG_USER_ONLY ) +/* Return true is floating point support is currently enabled */ +bool riscv_cpu_fp_enabled (CPURISCVState *env) +{ +if (env->mstatus &MSTATUS_FS ) { +return true; + } + +return false; +} +int riscv_cpu_claim_interrupts (RISCVCPU *cpu,uint32_t interrupts) {CPURISCVState *env = &cpu->env;target/riscv/csr.c +1
-1@@ -46,7 +46,7 @@ void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops)static int fs (CPURISCVState *env,int csrno) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif
target/riscv/csr.c +2
-2@@ -108,7 +108,7 @@ static int pmp(CPURISCVState *env, int csrno)static int read_fflags (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -119,7 +119,7 @@ static int read_fflags(CPURISCVState *env, int csrno, target_ulong *val)static int write_fflags (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }env->mstatus |= MSTATUS_FS;target/riscv/csr.c +1
-1@@ -131,7 +131,7 @@ static int write_fflags(CPURISCVState *env, int csrno, target_ulong val)static int read_frm (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif
target/riscv/csr.c +1
-1@@ -142,7 +142,7 @@ static int read_frm(CPURISCVState *env, int csrno, target_ulong *val)static int write_frm (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }env->mstatus |= MSTATUS_FS;target/riscv/csr.c +1
-1@@ -154,7 +154,7 @@ static int write_frm(CPURISCVState *env, int csrno, target_ulong val)static int read_fcsr (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif
target/riscv/csr.c +1
-1@@ -166,7 +166,7 @@ static int read_fcsr(CPURISCVState *env, int csrno, target_ulong *val)static int write_fcsr (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }env->mstatus |= MSTATUS_FS;target/riscv/csr.c +1
-0@@ -307,6 +307,7 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) {target_ulong mstatus = env->mstatus;target_ulong mask =0 ; +int dirty;/* flush tlb on mstatus fields that affect VM */ if (env->priv_ver <= PRIV_VERSION_1_09_1) {
target/riscv/csr.c static int write_mstatus(CPURISCVState *env, int · line 347
Quoted 9 lines
target/riscv/csr.c +2
-2@@ -340,8 +341,8 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) mstatus = (mstatus & ~mask) | (val & mask); -int dirty = ((mstatus &MSTATUS_FS ) ==MSTATUS_FS ) | - ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); + dirty =riscv_cpu_fp_enabled (env) | + ((mstatus &MSTATUS_XS ) ==MSTATUS_XS );
FS are 2bits
original:
only 3 is true
new
1, 2, 3 all make dirty true
Since only 3 means dirty, keeps this part unchanged should be reasonable.
chihmin
mstatus = set_field(mstatus, MSTATUS_SD, dirty);
Quoted 7 lines
target/riscv/csr.c +0
-1env->mstatus = mstatus; --2.22.0
Re: [Qemu-devel] [Qemu-riscv] [PATCH-4.2 v1 4/6] target/riscv: Create function to test if FP is enabled
Quoted 155 lines · Chih-Min Chao
Quoted 136 lines · Alistair Francis
Let's creaate a function that tests if floating point support is enabled. We can then protect all floating point operations based on if they are enabled.
This patch so far doesn't change anything, it's just preparing for the Hypervisor support for floating point operations.
Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu.h | 6 +++++- target/riscv/cpu_helper.c | 10 ++++++++++ target/riscv/csr.c | 19 ++++++++++--------- 3 files changed, 25 insertions(+), 10 deletions(-)target/riscv/cpu.h +5
-1@@ -255,6 +255,7 @@ void riscv_cpu_do_interrupt(CPUState *cpu);int riscv_cpu_gdb_read_register (CPUState *cpu,uint8_t *buf,int reg);int riscv_cpu_gdb_write_register (CPUState *cpu,uint8_t *buf,int reg);bool riscv_cpu_exec_interrupt (CPUState *cs,int interrupt_request); +bool riscv_cpu_fp_enabled (CPURISCVState *env);int riscv_cpu_mmu_index (CPURISCVState *env,bool ifetch);hwaddr riscv_cpu_get_phys_page_debug (CPUState *cpu,vaddr addr); void riscv_cpu_do_unaligned_access(CPUState *cs, vaddr addr,@@ -298,7 +299,10 @@ static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,#ifdef CONFIG_USER_ONLY *flags =TB_FLAGS_MSTATUS_FS ;#else - *flags =cpu_mmu_index (env,0 ) | (env->mstatus &MSTATUS_FS ); + *flags =cpu_mmu_index (env,0 ); +if (riscv_cpu_fp_enabled (env)) { + *flags |= env->mstatus &MSTATUS_FS ; + }#endif }target/riscv/cpu_helper.c +10
-0@@ -71,6 +71,16 @@ bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request)#if !defined(CONFIG_USER_ONLY ) +/* Return true is floating point support is currently enabled */ +bool riscv_cpu_fp_enabled (CPURISCVState *env) +{ +if (env->mstatus &MSTATUS_FS ) { +return true; + } + +return false; +} +int riscv_cpu_claim_interrupts (RISCVCPU *cpu,uint32_t interrupts) {CPURISCVState *env = &cpu->env;target/riscv/csr.c +8
-7@@ -46,7 +46,7 @@ void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops)static int fs (CPURISCVState *env,int csrno) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -108,7 +108,7 @@ static int pmp(CPURISCVState *env, int csrno)static int read_fflags (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -119,7 +119,7 @@ static int read_fflags(CPURISCVState *env, int csrno, target_ulong *val)static int write_fflags (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -131,7 +131,7 @@ static int write_fflags(CPURISCVState *env, int csrno, target_ulong val)static int read_frm (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -142,7 +142,7 @@ static int read_frm(CPURISCVState *env, int csrno, target_ulong *val)static int write_frm (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -154,7 +154,7 @@ static int write_frm(CPURISCVState *env, int csrno, target_ulong val)static int read_fcsr (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -166,7 +166,7 @@ static int read_fcsr(CPURISCVState *env, int csrno, target_ulong *val)static int write_fcsr (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -307,6 +307,7 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) {target_ulong mstatus = env->mstatus;target_ulong mask =0 ; +int dirty;/* flush tlb on mstatus fields that affect VM */ if (env->priv_ver <=PRIV_VERSION_1_09_1 ) {target/riscv/csr.c static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) · line 345
Quoted 7 lines
target/riscv/csr.c +2
-2@@ -340,8 +341,8 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) mstatus = (mstatus & ~mask) | (val & mask); -int dirty = ((mstatus &MSTATUS_FS ) ==MSTATUS_FS ) | - ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); + dirty =riscv_cpu_fp_enabled (env) | + ((mstatus &MSTATUS_XS ) ==MSTATUS_XS );FS are 2bits
original:
only 3 is true
new
1, 2, 3 all make dirty true
Quoted 1 line
Since only 3 means dirty, keeps this part unchanged should be reasonable.
Good point, I have updated this.
Alistair
Quoted 10 lines
chihmin
Quoted 7 lines
target/riscv/csr.c +0
-1mstatus =set_field (mstatus,MSTATUS_SD , dirty); env->mstatus = mstatus; --2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 4/6] target/riscv: Create function to test if FP is enabled
In reply to Alistair FrancisQuoted 1 line · Alistair Francis
Let's creaate a function that tests if floating point support is
Typo: create
Quoted 43 lines
enabled. We can then protect all floating point operations based on if they are enabled.
This patch so far doesn't change anything, it's just preparing for the Hypervisor support for floating point operations.
Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu.h | 6 +++++- target/riscv/cpu_helper.c | 10 ++++++++++ target/riscv/csr.c | 19 ++++++++++--------- 3 files changed, 25 insertions(+), 10 deletions(-)target/riscv/cpu.h +5
-1@@ -255,6 +255,7 @@ void riscv_cpu_do_interrupt(CPUState *cpu);int riscv_cpu_gdb_read_register (CPUState *cpu,uint8_t *buf,int reg);int riscv_cpu_gdb_write_register (CPUState *cpu,uint8_t *buf,int reg);bool riscv_cpu_exec_interrupt (CPUState *cs,int interrupt_request); +bool riscv_cpu_fp_enabled (CPURISCVState *env);int riscv_cpu_mmu_index (CPURISCVState *env,bool ifetch);hwaddr riscv_cpu_get_phys_page_debug (CPUState *cpu,vaddr addr); void riscv_cpu_do_unaligned_access(CPUState *cs, vaddr addr,@@ -298,7 +299,10 @@ static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,#ifdef CONFIG_USER_ONLY *flags =TB_FLAGS_MSTATUS_FS ;#else - *flags =cpu_mmu_index (env,0 ) | (env->mstatus &MSTATUS_FS ); + *flags =cpu_mmu_index (env,0 ); +if (riscv_cpu_fp_enabled (env)) { + *flags |= env->mstatus &MSTATUS_FS ; + }#endif }target/riscv/cpu_helper.c +0
-0@@ -71,6 +71,16 @@ bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request)
target/riscv/cpu_helper.c bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request) · line 81
Quoted 10 lines
target/riscv/cpu_helper.c +8
-0#if !defined(CONFIG_USER_ONLY ) +/* Return true is floating point support is currently enabled */ +bool riscv_cpu_fp_enabled (CPURISCVState *env) +{ +if (env->mstatus &MSTATUS_FS ) { +return true; + } + +return false;
Will there be more conditions that lead to the "true" case? If not, please consider making it a one-liner for readability, e.g.
return env->mstatus & MSTATUS_FS;
(just a personal preference, feel free to ignore)
Quoted 10 lines
target/riscv/cpu_helper.c +2
-0+} +int riscv_cpu_claim_interrupts (RISCVCPU *cpu,uint32_t interrupts) {CPURISCVState *env = &cpu->env;target/riscv/csr.c +0
-0
target/riscv/csr.c void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops) · line 49
Quoted 5 lines
target/riscv/csr.c +1
-1@@ -46,7 +46,7 @@ void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops)static int fs (CPURISCVState *env,int csrno) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {
This was existing behavior, but I'm curious why all these tests are disabled when env->debugger is set. This was introduced in 753e3fe20, but I see no rationale in the commit message. I find it odd, maybe even suspicious, that activating the debugger would change the behavior :-)
Quoted 66 lines
target/riscv/csr.c +7
-6return -1; }#endif @@ -108,7 +108,7 @@ static int pmp(CPURISCVState *env, int csrno)static int read_fflags (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -119,7 +119,7 @@ static int read_fflags(CPURISCVState *env, int csrno, target_ulong *val)static int write_fflags (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -131,7 +131,7 @@ static int write_fflags(CPURISCVState *env, int csrno, target_ulong val)static int read_frm (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -142,7 +142,7 @@ static int read_frm(CPURISCVState *env, int csrno, target_ulong *val)static int write_frm (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -154,7 +154,7 @@ static int write_frm(CPURISCVState *env, int csrno, target_ulong val)static int read_fcsr (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -166,7 +166,7 @@ static int read_fcsr(CPURISCVState *env, int csrno, target_ulong *val)static int write_fcsr (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -307,6 +307,7 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) {target_ulong mstatus = env->mstatus;target_ulong mask =0 ; +int dirty;/* flush tlb on mstatus fields that affect VM */ if (env->priv_ver <=PRIV_VERSION_1_09_1 ) {
target/riscv/csr.c static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) · line 347
Quoted 9 lines
target/riscv/csr.c +2
-2@@ -340,8 +341,8 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) mstatus = (mstatus & ~mask) | (val & mask); -int dirty = ((mstatus &MSTATUS_FS ) ==MSTATUS_FS ) | - ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); + dirty =riscv_cpu_fp_enabled (env) | + ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); mstatus =set_field (mstatus,MSTATUS_SD , dirty); env->mstatus = mstatus;
--
Cheers,
Christophe de Dinechin (IRC c3d)
Re: [Qemu-devel] [PATCH-4.2 v1 4/6] target/riscv: Create function to test if FP is enabled
Quoted 7 lines · Christophe de Dinechin
Quoted 1 line · Alistair Francis
Let's creaate a function that tests if floating point support is
Typo: create
Fixed
Quoted 58 lines
Quoted 44 lines
enabled. We can then protect all floating point operations based on if they are enabled.
This patch so far doesn't change anything, it's just preparing for the Hypervisor support for floating point operations.
Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu.h | 6 +++++- target/riscv/cpu_helper.c | 10 ++++++++++ target/riscv/csr.c | 19 ++++++++++--------- 3 files changed, 25 insertions(+), 10 deletions(-)target/riscv/cpu.h +5
-1@@ -255,6 +255,7 @@ void riscv_cpu_do_interrupt(CPUState *cpu);int riscv_cpu_gdb_read_register (CPUState *cpu,uint8_t *buf,int reg);int riscv_cpu_gdb_write_register (CPUState *cpu,uint8_t *buf,int reg);bool riscv_cpu_exec_interrupt (CPUState *cs,int interrupt_request); +bool riscv_cpu_fp_enabled (CPURISCVState *env);int riscv_cpu_mmu_index (CPURISCVState *env,bool ifetch);hwaddr riscv_cpu_get_phys_page_debug (CPUState *cpu,vaddr addr); void riscv_cpu_do_unaligned_access(CPUState *cs, vaddr addr,@@ -298,7 +299,10 @@ static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,#ifdef CONFIG_USER_ONLY *flags =TB_FLAGS_MSTATUS_FS ;#else - *flags =cpu_mmu_index (env,0 ) | (env->mstatus &MSTATUS_FS ); + *flags =cpu_mmu_index (env,0 ); +if (riscv_cpu_fp_enabled (env)) { + *flags |= env->mstatus &MSTATUS_FS ; + }#endif }target/riscv/cpu_helper.c +0
-0@@ -71,6 +71,16 @@ bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request)target/riscv/cpu_helper.c bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request) · line 81
Quoted 10 lines
target/riscv/cpu_helper.c +8
-0#if !defined(CONFIG_USER_ONLY ) +/* Return true is floating point support is currently enabled */ +bool riscv_cpu_fp_enabled (CPURISCVState *env) +{ +if (env->mstatus &MSTATUS_FS ) { +return true; + } + +return false;Will there be more conditions that lead to the "true" case? If not, please consider making it a one-liner for readability, e.g.
Quoted 3 lines
return env->mstatus & MSTATUS_FS;(just a personal preference, feel free to ignore)
There are going to be more conditionals when adding the Hypervisor extension.
Quoted 14 lines
Quoted 14 lines
target/riscv/cpu_helper.c +2
-0+} +int riscv_cpu_claim_interrupts (RISCVCPU *cpu,uint32_t interrupts) {CPURISCVState *env = &cpu->env;target/riscv/csr.c +0
-0@@ -46,7 +46,7 @@ void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops)static int fs (CPURISCVState *env,int csrno) {#if !defined(CONFIG_USER_ONLY )
Quoted 8 lines
target/riscv/csr.c void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops) · line 49
Quoted 2 lines
target/riscv/csr.c +1
-1-if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {This was existing behavior, but I'm curious why all these tests are disabled when
env->debuggeris set. This was introduced in753e3fe20, but I see no rationale in the commit message. I find it odd, maybe even suspicious, that activating the debugger would change the behavior :-)
This is to allow GDB to access the registers.
Quoted 70 lines
Quoted 70 lines
target/riscv/csr.c +7
-6return -1; }#endif @@ -108,7 +108,7 @@ static int pmp(CPURISCVState *env, int csrno)static int read_fflags (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -119,7 +119,7 @@ static int read_fflags(CPURISCVState *env, int csrno, target_ulong *val)static int write_fflags (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -131,7 +131,7 @@ static int write_fflags(CPURISCVState *env, int csrno, target_ulong val)static int read_frm (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -142,7 +142,7 @@ static int read_frm(CPURISCVState *env, int csrno, target_ulong *val)static int write_frm (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -154,7 +154,7 @@ static int write_frm(CPURISCVState *env, int csrno, target_ulong val)static int read_fcsr (CPURISCVState *env,int csrno,target_ulong *val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; }#endif @@ -166,7 +166,7 @@ static int read_fcsr(CPURISCVState *env, int csrno, target_ulong *val)static int write_fcsr (CPURISCVState *env,int csrno,target_ulong val) {#if !defined(CONFIG_USER_ONLY ) -if (!env->debugger && !(env->mstatus &MSTATUS_FS )) { +if (!env->debugger && !riscv_cpu_fp_enabled (env)) {return -1 ; } env->mstatus |=MSTATUS_FS ;@@ -307,6 +307,7 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) {target_ulong mstatus = env->mstatus;target_ulong mask =0 ; +int dirty;/* flush tlb on mstatus fields that affect VM */ if (env->priv_ver <=PRIV_VERSION_1_09_1 ) {@@ -340,8 +341,8 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) mstatus = (mstatus & ~mask) | (val & mask);
Quoted 8 lines
target/riscv/csr.c static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) · line 347
Quoted 6 lines
target/riscv/csr.c +2
-2-int dirty = ((mstatus &MSTATUS_FS ) ==MSTATUS_FS ) | - ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); + dirty =riscv_cpu_fp_enabled (env) | + ((mstatus &MSTATUS_XS ) ==MSTATUS_XS ); mstatus =set_field (mstatus,MSTATUS_SD , dirty); env->mstatus = mstatus;Reviewed-by Christophe de Dinechin <[email protected]>
Thanks!
Alistair
Quoted 4 lines
--
Cheers,
Christophe de Dinechin (IRC c3d)
PATCH 5/6 target/riscv: Update the Hypervisor CSRs to v0.4
In reply to Alistair FrancisUpdate the Hypervisor CSR addresses to match the v0.4 spec.
--- target/riscv/cpu_bits.h | 35 ++++++++++++++++++----------------- 1 file changed, 18 insertions(+), 17 deletions(-)
target/riscv/cpu_bits.h +18 -17
@@ -173,6 +173,24 @@#define CSR_SPTBR 0x180#define CSR_SATP 0x180 +/* Hpervisor CSRs */ +#define CSR_HSTATUS 0x600 +#define CSR_HEDELEG 0x602 +#define CSR_HIDELEG 0x603 +#define CSR_HCOUNTERNEN 0x606 +#define CSR_HGATP 0x680 + +#if defined(TARGET_RISCV32 ) +#define HGATP_MODE SATP32_MODE +#define HGATP_ASID SATP32_ASID +#define HGATP_PPN SATP32_PPN +#endif +#if defined(TARGET_RISCV64 ) +#define HGATP_MODE SATP64_MODE +#define HGATP_ASID SATP64_ASID +#define HGATP_PPN SATP64_PPN +#endif +/* Physical Memory Protection */ #define CSR_PMPCFG0 0x3a0#define CSR_PMPCFG1 0x3a1
@@ -206,23 +224,6 @@#define CSR_DPC 0x7b1#define CSR_DSCRATCH 0x7b2 -/* Hpervisor CSRs */ -#define CSR_HSTATUS 0xa00 -#define CSR_HEDELEG 0xa02 -#define CSR_HIDELEG 0xa03 -#define CSR_HGATP 0xa80 - -#if defined(TARGET_RISCV32 ) -#define HGATP_MODE SATP32_MODE -#define HGATP_ASID SATP32_ASID -#define HGATP_PPN SATP32_PPN -#endif -#if defined(TARGET_RISCV64 ) -#define HGATP_MODE SATP64_MODE -#define HGATP_ASID SATP64_ASID -#define HGATP_PPN SATP64_PPN -#endif -/* Performance Counters */ #define CSR_MHPMCOUNTER3 0xb03#define CSR_MHPMCOUNTER4 0xb04
Signature · 1 line
2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 5/6] target/riscv: Update the Hypervisor CSRs to v0.4
Quoted 24 lines · Alistair Francis
Update the Hypervisor CSR addresses to match the v0.4 spec.
Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu_bits.h | 35 ++++++++++++++++++----------------- 1 file changed, 18 insertions(+), 17 deletions(-)target/riscv/cpu_bits.h +9
-0@@ -173,6 +173,24 @@#define CSR_SPTBR 0x180#define CSR_SATP 0x180 +/* Hpervisor CSRs */ +#define CSR_HSTATUS 0x600 +#define CSR_HEDELEG 0x602 +#define CSR_HIDELEG 0x603 +#define CSR_HCOUNTERNEN 0x606 +#define CSR_HGATP 0x680 + +#if defined(TARGET_RISCV32 ) +#define HGATP_MODE SATP32_MODE
target/riscv/cpu_bits.h line 194
Quoted 10 lines
target/riscv/cpu_bits.h +9
-0+#define HGATP_ASID SATP32_ASID +#define HGATP_PPN SATP32_PPN +#endif +#if defined(TARGET_RISCV64 ) +#define HGATP_MODE SATP64_MODE +#define HGATP_ASID SATP64_ASID +#define HGATP_PPN SATP64_PPN +#endif +
Basd on spec, is HGATP_VMID preferable ?
chihmin
Quoted 32 lines
target/riscv/cpu_bits.h +0
-0/* Physical Memory Protection */ #define CSR_PMPCFG0 0x3a0#define
CSR_PMPCFG10x3a1target/riscv/cpu_bits.h +0
-18@@ -206,23 +224,6 @@#define CSR_DPC 0x7b1#define CSR_DSCRATCH 0x7b2 -/* Hpervisor CSRs */ -#define CSR_HSTATUS 0xa00 -#define CSR_HEDELEG 0xa02 -#define CSR_HIDELEG 0xa03 -#define CSR_HGATP 0xa80 - -#if defined(TARGET_RISCV32 ) -#define HGATP_MODE SATP32_MODE -#define HGATP_ASID SATP32_ASID -#define HGATP_PPN SATP32_PPN -#endif -#if defined(TARGET_RISCV64 ) -#define HGATP_MODE SATP64_MODE -#define HGATP_ASID SATP64_ASID -#define HGATP_PPN SATP64_PPN -#endif -/* Performance Counters */ #define CSR_MHPMCOUNTER3 0xb03#define CSR_MHPMCOUNTER4 0xb04 --2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 5/6] target/riscv: Update the Hypervisor CSRs to v0.4
Quoted 33 lines · Chih-Min Chao
Quoted 29 lines · Alistair Francis
Update the Hypervisor CSR addresses to match the v0.4 spec.
Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu_bits.h | 35 ++++++++++++++++++----------------- 1 file changed, 18 insertions(+), 17 deletions(-)target/riscv/cpu_bits.h +13
-0@@ -173,6 +173,24 @@#define CSR_SPTBR 0x180#define CSR_SATP 0x180 +/* Hpervisor CSRs */ +#define CSR_HSTATUS 0x600 +#define CSR_HEDELEG 0x602 +#define CSR_HIDELEG 0x603 +#define CSR_HCOUNTERNEN 0x606 +#define CSR_HGATP 0x680 + +#if defined(TARGET_RISCV32 ) +#define HGATP_MODE SATP32_MODE +#define HGATP_ASID SATP32_ASID +#define HGATP_PPN SATP32_PPN +#endif +#if defined(TARGET_RISCV64 )
Quoted 8 lines
target/riscv/cpu_bits.h line 193
Quoted 5 lines
target/riscv/cpu_bits.h +5
-0+#define HGATP_MODE SATP64_MODE +#define HGATP_ASID SATP64_ASID +#define HGATP_PPN SATP64_PPN +#endif +Basd on spec, is
HGATP_VMIDpreferable ?
Yep, updated.
Alistair
Quoted 34 lines
chihmin
Quoted 32 lines
target/riscv/cpu_bits.h +0
-18/* Physical Memory Protection */ #define CSR_PMPCFG0 0x3a0#define CSR_PMPCFG1 0x3a1@@ -206,23 +224,6 @@#define CSR_DPC 0x7b1#define CSR_DSCRATCH 0x7b2 -/* Hpervisor CSRs */ -#define CSR_HSTATUS 0xa00 -#define CSR_HEDELEG 0xa02 -#define CSR_HIDELEG 0xa03 -#define CSR_HGATP 0xa80 - -#if defined(TARGET_RISCV32 ) -#define HGATP_MODE SATP32_MODE -#define HGATP_ASID SATP32_ASID -#define HGATP_PPN SATP32_PPN -#endif -#if defined(TARGET_RISCV64 ) -#define HGATP_MODE SATP64_MODE -#define HGATP_ASID SATP64_ASID -#define HGATP_PPN SATP64_PPN -#endif -/* Performance Counters */ #define CSR_MHPMCOUNTER3 0xb03#define CSR_MHPMCOUNTER4 0xb04 --2.22.0
PATCH 6/6 target/riscv: Fix Floating Point register names
In reply to Alistair FrancisFrom: Atish Patra <[email protected]>
As per the RISC-V spec, Floating Point registers are named as f0..f31 so lets fix the register names accordingly.
--- target/riscv/cpu.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
target/riscv/cpu.c +4 -4
@@ -40,10 +40,10 @@ const char * const riscv_int_regnames[] = { };const char *const riscv_fpr_regnames[] = { -"ft0" ,"ft1" ,"ft2" ,"ft3" ,"ft4" ,"ft5" ,"ft6" ,"ft7" , -"fs0" ,"fs1" ,"fa0" ,"fa1" ,"fa2" ,"fa3" ,"fa4" ,"fa5" , -"fa6" ,"fa7" ,"fs2" ,"fs3" ,"fs4" ,"fs5" ,"fs6" ,"fs7" , -"fs8" ,"fs9" ,"fs10" ,"fs11" ,"ft8" ,"ft9" ,"ft10" ,"ft11" +"f0" ,"f1" ,"f2" ,"f3" ,"f4" ,"f5" ,"f6" ,"f7" , +"f8" ,"f9" ,"f10" ,"f11" ,"f12" ,"f13" ,"f14" ,"f15" , +"f16" ,"f17" ,"f18" ,"f19" ,"f20" ,"f21" ,"f22" ,"f23" , +"f24" ,"f25" ,"f26" ,"f27" ,"f28" ,"f29" ,"f30" ,"f31" };const char *const riscv_excp_names[] = {
Signature · 1 line
2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 6/6] target/riscv: Fix Floating Point register names
Quoted 19 lines · Alistair Francis
From: Atish Patra <[email protected]>
As per the RISC-V spec, Floating Point registers are named as f0..f31 so lets fix the register names accordingly.
Signed-off-by Atish Patra <[email protected]>Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)target/riscv/cpu.c +0
-0@@ -40,10 +40,10 @@ const char * const riscv_int_regnames[] = { };const char *const riscv_fpr_regnames[] = {
target/riscv/cpu.c const char * const riscv_int_regnames[] = { · line 48
Quoted 10 lines
target/riscv/cpu.c +4
-4-"ft0" ,"ft1" ,"ft2" ,"ft3" ,"ft4" ,"ft5" ,"ft6" ,"ft7" , -"fs0" ,"fs1" ,"fa0" ,"fa1" ,"fa2" ,"fa3" ,"fa4" ,"fa5" , -"fa6" ,"fa7" ,"fs2" ,"fs3" ,"fs4" ,"fs5" ,"fs6" ,"fs7" , -"fs8" ,"fs9" ,"fs10" ,"fs11" ,"ft8" ,"ft9" ,"ft10" ,"ft11" +"f0" ,"f1" ,"f2" ,"f3" ,"f4" ,"f5" ,"f6" ,"f7" , +"f8" ,"f9" ,"f10" ,"f11" ,"f12" ,"f13" ,"f14" ,"f15" , +"f16" ,"f17" ,"f18" ,"f19" ,"f20" ,"f21" ,"f22" ,"f23" , +"f24" ,"f25" ,"f26" ,"f27" ,"f28" ,"f29" ,"f30" ,"f31" };
Could you indicate the section of the spec ?
By chapter 20 of user spec, the patch changes the floating register name to architecture name but leave the integer register use the ABI name.
chihmin
Quoted 6 lines
target/riscv/cpu.c +0
-1const char *const riscv_excp_names[] = { --2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 6/6] target/riscv: Fix Floating Point register names
Quoted 22 lines · Chih-Min Chao
Quoted 19 lines · Alistair Francis
From: Atish Patra <[email protected]>
As per the RISC-V spec, Floating Point registers are named as f0..f31 so lets fix the register names accordingly.
Signed-off-by Atish Patra <[email protected]>Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)target/riscv/cpu.c +0
-0@@ -40,10 +40,10 @@ const char * const riscv_int_regnames[] = { };
Quoted 13 lines
target/riscv/cpu.c const char * const riscv_int_regnames[] = { · line 47
Quoted 10 lines
target/riscv/cpu.c +4
-4const char *const riscv_fpr_regnames[] = { -"ft0" ,"ft1" ,"ft2" ,"ft3" ,"ft4" ,"ft5" ,"ft6" ,"ft7" , -"fs0" ,"fs1" ,"fa0" ,"fa1" ,"fa2" ,"fa3" ,"fa4" ,"fa5" , -"fa6" ,"fa7" ,"fs2" ,"fs3" ,"fs4" ,"fs5" ,"fs6" ,"fs7" , -"fs8" ,"fs9" ,"fs10" ,"fs11" ,"ft8" ,"ft9" ,"ft10" ,"ft11" +"f0" ,"f1" ,"f2" ,"f3" ,"f4" ,"f5" ,"f6" ,"f7" , +"f8" ,"f9" ,"f10" ,"f11" ,"f12" ,"f13" ,"f14" ,"f15" , +"f16" ,"f17" ,"f18" ,"f19" ,"f20" ,"f21" ,"f22" ,"f23" , +"f24" ,"f25" ,"f26" ,"f27" ,"f28" ,"f29" ,"f30" ,"f31" };Could you indicate the section of the spec ?
Chapter 11: "“F” Standard Extension for Single-Precision Floating-Point, Version 2.2", section 11.1, Figure 11.1 shows f0 - f32.
Quoted 1 line
By chapter 20 of user spec, the patch changes the floating register name to architecture name but leave the integer register use the ABI name.
You mean the Packed-SIMD extension?
Alistair
Quoted 8 lines
chihmin
Quoted 6 lines
target/riscv/cpu.c +0
-1const char *const riscv_excp_names[] = { --2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 6/6] target/riscv: Fix Floating Point register names
Quoted 43 lines · Alistair Francis
Quoted 3 lines · Chih-Min Chao
On Fri, Jul 26, 2019 at 2:56 AM Alistair Francis <
Quoted 19 lines · [email protected]> wrote:
Quoted 19 lines
From: Atish Patra <[email protected]>
As per the RISC-V spec, Floating Point registers are named as f0..f31 so lets fix the register names accordingly.
Signed-off-by Atish Patra <[email protected]>Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)target/riscv/cpu.c +0
-0@@ -40,10 +40,10 @@ const char * const riscv_int_regnames[] = { };Quoted 13 lines
target/riscv/cpu.c const char * const riscv_int_regnames[] = { · line 47
Quoted 10 lines
target/riscv/cpu.c +4
-4const char *const riscv_fpr_regnames[] = { -"ft0" ,"ft1" ,"ft2" ,"ft3" ,"ft4" ,"ft5" ,"ft6" ,"ft7" , -"fs0" ,"fs1" ,"fa0" ,"fa1" ,"fa2" ,"fa3" ,"fa4" ,"fa5" , -"fa6" ,"fa7" ,"fs2" ,"fs3" ,"fs4" ,"fs5" ,"fs6" ,"fs7" , -"fs8" ,"fs9" ,"fs10" ,"fs11" ,"ft8" ,"ft9" ,"ft10" ,"ft11" +"f0" ,"f1" ,"f2" ,"f3" ,"f4" ,"f5" ,"f6" ,"f7" , +"f8" ,"f9" ,"f10" ,"f11" ,"f12" ,"f13" ,"f14" ,"f15" , +"f16" ,"f17" ,"f18" ,"f19" ,"f20" ,"f21" ,"f22" ,"f23" , +"f24" ,"f25" ,"f26" ,"f27" ,"f28" ,"f29" ,"f30" ,"f31" };Could you indicate the section of the spec ?
Chapter 11: "“F” Standard Extension for Single-Precision Floating-Point, Version 2.2", section 11.1, Figure 11.1 shows f0 - f32.
Quoted 7 lines
Quoted 1 line
By chapter 20 of user spec, the patch changes the floating register name
to architecture name but leave the integer register use the ABI name.
You mean the Packed-SIMD extension?
Alistair
I means "Chapter 20RISC-V Assembly Programmer’s Handbook". There is an table, "Table 20.1: Assembler mnemonics for RISC-V integer and floating-point registers.", describes
the architecture name and ABI name for integer and floating-point register.
By the way, I reference the riscv-spec-2.2
chihmin
Quoted 9 lines
Quoted 8 lines
chihmin
Quoted 6 lines
target/riscv/cpu.c +0
-1const char *const riscv_excp_names[] = { --2.22.0
Re: [Qemu-devel] [PATCH-4.2 v1 6/6] target/riscv: Fix Floating Point register names
Quoted 48 lines · Chih-Min Chao
Quoted 44 lines · Alistair Francis
Quoted 22 lines · Chih-Min Chao
Quoted 19 lines · Alistair Francis
From: Atish Patra <[email protected]>
As per the RISC-V spec, Floating Point registers are named as f0..f31 so lets fix the register names accordingly.
Signed-off-by Atish Patra <[email protected]>Signed-off-by Alistair Francis <[email protected]>--- target/riscv/cpu.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)target/riscv/cpu.c +0
-0@@ -40,10 +40,10 @@ const char * const riscv_int_regnames[] = { };Quoted 13 lines
target/riscv/cpu.c const char * const riscv_int_regnames[] = { · line 47
Quoted 10 lines
target/riscv/cpu.c +4
-4const char *const riscv_fpr_regnames[] = { -"ft0" ,"ft1" ,"ft2" ,"ft3" ,"ft4" ,"ft5" ,"ft6" ,"ft7" , -"fs0" ,"fs1" ,"fa0" ,"fa1" ,"fa2" ,"fa3" ,"fa4" ,"fa5" , -"fa6" ,"fa7" ,"fs2" ,"fs3" ,"fs4" ,"fs5" ,"fs6" ,"fs7" , -"fs8" ,"fs9" ,"fs10" ,"fs11" ,"ft8" ,"ft9" ,"ft10" ,"ft11" +"f0" ,"f1" ,"f2" ,"f3" ,"f4" ,"f5" ,"f6" ,"f7" , +"f8" ,"f9" ,"f10" ,"f11" ,"f12" ,"f13" ,"f14" ,"f15" , +"f16" ,"f17" ,"f18" ,"f19" ,"f20" ,"f21" ,"f22" ,"f23" , +"f24" ,"f25" ,"f26" ,"f27" ,"f28" ,"f29" ,"f30" ,"f31" };Could you indicate the section of the spec ?
Chapter 11: "“F” Standard Extension for Single-Precision Floating-Point, Version 2.2", section 11.1, Figure 11.1 shows f0 - f32.
Quoted 1 line
By chapter 20 of user spec, the patch changes the floating register name to architecture name but leave the integer register use the ABI name.
Quoted 8 lines
Quoted 3 lines
You mean the Packed-SIMD extension?
Alistair
I means "Chapter 20RISC-V Assembly Programmer’s Handbook".
There is an table, "Table 20.1: Assembler mnemonics for RISC-V integer and floating-point registers.", describes the architecture name and ABI name for integer and floating-point register.
Ah ok. In general I think it makes sense to base the names on the spec and not other sources.
Alistair
Quoted 15 lines
By the way, I reference the riscv-spec-2.2
chihmin
Quoted 9 lines
Quoted 9 lines
chihmin
Quoted 6 lines
target/riscv/cpu.c +0
-1const char *const riscv_excp_names[] = { --2.22.0