target/riscv: PMP violation due to wrong size parameter
PATCH 1/1 target/riscv: PMP violation due to wrong size parameter
riscv_cpu_tlb_fill() uses the size parameter to check PMP violation using pmp_hart_has_privs().
However, if the size is unknown (=0), the ending address will be addr - 1 as it is addr + size - 1 in pmp_hart_has_privs(). This always causes a false PMP violation on the starting address of the range, as addr - 1 is not in the range.
In order to fix, we just assume that all bytes from addr to the end of the page will be accessed if the size is unknown.
--- target/riscv/cpu_helper.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)
target/riscv/cpu_helper.c +12 -1
@@ -441,6 +441,7 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,CPURISCVState *env = &cpu->env;hwaddr pa =0 ;int prot; +int pmp_size =0 ;bool pmp_violation = false;int ret =TRANSLATE_FAIL ;int mode = mmu_idx;
@@ -460,9 +461,19 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,"%s address=%" VADDR_PRIx" ret %d physical " TARGET_FMT_plx " prot %d\n", __func__, address, ret, pa, prot); +/* +* if size is unknown (0), assume that all bytes +* from addr to the end of the page will be accessed. +*/ +if (size ==0 ) { + pmp_size = -(address |TARGET_PAGE_MASK ); + }else { + pmp_size = size; + } +if (riscv_feature (env,RISCV_FEATURE_PMP ) && (ret ==TRANSLATE_SUCCESS ) && - !pmp_hart_has_privs (env, pa, size,1 << access_type, mode)) { + !pmp_hart_has_privs (env, pa, pmp_size,1 << access_type, mode)) { ret =TRANSLATE_PMP_FAIL ; }if (ret ==TRANSLATE_PMP_FAIL ) {
Signature · 1 line
2.20.1
How do you know that the access won't straddle a page boundary? Is there a guarantee somewhere that size=0 means that the access is naturally aligned?
Jonathan
Quoted 57 lines · Dayeol Lee
riscv_cpu_tlb_fill()uses thesizeparameter to check PMP violation usingpmp_hart_has_privs().
However, if the size is unknown (=0), the ending address will beaddr - 1as it isaddr + size - 1inpmp_hart_has_privs(). This always causes a false PMP violation on the starting address of the range, asaddr - 1is not in the range.In order to fix, we just assume that all bytes from addr to the end of the page will be accessed if the size is unknown.
Signed-off-by Dayeol Lee <[email protected]>Reviewed-by Richard Henderson <[email protected]>--- target/riscv/cpu_helper.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)target/riscv/cpu_helper.c +1
-0@@ -441,6 +441,7 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address,int size, CPURISCVState *env = &cpu->env;hwaddr pa =0 ;int prot; +int pmp_size =0 ;bool pmp_violation = false;int ret =TRANSLATE_FAIL ;int mode = mmu_idx;target/riscv/cpu_helper.c +0
-0@@ -460,9 +461,19 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,"%s address=%" VADDR_PRIx" ret %d physical "
TARGET_FMT_plx" prot %d\n", __func__, address, ret, pa, prot);target/riscv/cpu_helper.c +11
-2+/* +* if size is unknown (0), assume that all bytes +* from addr to the end of the page will be accessed. +*/ +if (size ==0 ) { + pmp_size = -(address |TARGET_PAGE_MASK ); + }else { + pmp_size = size; + } +if (riscv_feature (env,RISCV_FEATURE_PMP ) && (ret ==TRANSLATE_SUCCESS ) && - !pmp_hart_has_privs (env, pa, size,1 << access_type, mode)) { + !pmp_hart_has_privs (env, pa, pmp_size,1 << access_type, mode)) { ret =TRANSLATE_PMP_FAIL ; }if (ret ==TRANSLATE_PMP_FAIL ) { --2.20.1
No it doesn't mean that.
But the following code will make the size TARGET_PAGE_SIZE - (page offset) if the address is not aligned.
pmp_size = -(address | TARGET_PAGE_MASK)
Quoted 65 lines · Jonathan Behrens
How do you know that the access won't straddle a page boundary? Is there a guarantee somewhere that size=0 means that the access is naturally aligned?
Jonathan
Quoted 57 lines · Dayeol Lee
riscv_cpu_tlb_fill()uses thesizeparameter to check PMP violation usingpmp_hart_has_privs().
However, if the size is unknown (=0), the ending address will beaddr - 1as it isaddr + size - 1inpmp_hart_has_privs(). This always causes a false PMP violation on the starting address of the range, asaddr - 1is not in the range.In order to fix, we just assume that all bytes from addr to the end of the page will be accessed if the size is unknown.
Signed-off-by Dayeol Lee <[email protected]>Reviewed-by Richard Henderson <[email protected]>--- target/riscv/cpu_helper.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)target/riscv/cpu_helper.c +1
-0@@ -441,6 +441,7 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address,int size, CPURISCVState *env = &cpu->env;hwaddr pa =0 ;int prot; +int pmp_size =0 ;bool pmp_violation = false;int ret =TRANSLATE_FAIL ;int mode = mmu_idx;target/riscv/cpu_helper.c +0
-0@@ -460,9 +461,19 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,"%s address=%" VADDR_PRIx" ret %d physical "
TARGET_FMT_plx" prot %d\n", __func__, address, ret, pa, prot);target/riscv/cpu_helper.c +11
-2+/* +* if size is unknown (0), assume that all bytes +* from addr to the end of the page will be accessed. +*/ +if (size ==0 ) { + pmp_size = -(address |TARGET_PAGE_MASK ); + }else { + pmp_size = size; + } +if (riscv_feature (env,RISCV_FEATURE_PMP ) && (ret ==TRANSLATE_SUCCESS ) && - !pmp_hart_has_privs (env, pa, size,1 << access_type, mode)) { + !pmp_hart_has_privs (env, pa, pmp_size,1 << access_type, mode)) { ret =TRANSLATE_PMP_FAIL ; }if (ret ==TRANSLATE_PMP_FAIL ) { --2.20.1
Hi,
Could this patch go through?
If not please let me know so that I can fix.
Thank you!
Dayeol
Quoted 76 lines · Dayeol Lee
No it doesn't mean that.
But the following code will make the sizeTARGET_PAGE_SIZE- (page offset) if the address is not aligned.
pmp_size= -(address |TARGET_PAGE_MASK)Quoted 67 lines · Jonathan Behrens
How do you know that the access won't straddle a page boundary? Is there a guarantee somewhere that size=0 means that the access is naturally aligned?
Jonathan
Quoted 58 lines · Dayeol Lee
riscv_cpu_tlb_fill()uses thesizeparameter to check PMP violation usingpmp_hart_has_privs().
However, if the size is unknown (=0), the ending address will beaddr - 1as it isaddr + size - 1inpmp_hart_has_privs(). This always causes a false PMP violation on the starting address of the range, asaddr - 1is not in the range.In order to fix, we just assume that all bytes from addr to the end of the page will be accessed if the size is unknown.
Signed-off-by Dayeol Lee <[email protected]>Reviewed-by Richard Henderson <[email protected]>--- target/riscv/cpu_helper.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)target/riscv/cpu_helper.c +1
-0@@ -441,6 +441,7 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address,int size, CPURISCVState *env = &cpu->env;hwaddr pa =0 ;int prot; +int pmp_size =0 ;bool pmp_violation = false;int ret =TRANSLATE_FAIL ;int mode = mmu_idx;target/riscv/cpu_helper.c +0
-0@@ -460,9 +461,19 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,"%s address=%" VADDR_PRIx" ret %d physical "
TARGET_FMT_plx" prot %d\n", __func__, address, ret, pa, prot);target/riscv/cpu_helper.c +11
-1+/* +* if size is unknown (0), assume that all bytes +* from addr to the end of the page will be accessed. +*/ +if (size ==0 ) { + pmp_size = -(address |TARGET_PAGE_MASK ); + }else { + pmp_size = size; + } +if (riscv_feature (env,RISCV_FEATURE_PMP ) && (ret ==TRANSLATE_SUCCESS ) && - !pmp_hart_has_privs (env, pa, size,1 << access_type, mode)) { + !pmp_hart_has_privs (env, pa, pmp_size,1 << access_type, mode)){ ret = TRANSLATE_PMP_FAIL; } if (ret == TRANSLATE_PMP_FAIL) {--
2.20.1
Quoted 5 lines · On Tue, 15 Oct 2019 10:04:32 PDT (-0700), [email protected] wrote:
Hi,
Could this patch go through?
If not please let me know so that I can fix.
Thank you!
Sorry, I dropped this one. It's in the patch queue now. We should also check for size==0 in pmp_hart_has_privs(), as that won't work. LMK if you want to send a patch for that.
Quoted 82 lines
Dayeol
Quoted 76 lines · Dayeol Lee
No it doesn't mean that.
But the following code will make the sizeTARGET_PAGE_SIZE- (page offset) if the address is not aligned.
pmp_size= -(address |TARGET_PAGE_MASK)Quoted 67 lines · Jonathan Behrens
How do you know that the access won't straddle a page boundary? Is there a guarantee somewhere that size=0 means that the access is naturally aligned?
Jonathan
Quoted 58 lines · Dayeol Lee
riscv_cpu_tlb_fill()uses thesizeparameter to check PMP violation usingpmp_hart_has_privs().
However, if the size is unknown (=0), the ending address will beaddr - 1as it isaddr + size - 1inpmp_hart_has_privs(). This always causes a false PMP violation on the starting address of the range, asaddr - 1is not in the range.In order to fix, we just assume that all bytes from addr to the end of the page will be accessed if the size is unknown.
Signed-off-by Dayeol Lee <[email protected]>Reviewed-by Richard Henderson <[email protected]>--- target/riscv/cpu_helper.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)target/riscv/cpu_helper.c +1
-0@@ -441,6 +441,7 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address,int size, CPURISCVState *env = &cpu->env;hwaddr pa =0 ;int prot; +int pmp_size =0 ;bool pmp_violation = false;int ret =TRANSLATE_FAIL ;int mode = mmu_idx;target/riscv/cpu_helper.c +0
-0@@ -460,9 +461,19 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,"%s address=%" VADDR_PRIx" ret %d physical "
TARGET_FMT_plx" prot %d\n", __func__, address, ret, pa, prot);target/riscv/cpu_helper.c +11
-1+/* +* if size is unknown (0), assume that all bytes +* from addr to the end of the page will be accessed. +*/ +if (size ==0 ) { + pmp_size = -(address |TARGET_PAGE_MASK ); + }else { + pmp_size = size; + } +if (riscv_feature (env,RISCV_FEATURE_PMP ) && (ret ==TRANSLATE_SUCCESS ) && - !pmp_hart_has_privs (env, pa, size,1 << access_type, mode)) { + !pmp_hart_has_privs (env, pa, pmp_size,1 << access_type, mode)){ ret = TRANSLATE_PMP_FAIL; } if (ret == TRANSLATE_PMP_FAIL) {--
2.20.1
I'll move the entire check into pmp_hart_has_privs as it makes more sense.
Thanks!
Quoted 103 lines · Palmer Dabbelt
Quoted 5 lines · On Tue, 15 Oct 2019 10:04:32 PDT (-0700), [email protected] wrote:
Hi,
Could this patch go through?
If not please let me know so that I can fix.
Thank you!Sorry, I dropped this one. It's in the patch queue now. We should also check
for size==0 inpmp_hart_has_privs(), as that won't work. LMK if you want to
send a patch for that.Quoted 8 lines
Dayeol
Quoted 2 lines · Dayeol Lee
No it doesn't mean that.
But the following code will make the sizeTARGET_PAGE_SIZE- (pageoffset)
Quoted 6 lines
Quoted 6 lines
if the address is not aligned.
pmp_size= -(address |TARGET_PAGE_MASK)On Fri, Oct 11, 2019, 7:37 PM Jonathan Behrens <[email protected]>
Quoted 2 lines · wrote:
Quoted 2 lines
Quoted 2 lines
How do you know that the access won't straddle a page boundary? Is
there
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
a guarantee somewhere that size=0 means that the access is naturally aligned?
Jonathan
On Fri, Oct 11, 2019 at 7:14 PM Dayeol Lee <[email protected]>
Quoted 6 lines · wrote:
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
riscv_cpu_tlb_fill()uses thesizeparameter to check PMP violation usingpmp_hart_has_privs().
However, if the size is unknown (=0), the ending address will beaddr - 1as it isaddr + size - 1inpmp_hart_has_privs(). This always causes a false PMP violation on the starting address ofthe
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
range, as
addr - 1is not in the range.In order to fix, we just assume that all bytes from addr to the end of the page will be accessed if the size is unknown.
Signed-off-by Dayeol Lee <[email protected]>Reviewed-by Richard Henderson <[email protected]>--- target/riscv/cpu_helper.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
target/riscv/cpu_helper.c bool riscv_cpu_tlb_fill(CPUState *cs, vaddr
Quoted 5 lines
target/riscv/cpu_helper.c +0
-0@@ -441,6 +441,7 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddraddress,
Quoted 14 lines
Quoted 14 lines
Quoted 14 lines
Quoted 14 lines
target/riscv/cpu_helper.c +1
-0int size, CPURISCVState *env = &cpu->env;hwaddr pa =0 ;int prot; +int pmp_size =0 ;bool pmp_violation = false;int ret =TRANSLATE_FAIL ;int mode = mmu_idx;target/riscv/cpu_helper.c +0
-0@@ -460,9 +461,19 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,"%s address=%" VADDR_PRIx" ret %d physical "
TARGET_FMT_plx" prot %d\n", __func__, address, ret, pa, prot);Quoted 14 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
target/riscv/cpu_helper.c +6
-0+/* +* if size is unknown (0), assume that all bytes +* from addr to the end of the page will be accessed. +*/ +if (size ==0 ) { + pmp_size = -(address |TARGET_PAGE_MASK );Quoted 8 lines
Quoted 8 lines
target/riscv/cpu_helper.c bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,
Quoted 8 lines
target/riscv/cpu_helper.c +5
-1+ } else { + pmp_size = size; + } +if (riscv_feature (env,RISCV_FEATURE_PMP ) && (ret ==TRANSLATE_SUCCESS ) && - !pmp_hart_has_privs (env, pa, size,1 << access_type, mode)) { + !pmp_hart_has_privs(env, pa, pmp_size, 1 << access_type,mode))
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
{ ret = TRANSLATE_PMP_FAIL; } if (ret == TRANSLATE_PMP_FAIL) {--
2.20.1