[3/4] arm64/mm: Consolidate page fault information capture
diff mbox series

Message ID 1559133285-27986-4-git-send-email-anshuman.khandual@arm.com
State New
Headers show
Series
  • arm64/mm: Fixes and cleanups for do_page_fault()
Related show

Commit Message

Anshuman Khandual May 29, 2019, 12:34 p.m. UTC
This consolidates page fault information capture and move them bit earlier.
While here it also adds an wrapper is_write_abort(). It also saves some
cycles by replacing multiple user_mode() calls into a single one earlier
during the fault.

Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: James Morse <james.morse@arm.com> 
Cc: Andrey Konovalov <andreyknvl@google.com>
---
 arch/arm64/mm/fault.c | 22 +++++++++++++++-------
 1 file changed, 15 insertions(+), 7 deletions(-)

Comments

Mark Rutland May 29, 2019, 2:53 p.m. UTC | #1
On Wed, May 29, 2019 at 06:04:44PM +0530, Anshuman Khandual wrote:
> This consolidates page fault information capture and move them bit earlier.
> While here it also adds an wrapper is_write_abort(). It also saves some
> cycles by replacing multiple user_mode() calls into a single one earlier
> during the fault.

To be honest, I doubt this has any measureable impact, but I agree that
using variables _may_ make the flow control easier to understand.

> 
> Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
> Cc: Catalin Marinas <catalin.marinas@arm.com>
> Cc: Will Deacon <will.deacon@arm.com>
> Cc: Mark Rutland <mark.rutland@arm.com>
> Cc: James Morse <james.morse@arm.com> 
> Cc: Andrey Konovalov <andreyknvl@google.com>
> ---
>  arch/arm64/mm/fault.c | 22 +++++++++++++++-------
>  1 file changed, 15 insertions(+), 7 deletions(-)
> 
> diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
> index da02678..170c71f 100644
> --- a/arch/arm64/mm/fault.c
> +++ b/arch/arm64/mm/fault.c
> @@ -435,6 +435,11 @@ static bool is_el0_instruction_abort(unsigned int esr)
>  	return ESR_ELx_EC(esr) == ESR_ELx_EC_IABT_LOW;
>  }
>  
> +static bool is_write_abort(unsigned int esr)
> +{
> +	return (esr & ESR_ELx_WNR) && !(esr & ESR_ELx_CM);
> +}

In off-list review, I mentioned that this isn't true for EL1, and I
think that we should name this 'is_el0_write_abort()' or add a comment
explaining the caveats if factored into a helper.

Thanks,
Mark.

> +
>  static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
>  				   struct pt_regs *regs)
>  {
> @@ -443,6 +448,9 @@ static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
>  	vm_fault_t fault, major = 0;
>  	unsigned long vm_flags = VM_READ | VM_WRITE;
>  	unsigned int mm_flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
> +	bool is_user = user_mode(regs);
> +	bool is_el0_exec = is_el0_instruction_abort(esr);
> +	bool is_write = is_write_abort(esr);
>  
>  	if (notify_page_fault(regs, esr))
>  		return 0;
> @@ -454,12 +462,12 @@ static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
>  	if (faulthandler_disabled() || !mm)
>  		goto no_context;
>  
> -	if (user_mode(regs))
> +	if (is_user)
>  		mm_flags |= FAULT_FLAG_USER;
>  
> -	if (is_el0_instruction_abort(esr)) {
> +	if (is_el0_exec) {
>  		vm_flags = VM_EXEC;
> -	} else if ((esr & ESR_ELx_WNR) && !(esr & ESR_ELx_CM)) {
> +	} else if (is_write) {
>  		vm_flags = VM_WRITE;
>  		mm_flags |= FAULT_FLAG_WRITE;
>  	}
> @@ -487,7 +495,7 @@ static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
>  	 * we can bug out early if this is from code which shouldn't.
>  	 */
>  	if (!down_read_trylock(&mm->mmap_sem)) {
> -		if (!user_mode(regs) && !search_exception_tables(regs->pc))
> +		if (!is_user && !search_exception_tables(regs->pc))
>  			goto no_context;
>  retry:
>  		down_read(&mm->mmap_sem);
> @@ -498,7 +506,7 @@ static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
>  		 */
>  		might_sleep();
>  #ifdef CONFIG_DEBUG_VM
> -		if (!user_mode(regs) && !search_exception_tables(regs->pc)) {
> +		if (!is_user && !search_exception_tables(regs->pc)) {
>  			up_read(&mm->mmap_sem);
>  			goto no_context;
>  		}
> @@ -516,7 +524,7 @@ static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
>  		 * in __lock_page_or_retry in mm/filemap.c.
>  		 */
>  		if (fatal_signal_pending(current)) {
> -			if (!user_mode(regs))
> +			if (!is_user)
>  				goto no_context;
>  			return 0;
>  		}
> @@ -561,7 +569,7 @@ static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
>  	 * If we are in kernel mode at this point, we have no context to
>  	 * handle this fault with.
>  	 */
> -	if (!user_mode(regs))
> +	if (!is_user)
>  		goto no_context;
>  
>  	if (fault & VM_FAULT_OOM) {
> -- 
> 2.7.4
>
Anshuman Khandual May 31, 2019, 9:10 a.m. UTC | #2
On 05/29/2019 08:23 PM, Mark Rutland wrote:
> On Wed, May 29, 2019 at 06:04:44PM +0530, Anshuman Khandual wrote:
>> This consolidates page fault information capture and move them bit earlier.
>> While here it also adds an wrapper is_write_abort(). It also saves some
>> cycles by replacing multiple user_mode() calls into a single one earlier
>> during the fault.
> 
> To be honest, I doubt this has any measureable impact, but I agree that
> using variables _may_ make the flow control easier to understand.
> 
>>
>> Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
>> Cc: Catalin Marinas <catalin.marinas@arm.com>
>> Cc: Will Deacon <will.deacon@arm.com>
>> Cc: Mark Rutland <mark.rutland@arm.com>
>> Cc: James Morse <james.morse@arm.com> 
>> Cc: Andrey Konovalov <andreyknvl@google.com>
>> ---
>>  arch/arm64/mm/fault.c | 22 +++++++++++++++-------
>>  1 file changed, 15 insertions(+), 7 deletions(-)
>>
>> diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
>> index da02678..170c71f 100644
>> --- a/arch/arm64/mm/fault.c
>> +++ b/arch/arm64/mm/fault.c
>> @@ -435,6 +435,11 @@ static bool is_el0_instruction_abort(unsigned int esr)
>>  	return ESR_ELx_EC(esr) == ESR_ELx_EC_IABT_LOW;
>>  }
>>  
>> +static bool is_write_abort(unsigned int esr)
>> +{
>> +	return (esr & ESR_ELx_WNR) && !(esr & ESR_ELx_CM);
>> +}
> 
> In off-list review, I mentioned that this isn't true for EL1, and I
> think that we should name this 'is_el0_write_abort()' or add a comment
> explaining the caveats if factored into a helper.
> 
> Thanks,
> Mark.

Okay will change the wrapper name to is_el0_write_abort() and add a comment
explaining how this is only applicable to aborts originating from EL0.

Patch
diff mbox series

diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
index da02678..170c71f 100644
--- a/arch/arm64/mm/fault.c
+++ b/arch/arm64/mm/fault.c
@@ -435,6 +435,11 @@  static bool is_el0_instruction_abort(unsigned int esr)
 	return ESR_ELx_EC(esr) == ESR_ELx_EC_IABT_LOW;
 }
 
+static bool is_write_abort(unsigned int esr)
+{
+	return (esr & ESR_ELx_WNR) && !(esr & ESR_ELx_CM);
+}
+
 static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
 				   struct pt_regs *regs)
 {
@@ -443,6 +448,9 @@  static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
 	vm_fault_t fault, major = 0;
 	unsigned long vm_flags = VM_READ | VM_WRITE;
 	unsigned int mm_flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
+	bool is_user = user_mode(regs);
+	bool is_el0_exec = is_el0_instruction_abort(esr);
+	bool is_write = is_write_abort(esr);
 
 	if (notify_page_fault(regs, esr))
 		return 0;
@@ -454,12 +462,12 @@  static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
 	if (faulthandler_disabled() || !mm)
 		goto no_context;
 
-	if (user_mode(regs))
+	if (is_user)
 		mm_flags |= FAULT_FLAG_USER;
 
-	if (is_el0_instruction_abort(esr)) {
+	if (is_el0_exec) {
 		vm_flags = VM_EXEC;
-	} else if ((esr & ESR_ELx_WNR) && !(esr & ESR_ELx_CM)) {
+	} else if (is_write) {
 		vm_flags = VM_WRITE;
 		mm_flags |= FAULT_FLAG_WRITE;
 	}
@@ -487,7 +495,7 @@  static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
 	 * we can bug out early if this is from code which shouldn't.
 	 */
 	if (!down_read_trylock(&mm->mmap_sem)) {
-		if (!user_mode(regs) && !search_exception_tables(regs->pc))
+		if (!is_user && !search_exception_tables(regs->pc))
 			goto no_context;
 retry:
 		down_read(&mm->mmap_sem);
@@ -498,7 +506,7 @@  static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
 		 */
 		might_sleep();
 #ifdef CONFIG_DEBUG_VM
-		if (!user_mode(regs) && !search_exception_tables(regs->pc)) {
+		if (!is_user && !search_exception_tables(regs->pc)) {
 			up_read(&mm->mmap_sem);
 			goto no_context;
 		}
@@ -516,7 +524,7 @@  static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
 		 * in __lock_page_or_retry in mm/filemap.c.
 		 */
 		if (fatal_signal_pending(current)) {
-			if (!user_mode(regs))
+			if (!is_user)
 				goto no_context;
 			return 0;
 		}
@@ -561,7 +569,7 @@  static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
 	 * If we are in kernel mode at this point, we have no context to
 	 * handle this fault with.
 	 */
-	if (!user_mode(regs))
+	if (!is_user)
 		goto no_context;
 
 	if (fault & VM_FAULT_OOM) {