[PATCH] mm/vmalloc: fix align value calculation error

[PATCH] mm/vmalloc: fix align value calculation error

[zijun_hu]

Re: [PATCH] mm/vmalloc: fix align value calculation error

[zijun_hu]

On 08/04/2016 04:02 PM, zijun_hu wrote:
>>From e40d1066f61394992e0167f259001ae9d2581dc1 Mon Sep 17 00:00:00 2001
> From: zijun_hu <zijun_hu@htc.com>
> Date: Thu, 4 Aug 2016 14:22:52 +0800
> Subject: [PATCH] mm/vmalloc: fix align value calculation error
> 
> it causes double align requirement for __get_vm_area_node() if parameter
> size is power of 2 and VM_IOREMAP is set in parameter flags
> 
> it is fixed by using order_base_2 instead of fls_long() due to lack of
> get_count_order() for long parameter
> 
> Signed-off-by: zijun_hu <zijun_hu@htc.com>
> ---
>  mm/vmalloc.c | 14 +++++++++++---
>  1 file changed, 11 insertions(+), 3 deletions(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index 91f44e7..8b17c51 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -1357,11 +1357,19 @@ static struct vm_struct *__get_vm_area_node(unsigned long size,
>  {
>  	struct vmap_area *va;
>  	struct vm_struct *area;
> +	int ioremap_size_order;
>  
>  	BUG_ON(in_interrupt());
> -	if (flags & VM_IOREMAP)
> -		align = 1ul << clamp_t(int, fls_long(size),
> -				       PAGE_SHIFT, IOREMAP_MAX_ORDER);
> +	if (flags & VM_IOREMAP) {
> +		if (unlikely(size < 2))
> +			ioremap_size_order = size;
> +		else if (unlikely((signed long)size < 0))
> +			ioremap_size_order = sizeof(size) * 8;
> +		else
> +			ioremap_size_order = order_base_2(size);
> +		align = 1ul << clamp_t(int, ioremap_size_order, PAGE_SHIFT,
> +				IOREMAP_MAX_ORDER);
> +	}
>  
>  	size = PAGE_ALIGN(size);
>  	if (unlikely(!size))
> 
another fix approach is shown as follows

From: zijun_hu <zijun_hu@htc.com>
Date: Thu, 4 Aug 2016 14:22:52 +0800
Subject: [PATCH] mm/vmalloc: fix align value calculation error

it causes double align requirement for __get_vm_area_node() if parameter
size is power of 2 and VM_IOREMAP is set in parameter flags

it is fixed by handling the specail case manually due to lack of
get_count_order() for long parameter

Signed-off-by: zijun_hu <zijun_hu@htc.com>
---
 mm/vmalloc.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index 91f44e7..dbbca8a 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -1357,11 +1357,16 @@ static struct vm_struct *__get_vm_area_node(unsigned long size,
 {
 	struct vmap_area *va;
 	struct vm_struct *area;
+	int ioremap_size_order;
 
 	BUG_ON(in_interrupt());
-	if (flags & VM_IOREMAP)
-		align = 1ul << clamp_t(int, fls_long(size),
-				       PAGE_SHIFT, IOREMAP_MAX_ORDER);
+	if (flags & VM_IOREMAP) {
+		ioremap_size_order = fls_long(size);
+		if (is_power_of_2(size) && size != 1)
+			ioremap_size_order--;
+		align = 1ul << clamp_t(int, ioremap_size_order, PAGE_SHIFT,
+				IOREMAP_MAX_ORDER);
+	}
 
 	size = PAGE_ALIGN(size);
 	if (unlikely(!size))
-- 
1.9.1

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] mm/vmalloc: fix align value calculation error

[Andrew Morton]

> 
> it causes double align requirement for __get_vm_area_node() if parameter
> size is power of 2 and VM_IOREMAP is set in parameter flags
> 
> it is fixed by handling the specail case manually due to lack of
> get_count_order() for long parameter
> 
> ...
>
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -1357,11 +1357,16 @@ static struct vm_struct *__get_vm_area_node(unsigned long size,
>  {
>  	struct vmap_area *va;
>  	struct vm_struct *area;
> +	int ioremap_size_order;
>  
>  	BUG_ON(in_interrupt());
> -	if (flags & VM_IOREMAP)
> -		align = 1ul << clamp_t(int, fls_long(size),
> -				       PAGE_SHIFT, IOREMAP_MAX_ORDER);
> +	if (flags & VM_IOREMAP) {
> +		ioremap_size_order = fls_long(size);
> +		if (is_power_of_2(size) && size != 1)
> +			ioremap_size_order--;
> +		align = 1ul << clamp_t(int, ioremap_size_order, PAGE_SHIFT,
> +				IOREMAP_MAX_ORDER);
> +	}
>  
>  	size = PAGE_ALIGN(size);
>  	if (unlikely(!size))

I'm having trouble with this, and a more complete description would
have helped!

As far as I can tell, the current code will decide the following:

size=0x10000: alignment=0x10000
size=0x0f000: alignment=0x8000

And your patch will change it so that

size=0x10000: alignment=0x8000
size=0x0f000: alignment=0x8000

Correct?

If so, I'm struggling to see the sense in this.  Shouldn't we be
changing things so that

size=0x10000: alignment=0x10000
size=0x0f000: alignment=0x10000

?

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] mm/vmalloc: fix align value calculation error

[zijun_hu]

On 08/05/2016 05:24 AM, Andrew Morton wrote:
>>
>> it causes double align requirement for __get_vm_area_node() if parameter
>> size is power of 2 and VM_IOREMAP is set in parameter flags
>>
>> it is fixed by handling the specail case manually due to lack of
>> get_count_order() for long parameter
>>
>> ...
>>
>> --- a/mm/vmalloc.c
>> +++ b/mm/vmalloc.c
>> @@ -1357,11 +1357,16 @@ static struct vm_struct *__get_vm_area_node(unsigned long size,
>>  {
>>  	struct vmap_area *va;
>>  	struct vm_struct *area;
>> +	int ioremap_size_order;
>>  
>>  	BUG_ON(in_interrupt());
>> -	if (flags & VM_IOREMAP)
>> -		align = 1ul << clamp_t(int, fls_long(size),
>> -				       PAGE_SHIFT, IOREMAP_MAX_ORDER);
>> +	if (flags & VM_IOREMAP) {
>> +		ioremap_size_order = fls_long(size);
>> +		if (is_power_of_2(size) && size != 1)
>> +			ioremap_size_order--;
>> +		align = 1ul << clamp_t(int, ioremap_size_order, PAGE_SHIFT,
>> +				IOREMAP_MAX_ORDER);
>> +	}
>>  
>>  	size = PAGE_ALIGN(size);
>>  	if (unlikely(!size))
> 
> I'm having trouble with this, and a more complete description would
> have helped!
> 
> As far as I can tell, the current code will decide the following:
> 
> size=0x10000: alignment=0x10000
> size=0x0f000: alignment=0x8000
>
no, the current code doesn't achieve the above results as shown below
size=0x10000 -> fls_long(0x10000)=17 -> alignment=0x20000
size=0x0f000 -> fls_long(0x0f000)=16 -> alignment=0x10000
it is wrong for power of 2 value such as size=0x10000
 
> And your patch will change it so that
> 
> size=0x10000: alignment=0x8000
> size=0x0f000: alignment=0x8000
> 
> Correct?
>
no, my patch will results in the following calculations
size=0x10000: alignment=0x10000
size=0x0f000: alignment=0x10000

> If so, I'm struggling to see the sense in this.  Shouldn't we be
> changing things so that
> 
> size=0x10000: alignment=0x10000
> size=0x0f000: alignment=0x10000
> 
> ?
okay, it is the aim of my patch as explained above
> 
> 


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] mm/vmalloc: fix align value calculation error

[zijun_hu]

On 2016/8/5 5:24, Andrew Morton wrote:
>>
>> it causes double align requirement for __get_vm_area_node() if parameter
>> size is power of 2 and VM_IOREMAP is set in parameter flags
>>
>> it is fixed by handling the specail case manually due to lack of
>> get_count_order() for long parameter
>>
>> ...
>>
>> --- a/mm/vmalloc.c
>> +++ b/mm/vmalloc.c
>> @@ -1357,11 +1357,16 @@ static struct vm_struct *__get_vm_area_node(unsigned long size,
>>  {
>>  	struct vmap_area *va;
>>  	struct vm_struct *area;
>> +	int ioremap_size_order;
>>  
>>  	BUG_ON(in_interrupt());
>> -	if (flags & VM_IOREMAP)
>> -		align = 1ul << clamp_t(int, fls_long(size),
>> -				       PAGE_SHIFT, IOREMAP_MAX_ORDER);
>> +	if (flags & VM_IOREMAP) {
>> +		ioremap_size_order = fls_long(size);
>> +		if (is_power_of_2(size) && size != 1)
>> +			ioremap_size_order--;
>> +		align = 1ul << clamp_t(int, ioremap_size_order, PAGE_SHIFT,
>> +				IOREMAP_MAX_ORDER);
>> +	}
>>  
>>  	size = PAGE_ALIGN(size);
>>  	if (unlikely(!size))
> 
> I'm having trouble with this, and a more complete description would
> have helped!
> 
> As far as I can tell, the current code will decide the following:
> 
> size=0x10000: alignment=0x10000
> size=0x0f000: alignment=0x8000
> 

no, the current code doesn't achieve the above results as shown below
size=0x10000 -> fls_long(0x10000)=17 -> alignment=0x20000
size=0x0f000 -> fls_long(0x0f000)=16 -> alignment=0x10000
it is wrong for power of 2 value such as size=0x10000

> And your patch will change it so that
> 
> size=0x10000: alignment=0x8000
> size=0x0f000: alignment=0x8000
> 
> Correct?
>

no, my patch will results in the following calculations
size=0x10000: alignment=0x10000
size=0x0f000: alignment=0x10000

> If so, I'm struggling to see the sense in this.  Shouldn't we be
> changing things so that
> 
> size=0x10000: alignment=0x10000
> size=0x0f000: alignment=0x10000
> 
> ?
okay, it is the aim of my patch as explained above
i provide another solution as shown below
i appreciate it since it is more canonical
please help to review and apply it kindly

Re: [PATCH] mm/vmalloc: fix align value calculation error

[Andrew Morton]

On Fri, 5 Aug 2016 23:48:21 +0800 zijun_hu <zijun_hu@zoho.com> wrote:

> From: zijun_hu <zijun_hu@htc.com>
> Date: Fri, 5 Aug 2016 22:10:07 +0800
> Subject: [PATCH 1/1] mm/vmalloc: fix align value calculation error
> 
> it causes double align requirement for __get_vm_area_node() if parameter
> size is power of 2 and VM_IOREMAP is set in parameter flags
> 
> get_order_long() is implemented and used instead of fls_long() for
> fixing the bug

Makes sense.  I think.

> --- a/include/linux/bitops.h
> +++ b/include/linux/bitops.h
> @@ -192,6 +192,23 @@ static inline unsigned fls_long(unsigned long l)
>  }
>  
>  /**
> + * get_order_long - get order after rounding @l up to power of 2
> + * @l: parameter
> + *
> + * it is same as get_count_order() but long type parameter
> + * or 0 is returned if @l == 0UL
> + */
> +static inline int get_order_long(unsigned long l)
> +{
> +	if (l == 0UL)
> +		return 0;
> +	else if (l & (l - 1UL))
> +		return fls_long(l);
> +	else
> +		return fls_long(l) - 1;
> +}
> +
> +/**
>   * __ffs64 - find first set bit in a 64 bit word
>   * @word: The 64 bit word
>   *
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index 91f44e7..7d717f3 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -1360,7 +1360,7 @@ static struct vm_struct *__get_vm_area_node(unsigned long size,
>  
>  	BUG_ON(in_interrupt());
>  	if (flags & VM_IOREMAP)
> -		align = 1ul << clamp_t(int, fls_long(size),
> +		align = 1ul << clamp_t(int, get_order_long(size),
>  				       PAGE_SHIFT, IOREMAP_MAX_ORDER);
>  
>  	size = PAGE_ALIGN(size);

It would be better to call this get_count_order_long(), I think?  To
match get_count_order()?

get_count_order() is a weird name and perhaps both of these should be
renamed to things which actually make sense.  That's a separate issue.

--- a/include/linux/bitops.h~mm-vmalloc-fix-align-value-calculation-error-fix
+++ a/include/linux/bitops.h
@@ -75,6 +75,23 @@ static inline int get_count_order(unsign
 	return order;
 }
 
+/**
+ * get_count_order_long - get order after rounding @l up to power of 2
+ * @l: parameter
+ *
+ * The same as get_count_order() but accepts a long type parameter
+ * or 0 is returned if @l == 0UL
+ */
+static inline int get_count_order_long(unsigned long l)
+{
+	if (l == 0UL)
+		return 0;
+	else if (l & (l - 1UL))
+		return fls_long(l);
+	else
+		return fls_long(l) - 1;
+}
+
 static __always_inline unsigned long hweight_long(unsigned long w)
 {
 	return sizeof(w) == 4 ? hweight32(w) : hweight64(w);
@@ -192,23 +209,6 @@ static inline unsigned fls_long(unsigned
 }
 
 /**
- * get_order_long - get order after rounding @l up to power of 2
- * @l: parameter
- *
- * it is same as get_count_order() but long type parameter
- * or 0 is returned if @l == 0UL
- */
-static inline int get_order_long(unsigned long l)
-{
-	if (l == 0UL)
-		return 0;
-	else if (l & (l - 1UL))
-		return fls_long(l);
-	else
-		return fls_long(l) - 1;
-}
-
-/**
  * __ffs64 - find first set bit in a 64 bit word
  * @word: The 64 bit word
  *
--- a/mm/vmalloc.c~mm-vmalloc-fix-align-value-calculation-error-fix
+++ a/mm/vmalloc.c
@@ -1360,7 +1360,7 @@ static struct vm_struct *__get_vm_area_n
 
 	BUG_ON(in_interrupt());
 	if (flags & VM_IOREMAP)
-		align = 1ul << clamp_t(int, get_order_long(size),
+		align = 1ul << clamp_t(int, get_count_order_long(size),
 				       PAGE_SHIFT, IOREMAP_MAX_ORDER);
 
 	size = PAGE_ALIGN(size);
_

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] mm/vmalloc: fix align value calculation error

[zijun_hu]

On 08/10/2016 05:28 AM, Andrew Morton wrote:
> On Fri, 5 Aug 2016 23:48:21 +0800 zijun_hu <zijun_hu@zoho.com> wrote:
> 
>> From: zijun_hu <zijun_hu@htc.com>
>> Date: Fri, 5 Aug 2016 22:10:07 +0800
>> Subject: [PATCH 1/1] mm/vmalloc: fix align value calculation error
>>
>> it causes double align requirement for __get_vm_area_node() if parameter
>> size is power of 2 and VM_IOREMAP is set in parameter flags
>>
>> get_order_long() is implemented and used instead of fls_long() for
>> fixing the bug
> 
> Makes sense.  I think.
> 
>> --- a/include/linux/bitops.h
>> +++ b/include/linux/bitops.h
>> @@ -192,6 +192,23 @@ static inline unsigned fls_long(unsigned long l)
>>  }
>>  
>>  /**
>> + * get_order_long - get order after rounding @l up to power of 2
>> + * @l: parameter
>> + *
>> + * it is same as get_count_order() but long type parameter
>> + * or 0 is returned if @l == 0UL
>> + */
>> +static inline int get_order_long(unsigned long l)
>> +{
>> +	if (l == 0UL)
>> +		return 0;
>> +	else if (l & (l - 1UL))
>> +		return fls_long(l);
>> +	else
>> +		return fls_long(l) - 1;
>> +}
>> +
>> +/**
>>   * __ffs64 - find first set bit in a 64 bit word
>>   * @word: The 64 bit word
>>   *
>> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
>> index 91f44e7..7d717f3 100644
>> --- a/mm/vmalloc.c
>> +++ b/mm/vmalloc.c
>> @@ -1360,7 +1360,7 @@ static struct vm_struct *__get_vm_area_node(unsigned long size,
>>  
>>  	BUG_ON(in_interrupt());
>>  	if (flags & VM_IOREMAP)
>> -		align = 1ul << clamp_t(int, fls_long(size),
>> +		align = 1ul << clamp_t(int, get_order_long(size),
>>  				       PAGE_SHIFT, IOREMAP_MAX_ORDER);
>>  
>>  	size = PAGE_ALIGN(size);
> 
> It would be better to call this get_count_order_long(), I think?  To
> match get_count_order()?
> 
yes, i agree with you to correct function name

i provide another patch called v2 based on your suggestion as shown below
it have following correction against original patch v1
1) use name get_count_order_long() instead of get_order_long()
2) return -1 if @l == 0 to consist with get_order_long()
3) cast type to int before returning from get_count_order_long()
4) move up function parameter checking for __get_vm_area_node()
5) more commit message is offered to make issue and approach clear
any comments about new patch is welcome

this new patch called patch v2 is shown below