Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1572842 > unrolled thread

[PATCH v5 2/4] device property: constify property arrays values

Started byDmitry Torokhov <dmitry.torokhov@gmail.com>
First post2017-02-03 02:50 +0100
Last post2017-02-03 16:20 +0100
Articles 3 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  [PATCH v5 2/4] device property: constify property arrays values Dmitry Torokhov <dmitry.torokhov@gmail.com> - 2017-02-03 02:50 +0100
    Re: [PATCH v5 2/4] device property: constify property arrays values Andy Shevchenko <andriy.shevchenko@linux.intel.com> - 2017-02-03 12:50 +0100
      Re: [PATCH v5 2/4] device property: constify property arrays values Dmitry Torokhov <dmitry.torokhov@gmail.com> - 2017-02-03 16:20 +0100

#1572842 — [PATCH v5 2/4] device property: constify property arrays values

FromDmitry Torokhov <dmitry.torokhov@gmail.com>
Date2017-02-03 02:50 +0100
Subject[PATCH v5 2/4] device property: constify property arrays values
Message-ID<t6BaG-3o4-9@gated-at.bofh.it>
Data that is fed into property arrays should not be modified, so let's mark
relevant pointers as const. This will allow us making source arrays as
const/__initconst.

Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
---
 drivers/base/property.c  | 10 +++++-----
 include/linux/property.h | 12 ++++++------
 2 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/drivers/base/property.c b/drivers/base/property.c
index e9fa75645d69..31b942a29fdc 100644
--- a/drivers/base/property.c
+++ b/drivers/base/property.c
@@ -718,7 +718,8 @@ static void pset_free_set(struct property_set *pset)
 static int pset_copy_entry(struct property_entry *dst,
 			   const struct property_entry *src)
 {
-	const char **d, **s;
+	const char * const *s;
+	char **d;
 	size_t i, nval;
 
 	dst->name = kstrdup(src->name, GFP_KERNEL);
@@ -731,12 +732,11 @@ static int pset_copy_entry(struct property_entry *dst,
 
 		if (src->is_string) {
 			nval = src->length / sizeof(const char *);
-			dst->pointer.str = kcalloc(nval, sizeof(const char *),
-						   GFP_KERNEL);
-			if (!dst->pointer.str)
+			d = kcalloc(nval, sizeof(const char *), GFP_KERNEL);
+			if (!d)
 				return -ENOMEM;
 
-			d = dst->pointer.str;
+			dst->pointer.raw_data = d;
 			s = src->pointer.str;
 			for (i = 0; i < nval; i++) {
 				d[i] = kstrdup(s[i], GFP_KERNEL);
diff --git a/include/linux/property.h b/include/linux/property.h
index d37a4498b3ac..7a0a1cce5165 100644
--- a/include/linux/property.h
+++ b/include/linux/property.h
@@ -160,12 +160,12 @@ struct property_entry {
 	bool is_string;
 	union {
 		union {
-			void *raw_data;
-			u8 *u8_data;
-			u16 *u16_data;
-			u32 *u32_data;
-			u64 *u64_data;
-			const char **str;
+			const void *raw_data;
+			const u8 *u8_data;
+			const u16 *u16_data;
+			const u32 *u32_data;
+			const u64 *u64_data;
+			const char * const *str;
 		} pointer;
 		union {
 			unsigned long long raw_data;
-- 
2.11.0.483.g087da7b7c-goog

[toc] | [next] | [standalone]


#1573040

FromAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Date2017-02-03 12:50 +0100
Message-ID<t6Kxk-149-5@gated-at.bofh.it>
In reply to#1572842
On Thu, 2017-02-02 at 17:41 -0800, Dmitry Torokhov wrote:
> Data that is fed into property arrays should not be modified, so let's
> mark
> relevant pointers as const. This will allow us making source arrays as
> const/__initconst.
> 

> @@ -718,7 +718,8 @@ static void pset_free_set(struct property_set
> *pset)
>  static int pset_copy_entry(struct property_entry *dst,
>  			   const struct property_entry *src)
>  {
> -	const char **d, **s;
> +	const char * const *s;
> +	char **d;

You removed const here

>  	size_t i, nval;
>  
>  	dst->name = kstrdup(src->name, GFP_KERNEL);
> @@ -731,12 +732,11 @@ static int pset_copy_entry(struct property_entry
> *dst,
>  
>  		if (src->is_string) {
>  			nval = src->length / sizeof(const char *);
> -			dst->pointer.str = kcalloc(nval, sizeof(const
> char *),
> -						   GFP_KERNEL);
> -			if (!dst->pointer.str)
> +			d = kcalloc(nval, sizeof(const char *),
> GFP_KERNEL);

But left it here. Do we need to remove const?

> +			if (!d)
>  				return -ENOMEM;
>  
> -			d = dst->pointer.str;
> +			dst->pointer.raw_data = d;
>  			s = src->pointer.str;

So, overall, do we need these changes at all? Nothing in commit message
sheds a light on it.

>  			for (i = 0; i < nval; i++) {
>  				d[i] = kstrdup(s[i], GFP_KERNEL);

-- 
Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Intel Finland Oy

[toc] | [prev] | [next] | [standalone]


#1573143

FromDmitry Torokhov <dmitry.torokhov@gmail.com>
Date2017-02-03 16:20 +0100
Message-ID<t6NOy-3bO-17@gated-at.bofh.it>
In reply to#1573040
On Fri, Feb 03, 2017 at 01:43:03PM +0200, Andy Shevchenko wrote:
> On Thu, 2017-02-02 at 17:41 -0800, Dmitry Torokhov wrote:
> > Data that is fed into property arrays should not be modified, so let's
> > mark
> > relevant pointers as const. This will allow us making source arrays as
> > const/__initconst.
> > 
> 
> > @@ -718,7 +718,8 @@ static void pset_free_set(struct property_set
> > *pset)
> >  static int pset_copy_entry(struct property_entry *dst,
> >  			   const struct property_entry *src)
> >  {
> > -	const char **d, **s;
> > +	const char * const *s;
> > +	char **d;
> 
> You removed const here

Yes I did. It is hard to assign value to a constant otherwise.

> 
> >  	size_t i, nval;
> >  
> >  	dst->name = kstrdup(src->name, GFP_KERNEL);
> > @@ -731,12 +732,11 @@ static int pset_copy_entry(struct property_entry
> > *dst,
> >  
> >  		if (src->is_string) {
> >  			nval = src->length / sizeof(const char *);
> > -			dst->pointer.str = kcalloc(nval, sizeof(const
> > char *),
> > -						   GFP_KERNEL);
> > -			if (!dst->pointer.str)
> > +			d = kcalloc(nval, sizeof(const char *),
> > GFP_KERNEL);
> 
> But left it here. Do we need to remove const?

I do not know why we had it in the first place: the size is the samei
between constant and variable of the same type.

Ideally we'd use sizeof(*d), I can do it after this batch is accepted.

> 
> > +			if (!d)
> >  				return -ENOMEM;
> >  
> > -			d = dst->pointer.str;
> > +			dst->pointer.raw_data = d;
> >  			s = src->pointer.str;
> 
> So, overall, do we need these changes at all? Nothing in commit message
> sheds a light on it.

The compiler insists in them though.

> 
> >  			for (i = 0; i < nval; i++) {
> >  				d[i] = kstrdup(s[i], GFP_KERNEL);

Thanks.

-- 
Dmitry

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web