Skip to contentSkip to conversation

lib/vsprintf: Replace hidden BUILD_BUG_ON() with static_assert()

PATCH Superseded Andy Shevchenko 3 participants 10 messages Jul 31, 2020

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Andy Shevchenko

author reviewer
#0 lore raw

PATCH 1/3 lib/vsprintf: Replace hidden BUILD_BUG_ON() with static_assert()

First of all, there is no compile time check for the SMALL to be ' ' (0x20, i.e. space). Second, for ZEROPAD the check is hidden in the code.

For better maintenance replace BUILD_BUG_ON() with static_assert() for ZEROPAD and move it closer to the definition. While at it, introduce check for SMALL.

---
 lib/vsprintf.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
lib/vsprintf.c +4 -1
index 31a674dd2674..8a8ac7ce0289 100644
@@ -381,6 +381,9 @@ int num_to_str(char *buf, int size, unsigned long long num, unsigned int width)
 #define SMALL	32		/* use lowercase in hex (must be 32 == 0x20) */
 #define SPECIAL	64		/* prefix hex with "0x", octal with "0" */
 
+static_assert(ZEROPAD + ' ' == '0');
+static_assert(SMALL == ' ');
+
 enum format_type {
 	FORMAT_TYPE_NONE, /* Just a string part */
 	FORMAT_TYPE_WIDTH,
@@ -507,7 +510,7 @@ char *number(char *buf, char *end, unsigned long long num,
 	/* zero or space padding */
 	if (!(spec.flags & LEFT)) {
 		char c = ' ' + (spec.flags & ZEROPAD);
-		BUILD_BUG_ON(' ' + ZEROPAD != '0');
+
 		while (--field_width >= 0) {
 			if (buf < end)
 				*buf = c;
Signature · 1 line
2.27.0
author reviewer
#1 lore raw

PATCH 2/3 lib/vsprintf: Replace custom spec to print decimals with generic one

When printing phandle in %pOF the custom spec is used. First of all, it has SMALL flag which makes no sense for decimal numbers. Second, we have already default spec for decimal numbers. Use the latter in %pOF case as well.

---
 lib/vsprintf.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
lib/vsprintf.c +1 -7
index 8a8ac7ce0289..90d818ef03c5 100644
@@ -1979,12 +1979,6 @@ char *device_node_string(char *buf, char *end, struct device_node *dn,
 	char *buf_start = buf;
 	struct property *prop;
 	bool has_mult, pass;
-	static const struct printf_spec num_spec = {
-		.flags = SMALL,
-		.field_width = -1,
-		.precision = -1,
-		.base = 10,
-	};
 
 	struct printf_spec str_spec = spec;
 	str_spec.field_width = -1;
@@ -2024,7 +2018,7 @@ char *device_node_string(char *buf, char *end, struct device_node *dn,
 			str_spec.precision = precision;
 			break;
 		case 'p':	/* phandle */
-			buf = number(buf, end, (unsigned int)dn->phandle, num_spec);
+			buf = number(buf, end, (unsigned int)dn->phandle, default_dec_spec);
 			break;
 		case 'P':	/* path-spec */
 			p = fwnode_get_name(of_fwnode_handle(dn));
Signature · 1 line
2.27.0
maintainer Reviewed-by on v1 2/3
#2 lore raw

Re: [PATCH v1 2/3] lib/vsprintf: Replace custom spec to print decimals with generic one

Quoted 4 lines · On Fri, 31 Jul 2020 15:31:44 +0300 Andy Shevchenko…

When printing phandle in %pOF the custom spec is used. First of all, it has SMALL flag which makes no sense for decimal numbers. Second, we have already default spec for decimal numbers. Use the latter in %pOF case as well.

The change log threw me off, as this is specifically for %pOFp. This only affects the phandle ('p') portion. Please update the change log to reflect that.

Also, I would add in Cc those that added this code and looked it over:

Cc Pantelis Antoniou <[email protected]>
Cc Joe Perches <[email protected]>
Cc Rob Herring <[email protected]>

And mentioned in that code's change log:

Cc Grant Likely <[email protected]>

Otherwise,

-- Steve

Quoted 32 lines
---
 lib/vsprintf.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
lib/vsprintf.c +1 -7
index 8a8ac7ce0289..90d818ef03c5 100644
@@ -1979,12 +1979,6 @@ char *device_node_string(char *buf, char *end, struct device_node *dn,
 	char *buf_start = buf;
 	struct property *prop;
 	bool has_mult, pass;
-	static const struct printf_spec num_spec = {
-		.flags = SMALL,
-		.field_width = -1,
-		.precision = -1,
-		.base = 10,
-	};
 
 	struct printf_spec str_spec = spec;
 	str_spec.field_width = -1;
@@ -2024,7 +2018,7 @@ char *device_node_string(char *buf, char *end, struct device_node *dn,
 			str_spec.precision = precision;
 			break;
 		case 'p':	/* phandle */
-			buf = number(buf, end, (unsigned int)dn->phandle, num_spec);
+			buf = number(buf, end, (unsigned int)dn->phandle, default_dec_spec);
 			break;
 		case 'P':	/* path-spec */
 			p = fwnode_get_name(of_fwnode_handle(dn));
#3 lore raw

Re: [PATCH v1 2/3] lib/vsprintf: Replace custom spec to print decimals with generic one

Quoted 4 lines · Steven Rostedt

On Fri, 31 Jul 2020 15:31:44 +0300
Andy Shevchenko <[email protected]> wrote:

Quoted 8 lines
Quoted 4 lines

When printing phandle in %pOF the custom spec is used. First of all, it has SMALL flag which makes no sense for decimal numbers. Second, we have already default spec for decimal numbers. Use the latter in %pOF case as well.

The change log threw me off, as this is specifically for %pOFp. This only affects the phandle ('p') portion. Please update the change log to reflect that.

I think phandle implies this, but I update for sake of clarification.

Quoted 9 lines

Also, I would add in Cc those that added this code and looked it over:

Cc Pantelis Antoniou <[email protected]>
Cc Joe Perches <[email protected]>
Cc Rob Herring <[email protected]>

And mentioned in that code's change log:

Cc Grant Likely <[email protected]>

Okay, will do.
Also I have to mention that this (actually entire series) has been tested against test_printf and OF unittest modules.

Quoted 41 lines

Otherwise,

-- Steve

Quoted 32 lines
---
 lib/vsprintf.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
lib/vsprintf.c +1 -7
index 8a8ac7ce0289..90d818ef03c5 100644
@@ -1979,12 +1979,6 @@ char *device_node_string(char *buf, char *end, struct device_node *dn,
      char *buf_start = buf;
      struct property *prop;
      bool has_mult, pass;
-     static const struct printf_spec num_spec = {
-             .flags = SMALL,
-             .field_width = -1,
-             .precision = -1,
-             .base = 10,
-     };
 
      struct printf_spec str_spec = spec;
      str_spec.field_width = -1;
@@ -2024,7 +2018,7 @@ char *device_node_string(char *buf, char *end, struct device_node *dn,
                      str_spec.precision = precision;
                      break;
              case 'p':       /* phandle */
-                     buf = number(buf, end, (unsigned int)dn->phandle, num_spec);
+                     buf = number(buf, end, (unsigned int)dn->phandle, default_dec_spec);
                      break;
              case 'P':       /* path-spec */
                      p = fwnode_get_name(of_fwnode_handle(dn));
Signature · 2 lines
With Best Regards,
Andy Shevchenko
author reviewer
#4 lore raw

PATCH 3/3 lib/vsprintf: Force type of flags for gfp_t

In reply to Andy Shevchenko

Sparse is not happy about restricted type being assigned:
lib/vsprintf.c:1940:23: warning: incorrect type in assignment (different base types)
lib/vsprintf.c:1940:23: expected unsigned long [assigned] flags
lib/vsprintf.c:1940:23: got restricted gfp_t [usertype]

Force type of flags to make sparse happy.

---
 lib/vsprintf.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/vsprintf.c +1 -1
index 90d818ef03c5..118e2727d058 100644
@@ -1937,7 +1937,7 @@ char *flags_string(char *buf, char *end, void *flags_ptr,
 		names = vmaflag_names;
 		break;
 	case 'g':
-		flags = *(gfp_t *)flags_ptr;
+		flags = (__force typeof(flags))(*(gfp_t *)flags_ptr);
 		names = gfpflag_names;
 		break;
 	default:
Signature · 1 line
2.27.0
maintainer
#5 lore raw

Re: [PATCH v1 3/3] lib/vsprintf: Force type of flags for gfp_t

Quoted 17 lines · On Fri, 31 Jul 2020 15:31:45 +0300 Andy Shevchenko…

Sparse is not happy about restricted type being assigned:
lib/vsprintf.c:1940:23: warning: incorrect type in assignment (different base types)
lib/vsprintf.c:1940:23: expected unsigned long [assigned] flags
lib/vsprintf.c:1940:23: got restricted gfp_t [usertype]

Force type of flags to make sparse happy.

---
 lib/vsprintf.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/vsprintf.c +0 -0
index 90d818ef03c5..118e2727d058 100644

lib/vsprintf.c char *flags_string(char *buf, char *end, void *flags_ptr, · line 1940

Quoted 5 lines
lib/vsprintf.c +1 -1
@@ -1937,7 +1937,7 @@ char *flags_string(char *buf, char *end, void *flags_ptr,
 		names = vmaflag_names;
 		break;
 	case 'g':
-		flags = *(gfp_t *)flags_ptr;
+		flags = (__force typeof(flags))(*(gfp_t *)flags_ptr);

Do we really need to say "typeof(flags)" ? What about simply using flags' type?

		flags = (__force unsigned long)(*(gfp_t *)flags_ptr);

?

I mean, it's not like flags is a global. It's defined a few lines above.

-- Steve

Quoted 3 lines
lib/vsprintf.c +0 -0
 		names = gfpflag_names;
 		break;
 	default:
#6 lore raw

Re: [PATCH v1 3/3] lib/vsprintf: Force type of flags for gfp_t

Quoted 16 lines · Steven Rostedt
Quoted 12 lines · On Fri, 31 Jul 2020 15:31:45 +0300 Andy Shevchenko…

Sparse is not happy about restricted type being assigned:
lib/vsprintf.c:1940:23: warning: incorrect type in assignment (different base types)
lib/vsprintf.c:1940:23: expected unsigned long [assigned] flags
lib/vsprintf.c:1940:23: got restricted gfp_t [usertype]

Force type of flags to make sparse happy.

---
 lib/vsprintf.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 13 lines

lib/vsprintf.c char *flags_string(char *buf, char *end, void *flags_ptr, · line 1940

Quoted 10 lines
lib/vsprintf.c +1 -1
index 90d818ef03c5..118e2727d058 100644
@@ -1937,7 +1937,7 @@ char *flags_string(char *buf, char *end, void *flags_ptr,
              names = vmaflag_names;
              break;
      case 'g':
-             flags = *(gfp_t *)flags_ptr;
+             flags = (__force typeof(flags))(*(gfp_t *)flags_ptr);

Do we really need to say "typeof(flags)" ? What about simply using flags' type?

Whatever you prefer. I actually came with the latter and switched to the former. So, I'll switch back for v2.

Quoted 14 lines
                flags = (__force unsigned long)(*(gfp_t *)flags_ptr);

?

I mean, it's not like flags is a global. It's defined a few lines above.

-- Steve

Quoted 3 lines
lib/vsprintf.c +0 -0
              names = gfpflag_names;
              break;
      default:
Signature · 2 lines
With Best Regards,
Andy Shevchenko
maintainer
#7 lore raw

Re: [PATCH v1 3/3] lib/vsprintf: Force type of flags for gfp_t

Quoted 11 lines · On Fri, 31 Jul 2020 18:49:46 +0300 Andy Shevchenko <[email protected]>…
Quoted 10 lines

lib/vsprintf.c char *flags_string(char *buf, char *end, void *flags_ptr, · line 1940

Quoted 7 lines

+++ b/lib/vsprintf.c

lib/vsprintf.c +1 -1
@@ -1937,7 +1937,7 @@ char *flags_string(char *buf, char *end, void *flags_ptr,
              names = vmaflag_names;
              break;
      case 'g':
-             flags = *(gfp_t *)flags_ptr;
+             flags = (__force typeof(flags))(*(gfp_t *)flags_ptr);  

Do we really need to say "typeof(flags)" ? What about simply using flags' type?

Quoted 2 lines

Whatever you prefer. I actually came with the latter and switched to the former. So, I'll switch back for v2.

For this instance, I prefer spelling out the type.

Thanks!

-- Steve

maintainer Reviewed-by on v1 1/3
#8 lore raw
In reply to Andy Shevchenko
Quoted 18 lines · On Fri, 31 Jul 2020 15:31:43 +0300 Andy Shevchenko…

First of all, there is no compile time check for the SMALL to be ' ' (0x20, i.e. space). Second, for ZEROPAD the check is hidden in the code.

For better maintenance replace BUILD_BUG_ON() with static_assert() for ZEROPAD and move it closer to the definition. While at it, introduce check for SMALL.

---
 lib/vsprintf.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
lib/vsprintf.c +0 -0
index 31a674dd2674..8a8ac7ce0289 100644

lib/vsprintf.c int num_to_str(char *buf, int size, unsigned long long num, unsigned int width) · line 384

Quoted 4 lines
lib/vsprintf.c +1 -0
@@ -381,6 +381,9 @@ int num_to_str(char *buf, int size, unsigned long long num, unsigned int width)
 #define SMALL	32		/* use lowercase in hex (must be 32 == 0x20) */
 #define SPECIAL	64		/* prefix hex with "0x", octal with "0" */
 
+static_assert(ZEROPAD + ' ' == '0');

I would have it match the comment above:

static_assert(ZEROPAD == ('0' - ' '));

But other than that:

-- Steve

Quoted 14 lines
lib/vsprintf.c +3 -1
+static_assert(SMALL == ' ');
+
 enum format_type {
 	FORMAT_TYPE_NONE, /* Just a string part */
 	FORMAT_TYPE_WIDTH,
@@ -507,7 +510,7 @@ char *number(char *buf, char *end, unsigned long long num,
 	/* zero or space padding */
 	if (!(spec.flags & LEFT)) {
 		char c = ' ' + (spec.flags & ZEROPAD);
-		BUILD_BUG_ON(' ' + ZEROPAD != '0');
+
 		while (--field_width >= 0) {
 			if (buf < end)
 				*buf = c;
#9 lore raw
Quoted 25 lines · Steven Rostedt
Quoted 21 lines · On Fri, 31 Jul 2020 15:31:43 +0300 Andy Shevchenko…

First of all, there is no compile time check for the SMALL to be ' ' (0x20, i.e. space). Second, for ZEROPAD the check is hidden in the code.

For better maintenance replace BUILD_BUG_ON() with static_assert() for ZEROPAD and move it closer to the definition. While at it, introduce check for SMALL.

---
 lib/vsprintf.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
lib/vsprintf.c +0 -0
index 31a674dd2674..8a8ac7ce0289 100644
@@ -381,6 +381,9 @@ int num_to_str(char *buf, int size, unsigned long long num, unsigned int width)
 #define SMALL        32              /* use lowercase in hex (must be 32 == 0x20) */
 #define SPECIAL      64              /* prefix hex with "0x", octal with "0" */
 
Quoted 5 lines

lib/vsprintf.c int num_to_str(char *buf, int size, unsigned long long num, unsigned int width) · line 384

Quoted 1 line
lib/vsprintf.c +1 -0
+static_assert(ZEROPAD + ' ' == '0');

I would have it match the comment above:

static_assert(ZEROPAD == ('0' - ' '));

Okay, makes sense.
Will do in v2.

Quoted 3 lines

But other than that:

Thanks!

Quoted 18 lines

-- Steve

Quoted 14 lines
lib/vsprintf.c +3 -1
+static_assert(SMALL == ' ');
+
 enum format_type {
      FORMAT_TYPE_NONE, /* Just a string part */
      FORMAT_TYPE_WIDTH,
@@ -507,7 +510,7 @@ char *number(char *buf, char *end, unsigned long long num,
      /* zero or space padding */
      if (!(spec.flags & LEFT)) {
              char c = ' ' + (spec.flags & ZEROPAD);
-             BUILD_BUG_ON(' ' + ZEROPAD != '0');
+
              while (--field_width >= 0) {
                      if (buf < end)
                              *buf = c;
Signature · 2 lines
With Best Regards,
Andy Shevchenko