Skip to contentSkip to conversation

platform/x86: wmi: Replace kmalloc + sprintf() with kasprintf()

PATCH In review Andy Shevchenko 4 participants 4 messages Feb 16, 2018 x86 platform drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 platform/x86: wmi: Replace kmalloc + sprintf() with kasprintf()

kasprintf() does the job of two: kmalloc() and sprintf(). Replace two calls with one.

---
 drivers/platform/x86/wmi.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/platform/x86/wmi.c +1 -2
index c0c8945603cb..4e76ffcb5394 100644
@@ -933,12 +933,11 @@ static int wmi_dev_probe(struct device *dev)
 			goto probe_failure;
 		}
 
-		buf = kmalloc(strlen(wdriver->driver.name) + 5, GFP_KERNEL);
+		buf = kasprintf(GFP_KERNEL, "wmi/%s", wdriver->driver.name);
 		if (!buf) {
 			ret = -ENOMEM;
 			goto probe_string_failure;
 		}
-		sprintf(buf, "wmi/%s", wdriver->driver.name);
 		wblock->char_dev.minor = MISC_DYNAMIC_MINOR;
 		wblock->char_dev.name = buf;
 		wblock->char_dev.fops = &wmi_fops;
Signature · 1 line
2.16.1
#1 lore raw

From: Andy Shevchenko

Quoted 4 lines

Sent: 16 February 2018 15:40

kasprintf() does the job of two: kmalloc() and sprintf(). Replace two calls with one.

...

drivers/platform/x86/wmi.c static int wmi_dev_probe(struct device *dev) · line 936

Quoted 2 lines
drivers/platform/x86/wmi.c +1 -1
-		buf = kmalloc(strlen(wdriver->driver.name) + 5, GFP_KERNEL);
+		buf = kasprintf(GFP_KERNEL, "wmi/%s", wdriver->driver.name);

...

Except that kasprintf() has no idea how long a buffer is needed. It might even do the printf twice just to get the length.

David

#2 lore raw
Quoted 8 lines · David Laight
Quoted 2 lines

kasprintf() does the job of two: kmalloc() and sprintf(). Replace two calls with one.

...

drivers/platform/x86/wmi.c static int wmi_dev_probe(struct device *dev) · line 936

Quoted 2 lines
drivers/platform/x86/wmi.c +1 -1
-             buf = kmalloc(strlen(wdriver->driver.name) + 5, GFP_KERNEL);
+             buf = kasprintf(GFP_KERNEL, "wmi/%s", wdriver->driver.name);

...

Except that kasprintf() has no idea how long a buffer is needed.

Hmm...

Quoted 1 line

It might even do the printf twice just to get the length.

It does exactly that. So what?

Signature · 2 lines
With Best Regards,
Andy Shevchenko
Reviewed-by
#3 lore raw
In reply to David Laight
Quoted 5 lines · David Laight

From: Andy Shevchenko

Quoted 4 lines

Sent: 16 February 2018 15:40

kasprintf() does the job of two: kmalloc() and sprintf(). Replace two calls with one.

Quoted 7 lines

...

drivers/platform/x86/wmi.c static int wmi_dev_probe(struct device *dev) · line 936

Quoted 2 lines
drivers/platform/x86/wmi.c +1 -1
-		buf = kmalloc(strlen(wdriver->driver.name) + 5, GFP_KERNEL);
+		buf = kasprintf(GFP_KERNEL, "wmi/%s", wdriver->driver.name);

...

Except that kasprintf() has no idea how long a buffer is needed. It might even do the printf twice just to get the length.

Sure, but this is one-time and non-critical path. Unfortunately there isn't any guidance in Documentation here. Of the 520 or so instances of kasprintf usages in the kernel, device name or similar is a very common pattern. Eliminating manual counting of characters for buffer allocation seems like a good plan to me.

So unless we want to argue that all those use cases are wrong, this would appear to be at least common practice, if not best practice for this particular pattern.

Signature · 2 lines
Darren Hart
VMware Open Source Technology Center