Skip to contentSkip to conversation

missing vprintf in kernel.api. Interest in patch?

DISCUSSION [email protected] 2 participants 2 messages Feb 10, 2004 – Feb 11, 2004

lore
Share & tools
lore
#0 lore raw

Hi,

while writing a new kernel module I needed vprintf behaviour to send
additional info to printk (e.g. __LINE__ and __FUNCTION__ preprocessor macros). While the kernel API contains vsnprintf and vsprintf functions, it does not contain a vprintf and since we have no access to the printk_buf structure it is therefore impossible to directly write a formatted argument using a va_list to the log.

As a quick fix I wrote my own function using vsnprintf and printk, but this is far from ideal when more people need such functionality.

Indeed, doing a quick grep on the (2.6.1) tree I noticed that I am not the only one creating my own vprintf. In arch/um/kernel/tt/tracer.c, drivers/acpi/osl.c and drivers/acpi/utilities/utdebug.c others have also written similar functions. Not to mention all the DPRINTF, YYDPRINTF, etc. macros declared everywhere

Therefore my question is this: is there any interest in

(1) a small patch that extracts the vsnprintf(printk_buf,...) from printk and thus creates a vprintf function (trivial change, perhaps +5 lines of code). and/or
(2) a kernel wide DPRINTK (or something) macro that adds the __FILE__, __FUNCTION__ and __LINE__ macros to the output.
and/or
(3) an extension of 2 that also prints out clockcycles or timing statistics.

IMHO, everyone debugging and optimizing his code needs these features. It might help clean up the codebase if everyone uses the same helper functions. And since I already have such code laying around, sending in a patch is a trivial task right now.

Please share your views, I don't want to create a patch if it'll be rejected instantaneously because of some unspoken rule I haven't heard of.

cheers,

Willem de Bruijn

#1 lore raw
Quoted 4 lines · On Tue, Feb 10, 2004 at 11:22:59AM +0100, [email protected] wrote:

Therefore my question is this: is there any interest in

(1) a small patch that extracts the vsnprintf(printk_buf,...) from printk and thus creates a vprintf function (trivial change, perhaps +5 lines of code).

I already have this patch in my -tiny tree as well as some other patches to convert some would-be users like ext3_error. Andrew, interested in these?

Add vprintk call

 tiny-mpm/include/linux/kernel.h |    1 +
 tiny-mpm/kernel/printk.c        |   14 ++++++++++++--
 2 files changed, 13 insertions(+), 2 deletions(-)
include/linux/kernel.h renamed from include/linux/kernel.h~vprintk +1 -0
diff -puN include/linux/kernel.h~vprintk include/linux/kernel.h
@@ -84,6 +84,7 @@ extern unsigned long long memparse(char 
 extern int kernel_text_address(unsigned long addr);
 extern int session_of_pgrp(int pgrp);
 
+asmlinkage int vprintk(const char *fmt, va_list args);
 asmlinkage int printk(const char * fmt, ...)
 	__attribute__ ((format (printf, 1, 2)));
 
kernel/printk.c renamed from kernel/printk.c~vprintk +12 -2
diff -puN kernel/printk.c~vprintk kernel/printk.c
@@ -474,6 +474,17 @@ static void emit_log_char(char c)
 asmlinkage int printk(const char *fmt, ...)
 {
 	va_list args;
+	int r;
+
+	va_start(args, fmt);
+	r = vprintk(fmt, args);
+	va_end(args);
+
+	return r;
+}
+
+asmlinkage int vprintk(const char *fmt, va_list args)
+{
 	unsigned long flags;
 	int printed_len;
 	char *p;
@@ -491,9 +502,7 @@ asmlinkage int printk(const char *fmt, .
 	spin_lock_irqsave(&logbuf_lock, flags);
 
 	/* Emit the output into the temporary buffer */
-	va_start(args, fmt);
 	printed_len = vsnprintf(printk_buf, sizeof(printk_buf), fmt, args);
-	va_end(args);
 
 	/*
 	 * Copy the output into log_buf.  If the caller didn't provide
@@ -543,6 +552,7 @@ out:
 	return printed_len;
 }
 EXPORT_SYMBOL(printk);
+EXPORT_SYMBOL(vprintk);
 
 /**
  * acquire_console_sem - lock the console system for exclusive use.

_

Signature · 1 line
Matt Mackall : http://www.selenic.com : Linux development and consulting