Hi,
I have an RT-Linux (2.2.14-rtl2.2) program that is doing quite a few
rtl_printf()s (for tracing), that (after a moment) panic'd the
kernel (consistently). Disabling a subset of these print statements
seemed to make the problem go away. Is this a known problem, or
should I start learning how to debug the kernel crashes? (Yeah, I
know, I should probably learn that anyway...)
If anybody is interested, below is the debug macro I cooked up that
I use from my code:
#ifdef DEBUG
extern int DebugLevel; /* define this in your RT module */
#define dprintf(level, fmt, args...) do {\
if (level < DebugLevel) { \
rtl_printf("<"#level">" "!" fmt, ## args); \
} } while(0)
#else
#define dprintf(level, fmt, args...) do {/*nothing*/} while(0)
#endif /* DEBUG */
Explanations for the curious:
1. The if statement lets me keep from spending CPU on the rtl_printf()
call without actually compiling out the code. For more detailed
debugging, you may want to use DebugLevel as a bitmask.
2. The rtl_printf() call uses the same params as the kernels
printk() routine, so I stuff the debug level into the string "<n>"
so klogd can do the right thing.
3. The "!" is used just so I can distinguish RTL code from other
kernel printk() calls.
4. In the #else case, I put a bogus statement (that the compiler
should optimize out) as leaving a blank semicolon has in the past
bitten me.
5. Note the use of (GCC specific!) varargs in a macro - beware
trying this on none-gcc platforms.
Regards,
Eric
--
Given a choice between grief and nothing, I'd choose grief.
-- William Faulkner
Eric Peterson WB6PYK (805)370-3046 mailto:[EMAIL PROTECTED]
http://www.troikanetworks.com
-- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
--
For more information on Real-Time Linux see:
http://www.rtlinux.org/rtlinux/
