Re: [Ltib] Forcing DEBUG for pr_debug in kernel driver through LTIB

[Stuart Hughes]

Hi Franz,

As far as ltib goes, you can do: ./ltib -p kernel -c and that will 
unpack the kernel sources and drop you to the kernel config screen.  If 
the kernel config system has a configuration pooint to set KERN_DEBUG 
then you can set it there, save and exit and the kernel will build.  If 
there is no config point for it, you'll have to edit the file manually.

Regards, Stuart

Franz Trierweiler wrote:
> Hello everybody
>
> Freescale has just issued a new PDK (PDK 1.6) for the iMX25 3-stack 
> evalboard. I have hardwired a SIM interface to my evalboard so that I 
> can scope all electrical signals coming from that cell. The fact is that 
> I am trying to debug the mxc_sim.c source file for this SIM interface. 
> mxc_sim is a SIM driver. In order to achieve this, I would like to 
> enable pr_debug (...) output to my console when this driver is running 
> and watch all debug messages displayed.
>
> The kernel.h shows the pr_debug macro like this:
>
> /* If you are writing a driver, please use dev_dbg instead */
> #if defined(CONFIG_DYNAMIC_PRINTK_DEBUG)
> #define pr_debug(fmt, ...) do { \
>    dynamic_pr_debug(pr_fmt(fmt), ##__VA_ARGS__); \
>    } while (0)
> #elif defined(DEBUG)
> #define pr_debug(fmt, ...) \
>    printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__)
> #else
> #define pr_debug(fmt, ...) \
>    ({ if (0) printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__); 0; })
> #endif
>
> There seems that if I #define DEBUG somewhere then I will have the 
> messages displayed (or logged) somewhere.
>
> Is there a way of doing this properly in the LTIB configuration when 
> recompiling the kernel ?
>
> Or should I simply #define DEBUG in kernel.h? (This really does not 
> sound the right way).
>
> Or should I simply put DEBUG=ON (or something like this) in the LTIB 
> perl script? (sounds also very dirty).
>
> I really would like to make this in a clean way.
>
> Regards,
> Franz