Re: [Xen-devel] [PATCH v2 1/4] xen: indicate gcov in log messages

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 06.09.16 at 16:06, <wei.liu2@xxxxxxxxxx> wrote:
> On Tue, Sep 06, 2016 at 07:34:26AM -0600, Jan Beulich wrote:
>> >>> On 06.09.16 at 14:47, <andrew.cooper3@xxxxxxxxxx> wrote:
>> > On 06/09/16 13:41, Wei Liu wrote:
>> >> --- a/xen/drivers/char/console.c
>> >> +++ b/xen/drivers/char/console.c
>> >> @@ -735,10 +735,11 @@ void __init console_init_preirq(void)
>> >>      spin_lock(&console_lock);
>> >>      __putstr(xen_banner());
>> >>      spin_unlock(&console_lock);
>> >> -    printk("Xen version %d.%d%s (%s@%s) (%s) debug=%c %s\n",
>> >> +    printk("Xen version %d.%d%s (%s@%s) (%s) debug=%c gcov=%c %s\n",
>> > 
>> > Please instead use
>> > 
>> > printk("Xen version %d.%d%s (%s@%s) (%s) debug=%c"
>> > #ifdef CONFIG_GCOV
>> > " gcov=y"
>> > #endif
>> > " %s\n",
>> > 
>> > Enabling gcov will be extremely rare in the grand scheme of things, and
>> > this method causes zero overhead for the case where gcov is compiled
>> > out.  (Similarly elsewhere.)
>> 
>> How about instead of the kind of ugly #ifdef-ery above, instead of
>> 
>> +#ifdef CONFIG_GCOV
>> +#define gcov_build() 1
>> +#else
>> +#define gcov_build() 0
>> +#endif
>> 
>> we'd have
>> 
>> +#ifdef CONFIG_GCOV
>> +#define gcov_string "gcov=y"
>> +#else
>> +#define gcov_string ""
>> +#endif
>> 
>> (with the identifier name open for improvement), which can then be
>> embedded easily into any string literal?
>> 
> 
> Now this patch looks like

And this way it can have my ack.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel