Re: linux-next: Tree for March 11 (tracing)

[Randy Dunlap <randy.dunlap@oracle.com>]

Ingo Molnar wrote:
> * Randy Dunlap <randy.dunlap@oracle.com> wrote:
> 
>> Ingo Molnar wrote:
>>> * Randy Dunlap <randy.dunlap@oracle.com> wrote:
>>>
>>>> On Thu, 12 Mar 2009 12:26:21 -0400 Steven Rostedt wrote:
>>>>
>>>>> On Thu, 2009-03-12 at 09:17 -0700, Randy Dunlap wrote:
>>>>>> [adding cc:s]
>>>>>>
>>>>>> [same report for March 12]
>>>>>>
>>>>>> Randy Dunlap wrote:
>>>>>>> Stephen Rothwell wrote:
>>>>>>>> Hi all,
>>>>>>>>
>>>>>>>> Changes since 20090310:
>>>>>>> Building on i386 generates a ton of printk format warnings:
>>>>>>>
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 5 has type 'unsigned int'
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 6 has type 'unsigned int'
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 9 has type 'unsigned int'
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 10 has type 'unsigned int'
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 13 has type 'unsigned int'
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 14 has type 'unsigned int'
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 17 has type 'unsigned int'
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 18 has type 'unsigned int'
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 21 has type 'unsigned int'
>>>>>>> kernel/trace/trace_events.c:470: warning: format '%lu' expects type 'long unsigned int', but argument 22 has type 'unsigned int'
>>>>>>>
>>>>> I believe this is corrected in Ingo's tip tree. I changed %lu to %zu to
>>>>> handle the "sizeof()" case. The fix was suggested by Andrew Morton.
>>>> This build warning is still around (20090318).
>>>> Is the fix not in some branch that is imported into linux-next or what?
>>> be patient.
>>>
>>> 	Ingo
>> I think that 7 days is being patient for a simple build fix.
> 
> s/build fix/harmless build warning fix

180+ lines of noise in a build log.

> If you are interested in having a resolution you can git-merge the 
> latest development tree yourself and you can get rid of that 
> warning.
> 
> Of course that way you'd expose yourself to even fresher code, 
> potentially with much more serious breakages.
> 
> It's a balance of freshness versus stability, and that balance is 
> kept by maintainers.
> 
> If you want the latest development code - go engage with the 
> development trees directly.
> 
> If you want something that is relatively new (i.e. 1-2 weeks fresh) 
> but works on the range of systems we test, use what you get in 
> linux-next.
> 
> It's your choice which one you pick.
> 
> But you cannot have both.
> 
> If you genuinely think you can have it both, by all means i 
> encourage you to try it - it's all open source so you can run your 
> own tree. Just please dont feel entitled to demand it from others.

Thanks for the explanation.  That's what I tried to ask for
to begin with.  I guess that I have a language problem.

-- 
~Randy