Re: [PATCH 3/3] ftrace/MIPS: Enable C Version of recordmcount

["Maciej W. Rozycki" <macro@linux-mips.org>]

On Wed, 27 Oct 2010, Steven Rostedt wrote:

> On Wed, 2010-10-27 at 18:59 +0800, Wu Zhangjin wrote:
> > From: Wu Zhangjin <wuzhangjin@gmail.com>
> > 
> > Selects HAVE_C_RECORDMCOUNT to use the C version of the recordmcount
> > intead of the old Perl Version of recordmcount.
> > 
> > Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
> 
> I'd like to get an Acked-by from Ralf and Maciej on this.

Acked-by: Maciej W. Rozycki <macro@linux-mips.org>

 I have looked through it and spotted nothing obviously wrong, but I can't 
afford any further testing, especially at run time, sorry.

 One point to note -- it seems to me the code currently assumes a 32-bit 
model, i.e. the use of the -msym32 GCC option suitable for a 64-bit kernel 
loaded to a CKSEG0 address (cf. KBUILD_SYM32 in arch/mips/Makefile), tools 
support permitting.  That means it does not (correctly) support kernels 
loaded to an XPHYS address as required for some platforms (or otherwise 
chosen for testing such a configuration; modulo some processor errata and 
bootloader limitations, it is generally OK to run the kernel from XPHYS on 
64-bit chips even if the entire RAM fits into CKSEG0).

 For the avoidance of doubt -- I'm just mentioning it to emphasise a 
possible future direction for improvement of this code -- not an objection 
against this submission, that is certainly a good foundation for future 
development.

 Thanks to everybody involved.

  Maciej