Re: [PATCH] init/version.c: Define version_string only if CONFIG_KALLSYMS is not defined.

[Daniel Guilak <guilak@linux.vnet.ibm.com>]

On Mon, 2008-07-14 at 15:07 -0700, Randy Dunlap wrote:
> On Mon, 14 Jul 2008 14:20:38 -0700 Daniel Guilak wrote:
> 
> > On Mon, 2008-07-14 at 12:38 -0700, Randy Dunlap wrote:
> > > On Mon, 14 Jul 2008 12:04:46 -0700 Daniel Guilak wrote:
> > > 
> > > > int Version_* is only used with ksymoops, which is only needed (according to
> > > > README and Documentation/Changes) if CONFIG_KALLSYMS is NOT defined.  Therefore
> > > > this patch defines version_string only if CONFIG_KALLSYMS is not defined.
> > > > 
> > > > Signed-off-by: Daniel Guilak <daniel@danielguilak.com>
> > > > ---
> > > >  init/version.c |    2 ++
> > > >  1 files changed, 2 insertions(+), 0 deletions(-)
> > > > 
> > > > Depends on "[PATCH] init/version.c: Silenced sparse warning by declaring the version string."
> > > > 
> > > > diff --git a/init/version.c b/init/version.c
> > > > index 041fd82..52a8b98 100644
> > > > --- a/init/version.c
> > > > +++ b/init/version.c
> > > > @@ -13,11 +13,13 @@
> > > >  #include <linux/utsrelease.h>
> > > >  #include <linux/version.h>
> > > >  
> > > > +#ifndef CONFIG_KALLSYMS
> > > >  #define version(a) Version_ ## a
> > > >  #define version_string(a) version(a)
> > > >  
> > > >  extern int version_string(LINUX_VERSION_CODE);
> > > >  int version_string(LINUX_VERSION_CODE);
> > > > +#endif
> > > >  
> > > >  struct uts_namespace init_uts_ns = {
> > > >  	.kref = {
> > > > -- 
> > > 
> > > Does not apply cleanly to linux-2.6.26, linux-next-20080714, or Linus-git-head.
> > > 
> > 
> > Did you apply "[PATCH] init/version.c: Silenced sparse warning by
> > declaring the version string." beforehand? I just applied it on the most
> > recent tree and I didn't have any issues.
> 
> Clearly I didn't apply that patch.  I missed anything saying that this was
> patch 2/2 in a dependent series.

Sorry about that, I should have been a bit more clear with it. Next
time I'll make sure to state that in the subject line. I stated the
dependency right before the diff message.

> > > and why do we ned both extern int version_string()
> > > and int version_string() ?
> > Because sparse was complaining that it wasn't prototyped or static.
> > There isn't a header file to put it in since it only exists to be a
> > kernel.
> 
> so the extern makes it be published in the symbol table, whereas a static
> would keep it private.  Is that correct?
> 
> and the second line (int version_string(LINUX_VERSION_CODE)) actually
> defines it.

Yes, that's exactly the idea behind it.

--Daniel