Re: [Patch 1/1] Fix genksyms handling of DEFINE_PER_CPU(struct foo_s *, bar);

Re: [Patch 1/1] Fix genksyms handling of DEFINE_PER_CPU(struct foo_s *, bar);

[Sam Ravnborg]

On Wed, Dec 21, 2005 at 02:36:01PM -0600, Robin Holt wrote:
> This is a one-line change to parse.y.  It results in rebuilding the
> scripts/genksyms/*_shipped files.  Those are the next four patches.
Does this differ from the first patch-set you sent out?
I plan to apply these so they will be part of next round of kbuild
updates - which will take place during next merge window.

	Sam

[Patch 1/1] Fix genksyms handling of DEFINE_PER_CPU(struct foo_s *, bar);

[Robin Holt]

[-- Attachment #1: Type: text/plain, Size: 893 bytes --]

This is a one-line change to parse.y.  It results in rebuilding the
scripts/genksyms/*_shipped files.  Those are the next four patches.

When a .c file contains:
DEFINE_PER_CPU(struct foo_s *, bar);

the .cpp output looks like:
__attribute__((__section__(".data.percpu"))) __typeof__(struct foo_s *) per_cpu__bar;

With the existing parse.y, the value inside the paranthesis of
__typeof__() does not evaluate as a type_specifier and therefore
per_cpu__bar does not get assigned a type for genksyms which results in
the EXPORT_PER_CPU_SYMBOL() not generating a CRC value.

I have compared the Modules.symvers with and without this
patch and for ia64's defconfig, the only change is:
Before 0x00000000    per_cpu____sn_nodepda   vmlinux
After  0x9d3f3faa    per_cpu____sn_nodepda   vmlinux

per_cpu____sn_nodepda was the original source of my problems.

Signed-off-by: Robin Holt <holt@sgi.com>

[-- Attachment #2: genksyms_typeof_pointer.bz2 --]
[-- Type: application/x-bzip2, Size: 22664 bytes --]

Re: [Patch 1/1] Fix genksyms handling of DEFINE_PER_CPU(struct foo_s *, bar);

[Robin Holt]

On Wed, Dec 21, 2005 at 09:23:56PM +0100, Sam Ravnborg wrote:
> On Wed, Dec 21, 2005 at 02:36:01PM -0600, Robin Holt wrote:
> > This is a one-line change to parse.y.  It results in rebuilding the
> > scripts/genksyms/*_shipped files.  Those are the next four patches.
> Does this differ from the first patch-set you sent out?
> I plan to apply these so they will be part of next round of kbuild
> updates - which will take place during next merge window.

They are the same.  It took me four to finally get all the
parts out and on the lkml.

Thanks,
Robin

Re: [Patch 1/1] Fix genksyms handling of DEFINE_PER_CPU(struct foo_s *, bar);

[Sam Ravnborg]

On Wed, Dec 21, 2005 at 04:02:51PM -0600, Robin Holt wrote:
> On Wed, Dec 21, 2005 at 09:23:56PM +0100, Sam Ravnborg wrote:
> > On Wed, Dec 21, 2005 at 02:36:01PM -0600, Robin Holt wrote:
> > > This is a one-line change to parse.y.  It results in rebuilding the
> > > scripts/genksyms/*_shipped files.  Those are the next four patches.
> > Does this differ from the first patch-set you sent out?
> > I plan to apply these so they will be part of next round of kbuild
> > updates - which will take place during next merge window.
> 
> They are the same.  It took me four to finally get all the
> parts out and on the lkml.

Applied now. I created new _shipped files with the tools I have
installed.

bison (GNU Bison) 2.0
flex version 2.5.4
GNU gperf 3.0.1

This is (almost) a match of the same tools used for kconfig.

	Sam

Re: [Patch 1/1] Fix genksyms handling of DEFINE_PER_CPU(struct foo_s *, bar);

[Sam Ravnborg]

On Tue, Dec 27, 2005 at 06:38:49AM -0600, Robin Holt wrote:
> 
> Thanks.  When do you think that will be applied to Linus' tree?

When 2.6.16 opens up. I have not planned to push to get it into
2.6.15-rc since noone but you have noticed.
And also the change of bison/flex version deserve a bit wider testing.

	Sam

Re: [Patch 1/1] Fix genksyms handling of DEFINE_PER_CPU(struct foo_s *, bar);

[Robin Holt]

On Mon, Dec 26, 2005 at 10:55:42PM +0100, Sam Ravnborg wrote:
> On Wed, Dec 21, 2005 at 04:02:51PM -0600, Robin Holt wrote:
> > On Wed, Dec 21, 2005 at 09:23:56PM +0100, Sam Ravnborg wrote:
> > > On Wed, Dec 21, 2005 at 02:36:01PM -0600, Robin Holt wrote:
> > > > This is a one-line change to parse.y.  It results in rebuilding the
> > > > scripts/genksyms/*_shipped files.  Those are the next four patches.
> > > Does this differ from the first patch-set you sent out?
> > > I plan to apply these so they will be part of next round of kbuild
> > > updates - which will take place during next merge window.
> > 
> > They are the same.  It took me four to finally get all the
> > parts out and on the lkml.
> 
> Applied now. I created new _shipped files with the tools I have
> installed.
> 
> bison (GNU Bison) 2.0
> flex version 2.5.4
> GNU gperf 3.0.1
> 
> This is (almost) a match of the same tools used for kconfig.

Thanks.  When do you think that will be applied to Linus' tree?

Robin