building ppc_htab?

building ppc_htab?

[Kumar Gala]

Is there any reason not to make the following change so we build 
ppc_htab only on PPC STD MMU?  This would also require fixing up 
fs/proc/proc_misc.c.

--- 1.52/arch/ppc/kernel/Makefile       2005-03-13 17:29:44 -06:00
+++ edited/Makefile     2005-03-16 00:44:17 -06:00
@@ -2,7 +2,7 @@
  # Makefile for the linux kernel.
  #

-extra-$(CONFIG_PPC_STD_MMU)    := head.o
+extra-$(CONFIG_PPC_STD_MMU)    := head.o ppc_htab.o
  extra-$(CONFIG_40x)            := head_4xx.o
  extra-$(CONFIG_44x)            := head_44x.o
  extra-$(CONFIG_E500)           := head_e500.o
@@ -14,7 +14,7 @@
  obj-y                          := entry.o traps.o irq.o idle.o time.o 
misc.o \
                                         process.o signal.o ptrace.o 
align.o \
                                         semaphore.o syscalls.o setup.o \
-                                       cputable.o ppc_htab.o perfmon.o
+                                       cputable.o perfmon.o
  obj-$(CONFIG_6xx)              += l2cr.o cpu_setup_6xx.o
  obj-$(CONFIG_SOFTWARE_SUSPEND) += swsusp.o
  obj-$(CONFIG_POWER4)           += cpu_setup_power4.o

- kumar

Re: building ppc_htab?

[Dan Malek]

On Mar 16, 2005, at 1:57 AM, Kumar Gala wrote:

> Is there any reason not to make the following change so we build 
> ppc_htab only on PPC STD MMU?  This would also require fixing up 
> fs/proc/proc_misc.c.

Does it actually build properly this way?  There used to be 
functions/variables
defined there, needed for linking resolution at least.  There were also 
some
generally called functions, that usually didn't do anything unless 
there was
a hash table used.  If that is fixed up, then this is fine :-)

Thanks.


	-- Dan

Re: building ppc_htab?

[Tom Rini]

On Wed, Mar 16, 2005 at 12:57:24AM -0600, Kumar Gala wrote:

> Is there any reason not to make the following change so we build 
> ppc_htab only on PPC STD MMU?  This would also require fixing up 
> fs/proc/proc_misc.c.

I could have sworn, but maybe my memory sucks, that we no longer messed
w/ fs/proc/proc_misc.c for any of the PPC-specific things (since in 2.6
you can do all of that proc magic in your 'driver').  So this patch
would make a great deal of sense to do, and if it lacks what I'm talking
about, we should do that part as well :)

-- 
Tom Rini
http://gate.crashing.org/~trini/

Re: building ppc_htab?

[Kumar Gala]

On Mar 16, 2005, at 4:11 PM, Tom Rini wrote:

> On Wed, Mar 16, 2005 at 12:57:24AM -0600, Kumar Gala wrote:
>
> > Is there any reason not to make the following change so we build
> > ppc_htab only on PPC STD MMU?=A0 This would also require fixing up
> > fs/proc/proc_misc.c.
>
> I could have sworn, but maybe my memory sucks, that we no longer =
messed
> w/ fs/proc/proc_misc.c for any of the PPC-specific things (since in =
2.6
>  you can do all of that proc magic in your 'driver').=A0 So this patch
>  would make a great deal of sense to do, and if it lacks what I'm=20
> talking
>  about, we should do that part as well :)

I'm not exactly clear on what you are talking about.  There is this=20
snippet in fs/proc/proc_misc.c:

#ifdef CONFIG_PPC32
         {
                 extern struct file_operations ppc_htab_operations;
                 entry =3D create_proc_entry("ppc_htab", =
S_IRUGO|S_IWUSR,=20
NULL);
                 if (entry)
                         entry->proc_fops =3D &ppc_htab_operations;
         }
#endif

Are you suggesting this should live elsewhere?  I'm not that familiar=20
with proc code.

- kumar

Re: building ppc_htab?

[Tom Rini]

On Thu, Mar 17, 2005 at 11:34:14AM -0600, Kumar Gala wrote:
> 
> On Mar 16, 2005, at 4:11 PM, Tom Rini wrote:
> 
> >On Wed, Mar 16, 2005 at 12:57:24AM -0600, Kumar Gala wrote:
> >
> >> Is there any reason not to make the following change so we build
> >> ppc_htab only on PPC STD MMU?? This would also require fixing up
> >> fs/proc/proc_misc.c.
> >
> >I could have sworn, but maybe my memory sucks, that we no longer messed
> >w/ fs/proc/proc_misc.c for any of the PPC-specific things (since in 2.6
> > you can do all of that proc magic in your 'driver').? So this patch
> > would make a great deal of sense to do, and if it lacks what I'm 
> >talking
> > about, we should do that part as well :)
> 
> I'm not exactly clear on what you are talking about.  There is this 
> snippet in fs/proc/proc_misc.c:

Crap :(

> #ifdef CONFIG_PPC32
>         {
>                 extern struct file_operations ppc_htab_operations;
>                 entry = create_proc_entry("ppc_htab", S_IRUGO|S_IWUSR, 
> NULL);
>                 if (entry)
>                         entry->proc_fops = &ppc_htab_operations;
>         }
> #endif
> 
> Are you suggesting this should live elsewhere?  I'm not that familiar 
> with proc code.

Yes, it should live in ppc_htab.c.  See the CONFIG_SYSCTL stuff already
in ppc_htab.c for l2crvec.

-- 
Tom Rini
http://gate.crashing.org/~trini/

Re: building ppc_htab?

[Tom Rini]

On Thu, Mar 17, 2005 at 10:41:37AM -0700, Tom Rini wrote:
> On Thu, Mar 17, 2005 at 11:34:14AM -0600, Kumar Gala wrote:
> > 
> > On Mar 16, 2005, at 4:11 PM, Tom Rini wrote:
> > 
> > >On Wed, Mar 16, 2005 at 12:57:24AM -0600, Kumar Gala wrote:
> > >
> > >> Is there any reason not to make the following change so we build
> > >> ppc_htab only on PPC STD MMU?? This would also require fixing up
> > >> fs/proc/proc_misc.c.
> > >
> > >I could have sworn, but maybe my memory sucks, that we no longer messed
> > >w/ fs/proc/proc_misc.c for any of the PPC-specific things (since in 2.6
> > > you can do all of that proc magic in your 'driver').? So this patch
> > > would make a great deal of sense to do, and if it lacks what I'm 
> > >talking
> > > about, we should do that part as well :)
> > 
> > I'm not exactly clear on what you are talking about.  There is this 
> > snippet in fs/proc/proc_misc.c:
> 
> Crap :(
> 
> > #ifdef CONFIG_PPC32
> >         {
> >                 extern struct file_operations ppc_htab_operations;
> >                 entry = create_proc_entry("ppc_htab", S_IRUGO|S_IWUSR, 
> > NULL);
> >                 if (entry)
> >                         entry->proc_fops = &ppc_htab_operations;
> >         }
> > #endif
> > 
> > Are you suggesting this should live elsewhere?  I'm not that familiar 
> > with proc code.
> 
> Yes, it should live in ppc_htab.c.  See the CONFIG_SYSCTL stuff already
> in ppc_htab.c for l2crvec.

Der, that's not the same thing!  I still swear there's a way to do this
w/o messing w/ fs/proc/proc_misc.c, but I don't recall a good example.

-- 
Tom Rini
http://gate.crashing.org/~trini/

Re: building ppc_htab?

[Kumar Gala]

> > Crap :(
>  >
> > > #ifdef CONFIG_PPC32
> > >=A0=A0=A0=A0=A0=A0=A0=A0 {
>  > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 extern struct =
file_operations ppc_htab_operations;
> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 entry =3D =
create_proc_entry("ppc_htab",=20
> S_IRUGO|S_IWUSR,
>  > > NULL);
>  > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 if (entry)
>  > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=
=A0 entry->proc_fops =3D &ppc_htab_operations;
> > >=A0=A0=A0=A0=A0=A0=A0=A0 }
>  > > #endif
>  > >
> > > Are you suggesting this should live elsewhere?=A0 I'm not that=20
> familiar
> > > with proc code.
>  >
> > Yes, it should live in ppc_htab.c.=A0 See the CONFIG_SYSCTL stuff=20
> already
>  > in ppc_htab.c for l2crvec.
>
> Der, that's not the same thing!=A0 I still swear there's a way to do =
this
>  w/o messing w/ fs/proc/proc_misc.c, but I don't recall a good =
example.

Will, for now I will add a conditional on CONFIG_PPC_STD_MMU.  We can=20
move this out once we figure out how to do it outside of=20
fs/proc/proc_misc.c.

- kumar