[2.6 patch] move EXPORT_SYMBOL's away from sound/pci/emu10k1/emu10k1_main.c

[2.6 patch] move EXPORT_SYMBOL's away from sound/pci/emu10k1/emu10k1_main.c

[Adrian Bunk]

This patch moves the EXPORT_SYMBOL's from 
sound/pci/emu10k1/emu10k1_main.c to the files with the actual functions.

Signed-off-by: Adrian Bunk <bunk@stusta.de>

---

 sound/pci/emu10k1/emu10k1_main.c |   12 ------------
 sound/pci/emu10k1/io.c           |    2 ++
 sound/pci/emu10k1/memory.c       |    6 ++++++
 sound/pci/emu10k1/voice.c        |    3 +++
 4 files changed, 11 insertions(+), 12 deletions(-)

--- linux-2.6.17-rc1-mm1-full/sound/pci/emu10k1/emu10k1_main.c.old	2006-04-07 01:08:22.000000000 +0200
+++ linux-2.6.17-rc1-mm1-full/sound/pci/emu10k1/emu10k1_main.c	2006-04-07 01:08:32.000000000 +0200
@@ -1419,15 +1419,3 @@
 }
 #endif
 
-/* memory.c */
-EXPORT_SYMBOL(snd_emu10k1_synth_alloc);
-EXPORT_SYMBOL(snd_emu10k1_synth_free);
-EXPORT_SYMBOL(snd_emu10k1_synth_bzero);
-EXPORT_SYMBOL(snd_emu10k1_synth_copy_from_user);
-EXPORT_SYMBOL(snd_emu10k1_memblk_map);
-/* voice.c */
-EXPORT_SYMBOL(snd_emu10k1_voice_alloc);
-EXPORT_SYMBOL(snd_emu10k1_voice_free);
-/* io.c */
-EXPORT_SYMBOL(snd_emu10k1_ptr_read);
-EXPORT_SYMBOL(snd_emu10k1_ptr_write);
--- linux-2.6.17-rc1-mm1-full/sound/pci/emu10k1/memory.c.old	2006-04-07 01:09:12.000000000 +0200
+++ linux-2.6.17-rc1-mm1-full/sound/pci/emu10k1/memory.c	2006-04-07 01:10:01.000000000 +0200
@@ -286,6 +286,7 @@
 	spin_unlock_irqrestore(&emu->memblk_lock, flags);
 	return err;
 }
+EXPORT_SYMBOL(snd_emu10k1_memblk_map);
 
 /*
  * page allocation for DMA
@@ -386,6 +387,7 @@
 	mutex_unlock(&hdr->block_mutex);
 	return (struct snd_util_memblk *)blk;
 }
+EXPORT_SYMBOL(snd_emu10k1_synth_alloc);
 
 
 /*
@@ -408,6 +410,7 @@
 	mutex_unlock(&hdr->block_mutex);
 	return 0;
 }
+EXPORT_SYMBOL(snd_emu10k1_synth_free);
 
 
 /* check new allocation range */
@@ -539,6 +542,7 @@
 	} while (offset < end_offset);
 	return 0;
 }
+EXPORT_SYMBOL(snd_emu10k1_synth_bzero);
 
 /*
  * copy_from_user(blk + offset, data, size)
@@ -568,3 +572,5 @@
 	} while (offset < end_offset);
 	return 0;
 }
+EXPORT_SYMBOL(snd_emu10k1_synth_copy_from_user);
+
--- linux-2.6.17-rc1-mm1-full/sound/pci/emu10k1/voice.c.old	2006-04-07 01:10:14.000000000 +0200
+++ linux-2.6.17-rc1-mm1-full/sound/pci/emu10k1/voice.c	2006-04-07 01:10:35.000000000 +0200
@@ -138,6 +138,7 @@
 
 	return result;
 }
+EXPORT_SYMBOL(snd_emu10k1_voice_alloc);
 
 int snd_emu10k1_voice_free(struct snd_emu10k1 *emu,
 			   struct snd_emu10k1_voice *pvoice)
@@ -153,3 +154,5 @@
 	spin_unlock_irqrestore(&emu->voice_lock, flags);
 	return 0;
 }
+EXPORT_SYMBOL(snd_emu10k1_voice_free);
+
--- linux-2.6.17-rc1-mm1-full/sound/pci/emu10k1/io.c.old	2006-04-07 01:10:46.000000000 +0200
+++ linux-2.6.17-rc1-mm1-full/sound/pci/emu10k1/io.c	2006-04-07 02:19:43.000000000 +0200
@@ -61,6 +61,7 @@
 		return val;
 	}
 }
+EXPORT_SYMBOL(snd_emu10k1_ptr_read);
 
 void snd_emu10k1_ptr_write(struct snd_emu10k1 *emu, unsigned int reg, unsigned int chn, unsigned int data)
 {
@@ -91,6 +92,7 @@
 		spin_unlock_irqrestore(&emu->emu_lock, flags);
 	}
 }
+EXPORT_SYMBOL(snd_emu10k1_ptr_write);
 
 unsigned int snd_emu10k1_ptr20_read(struct snd_emu10k1 * emu, 
 					  unsigned int reg, 

Re: [2.6 patch] move EXPORT_SYMBOL's away from sound/pci/emu10k1/emu10k1_main.c

[Takashi Iwai]

At Fri, 7 Apr 2006 02:31:05 +0200,
Adrian Bunk wrote:
> 
> This patch moves the EXPORT_SYMBOL's from 
> sound/pci/emu10k1/emu10k1_main.c to the files with the actual functions.

What is the merit of this movement?


Takashi

Re: [2.6 patch] move EXPORT_SYMBOL's away from sound/pci/emu10k1/emu10k1_main.c

[Sam Ravnborg]

On Fri, Apr 07, 2006 at 02:34:56PM +0200, Takashi Iwai wrote:
> At Fri, 7 Apr 2006 02:31:05 +0200,
> Adrian Bunk wrote:
> > 
> > This patch moves the EXPORT_SYMBOL's from 
> > sound/pci/emu10k1/emu10k1_main.c to the files with the actual functions.
> 
> What is the merit of this movement?

1) Documentation - it is obvious that the function/data is exported so
be a bit mroe careful when introducing changes
2) Style. In 2.6 the preferred style is to put the EXPORT_SYMBOL on the
line following the closing } of the exported function.
3) Keep changes local. If one removes a previously exported symbol less
files needs to be touched.

	Sam

Re: [2.6 patch] move EXPORT_SYMBOL's away from sound/pci/emu10k1/emu10k1_main.c

[Takashi Iwai]

At Fri, 7 Apr 2006 20:49:09 +0200,
Sam Ravnborg wrote:
> 
> On Fri, Apr 07, 2006 at 02:34:56PM +0200, Takashi Iwai wrote:
> > At Fri, 7 Apr 2006 02:31:05 +0200,
> > Adrian Bunk wrote:
> > > 
> > > This patch moves the EXPORT_SYMBOL's from 
> > > sound/pci/emu10k1/emu10k1_main.c to the files with the actual functions.
> > 
> > What is the merit of this movement?
> 
> 1) Documentation - it is obvious that the function/data is exported so
> be a bit mroe careful when introducing changes
> 2) Style. In 2.6 the preferred style is to put the EXPORT_SYMBOL on the
> line following the closing } of the exported function.
> 3) Keep changes local. If one removes a previously exported symbol less
> files needs to be touched.

I know the above for the new codes, yes.  But my question is wheter do
we get a good enough benifit by changing the existing code.

I'm not against such an action but just wornder whether it's really
needed.  If yes, we should do it over the whole tree.


Takashi

Re: [2.6 patch] move EXPORT_SYMBOL's away from sound/pci/emu10k1/emu10k1_main.c

[Arjan van de Ven]

> I know the above for the new codes, yes.  But my question is wheter do
> we get a good enough benifit by changing the existing code.
> 
> I'm not against such an action but just wornder whether it's really
> needed.  If yes, we should do it over the whole tree.
> 

that's what Adrian is doing, one subsystem at a time ;)
it was just your turn...

Re: [2.6 patch] move EXPORT_SYMBOL's away from sound/pci/emu10k1/emu10k1_main.c

[Takashi Iwai]

At Mon, 10 Apr 2006 18:21:54 +0200,
Arjan van de Ven wrote:
> 
> 
> > I know the above for the new codes, yes.  But my question is wheter do
> > we get a good enough benifit by changing the existing code.
> > 
> > I'm not against such an action but just wornder whether it's really
> > needed.  If yes, we should do it over the whole tree.
> > 
> 
> that's what Adrian is doing, one subsystem at a time ;)
> it was just your turn...

OK, then I'll check over the subtree, too.


thanks,

Takashi

Re: [2.6 patch] move EXPORT_SYMBOL's away from sound/pci/emu10k1/emu10k1_main.c

[Sam Ravnborg]

On Mon, Apr 10, 2006 at 11:55:22AM +0200, Takashi Iwai wrote:
 > > > This patch moves the EXPORT_SYMBOL's from 
> > > > sound/pci/emu10k1/emu10k1_main.c to the files with the actual functions.
> > > 
> > > What is the merit of this movement?
> > 
> > 1) Documentation - it is obvious that the function/data is exported so
> > be a bit mroe careful when introducing changes
> > 2) Style. In 2.6 the preferred style is to put the EXPORT_SYMBOL on the
> > line following the closing } of the exported function.
> > 3) Keep changes local. If one removes a previously exported symbol less
> > files needs to be touched.
> 
> I know the above for the new codes, yes.  But my question is wheter do
> we get a good enough benifit by changing the existing code.
> 
> I'm not against such an action but just wornder whether it's really
> needed.  If yes, we should do it over the whole tree.
That has been done to a lot of existing code also.
In the 2.4 days it made sense to have them grouped due to build system
limitations - that is no longer the case with 2.6 so it is preferred to
do it correct.

	Sam