[PATCH] tulip: de2104x, fix init. sections

[PATCH] tulip: de2104x, fix init. sections

[Randy.Dunlap]

tulip/de2104x: some __devinit functions were calling __init
  functions, made the latter __devinit also;

Error: ./drivers/net/tulip/de2104x.o .text refers to 000000000000176d R_X86_64_PC32     .init.text+0xfffffffffffffffc
Error: ./drivers/net/tulip/de2104x.o .text refers to 0000000000001798 R_X86_64_PC32     .init.text+0xfffffffffffffffc

Signed-off-by: Randy Dunlap <rddunlap@osdl.org>

diffstat:=
 drivers/net/tulip/de2104x.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)

diff -Naurp ./drivers/net/tulip/de2104x.c~de2104x_sections ./drivers/net/tulip/de2104x.c
--- ./drivers/net/tulip/de2104x.c~de2104x_sections	2005-02-27 12:54:05.830037144 -0800
+++ ./drivers/net/tulip/de2104x.c	2005-03-01 21:41:19.354417488 -0800
@@ -1691,7 +1691,7 @@ static struct ethtool_ops de_ethtool_ops
 	.get_regs		= de_get_regs,
 };
 
-static void __init de21040_get_mac_address (struct de_private *de)
+static void __devinit de21040_get_mac_address (struct de_private *de)
 {
 	unsigned i;
 
@@ -1709,7 +1709,7 @@ static void __init de21040_get_mac_addre
 	}
 }
 
-static void __init de21040_get_media_info(struct de_private *de)
+static void __devinit de21040_get_media_info(struct de_private *de)
 {
 	unsigned int i;
 
@@ -1736,7 +1736,7 @@ static void __init de21040_get_media_inf
 }
 
 /* Note: this routine returns extra data bits for size detection. */
-static unsigned __init tulip_read_eeprom(void __iomem *regs, int location, int addr_len)
+static unsigned __devinit tulip_read_eeprom(void __iomem *regs, int location, int addr_len)
 {
 	int i;
 	unsigned retval = 0;
@@ -1771,7 +1771,7 @@ static unsigned __init tulip_read_eeprom
 	return retval;
 }
 
-static void __init de21041_get_srom_info (struct de_private *de)
+static void __devinit de21041_get_srom_info (struct de_private *de)
 {
 	unsigned i, sa_offset = 0, ofs;
 	u8 ee_data[DE_EEPROM_SIZE + 6] = {};

---

Re: [PATCH] tulip: de2104x, fix init. sections

[Jeff Garzik]

Randy.Dunlap wrote:
> tulip/de2104x: some __devinit functions were calling __init
>   functions, made the latter __devinit also;

Should go the other way:  all the stuff in de2104x should be __init, not 
__devinit.

Nobody ever hotplugs a de2104x.

	Jeff

Re: [PATCH] tulip: de2104x, fix init. sections

[Linus Torvalds]

On Wed, 2 Mar 2005, Jeff Garzik wrote:
> 
> Nobody ever hotplugs a de2104x.

Are you sure? I'm pretty certain some of the Cardbus cards are based on
that chipset. In fact, I'm pretty sure that's exactly that I used to have
(and actually used when I did the "CardBus as a real PCI bridge" stuff).

CardBus tulip network cards may be less interesting these days, but I'd be
surprised if they are all gone. The tulip chip used to be the most common
one.

(And don't be fooled by the Xircom config stuff - those have a special
driver partly because Xircom did something wrong, and probably partly
because the old PCMCIA layer just was too damn confused, and thought that
CardBus devices needed something else than a PCI driver. Most tulip
cardbus cards used just the bog-standard tulip driver after my Cardbus
rewrite, I do believe).

		Linus

Re: [PATCH] tulip: de2104x, fix init. sections

[Jeff Garzik]

On Tue, Mar 01, 2005 at 11:02:16PM -0800, Linus Torvalds wrote:
> 
> 
> On Wed, 2 Mar 2005, Jeff Garzik wrote:
> > 
> > Nobody ever hotplugs a de2104x.
> 
> Are you sure? I'm pretty certain some of the Cardbus cards are based on
> that chipset. In fact, I'm pretty sure that's exactly that I used to have
> (and actually used when I did the "CardBus as a real PCI bridge" stuff).

Nope -- You are thinking of the 2114x stuff.  You even bought me a
cardbus card of my own to get things going on that, way back when.  :)

2104x is truly ancient.

	Jeff

Re: [PATCH] tulip: de2104x, fix init. sections

[Linus Torvalds]

On Wed, 2 Mar 2005, Jeff Garzik wrote:
> 
> Nope -- You are thinking of the 2114x stuff.  You even bought me a
> cardbus card of my own to get things going on that, way back when.  :)
> 
> 2104x is truly ancient.

Ahh, ok. Goodie.

		Linus

Re: [PATCH] tulip: de2104x, fix init. sections

[Christoph Hellwig]

On Wed, Mar 02, 2005 at 01:09:54AM -0500, Jeff Garzik wrote:
> Randy.Dunlap wrote:
> >tulip/de2104x: some __devinit functions were calling __init
> >  functions, made the latter __devinit also;
> 
> Should go the other way:  all the stuff in de2104x should be __init, not 
> __devinit.
> 
> Nobody ever hotplugs a de2104x.

How comes that you want to dictate what cards I put into my hotplug
PCI slots?

Re: [PATCH] tulip: de2104x, fix init. sections

[Jeff Garzik]

Christoph Hellwig wrote:
> On Wed, Mar 02, 2005 at 01:09:54AM -0500, Jeff Garzik wrote:
> 
>>Randy.Dunlap wrote:
>>
>>>tulip/de2104x: some __devinit functions were calling __init
>>> functions, made the latter __devinit also;
>>
>>Should go the other way:  all the stuff in de2104x should be __init, not 
>>__devinit.
>>
>>Nobody ever hotplugs a de2104x.
> 
> 
> How comes that you want to dictate what cards I put into my hotplug
> PCI slots?

When someone proves this is possible, I will accept the code change.  I 
doubt this will ever happen, for two reasons:

1) Hotplug PCI slots are highly likely to choke on the 2104x's very 
early and slightly weird PCI hardware implementation.

2) It's an ancient card and nobody will ever bother trying to hotplug it.

	Jeff

Re: [PATCH] tulip: de2104x, fix init. sections

[Christoph Hellwig]

On Wed, Mar 02, 2005 at 02:30:17AM -0500, Jeff Garzik wrote:
> When someone proves this is possible, I will accept the code change.  I 
> doubt this will ever happen, for two reasons:
> 
> 1) Hotplug PCI slots are highly likely to choke on the 2104x's very 
> early and slightly weird PCI hardware implementation.
> 
> 2) It's an ancient card and nobody will ever bother trying to hotplug it.

No, it's a matter of correctness.  The pci core can call ->probe all the
time and every driver must be prepared.  Even if no one tries it physically
it's easily done with Greg's fake hotplug driver.

Re: [PATCH] tulip: de2104x, fix init. sections

[Jeff Garzik]

Christoph Hellwig wrote:
> On Wed, Mar 02, 2005 at 02:30:17AM -0500, Jeff Garzik wrote:
> 
>>When someone proves this is possible, I will accept the code change.  I 
>>doubt this will ever happen, for two reasons:
>>
>>1) Hotplug PCI slots are highly likely to choke on the 2104x's very 
>>early and slightly weird PCI hardware implementation.
>>
>>2) It's an ancient card and nobody will ever bother trying to hotplug it.
> 
> 
> No, it's a matter of correctness.  The pci core can call ->probe all the
> time and every driver must be prepared.  Even if no one tries it physically
> it's easily done with Greg's fake hotplug driver.

The change can be made when it's actually useful to someone.  Making 
janitors and Christoph happy is not a good enough reason in my book.

	Jeff

Re: [PATCH] tulip: de2104x, fix init. sections

[Jeff Garzik]

For what it's worth, I'm not just being obstinate :)  I build this 
driver in hotplug kernels, and I know it will never be hotplugged.  In 
other words, I like it like that.

	Jeff