[PATCH] 8390: Fix ne.c build

[PATCH] 8390: Fix ne.c build

[Atsushi Nemoto]

Since commit 055e5110ae0c0c1176a75b78d789294f2ff2f7af ("8390: Split
8390 support into a pausing and a non pausing driver core"), ne.c
cause this build failure:

drivers/built-in.o: In function `ne_probe1':
ne.c:(.init.text+0x2464): undefined reference to `NS8390_init'
drivers/built-in.o: In function `ne_block_output':
ne.c:(.text.ne_block_output+0x1b0): undefined reference to `NS8390_init'

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
---
diff --git a/drivers/net/ne.c b/drivers/net/ne.c
index 1412697..3df231d 100644
--- a/drivers/net/ne.c
+++ b/drivers/net/ne.c
@@ -536,7 +536,7 @@ static int __init ne_probe1(struct net_device *dev, unsigned long ioaddr)
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	dev->poll_controller = eip_poll;
 #endif
-	NS8390_init(dev, 0);
+	NS8390p_init(dev, 0);
 
 	ret = register_netdev(dev);
 	if (ret)
@@ -794,7 +794,7 @@ retry:
 		if (time_after(jiffies, dma_start + 2*HZ/100)) {		/* 20ms */
 			printk(KERN_WARNING "%s: timeout waiting for Tx RDC.\n", dev->name);
 			ne_reset_8390(dev);
-			NS8390_init(dev,1);
+			NS8390p_init(dev,1);
 			break;
 		}
 
@@ -855,7 +855,7 @@ static int ne_drv_resume(struct platform_device *pdev)
 
 	if (netif_running(dev)) {
 		ne_reset_8390(dev);
-		NS8390_init(dev, 1);
+		NS8390p_init(dev, 1);
 		netif_device_attach(dev);
 	}
 	return 0;

[build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Ingo Molnar]

* Atsushi Nemoto <anemo@mba.ocn.ne.jp> wrote:

> Since commit 055e5110ae0c0c1176a75b78d789294f2ff2f7af ("8390: Split
> 8390 support into a pausing and a non pausing driver core"), ne.c
> cause this build failure:
> 
> drivers/built-in.o: In function `ne_probe1':
> ne.c:(.init.text+0x2464): undefined reference to `NS8390_init'
> drivers/built-in.o: In function `ne_block_output':
> ne.c:(.text.ne_block_output+0x1b0): undefined reference to `NS8390_init'

i just triggered a similar build failure, but in a different 
ne2000-derived file:

 drivers/built-in.o: In function `hpp_close':
 hp-plus.c:(.text+0xd7c1a): undefined reference to `ei_close'
 drivers/built-in.o: In function `hpp_open':
 hp-plus.c:(.text+0xd7c3c): undefined reference to `ei_interrupt'
 hp-plus.c:(.text+0xd7ca9): undefined reference to `ei_open'
 drivers/built-in.o: In function `hpp_probe1':
 hp-plus.c:(.init.text+0xab54): undefined reference to `ei_poll'
 hp-plus.c:(.init.text+0xabed): undefined reference to `NS8390_init'
 drivers/built-in.o: In function `hp_plus_probe':
 (.init.text+0xac4c): undefined reference to `__alloc_ei_netdev'
 make: *** [.tmp_vmlinux1] Error 1

with this config:

  http://redhat.com/~mingo/misc/config-Mon_Jul_21_20_21_08_CEST_2008.bad

	Ingo

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Alan Cox]

On Mon, 21 Jul 2008 21:15:19 +0200
Ingo Molnar <mingo@elte.hu> wrote:

> 
> * Atsushi Nemoto <anemo@mba.ocn.ne.jp> wrote:
> 
> > Since commit 055e5110ae0c0c1176a75b78d789294f2ff2f7af ("8390: Split
> > 8390 support into a pausing and a non pausing driver core"), ne.c
> > cause this build failure:
> > 
> > drivers/built-in.o: In function `ne_probe1':
> > ne.c:(.init.text+0x2464): undefined reference to `NS8390_init'
> > drivers/built-in.o: In function `ne_block_output':
> > ne.c:(.text.ne_block_output+0x1b0): undefined reference to `NS8390_init'
> 
> i just triggered a similar build failure, but in a different 
> ne2000-derived file:

I'll take a look. Various odd things happened with the Makefile for this
patch and the follow up fixes. If the followup didn't get applied then
that might explain the breakage.

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Ingo Molnar]

* Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> On Mon, 21 Jul 2008 21:15:19 +0200
> Ingo Molnar <mingo@elte.hu> wrote:
> 
> > 
> > * Atsushi Nemoto <anemo@mba.ocn.ne.jp> wrote:
> > 
> > > Since commit 055e5110ae0c0c1176a75b78d789294f2ff2f7af ("8390: Split
> > > 8390 support into a pausing and a non pausing driver core"), ne.c
> > > cause this build failure:
> > > 
> > > drivers/built-in.o: In function `ne_probe1':
> > > ne.c:(.init.text+0x2464): undefined reference to `NS8390_init'
> > > drivers/built-in.o: In function `ne_block_output':
> > > ne.c:(.text.ne_block_output+0x1b0): undefined reference to `NS8390_init'
> > 
> > i just triggered a similar build failure, but in a different 
> > ne2000-derived file:
> 
> I'll take a look. Various odd things happened with the Makefile for 
> this patch and the follow up fixes. If the followup didn't get applied 
> then that might explain the breakage.

ok. I also got:

 drivers/built-in.o: In function `hp_probe1':
 hp.c:(.init.text+0xa280): undefined reference to `NS8390_init'

so i suspect most of the 8390p.o library using ne2000 drivers are 
affected?

i think instead of a manual review i think this is better to fix all the 
problems:

cd drivers/net/; sed -i 's/NS8390_/NS8390p_/g' \
  $(grep -l NS8390_ $(grep 8390p.o Makefile | cut -d' ' -f3 | \
     sed 's/.o$/.c/g'))

find the full tested patch below. It solves the build failures i 
triggered so far.

	Ingo

------------->
commit 79f2b8594df44a4f3f11085df656883525b74abd
Author: Ingo Molnar <mingo@elte.hu>
Date:   Tue Jul 22 09:19:31 2008 +0200

    net 8390p: fix interface usage
    
    various drivers were using the wrong APIs:
    
     drivers/built-in.o: In function `hp_probe1':
     hp.c:(.init.text+0xa280): undefined reference to `NS8390_init'
    
    fixed via:
    
    cd drivers/net/; sed -i 's/NS8390_/NS8390p_/g' \
      $(grep -l NS8390_ $(grep 8390p.o Makefile | cut -d' ' -f3 | \
          sed 's/.o$/.c/g'))
    
    Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 drivers/net/hp-plus.c |    2 +-
 drivers/net/hp.c      |    2 +-
 drivers/net/ne.c      |    2 +-
 drivers/net/ne2.c     |    2 +-
 4 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/net/hp-plus.c b/drivers/net/hp-plus.c
index c2c4f49..8239939 100644
--- a/drivers/net/hp-plus.c
+++ b/drivers/net/hp-plus.c
@@ -262,7 +262,7 @@ static int __init hpp_probe1(struct net_device *dev, int ioaddr)
 	}
 
 	outw(Perf_Page, ioaddr + HP_PAGING);
-	NS8390_init(dev, 0);
+	NS8390p_init(dev, 0);
 	/* Leave the 8390 and HP chip reset. */
 	outw(inw(ioaddr + HPP_OPTION) & ~EnableIRQ, ioaddr + HPP_OPTION);
 
diff --git a/drivers/net/hp.c b/drivers/net/hp.c
index 8281209..0a8c649 100644
--- a/drivers/net/hp.c
+++ b/drivers/net/hp.c
@@ -389,7 +389,7 @@ static void __init
 hp_init_card(struct net_device *dev)
 {
 	int irq = dev->irq;
-	NS8390_init(dev, 0);
+	NS8390p_init(dev, 0);
 	outb_p(irqmap[irq&0x0f] | HP_RUN,
 		   dev->base_addr - NIC_OFFSET + HP_CONFIGURE);
 	return;
diff --git a/drivers/net/ne.c b/drivers/net/ne.c
index 3df231d..4a8a4b1 100644
--- a/drivers/net/ne.c
+++ b/drivers/net/ne.c
@@ -355,7 +355,7 @@ static int __init ne_probe1(struct net_device *dev, unsigned long ioaddr)
 	}
 
 	/* Read the 16 bytes of station address PROM.
-	   We must first initialize registers, similar to NS8390_init(eifdev, 0).
+	   We must first initialize registers, similar to NS8390p_init(eifdev, 0).
 	   We can't reliably read the SAPROM address without this.
 	   (I learned the hard way!). */
 	{
diff --git a/drivers/net/ne2.c b/drivers/net/ne2.c
index 8f72563..332df75 100644
--- a/drivers/net/ne2.c
+++ b/drivers/net/ne2.c
@@ -404,7 +404,7 @@ static int __init ne2_probe1(struct net_device *dev, int slot)
 
 	/* Read the 16 bytes of station address PROM.
 	   We must first initialize registers, similar to
-	   NS8390_init(eifdev, 0).
+	   NS8390p_init(eifdev, 0).
 	   We can't reliably read the SAPROM address without this.
 	   (I learned the hard way!). */
 	{

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Alan Cox]

> find the full tested patch below. It solves the build failures i 
> triggered so far.

Yes - that matches the patch I sent Jeff on the 8th July

Subject: Re: linux-next: Tree for July 8 (ns8390)

From: Alan Cox <alan@redhat.com>

Please try this and if it sorts it fold it into the driver. This is a diff
between my tree and the linux-next tree
---

 drivers/net/hp-plus.c |    2 +-
 drivers/net/hp.c      |    2 +-
 drivers/net/ne.c      |    8 ++++----
 drivers/net/wd.c      |    2 +-
 4 files changed, 7 insertions(+), 7 deletions(-)

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Jeff Garzik]

Ingo Molnar wrote:
> * Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> 
>> On Mon, 21 Jul 2008 21:15:19 +0200
>> Ingo Molnar <mingo@elte.hu> wrote:
>>
>>> * Atsushi Nemoto <anemo@mba.ocn.ne.jp> wrote:
>>>
>>>> Since commit 055e5110ae0c0c1176a75b78d789294f2ff2f7af ("8390: Split
>>>> 8390 support into a pausing and a non pausing driver core"), ne.c
>>>> cause this build failure:
>>>>
>>>> drivers/built-in.o: In function `ne_probe1':
>>>> ne.c:(.init.text+0x2464): undefined reference to `NS8390_init'
>>>> drivers/built-in.o: In function `ne_block_output':
>>>> ne.c:(.text.ne_block_output+0x1b0): undefined reference to `NS8390_init'
>>> i just triggered a similar build failure, but in a different 
>>> ne2000-derived file:
>> I'll take a look. Various odd things happened with the Makefile for 
>> this patch and the follow up fixes. If the followup didn't get applied 
>> then that might explain the breakage.
> 
> ok. I also got:
> 
>  drivers/built-in.o: In function `hp_probe1':
>  hp.c:(.init.text+0xa280): undefined reference to `NS8390_init'
> 
> so i suspect most of the 8390p.o library using ne2000 drivers are 
> affected?
> 
> i think instead of a manual review i think this is better to fix all the 
> problems:
> 
> cd drivers/net/; sed -i 's/NS8390_/NS8390p_/g' \
>   $(grep -l NS8390_ $(grep 8390p.o Makefile | cut -d' ' -f3 | \
>      sed 's/.o$/.c/g'))
> 
> find the full tested patch below. It solves the build failures i 
> triggered so far.
> 
> 	Ingo
> 
> ------------->
> commit 79f2b8594df44a4f3f11085df656883525b74abd
> Author: Ingo Molnar <mingo@elte.hu>
> Date:   Tue Jul 22 09:19:31 2008 +0200
> 
>     net 8390p: fix interface usage
>     
>     various drivers were using the wrong APIs:
>     
>      drivers/built-in.o: In function `hp_probe1':
>      hp.c:(.init.text+0xa280): undefined reference to `NS8390_init'
>     
>     fixed via:
>     
>     cd drivers/net/; sed -i 's/NS8390_/NS8390p_/g' \
>       $(grep -l NS8390_ $(grep 8390p.o Makefile | cut -d' ' -f3 | \
>           sed 's/.o$/.c/g'))
>     
>     Signed-off-by: Ingo Molnar <mingo@elte.hu>
> ---
>  drivers/net/hp-plus.c |    2 +-
>  drivers/net/hp.c      |    2 +-
>  drivers/net/ne.c      |    2 +-
>  drivers/net/ne2.c     |    2 +-
>  4 files changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/net/hp-plus.c b/drivers/net/hp-plus.c
> index c2c4f49..8239939 100644
> --- a/drivers/net/hp-plus.c
> +++ b/drivers/net/hp-plus.c
> @@ -262,7 +262,7 @@ static int __init hpp_probe1(struct net_device *dev, int ioaddr)
>  	}
>  
>  	outw(Perf_Page, ioaddr + HP_PAGING);
> -	NS8390_init(dev, 0);
> +	NS8390p_init(dev, 0);
>  	/* Leave the 8390 and HP chip reset. */
>  	outw(inw(ioaddr + HPP_OPTION) & ~EnableIRQ, ioaddr + HPP_OPTION);
>  
> diff --git a/drivers/net/hp.c b/drivers/net/hp.c
> index 8281209..0a8c649 100644
> --- a/drivers/net/hp.c
> +++ b/drivers/net/hp.c
> @@ -389,7 +389,7 @@ static void __init
>  hp_init_card(struct net_device *dev)
>  {
>  	int irq = dev->irq;
> -	NS8390_init(dev, 0);
> +	NS8390p_init(dev, 0);
>  	outb_p(irqmap[irq&0x0f] | HP_RUN,
>  		   dev->base_addr - NIC_OFFSET + HP_CONFIGURE);
>  	return;
> diff --git a/drivers/net/ne.c b/drivers/net/ne.c
> index 3df231d..4a8a4b1 100644
> --- a/drivers/net/ne.c
> +++ b/drivers/net/ne.c
> @@ -355,7 +355,7 @@ static int __init ne_probe1(struct net_device *dev, unsigned long ioaddr)
>  	}
>  
>  	/* Read the 16 bytes of station address PROM.
> -	   We must first initialize registers, similar to NS8390_init(eifdev, 0).
> +	   We must first initialize registers, similar to NS8390p_init(eifdev, 0).
>  	   We can't reliably read the SAPROM address without this.
>  	   (I learned the hard way!). */
>  	{
> diff --git a/drivers/net/ne2.c b/drivers/net/ne2.c
> index 8f72563..332df75 100644
> --- a/drivers/net/ne2.c
> +++ b/drivers/net/ne2.c
> @@ -404,7 +404,7 @@ static int __init ne2_probe1(struct net_device *dev, int slot)
>  
>  	/* Read the 16 bytes of station address PROM.
>  	   We must first initialize registers, similar to
> -	   NS8390_init(eifdev, 0).
> +	   NS8390p_init(eifdev, 0).
>  	   We can't reliably read the SAPROM address without this.
>  	   (I learned the hard way!). */

applied

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Ingo Molnar]

* Jeff Garzik <jeff@garzik.org> wrote:

>>>>> drivers/built-in.o: In function `ne_probe1':
>>>>> ne.c:(.init.text+0x2464): undefined reference to `NS8390_init'
>>>>> drivers/built-in.o: In function `ne_block_output':
>>>>> ne.c:(.text.ne_block_output+0x1b0): undefined reference to `NS8390_init'
>>>> i just triggered a similar build failure, but in a different  
>>>> ne2000-derived file:
>>> I'll take a look. Various odd things happened with the Makefile for  
>>> this patch and the follow up fixes. If the followup didn't get 
>>> applied then that might explain the breakage.
>>
>> ok. I also got:
>>
>>  drivers/built-in.o: In function `hp_probe1':
>>  hp.c:(.init.text+0xa280): undefined reference to `NS8390_init'
>>
>> so i suspect most of the 8390p.o library using ne2000 drivers are  
>> affected?
>>
>> i think instead of a manual review i think this is better to fix all 
>> the problems:
>>
>> cd drivers/net/; sed -i 's/NS8390_/NS8390p_/g' \
>>   $(grep -l NS8390_ $(grep 8390p.o Makefile | cut -d' ' -f3 | \
>>      sed 's/.o$/.c/g'))
>>
>> find the full tested patch below. It solves the build failures i  
>> triggered so far.
>>
>> 	Ingo
>>
>> ------------->
>> commit 79f2b8594df44a4f3f11085df656883525b74abd
>> Author: Ingo Molnar <mingo@elte.hu>
>> Date:   Tue Jul 22 09:19:31 2008 +0200
>>
>>     net 8390p: fix interface usage
>>         various drivers were using the wrong APIs:
>>          drivers/built-in.o: In function `hp_probe1':
>>      hp.c:(.init.text+0xa280): undefined reference to `NS8390_init'
>>         fixed via:
>>         cd drivers/net/; sed -i 's/NS8390_/NS8390p_/g' \
>>       $(grep -l NS8390_ $(grep 8390p.o Makefile | cut -d' ' -f3 | \
>>           sed 's/.o$/.c/g'))
>>         Signed-off-by: Ingo Molnar <mingo@elte.hu>
>> ---
>>  drivers/net/hp-plus.c |    2 +-
>>  drivers/net/hp.c      |    2 +-
>>  drivers/net/ne.c      |    2 +-
>>  drivers/net/ne2.c     |    2 +-
>>  4 files changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/net/hp-plus.c b/drivers/net/hp-plus.c
>> index c2c4f49..8239939 100644
>> --- a/drivers/net/hp-plus.c
>> +++ b/drivers/net/hp-plus.c
>> @@ -262,7 +262,7 @@ static int __init hpp_probe1(struct net_device *dev, int ioaddr)
>>  	}
>>   	outw(Perf_Page, ioaddr + HP_PAGING);
>> -	NS8390_init(dev, 0);
>> +	NS8390p_init(dev, 0);
>>  	/* Leave the 8390 and HP chip reset. */
>>  	outw(inw(ioaddr + HPP_OPTION) & ~EnableIRQ, ioaddr + HPP_OPTION);
>>  diff --git a/drivers/net/hp.c b/drivers/net/hp.c
>> index 8281209..0a8c649 100644
>> --- a/drivers/net/hp.c
>> +++ b/drivers/net/hp.c
>> @@ -389,7 +389,7 @@ static void __init
>>  hp_init_card(struct net_device *dev)
>>  {
>>  	int irq = dev->irq;
>> -	NS8390_init(dev, 0);
>> +	NS8390p_init(dev, 0);
>>  	outb_p(irqmap[irq&0x0f] | HP_RUN,
>>  		   dev->base_addr - NIC_OFFSET + HP_CONFIGURE);
>>  	return;
>> diff --git a/drivers/net/ne.c b/drivers/net/ne.c
>> index 3df231d..4a8a4b1 100644
>> --- a/drivers/net/ne.c
>> +++ b/drivers/net/ne.c
>> @@ -355,7 +355,7 @@ static int __init ne_probe1(struct net_device *dev, unsigned long ioaddr)
>>  	}
>>   	/* Read the 16 bytes of station address PROM.
>> -	   We must first initialize registers, similar to NS8390_init(eifdev, 0).
>> +	   We must first initialize registers, similar to NS8390p_init(eifdev, 0).
>>  	   We can't reliably read the SAPROM address without this.
>>  	   (I learned the hard way!). */
>>  	{
>> diff --git a/drivers/net/ne2.c b/drivers/net/ne2.c
>> index 8f72563..332df75 100644
>> --- a/drivers/net/ne2.c
>> +++ b/drivers/net/ne2.c
>> @@ -404,7 +404,7 @@ static int __init ne2_probe1(struct net_device *dev, int slot)
>>   	/* Read the 16 bytes of station address PROM.
>>  	   We must first initialize registers, similar to
>> -	   NS8390_init(eifdev, 0).
>> +	   NS8390p_init(eifdev, 0).
>>  	   We can't reliably read the SAPROM address without this.
>>  	   (I learned the hard way!). */
>
> applied

thanks Jeff. I think the fix below from Atsushi-san might still be 
missing from upstream?

	Ingo

-------------------->
commit 994cd3a79a431f5ce234478e951856357f4bc94c
Author: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Date:   Mon Jul 21 23:35:12 2008 +0900

    8390: Fix ne.c build
    
    Since commit 055e5110ae0c0c1176a75b78d789294f2ff2f7af ("8390: Split
    8390 support into a pausing and a non pausing driver core"), ne.c
    cause this build failure:
    
    drivers/built-in.o: In function `ne_probe1':
    ne.c:(.init.text+0x2464): undefined reference to `NS8390_init'
    drivers/built-in.o: In function `ne_block_output':
    ne.c:(.text.ne_block_output+0x1b0): undefined reference to `NS8390_init'
    
    Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
    Cc: jgarzik@redhat.com
    Cc: akpm@linux-foundation.org
    Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
    Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 drivers/net/ne.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ne.c b/drivers/net/ne.c
index 2fec612..4a8a4b1 100644
--- a/drivers/net/ne.c
+++ b/drivers/net/ne.c
@@ -536,7 +536,7 @@ static int __init ne_probe1(struct net_device *dev, unsigned long ioaddr)
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	dev->poll_controller = eip_poll;
 #endif
-	NS8390_init(dev, 0);
+	NS8390p_init(dev, 0);
 
 	ret = register_netdev(dev);
 	if (ret)
@@ -794,7 +794,7 @@ retry:
 		if (time_after(jiffies, dma_start + 2*HZ/100)) {		/* 20ms */
 			printk(KERN_WARNING "%s: timeout waiting for Tx RDC.\n", dev->name);
 			ne_reset_8390(dev);
-			NS8390_init(dev,1);
+			NS8390p_init(dev,1);
 			break;
 		}
 
@@ -855,7 +855,7 @@ static int ne_drv_resume(struct platform_device *pdev)
 
 	if (netif_running(dev)) {
 		ne_reset_8390(dev);
-		NS8390_init(dev, 1);
+		NS8390p_init(dev, 1);
 		netif_device_attach(dev);
 	}
 	return 0;

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Ingo Molnar]

not sure it's known but despite all the fixes i still get:

 drivers/built-in.o: In function `ne_drv_resume':
 ne.c:(.text+0x146ba4): undefined reference to `NS8390_init'
 drivers/built-in.o: In function `ne_block_output':
 ne.c:(.text+0x146e00): undefined reference to `NS8390_init'
 drivers/built-in.o: In function `hpp_close':
 hp-plus.c:(.text+0x148c3b): undefined reference to `ei_close'
 drivers/built-in.o: In function `hpp_open':
 hp-plus.c:(.text+0x148c77): undefined reference to `ei_interrupt'
 hp-plus.c:(.text+0x148ce7): undefined reference to `ei_open'
 drivers/built-in.o: In function `ne_probe1':
 ne.c:(.init.text+0xb532): undefined reference to `NS8390_init'
 drivers/built-in.o: In function `hpp_probe1':
 hp-plus.c:(.init.text+0xc033): undefined reference to `ei_poll'
 drivers/built-in.o: In function `hp_plus_probe':
 (.init.text+0xc134): undefined reference to `__alloc_ei_netdev'
 make: *** [.tmp_vmlinux1] Error 1

with this config and latest -git (v2.6.26-6371-g338b9bb):

 http://redhat.com/~mingo/misc/config-Thu_Jul_24_07_13_29_CEST_2008.bad

	Ingo

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Alan Cox]

On Thu, 24 Jul 2008 07:26:03 +0200
Ingo Molnar <mingo@elte.hu> wrote:

> 
> not sure it's known but despite all the fixes i still get:

I'll have a look at this next week or possibly the week after. The
original patches worked but they spent 3 months going round in circles
and I've completely given up even trying to work out what happened or why
all the previous updates I sent were ignored.

Alan

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Ingo Molnar]

* Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> On Thu, 24 Jul 2008 07:26:03 +0200
> Ingo Molnar <mingo@elte.hu> wrote:
> 
> > 
> > not sure it's known but despite all the fixes i still get:
> 
> I'll have a look at this next week or possibly the week after. The 
> original patches worked but they spent 3 months going round in circles 
> and I've completely given up even trying to work out what happened or 
> why all the previous updates I sent were ignored.

ok - meanwhile i'm using the workaround hack below.

	Ingo

--------------------->
commit 91de2dc9031b92157f067d78aaf2c8e58a965c0d
Author: Ingo Molnar <mingo@elte.hu>
Date:   Mon Jul 21 21:18:29 2008 +0200

    hp-plus.c: work around build failure
    
    Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 drivers/net/Kconfig |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index fa533c2..1c809e4 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -1129,6 +1129,8 @@ config EEXPRESS_PRO
 config HPLAN_PLUS
 	tristate "HP PCLAN+ (27247B and 27252A) support"
 	depends on NET_ISA
+	# broken build with config-Mon_Jul_21_20_21_08_CEST_2008.bad
+	depends on 0
 	select CRC32
 	help
 	  If you have a network (Ethernet) card of this type, say Y and read

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Jeff Garzik]

Alan Cox wrote:
> On Thu, 24 Jul 2008 07:26:03 +0200
> Ingo Molnar <mingo@elte.hu> wrote:
> 
>> not sure it's known but despite all the fixes i still get:
> 
> I'll have a look at this next week or possibly the week after. The
> original patches worked but they spent 3 months going round in circles
> and I've completely given up even trying to work out what happened or why
> all the previous updates I sent were ignored.

My apologies.  FWIW, from my point of view the same patch(es) kept 
getting fixed and updated and going around in circles, so I kept 
waiting.  Since it was was a net-next issue (until now), I lowered the 
priority and waited to let you guys (you, Andrew, Randy mainly) sort 
things out.

So far the only remaining patch I have outstanding and needing to be 
applied is from Atsushi-san via Ingo (message id 
<20080723104520.GA20672@elte.hu>)...  is that the last of it?

	Jeff

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Randy.Dunlap]

On Thu, 24 Jul 2008, Jeff Garzik wrote:

> Alan Cox wrote:
> > On Thu, 24 Jul 2008 07:26:03 +0200
> > Ingo Molnar <mingo@elte.hu> wrote:
> > 
> > > not sure it's known but despite all the fixes i still get:
> > 
> > I'll have a look at this next week or possibly the week after. The
> > original patches worked but they spent 3 months going round in circles
> > and I've completely given up even trying to work out what happened or why
> > all the previous updates I sent were ignored.
> 
> My apologies.  FWIW, from my point of view the same patch(es) kept getting
> fixed and updated and going around in circles, so I kept waiting.  Since it
> was was a net-next issue (until now), I lowered the priority and waited to let
> you guys (you, Andrew, Randy mainly) sort things out.

Randy got frustrated and gave up on this particular issue & patch(es).
This one has gone on far too long...

> So far the only remaining patch I have outstanding and needing to be applied
> is from Atsushi-san via Ingo (message id <20080723104520.GA20672@elte.hu>)...
> is that the last of it?

-- 
~Randy

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Alan Cox]

> My apologies.  FWIW, from my point of view the same patch(es) kept 
> getting fixed and updated and going around in circles, so I kept 
> waiting.  Since it was was a net-next issue (until now), I lowered the 
> priority and waited to let you guys (you, Andrew, Randy mainly) sort 
> things out.
> 
> So far the only remaining patch I have outstanding and needing to be 
> applied is from Atsushi-san via Ingo (message id 
> <20080723104520.GA20672@elte.hu>)...  is that the last of it?

I've got a reference tree that I will regenerate and diff against the
upstream GIT tree to check. It looks like some of the ei_ to eip_ changes
have gone for a walk.

Its on my TODO list and now IT821x is working again the panic mode stuff
is over I hope

Alan

Re: [build failure] hp-plus.c: (.text+0xd7c1a): undefined reference to `ei_close'

[Alan Cox]

On Thu, 24 Jul 2008 07:26:03 +0200
Ingo Molnar <mingo@elte.hu> wrote:

> 
> not sure it's known but despite all the fixes i still get:
> 
>  drivers/built-in.o: In function `ne_drv_resume':
>  ne.c:(.text+0x146ba4): undefined reference to `NS8390_init'
>  drivers/built-in.o: In function `ne_block_output':
>  ne.c:(.text+0x146e00): undefined reference to `NS8390_init'

This should do the trick. I can't test it currently as that configuration
now fails to build with mm errors in current git head..

mm/hugetlb.c: In function ‘hugetlb_acct_memory’: mm/hugetlb.c:1507:
error: implicit declaration of function ‘cpuset_mems_nr’

8390: Fix various compile errors

From: Alan Cox <alan@redhat.com>

Some of the original patches got lost along the way, so sort out the
remaining stuff
---

 drivers/net/Makefile |    2 +-
 drivers/net/ne.c     |    6 +++---
 2 files changed, 4 insertions(+), 4 deletions(-)


diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index 7629c90..f66b79b 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -111,7 +111,7 @@ obj-$(CONFIG_EL2) += 3c503.o 8390p.o
 obj-$(CONFIG_NE2000) += ne.o 8390p.o
 obj-$(CONFIG_NE2_MCA) += ne2.o 8390p.o
 obj-$(CONFIG_HPLAN) += hp.o 8390p.o
-obj-$(CONFIG_HPLAN_PLUS) += hp-plus.o 8390p.o
+obj-$(CONFIG_HPLAN_PLUS) += hp-plus.o 8390.o
 obj-$(CONFIG_ULTRA) += smc-ultra.o 8390.o
 obj-$(CONFIG_ULTRAMCA) += smc-mca.o 8390.o
 obj-$(CONFIG_ULTRA32) += smc-ultra32.o 8390.o
diff --git a/drivers/net/ne.c b/drivers/net/ne.c
index 2fec612..4a8a4b1 100644
--- a/drivers/net/ne.c
+++ b/drivers/net/ne.c
@@ -536,7 +536,7 @@ static int __init ne_probe1(struct net_device *dev, unsigned long ioaddr)
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	dev->poll_controller = eip_poll;
 #endif
-	NS8390_init(dev, 0);
+	NS8390p_init(dev, 0);
 
 	ret = register_netdev(dev);
 	if (ret)
@@ -794,7 +794,7 @@ retry:
 		if (time_after(jiffies, dma_start + 2*HZ/100)) {		/* 20ms */
 			printk(KERN_WARNING "%s: timeout waiting for Tx RDC.\n", dev->name);
 			ne_reset_8390(dev);
-			NS8390_init(dev,1);
+			NS8390p_init(dev,1);
 			break;
 		}
 
@@ -855,7 +855,7 @@ static int ne_drv_resume(struct platform_device *pdev)
 
 	if (netif_running(dev)) {
 		ne_reset_8390(dev);
-		NS8390_init(dev, 1);
+		NS8390p_init(dev, 1);
 		netif_device_attach(dev);
 	}
 	return 0;

[PATCH] hugetlb: fix CONFIG_SYSCTL=n build

[Nishanth Aravamudan]

On 25.07.2008 [18:09:35 +0100], Alan Cox wrote:
> On Thu, 24 Jul 2008 07:26:03 +0200
> Ingo Molnar <mingo@elte.hu> wrote:
> 
> > 
> > not sure it's known but despite all the fixes i still get:
> > 
> >  drivers/built-in.o: In function `ne_drv_resume':
> >  ne.c:(.text+0x146ba4): undefined reference to `NS8390_init'
> >  drivers/built-in.o: In function `ne_block_output':
> >  ne.c:(.text+0x146e00): undefined reference to `NS8390_init'
> 
> This should do the trick. I can't test it currently as that configuration
> now fails to build with mm errors in current git head..
> 
> mm/hugetlb.c: In function ???hugetlb_acct_memory???: mm/hugetlb.c:1507:
> error: implicit declaration of function ???cpuset_mems_nr???

When moving the intialization functions around, I did not remove the
dependency of some helpers on CONFIG_SYSCTL. This led to a large chunk
of code being elided even though it is needed to build hugetlb.o.

Fixes a build failure reported by Alan Cox:

mm/hugetlb.c: In function ‘hugetlb_acct_memory’: mm/hugetlb.c:1507:
error: implicit declaration of function ‘cpuset_mems_nr’

Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 41341c4..3be79dc 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -1026,7 +1026,6 @@ static void __init report_hugepages(void)
 	}
 }
 
-#ifdef CONFIG_SYSCTL
 #ifdef CONFIG_HIGHMEM
 static void try_to_free_low(struct hstate *h, unsigned long count)
 {
@@ -1386,6 +1385,7 @@ static unsigned int cpuset_mems_nr(unsigned int *array)
 	return nr;
 }
 
+#ifdef CONFIG_SYSCTL
 int hugetlb_sysctl_handler(struct ctl_table *table, int write,
 			   struct file *file, void __user *buffer,
 			   size_t *length, loff_t *ppos)

-- 
Nishanth Aravamudan <nacc@us.ibm.com>
IBM Linux Technology Center

[PATCH RESEND] hugetlb: fix CONFIG_SYSCTL=n build

[Nishanth Aravamudan]

[resending with Linus' address fixed]

On 25.07.2008 [18:09:35 +0100], Alan Cox wrote:
> On Thu, 24 Jul 2008 07:26:03 +0200
> Ingo Molnar <mingo@elte.hu> wrote:
> 
> > 
> > not sure it's known but despite all the fixes i still get:
> > 
> >  drivers/built-in.o: In function `ne_drv_resume':
> >  ne.c:(.text+0x146ba4): undefined reference to `NS8390_init'
> >  drivers/built-in.o: In function `ne_block_output':
> >  ne.c:(.text+0x146e00): undefined reference to `NS8390_init'
> 
> This should do the trick. I can't test it currently as that configuration
> now fails to build with mm errors in current git head..
> 
> mm/hugetlb.c: In function ???hugetlb_acct_memory???: mm/hugetlb.c:1507:
> error: implicit declaration of function ???cpuset_mems_nr???

When moving the intialization functions around, I did not remove the
dependency of some helpers on CONFIG_SYSCTL. This led to a large chunk
of code being elided even though it is needed to build hugetlb.o.

Fixes a build failure reported by Alan Cox:

mm/hugetlb.c: In function ‘hugetlb_acct_memory’: mm/hugetlb.c:1507:
error: implicit declaration of function ‘cpuset_mems_nr’

Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 41341c4..3be79dc 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -1026,7 +1026,6 @@ static void __init report_hugepages(void)
 	}
 }
 
-#ifdef CONFIG_SYSCTL
 #ifdef CONFIG_HIGHMEM
 static void try_to_free_low(struct hstate *h, unsigned long count)
 {
@@ -1386,6 +1385,7 @@ static unsigned int cpuset_mems_nr(unsigned int *array)
 	return nr;
 }
 
+#ifdef CONFIG_SYSCTL
 int hugetlb_sysctl_handler(struct ctl_table *table, int write,
 			   struct file *file, void __user *buffer,
 			   size_t *length, loff_t *ppos)

-- 
Nishanth Aravamudan <nacc@us.ibm.com>
IBM Linux Technology Center

Re: [PATCH RESEND] hugetlb: fix CONFIG_SYSCTL=n build

[Andrew Morton]

On Fri, 25 Jul 2008 13:33:51 -0700
Nishanth Aravamudan <nacc@us.ibm.com> wrote:

> [resending with Linus' address fixed]
> 
> On 25.07.2008 [18:09:35 +0100], Alan Cox wrote:
> > On Thu, 24 Jul 2008 07:26:03 +0200
> > Ingo Molnar <mingo@elte.hu> wrote:
> > 
> > > 
> > > not sure it's known but despite all the fixes i still get:
> > > 
> > >  drivers/built-in.o: In function `ne_drv_resume':
> > >  ne.c:(.text+0x146ba4): undefined reference to `NS8390_init'
> > >  drivers/built-in.o: In function `ne_block_output':
> > >  ne.c:(.text+0x146e00): undefined reference to `NS8390_init'
> > 
> > This should do the trick. I can't test it currently as that configuration
> > now fails to build with mm errors in current git head..
> > 
> > mm/hugetlb.c: In function ???hugetlb_acct_memory???: mm/hugetlb.c:1507:
> > error: implicit declaration of function ???cpuset_mems_nr???

Remind me to give a wedgie to the next gcc developer I meet.

> 
> When moving the intialization functions around, I did not remove the
> dependency of some helpers on CONFIG_SYSCTL. This led to a large chunk
> of code being elided even though it is needed to build hugetlb.o.
> 
> Fixes a build failure reported by Alan Cox:
> 
> mm/hugetlb.c: In function ______hugetlb_acct_memory______: mm/hugetlb.c:1507:
> error: implicit declaration of function ______cpuset_mems_nr______

And the next one.

LANG=C stomps this silliness.

> Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
> 
> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> index 41341c4..3be79dc 100644
> --- a/mm/hugetlb.c
> +++ b/mm/hugetlb.c
> @@ -1026,7 +1026,6 @@ static void __init report_hugepages(void)
>  	}
>  }
>  
> -#ifdef CONFIG_SYSCTL
>  #ifdef CONFIG_HIGHMEM
>  static void try_to_free_low(struct hstate *h, unsigned long count)
>  {
> @@ -1386,6 +1385,7 @@ static unsigned int cpuset_mems_nr(unsigned int *array)
>  	return nr;
>  }
>  
> +#ifdef CONFIG_SYSCTL
>  int hugetlb_sysctl_handler(struct ctl_table *table, int write,
>  			   struct file *file, void __user *buffer,
>  			   size_t *length, loff_t *ppos)
> 

OK, we had a fix from Ingo but this one is better.

Re: [PATCH RESEND] hugetlb: fix CONFIG_SYSCTL=n build

[Nishanth Aravamudan]

On 25.07.2008 [14:29:21 -0700], Andrew Morton wrote:
> On Fri, 25 Jul 2008 13:33:51 -0700
> Nishanth Aravamudan <nacc@us.ibm.com> wrote:
> 
> > [resending with Linus' address fixed]
> > 
> > On 25.07.2008 [18:09:35 +0100], Alan Cox wrote:
> > > On Thu, 24 Jul 2008 07:26:03 +0200
> > > Ingo Molnar <mingo@elte.hu> wrote:
> > > 
> > > > 
> > > > not sure it's known but despite all the fixes i still get:
> > > > 
> > > >  drivers/built-in.o: In function `ne_drv_resume':
> > > >  ne.c:(.text+0x146ba4): undefined reference to `NS8390_init'
> > > >  drivers/built-in.o: In function `ne_block_output':
> > > >  ne.c:(.text+0x146e00): undefined reference to `NS8390_init'
> > > 
> > > This should do the trick. I can't test it currently as that configuration
> > > now fails to build with mm errors in current git head..
> > > 
> > > mm/hugetlb.c: In function ???hugetlb_acct_memory???: mm/hugetlb.c:1507:
> > > error: implicit declaration of function ???cpuset_mems_nr???
> 
> Remind me to give a wedgie to the next gcc developer I meet.

This may have been C&P error on my part, sorry about that.

> > When moving the intialization functions around, I did not remove the
> > dependency of some helpers on CONFIG_SYSCTL. This led to a large chunk
> > of code being elided even though it is needed to build hugetlb.o.
> > 
> > Fixes a build failure reported by Alan Cox:
> > 
> > mm/hugetlb.c: In function ______hugetlb_acct_memory______: mm/hugetlb.c:1507:
> > error: implicit declaration of function ______cpuset_mems_nr______
> 
> And the next one.
> 
> LANG=C stomps this silliness.

Yep, I've fixed the configuration of this box (I hope).

Thanks,
Nish

-- 
Nishanth Aravamudan <nacc@us.ibm.com>
IBM Linux Technology Center

Re: [PATCH RESEND] hugetlb: fix CONFIG_SYSCTL=n build

[Linus Torvalds]

On Fri, 25 Jul 2008, Nishanth Aravamudan wrote:
>
> [resending with Linus' address fixed]

Hmm, this doesn't apply for me any more. Can you check the current -git 
tree?

		Linus

Re: [PATCH RESEND] hugetlb: fix CONFIG_SYSCTL=n build

[Andrew Morton]

On Fri, 25 Jul 2008 17:32:38 -0700 (PDT)
Linus Torvalds <torvalds@linux-foundation.org> wrote:

> 
> 
> On Fri, 25 Jul 2008, Nishanth Aravamudan wrote:
> >
> > [resending with Linus' address fixed]
> 
> Hmm, this doesn't apply for me any more. Can you check the current -git 
> tree?
> 

Is OK, I have a fixed-up version here which I'm about to start testing.

Re: [PATCH RESEND] hugetlb: fix CONFIG_SYSCTL=n build

[Nishanth Aravamudan]

On 25.07.2008 [17:44:56 -0700], Andrew Morton wrote:
> On Fri, 25 Jul 2008 17:32:38 -0700 (PDT)
> Linus Torvalds <torvalds@linux-foundation.org> wrote:
> 
> > 
> > 
> > On Fri, 25 Jul 2008, Nishanth Aravamudan wrote:
> > >
> > > [resending with Linus' address fixed]
> > 
> > Hmm, this doesn't apply for me any more. Can you check the current
> > -git tree?
> > 
> 
> Is OK, I have a fixed-up version here which I'm about to start testing.

Ah, I see that Linus had already picked up Ingo's fix
(e44d1b2998d62a1f2f4d7eb17b56ba396535509f). While that does fix the
build, it isn't really the right solution. I'm guessing Andrew's version
will be a revert (either real or in patch-form) of that commit and an
application of mine?

Thanks,
Nish

-- 
Nishanth Aravamudan <nacc@us.ibm.com>
IBM Linux Technology Center

Re: [PATCH RESEND] hugetlb: fix CONFIG_SYSCTL=n build

[Andrew Morton]

On Fri, 25 Jul 2008 18:25:49 -0700 Nishanth Aravamudan <nacc@us.ibm.com> wrote:

> On 25.07.2008 [17:44:56 -0700], Andrew Morton wrote:
> > On Fri, 25 Jul 2008 17:32:38 -0700 (PDT)
> > Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > 
> > > 
> > > 
> > > On Fri, 25 Jul 2008, Nishanth Aravamudan wrote:
> > > >
> > > > [resending with Linus' address fixed]
> > > 
> > > Hmm, this doesn't apply for me any more. Can you check the current
> > > -git tree?
> > > 
> > 
> > Is OK, I have a fixed-up version here which I'm about to start testing.
> 
> Ah, I see that Linus had already picked up Ingo's fix
> (e44d1b2998d62a1f2f4d7eb17b56ba396535509f).

Yeah, that fixes the build but tosses out a storm of
unused-static-function warnings.

> While that does fix the
> build, it isn't really the right solution. I'm guessing Andrew's version
> will be a revert (either real or in patch-form) of that commit and an
> application of mine?

I'll fix it up.

Re: [PATCH] 8390: Fix ne.c build

[Alan Cox]

On Mon, 21 Jul 2008 23:35:12 +0900 (JST)
Atsushi Nemoto <anemo@mba.ocn.ne.jp> wrote:

> Since commit 055e5110ae0c0c1176a75b78d789294f2ff2f7af ("8390: Split
> 8390 support into a pausing and a non pausing driver core"), ne.c
> cause this build failure:
> 
> drivers/built-in.o: In function `ne_probe1':
> ne.c:(.init.text+0x2464): undefined reference to `NS8390_init'
> drivers/built-in.o: In function `ne_block_output':
> ne.c:(.text.ne_block_output+0x1b0): undefined reference to `NS8390_init'

Yes - I sent the same to Jeff Garzik early July.

Alan