Latest e1000 7.3.20 driver busted with 2.6.20.4

Latest e1000 7.3.20 driver busted with 2.6.20.4

[Jeff V. Merkey]

CC [M]  drivers/net/chelsio/mv88x201x.o
  CC [M]  drivers/net/chelsio/my3126.o
  LD [M]  drivers/net/chelsio/cxgb.o
  CC      drivers/net/e1000/e1000_main.o
drivers/net/e1000/e1000_main.c:1185:45: error: macro "INIT_WORK" passed 
3 arguments, but takes just 2
ared (first use in this function)
drivers/net/e1000/e1000_main.c:1184: error: (Each undeclared identifier 
is reported only once
drivers/net/e1000/e1000_main.c:1184: error: for each function it appears 
in.)
make[3]: *** [drivers/net/e1000/e1000_main.o] Error 1
make[2]: *** [drivers/net/e1000] Error 2
make[1]: *** [drivers/net] Error 2
make: *** [drivers] Error 2

Jeff

Re: Latest e1000 7.3.20 driver busted with 2.6.20.4

[Kok, Auke]

Jeff V. Merkey wrote:
> CC [M]  drivers/net/chelsio/mv88x201x.o
>   CC [M]  drivers/net/chelsio/my3126.o
>   LD [M]  drivers/net/chelsio/cxgb.o
>   CC      drivers/net/e1000/e1000_main.o
> drivers/net/e1000/e1000_main.c:1185:45: error: macro "INIT_WORK" passed 
> 3 arguments, but takes just 2
> ared (first use in this function)
> drivers/net/e1000/e1000_main.c:1184: error: (Each undeclared identifier 
> is reported only once
> drivers/net/e1000/e1000_main.c:1184: error: for each function it appears 
> in.)
> make[3]: *** [drivers/net/e1000/e1000_main.o] Error 1
> make[2]: *** [drivers/net/e1000] Error 2
> make[1]: *** [drivers/net] Error 2
> make: *** [drivers] Error 2

this is hardly a thread for lkml as you're posting about out out-of-tree driver. 
  This issue has also been fixed in 7.4.35 which was released last week on 
e1000.sf.net. In the future, please post this stuff to 
e1000-devel@lists.sourceforge.net where we discuss the out-of-tree driver.

Cheers,

Auke

Re: Latest e1000 7.3.20 driver busted with 2.6.20.4

[Jeff V. Merkey]

Kok, Auke wrote:

> Jeff V. Merkey wrote:
>
>> CC [M]  drivers/net/chelsio/mv88x201x.o
>>   CC [M]  drivers/net/chelsio/my3126.o
>>   LD [M]  drivers/net/chelsio/cxgb.o
>>   CC      drivers/net/e1000/e1000_main.o
>> drivers/net/e1000/e1000_main.c:1185:45: error: macro "INIT_WORK" 
>> passed 3 arguments, but takes just 2
>> ared (first use in this function)
>> drivers/net/e1000/e1000_main.c:1184: error: (Each undeclared 
>> identifier is reported only once
>> drivers/net/e1000/e1000_main.c:1184: error: for each function it 
>> appears in.)
>> make[3]: *** [drivers/net/e1000/e1000_main.o] Error 1
>> make[2]: *** [drivers/net/e1000] Error 2
>> make[1]: *** [drivers/net] Error 2
>> make: *** [drivers] Error 2
>
>
> this is hardly a thread for lkml as you're posting about out 
> out-of-tree driver.  This issue has also been fixed in 7.4.35 which 
> was released last week on e1000.sf.net. In the future, please post 
> this stuff to e1000-devel@lists.sourceforge.net where we discuss the 
> out-of-tree driver.
>
> Cheers,
>
> Auke
>
Thanks.  I'll download Intel's latest driver and test it.   I wasn't 
discussing the out of tree driver (why do you have a broken Linux out of 
tree driver, BTW?   If its for Linux, shouldn't it be updated here 
first) but the driver in 2.6.20.4 (which is missing support four 
additional chipsets). 

Jeff

Re: Latest e1000 7.3.20 driver busted with 2.6.20.4

[Kok, Auke]

Jeff V. Merkey wrote:
> Kok, Auke wrote:
> 
>> Jeff V. Merkey wrote:
>>
>>> CC [M]  drivers/net/chelsio/mv88x201x.o
>>>   CC [M]  drivers/net/chelsio/my3126.o
>>>   LD [M]  drivers/net/chelsio/cxgb.o
>>>   CC      drivers/net/e1000/e1000_main.o
>>> drivers/net/e1000/e1000_main.c:1185:45: error: macro "INIT_WORK" 
>>> passed 3 arguments, but takes just 2
>>> ared (first use in this function)
>>> drivers/net/e1000/e1000_main.c:1184: error: (Each undeclared 
>>> identifier is reported only once
>>> drivers/net/e1000/e1000_main.c:1184: error: for each function it 
>>> appears in.)
>>> make[3]: *** [drivers/net/e1000/e1000_main.o] Error 1
>>> make[2]: *** [drivers/net/e1000] Error 2
>>> make[1]: *** [drivers/net] Error 2
>>> make: *** [drivers] Error 2
>>
>> this is hardly a thread for lkml as you're posting about out 
>> out-of-tree driver.  This issue has also been fixed in 7.4.35 which 
>> was released last week on e1000.sf.net. In the future, please post 
>> this stuff to e1000-devel@lists.sourceforge.net where we discuss the 
>> out-of-tree driver.
>>
>> Cheers,
>>
>> Auke
>>
> Thanks.  I'll download Intel's latest driver and test it.   I wasn't 
> discussing the out of tree driver (why do you have a broken Linux out of 
> tree driver, BTW?   If its for Linux, shouldn't it be updated here 
> first) but the driver in 2.6.20.4 (which is missing support four 
> additional chipsets). 

Our hardware release schedule is not aligned with kernel releases, and customers 
need to have a working (and tested) driver once the hardware is released. We try 
to get the updates into the kernel as fast as we can, but often meet resistance, 
as the two parties involved just have different interests.

Are you saying that the in-kernel driver in 2.6.20.4 is broken? That is bizarre, 
and should never happen. It compiles fine for me here...

Auke

Re: Latest e1000 7.3.20 driver busted with 2.6.20.4

[Jeff V. Merkey]

Kok, Auke wrote:

> Jeff V. Merkey wrote:
>
>> Kok, Auke wrote:
>>
>>> Jeff V. Merkey wrote:
>>>
>>>> CC [M]  drivers/net/chelsio/mv88x201x.o
>>>>   CC [M]  drivers/net/chelsio/my3126.o
>>>>   LD [M]  drivers/net/chelsio/cxgb.o
>>>>   CC      drivers/net/e1000/e1000_main.o
>>>> drivers/net/e1000/e1000_main.c:1185:45: error: macro "INIT_WORK" 
>>>> passed 3 arguments, but takes just 2
>>>> ared (first use in this function)
>>>> drivers/net/e1000/e1000_main.c:1184: error: (Each undeclared 
>>>> identifier is reported only once
>>>> drivers/net/e1000/e1000_main.c:1184: error: for each function it 
>>>> appears in.)
>>>> make[3]: *** [drivers/net/e1000/e1000_main.o] Error 1
>>>> make[2]: *** [drivers/net/e1000] Error 2
>>>> make[1]: *** [drivers/net] Error 2
>>>> make: *** [drivers] Error 2
>>>
>>>
>>> this is hardly a thread for lkml as you're posting about out 
>>> out-of-tree driver.  This issue has also been fixed in 7.4.35 which 
>>> was released last week on e1000.sf.net. In the future, please post 
>>> this stuff to e1000-devel@lists.sourceforge.net where we discuss the 
>>> out-of-tree driver.
>>>
>>> Cheers,
>>>
>>> Auke
>>>
>> Thanks.  I'll download Intel's latest driver and test it.   I wasn't 
>> discussing the out of tree driver (why do you have a broken Linux out 
>> of tree driver, BTW?   If its for Linux, shouldn't it be updated here 
>> first) but the driver in 2.6.20.4 (which is missing support four 
>> additional chipsets). 
>
>
> Our hardware release schedule is not aligned with kernel releases, and 
> customers need to have a working (and tested) driver once the hardware 
> is released. We try to get the updates into the kernel as fast as we 
> can, but often meet resistance, as the two parties involved just have 
> different interests.
>
> Are you saying that the in-kernel driver in 2.6.20.4 is broken? That 
> is bizarre, and should never happen. It compiles fine for me here...
>
>
It's missing support for some of the newer cards.  Your sales folks 
shipped us some newer hardware with the improved e1000, so I will have 
to adapt the newer drivers to 2.4.20.4.   Sounds like you are on top of 
it and things will sync up eventually.  :-)

Jeff