Why each time of kernel compilation must be done completely?

Why each time of kernel compilation must be done completely?

[秦弋戈]

Hi,


This problem has puzzled me for a long time. A normal C make file could parse the reliance and just regenerate the modified source file. But each time when I modified the kernel and recompile it, it does a complete compilation, no matter whether make mrproper is used or not. It's quite time consuming and inefficient.


So why kernel make file couldn't do what a normal C make file could do? Or maybe it could but I got something wrong? 


If a complete compilation is inevitable, how to make it faster? I only know the -jX option.


Best regards
Michael
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20141120/a1a2b38d/attachment.html 

Why each time of kernel compilation must be done completely?

[Greg KH]

On Thu, Nov 20, 2014 at 10:05:04AM +0800, ??? wrote:
> Hi,
> 
> This problem has puzzled me for a long time. A normal C make file could parse
> the reliance and just regenerate the modified source file. But each time when I
> modified the kernel and recompile it, it does a complete compilation, no matter
> whether make mrproper is used or not. It's quite time consuming and
> inefficient.
> 
> So why kernel make file couldn't do what a normal C make file could do? Or
> maybe it could but I got something wrong? 

I think you are doing something wrong, unless you are changing a
configuration option, version of gcc, or are changing a .h file that all
files include.

What are you changing and doing that causes a full rebuild?  Can you
show us exactly?

> If a complete compilation is inevitable, how to make it faster? I only know the
> -jX option.

-jX is all you need.  You can use ccache if you are rebuilding the same
thing all the time, or distcc if you have other machines on your network
to take advantage of.

thanks,

greg k-h

Why each time of kernel compilation must be done completely?

[Valdis.Kletnieks at vt.edu]

On Wed, 19 Nov 2014 18:43:26 -0800, Greg KH said:

> I think you are doing something wrong, unless you are changing a
> configuration option, version of gcc, or are changing a .h file that all
> files include.

Actually, that reminds me of a patch I probably need to polish up and
make git bisect a lot less painful :)

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 848 bytes
Desc: not available
Url : http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20141119/a91e5abc/attachment.bin 

Why each time of kernel compilation must be done completely?

[秦弋戈]

Thank you very much for the help. It should be the fact that I changed the configuration before I make a compilation, which makes me thought that a complete compilation is inevitable.


Thanks again. : )


Regards
Michael






At 2014-11-20 10:43:26, "Greg KH" <greg@kroah.com> wrote:
>On Thu, Nov 20, 2014 at 10:05:04AM +0800, ??? wrote:
>> Hi,
>> 
>> This problem has puzzled me for a long time. A normal C make file could parse
>> the reliance and just regenerate the modified source file. But each time when I
>> modified the kernel and recompile it, it does a complete compilation, no matter
>> whether make mrproper is used or not. It's quite time consuming and
>> inefficient.
>> 
>> So why kernel make file couldn't do what a normal C make file could do? Or
>> maybe it could but I got something wrong? 
>
>I think you are doing something wrong, unless you are changing a
>configuration option, version of gcc, or are changing a .h file that all
>files include.
>
>What are you changing and doing that causes a full rebuild?  Can you
>show us exactly?
>
>> If a complete compilation is inevitable, how to make it faster? I only know the
>> -jX option.
>
>-jX is all you need.  You can use ccache if you are rebuilding the same
>thing all the time, or distcc if you have other machines on your network
>to take advantage of.
>
>thanks,
>
>greg k-h
>
>_______________________________________________
>Kernelnewbies mailing list
>Kernelnewbies at kernelnewbies.org
>http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20141121/fdad634e/attachment.html