Building vmlinux without date/timestamp bits

Building vmlinux without date/timestamp bits

[John Daiker]

I'm working on various trivial patches (mainly whitespace and 
checkpatch.pl cleanups).

I read quite some time ago that it's best to md5sum the vmlinux file 
before and after these trivial patches are applied (to ensure that no 
executable code has changed, modules on a binary level are identical, etc).

So far, I have no problems producing identical .o and .ko files with my 
trivial patches.  My problem lies in the resulting vmlinux and vmlinux.o 
file, though.

I know the files are different due to date/timestamping the kernel, but 
I'm not sure where this happens (and how I can prevent it).  Are there 
pieces of code and/or Makefile that I can comment out so that the 
resulting vmlinux file isn't timestamped (and thus would be an exactly 
identical before and after).  I know it can be done... I've done it 
before (years ago), but seem to have forgotten the trick.

Thanks for the help!

John Daiker

Building vmlinux without date/timestamp bits

[John Daiker]

I'm working on various trivial patches (mainly whitespace and 
checkpatch.pl cleanups).

I read quite some time ago that it's best to md5sum the vmlinux file 
before and after these trivial patches are applied (to ensure that no 
executable code has changed, modules on a binary level are identical, etc).

So far, I have no problems producing identical .o and .ko files with my 
trivial patches.  My problem lies in the resulting vmlinux and vmlinux.o 
file, though.

I know the files are different due to date/timestamping the kernel, but 
I'm not sure where this happens (and how I can prevent it).  Are there 
pieces of code and/or Makefile that I can comment out so that the 
resulting vmlinux file isn't timestamped (and thus would be an exactly 
identical before and after).  I know it can be done... I've done it 
before (years ago), but seem to have forgotten the trick.

Thanks for the help!

John Daiker

Re: Building vmlinux without date/timestamp bits

[Peter Huewe]

Hi John,

check include/linux/compile.h and scripts/mkcompile_h

In compile.h is exactly your timestamped information which gets built into the 
kernel (cmp. output of uname -s)

If the include/linux/compile.h is not present or needs an update it is 
generated using the scripts/mkcompile_h script.

--> You can either supply an identical compile.h for both or change the 
behavior of the mkcompile_h script.

Best regards,
Peter




Am Samstag 13 Juni 2009 00:15:26 schrieb John Daiker:
> I'm working on various trivial patches (mainly whitespace and
> checkpatch.pl cleanups).
>
> I read quite some time ago that it's best to md5sum the vmlinux file
> before and after these trivial patches are applied (to ensure that no
> executable code has changed, modules on a binary level are identical, etc).
>
> So far, I have no problems producing identical .o and .ko files with my
> trivial patches.  My problem lies in the resulting vmlinux and vmlinux.o
> file, though.
>
> I know the files are different due to date/timestamping the kernel, but
> I'm not sure where this happens (and how I can prevent it).  Are there
> pieces of code and/or Makefile that I can comment out so that the
> resulting vmlinux file isn't timestamped (and thus would be an exactly
> identical before and after).  I know it can be done... I've done it
> before (years ago), but seem to have forgotten the trick.
>
> Thanks for the help!
>
> John Daiker
> --
> To unsubscribe from this list: send the line "unsubscribe kernel-janitors"
> in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Building vmlinux without date/timestamp bits

[Peter Huewe]

Hi John,

check include/linux/compile.h and scripts/mkcompile_h

In compile.h is exactly your timestamped information which gets built into the 
kernel (cmp. output of uname -s)

If the include/linux/compile.h is not present or needs an update it is 
generated using the scripts/mkcompile_h script.

--> You can either supply an identical compile.h for both or change the 
behavior of the mkcompile_h script.

Best regards,
Peter




Am Samstag 13 Juni 2009 00:15:26 schrieb John Daiker:
> I'm working on various trivial patches (mainly whitespace and
> checkpatch.pl cleanups).
>
> I read quite some time ago that it's best to md5sum the vmlinux file
> before and after these trivial patches are applied (to ensure that no
> executable code has changed, modules on a binary level are identical, etc).
>
> So far, I have no problems producing identical .o and .ko files with my
> trivial patches.  My problem lies in the resulting vmlinux and vmlinux.o
> file, though.
>
> I know the files are different due to date/timestamping the kernel, but
> I'm not sure where this happens (and how I can prevent it).  Are there
> pieces of code and/or Makefile that I can comment out so that the
> resulting vmlinux file isn't timestamped (and thus would be an exactly
> identical before and after).  I know it can be done... I've done it
> before (years ago), but seem to have forgotten the trick.
>
> Thanks for the help!
>
> John Daiker
> --
> To unsubscribe from this list: send the line "unsubscribe kernel-janitors"
> in the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html