almost time to run my kernel source cleaning scripts again?

almost time to run my kernel source cleaning scripts again?

[Robert P. J. Day]

  it's been a while so i was thinking of firing up my scripts to
peruse the source tree looking for oddities/inconsistencies/whatever
(unless someone else is already doing that).

  one example is looking for preprocessor tests of symbols that aren't
defined in any Kconfig file, and i can focus on any part of the source
tree.  for example, if i run that script against drivers/mmc, i get:

>>>>> MMC_MSM7X00A_RESUME_IN_WQ
drivers/mmc/host/msm_sdcc.c:1408:#ifdef CONFIG_MMC_MSM7X00A_RESUME_IN_WQ

telling me that a source file is testing a CONFIG variable that
doesn't exist.

  the scripts are available online (maybe i better test them first),
so anyone's free to run them against any part of the tree they want.

  running this script against drivers/staging is always good for
entertainment:

>>>>> CAVIUM_OCTEON_NUM_PACKET_BUFFERS
drivers/staging/octeon/ethernet-defines.h:36: *  CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS
drivers/staging/octeon/ethernet.c:56:#if defined(CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS) \
drivers/staging/octeon/ethernet.c:57:	&& CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS
drivers/staging/octeon/ethernet.c:58:int num_packet_buffers = CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS;
>>>>> CONSOLE_115200
drivers/staging/serial/68360serial.c:310:#elif defined(CONFIG_CONSOLE_115200)
>>>>> CONSOLE_19200
drivers/staging/serial/68360serial.c:308:#elif defined(CONFIG_CONSOLE_19200)
>>>>> CONSOLE_9600
drivers/staging/serial/68360serial.c:306:#if defined(CONFIG_CONSOLE_9600)
>>>>> FORCE_HARD_FLOAT
drivers/staging/rtl8192u/r8192U_core.c:27:#ifndef CONFIG_FORCE_HARD_FLOAT
... lots more ...

  i'll make sure the scripts work and maybe let them loose to find
stuff in a day or two.

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================

Re: almost time to run my kernel source cleaning scripts again?

[Jesper Juhl]

On Wed, 11 Apr 2012, Robert P. J. Day wrote:

> 
>   it's been a while so i was thinking of firing up my scripts to
> peruse the source tree looking for oddities/inconsistencies/whatever
> (unless someone else is already doing that).
> 
>   one example is looking for preprocessor tests of symbols that aren't
> defined in any Kconfig file, and i can focus on any part of the source
> tree.  for example, if i run that script against drivers/mmc, i get:
> 
> >>>>> MMC_MSM7X00A_RESUME_IN_WQ
> drivers/mmc/host/msm_sdcc.c:1408:#ifdef CONFIG_MMC_MSM7X00A_RESUME_IN_WQ
> 
> telling me that a source file is testing a CONFIG variable that
> doesn't exist.
> 
>   the scripts are available online 

I'd be interrested in taking a look at those. Got a URL where I can grab 
them?


[...]

-- 
Jesper Juhl <jj@chaosbits.net>       http://www.chaosbits.net/
Don't top-post http://www.catb.org/jargon/html/T/top-post.html
Plain text mails only, please.

Re: almost time to run my kernel source cleaning scripts again?

[Robert P. J. Day]

On Wed, 11 Apr 2012, Jesper Juhl wrote:

> On Wed, 11 Apr 2012, Robert P. J. Day wrote:
>
> >
> >   it's been a while so i was thinking of firing up my scripts to
> > peruse the source tree looking for oddities/inconsistencies/whatever
> > (unless someone else is already doing that).
> >
> >   one example is looking for preprocessor tests of symbols that aren't
> > defined in any Kconfig file, and i can focus on any part of the source
> > tree.  for example, if i run that script against drivers/mmc, i get:
> >
> > >>>>> MMC_MSM7X00A_RESUME_IN_WQ
> > drivers/mmc/host/msm_sdcc.c:1408:#ifdef CONFIG_MMC_MSM7X00A_RESUME_IN_WQ
> >
> > telling me that a source file is testing a CONFIG variable that
> > doesn't exist.
> >
> >   the scripts are available online
>
> I'd be interrested in taking a look at those. Got a URL where I can grab
> them?

  as long as you promise not to snicker at them, here are my most
"popular" ones:

http://www.crashcourse.ca/wiki/index.php/Kernel_cleanup_scripts

  i typically used the word "unused" to refer to variables defined in
Kconfig files that simply aren't referenced anymore -- not fatal in
any sense, just leftover detritus when people removed code but forgot
to take out the corresponding Kconfig entry.

  i use the word "badref" for instances where a CONFIG_ variable *is*
tested in a source file, but doesn't exist in a Kconfig file -- that's
typically more serious.  or sometimes, it's just a typo.

  just now, i ran the "find_badref_configs" script against the entire
drivers/ directory -- first few lines of output were:

>>>>> ACORNSCSI_CONSTANTS
drivers/scsi/arm/acornscsi.c:106:#undef CONFIG_ACORNSCSI_CONSTANTS
drivers/scsi/arm/acornscsi.c:413:#ifdef CONFIG_ACORNSCSI_CONSTANTS
drivers/scsi/arm/acornscsi.c:491:#ifdef CONFIG_ACORNSCSI_CONSTANTS
... snip ...

which i recognize by now since that's been there, like, *forever*.

  anyway, feel free to play; i might do some tweaking on those over
the next few days.

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================

Re: almost time to run my kernel source cleaning scripts again?

[David Daney]

On 04/11/2012 02:07 PM, Robert P. J. Day wrote:
>
>    it's been a while so i was thinking of firing up my scripts to
> peruse the source tree looking for oddities/inconsistencies/whatever
> (unless someone else is already doing that).
>
>    one example is looking for preprocessor tests of symbols that aren't
> defined in any Kconfig file, and i can focus on any part of the source
> tree.  for example, if i run that script against drivers/mmc, i get:
[...]
>
>>>>>> CAVIUM_OCTEON_NUM_PACKET_BUFFERS
> drivers/staging/octeon/ethernet-defines.h:36: *  CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS
> drivers/staging/octeon/ethernet.c:56:#if defined(CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS) \
> drivers/staging/octeon/ethernet.c:57:	&&  CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS
> drivers/staging/octeon/ethernet.c:58:int num_packet_buffers = CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS;

I am aware of these and am working on fixing them.

Thanks,
David Daney

Re: almost time to run my kernel source cleaning scripts again?

[Robert P. J. Day]

On Wed, 11 Apr 2012, David Daney wrote:

> On 04/11/2012 02:07 PM, Robert P. J. Day wrote:
> >
> >    it's been a while so i was thinking of firing up my scripts to
> > peruse the source tree looking for oddities/inconsistencies/whatever
> > (unless someone else is already doing that).
> >
> >    one example is looking for preprocessor tests of symbols that aren't
> > defined in any Kconfig file, and i can focus on any part of the source
> > tree.  for example, if i run that script against drivers/mmc, i get:
> [...]
> >
> > > > > > > CAVIUM_OCTEON_NUM_PACKET_BUFFERS
> > drivers/staging/octeon/ethernet-defines.h:36: *
> > CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS
> > drivers/staging/octeon/ethernet.c:56:#if
> > defined(CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS) \
> > drivers/staging/octeon/ethernet.c:57:	&&
> > CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS
> > drivers/staging/octeon/ethernet.c:58:int num_packet_buffers =
> > CONFIG_CAVIUM_OCTEON_NUM_PACKET_BUFFERS;
>
> I am aware of these and am working on fixing them.

  sorry, i wasn't trying to pick on anyone, it was just a random
example.

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================