any value in running my kernel scanning scripts before release?

any value in running my kernel scanning scripts before release?

[Robert P. J. Day]

  normally i wait for the merge window, then run my kernel scanning
scripts and post results.  any interest in me doing that earlier?  i
just did a quick test and found a number of incongruities like, for
example:

$ grep -r "config QT2160" *
drivers/input/keyboard/Kconfig:config QT2160
$

$ grep -r KEYBOARD_QT2160 *
drivers/input/keyboard/Makefile:obj-$(CONFIG_KEYBOARD_QT2160) += qt2160.o
$

  which is a fairly obvious mismatch between the config variable and
the test, yes?

  anyway, i'm open to suggestions.

rday

-- 

========================================================================
Robert P. J. Day                               Waterloo, Ontario, CANADA

        Top-notch, inexpensive online Linux/OSS/kernel courses
                        http://crashcourse.ca

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

Re: any value in running my kernel scanning scripts before release?

[Michal Marek]

On 07/15/2010 01:32 PM, Robert P. J. Day wrote:
> 
>   normally i wait for the merge window, then run my kernel scanning
> scripts and post results.  any interest in me doing that earlier?  i
> just did a quick test and found a number of incongruities like, for
> example:
> 
> $ grep -r "config QT2160" *
> drivers/input/keyboard/Kconfig:config QT2160
> $
> 
> $ grep -r KEYBOARD_QT2160 *
> drivers/input/keyboard/Makefile:obj-$(CONFIG_KEYBOARD_QT2160) += qt2160.o
> $

These are plain bugs, so it's IMO OK to send fixes for these at any time
before the final release. In the worst case, it will be merged after the
final release.

just my two cents,
Michal

Re: any value in running my kernel scanning scripts before release?

[Robert P. J. Day]

On Thu, 15 Jul 2010, Michal Marek wrote:

> On 07/15/2010 01:32 PM, Robert P. J. Day wrote:
> >
> >   normally i wait for the merge window, then run my kernel scanning
> > scripts and post results.  any interest in me doing that earlier?  i
> > just did a quick test and found a number of incongruities like, for
> > example:
> >
> > $ grep -r "config QT2160" *
> > drivers/input/keyboard/Kconfig:config QT2160
> > $
> >
> > $ grep -r KEYBOARD_QT2160 *
> > drivers/input/keyboard/Makefile:obj-$(CONFIG_KEYBOARD_QT2160) += qt2160.o
> > $
>
> These are plain bugs, so it's IMO OK to send fixes for these at any
> time before the final release. In the worst case, it will be merged
> after the final release.

  i wasn't suggesting that was the only example, only that it was an
obvious one.  there are others.  just for fun, here's the entire list
of references specifically from *Makefiles* to Kconfig variables that
don't exist (for whatever reason):

===== ATHEROS_AR71XX =====
./drivers/net/wireless/ath/ath9k/Makefile:ath9k-$(CONFIG_ATHEROS_AR71XX) += ahb.o
===== KEYBOARD_QT2160 =====
./drivers/input/keyboard/Makefile:obj-$(CONFIG_KEYBOARD_QT2160)		+= qt2160.o
===== PCMCIA_IOCTL =====
./drivers/pcmcia/Makefile:pcmcia-$(CONFIG_PCMCIA_IOCTL)			+= pcmcia_ioctl.o
===== USB_ETH_FUNCTIONFS =====
./drivers/usb/gadget/Makefile:obj-$(CONFIG_USB_ETH_FUNCTIONFS)	+= g_eth_ffs.o

  in other words, none of that content is ever going to get compiled.
mercifully, there are only four examples of that left in the tree.
used to be a lot more.  clean up whatever catches your eye.

rday

p.s.  i cc'ed greg kh since i suspect he'll want to do something about
that last one.  i need sleep.

-- 

========================================================================
Robert P. J. Day                               Waterloo, Ontario, CANADA

        Top-notch, inexpensive online Linux/OSS/kernel courses
                        http://crashcourse.ca

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

Re: any value in running my kernel scanning scripts before release?

[Greg KH]

On Thu, Jul 15, 2010 at 06:50:20PM -0400, Robert P. J. Day wrote:
> On Thu, 15 Jul 2010, Michal Marek wrote:
> 
> > On 07/15/2010 01:32 PM, Robert P. J. Day wrote:
> > >
> > >   normally i wait for the merge window, then run my kernel scanning
> > > scripts and post results.  any interest in me doing that earlier?  i
> > > just did a quick test and found a number of incongruities like, for
> > > example:
> > >
> > > $ grep -r "config QT2160" *
> > > drivers/input/keyboard/Kconfig:config QT2160
> > > $
> > >
> > > $ grep -r KEYBOARD_QT2160 *
> > > drivers/input/keyboard/Makefile:obj-$(CONFIG_KEYBOARD_QT2160) += qt2160.o
> > > $
> >
> > These are plain bugs, so it's IMO OK to send fixes for these at any
> > time before the final release. In the worst case, it will be merged
> > after the final release.
> 
>   i wasn't suggesting that was the only example, only that it was an
> obvious one.  there are others.  just for fun, here's the entire list
> of references specifically from *Makefiles* to Kconfig variables that
> don't exist (for whatever reason):
> 
> ===== ATHEROS_AR71XX =====
> ./drivers/net/wireless/ath/ath9k/Makefile:ath9k-$(CONFIG_ATHEROS_AR71XX) += ahb.o
> ===== KEYBOARD_QT2160 =====
> ./drivers/input/keyboard/Makefile:obj-$(CONFIG_KEYBOARD_QT2160)		+= qt2160.o
> ===== PCMCIA_IOCTL =====
> ./drivers/pcmcia/Makefile:pcmcia-$(CONFIG_PCMCIA_IOCTL)			+= pcmcia_ioctl.o
> ===== USB_ETH_FUNCTIONFS =====
> ./drivers/usb/gadget/Makefile:obj-$(CONFIG_USB_ETH_FUNCTIONFS)	+= g_eth_ffs.o
> 
>   in other words, none of that content is ever going to get compiled.
> mercifully, there are only four examples of that left in the tree.
> used to be a lot more.  clean up whatever catches your eye.
> 
> rday
> 
> p.s.  i cc'ed greg kh since i suspect he'll want to do something about
> that last one.  i need sleep.

Ah, ick, that usb functionfs one is a real bug, nice catch.  I too have
been getting no sleep, and will just wait for you to send me a patch :)

The joys of conference season...

thanks,

greg k-h

Re: any value in running my kernel scanning scripts before release?

[Robert P. J. Day]

On Thu, 15 Jul 2010, Greg KH wrote:

> On Thu, Jul 15, 2010 at 06:50:20PM -0400, Robert P. J. Day wrote:

> > ===== USB_ETH_FUNCTIONFS =====
> > ./drivers/usb/gadget/Makefile:obj-$(CONFIG_USB_ETH_FUNCTIONFS)	+= g_eth_ffs.o
>
> Ah, ick, that usb functionfs one is a real bug, nice catch.  I too
> have been getting no sleep, and will just wait for you to send me a
> patch :)

  just submitted patch on linux USB list ...

rday

-- 

========================================================================
Robert P. J. Day                               Waterloo, Ontario, CANADA

        Top-notch, inexpensive online Linux/OSS/kernel courses
                        http://crashcourse.ca

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