.o.cmd files wanted, static analysis

.o.cmd files wanted, static analysis

[Peter T. Breuer]

(was sent to Linux Testing Project list, but no answer there yet).


G'day all

I'm looking for a set of .*.o.cmd files for the 2.6 kernel. These are
generated when you compile the kernel. I have about 1600 of them from my
standard compilation choices, but I want the full 6000 odd so that I can
run static analysis tests.

See ftp://oboe.it.uc3m.es/pub/Programs/c-1.2.*.tgz

The currently enabled tests are for sleep under spinlock, double spinlock,
and access to kfree'd memory. I'll add more according to suggestions
received. About 3 errors in each class are detected for every 1000
kernel source files, with about three times that many false alarms
raised. It takes a few seconds per file, at about 10000 expanded lines
of code per second.

Does somebody out there have a pretty complete kernel compilation,
mainly all modules (though I don't care!), and could let me have their
 .*.o.cmd files? In fact, just the FIRST LINE of them.
 
That first line gives me the compile flags used.

Thanks.

Peter

Re: .o.cmd files wanted, static analysis

[Kyle Moffett]

On Sep 15, 2005, at 19:50:29, Peter T. Breuer wrote:
> Does somebody out there have a pretty complete kernel compilation,  
> mainly all modules (though I don't care!), and could let me have  
> their .*.o.cmd files? In fact, just the FIRST LINE of them.

doesn't "make allmodconfig" work for you?  That should generate a  
config with everything possible included as modules, and then you can  
compile that to generate .cmd files.


Cheers,
Kyle Moffett

--
There are two ways of constructing a software design. One way is to  
make it so simple that there are obviously no deficiencies. And the  
other way is to make it so complicated that there are no obvious  
deficiencies.  The first method is far more difficult.
   -- C.A.R. Hoare

Re: .o.cmd files wanted, static analysis

[Peter T. Breuer]

"Also sprach Kyle Moffett:"
> On Sep 15, 2005, at 19:50:29, Peter T. Breuer wrote:
> > Does somebody out there have a pretty complete kernel compilation,  
> > mainly all modules (though I don't care!), and could let me have  
> > their .*.o.cmd files? In fact, just the FIRST LINE of them.
> 
> doesn't "make allmodconfig" work for you?  That should generate a  

Thanks. I didn't know about allconfig. That sounds good.. But no,
it wouldn't exactly work for me just like that - I don't have space
to do the actual kernel compilation on any machine (nor time to do it,
since the fastest machine is my laptop at 1.13GHz, and that suffers from
overheating problems - the next fastest machine is a dual 550MHz, and I
will start that churning but space is tight everywhere  and I am pretty
sure that there is no chance to get it to go to a finish as is).

> config with everything possible included as modules, and then you can  
> compile that to generate .cmd files.

Has anyone already done it? I was hoping there was somebody running a
big compile every day in order to check consistency. Then I could
simply "borrow" their files!

Failing that, I'll try setting CC to something like FOO, and seeing if
I can generate the .o.cmd files without the compilation actually being
done. Hmmm ...

Peter

Re: .o.cmd files wanted, static analysis

[Peter T. Breuer]

"Also sprach Frank Sorenson:"
> Here are the *.o.cmd files from a "make allmodconfig" on my box.  Hope
> this helps.

That looks great. Thanks!

Peter

Re: .o.cmd files wanted, static analysis

[Peter T. Breuer]

"Also sprach Lion Vollnhals:"
> Hi!
> 
> Here are mine *.o.cmd files.

Thanks! I should be able to work with that.

Peter