On Tue, Sep 30, 2014 at 12:14:29AM +0300, Dan Carpenter wrote: > On Mon, Sep 29, 2014 at 08:42:23PM +0200, Silvan Jegen wrote: > > > Currently it only stores things like "s32min-s32max[$0 + $1]" but it > > > needs to be able to store "s32min-s32max[$0->len + $1]". > > > > That means I will just go and rerun smatch on the kernel then! :) > > > > You need to build the database first, to experience any of the new > goodness. The command is: > > ~/progs/smatch/devel/smatch_scripts/build_kernel_data.sh > > It takes a few hours some 15-20GB but building the DB is so worth it > though. The results will be more accurate since it has cross function > information. You can also use the > ~/progs/smatch/devel/smatch_data/db/smdb.py program to see how a > function is called. Also where a data type is allocated. Which kinds > of values are stored in a particular struct member. Loads of > interesting stuff. :) > > I don't think I've ever had any real feedback from people using the > cross function db. Let me know how it goes. I ran the build_kernel_data.sh script from within the linux source tree and ended up with a 7GB warns.txt file. I can run smdb.py in the same directory as the warns.txt file adding a random kernel function name as an argument but the script does not seem to return anything. If you point me to some information about the db-usage I will experiment some more! Cheers, Silvan -- To unsubscribe from this list: send the line "unsubscribe smatch" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
