On Mon, Jan 20, 2020 at 02:43:45PM +0000, Ido Schimmel wrote:
> On Mon, Jan 20, 2020 at 05:37:12PM +0300, Dan Carpenter wrote:
> > Hello Ido Schimmel,
> >
> > The patch 48bb9eb47b27: "netdevsim: fib: Add dummy implementation for
> > FIB offload" from Jan 14, 2020, leads to the following static checker
> > warning:
> >
> > drivers/net/netdevsim/fib.c:663 nsim_fib6_rt_insert()
> > error: 'fib6_rt' dereferencing possible ERR_PTR()
> >
> > drivers/net/netdevsim/fib.c
> > 460 nsim_fib6_rt_create(struct nsim_fib_data *data,
> > 461 struct fib6_entry_notifier_info *fen6_info)
> > 462 {
> > 463 struct fib6_info *iter, *rt = fen6_info->rt;
> > 464 struct nsim_fib6_rt *fib6_rt;
> > 465 int i = 0;
> > 466 int err;
> > 467
> > 468 fib6_rt = kzalloc(sizeof(*fib6_rt), GFP_ATOMIC);
> > 469 if (!fib6_rt)
> > 470 return NULL;
> > ^^^^^^^^^^^
>
> Dan, thank you very much for the report. It is already fixed in net-next
> thanks to Eric Dumazet:
> https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=41cdc741048b0d04604c02aad9ec19f7d9130b70
>
> Can you share the static checker you used so I can avoid these mistakes
> in the future? Alternatively, is it possible to register development
> trees for this service like with the kbuild robot?
>
> Thanks in advance.
It's a Smatch warning but you need to rebuild the kernel data.
~/path/to/smatch/smatch_scripts/build_kernel_data.sh
It takes a few hours (maybe four) to rebuild.
regards,
dan carpenter
