Re: [RFC PATCH] Fix: module signature vs tracepoints: add new TAINT_UNSIGNED_MODULE

[Arend van Spriel <arend@broadcom.com>]

On 02/13/2014 04:44 PM, Steven Rostedt wrote:
> On Thu, 13 Feb 2014 10:36:35 -0500
> fche@redhat.com (Frank Ch. Eigler) wrote:
> 
>>
>> rostedt wrote:
>>
>>> [...]
>>> Oh! You are saying that if the kernel only *supports* signed modules,
>>> and you load a module that is not signed, it will taint the kernel?
>>
>> Yes: this is the default for several distros.
>>
> 
> Rusty, Ingo,
> 
> This looks like a bug to me, as it can affect even in-tree kernel
> modules. If you have a kernel that supports signed modules, and you
> modify a module, recompile it, apply it, since it is no longer signed,
> then it sounds like we just tainted it. Worse yet, we just disabled any
> tracepoints on that module, which means it is even harder to debug that
> module (if that's the reason you recompiled it in the first place).

When I stumbled upon this issue a while ago on Fedora 19 I built my
kernel rpm packages which generates a signature key (.priv and .x509),
which I kept safe with the kernel headers. When building recompiling
modules I refer to it with MODSECKEY and MODPUBKEY, ie.

$ make MODSECKEY=bla MODPUBKEY=duh \
	M=drivers/net/wireless/brcm80211  modules

Or sign it manually using the sign-file perl script:

mod_sign_cmd = perl $(srctree)/scripts/sign-file \
	$(CONFIG_MODULE_SIG_HASH) $(MODSECKEY) $(MODPUBKEY)

Of course I could disable signed modules while building a new kernel,
but I was in it for the ride (I had better ones) ;-)

Gr. AvS

> -- Steve
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>