[RFC PATCH 2/7] static_call: deal with unexported keys without cluttering up the API
Ard Biesheuvel
ardb at kernel.org
Tue Nov 9 10:53:33 PST 2021
On Tue, 9 Nov 2021 at 19:49, Peter Zijlstra <peterz at infradead.org> wrote:
>
> On Tue, Nov 09, 2021 at 05:45:44PM +0100, Ard Biesheuvel wrote:
>
> > diff --git a/include/linux/static_call_types.h b/include/linux/static_call_types.h
> > index 5a00b8b2cf9f..0bb36294cce7 100644
> > --- a/include/linux/static_call_types.h
> > +++ b/include/linux/static_call_types.h
> > @@ -32,15 +32,20 @@
> > struct static_call_site {
> > s32 addr;
> > s32 key;
> > + s32 tramp;
> > };
>
> I can't say I'm thrilled at growing this thing, but the cleanup is nice.
> Perhaps we can increase alignment on struct static_call_key and instead
> frob it in .key still?
>
This is already a place-relative field, and one points into the data
section and the other into text. So I don't see how we can squeeze
enough bits out of it to make this fit.
> >
> > #define DECLARE_STATIC_CALL(name, func) \
> > - extern struct static_call_key STATIC_CALL_KEY(name); \
> > + extern __weak struct static_call_key STATIC_CALL_KEY(name); \
> > extern typeof(func) STATIC_CALL_TRAMP(name);
>
> I'm a little bit confused on how this actually works. What does a __weak
> extern data symbol do?
>
> A __weak function definition would create a module local instance of the
> function barring a strong override.
>
> But what does a __weak extern do?
>
It is simply a reference that is permitted to remain undefined.
That is why (in another patch) I do sth like
extern __weak foo;
if (&foo)
... use foo
else
... use sth else
> > diff --git a/kernel/static_call.c b/kernel/static_call.c
> > index 43ba0b1e0edb..360cc3cd0fbf 100644
> > --- a/kernel/static_call.c
> > +++ b/kernel/static_call.c
> > @@ -366,18 +366,18 @@ static int static_call_add_module(struct module *mod)
> > * means modules are allowed to call static_call_update() on
> > * it.
> > *
> > - * Otherwise, the key isn't exported, and 'addr' points to the
> > + * Otherwise, the key isn't exported, and 'tramp' points to the
> > * trampoline so we need to lookup the key.
> > *
> > * We go through this dance to prevent crazy modules from
> > * abusing sensitive static calls.
> > */
> > - if (!kernel_text_address(addr))
> > + if (addr)
> > continue;
>
> This seems to imply that the __weak extern symbol gets resolved to 0 at
> module link time.
>
Yes, if the referenced symbol is not exported, it just uses 0x0 for its value.
> >
> > - key = tramp_key_lookup(addr);
> > + key = tramp_key_lookup((unsigned long)offset_to_ptr(&site->tramp));
> > if (!key) {
> > - pr_warn("Failed to fixup __raw_static_call() usage at: %ps\n",
> > + pr_warn("Failed to fixup static_call() usage at: %ps\n",
> > static_call_addr(site));
> > return -EINVAL;
>
> > }
More information about the linux-arm-kernel
mailing list