| Message ID | 20211108101157.15189-40-bp@alien8.de (mailing list archive) |
|---|---|
| State | Not Applicable |
| Delegated to: | Herbert Xu |
| Headers | show |
| Series | None | expand |
On Mon, Nov 08, 2021 at 11:11:54AM +0100, Borislav Petkov wrote: > From: Borislav Petkov <bp@suse.de> > > Avoid homegrown notifier registration checks. > > No functional changes. > Hi Borislav, > Signed-off-by: Borislav Petkov <bp@suse.de> > Cc: linux-crypto@vger.kernel.org > --- > drivers/crypto/ccree/cc_fips.c | 4 +++- > 1 file changed, 3 insertions(+), 1 deletion(-) > > diff --git a/drivers/crypto/ccree/cc_fips.c b/drivers/crypto/ccree/cc_fips.c > index 702aefc21447..de842da7d51c 100644 > --- a/drivers/crypto/ccree/cc_fips.c > +++ b/drivers/crypto/ccree/cc_fips.c > @@ -146,7 +146,9 @@ int cc_fips_init(struct cc_drvdata *p_drvdata) > tasklet_init(&fips_h->tasklet, fips_dsr, (unsigned long)p_drvdata); > fips_h->drvdata = p_drvdata; > fips_h->nb.notifier_call = cc_ree_fips_failure; > - atomic_notifier_chain_register(&fips_fail_notif_chain, &fips_h->nb); > + > + if (atomic_notifier_chain_register(&fips_fail_notif_chain, &fips_h->nb)) > + pr_warn("Failure notifier already registered\n"); > Looking at the implementation of atomic_notifier_chain_register() and its internal helper down below, I can see that atomic_notifier_chain_register() ALWAYS return 0 (O_o) and anyway there is a WARN() in the notifiier core already to alert of double registrations. What is the aim of this patch ? It's not clear from the commit message what are you trying to achieve. Am I missing something ? (Is it to circumvent some static checker to avoid this false positive ? just guessing...) Thanks, Cristian -----8<---------- static int notifier_chain_register(struct notifier_block **nl, struct notifier_block *n) { while ((*nl) != NULL) { if (unlikely((*nl) == n)) { WARN(1, "double register detected"); return 0; } if (n->priority > (*nl)->priority) break; nl = &((*nl)->next); } n->next = *nl; rcu_assign_pointer(*nl, n); return 0; } /** * atomic_notifier_chain_register - Add notifier to an atomic notifier chain * @nh: Pointer to head of the atomic notifier chain * @n: New entry in notifier chain * * Adds a notifier to an atomic notifier chain. * * Currently always returns zero. */ int atomic_notifier_chain_register(struct atomic_notifier_head *nh, struct notifier_block *n) { unsigned long flags; int ret; spin_lock_irqsave(&nh->lock, flags); ret = notifier_chain_register(&nh->head, n); spin_unlock_irqrestore(&nh->lock, flags); return ret; } EXPORT_SYMBOL_GPL(atomic_notifier_chain_register); ------
On Fri, Nov 12, 2021 at 06:32:33PM +0000, Cristian Marussi wrote: > Looking at the implementation of atomic_notifier_chain_register() and > its internal helper down below, I can see that atomic_notifier_chain_register() > ALWAYS return 0 (O_o) and anyway there is a WARN() in the notifiier core already > to alert of double registrations. You can ignore it - we agreed (at least no one has complained against it yet :)) on doing this instead: https://lore.kernel.org/r/YYk1xi3eJdMJdjHC@zn.tnic
diff --git a/drivers/crypto/ccree/cc_fips.c b/drivers/crypto/ccree/cc_fips.c index 702aefc21447..de842da7d51c 100644 --- a/drivers/crypto/ccree/cc_fips.c +++ b/drivers/crypto/ccree/cc_fips.c @@ -146,7 +146,9 @@ int cc_fips_init(struct cc_drvdata *p_drvdata) tasklet_init(&fips_h->tasklet, fips_dsr, (unsigned long)p_drvdata); fips_h->drvdata = p_drvdata; fips_h->nb.notifier_call = cc_ree_fips_failure; - atomic_notifier_chain_register(&fips_fail_notif_chain, &fips_h->nb); + + if (atomic_notifier_chain_register(&fips_fail_notif_chain, &fips_h->nb)) + pr_warn("Failure notifier already registered\n"); cc_tee_handle_fips_error(p_drvdata);