net/psp/psp_nl.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
From: Wei Wang <weibunny@fb.com>
sparse reports:
net/psp/psp_nl.c:513:13: sparse: sparse: cast removes address space
'__rcu' of expression
cmpxchg() returns typeof(*ptr) and its internal casts strip the __rcu
annotation. Wrap it in unrcu_pointer(), the documented way to use an
__rcu pointer with xchg() and friends.
This was introduced by commit 06c2dce2d0f6 ("psp: add new netlink cmd
for dev-assoc and dev-disassoc").
No functional change intended.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202608080910.l9KvOH7O-lkp@intel.com/
Signed-off-by: Wei Wang <weibunny@fb.com>
---
net/psp/psp_nl.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/psp/psp_nl.c b/net/psp/psp_nl.c
index 43b066353c65..f91665748dde 100644
--- a/net/psp/psp_nl.c
+++ b/net/psp/psp_nl.c
@@ -533,7 +533,8 @@ int psp_nl_dev_assoc_doit(struct sk_buff *skb, struct genl_info *info)
}
/* Check if device is already associated with a PSP device */
- if (cmpxchg(&assoc_dev->psp_dev, NULL, RCU_INITIALIZER(psd))) {
+ if (unrcu_pointer(cmpxchg(&assoc_dev->psp_dev, NULL,
+ RCU_INITIALIZER(psd)))) {
NL_SET_ERR_MSG(info->extack,
"Device already associated with a PSP device");
err = -EBUSY;
--
2.53.0-Meta
On 8/13/26 3:34 PM, Wei Wang wrote:
> From: Wei Wang <weibunny@fb.com>
>
> sparse reports:
>
> net/psp/psp_nl.c:513:13: sparse: sparse: cast removes address space
> '__rcu' of expression
>
> cmpxchg() returns typeof(*ptr) and its internal casts strip the __rcu
> annotation. Wrap it in unrcu_pointer(), the documented way to use an
> __rcu pointer with xchg() and friends.
>
> This was introduced by commit 06c2dce2d0f6 ("psp: add new netlink cmd
> for dev-assoc and dev-disassoc").
>
> No functional change intended.
>
> Reported-by: kernel test robot <lkp@intel.com>
> Closes: https://lore.kernel.org/oe-kbuild-all/202608080910.l9KvOH7O-lkp@intel.com/
> Signed-off-by: Wei Wang <weibunny@fb.com>
Commits targeting the net tree need a fixes tag.
On Thu, Aug 13, 2026 at 12:57 PM Daniel Zahka <daniel.zahka@gmail.com> wrote:
>
> >
>
>
> On 8/13/26 3:34 PM, Wei Wang wrote:
> > From: Wei Wang <weibunny@fb.com>
> >
> > sparse reports:
> >
> > net/psp/psp_nl.c:513:13: sparse: sparse: cast removes address space
> > '__rcu' of expression
> >
> > cmpxchg() returns typeof(*ptr) and its internal casts strip the __rcu
> > annotation. Wrap it in unrcu_pointer(), the documented way to use an
> > __rcu pointer with xchg() and friends.
> >
> > This was introduced by commit 06c2dce2d0f6 ("psp: add new netlink cmd
> > for dev-assoc and dev-disassoc").
> >
> > No functional change intended.
> >
> > Reported-by: kernel test robot <lkp@intel.com>
> > Closes: https://lore.kernel.org/oe-kbuild-all/202608080910.l9KvOH7O-lkp@intel.com/
> > Signed-off-by: Wei Wang <weibunny@fb.com>
> Commits targeting the net tree need a fixes tag.
This is a sparse warning fix, my understanding is that it's not a
functional fix and therefore should not use a Fixes tag. Please let me
know if my understanding is wrong.
On Thu, 13 Aug 2026 13:37:23 -0700 Wei Wang wrote: > > > Reported-by: kernel test robot <lkp@intel.com> > > > Closes: https://lore.kernel.org/oe-kbuild-all/202608080910.l9KvOH7O-lkp@intel.com/ > > > Signed-off-by: Wei Wang <weibunny@fb.com> > > Commits targeting the net tree need a fixes tag. > > This is a sparse warning fix, my understanding is that it's not a > functional fix and therefore should not use a Fixes tag. Please let me > know if my understanding is wrong. I mentioned "no Fixes tag needed" in internal review, I should have been more clear that this implies net-next, sorry. No need to repost - we'll route it to net-next when applying.
On Fri, Aug 14, 2026 at 10:07 AM Jakub Kicinski <kuba@kernel.org> wrote: > > > > On Thu, 13 Aug 2026 13:37:23 -0700 Wei Wang wrote: > > > > Reported-by: kernel test robot <lkp@intel.com> > > > > Closes: https://lore.kernel.org/oe-kbuild-all/202608080910.l9KvOH7O-lkp@intel.com/ > > > > Signed-off-by: Wei Wang <weibunny@fb.com> > > > Commits targeting the net tree need a fixes tag. > > > > This is a sparse warning fix, my understanding is that it's not a > > functional fix and therefore should not use a Fixes tag. Please let me > > know if my understanding is wrong. > > I mentioned "no Fixes tag needed" in internal review, I should have > been more clear that this implies net-next, sorry. > > No need to repost - we'll route it to net-next when applying. Thanks! Noted for next time.
© 2016 - 2026 Red Hat, Inc.