On Tue, 2020-06-02 at 12:38 +0800, kbuild test robot wrote: > tree: https://github.com/ceph/ceph-client.git testing > head: 9c4e2af200ffe80d2f917b75f92ab57f0a091a17 > commit: 7833323363233c75fd8d10b5ceefbb9515cb3e32 [9/30] ceph: don't take i_ceph_lock in handle_cap_import > config: microblaze-randconfig-s031-20200602 (attached as .config) > compiler: microblaze-linux-gcc (GCC) 9.3.0 > reproduce: > # apt-get install sparse > # sparse version: v0.6.1-243-gc100a7ab-dirty > git checkout 7833323363233c75fd8d10b5ceefbb9515cb3e32 > # save the attached .config to linux build tree > make W=1 C=1 ARCH=microblaze CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' > > If you fix the issue, kindly add following tag as appropriate > Reported-by: kbuild test robot <lkp@xxxxxxxxx> > > > sparse warnings: (new ones prefixed by >>) > > fs/ceph/caps.c:3443:9: sparse: sparse: context imbalance in 'handle_cap_grant' - wrong count at exit > > > include/linux/spinlock.h:353:9: sparse: sparse: context imbalance in 'ceph_handle_caps' - unexpected unlock > > vim +/ceph_handle_caps +353 include/linux/spinlock.h > > de8f5e4f2dc1f0 Peter Zijlstra 2020-03-21 350 > 3490565b633c70 Denys Vlasenko 2015-07-13 351 static __always_inline void spin_lock(spinlock_t *lock) > c2f21ce2e31286 Thomas Gleixner 2009-12-02 352 { > c2f21ce2e31286 Thomas Gleixner 2009-12-02 @353 raw_spin_lock(&lock->rlock); > c2f21ce2e31286 Thomas Gleixner 2009-12-02 354 } > c2f21ce2e31286 Thomas Gleixner 2009-12-02 355 > > :::::: The code at line 353 was first introduced by commit > :::::: c2f21ce2e31286a0a32f8da0a7856e9ca1122ef3 locking: Implement new raw_spinlock > > :::::: TO: Thomas Gleixner <tglx@xxxxxxxxxxxxx> > :::::: CC: Thomas Gleixner <tglx@xxxxxxxxxxxxx> > > --- > 0-DAY CI Kernel Test Service, Intel Corporation > https://lists.01.org/hyperkitty/list/kbuild-all@xxxxxxxxxxxx I think this is a false-positive, but I'd welcome someone else to sanity check me here (Ilya?). My sparse says something a little different: fs/ceph/caps.c:4052:26: warning: context imbalance in 'ceph_handle_caps' - unexpected unlock ...and I get a similar warning when I move to the commit ahead of 7833323363233c as well. All that patch does is push the i_ceph_lock acquisition from handle_cap_import to ceph_handle_caps. I don't see how it would have introduced a locking imbalance, but the locking in this code really is quite confusing, so please do point it out if I'm wrong here. Thanks, -- Jeff Layton <jlayton@xxxxxxxxxx>
