CC: riscv, is there something special about cmpxchg() on that arch?
We don't see it on x86, and I don't see anything special in the code
for riscv, at least on 5.17.

On Mon, 28 Mar 2022 14:28:34 +0800 kernel test robot wrote:
> tree:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> head:   ae085d7f9365de7da27ab5c0d16b12d51ea7fca9
> commit: 625788b5844511cf4c30cffa7fa0bc3a69cebc82 net: add per-cpu storage and net->core_stats
> date:   2 weeks ago
> config: riscv-randconfig-s032-20220328 (https://download.01.org/0day-ci/archive/20220328/202203281436.hsNgYWWu-lkp@intel.com/config)
> compiler: riscv64-linux-gcc (GCC) 11.2.0
> reproduce:
>         wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
>         chmod +x ~/bin/make.cross
>         # apt-get install sparse
>         # sparse version: v0.6.4-dirty
>         # https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=625788b5844511cf4c30cffa7fa0bc3a69cebc82
>         git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
>         git fetch --no-tags linus master
>         git checkout 625788b5844511cf4c30cffa7fa0bc3a69cebc82
>         # save the config file to linux build tree
>         mkdir build_dir
>         COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-11.2.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' O=build_dir ARCH=riscv SHELL=/bin/bash net/core/
> 
> If you fix the issue, kindly add following tag as appropriate
> Reported-by: kernel test robot <lkp@intel.com>
> 
> 
> sparse warnings: (new ones prefixed by >>)
>    net/core/dev.c:3254:23: sparse: sparse: incorrect type in argument 4 (different base types) @@     expected restricted __wsum [usertype] csum @@     got unsigned int @@
>    net/core/dev.c:3254:23: sparse:     expected restricted __wsum [usertype] csum
>    net/core/dev.c:3254:23: sparse:     got unsigned int
>    net/core/dev.c:3254:23: sparse: sparse: cast from restricted __wsum
> >> net/core/dev.c:10290:18: sparse: sparse: cast removes address space '__percpu' of expression  
>    net/core/dev.c:3759:17: sparse: sparse: context imbalance in '__dev_queue_xmit' - different lock contexts for basic block
>    net/core/dev.c:4961:17: sparse: sparse: context imbalance in 'net_tx_action' - different lock contexts for basic block
> 
> vim +/__percpu +10290 net/core/dev.c
> 
>  10282	
>  10283	struct net_device_core_stats *netdev_core_stats_alloc(struct net_device *dev)
>  10284	{
>  10285		struct net_device_core_stats __percpu *p;
>  10286	
>  10287		p = alloc_percpu_gfp(struct net_device_core_stats,
>  10288				     GFP_ATOMIC | __GFP_NOWARN);
>  10289	
>  10290		if (p && cmpxchg(&dev->core_stats, NULL, p))
>  10291			free_percpu(p);
>  10292	
>  10293		/* This READ_ONCE() pairs with the cmpxchg() above */
>  10294		p = READ_ONCE(dev->core_stats);
>  10295		if (!p)
>  10296			return NULL;
>  10297	
>  10298		return this_cpu_ptr(p);
>  10299	}
>  10300	EXPORT_SYMBOL(netdev_core_stats_alloc);
>  10301	
>