Should shrinker_mutex be released on the error path?
julia
---------- Forwarded message ----------
Date: Fri, 29 Sep 2023 21:31:25 +0800
From: kernel test robot <lkp@xxxxxxxxx>
To: oe-kbuild@xxxxxxxxxxxxxxx
Cc: lkp@xxxxxxxxx, Julia Lawall <julia.lawall@xxxxxxxx>
Subject: [linux-next:master 8008/8345] mm/shrinker.c:100:1-7: preceding lock on
line 83
BCC: lkp@xxxxxxxxx
CC: oe-kbuild-all@xxxxxxxxxxxxxxx
CC: Linux Memory Management List <linux-mm@xxxxxxxxx>
TO: Qi Zheng <zhengqi.arch@xxxxxxxxxxxxx>
CC: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
CC: Linux Memory Management List <linux-mm@xxxxxxxxx>
CC: Muchun Song <songmuchun@xxxxxxxxxxxxx>
Hi Qi,
FYI, the error/warning was bisected to this commit, please ignore it if it's irrelevant.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
head: df964ce9ef9fea10cf131bf6bad8658fde7956f6
commit: 3b0ec351e30dfbe71675deb22da8a9fe1894c87f [8008/8345] mm: shrinker: convert shrinker_rwsem to mutex
:::::: branch date: 9 hours ago
:::::: commit date: 3 days ago
config: x86_64-randconfig-103-20230927 (https://download.01.org/0day-ci/archive/20230929/202309292114.VMjkyBBq-lkp@xxxxxxxxx/config)
compiler: gcc-12 (Debian 12.2.0-14) 12.2.0
reproduce: (https://download.01.org/0day-ci/archive/20230929/202309292114.VMjkyBBq-lkp@xxxxxxxxx/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@xxxxxxxxx>
| Reported-by: Julia Lawall <julia.lawall@xxxxxxxx>
| Closes: https://lore.kernel.org/r/202309292114.VMjkyBBq-lkp@xxxxxxxxx/
cocci warnings: (new ones prefixed by >>)
>> mm/shrinker.c:100:1-7: preceding lock on line 83
vim +100 mm/shrinker.c
2babd0d6e9cfca Qi Zheng 2023-09-11 76
2babd0d6e9cfca Qi Zheng 2023-09-11 77 int alloc_shrinker_info(struct mem_cgroup *memcg)
2babd0d6e9cfca Qi Zheng 2023-09-11 78 {
2babd0d6e9cfca Qi Zheng 2023-09-11 79 struct shrinker_info *info;
b6884b5f15cf5a Qi Zheng 2023-09-11 80 int nid, ret = 0;
b6884b5f15cf5a Qi Zheng 2023-09-11 81 int array_size = 0;
2babd0d6e9cfca Qi Zheng 2023-09-11 82
3b0ec351e30dfb Qi Zheng 2023-09-11 @83 mutex_lock(&shrinker_mutex);
b6884b5f15cf5a Qi Zheng 2023-09-11 84 array_size = shrinker_unit_size(shrinker_nr_max);
2babd0d6e9cfca Qi Zheng 2023-09-11 85 for_each_node(nid) {
b6884b5f15cf5a Qi Zheng 2023-09-11 86 info = kvzalloc_node(sizeof(*info) + array_size, GFP_KERNEL, nid);
b6884b5f15cf5a Qi Zheng 2023-09-11 87 if (!info)
b6884b5f15cf5a Qi Zheng 2023-09-11 88 goto err;
2babd0d6e9cfca Qi Zheng 2023-09-11 89 info->map_nr_max = shrinker_nr_max;
b6884b5f15cf5a Qi Zheng 2023-09-11 90 if (shrinker_unit_alloc(info, NULL, nid))
b6884b5f15cf5a Qi Zheng 2023-09-11 91 goto err;
2babd0d6e9cfca Qi Zheng 2023-09-11 92 rcu_assign_pointer(memcg->nodeinfo[nid]->shrinker_info, info);
2babd0d6e9cfca Qi Zheng 2023-09-11 93 }
3b0ec351e30dfb Qi Zheng 2023-09-11 94 mutex_unlock(&shrinker_mutex);
2babd0d6e9cfca Qi Zheng 2023-09-11 95
2babd0d6e9cfca Qi Zheng 2023-09-11 96 return ret;
b6884b5f15cf5a Qi Zheng 2023-09-11 97
b6884b5f15cf5a Qi Zheng 2023-09-11 98 err:
b6884b5f15cf5a Qi Zheng 2023-09-11 99 free_shrinker_info(memcg);
b6884b5f15cf5a Qi Zheng 2023-09-11 @100 return -ENOMEM;
2babd0d6e9cfca Qi Zheng 2023-09-11 101 }
2babd0d6e9cfca Qi Zheng 2023-09-11 102
:::::: The code at line 100 was first introduced by commit
:::::: b6884b5f15cf5acf6886ec0704d799035477ad7e mm: shrinker: add a secondary array for shrinker_info::{map, nr_deferred}
:::::: TO: Qi Zheng <zhengqi.arch@xxxxxxxxxxxxx>
:::::: CC: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
