KCSAN report there's a data race risk while using nr_threads. But according to the comment above it: ' /* * If multiple threads are within copy_process(), then this check * triggers too late. This doesn't hurt, the check is only there * to stop root fork bombs. */ ' The concurrency problem is not care. And we needn't to use READ_ONCE/atomic/etc to protect it. Meanwhile 'max_threads' is a sysctl variable which can be modified concurrently while being read, we can use 'data_race(nr_threads >= max_threads)' to mark both of then.
BUG: KCSAN: data-race in copy_process / copy_process write to 0xffffffff86205cf8 of 4 bytes by task 14779 on cpu 1: copy_process+0x2eba/0x3c40 kernel/fork.c:2273 _do_fork+0xfe/0x7a0 kernel/fork.c:2421 __do_sys_clone kernel/fork.c:2576 [inline] __se_sys_clone kernel/fork.c:2557 [inline] __x64_sys_clone+0x130/0x170 kernel/fork.c:2557 do_syscall_64+0xcc/0x3a0 arch/x86/entry/common.c:294 entry_SYSCALL_64_after_hwframe+0x44/0xa9 read to 0xffffffff86205cf8 of 4 bytes by task 6944 on cpu 0: copy_process+0x94d/0x3c40 kernel/fork.c:1954 _do_fork+0xfe/0x7a0 kernel/fork.c:2421 __do_sys_clone kernel/fork.c:2576 [inline] __se_sys_clone kernel/fork.c:2557 [inline] __x64_sys_clone+0x130/0x170 kernel/fork.c:2557 do_syscall_64+0xcc/0x3a0 arch/x86/entry/common.c:294 entry_SYSCALL_64_after_hwframe+0x44/0xa9 Link: https://groups.google.com/forum/#!msg/syzkaller-upstream-mo deration/thvp7AHs5Ew/aPdYLXfYBQAJ Reported-by: syzbot+52fced2d288f8ecd2...@syzkaller.appspotmail.com Cc: Qian Cai <c...@lca.pw> Cc: Oleg Nesterov <o...@redhat.com> Cc: Christian Brauner <christian.brau...@ubuntu.com> Cc: Marco Elver <el...@google.com> Signed-off-by: Zefan Li <lize...@huawei.com> Signed-off-by: Weilong Chen <chenweil...@huawei.com> --- kernel/fork.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/fork.c b/kernel/fork.c index 63c8fb2f5ca7..caa9c1f27444 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1976,7 +1976,7 @@ static __latent_entropy struct task_struct *copy_process( * to stop root fork bombs. */ retval = -EAGAIN; - if (nr_threads >= max_threads) + if (data_race(nr_threads >= max_threads)) goto bad_fork_cleanup_count; delayacct_tsk_init(p); /* Must remain after dup_task_struct() */ -- 2.17.1