linux-kernel - Using list_for_each_entry() in place of list_for_each_entry

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <201411012059.JHB90633.HtJOSOVLFFQFMO@I-love.SAKURA.ne.jp>
Date:	Sat, 1 Nov 2014 20:59:03 +0900
From:	Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
To:	paulmck@...ux.vnet.ibm.com
Cc:	linux-kernel@...r.kernel.org
Subject: Using list_for_each_entry() in place of list_for_each_entry_rcu() ?

Excuse me for FAQ, Paul.
I want to confirm one thing for code optimization in LSM stacking.
( https://marc.info/?l=linux-security-module&m=141481716931982&w=2 )

In the following code, is there race window for seeing invalid
"struct list_head"->next value if we used list_for_each_entry()
in place of list_for_each_entry_rcu() ?

----------
/* Definition and declaration */
DEFINE_SPINLOCK(my_lock);
LIST_HEAD(my_list);
struct my_struct {
    struct list_head list;
    const unsigned long value;
} v1 = { .value = 1 }, v2 = { .value = 2 }, v3 = { .value = 3 };

/* Writer side */
void add_entry(struct my_struct *p) {
    spin_lock(&my_lock);
    list_add_tail_rcu(&p->list, &my_list);
    spin_unlock(&my_lock);
}

void del_entry(struct my_struct *p) {
    spin_lock(&my_lock);
    list_del_rcu(&p->list);
    spin_unlock(&my_lock);
}

/* Reader side */
unsigned long reader(void) {
    struct my_struct *p;
    unsigned long sum = 0;
    list_for_each_entry_rcu(p, &my_list, list)
        sum += p->value;
    return sum;
}
----------

Assumptions are:

  (1) v1, v2, v3 are statically allocated variables inside module,
      while my_lock, my_list, add_entry(), del_entry(), reader()
      are built-in.

  (2) v1, v2, v3 are added to my_list only once upon module load

  (3) v1, v2, v3 might be removed from my_list some time later after
      module was loaded

Regards.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/