linux-kernel - [PATCH 1/3] kprobes: Fixed probe nodes not correctly removed when forcibly unoptimized

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Wed, 15 Feb 2023 19:54:28 +0800
From:   Yang Jihong <yangjihong1@...wei.com>
To:     <tglx@...utronix.de>, <mingo@...hat.com>, <bp@...en8.de>,
        <dave.hansen@...ux.intel.com>, <x86@...nel.org>, <hpa@...or.com>,
        <naveen.n.rao@...ux.ibm.com>, <anil.s.keshavamurthy@...el.com>,
        <davem@...emloft.net>, <mhiramat@...nel.org>, <ast@...nel.org>,
        <peterz@...radead.org>, <linux-kernel@...r.kernel.org>,
        <linux-trace-kernel@...r.kernel.org>
CC:     <yangjihong1@...wei.com>
Subject: [PATCH 1/3] kprobes: Fixed probe nodes not correctly removed when forcibly unoptimized

When unoptimize_kprobe forcibly unoptimize the kprobe, simply queue it in
the freeing_list, and do_free_cleaned_kprobes directly reclaims the kprobe
if unoptimizing_list is empty (see do_unoptimize_kprobes), which may cause
WARN or UAF problems.

The specific scenarios are as follows:

                          Thread1
arm_kprobe(p)
  mutex_lock(&kprobe_mutex)
  __arm_kprobe(kp)
    p = get_optimized_kprobe(p->addr)
    if (unlikely(_p))
      unoptimize_kprobe(_p, true)  // now _p is queued in freeing_list
  mutex_unlock(&kprobe_mutex)

                           Thread2
kprobe_optimizer
  mutex_lock(&kprobe_mutex)
  do_unoptimize_kprobes
    if (list_empty(&unoptimizing_list))
      return;  // here directly returned and does not process freeing_list.
  ...
  do_free_cleaned_kprobes
    foreach op in freeing_list:
      WARN_ON_ONCE(!kprobe_unused(&op->kp))  // WANR will be triggered here.
      free_aggr_kprobe((&op->kp)  // Delete op->kp directly, if access hash
                                  // list later, UAF problem will be triggered.
  mutex_unlock(&kprobe_mutex)

The freeing_list needs to be processed in do_unoptimize_kprobes regardless
of whether unoptimizing_list is empty.

Signed-off-by: Yang Jihong <yangjihong1@...wei.com>
---
 kernel/kprobes.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 1c18ecf9f98b..0730e595f4c1 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -556,10 +556,9 @@ static void do_unoptimize_kprobes(void)
 	lockdep_assert_cpus_held();
 
 	/* Unoptimization must be done anytime */
-	if (list_empty(&unoptimizing_list))
-		return;
+	if (!list_empty(&unoptimizing_list))
+		arch_unoptimize_kprobes(&unoptimizing_list, &freeing_list);
 
-	arch_unoptimize_kprobes(&unoptimizing_list, &freeing_list);
 	/* Loop on 'freeing_list' for disarming */
 	list_for_each_entry_safe(op, tmp, &freeing_list, list) {
 		/* Switching from detour code to origin */
-- 
2.30.GIT