linux-kernel - [PATCH -tip 1/2] x86/alternative: Sync bp_patching update for avoiding NULL pointer exception

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, 27 Nov 2019 14:56:52 +0900
From:   Masami Hiramatsu <mhiramat@...nel.org>
To:     Ingo Molnar <mingo@...nel.org>
Cc:     Steven Rostedt <rostedt@...dmis.org>,
        Peter Zijlstra <peterz@...radead.org>, x86@...nel.org,
        linux-kernel@...r.kernel.org, bristot@...hat.com,
        jbaron@...mai.com, torvalds@...ux-foundation.org,
        tglx@...utronix.de, namit@...are.com, hpa@...or.com,
        luto@...nel.org, ard.biesheuvel@...aro.org, jpoimboe@...hat.com,
        jeyu@...nel.org, alexei.starovoitov@...il.com,
        Masami Hiramatsu <mhiramat@...nel.org>
Subject: [PATCH -tip 1/2] x86/alternative: Sync bp_patching update for avoiding NULL pointer exception

ftracetest multiple_kprobes.tc testcase hit a following NULL pointer
exception.

BUG: kernel NULL pointer dereference, address: 0000000000000000
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
PGD 800000007bf60067 P4D 800000007bf60067 PUD 7bf5f067 PMD 0
Oops: 0000 [#1] PREEMPT SMP PTI
CPU: 6 PID: 0 Comm: swapper/6 Not tainted 5.4.0-rc8+ #23
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014
RIP: 0010:poke_int3_handler+0x39/0x100
Code: 5b 5d c3 f6 87 88 00 00 00 03 75 f2 48 8b 87 80 00 00 00 48 89 fb 48 8d 68 ff 48 8b 05 80 98 72 01 83 fa 01 0f 8f 93 00 00 00 <48> 63 10 48 81 c2 00 00 00 81 48 39 d5 75 c5 0f b6 50 08 8d 4a 34
RSP: 0018:ffffc900001a8eb8 EFLAGS: 00010046
RAX: 0000000000000000 RBX: ffffc900001a8ee8 RCX: ffffffff81a00b57
RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffffc900001a8ee8
RBP: ffffffff81027635 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
FS:  0000000000000000(0000) GS:ffff88807d980000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000000 CR3: 000000007a970000 CR4: 00000000000006a0
Call Trace:
 <IRQ>
 do_int3+0xd/0xf0
 int3+0x42/0x50
RIP: 0010:sched_clock+0x6/0x10

eu-addr2line told that poke_int3_handler+0x39 was alternatives:958.

static inline void *text_poke_addr(struct text_poke_loc *tp)
{
        return _stext + tp->rel_addr; <------ Here is line #958
}

This seems like caused by the tp (bp_patching.vec) was NULL but
bp_patching.nr_entries != 0. There is a small chance to do
this, because we have no sync after zeroing bp_patching.nr_entries
before clearing bp_patching.vec.

Steve suggested we could fix this by adding sync_core, because int3
is done with interrupts disabled, and the on_each_cpu() requires
all CPUs to have had their interrupts enabled.

Fixes: c0213b0ac03c ("x86/alternative: Batch of patch operations")
Suggested-by: Steven Rostedt (VMware) <rostedt@...dmis.org>
Signed-off-by: Masami Hiramatsu <mhiramat@...nel.org>
---
 arch/x86/kernel/alternative.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 4552795a8df4..9505096e2cd1 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -1134,8 +1134,14 @@ static void text_poke_bp_batch(struct text_poke_loc *tp, unsigned int nr_entries
 	 * sync_core() implies an smp_mb() and orders this store against
 	 * the writing of the new instruction.
 	 */
-	bp_patching.vec = NULL;
 	bp_patching.nr_entries = 0;
+	/*
+	 * This sync_core () ensures that all int3 handlers in progress
+	 * have finished. This allows poke_int3_handler () after this to
+	 * avoid touching bp_paching.vec by checking nr_entries == 0.
+	 */
+	text_poke_sync();
+	bp_patching.vec = NULL;
 }
 
 void text_poke_loc_init(struct text_poke_loc *tp, void *addr,