From: Petr Mladek <pmladek@suse.com>
To: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>,
Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <peterz@infradead.org>,
Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>,
Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>,
linux-kernel@vger.kernel.org, Petr Mladek <pmladek@suse.com>,
"4 . 13+" <stable@vger.kernel.org>
Subject: [PATCH] printk/nmi: Prevent deadlock when serializing NMI backtraces
Date: Thu, 17 May 2018 16:39:03 +0200 [thread overview]
Message-ID: <20180517143903.19339-1-pmladek@suse.com> (raw)
The commit 719f6a7040f1bdaf96fcc ("printk: Use the main logbuf in NMI when
logbuf_lock is available") tried to detect when logbuf_lock was taken
on another CPU. Then it looked safe to wait for the lock even in NMI.
It would be safe if other locks were not involved. Ironically the same
commit introduced an ABBA deadlock scenario. It added a spin lock into
nmi_cpu_backtrace() to serialize logs from different CPUs. The effect
is that also the NMI handlers are serialized. As a result, logbuf_lock
might be blocked by NMI on another CPU:
CPU0 CPU1 CPU2
printk()
vprintk_emit()
spin_lock(&logbuf_lock)
trigger_all_cpu_backtrace()
raise()
nmi_enter()
printk_nmi_enter()
if (this_cpu_read(printk_context)
& PRINTK_SAFE_CONTEXT_MASK)
// false
else
// looks safe to use printk_deferred()
this_cpu_or(printk_context,
PRINTK_NMI_DEFERRED_CONTEXT_MASK);
nmi_cpu_backtrace()
arch_spin_lock(&lock);
show_regs()
nmi_enter()
nmi_cpu_backtrace()
arch_spin_lock(&lock);
printk()
vprintk_func()
vprintk_deferred()
vprintk_emit()
spin_lock(&logbuf_lock)
DEADLOCK: between &logbuf_lock from vprintk_emit() and
&lock from nmi_cpu_backtrace().
CPU0 CPU1
lock(logbuf_lock) lock(lock)
lock(lock) lock(logbuf_lock)
I have found this problem when stress testing trigger_all_cpu_backtrace()
and the system frozen.
Note that lockdep is not able to detect these dependencies because
there is no support for NMI context. Let's stay on the safe side
and always use printk_safe buffers when logbuf_lock is taken
when entering NMI.
Fixes: 719f6a7040f1bdaf96fcc ("printk: Use the main logbuf in NMI when logbuf_lock is available")
Cc: 4.13+ <stable@vger.kernel.org> # v4.13+
Signed-off-by: Petr Mladek <pmladek@suse.com>
---
kernel/printk/printk_safe.c | 9 +++------
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/kernel/printk/printk_safe.c b/kernel/printk/printk_safe.c
index 449d67edfa4b..a2ebd749c053 100644
--- a/kernel/printk/printk_safe.c
+++ b/kernel/printk/printk_safe.c
@@ -310,15 +310,12 @@ void printk_nmi_enter(void)
{
/*
* The size of the extra per-CPU buffer is limited. Use it only when
- * the main one is locked. If this CPU is not in the safe context,
- * the lock must be taken on another CPU and we could wait for it.
+ * the main one is locked.
*/
- if ((this_cpu_read(printk_context) & PRINTK_SAFE_CONTEXT_MASK) &&
- raw_spin_is_locked(&logbuf_lock)) {
+ if (raw_spin_is_locked(&logbuf_lock))
this_cpu_or(printk_context, PRINTK_NMI_CONTEXT_MASK);
- } else {
+ else
this_cpu_or(printk_context, PRINTK_NMI_DEFERRED_CONTEXT_MASK);
- }
}
void printk_nmi_exit(void)
--
2.13.6
next reply other threads:[~2018-05-17 14:39 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-17 14:39 Petr Mladek [this message]
2018-05-18 2:07 ` Sergey Senozhatsky
2018-05-18 6:38 ` Sergey Senozhatsky
2018-05-18 8:10 ` Petr Mladek
2018-05-22 21:43 ` Steven Rostedt
2018-05-23 2:01 ` Sergey Senozhatsky
2018-05-28 12:27 ` Petr Mladek
2018-06-05 12:47 ` Petr Mladek
2018-06-06 5:10 ` Sergey Senozhatsky
2018-06-06 10:33 ` Sergey Senozhatsky
2018-06-08 10:48 ` Petr Mladek
2018-06-18 6:37 ` Sergey Senozhatsky
2018-06-18 9:39 ` Petr Mladek
2018-06-18 10:07 ` Sergey Senozhatsky
2018-06-19 7:52 ` Petr Mladek
2018-06-19 8:27 ` Sergey Senozhatsky
2018-06-19 13:23 ` Steven Rostedt
2018-06-20 1:58 ` Sergey Senozhatsky
2018-06-20 2:32 ` Steven Rostedt
2018-06-20 4:17 ` Sergey Senozhatsky
2018-06-06 11:15 ` Petr Mladek
2018-06-07 5:40 ` Sergey Senozhatsky
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180517143903.19339-1-pmladek@suse.com \
--to=pmladek@suse.com \
--cc=linux-kernel@vger.kernel.org \
--cc=penguin-kernel@I-love.SAKURA.ne.jp \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=sergey.senozhatsky.work@gmail.com \
--cc=sergey.senozhatsky@gmail.com \
--cc=stable@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®