]> git.karo-electronics.de Git - karo-tx-linux.git/commitdiff
panic: release stale console lock to always get the logbuf printed out
authorVitaly Kuznetsov <vkuznets@redhat.com>
Wed, 21 Oct 2015 22:04:01 +0000 (09:04 +1100)
committerStephen Rothwell <sfr@canb.auug.org.au>
Wed, 21 Oct 2015 22:04:01 +0000 (09:04 +1100)
In some cases we may end up killing the CPU holding the console lock
while still having valuable data in logbuf. E.g. I'm observing the
following:

- A crash is happening on one CPU and console_unlock() is being called on
  some other.

- console_unlock() tries to print out the buffer before releasing the lock
  and on slow console it takes time.

- in the meanwhile crashing CPU does lots of printk()-s with valuable data
  (which go to the logbuf) and sends IPIs to all other CPUs.

- console_unlock() finishes printing previous chunk and enables interrupts
  before trying to print out the rest, the CPU catches the IPI and never
  releases console lock.

This is not the only possible case: in VT/fb subsystems we have many other
console_lock()/console_unlock() users.  Non-masked interrupts (or
receiving NMI in case of extreme slowness) will have the same result.
Getting the whole console buffer printed out on crash should be top
priority.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Cc: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Jiri Kosina <jkosina@suse.cz>
Cc: Baoquan He <bhe@redhat.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Xie XiuQi <xiexiuqi@huawei.com>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: "K. Y. Srinivasan" <kys@microsoft.com>
Cc: Jan Kara <jack@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
kernel/panic.c

index 04e91ff7560b3a35445f006a5be6e1d786bde245..f94525f601c968b4bea971b07b3357c6e0d5cb60 100644 (file)
@@ -23,6 +23,7 @@
 #include <linux/sysrq.h>
 #include <linux/init.h>
 #include <linux/nmi.h>
+#include <linux/console.h>
 
 #define PANIC_TIMER_STEP 100
 #define PANIC_BLINK_SPD 18
@@ -147,6 +148,15 @@ void panic(const char *fmt, ...)
 
        bust_spinlocks(0);
 
+       /*
+        * We may have ended up killing the CPU holding the lock and still have
+        * some valuable data in console buffer. Try to acquire the lock and
+        * release it regardless of the result. The release will also print the
+        * buffers out.
+        */
+       console_trylock();
+       console_unlock();
+
        if (!panic_blink)
                panic_blink = no_blink;