sparc64: Eliminate obsolete __handle_softirq() function
authorPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Fri, 13 Apr 2012 03:35:13 +0000 (03:35 +0000)
committerDavid S. Miller <davem@davemloft.net>
Fri, 13 Apr 2012 18:58:38 +0000 (11:58 -0700)
commit3d3eeb2ef26112a200785e5fca58ec58dd33bf1e
tree9a7ba8c313355f9b9832bff425517c0457f0cf69
parent9e0daff30fd7ecf698e5d20b0fa7f851e427cca5
sparc64: Eliminate obsolete __handle_softirq() function

The invocation of softirq is now handled by irq_exit(), so there is no
need for sparc64 to invoke it on the trap-return path.  In fact, doing so
is a bug because if the trap occurred in the idle loop, this invocation
can result in lockdep-RCU failures.  The problem is that RCU ignores idle
CPUs, and the sparc64 trap-return path to the softirq handlers fails to
tell RCU that the CPU must be considered non-idle while those handlers
are executing.  This means that RCU is ignoring any RCU read-side critical
sections in those handlers, which in turn means that RCU-protected data
can be yanked out from under those read-side critical sections.

The shiny new lockdep-RCU ability to detect RCU read-side critical sections
that RCU is ignoring located this problem.

The fix is straightforward: Make sparc64 stop manually invoking the
softirq handlers.

Reported-by: Meelis Roos <mroos@linux.ee>
Suggested-by: David Miller <davem@davemloft.net>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Meelis Roos <mroos@linux.ee>
Cc: stable@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
arch/sparc/kernel/rtrap_64.S