Date: Thu, 17 Mar 2016 16:42:56 +0100
From: Torsten Duwe <duwe@lst.de>
To: Balbir Singh <bsingharora@gmail.com>
Cc: linuxppc-dev@ozlabs.org, linux-kernel@vger.kernel.org, rostedt@goodmis.org,
        kamalesh@linux.vnet.ibm.com, pmladek@suse.com, jeyu@redhat.com,
        jkosina@suse.cz, live-patching@vger.kernel.org, mbenes@suse.cz
Subject: Re: [PATCH][v6][RFC] livepatch/ppc: Enable livepatching on powerpc
Message-ID: <20160317154256.GA32575@lst.de>
References: <1457506780-19556-1-git-send-email-bsingharora@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1457506780-19556-1-git-send-email-bsingharora@gmail.com>
User-Agent: Mutt/1.5.17 (2007-11-01)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 786
Lines: 19

On Wed, Mar 09, 2016 at 05:59:40PM +1100, Balbir Singh wrote:
> 
> Changelog v6:
> 	1. Experimental changes -- need loads of testing
> 	   Based on the assumption that very far TOC and LR values
> 	   indicate the call happened through a stub and the
> 	   stub return works differently from a local call which
> 	   uses klp_return_helper.

Well, this is true, but the inverse is not.
Less than 2GiB between LR and TOC is a necessary but not a sufficient
condition for a local call.

I see the problem with your code that it hardly detects calls from one module
to another nearby. It will fail iff you mistake a global call to a module loaded
earlier to be local while patching AND you access global data / 64-bit constants
after that call before the TOC is sanitised again.

	Torsten