From mboxrd@z Thu Jan  1 00:00:00 1970
From: andreas.herrmann@calxeda.com (Andreas Herrmann)
Date: Fri, 27 Sep 2013 15:48:09 +0200
Subject: [PATCH 6/9] iommu/arm-smmu: Support buggy implemenations where
 all config accesses are secure
In-Reply-To: <20130927130527.GH9520@mudshark.cambridge.arm.com>
References: <1380234982-1677-1-git-send-email-andreas.herrmann@calxeda.com>
 <1380234982-1677-7-git-send-email-andreas.herrmann@calxeda.com>
 <20130927130527.GH9520@mudshark.cambridge.arm.com>
Message-ID: <20130927134809.GQ3315@alberich>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Fri, Sep 27, 2013 at 09:05:28AM -0400, Will Deacon wrote:
> On Thu, Sep 26, 2013 at 11:36:18PM +0100, Andreas Herrmann wrote:
> > In such a case we have to use secure aliases of some non-secure
> > registers.
> > 
> > This behaviour is controlled via a flag in smmu->bugs. It is set
> > based on DT information when probing an SMMU device.
> 
> :(
> 
> I guess my main comment is just some understanding on how things are wired.
> It's *just* the configuration access that is secure, right?

Correct.

> Device transactions are non-secure, so the existing TLB invalidation
> code will work correctly?

Yes, transactions are non-secure.

The TLB invalidation code (using SMMU_TLBIVMID for stage2) also seems
too work -- no problems triggered so far (neither with sata nor with
xgmac).


Andreas