From mboxrd@z Thu Jan  1 00:00:00 1970
From: Joerg.Roedel@amd.com (Roedel, Joerg)
Date: Mon, 5 Sep 2011 12:00:53 +0200
Subject: [RFC 6/7] iommu/core: add fault reporting
In-Reply-To: <1314984756-4400-7-git-send-email-ohad@wizery.com>
References: <1314984756-4400-1-git-send-email-ohad@wizery.com>
	<1314984756-4400-7-git-send-email-ohad@wizery.com>
Message-ID: <20110905100053.GP16691@amd.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Fri, Sep 02, 2011 at 01:32:35PM -0400, Ohad Ben-Cohen wrote:
> -struct iommu_domain *iommu_domain_alloc(void)
> +/**
> + * iommu_domain_alloc() - allocate and initialize a new iommu domain
> + * @handler: an optional pointer to a fault handler, or NULL if not needed
> + *
> + * Returns the new domain, or NULL on error.
> + */
> +struct iommu_domain *iommu_domain_alloc(iommu_fault_handler_t handler)

Please add a seperate function for setting the fault-handler. It is
optional, so no need to be a value of the alloc-function.

> +/**
> + * enum iommu_fault_types - iommu fault types
> + *
> + * @IOMMU_ERROR: Unrecoverable error
> + * @IOMMU_TLBMISS: TLB miss while the page table walker is disabled
> + * @IOMMU_NOPTE: TLB miss and no PTE for the requested address
> + */
> +enum iommu_fault_types {
> +	IOMMU_ERROR,
> +	IOMMU_TLBMISS,
> +	IOMMU_NOPTE,
> +};

Can you elaborate a bit on what the user of the api will do different
between IOMMU_TLBMISS and IOMMU_NOPTE?
My feeling is that those differences should be handled internally in the
IOMMU driver, but probably I miss a use-case.
Also, we need some flags to distinguish between the type of the fault
(read, write, ...).

	Joerg


-- 
AMD Operating System Research Center

Advanced Micro Devices GmbH Einsteinring 24 85609 Dornach
General Managers: Alberto Bozzo, Andrew Bowd
Registration: Dornach, Landkr. Muenchen; Registerger. Muenchen, HRB Nr. 43632