Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 8 Sep 2022 15:28:22 +0200
From:   Joerg Roedel <joro@8bytes.org>
To:     Jason Gunthorpe <jgg@nvidia.com>
Cc:     Nicolin Chen <nicolinc@nvidia.com>, will@kernel.org,
        robin.murphy@arm.com, alex.williamson@redhat.com,
        suravee.suthikulpanit@amd.com, marcan@marcan.st,
        sven@svenpeter.dev, alyssa@rosenzweig.io, robdclark@gmail.com,
        dwmw2@infradead.org, baolu.lu@linux.intel.com,
        mjrosato@linux.ibm.com, gerald.schaefer@linux.ibm.com,
        orsonzhai@gmail.com, baolin.wang@linux.alibaba.com,
        zhang.lyra@gmail.com, thierry.reding@gmail.com, vdumpa@nvidia.com,
        jonathanh@nvidia.com, jean-philippe@linaro.org, cohuck@redhat.com,
        tglx@linutronix.de, shameerali.kolothum.thodi@huawei.com,
        thunder.leizhen@huawei.com, christophe.jaillet@wanadoo.fr,
        yangyingliang@huawei.com, jon@solid-run.com, iommu@lists.linux.dev,
        linux-kernel@vger.kernel.org, asahi@lists.linux.dev,
        linux-arm-kernel@lists.infradead.org,
        linux-arm-msm@vger.kernel.org, linux-s390@vger.kernel.org,
        linux-tegra@vger.kernel.org,
        virtualization@lists.linux-foundation.org, kvm@vger.kernel.org,
        kevin.tian@intel.com
Subject: Re: [PATCH v6 1/5] iommu: Return -EMEDIUMTYPE for incompatible
 domain and device/group
Message-ID: <Yxnt9uQTmbqul5lf@8bytes.org>
References: <20220815181437.28127-1-nicolinc@nvidia.com>
 <20220815181437.28127-2-nicolinc@nvidia.com>
 <YxiRkm7qgQ4k+PIG@8bytes.org>
 <Yxig+zfA2Pr4vk6K@nvidia.com>
 <YxilZbRL0WBR97oi@8bytes.org>
 <YxjQiVnpU0dr7SHC@nvidia.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YxjQiVnpU0dr7SHC@nvidia.com>
Precedence: bulk

On Wed, Sep 07, 2022 at 02:10:33PM -0300, Jason Gunthorpe wrote:
> Sure, rust has all sorts of nice things. But the kernel doesn't follow
> rust idioms, and I don't think this is a great place to start
> experimenting with them.

It is actually a great place to start experimenting. The IOMMU
interfaces are rather domain specific and if we get something wrong the
damage is limited to a few callers. There are APIs much more exposed in
the kernel which would be worse for that.

But anyway, I am not insisting on it.

> It has been 3 months since EMEDIUMTYPE was first proposed and 6
> iterations of the series, don't you think it is a bit late in the game
> to try to experiment with rust error handling idioms?

If I am not mistaken, I am the person who gets blamed when crappy IOMMU
code is sent upstream. So it is also up to me to decide in which state
and how close to merging a given patch series is an whether it is
already 'late in the game'.

> So, again, would you be happy with a simple 
> 
>  #define IOMMU_EINCOMPATIBLE_DEVICE xx
> 
> to make it less "re-using random error codes"?

I am wondering if this can be solved by better defining what the return
codes mean and adjust the call-back functions to match the definition.
Something like:

	-ENODEV : Device not mapped my an IOMMU
	-EBUSY  : Device attached and domain can not be changed
	-EINVAL : Device and domain are incompatible
	...

That would be much more intuitive than using something obscure like
EMEDIUMTYPE.

Regards,

	Joerg