Re: [PATCH 2/2] uacce: add uacce module

[Greg Kroah-Hartman]

On Mon, Aug 26, 2019 at 12:10:42PM +0800, Kenneth Lee wrote:
> On Wed, Aug 21, 2019 at 09:05:42AM -0700, Greg Kroah-Hartman wrote:
> > Date: Wed, 21 Aug 2019 09:05:42 -0700
> > From: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > To: zhangfei <zhangfei.gao@xxxxxxxxxx>
> > CC: Arnd Bergmann <arnd@xxxxxxxx>, linux-accelerators@xxxxxxxxxxxxxxxx,
> >  linux-kernel@xxxxxxxxxxxxxxx, Kenneth Lee <liguozhu@xxxxxxxxxxxxx>, Zaibo
> >  Xu <xuzaibo@xxxxxxxxxx>, Zhou Wang <wangzhou1@xxxxxxxxxxxxx>
> > Subject: Re: [PATCH 2/2] uacce: add uacce module
> > User-Agent: Mutt/1.12.1 (2019-06-15)
> > Message-ID: <20190821160542.GA14760@xxxxxxxxx>
> > 
> > On Wed, Aug 21, 2019 at 10:30:22PM +0800, zhangfei wrote:
> > > 
> > > 
> > > On 2019/8/21 äå5:17, Greg Kroah-Hartman wrote:
> > > > On Wed, Aug 21, 2019 at 03:21:18PM +0800, zhangfei.gao@xxxxxxxxxxx wrote:
> > > > > Hi, Greg
> > > > > 
> > > > > On 2019/8/21 äå12:59, Greg Kroah-Hartman wrote:
> > > > > > On Tue, Aug 20, 2019 at 09:08:55PM +0800, zhangfei wrote:
> > > > > > > On 2019/8/15 äå10:13, Greg Kroah-Hartman wrote:
> > > > > > > > On Wed, Aug 14, 2019 at 05:34:25PM +0800, Zhangfei Gao wrote:
> > > > > > > > > +int uacce_register(struct uacce *uacce)
> > > > > > > > > +{
> > > > > > > > > +	int ret;
> > > > > > > > > +
> > > > > > > > > +	if (!uacce->pdev) {
> > > > > > > > > +		pr_debug("uacce parent device not set\n");
> > > > > > > > > +		return -ENODEV;
> > > > > > > > > +	}
> > > > > > > > > +
> > > > > > > > > +	if (uacce->flags & UACCE_DEV_NOIOMMU) {
> > > > > > > > > +		add_taint(TAINT_CRAP, LOCKDEP_STILL_OK);
> > > > > > > > > +		dev_warn(uacce->pdev,
> > > > > > > > > +			 "Register to noiommu mode, which export kernel data to user space and may vulnerable to attack");
> > > > > > > > > +	}
> > > > > > > > THat is odd, why even offer this feature then if it is a major issue?
> > > > > > > UACCE_DEV_NOIOMMU maybe confusing here.
> > > > > > > 
> > > > > > > In this mode, app use ioctl to get dma_handle from dma_alloc_coherent.
> > > > > > That's odd, why not use the other default apis to do that?
> > > > > > 
> > > > > > > It does not matter iommu is enabled or not.
> > > > > > > In case iommu is disabled, it maybe dangerous to kernel, so we added warning here, is it required?
> > > > > > You should use the other documentated apis for this, don't create your
> > > > > > own.
> > > > > I am sorry, not understand here.
> > > > > Do you mean there is a standard ioctl or standard api in user space, it can
> > > > > get dma_handle from dma_alloc_coherent from kernel?
> > > > There should be a standard way to get such a handle from userspace
> > > > today.  Isn't that what the ion interface does?  DRM also does this, as
> > > > does UIO I think.
> > > Thanks Greg,
> > > Still not find it, will do more search.
> > > But this may introduce dependency in our lib, like depend on ion?
> > > > Do you have a spec somewhere that shows exactly what you are trying to
> > > > do here, along with example userspace code?  It's hard to determine it
> > > > given you only have one "half" of the code here and no users of the apis
> > > > you are creating.
> > > > 
> > > The purpose is doing dma in user space.
> > 
> > Oh no, please no.  Are you _SURE_ you want to do this?
> > 
> > Again, look at how ION does this and how the DMAbuff stuff is replacing
> > it.  Use that api please instead, otherwise you will get it wrong and we
> > don't want to duplicate efforts.
> > 
> > thanks,
> > 
> > greg k-h
> 
> Dear Greg. I wrote a blog to explain the intention of WarpDrive here:
> https://zhuanlan.zhihu.com/p/79680889.

Putting that information into the changelog and kernel documentation is
a much better idea than putting it there.

> Sharing data is not our intention, Sharing address is. NOIOMMU mode is just a
> temporary solution to let some hardware which does not care the security issue
> to try WarpDrive for the first step. Some user do not care this much in embedded
> scenario. We saw VFIO use the same model so we also want to make a try. If you
> insist this is risky, we can remove it.

Why not just use vfio then?

And yes, for now, please remove it, if you are not requiring it.

thanks,

greg k-h