[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <f5606802-2ff3-2f54-0ff1-c1f1dd59f52c@intel.com>
Date: Wed, 26 May 2021 18:15:17 -0700
From: Dave Jiang <dave.jiang@...el.com>
To: Jason Gunthorpe <jgg@...dia.com>
Cc: alex.williamson@...hat.com, kwankhede@...dia.com,
tglx@...utronix.de, vkoul@...nel.org, megha.dey@...el.com,
jacob.jun.pan@...el.com, ashok.raj@...el.com, yi.l.liu@...el.com,
baolu.lu@...el.com, kevin.tian@...el.com, sanjay.k.kumar@...el.com,
tony.luck@...el.com, dan.j.williams@...el.com,
eric.auger@...hat.com, pbonzini@...hat.com,
dmaengine@...r.kernel.org, linux-kernel@...r.kernel.org,
kvm@...r.kernel.org
Subject: Re: [PATCH v6 15/20] vfio/mdev: idxd: ims domain setup for the vdcm
On 5/26/2021 5:54 PM, Jason Gunthorpe wrote:
> On Wed, May 26, 2021 at 05:22:22PM -0700, Dave Jiang wrote:
>> On 5/23/2021 4:50 PM, Jason Gunthorpe wrote:
>>> On Fri, May 21, 2021 at 05:20:37PM -0700, Dave Jiang wrote:
>>>> @@ -77,8 +80,18 @@ int idxd_mdev_host_init(struct idxd_device *idxd, struct mdev_driver *drv)
>>>> return rc;
>>>> }
>>>> + ims_info.max_slots = idxd->ims_size;
>>>> + ims_info.slots = idxd->reg_base + idxd->ims_offset;
>>>> + idxd->ims_domain = pci_ims_array_create_msi_irq_domain(idxd->pdev, &ims_info);
>>>> + if (!idxd->ims_domain) {
>>>> + dev_warn(dev, "Fail to acquire IMS domain\n");
>>>> + iommu_dev_disable_feature(dev, IOMMU_DEV_FEAT_AUX);
>>>> + return -ENODEV;
>>>> + }
>>> I'm quite surprised that every mdev doesn't create its own ims_domain
>>> in its probe function.
>>>
>>> This places a global total limit on the # of vectors which makes me
>>> ask what was the point of using IMS in the first place ?
>>>
>>> The entire idea for IMS was to make the whole allocation system fully
>>> dynamic based on demand.
>> Hi Jason, thank you for the review of the series.
>>
>> My understanding is that the driver creates a single IMS domain for the
>> device and provides the address base and IMS numbers for the domain based on
>> device IMS resources. So the IMS region needs to be contiguous. Each mdev
>> can call msi_domain_alloc_irqs() and acquire the number of IMS vectors it
>> desires and the DEV MSI core code will keep track of which vectors are being
>> used. This allows the mdev devices to dynamically allocate based on demand.
>> If the driver allocates a domain per mdev, it'll needs to do internal
>> accounting of the base and vector numbers for each of those domains that the
>> MSI core already provides. Isn't that what we are trying to avoid? As mdevs
>> come and go, that partitioning will become fragmented.
> I suppose it depends entirely on how the HW works.
>
> If the HW has a fixed number of interrupt vectors organized in a
> single table then by all means allocate a single domain that spans the
> entire fixed HW vector space. But then why do we have a ims_size
> variable here??
>
> However, that really begs the question of why the HW is using IMS at
> all? I'd expect needing 2x-10x the max MSI-X vector size before
> reaching for IMS.
>
> So does IDXD really have like a 4k - 40k entry linear IMS vector table
> to wrap a shared domain around?
>
> Basically, that isn't really "scalable" it is just "bigger".
>
> Fully scalable would be for every mdev to point to its own 2k entry
> IMS table that is allocated on the fly. Every mdev gets a domain and
> every domain is fully utilized by the mdev in emulating
> MSI-X. Basically for a device like idxd every PASID would have to map
> to a IMS vector table array.
>
> I suppose that was not what was done?
At least not for first gen of hardware. DSA 1.0 has 2k of IMS entries
total. ims_size is what is read from the device cap register. For MSIX,
the device only has 1 misc vector and 8 I/O vectors. That's why IMS is
being used for mdevs. We will discuss with our hardware people your
suggestion.
>
> Jason
Powered by blists - more mailing lists