public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Yi Liu <yi.l.liu@intel.com>
To: Jason Gunthorpe <jgg@nvidia.com>
Cc: Lu Baolu <baolu.lu@linux.intel.com>,
	Bjorn Helgaas <bhelgaas@google.com>,
	David Woodhouse <dwmw2@infradead.org>, <iommu@lists.linux.dev>,
	Joerg Roedel <joro@8bytes.org>, <kvm@vger.kernel.org>,
	<linux-arm-kernel@lists.infradead.org>,
	<linux-pci@vger.kernel.org>,
	"Robin Murphy" <robin.murphy@arm.com>,
	Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>,
	Will Deacon <will@kernel.org>, <patches@lists.linux.dev>
Subject: Re: [PATCH] iommu: Allow ATS to work on VFs when the PF uses IDENTITY
Date: Wed, 14 Aug 2024 07:21:36 +0800	[thread overview]
Message-ID: <4bb6fcee-e16a-4a31-a69a-655bedd952b4@intel.com> (raw)
In-Reply-To: <20240813123041.GU8378@nvidia.com>

On 2024/8/13 20:30, Jason Gunthorpe wrote:
> On Tue, Aug 13, 2024 at 11:11:01AM +0800, Yi Liu wrote:
> 
>>> The simplest solution is to have the iommu driver set the ATS STU when it
>>> probes the device. This way the ATS STU is loaded immediately at boot time
>>> to all PFs and there is no issue when a VF comes to use it.
>>
>> This only sets STU without setting the ATS_CTRL.E bit. Is it possible that
>> VF considers the PF's STU field as valid only if PF's ATS_CTRL.E bit is
>> set?
> 
> That doesn't seem to be the case. Do you see something in the spec
> that says so?

no, I didn't find any word that describe it. So it is highly possible this
is not defined and up to vendors. This means there is possibility for what
I typed in the above :( I'm not a hw guy, but it seems reasonable to treat
something valid only when the cap is enabled?

>>> @@ -4091,6 +4091,7 @@ static struct iommu_device *intel_iommu_probe_device(struct device *dev)
>>>    	dev_iommu_priv_set(dev, info);
>>>    	if (pdev && pci_ats_supported(pdev)) {
>>> +		pci_prepare_ats(pdev, VTD_PAGE_SHIFT);
>>
>> perhaps just do it for PFs? :)
> 
> That check is inside pci_perpare_ats(), no reason to duplicate it in
> all the callers.

got it.

>>> +int pci_prepare_ats(struct pci_dev *dev, int ps)
>>> +{
>>> +	u16 ctrl;
>>> +
>>> +	if (!pci_ats_supported(dev))
>>> +		return -EINVAL;
>>> +
>>> +	if (WARN_ON(dev->ats_enabled))
>>> +		return -EBUSY;
>>> +
>>> +	if (ps < PCI_ATS_MIN_STU)
>>> +		return -EINVAL;
>>> +
>>> +	if (dev->is_virtfn)
>>> +		return 0;
>>> +
>>> +	dev->ats_stu = ps;
>>> +	ctrl = PCI_ATS_CTRL_STU(dev->ats_stu - PCI_ATS_MIN_STU);
>>> +	pci_write_config_word(dev, dev->ats_cap + PCI_ATS_CTRL, ctrl);
>>
>> Is it valuable to have a flag to mark if STU is set or not? Such way can
>> avoid setting STU multiple times.
> 
> We don't because we only do it for the PF due to the is_virtfn check

ok. Doing it only for PF is enough.

-- 
Regards,
Yi Liu

  reply	other threads:[~2024-08-13 23:17 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-07 18:19 [PATCH] iommu: Allow ATS to work on VFs when the PF uses IDENTITY Jason Gunthorpe
2024-08-09  2:36 ` Tian, Kevin
2024-08-09 13:28   ` Jason Gunthorpe
2024-08-12  0:09     ` Tian, Kevin
2024-08-12  9:03     ` Yi Liu
2024-08-12 11:35       ` Jason Gunthorpe
2024-08-13  3:00         ` Yi Liu
2024-08-09 19:10 ` Bjorn Helgaas
2024-08-12  2:20 ` Baolu Lu
2024-08-13  3:11 ` Yi Liu
2024-08-13 12:30   ` Jason Gunthorpe
2024-08-13 23:21     ` Yi Liu [this message]
2024-08-13  8:52 ` Joerg Roedel

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4bb6fcee-e16a-4a31-a69a-655bedd952b4@intel.com \
    --to=yi.l.liu@intel.com \
    --cc=baolu.lu@linux.intel.com \
    --cc=bhelgaas@google.com \
    --cc=dwmw2@infradead.org \
    --cc=iommu@lists.linux.dev \
    --cc=jgg@nvidia.com \
    --cc=joro@8bytes.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=patches@lists.linux.dev \
    --cc=robin.murphy@arm.com \
    --cc=suravee.suthikulpanit@amd.com \
    --cc=will@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox