From: Alexey Kardashevskiy <aik@amd.com>
To: Leon Romanovsky <leon@kernel.org>, Robin Murphy <robin.murphy@arm.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
"Rafael J. Wysocki" <rafael@kernel.org>,
Danilo Krummrich <dakr@kernel.org>,
Dmitry Torokhov <dmitry.torokhov@gmail.com>,
Andrew Lunn <andrew@lunn.ch>,
Heiner Kallweit <hkallweit1@gmail.com>,
Russell King <linux@armlinux.org.uk>,
"David S. Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
Johannes Berg <johannes@sipsolutions.net>,
Jonathan Corbet <corbet@lwn.net>,
Shuah Khan <skhan@linuxfoundation.org>,
Randy Dunlap <rdunlap@infradead.org>,
Bjorn Helgaas <bhelgaas@google.com>,
"Joerg Roedel (AMD)" <joro@8bytes.org>,
Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>,
Vasant Hegde <vasant.hegde@amd.com>,
Will Deacon <will@kernel.org>,
David Woodhouse <dwmw2@infradead.org>,
Lu Baolu <baolu.lu@linux.intel.com>,
Johan Hovold <johan@kernel.org>,
Eric Dumazet <edumazet@kernel.org>,
Marcel Holtmann <marcel@holtmann.org>,
Luiz Augusto von Dentz <luiz.dentz@gmail.com>,
Nipun Gupta <nipun.gupta@amd.com>,
Nikhil Agarwal <nikhil.agarwal@amd.com>,
Pieter Jansen van Vuuren <pieter.jansen-van-vuuren@amd.com>,
linux-kernel@vger.kernel.org, driver-core@lists.linux.dev,
linux-input@vger.kernel.org, netdev@vger.kernel.org,
linux-wireless@vger.kernel.org, linux-usb@vger.kernel.org,
linux-bluetooth@vger.kernel.org, linux-acpi@vger.kernel.org,
linux-doc@vger.kernel.org, linux-pci@vger.kernel.org,
iommu@lists.linux.dev, Jason Gunthorpe <jgg@ziepe.ca>
Subject: Re: [PATCH v2 18/19] PCI: Replace untrusted state with device trust
Date: Wed, 18 Nov 2026 18:33:45 +1100 [thread overview]
Message-ID: <fd25ed52-e3ea-41a2-9f62-2397d5d9e8c8@amd.com> (raw)
In-Reply-To: <20261117124801.GD124972@unreal>
On 17/11/26 23:48, Leon Romanovsky wrote:
> On Mon, Nov 16, 2026 at 05:13:59PM +0000, Robin Murphy wrote:
>> On 10/11/2026 12:22 pm, Leon Romanovsky wrote:
>>> On Mon, Nov 09, 2026 at 01:53:02PM +0000, Robin Murphy wrote:
>>>> On 09/11/2026 7:33 am, Leon Romanovsky wrote:
>>>> [...]
>>>>> diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
>>>>> index b486b8bbd1fcd..8be2a246d4903 100644
>>>>> --- a/drivers/iommu/iommu.c
>>>>> +++ b/drivers/iommu/iommu.c
>>>>> @@ -1879,7 +1879,7 @@ static int iommu_get_def_domain_type(struct iommu_group *group,
>>>>> static int iommu_get_default_domain_type(struct iommu_group *group,
>>>>> int target_type)
>>>>> {
>>>>> - struct device *untrusted = NULL;
>>>>> + struct device *restricted = NULL;
>>>>> struct group_device *gdev;
>>>>> int driver_type = 0;
>>>>> @@ -1901,14 +1901,15 @@ static int iommu_get_default_domain_type(struct iommu_group *group,
>>>>> driver_type = iommu_get_def_domain_type(group, gdev->dev,
>>>>> driver_type);
>>>>> - if (dev_is_pci(gdev->dev) && to_pci_dev(gdev->dev)->untrusted) {
>>>>> + if (dev_is_pci(gdev->dev) &&
>>>>> + pci_dev_default_is_restricted(to_pci_dev(gdev->dev))) {
>>>>
>>>> I think this should be the generic device_is_adversarial(gdev->dev) now - in
>>>> fact probably all the dev_is_pci()/to_pci_dev() stanzas aren't conceptually
>>>> PCI-specific, just an artefact of the current "untrusted" flag being so.
>>>
>>> I wanted to do the same, but iommu_get_default_domain_type() may be called
>>> before the PCI device's trust level has been set. At that point, we are
>>> still in "disabled" mode, whereas device_is_adversarial() checks for
>>> "adversarial" mode.
>>>
>>> The below is from AI, I asked to present me the path for that:
>>> --------------------------------------------------------------------------------
>>> For a newly discovered PCI device with an available IOMMU and a new IOMMU group:
>>> pci_scan_single_device()
>>> pci_device_add()
>>> device_add()
>>> bus_notify(dev, BUS_NOTIFY_ADD_DEVICE)
>>> blocking_notifier_call_chain()
>>> notifier_call_chain()
>>> iommu_bus_notifier()
>>> iommu_probe_device(dev)
>>> __iommu_probe_device(dev, NULL)
>>> iommu_setup_default_domain(group, 0)
>>> iommu_get_default_domain_type(group, 0)
>>>
>>> For a device whose initial policy is adversary, its state during that call is:
>>> dev->trust_policy = DEVICE_TRUST_POLICY_ADVERSARY
>>> dev->trust_level = DEVICE_TRUST_DISABLED
>>> dev->trust_committed_level = DEVICE_TRUST_DISABLED
>>> dev->driver = NULL
>>>
>>> device_is_adversarial(dev) = false
>>>
>>> The active level changes later, during binding:
>>> really_probe()
>>> device_trust_prepare()
>>> __device_trust_prepare()
>>> WRITE_ONCE(dev->trust_level, DEVICE_TRUST_ADVERSARY)
>>> dev->bus->dma_configure()
>>> call_driver_probe()
>>
>> OK, that seems less than ideal. Feels like it would be better if bus code
>> set the initial device state prior to device_add, so then it only ever
>> changes later as a result of an explicit attestation/acceptance process. If
>> the initial IOMMU probe still has to poke into bus-specific code to avoid
>> always making a worst-case assumption, doesn't that rather defeat the point
>> of having a generic abstraction?
>
> As Jason said, this is part of a very large series. Each series is an
> adventure in itself. Even Greg's seemingly simple request to convert the
> USB authorization knob to this new scheme has been extremely painful
> because of AI reviews. They keep pointing out bugs and gaps that are
> automatically attributed to me simply because I'm touching that part of
> the code.
>
>>
>> Even with the slightly dubious argument that this series only intends to
>> shuffle the existing PCI code around rather than fully introduce the generic
>> abstraction that it claims to, if the PCI behaviour is still just based on
>> propagating the internal "untrusted" flag plus the new global override, both
>> of which *are* known at add_device time into the common trust state, then
>> there seems little excuse for not doing that up-front so that the new API
>> *can* just work as expected.
>
> I would prefer to keep both PCI and USB out of this series on device trust
> levels. Let's see what people say if I drop them in v3.
I'd rather you did not as it is going to make experiments with CoCo VMs harder.
>
>>
>>>>
>>>>> /*
>>>>> - * No ARM32 using systems will set untrusted, it cannot
>>>>> - * work.
>>>>> + * No ARM32 using systems select a restricted trust
>>>>> + * policy; it cannot work.
>>>>> */
>>>>> if (WARN_ON(IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU)))
>>>>> return -1;
>>>>> - untrusted = gdev->dev;
>>>>> + restricted = gdev->dev;
>>>>> }
>>>>> }
>>>> [...]
>>>>> @@ -616,6 +608,28 @@ static inline struct pci_dev *to_pci_dev(const struct device *dev)
>>>>> {
>>>>> return container_of(dev, struct pci_dev, dev);
>>>>> }
>>>>> +
>>>>> +static inline bool
>>>>> +pci_dev_default_is_restricted(const struct pci_dev *pdev)
>>>>> +{
>>>>> + enum device_trust_policy policy = READ_ONCE(pdev->dev.trust_policy);
>>>>> +
>>>>> + return policy == DEVICE_TRUST_POLICY_DISABLED ||
>>>>> + policy == DEVICE_TRUST_POLICY_ADVERSARY;
>>>>> +}
>>>>
>>>> Unless I've wildly misunderstood, I'm not sure this is right - if the user
>>>> has set DEVICE_TRUST_POLICY_DISABLED to explicitly override the bus-default
>>>> behaviour, but that leads to the IOMMU layer special-casing all PCI devices
>>>> as adversarial and thus force-*enabling* all their trust-model restrictions,
>>>> isn't that pretty much the complete opposite of what was asked for?
>>>
>>> We have two levels: policy - that describes the default of all devices
>>> for that bus. level - actual decision on how that device is going to be
>>> operated.
>>>
>>> "Disabled" prevents driver probing until the device is authorized,
>>> "Adversary" permits probing and requests the available protections
>>> for operating an untrusted device.
>>>
>>> Both need conservative initial IOMMU setup: a disabled device may later
>>> be authorized as adversary, and authorization does not rebuild its IOMMU
>>> domain.
>>
>> Urgh, I can be absolutely certain I'm not the only one who is going to be
>> confused this way... I see "disabled" and instinctively take it to mean
>> disabling the device_trust *mechanism*, as the logical thing that people
>> will reach for in terms of "I don't care what this new thing is, just get it
>> out of my way and let my drivers work like they always used to; I want to
>> disable the new thing."
>
> I completely agree. I was under the impression that these names had already
> been decided, since I picked up where Dan left off.>>
>> From there it then also makes perfect self-consistent sense that
>> "device_trust=full" would be the opposite end of the scale as in "Apply the
>> explicit trust model 'fully', to all devices regardless of their bus-level
>> default."
>>
>> Now I guess you could name it something like "device.default_trust_level" to
>> attempt to make it clearer that it represents a value the value represents
>> the "trust" *of* each individual device itself, but TBH it's still liable to
>> look backwards to many users - similarly it's hard to guess that
>> FOO_POLICY_DISABLED might mean something entirely different from "don't
>> enforce the policy for FOO.", thus rather than try to contort the naming of
>> everything, I'd still be much more inclined to just flip the meanings to be
>> more intuitive from the outsider perspective.
>>
>> Say you upgrade your kernel and find your boot is broken because it now
>> flibbles, but at least there is a "flibble" parameter whose values are
>> "disabled", "wibble" and "frobble". Are you really going to spend time
>> straight off reading up on what flibbling is to understand what wibbles and
>> frobbles mean? Or is your instinct just going to be to try the seemingly
>> obvious way of disabling flibbling, and then getting positively annoyed if
>> that only makes it flibble harder?
>
> What about: blocked, adversarial and unrestricted
yes, that's better.
> instead of: disabled, adversarial and full
may be simplify: blocked, restricted, unrestricted? Thanks,
>
> Thanks
>
>>
>> Robin.
--
Alexey
next prev parent reply other threads:[~2026-11-18 7:34 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-09 7:33 [PATCH v2 00/19] driver core: Authorize device trust before driver binding Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 01/19] Input: gameport - attach manual drivers through the driver core Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 02/19] wifi: mac80211_hwsim: lock device for manual driver binding Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 03/19] USB: Lock registered interfaces during " Leon Romanovsky
2026-11-10 7:35 ` sashiko-bot
2026-11-10 14:19 ` Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 04/19] USB: Release interface claims refused during registration Leon Romanovsky
2026-11-10 7:35 ` sashiko-bot
2026-11-10 14:23 ` Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 05/19] PNP: Attach card devices through the driver core Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 06/19] cdx: Propagate DMA configuration errors Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 07/19] driver core: Report deferred probes through sysfs drivers_probe Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 08/19] driver core: Add a compatibility API for manual driver attachment Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 09/19] Input: serio - use the compatibility driver attachment API Leon Romanovsky
2026-11-10 7:35 ` sashiko-bot
2026-11-10 14:29 ` Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 10/19] net: phy: Use " Leon Romanovsky
2026-11-09 13:45 ` Andrew Lunn
2026-11-10 7:39 ` Leon Romanovsky
2026-11-10 14:31 ` Andrew Lunn
2026-11-10 14:58 ` Leon Romanovsky
2026-11-10 7:35 ` sashiko-bot
2026-11-09 7:33 ` [PATCH v2 11/19] driver core: Introduce a generic device trust level Leon Romanovsky
2026-11-10 7:35 ` sashiko-bot
2026-11-10 15:12 ` Leon Romanovsky
2026-11-13 11:30 ` Alexey Kardashevskiy
2026-11-16 12:11 ` Leon Romanovsky
2026-11-16 12:27 ` Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 12/19] driver core: Expose device trust authorization interfaces Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 13/19] driver core: Accept device trust policy on the kernel command line Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 14/19] driver core: Accept device trust policy by driver name Leon Romanovsky
2026-11-10 7:35 ` sashiko-bot
2026-11-09 7:33 ` [PATCH v2 15/19] driver core: Expose committed device trust level Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 16/19] driver core: Cover trust authorization and binding with KUnit Leon Romanovsky
2026-11-10 7:35 ` sashiko-bot
2026-11-09 7:33 ` [PATCH v2 17/19] driver core: Document trust admission and initrd authorization Leon Romanovsky
2026-11-09 7:33 ` [PATCH v2 18/19] PCI: Replace untrusted state with device trust Leon Romanovsky
2026-11-09 13:53 ` Robin Murphy
2026-11-10 12:22 ` Leon Romanovsky
2026-11-10 14:25 ` Jason Gunthorpe
2026-11-10 14:34 ` Leon Romanovsky
2026-11-16 17:13 ` Robin Murphy
2026-11-17 0:05 ` Jason Gunthorpe
2026-11-17 12:48 ` Leon Romanovsky
2026-11-18 7:33 ` Alexey Kardashevskiy [this message]
2026-11-10 7:35 ` sashiko-bot
2026-11-16 18:46 ` Kameron Carr
2026-11-16 19:04 ` Jason Gunthorpe
2026-11-16 19:24 ` Kameron Carr
2026-11-16 20:09 ` Jason Gunthorpe
2026-11-09 7:33 ` [PATCH v2 19/19] USB: Authorize devices and interfaces through " Leon Romanovsky
2026-11-10 7:35 ` sashiko-bot
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=fd25ed52-e3ea-41a2-9f62-2397d5d9e8c8@amd.com \
--to=aik@amd.com \
--cc=andrew@lunn.ch \
--cc=baolu.lu@linux.intel.com \
--cc=bhelgaas@google.com \
--cc=corbet@lwn.net \
--cc=dakr@kernel.org \
--cc=davem@davemloft.net \
--cc=dmitry.torokhov@gmail.com \
--cc=driver-core@lists.linux.dev \
--cc=dwmw2@infradead.org \
--cc=edumazet@kernel.org \
--cc=gregkh@linuxfoundation.org \
--cc=hkallweit1@gmail.com \
--cc=iommu@lists.linux.dev \
--cc=jgg@ziepe.ca \
--cc=johan@kernel.org \
--cc=johannes@sipsolutions.net \
--cc=joro@8bytes.org \
--cc=kuba@kernel.org \
--cc=leon@kernel.org \
--cc=linux-acpi@vger.kernel.org \
--cc=linux-bluetooth@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-input@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=linux-usb@vger.kernel.org \
--cc=linux-wireless@vger.kernel.org \
--cc=linux@armlinux.org.uk \
--cc=luiz.dentz@gmail.com \
--cc=marcel@holtmann.org \
--cc=netdev@vger.kernel.org \
--cc=nikhil.agarwal@amd.com \
--cc=nipun.gupta@amd.com \
--cc=pabeni@redhat.com \
--cc=pieter.jansen-van-vuuren@amd.com \
--cc=rafael@kernel.org \
--cc=rdunlap@infradead.org \
--cc=robin.murphy@arm.com \
--cc=skhan@linuxfoundation.org \
--cc=suravee.suthikulpanit@amd.com \
--cc=vasant.hegde@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;
as well as URLs for NNTP newsgroup(s).