Linux Confidential Computing Development
 help / color / mirror / Atom feed
From: Xu Yilun <yilun.xu@linux.intel.com>
To: Dan Williams <dan.j.williams@intel.com>
Cc: linux-coco@lists.linux.dev, Lukas Wunner <lukas@wunner.de>,
	Samuel Ortiz <sameo@rivosinc.com>,
	Alexey Kardashevskiy <aik@amd.com>,
	Bjorn Helgaas <bhelgaas@google.com>,
	linux-pci@vger.kernel.org, gregkh@linuxfoundation.org
Subject: Re: [PATCH 05/11] PCI/TSM: Authenticate devices via platform TSM
Date: Thu, 12 Dec 2024 17:50:18 +0800	[thread overview]
Message-ID: <Z1qx2nAHbZN72Ljf@yilunxu-OptiPlex-7050> (raw)
In-Reply-To: <173343742510.1074769.16552514658771224955.stgit@dwillia2-xfh.jf.intel.com>

> +static int pci_tsm_disconnect(struct pci_dev *pdev)
> +{
> +	struct pci_tsm *pci_tsm = pdev->tsm;
> +
> +	lockdep_assert_held(&pci_tsm_rwsem);
> +	if_not_guard(mutex_intr, &pci_tsm->lock)
> +		return -EINTR;
> +
> +	if (pci_tsm->state < PCI_TSM_CONNECT)
> +		return 0;
> +	if (pci_tsm->state < PCI_TSM_INIT)
> +		return -ENXIO;

Check PCI_TSM_INIT first, or this condition will never hit.

  if (pci_tsm->state < PCI_TSM_INIT)
	return -ENXIO;
  if (pci_tsm->state < PCI_TSM_CONNECT)
	return 0;

I suggest the same sequence for pci_tsm_connect().

> +
> +	tsm_ops->disconnect(pdev);
> +	pci_tsm->state = PCI_TSM_INIT;
> +
> +	return 0;
> +}
> +
> +static int pci_tsm_connect(struct pci_dev *pdev)
> +{
> +	struct pci_tsm *pci_tsm = pdev->tsm;
> +	int rc;
> +
> +	lockdep_assert_held(&pci_tsm_rwsem);
> +	if_not_guard(mutex_intr, &pci_tsm->lock)
> +		return -EINTR;
> +
> +	if (pci_tsm->state >= PCI_TSM_CONNECT)
> +		return 0;
> +	if (pci_tsm->state < PCI_TSM_INIT)
> +		return -ENXIO;
> +
> +	rc = tsm_ops->connect(pdev);
> +	if (rc)
> +		return rc;
> +	pci_tsm->state = PCI_TSM_CONNECT;
> +	return 0;
> +}
> +

[...]

> +
> +static void __pci_tsm_init(struct pci_dev *pdev)
> +{
> +	bool tee_cap;
> +
> +	if (!is_physical_endpoint(pdev))
> +		return;

This Filters out virtual functions, just because not ready for support,
is it?

> +
> +	tee_cap = pdev->devcap & PCI_EXP_DEVCAP_TEE;
> +
> +	if (!(pdev->ide_cap || tee_cap))
> +		return;
> +
> +	lockdep_assert_held_write(&pci_tsm_rwsem);
> +	if (!tsm_ops)
> +		return;
> +
> +	struct pci_tsm *pci_tsm __free(kfree) = kzalloc(sizeof(*pci_tsm), GFP_KERNEL);
> +	if (!pci_tsm)
> +		return;
> +
> +	/*
> +	 * If a physical device has any security capabilities it may be
> +	 * a candidate to connect with the platform TSM
> +	 */
> +	struct pci_dsm *dsm __free(dsm_remove) = tsm_ops->probe(pdev);

IIUC, pdev->tsm should be for every pci function (physical or virtual),
pdev->tsm->dsm should be only for physical functions, is it?

> +
> +	pci_dbg(pdev, "Device security capabilities detected (%s%s ), TSM %s\n",
> +		pdev->ide_cap ? " ide" : "", tee_cap ? " tee" : "",
> +		dsm ? "attach" : "skip");
> +
> +	if (!dsm)
> +		return;
> +
> +	mutex_init(&pci_tsm->lock);
> +	pci_tsm->doe_mb = pci_find_doe_mailbox(pdev, PCI_VENDOR_ID_PCI_SIG,
> +					       PCI_DOE_PROTO_CMA);
> +	if (!pci_tsm->doe_mb) {
> +		pci_warn(pdev, "TSM init failure, no CMA mailbox\n");
> +		return;
> +	}
> +
> +	pci_tsm->state = PCI_TSM_INIT;
> +	pci_tsm->dsm = no_free_ptr(dsm);
> +	pdev->tsm = no_free_ptr(pci_tsm);
> +	sysfs_update_group(&pdev->dev.kobj, &pci_tsm_auth_attr_group);
> +	sysfs_update_group(&pdev->dev.kobj, &pci_tsm_attr_group);
> +	if (pci_tsm_owner_attr_group)
> +		sysfs_merge_group(&pdev->dev.kobj, pci_tsm_owner_attr_group);
> +}

  parent reply	other threads:[~2024-12-12  9:53 UTC|newest]

Thread overview: 125+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-12-05 22:23 [PATCH 00/11] PCI/TSM: Core infrastructure for PCI device security (TDISP) Dan Williams
2024-12-05 22:23 ` [PATCH 01/11] configfs-tsm: Namespace TSM report symbols Dan Williams
2024-12-10  6:08   ` Alexey Kardashevskiy
2024-12-11 13:55   ` Suzuki K Poulose
2024-12-05 22:23 ` [PATCH 02/11] coco/guest: Move shared guest CC infrastructure to drivers/virt/coco/guest/ Dan Williams
2024-12-10  6:09   ` Alexey Kardashevskiy
2024-12-05 22:23 ` [PATCH 03/11] coco/tsm: Introduce a class device for TEE Security Managers Dan Williams
2025-01-28 12:17   ` Jonathan Cameron
2025-02-25 21:08     ` Dan Williams
2024-12-05 22:23 ` [PATCH 04/11] PCI/IDE: Selective Stream IDE enumeration Dan Williams
2024-12-10  3:08   ` Aneesh Kumar K.V
2024-12-12  6:32     ` Xu Yilun
2025-02-22  0:42       ` Dan Williams
2025-02-20  3:17     ` Dan Williams
2024-12-10  6:18   ` Alexey Kardashevskiy
2025-02-20  3:59     ` Dan Williams
2024-12-10  7:05   ` Alexey Kardashevskiy
2024-12-12  6:06     ` Xu Yilun
2024-12-18 10:35       ` Alexey Kardashevskiy
2025-02-22  0:30       ` Dan Williams
2025-02-20 18:07     ` Dan Williams
2025-02-21  0:53       ` Alexey Kardashevskiy
2025-02-27 23:46         ` Dan Williams
2024-12-10 19:24   ` Bjorn Helgaas
2025-02-22  0:13     ` Dan Williams
2025-01-30 10:45   ` Jonathan Cameron
2025-02-26  0:21     ` Dan Williams
2024-12-05 22:23 ` [PATCH 05/11] PCI/TSM: Authenticate devices via platform TSM Dan Williams
2024-12-10 10:18   ` Alexey Kardashevskiy
2025-02-21  8:13     ` Aneesh Kumar K.V
2025-02-25  7:17       ` Xu Yilun
2025-02-26 12:10         ` Aneesh Kumar K.V
2025-02-26 12:13           ` [RFC PATCH 1/7] tsm: Select PCI_DOE which is required for PCI_TSM Aneesh Kumar K.V (Arm)
2025-02-26 12:13             ` [RFC PATCH 2/7] tsm: Move tsm core outside the host directory Aneesh Kumar K.V (Arm)
2025-02-26 12:13             ` [RFC PATCH 3/7] tsm: vfio: Add tsm bind/unbind support Aneesh Kumar K.V (Arm)
2025-02-26 12:13             ` [RFC PATCH 4/7] tsm: Allow tsm ops function to be called for multi-function devices Aneesh Kumar K.V (Arm)
2025-02-26 12:13             ` [RFC PATCH 5/7] tsm: Don't error out for doe mailbox failure Aneesh Kumar K.V (Arm)
2025-02-26 12:13             ` [RFC PATCH 6/7] tsm: Allow tsm connect ops to be used for multiple operations Aneesh Kumar K.V (Arm)
2025-02-26 12:13             ` [RFC PATCH 7/7] tsm: Add secure SPDM support Aneesh Kumar K.V (Arm)
2025-02-27  6:50               ` Xu Yilun
2025-02-27  6:35           ` [PATCH 05/11] PCI/TSM: Authenticate devices via platform TSM Xu Yilun
2025-02-27 13:57             ` Aneesh Kumar K.V
2025-02-28  1:26               ` Xu Yilun
2025-02-28  9:48                 ` Aneesh Kumar K.V
2025-03-01  7:50                   ` Xu Yilun
2025-03-07  3:07                   ` Alexey Kardashevskiy
2025-02-27 19:53           ` Dan Williams
2025-02-28 10:06             ` Aneesh Kumar K.V
2025-02-21 20:42     ` Dan Williams
2025-02-25  4:45       ` Alexey Kardashevskiy
2025-02-28  3:09         ` Dan Williams
2024-12-10 18:52   ` Bjorn Helgaas
2025-02-21 22:32     ` Dan Williams
2024-12-12  9:50   ` Xu Yilun [this message]
2025-02-22  1:15     ` Dan Williams
2025-02-24 11:02       ` Xu Yilun
2025-02-28  0:15         ` Dan Williams
2025-02-28  9:39           ` Xu Yilun
2025-01-30 11:45   ` Jonathan Cameron
2025-02-26  0:50     ` Dan Williams
2024-12-05 22:23 ` [PATCH 06/11] samples/devsec: PCI device-security bus / endpoint sample Dan Williams
2024-12-06  4:23   ` kernel test robot
2024-12-09  3:40   ` kernel test robot
2025-01-30 13:21   ` Jonathan Cameron
2025-02-26  2:00     ` Dan Williams
2024-12-05 22:23 ` [PATCH 07/11] PCI: Add PCIe Device 3 Extended Capability enumeration Dan Williams
2024-12-09 13:17   ` Ilpo Järvinen
2025-02-20  3:05     ` Dan Williams
2025-02-20  3:09       ` Dan Williams
2024-12-10 19:21   ` Bjorn Helgaas
2024-12-11 13:22     ` Ilpo Järvinen
2025-02-22  0:15       ` Dan Williams
2025-02-24 15:09         ` Ilpo Järvinen
2025-02-28  0:29           ` Dan Williams
2025-02-21 23:34     ` Dan Williams
2025-02-25  2:25       ` Alexey Kardashevskiy
2024-12-05 22:24 ` [PATCH 08/11] PCI/IDE: Add IDE establishment helpers Dan Williams
2024-12-10  3:19   ` Aneesh Kumar K.V
2024-12-10  3:37     ` Aneesh Kumar K.V
2025-02-20  3:39       ` Dan Williams
2025-02-21 15:53         ` Aneesh Kumar K.V
2025-02-25  0:46           ` Dan Williams
2025-01-07 20:19     ` Xu Yilun
2025-01-10 13:25       ` Aneesh Kumar K.V
2025-02-24 22:31         ` Dan Williams
2025-02-25  2:29           ` Alexey Kardashevskiy
2025-02-20  3:28     ` Dan Williams
2024-12-10  7:07   ` Alexey Kardashevskiy
2025-02-20 21:44     ` Dan Williams
2024-12-10 18:47   ` Bjorn Helgaas
2025-02-21 22:02     ` Dan Williams
2024-12-12 10:50   ` Xu Yilun
2024-12-19  7:25   ` Alexey Kardashevskiy
2024-12-19 10:05     ` Alexey Kardashevskiy
2025-01-07 20:00       ` Xu Yilun
2025-01-09  2:35         ` Alexey Kardashevskiy
2025-01-09 21:28           ` Xu Yilun
2025-01-15  0:20             ` Alexey Kardashevskiy
2025-02-25  0:06               ` Dan Williams
2025-02-25  3:39                 ` Alexey Kardashevskiy
2025-02-28  2:26                   ` Dan Williams
2025-03-04  0:03                     ` Alexey Kardashevskiy
2025-03-04  0:57                       ` Dan Williams
2025-03-04  1:31                         ` Alexey Kardashevskiy
2025-03-04 17:59                           ` Dan Williams
2025-02-20  4:19             ` Alexey Kardashevskiy
2025-02-24 22:24         ` Dan Williams
2025-02-25  2:45           ` Xu Yilun
2025-02-24 20:28       ` Dan Williams
2025-02-26  1:54         ` Alexey Kardashevskiy
2025-02-24 20:24     ` Dan Williams
2025-02-25  5:01       ` Xu Yilun
2024-12-05 22:24 ` [PATCH 09/11] PCI/IDE: Report available IDE streams Dan Williams
2024-12-06  0:12   ` kernel test robot
2024-12-06  0:43   ` kernel test robot
2025-02-11  6:10   ` Alexey Kardashevskiy
2025-02-27 23:35     ` Dan Williams
2024-12-05 22:24 ` [PATCH 10/11] PCI/TSM: Report active " Dan Williams
2024-12-10 18:49   ` Bjorn Helgaas
2025-02-21 22:28     ` Dan Williams
2024-12-05 22:24 ` [PATCH 11/11] samples/devsec: Add sample IDE establishment Dan Williams
2025-01-30 13:39   ` Jonathan Cameron
2025-02-27 23:27     ` Dan Williams
2024-12-06  6:05 ` [PATCH 00/11] PCI/TSM: Core infrastructure for PCI device security (TDISP) Greg KH
2024-12-06  8:44   ` Dan Williams

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=Z1qx2nAHbZN72Ljf@yilunxu-OptiPlex-7050 \
    --to=yilun.xu@linux.intel.com \
    --cc=aik@amd.com \
    --cc=bhelgaas@google.com \
    --cc=dan.j.williams@intel.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=linux-coco@lists.linux.dev \
    --cc=linux-pci@vger.kernel.org \
    --cc=lukas@wunner.de \
    --cc=sameo@rivosinc.com \
    /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