Linux IOMMU Development
 help / color / mirror / Atom feed
From: Vasant Hegde <vasant.hegde@amd.com>
To: <iommu@lists.linux.dev>, <joro@8bytes.org>
Cc: <suravee.suthikulpanit@amd.com>, <wei.huang2@amd.com>,
	<jsnitsel@redhat.com>, <jgg@ziepe.ca>,
	Vasant Hegde <vasant.hegde@amd.com>,
	Jason Gunthorpe <jgg@nvidia.com>
Subject: [PATCH v6 12/14] iommu/amd: Introduce iommu_dev_data.flags to track device capabilities
Date: Thu, 21 Sep 2023 09:21:45 +0000	[thread overview]
Message-ID: <20230921092147.5930-13-vasant.hegde@amd.com> (raw)
In-Reply-To: <20230921092147.5930-1-vasant.hegde@amd.com>

Currently we use struct iommu_dev_data.iommu_v2 to keep track of the device
ATS, PRI, and PASID capabilities. But these capabilities can be enabled
independently (except PRI requires ATS support). Hence, replace
the iommu_v2 variable with a flags variable, which keep track of the device
capabilities.

From commit 9bf49e36d718 ("PCI/ATS: Handle sharing of PF PRI Capability
with all VFs"), device PRI/PASID is shared between PF and any associated
VFs. Hence use pci_pri_supported() and pci_pasid_features() instead of
pci_find_ext_capability() to check device PRI/PASID support.

Signed-off-by: Vasant Hegde <vasant.hegde@amd.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Jerry Snitselaar <jsnitsel@redhat.com>
---
 drivers/iommu/amd/amd_iommu_types.h |  3 +-
 drivers/iommu/amd/iommu.c           | 46 ++++++++++++++++++-----------
 2 files changed, 30 insertions(+), 19 deletions(-)

diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h
index 87fa39c490e8..07f8919ee34e 100644
--- a/drivers/iommu/amd/amd_iommu_types.h
+++ b/drivers/iommu/amd/amd_iommu_types.h
@@ -811,7 +811,8 @@ struct iommu_dev_data {
 	struct protection_domain *domain; /* Domain the device is bound to */
 	struct device *dev;
 	u16 devid;			  /* PCI Device ID */
-	bool iommu_v2;			  /* Device can make use of IOMMUv2 */
+
+	u32 flags;			  /* Holds AMD_IOMMU_DEVICE_FLAG_<*> */
 	int ats_qdep;
 	u8 ats_enabled  :1;		  /* ATS state */
 	u8 pri_tlp      :1;		  /* PASID TLB required for
diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c
index 4ee69ccdc17b..151f405686e3 100644
--- a/drivers/iommu/amd/iommu.c
+++ b/drivers/iommu/amd/iommu.c
@@ -319,24 +319,34 @@ static struct iommu_group *acpihid_device_group(struct device *dev)
 	return entry->group;
 }
 
-static bool pci_iommuv2_capable(struct pci_dev *pdev)
+static inline bool pdev_pasid_supported(struct iommu_dev_data *dev_data)
 {
-	static const int caps[] = {
-		PCI_EXT_CAP_ID_PRI,
-		PCI_EXT_CAP_ID_PASID,
-	};
-	int i, pos;
+	return (dev_data->flags & AMD_IOMMU_DEVICE_FLAG_PASID_SUP);
+}
 
-	if (!pci_ats_supported(pdev))
-		return false;
+static u32 pdev_get_caps(struct pci_dev *pdev)
+{
+	int features;
+	u32 flags = 0;
+
+	if (pci_ats_supported(pdev))
+		flags |= AMD_IOMMU_DEVICE_FLAG_ATS_SUP;
+
+	if (pci_pri_supported(pdev))
+		flags |= AMD_IOMMU_DEVICE_FLAG_PRI_SUP;
 
-	for (i = 0; i < 2; ++i) {
-		pos = pci_find_ext_capability(pdev, caps[i]);
-		if (pos == 0)
-			return false;
+	features = pci_pasid_features(pdev);
+	if (features >= 0) {
+		flags |= AMD_IOMMU_DEVICE_FLAG_PASID_SUP;
+
+		if (features & PCI_PASID_CAP_EXEC)
+			flags |= AMD_IOMMU_DEVICE_FLAG_EXEC_SUP;
+
+		if (features & PCI_PASID_CAP_PRIV)
+			flags |= AMD_IOMMU_DEVICE_FLAG_PRIV_SUP;
 	}
 
-	return true;
+	return flags;
 }
 
 /*
@@ -396,8 +406,8 @@ static int iommu_init_device(struct amd_iommu *iommu, struct device *dev)
 	 * it'll be forced to go into translation mode.
 	 */
 	if ((iommu_default_passthrough() || !amd_iommu_force_isolation) &&
-	    dev_is_pci(dev) && pci_iommuv2_capable(to_pci_dev(dev))) {
-		dev_data->iommu_v2 = amd_iommu_gt_ppr_supported();
+	    dev_is_pci(dev) && amd_iommu_gt_ppr_supported()) {
+		dev_data->flags = pdev_get_caps(to_pci_dev(dev));
 	}
 
 	dev_iommu_priv_set(dev, dev_data);
@@ -1850,7 +1860,7 @@ static int attach_device(struct device *dev,
 			goto out;
 		}
 
-		if (dev_data->iommu_v2) {
+		if (pdev_pasid_supported(dev_data)) {
 			if (pdev_pri_ats_enable(pdev) != 0)
 				goto out;
 
@@ -1916,7 +1926,7 @@ static void detach_device(struct device *dev)
 	if (!dev_is_pci(dev))
 		goto out;
 
-	if (domain->flags & PD_IOMMUV2_MASK && dev_data->iommu_v2)
+	if (domain->flags & PD_IOMMUV2_MASK && pdev_pasid_supported(dev_data))
 		pdev_iommuv2_disable(to_pci_dev(dev));
 	else if (dev_data->ats_enabled)
 		pci_disable_ats(to_pci_dev(dev));
@@ -2471,7 +2481,7 @@ static int amd_iommu_def_domain_type(struct device *dev)
 	 *    and require remapping.
 	 *  - SNP is enabled, because it prohibits DTE[Mode]=0.
 	 */
-	if (dev_data->iommu_v2 &&
+	if (pdev_pasid_supported(dev_data) &&
 	    !cc_platform_has(CC_ATTR_MEM_ENCRYPT) &&
 	    !amd_iommu_snp_en) {
 		return IOMMU_DOMAIN_IDENTITY;
-- 
2.31.1


  parent reply	other threads:[~2023-09-21  9:28 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-21  9:21 [PATCH v6 00/14] iommu/amd: SVA Support (Part 1) - cleanup/refactoring Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 01/14] iommu/amd: Remove unused amd_io_pgtable.pt_root variable Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 02/14] iommu/amd: Consolidate timeout pre-define to amd_iommu_type.h Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 03/14] iommu/amd: Consolidate logic to allocate protection domain Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 04/14] iommu/amd: Refactor protection domain allocation code Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 05/14] iommu/amd: Introduce helper functions for managing GCR3 table Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 06/14] iommu/amd: Do not set amd_iommu_pgtable in pass-through mode Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 07/14] iommu/amd: Miscellaneous clean up when free domain Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 08/14] iommu/amd: Consolidate feature detection and reporting logic Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 09/14] iommu/amd: Modify logic for checking GT and PPR features Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 10/14] iommu/amd: Rename ats related variables Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 11/14] iommu/amd: Introduce iommu_dev_data.ppr Vasant Hegde
2023-09-21  9:21 ` Vasant Hegde [this message]
2023-09-21  9:21 ` [PATCH v6 13/14] iommu/amd: Enable device ATS/PASID/PRI capabilities independently Vasant Hegde
2023-09-21  9:21 ` [PATCH v6 14/14] iommu/amd: Initialize iommu_device->max_pasids Vasant Hegde
2023-09-21  9:49 ` [PATCH v6 00/14] iommu/amd: SVA Support (Part 1) - cleanup/refactoring Vasant Hegde
2023-09-21 17:21   ` Jason Gunthorpe
2023-09-25 10:40     ` Joerg Roedel
2023-09-25 11:02       ` Vasant Hegde

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=20230921092147.5930-13-vasant.hegde@amd.com \
    --to=vasant.hegde@amd.com \
    --cc=iommu@lists.linux.dev \
    --cc=jgg@nvidia.com \
    --cc=jgg@ziepe.ca \
    --cc=joro@8bytes.org \
    --cc=jsnitsel@redhat.com \
    --cc=suravee.suthikulpanit@amd.com \
    --cc=wei.huang2@amd.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