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>
Subject: [PATCH v7 00/17] iommu/amd: SVA Support (part 3) - refactor support for GCR3 table
Date: Mon, 5 Feb 2024 11:55:58 +0000	[thread overview]
Message-ID: <20240205115615.6053-1-vasant.hegde@amd.com> (raw)

This is part 3 of the 4-part series to introduce Share Virtual
Address (SVA) support, which focuses on refactoring GCR3 table.
This moves GCR3 related information from protection domain
structure to iommu_dev_data (per device) structure.

It contains the following enhancements:

* Patch 1 - 3:
  Cleanup and introduce helper function

* Patch 4 - 13:
  Introduce per device GCR3 table, invalidation based on per device and
  code refactoring. Also introduce iommu_group_mutex_assert.

* Patch 14 - 16:
  Remove unused variable, rearrange functions to make it easy to maintain

* Patch 17:
  Introduce per device domain ID.

This patch series is based on top of iommu/next branch (commit d2d00e15808).

This is also available at github :
  https://github.com/AMDESE/linux/tree/iommu_sva_part3_v7_v6.8_rc2

Thanks Jason for reviewing previous versions and providing valuable feedbacks.

Changes from v6 -> v7:
  - Added Reviewed-by tags
  - Rebased on top of iommu/next branch
  - iommu: Introduce iommu_group_mutex_assert()
    Removed 'extern' from iommu_group_mutex_assert() declaration
  - iommu/amd: Refactor helper function for setting / clearing GCR3
    Followed-success oriented flow
  - Minor update to patch 11/17 description

v6: https://lore.kernel.org/linux-iommu/20240125121135.8217-1-vasant.hegde@amd.com/T/#t

Changes from v5 -> v6:
  - Moved per-device-domain-ID patch to end
  - Added condition check for iommu_group_mutex_assert()
  - Moving per-device-domain-id allocation/free to setup_gcr3_table()/free_gcr3_table
  - Pass iommu to setup_gcr3_table instead of passing nid

v5: https://lore.kernel.org/linux-iommu/20240116165335.6043-1-vasant.hegde@amd.com/T/#t

Changes from v4 -> v5:
  - Added get_amd_iommu_from_dev_data()
  - pass gcr3_info instead of dev_data to setup/free_gcr3_table()
  - Added path to EXPORT iommu_group_mutex_assert
  - Moved dev_data->domid to gcr3_info->domid


v4: https://lore.kernel.org/linux-iommu/20231212085224.6985-1-vasant.hegde@amd.com/T/#t

Changes from v3 -> v4:
  - Added new patch to enable IOMMU GT feature early
  - Fixed dirty tracking path to use get_amd_iommu_from_dev()
  - Fixed duplicate flush issue in do_attach() path
  - Removed gcr3_info->giov variable. Instead we use domain page table mode to decide
    DTE[GIOV] bit.
  - Free gcr3_table before calling set_dte_entry (otherwise there is a possibility
    of its setting up gcr3 table with glx 0)
  - Adjusted gfp_t flag for GCR3 table allocation (GPF_ATOMIC -> GFP_KERNEL)
  - Moved below patches from invalidation improvement series to this series
    * Remove unused flush_pasid() related functions
      (Originally it was part of Invalidation improvement series, I had to push it to
       this series as set_gcr3() was still using __amd_iommu_flush_tlb()).
    * Added patch to Rearrange device flush code

v3 : https://lore.kernel.org/linux-iommu/20231013151652.6008-1-vasant.hegde@amd.com/T/#t

Changes from v2 -> v3:
  - Dropped "iommu/amd: Use struct protection_domain in helper functions" as flush related
    functions are already fixed and set_gcr3 related functions is reworked in this series.
  - Removed moving to_pdomain() function from this series. It will be introduced
    in SVA series where it will be used.
  - Added per device domain ID and per device invalidation functions
  - Added functions to flush TLB for the given device
  - Added Review-by tags

v2 : https://lore.kernel.org/linux-iommu/20230816174031.634453-1-vasant.hegde@amd.com/T/#t

Changes from v1 - v2:
  - Dropped iommu_v2 module related support as newly introduced Part2
    removed iommu_v2 module.
  - Moved Patch 'Use struct protection_domain in helper functions' from Part1 to
    Part3
  - Updated get_amd_iommu_from_dev() to retrieve iommu from device structure
  - Removed 'PD_MODE_PT'

v1 : https://lore.kernel.org/linux-iommu/20230808100232.5977-1-vasant.hegde@amd.com/T/#t

Thank you,
Vasant / Suravee

Suravee Suthikulpanit (8):
  iommu/amd: Introduce get_amd_iommu_from_dev()
  iommu/amd: Introduce struct protection_domain.pd_mode
  iommu/amd: Introduce per-device GCR3 table
  iommu/amd: Refactor helper function for setting / clearing GCR3
  iommu/amd: Refactor attaching / detaching device functions
  iommu/amd: Refactor protection_domain helper functions
  iommu/amd: Refactor GCR3 table helper functions
  iommu/amd: Remove unused GCR3 table parameters from struct protection_domain

Vasant Hegde (9):
  iommu/amd: Pass struct iommu_dev_data to set_dte_entry()
  iommu/amd: Enable Guest Translation before registering devices
  iommu/amd: Use protection_domain.flags to check page table mode
  iommu/amd: Add support for device based TLB invalidation
  iommu/amd: Rearrange GCR3 table setup code
  iommu: Introduce iommu_group_mutex_assert()
  iommu/amd: Remove unused flush pasid functions
  iommu/amd: Rearrange device flush code
  iommu/amd: Introduce per-device domain ID to fix potential TLB
    aliasing issue

 drivers/iommu/amd/amd_iommu.h       |  29 +-
 drivers/iommu/amd/amd_iommu_types.h |  26 +-
 drivers/iommu/amd/init.c            |   6 +-
 drivers/iommu/amd/io_pgtable_v2.c   |  21 +-
 drivers/iommu/amd/iommu.c           | 627 +++++++++++++---------------
 drivers/iommu/iommu.c               |  19 +
 include/linux/iommu.h               |  24 ++
 7 files changed, 376 insertions(+), 376 deletions(-)

-- 
2.31.1


             reply	other threads:[~2024-02-05 11:56 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-02-05 11:55 Vasant Hegde [this message]
2024-02-05 11:55 ` [PATCH v7 01/17] iommu/amd: Pass struct iommu_dev_data to set_dte_entry() Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 02/17] iommu/amd: Enable Guest Translation before registering devices Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 03/17] iommu/amd: Introduce get_amd_iommu_from_dev() Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 04/17] iommu/amd: Introduce struct protection_domain.pd_mode Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 05/17] iommu/amd: Introduce per-device GCR3 table Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 06/17] iommu/amd: Use protection_domain.flags to check page table mode Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 07/17] iommu/amd: Add support for device based TLB invalidation Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 08/17] iommu/amd: Rearrange GCR3 table setup code Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 09/17] iommu: Introduce iommu_group_mutex_assert() Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 10/17] iommu/amd: Refactor helper function for setting / clearing GCR3 Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 11/17] iommu/amd: Refactor attaching / detaching device functions Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 12/17] iommu/amd: Refactor protection_domain helper functions Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 13/17] iommu/amd: Refactor GCR3 table " Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 14/17] iommu/amd: Remove unused flush pasid functions Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 15/17] iommu/amd: Rearrange device flush code Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 16/17] iommu/amd: Remove unused GCR3 table parameters from struct protection_domain Vasant Hegde
2024-02-05 11:56 ` [PATCH v7 17/17] iommu/amd: Introduce per-device domain ID to fix potential TLB aliasing issue Vasant Hegde
2024-02-09 10:44 ` [PATCH v7 00/17] iommu/amd: SVA Support (part 3) - refactor support for GCR3 table 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=20240205115615.6053-1-vasant.hegde@amd.com \
    --to=vasant.hegde@amd.com \
    --cc=iommu@lists.linux.dev \
    --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