Skip to contentSkip to conversation

IOMMU Updates for Linux v4.1

GIT PULL Joerg Roedel 4 participants 7 messages Apr 20, 2015 – May 6, 2015

lore
Share & tools
lore
#0 lore raw

Hi Linus,

The following changes since commit e42391cd048809d903291d07f86ed3934ce138e9:

Linux 4.0-rc6 (2015-03-29 15:26:31 -0700)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu.git tags/iommu-updates-v4.1

for you to fetch changes up to 7f65ef01e131650d455875598099cd06fea6096b:

Merge branches 'iommu/fixes', 'x86/vt-d', 'x86/amd', 'arm/smmu', 'arm/tegra' and 'core' into next (2015-04-02 13:33:19 +0200)

----------------------------------------------------------------

IOMMU Updates for Linux v4.1

Not much this time, but the changes include:

* Moving domain allocation into the iommu drivers to prepare for the introduction of default domains for devices

* Fixing the IO page-table code in the AMD IOMMU driver to correctly encode large page sizes

* Extension of the PCI support in the ARM-SMMU driver

* Various fixes and cleanups

----------------------------------------------------------------
Dan Carpenter (1):
iommu/amd: Small cleanup in mn_release()

Joerg Roedel (27):
iommu: Introduce domain_alloc and domain_free iommu_ops
iommu: Introduce iommu domain types
iommu: Only allow iommu_map/unmap for paging domains
iommu/amd: Make use of domain_alloc and domain_free
iommu/vt-d: Make use of domain_alloc and domain_free
iommu/omap: Make use of domain_alloc and domain_free
iommu/arm-smmu: Make use of domain_alloc and domain_free
iommu/exynos: Make use of domain_alloc and domain_free
iommu/tegra-smmu: Make use of domain_alloc and domain_free
iommu/tegra-gart: Make use of domain_alloc and domain_free
iommu/msm: Make use of domain_alloc and domain_free
iommu/shmobile: Make use of domain_alloc and domain_free
iommu/ipmmu-vmsa: Make use of domain_alloc and domain_free
iommu/rockchip: Make use of domain_alloc and domain_free
iommu/fsl: Make use of domain_alloc and domain_free
iommu: Remove domain_init and domain_free iommu_ops
Merge branch 'for-joerg/arm-smmu/updates' of git://git.kernel.org/.../will/linux into arm/smmu iommu/amd: Use BUS_NOTIFY_REMOVED_DEVICE
iommu/amd: Ignore BUS_NOTIFY_UNBOUND_DRIVER event
iommu/amd: Don't allocate with __GFP_ZERO in alloc_coherent
iommu/amd: Add support for contiguous dma allocator
iommu/amd: Return the pte page-size in fetch_pte
iommu/amd: Optimize iommu_unmap_page for new fetch_pte interface
iommu/amd: Optimize alloc_new_range for new fetch_pte interface
iommu/amd: Optimize amd_iommu_iova_to_phys for new fetch_pte interface iommu/amd: Correctly encode huge pages in iommu page tables
Merge branches 'iommu/fixes', 'x86/vt-d', 'x86/amd', 'arm/smmu', 'arm/tegra' and 'core' into next

Quentin Lambert (1):
iommu/vt-d: Convert non-returned local variable to boolean when relevant

Robin Murphy (1):
iommu/arm-smmu: set a more appropriate DMA mask

Thierry Reding (3):
iommu/tegra: Setup aperture
iommu/tegra: gart: Set aperture at domain initialization time iommu/tegra: smmu: Compute PFN mask at runtime

Will Deacon (3):
iommu/arm-smmu: ensure CBA2R is initialised before CBAR on SMMUv2 iommu/arm-smmu: handle multi-alias IOMMU groups for PCI devices iommu/io-pgtable-arm: avoid speculative walks through TTBR1

 drivers/iommu/amd_iommu.c           | 250 ++++++++++++++++++------------------
 drivers/iommu/amd_iommu_types.h     |  13 +-
 drivers/iommu/amd_iommu_v2.c        |   2 +-
 drivers/iommu/arm-smmu.c            | 180 ++++++++++++++++----------
 drivers/iommu/exynos-iommu.c        |  87 +++++++------
 drivers/iommu/fsl_pamu_domain.c     |  60 +++++----
 drivers/iommu/fsl_pamu_domain.h     |   2 +-
 drivers/iommu/intel-iommu.c         |  68 +++++-----
 drivers/iommu/intel_irq_remapping.c |  12 +-
 drivers/iommu/io-pgtable-arm.c      |   5 +
 drivers/iommu/iommu.c               |  26 ++--
 drivers/iommu/ipmmu-vmsa.c          |  42 +++---
 drivers/iommu/msm_iommu.c           |  73 ++++++-----
 drivers/iommu/omap-iommu.c          |  49 ++++---
 drivers/iommu/rockchip-iommu.c      |  40 +++---
 drivers/iommu/shmobile-iommu.c      |  39 +++---
 drivers/iommu/tegra-gart.c          |  88 ++++++++-----
 drivers/iommu/tegra-smmu.c          |  59 +++++----
 include/linux/iommu.h               |  33 ++++-
 19 files changed, 644 insertions(+), 484 deletions(-)

Please pull.

Thanks,

Joerg

#1 lore raw

Hi Joerg,

I sent to you (cc iommu mailing list) an important fix on Apr-16 for a bug in amd_iommu_v2.c (from kernel 4.0). See:
http://lists.linuxfoundation.org/pipermail/iommu/2015-April/012757.html

Could you please add that to the pull request ? Its creating problems when running HSA applications.

Thanks,

Oded

Quoted 106 lines · Joerg Roedel

Hi Linus,

The following changes since commit e42391cd048809d903291d07f86ed3934ce138e9:

   Linux 4.0-rc6 (2015-03-29 15:26:31 -0700)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu.git tags/iommu-updates-v4.1

for you to fetch changes up to 7f65ef01e131650d455875598099cd06fea6096b:

   Merge branches 'iommu/fixes', 'x86/vt-d', 'x86/amd', 'arm/smmu', 'arm/tegra' and 'core' into next (2015-04-02 13:33:19 +0200)

----------------------------------------------------------------

IOMMU Updates for Linux v4.1

Not much this time, but the changes include:

* Moving domain allocation into the iommu drivers to prepare for the introduction of default domains for devices

* Fixing the IO page-table code in the AMD IOMMU driver to correctly encode large page sizes

* Extension of the PCI support in the ARM-SMMU driver

* Various fixes and cleanups

----------------------------------------------------------------
Dan Carpenter (1):
iommu/amd: Small cleanup in mn_release()

Joerg Roedel (27):
iommu: Introduce domain_alloc and domain_free iommu_ops
iommu: Introduce iommu domain types
iommu: Only allow iommu_map/unmap for paging domains
iommu/amd: Make use of domain_alloc and domain_free
iommu/vt-d: Make use of domain_alloc and domain_free
iommu/omap: Make use of domain_alloc and domain_free
iommu/arm-smmu: Make use of domain_alloc and domain_free
iommu/exynos: Make use of domain_alloc and domain_free
iommu/tegra-smmu: Make use of domain_alloc and domain_free
iommu/tegra-gart: Make use of domain_alloc and domain_free
iommu/msm: Make use of domain_alloc and domain_free
iommu/shmobile: Make use of domain_alloc and domain_free
iommu/ipmmu-vmsa: Make use of domain_alloc and domain_free
iommu/rockchip: Make use of domain_alloc and domain_free
iommu/fsl: Make use of domain_alloc and domain_free
iommu: Remove domain_init and domain_free iommu_ops
Merge branch 'for-joerg/arm-smmu/updates' of git://git.kernel.org/.../will/linux into arm/smmu iommu/amd: Use BUS_NOTIFY_REMOVED_DEVICE
iommu/amd: Ignore BUS_NOTIFY_UNBOUND_DRIVER event
iommu/amd: Don't allocate with __GFP_ZERO in alloc_coherent
iommu/amd: Add support for contiguous dma allocator
iommu/amd: Return the pte page-size in fetch_pte
iommu/amd: Optimize iommu_unmap_page for new fetch_pte interface
iommu/amd: Optimize alloc_new_range for new fetch_pte interface
iommu/amd: Optimize amd_iommu_iova_to_phys for new fetch_pte interface iommu/amd: Correctly encode huge pages in iommu page tables
Merge branches 'iommu/fixes', 'x86/vt-d', 'x86/amd', 'arm/smmu', 'arm/tegra' and 'core' into next

Quentin Lambert (1):
iommu/vt-d: Convert non-returned local variable to boolean when relevant

Robin Murphy (1):
iommu/arm-smmu: set a more appropriate DMA mask

Thierry Reding (3):
iommu/tegra: Setup aperture
iommu/tegra: gart: Set aperture at domain initialization time iommu/tegra: smmu: Compute PFN mask at runtime

Will Deacon (3):
iommu/arm-smmu: ensure CBA2R is initialised before CBAR on SMMUv2 iommu/arm-smmu: handle multi-alias IOMMU groups for PCI devices iommu/io-pgtable-arm: avoid speculative walks through TTBR1

  drivers/iommu/amd_iommu.c           | 250 ++++++++++++++++++------------------
  drivers/iommu/amd_iommu_types.h     |  13 +-
  drivers/iommu/amd_iommu_v2.c        |   2 +-
  drivers/iommu/arm-smmu.c            | 180 ++++++++++++++++----------
  drivers/iommu/exynos-iommu.c        |  87 +++++++------
  drivers/iommu/fsl_pamu_domain.c     |  60 +++++----
  drivers/iommu/fsl_pamu_domain.h     |   2 +-
  drivers/iommu/intel-iommu.c         |  68 +++++-----
  drivers/iommu/intel_irq_remapping.c |  12 +-
  drivers/iommu/io-pgtable-arm.c      |   5 +
  drivers/iommu/iommu.c               |  26 ++--
  drivers/iommu/ipmmu-vmsa.c          |  42 +++---
  drivers/iommu/msm_iommu.c           |  73 ++++++-----
  drivers/iommu/omap-iommu.c          |  49 ++++---
  drivers/iommu/rockchip-iommu.c      |  40 +++---
  drivers/iommu/shmobile-iommu.c      |  39 +++---
  drivers/iommu/tegra-gart.c          |  88 ++++++++-----
  drivers/iommu/tegra-smmu.c          |  59 +++++----
  include/linux/iommu.h               |  33 ++++-
  19 files changed, 644 insertions(+), 484 deletions(-)

Please pull.

Thanks,

Joerg

#2 lore raw

Hello Linus,

I've been trying for almost 3 weeks to get the patch at http://lists.linuxfoundation.org/pipermail/iommu/2015-April/012757.html into the IOMMU tree.

However, Joerg is not answering my emails for some reason (vacation/sick?)

This is a small patch to amd_iommu_v2.c, but it fixes a *big* bug, which causes a kernel thread to be stuck and resources to not be freed.

As amdkfd (AMD HSA kernel driver) is the only client for amd_iommu_v2 driver, it effects only HSA systems. I would like to get this into 4.1 and I also marked it as stable for 4.0.

Could you please take a look, ack it and pull it to -rc3/4 ? I can send you the patch again with git or I can also send it through Dave's drm pull request if that's more convenient for you.

Thanks,

Oded

Quoted 128 lines · Oded Gabbay

Hi Joerg,

I sent to you (cc iommu mailing list) an important fix on Apr-16 for a bug in amd_iommu_v2.c (from kernel 4.0). See:
http://lists.linuxfoundation.org/pipermail/iommu/2015-April/012757.html

Could you please add that to the pull request ? Its creating problems when running HSA applications.

Thanks,

Oded

Quoted 114 lines · Joerg Roedel

Hi Linus,

The following changes since commit
e42391cd048809d903291d07f86ed3934ce138e9:

   Linux 4.0-rc6 (2015-03-29 15:26:31 -0700)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu.git
tags/iommu-updates-v4.1

for you to fetch changes up to 7f65ef01e131650d455875598099cd06fea6096b:

   Merge branches 'iommu/fixes', 'x86/vt-d', 'x86/amd', 'arm/smmu',

'arm/tegra' and 'core' into next (2015-04-02 13:33:19 +0200)

----------------------------------------------------------------

IOMMU Updates for Linux v4.1

Not much this time, but the changes include:

* Moving domain allocation into the iommu drivers to prepare for the introduction of default domains for devices

* Fixing the IO page-table code in the AMD IOMMU driver to correctly encode large page sizes

* Extension of the PCI support in the ARM-SMMU driver

* Various fixes and cleanups

----------------------------------------------------------------
Dan Carpenter (1):
iommu/amd: Small cleanup in mn_release()

Joerg Roedel (27):
iommu: Introduce domain_alloc and domain_free iommu_ops
iommu: Introduce iommu domain types
iommu: Only allow iommu_map/unmap for paging domains iommu/amd: Make use of domain_alloc and domain_free iommu/vt-d: Make use of domain_alloc and domain_free iommu/omap: Make use of domain_alloc and domain_free iommu/arm-smmu: Make use of domain_alloc and domain_free iommu/exynos: Make use of domain_alloc and domain_free iommu/tegra-smmu: Make use of domain_alloc and domain_free iommu/tegra-gart: Make use of domain_alloc and domain_free iommu/msm: Make use of domain_alloc and domain_free iommu/shmobile: Make use of domain_alloc and domain_free iommu/ipmmu-vmsa: Make use of domain_alloc and domain_free iommu/rockchip: Make use of domain_alloc and domain_free iommu/fsl: Make use of domain_alloc and domain_free
iommu: Remove domain_init and domain_free iommu_ops Merge branch 'for-joerg/arm-smmu/updates' of
git://git.kernel.org/.../will/linux into arm/smmu
iommu/amd: Use BUS_NOTIFY_REMOVED_DEVICE
iommu/amd: Ignore BUS_NOTIFY_UNBOUND_DRIVER event iommu/amd: Don't allocate with __GFP_ZERO in alloc_coherent iommu/amd: Add support for contiguous dma allocator iommu/amd: Return the pte page-size in fetch_pte iommu/amd: Optimize iommu_unmap_page for new fetch_pte interface iommu/amd: Optimize alloc_new_range for new fetch_pte interface iommu/amd: Optimize amd_iommu_iova_to_phys for new fetch_pte
interface
iommu/amd: Correctly encode huge pages in iommu page tables Merge branches 'iommu/fixes', 'x86/vt-d', 'x86/amd',
'arm/smmu', 'arm/tegra' and 'core' into next

Quentin Lambert (1):
iommu/vt-d: Convert non-returned local variable to boolean when
relevant

Robin Murphy (1):
iommu/arm-smmu: set a more appropriate DMA mask

Thierry Reding (3):
iommu/tegra: Setup aperture
iommu/tegra: gart: Set aperture at domain initialization time iommu/tegra: smmu: Compute PFN mask at runtime

Will Deacon (3):
iommu/arm-smmu: ensure CBA2R is initialised before CBAR on SMMUv2 iommu/arm-smmu: handle multi-alias IOMMU groups for PCI devices iommu/io-pgtable-arm: avoid speculative walks through TTBR1

  drivers/iommu/amd_iommu.c           | 250
++++++++++++++++++------------------
  drivers/iommu/amd_iommu_types.h     |  13 +-
  drivers/iommu/amd_iommu_v2.c        |   2 +-
  drivers/iommu/arm-smmu.c            | 180 ++++++++++++++++----------
  drivers/iommu/exynos-iommu.c        |  87 +++++++------
  drivers/iommu/fsl_pamu_domain.c     |  60 +++++----
  drivers/iommu/fsl_pamu_domain.h     |   2 +-
  drivers/iommu/intel-iommu.c         |  68 +++++-----
  drivers/iommu/intel_irq_remapping.c |  12 +-
  drivers/iommu/io-pgtable-arm.c      |   5 +
  drivers/iommu/iommu.c               |  26 ++--
  drivers/iommu/ipmmu-vmsa.c          |  42 +++---
  drivers/iommu/msm_iommu.c           |  73 ++++++-----
  drivers/iommu/omap-iommu.c          |  49 ++++---
  drivers/iommu/rockchip-iommu.c      |  40 +++---
  drivers/iommu/shmobile-iommu.c      |  39 +++---
  drivers/iommu/tegra-gart.c          |  88 ++++++++-----
  drivers/iommu/tegra-smmu.c          |  59 +++++----
  include/linux/iommu.h               |  33 ++++-
  19 files changed, 644 insertions(+), 484 deletions(-)

Please pull.

Thanks,

Joerg

#3 lore raw

Hi Oded,

Quoted 1 line · Oded Gabbay

However, Joerg is not answering my emails for some reason (vacation/sick?)

Sorry, I was travelling for the last 3 weeks and had bad internet access. But I am back and will take care of this on Monday.

Joerg

#4 lore raw
Quoted 11 lines · Joerg Roedel

Hi Oded,

Quoted 1 line · Oded Gabbay

However, Joerg is not answering my emails for some reason (vacation/sick?)

Sorry, I was travelling for the last 3 weeks and had bad internet access. But I am back and will take care of this on Monday.

Joerg

Glad to hear it was vacation and not my other guess :)
Usually I'd be more patient, but as Fedora 22 already uses kernel 4.0, I wanted to get this fix asap into stable.

Oded

#5 lore raw
Quoted 16 lines · Oded Gabbay
Quoted 13 lines · Joerg Roedel

Hi Oded,

Quoted 3 lines · Oded Gabbay

However, Joerg is not answering my emails for some reason (vacation/sick?)

Sorry, I was travelling for the last 3 weeks and had bad internet access. But I am back and will take care of this on Monday.

Quoted 5 lines
Quoted 2 lines

Joerg

Glad to hear it was vacation and not my other guess :) Usually I'd be more patient, but as Fedora 22 already uses kernel 4.0, I wanted to get this fix asap into stable.

Is Fedora a particular concern here, or did you just happen to notice we're shipping with 4.0? I'm curious if there is a Fedora bug that was reported for the issue you're concerned about that we may have missed.

josh

#6 lore raw
Quoted 28 lines · Josh Boyer
Quoted 16 lines · Oded Gabbay
Quoted 13 lines · Joerg Roedel

Hi Oded,

Quoted 3 lines · Oded Gabbay

However, Joerg is not answering my emails for some reason (vacation/sick?)

Sorry, I was travelling for the last 3 weeks and had bad internet access. But I am back and will take care of this on Monday.

Quoted 5 lines
Quoted 2 lines

Joerg

Glad to hear it was vacation and not my other guess :) Usually I'd be more patient, but as Fedora 22 already uses kernel 4.0, I wanted to get this fix asap into stable.

Is Fedora a particular concern here, or did you just happen to notice we're shipping with 4.0? I'm curious if there is a Fedora bug that was reported for the issue you're concerned about that we may have missed.

Quoted 5 lines

josh
_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu

AFAIK, this is not a reported bug.

When I encountered the bug, I knew F22 uses 4.0, so I'm assuming that any user which installs F22 on Kaveri system AND executes an HSA application, will encounter this bug.

Signature · 1 line
          Oded