Skip to contentSkip to conversation

VT-d fixes and ACPI device mapping support for 3.15

GIT PULL David Woodhouse 2 participants 3 messages Mar 24, 2014 – Apr 1, 2014 1 message not in the archive

lore
Share & tools
lore
#0 lore raw

Jörg, please pull into your x86/vt-d branch from

	git://git.infradead.org/iommu-2.6

This contains the two patch sets I've posted in the last week; first the set which fixes the use-after-free issues with IOTLB flushes, adds some missing IOTLB flushes and generally improves unmap performance. And then the support for DMA mapping of ACPI devices.

This contains a couple of fixes triggered by Fengguang's build-test robots; one of which was posted as [PATCH 34/33] a few hours ago, and the other of which just moves the 'segment' member of struct intel_iommu outside the #ifdef CONFIG_INTEL_IOMMU since it's accessed from dmar.c without that config option set. I've merged both of those fixes back into the original offending patch, rather than putting them on top.

 drivers/iommu/dmar.c        | 166 ++++++++---
 drivers/iommu/intel-iommu.c | 979 ++++++++++++++++++++++++++++++++++++++++-------------------------
 include/acpi/actbl2.h       |  15 +-
 include/linux/dmar.h        |  18 +-
 include/linux/intel-iommu.h |   1 +
 5 files changed, 759 insertions(+), 420 deletions(-)

David Woodhouse (37):
iommu/vt-d: Clean up size handling for intel_iommu_unmap() iommu/vt-d: Clean up and fix page table clear/free behaviour iommu/vt-d: Honour intel_iommu=sp_off for non-VMM domains
iommu/vt-d: Be less pessimistic about domain coherency where possible iommu/vt-d: Add ACPI namespace device reporting structures iommu/vt-d: Parse ANDD records
iommu/vt-d: Allocate space for ACPI devices
iommu/vt-d: Change scope lists to struct device, bus, devfn iommu/vt-d: Add ACPI devices into dmaru->devices[] array
iommu/vt-d: Make iommu_dummy() take struct device instead of struct pci_dev iommu/vt-d: Make dmar_insert_dev_info() take struct device instead of struct pci_dev iommu/vt-d: Use struct device in device_domain_info, not struct pci_dev iommu/vt-d: Pass iommu to domain_context_mapping_one() and iommu_support_dev_iotlb() iommu/vt-d: Stop dmar_insert_dev_info() freeing domains on losing race iommu/vt-d: use dmar_insert_dev_info() from dma_add_dev_info() iommu/vt-d: Use domain_remove_one_dev_info() in domain_add_dev_info() error path iommu/vt-d: Always store iommu in device_domain_info
iommu/vt-d: Simplify iommu check in domain_remove_one_dev_info() iommu/vt-d: Remove device_to_iommu() call from domain_remove_dev_info() iommu/vt-d: Store PCI segment number in struct intel_iommu iommu/vt-d: Remove segment from struct device_domain_info() iommu/vt-d: Make identity_mapping() take struct device not struct pci_dev iommu/vt-d: Make device_to_iommu() cope with non-PCI devices iommu/vt-d: Make domain_context_mapp{ed,ing}() take struct device iommu/vt-d: Make get_domain_for_dev() take struct device
iommu/vt-d: Handle RMRRs for non-PCI devices
iommu/vt-d: Make iommu_should_identity_map() take struct device iommu/vt-d: Make get_valid_domain_for_dev() take struct device iommu/vt-d: Remove some pointless to_pci_dev() calls
iommu/vt-d: Rename 'hwdev' variables to 'dev' now that that's the norm iommu/vt-d: Make domain_remove_one_dev_info() take struct device iommu/vt-d: Make domain_add_dev_info() take struct device
iommu/vt-d: Remove pdev from iommu_no_mapping()
iommu/vt-d: Remove pdev from intel_iommu_attach_device()
iommu/vt-d: Remove to_pci_dev() in intel_map_page()
iommu/vt-d: Finally enable translation for non-PCI devices iommu/vt-d: Include ACPI devices in iommu=pt

Signature · 1 line
dwmw2

A message that is not in the archive (sent off-list or not yet received): 1395674279.4880.236.camel-W2I5cNIroUsVm/[email protected]

#2 lore raw
Quoted 16 lines · David Woodhouse

Jörg, please pull into your x86/vt-d branch from

	git://git.infradead.org/iommu-2.6

This contains the two patch sets I've posted in the last week; first the set which fixes the use-after-free issues with IOTLB flushes, adds some missing IOTLB flushes and generally improves unmap performance. And then the support for DMA mapping of ACPI devices.

This contains a couple of fixes triggered by Fengguang's build-test robots; one of which was posted as [PATCH 34/33] a few hours ago, and the other of which just moves the 'segment' member of struct intel_iommu outside the #ifdef CONFIG_INTEL_IOMMU since it's accessed from dmar.c without that config option set. I've merged both of those fixes back into the original offending patch, rather than putting them on top.

Quoted 6 lines
 drivers/iommu/dmar.c        | 166 ++++++++---
 drivers/iommu/intel-iommu.c | 979 ++++++++++++++++++++++++++++++++++++++++-------------------------
 include/acpi/actbl2.h       |  15 +-
 include/linux/dmar.h        |  18 +-
 include/linux/intel-iommu.h |   1 +
 5 files changed, 759 insertions(+), 420 deletions(-)

Pulled, thanks David.

#3 lore raw
In reply to a message not in the archive
Quoted 3 lines · David Woodhouse

Jörg, please pull into your x86/vt-d branch from

        git://git.infradead.org/iommu-2.6

Again, please. A few minor fixes...

Dan Carpenter (1):
iommu/vt-d: returning free pointer in get_domain_for_dev()

David Woodhouse (2):
iommu/vt-d: Only call dmar_acpi_dev_scope_init() if DRHD units present iommu/vt-d: Fix error handling in ANDD processing

 drivers/iommu/dmar.c        | 12 +++++-------
 drivers/iommu/intel-iommu.c |  1 +
 2 files changed, 6 insertions(+), 7 deletions(-)
Signature · 2 lines
David Woodhouse                            Open Source Technology Centre
[email protected]                              Intel Corporation