Message-Id: <20190213175454.7506-1-logang@deltatee.com>
Date:   Wed, 13 Feb 2019 10:54:42 -0700
From:   Logan Gunthorpe <logang@...tatee.com>
To:     linux-kernel@...r.kernel.org, linux-ntb@...glegroups.com,
        linux-pci@...r.kernel.org, iommu@...ts.linux-foundation.org,
        linux-kselftest@...r.kernel.org, Jon Mason <jdmason@...zu.us>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Joerg Roedel <joro@...tes.org>
Cc:     Allen Hubbe <allenbh@...il.com>, Dave Jiang <dave.jiang@...el.com>,
        Serge Semin <fancer.lancer@...il.com>,
        Eric Pilmore <epilmore@...aio.com>,
        Logan Gunthorpe <logang@...tatee.com>
Subject: [PATCH v2 00/12] Support using MSI interrupts in ntb_transport

Note: this version will likely trivially conflict with some cleanup
patches I sent to Bjorn. So this is meant for review purposes only.
If there are no objections, I'd like to look at getting it merged in
the next cycle through the NTB tree.

--

Changes in v2:

* Cleaned up the changes in intel_irq_remapping.c to make them
  less confusing and add a comment. (Per discussion with Jacob and
  Joerg)

* Fixed a nit from Bjorn and collected his Ack

* Added a Kconfig dependancy on CONFIG_PCI_MSI for CONFIG_NTB_MSI
  as the Kbuild robot hit a random config that didn't build
  without it.

* Worked in a callback for when the MSI descriptor changes so that
  the clients can resend the new address and data values to the peer.
  On my test system this was never necessary, but there may be
  other platforms where this can occur. I tested this by hacking
  in a path to rewrite the MSI descriptor when I change the cpu
  affinity of an IRQ. There's a bit of uncertainty over the latency
  of the change, but without hardware this can acctually occur on
  we can't test this. This was the result of a discussion with Dave.

--

This patch series adds optional support for using MSI interrupts instead
of NTB doorbells in ntb_transport. This is desirable seeing doorbells on
current hardware are quite slow and therefore switching to MSI interrupts
provides a significant performance gain. On switchtec hardware, a simple
apples-to-apples comparison shows ntb_netdev/iperf numbers going from
3.88Gb/s to 14.1Gb/s when switching to MSI interrupts.

To do this, a couple changes are required outside of the NTB tree:

1) The IOMMU must know to accept MSI requests from aliased bused numbers
seeing NTB hardware typically sends proxied request IDs through
additional requester IDs. The first patch in this series adds support
for the Intel IOMMU. A quirk to add these aliases for switchtec hardware
was already accepted. See commit ad281ecf1c7d ("PCI: Add DMA alias quirk
for Microsemi Switchtec NTB") for a description of NTB proxy IDs and why
this is necessary.

2) NTB transport (and other clients) may often need more MSI interrupts
than the NTB hardware actually advertises support for. However, seeing
these interrupts will not be triggered by the hardware but through an
NTB memory window, the hardware does not actually need support or need
to know about them. Therefore we add the concept of Virtual MSI
interrupts which are allocated just like any other MSI interrupt but
are not programmed into the hardware's MSI table. This is done in
Patch 2 and then made use of in Patch 3.

The remaining patches in this series add a library for dealing with MSI
interrupts, a test client and finally support in ntb_transport.

The series is based off of v5.0-rc4 and I've tested it on top of a
of the patches I've already sent to the NTB tree (though they are
independent changes). A git repo is available here:

https://github.com/sbates130272/linux-p2pmem/ ntb_transport_msi_v2

Thanks,

Logan

--

Logan Gunthorpe (12):
  iommu/vt-d: Implement dma_[un]map_resource()
  NTB: ntb_transport: Ensure the destination buffer is mapped for TX DMA
  iommu/vt-d: Add helper to set an IRTE to verify only the bus number
  iommu/vt-d: Allow interrupts from the entire bus for aliased devices
  PCI/MSI: Support allocating virtual MSI interrupts
  PCI/switchtec: Add module parameter to request more interrupts
  NTB: Introduce functions to calculate multi-port resource index
  NTB: Rename ntb.c to support multiple source files in the module
  NTB: Introduce MSI library
  NTB: Introduce NTB MSI Test Client
  NTB: Add ntb_msi_test support to ntb_test
  NTB: Add MSI interrupt support to ntb_transport

 drivers/iommu/intel-iommu.c             |  23 +-
 drivers/iommu/intel_irq_remapping.c     |  32 +-
 drivers/ntb/Kconfig                     |  11 +
 drivers/ntb/Makefile                    |   3 +
 drivers/ntb/{ntb.c => core.c}           |   0
 drivers/ntb/msi.c                       | 415 +++++++++++++++++++++++
 drivers/ntb/ntb_transport.c             | 197 ++++++++++-
 drivers/ntb/test/Kconfig                |   9 +
 drivers/ntb/test/Makefile               |   1 +
 drivers/ntb/test/ntb_msi_test.c         | 433 ++++++++++++++++++++++++
 drivers/pci/msi.c                       |  55 ++-
 drivers/pci/switch/switchtec.c          |  12 +-
 include/linux/msi.h                     |   8 +
 include/linux/ntb.h                     | 143 ++++++++
 include/linux/pci.h                     |   9 +
 tools/testing/selftests/ntb/ntb_test.sh |  54 ++-
 16 files changed, 1379 insertions(+), 26 deletions(-)
 rename drivers/ntb/{ntb.c => core.c} (100%)
 create mode 100644 drivers/ntb/msi.c
 create mode 100644 drivers/ntb/test/ntb_msi_test.c

--
2.19.0

Powered by blists - more mailing lists