linux-stable.git/drivers/ntb, branch v5.7.7

NTB: perf: Fix race condition when run with ntb_test

2020-06-24T15:49:07+00:00

[ Upstream commit 34d8673a01b053b6231a995a4eec9341163d63be ]

When running ntb_test, the script tries to run the ntb_perf test
immediately after probing the modules. Since adding multi-port support,
this fails seeing the new initialization procedure in ntb_perf
can not complete instantly.

To fix this we add a completion which is waited on when a test is
started. In this way, run can be written any time after the module is
loaded and it will wait for the initialization to complete instead of
sending an error.

Fixes: 5648e56d03fa ("NTB: ntb_perf: Add full multi-port NTB API support")
Signed-off-by: Logan Gunthorpe 
Acked-by: Allen Hubbe 
Tested-by: Alexander Fomichev 
Signed-off-by: Jon Mason 
Signed-off-by: Sasha Levin

NTB: perf: Fix support for hardware that doesn't have port numbers

2020-06-24T15:49:06+00:00

[ Upstream commit b54369a248c2e033bfcf5d6917e08cf9d73d54a6 ]

Legacy drivers do not have port numbers (but is reliably only two ports)
and was broken by the recent commit that added mult-port support to
ntb_perf. This is especially important to support the cross link
topology which is perfectly symmetric and cannot assign unique port
numbers easily.

Hardware that returns zero for both the local port and the peer should
just always use gidx=0 for the only peer.

Fixes: 5648e56d03fa ("NTB: ntb_perf: Add full multi-port NTB API support")
Signed-off-by: Logan Gunthorpe 
Acked-by: Allen Hubbe 
Tested-by: Alexander Fomichev 
Signed-off-by: Jon Mason 
Signed-off-by: Sasha Levin

NTB: perf: Don't require one more memory window than number of peers

2020-06-24T15:49:06+00:00

[ Upstream commit a9c4211ac918ade1522aced6b5acfbe824722f7d ]

ntb_perf should not require more than one memory window per peer. This
was probably an off-by-one error.

Fixes: 5648e56d03fa ("NTB: ntb_perf: Add full multi-port NTB API support")
Signed-off-by: Logan Gunthorpe 
Acked-by: Allen Hubbe 
Tested-by: Alexander Fomichev 
Signed-off-by: Jon Mason 
Signed-off-by: Sasha Levin

NTB: Revert the change to use the NTB device dev for DMA allocations

2020-06-24T15:49:06+00:00

[ Upstream commit 40da7d9a93c8941737ef4a1208d32c13ce017fe1 ]

Commit 417cf39cfea9 ("NTB: Set dma mask and dma coherent mask to NTB
devices") started using the NTB device for DMA allocations which was
turns out was wrong. If the IOMMU is enabled, such alloctanions will
always fail with messages such as:

  DMAR: Allocating domain for 0000:02:00.1 failed

This is because the IOMMU has not setup the device for such use.

Change the tools back to using the PCI device for allocations seeing
it doesn't make sense to add an IOMMU group for the non-physical NTB
device. Also remove the code that sets the DMA mask as it no longer
makes sense to do this.

Fixes: 7f46c8b3a552 ("NTB: ntb_tool: Add full multi-port NTB API support")
Signed-off-by: Logan Gunthorpe 
Tested-by: Alexander Fomichev 
Signed-off-by: Jon Mason 
Signed-off-by: Sasha Levin

NTB: ntb_tool: reading the link file should not end in a NULL byte

2020-06-24T15:49:06+00:00

[ Upstream commit 912e12813dd03c602e4922fc34709ec4d4380cf0 ]

When running ntb_test this warning is issued:

./ntb_test.sh: line 200: warning: command substitution: ignored null
byte in input

This is caused by the kernel returning one more byte than is necessary
when reading the link file.

Reduce the number of bytes read back to 2 as it was before the
commit that regressed this.

Fixes: 7f46c8b3a552 ("NTB: ntb_tool: Add full multi-port NTB API support")
Signed-off-by: Logan Gunthorpe 
Acked-by: Allen Hubbe 
Tested-by: Alexander Fomichev 
Signed-off-by: Jon Mason 
Signed-off-by: Sasha Levin

ntb_tool: pass correct struct device to dma_alloc_coherent

2020-06-24T15:49:06+00:00

[ Upstream commit 433efe720674efd9fdbcef78be75793393cf05db ]

Currently, ntb->dev is passed to dma_alloc_coherent
and dma_free_coherent calls. The returned dma_addr_t
is the CPU physical address. This works fine as long
as IOMMU is disabled. But when IOMMU is enabled, we
need to make sure that IOVA is returned for dma_addr_t.
So the correct way to achieve this is by changing the
first parameter of dma_alloc_coherent() as ntb->pdev->dev
instead.

Fixes: 5648e56d03fa ("NTB: ntb_perf: Add full multi-port NTB API support")
Signed-off-by: Sanjay R Mehta 
Signed-off-by: Arindam Nath 
Signed-off-by: Jon Mason 
Signed-off-by: Sasha Levin

ntb_perf: pass correct struct device to dma_alloc_coherent

2020-06-24T15:49:06+00:00

[ Upstream commit 98f4e140264eeb52f22ff05be6b6dd48237255ac ]

Currently, ntb->dev is passed to dma_alloc_coherent
and dma_free_coherent calls. The returned dma_addr_t
is the CPU physical address. This works fine as long
as IOMMU is disabled. But when IOMMU is enabled, we
need to make sure that IOVA is returned for dma_addr_t.
So the correct way to achieve this is by changing the
first parameter of dma_alloc_coherent() as ntb->pdev->dev
instead.

Fixes: 5648e56d03fa ("NTB: ntb_perf: Add full multi-port NTB API support")
Signed-off-by: Logan Gunthorpe 
Signed-off-by: Sanjay R Mehta 
Signed-off-by: Arindam Nath 
Signed-off-by: Jon Mason 
Signed-off-by: Sasha Levin

NTB: Fix the default port and peer numbers for legacy drivers

2020-06-24T15:48:30+00:00

[ Upstream commit fc8b086d9dbd57458d136c4fa70ee26f832c3a2e ]

When the commit adding ntb_default_port_number() and
ntb_default_peer_port_number()  entered the kernel there was no
users of it so it was impossible to tell what the API needed.

When a user finally landed a year later (ntb_pingpong) there were
more NTB topologies were created and no consideration was considered
to how other drivers had changed.

Now that there is a user it can be fixed to provide a sensible default
for the legacy drivers that do not implement ntb_{peer_}port_number().
Seeing ntb_pingpong doesn't check error codes returning EINVAL was also
not sensible.

Patches for ntb_pingpong and ntb_perf follow (which are broken
otherwise) to support hardware that doesn't have port numbers. This is
important not only to not break support with existing drivers but for
the cross link topology which, due to its perfect symmetry, cannot
assign unique port numbers to each side.

Fixes: 1e5301196a88 ("NTB: Add indexed ports NTB API")
Signed-off-by: Logan Gunthorpe 
Acked-by: Allen Hubbe 
Tested-by: Alexander Fomichev 
Signed-off-by: Jon Mason 
Signed-off-by: Sasha Levin

NTB: ntb_pingpong: Choose doorbells based on port number

2020-06-24T15:48:30+00:00

[ Upstream commit ca93c45755da98302c93abdd788fc09113baf9e0 ]

This commit fixes pingpong support for existing drivers that do not
implement ntb_default_port_number() and ntb_default_peer_port_number().
This is required for hardware (like the crosslink topology of
switchtec) which cannot assign reasonable port numbers to each port due
to its perfect symmetry.

Instead of picking the doorbell to use based on the the index of the
peer, we use the peer's port number. This is a bit clearer and easier
to understand.

Fixes: c7aeb0afdcc2 ("NTB: ntb_pp: Add full multi-port NTB API support")
Signed-off-by: Logan Gunthorpe 
Acked-by: Allen Hubbe 
Tested-by: Alexander Fomichev 
Signed-off-by: Jon Mason 
Signed-off-by: Sasha Levin

Merge tag 'pci-v5.7-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci

2020-04-03T21:25:02+00:00

Pull pci updates from Bjorn Helgaas:
 "Enumeration:

   - Revert sysfs "rescan" renames that broke apps (Kelsey Skunberg)

   - Add more 32 GT/s link speed decoding and improve the implementation
     (Yicong Yang)

  Resource management:

   - Add support for sizing programmable host bridge apertures and fix a
     related alpha Nautilus regression (Ivan Kokshaysky)

  Interrupts:

   - Add boot interrupt quirk mechanism for Xeon chipsets and document
     boot interrupts (Sean V Kelley)

  PCIe native device hotplug:

   - When possible, disable in-band presence detect and use PDS
     (Alexandru Gagniuc)

   - Add DMI table for devices that don't use in-band presence detection
     but don't advertise that correctly (Stuart Hayes)

   - Fix hang when powering slots up/down via sysfs (Lukas Wunner)

   - Fix an MSI interrupt race (Stuart Hayes)

  Virtualization:

   - Add ACS quirks for Zhaoxin devices (Raymond Pang)

  Error handling:

   - Add Error Disconnect Recover (EDR) support so firmware can report
     devices disconnected via DPC and we can try to recover (Kuppuswamy
     Sathyanarayanan)

  Peer-to-peer DMA:

   - Add Intel Sky Lake-E Root Ports B, C, D to the whitelist (Andrew
     Maier)

  ASPM:

   - Reduce severity of common clock config message (Chris Packham)

   - Clear the correct bits when enabling L1 substates, so we don't go
     to the wrong state (Yicong Yang)

  Endpoint framework:

   - Replace EPF linkup ops with notifier call chain and improve locking
     (Kishon Vijay Abraham I)

   - Fix concurrent memory allocation in OB address region (Kishon Vijay
     Abraham I)

   - Move PF function number assignment to EPC core to support multiple
     function creation methods (Kishon Vijay Abraham I)

   - Fix issue with clearing configfs "start" entry (Kunihiko Hayashi)

   - Fix issue with endpoint MSI-X ignoring BAR Indicator and Table
     Offset (Kishon Vijay Abraham I)

   - Add support for testing DMA transfers (Kishon Vijay Abraham I)

   - Add support for testing > 10 endpoint devices (Kishon Vijay Abraham I)

   - Add support for tests to clear IRQ (Kishon Vijay Abraham I)

   - Add common DT schema for endpoint controllers (Kishon Vijay Abraham I)

  Amlogic Meson PCIe controller driver:

   - Add DT bindings for AXG PCIe PHY, shared MIPI/PCIe analog PHY (Remi
     Pommarel)

   - Add Amlogic AXG PCIe PHY, AXG MIPI/PCIe analog PHY drivers (Remi
     Pommarel)

  Cadence PCIe controller driver:

   - Add Root Complex/Endpoint DT schema for Cadence PCIe (Kishon Vijay
     Abraham I)

  Intel VMD host bridge driver:

   - Add two VMD Device IDs that require bus restriction mode (Sushma
     Kalakota)

  Mobiveil PCIe controller driver:

   - Refactor and modularize mobiveil driver (Hou Zhiqiang)

   - Add support for Mobiveil GPEX Gen4 host (Hou Zhiqiang)

  Microsoft Hyper-V host bridge driver:

   - Add support for Hyper-V PCI protocol version 1.3 and
     PCI_BUS_RELATIONS2 (Long Li)

   - Refactor to prepare for virtual PCI on non-x86 architectures (Boqun
     Feng)

   - Fix memory leak in hv_pci_probe()'s error path (Dexuan Cui)

  NVIDIA Tegra PCIe controller driver:

   - Use pci_parse_request_of_pci_ranges() (Rob Herring)

   - Add support for endpoint mode and related DT updates (Vidya Sagar)

   - Reduce -EPROBE_DEFER error message log level (Thierry Reding)

  Qualcomm PCIe controller driver:

   - Restrict class fixup to specific Qualcomm devices (Bjorn Andersson)

  Synopsys DesignWare PCIe controller driver:

   - Refactor core initialization code for endpoint mode (Vidya Sagar)

   - Fix endpoint MSI-X to use correct table address (Kishon Vijay
     Abraham I)

  TI DRA7xx PCIe controller driver:

   - Fix MSI IRQ handling (Vignesh Raghavendra)

  TI Keystone PCIe controller driver:

   - Allow AM654 endpoint to raise MSI-X interrupt (Kishon Vijay Abraham I)

  Miscellaneous:

   - Quirk ASMedia XHCI USB to avoid "PME# from D0" defect (Kai-Heng
     Feng)

   - Use ioremap(), not phys_to_virt(), for platform ROM to fix video
     ROM mapping with CONFIG_HIGHMEM (Mikel Rychliski)"

* tag 'pci-v5.7-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (96 commits)
  misc: pci_endpoint_test: remove duplicate macro PCI_ENDPOINT_TEST_STATUS
  PCI: tegra: Print -EPROBE_DEFER error message at debug level
  misc: pci_endpoint_test: Use full pci-endpoint-test name in request_irq()
  misc: pci_endpoint_test: Fix to support > 10 pci-endpoint-test devices
  tools: PCI: Add 'e' to clear IRQ
  misc: pci_endpoint_test: Add ioctl to clear IRQ
  misc: pci_endpoint_test: Avoid using module parameter to determine irqtype
  PCI: keystone: Allow AM654 PCIe Endpoint to raise MSI-X interrupt
  PCI: dwc: Fix dw_pcie_ep_raise_msix_irq() to get correct MSI-X table address
  PCI: endpoint: Fix ->set_msix() to take BIR and offset as arguments
  misc: pci_endpoint_test: Add support to get DMA option from userspace
  tools: PCI: Add 'd' command line option to support DMA
  misc: pci_endpoint_test: Use streaming DMA APIs for buffer allocation
  PCI: endpoint: functions/pci-epf-test: Print throughput information
  PCI: endpoint: functions/pci-epf-test: Add DMA support to transfer data
  PCI: pciehp: Fix MSI interrupt race
  PCI: pciehp: Fix indefinite wait on sysfs requests
  PCI: endpoint: Fix clearing start entry in configfs
  PCI: tegra: Add support for PCIe endpoint mode in Tegra194
  PCI: sysfs: Revert "rescan" file renames
  ...