kernel.git/drivers/hwtracing/intel_th/core.c, branch linux-4.20.y

intel_th: Fix resource handling for ACPI glue layer

2018-09-18T14:08:38Z

The core of the driver expects the resource array from the glue layer to be indexed by even numbers, as is the case for 64-bit PCI resources. This doesn't hold true for others, ACPI in this instance, which leads to an out-of-bounds access and an ioremap() on whatever address that access fetches. This patch fixes the problem by reading resource array differently based on whether the 64-bit flag is set, which would indicate PCI glue layer. Signed-off-by: Alexander Shishkin Fixes: ebc57e399b8e ("intel_th: Add ACPI glue layer") CC: stable@vger.kernel.org # v4.17+ Signed-off-by: Greg Kroah-Hartman

intel_th: Fix device removal logic

2018-09-18T14:08:38Z

Commit a753bfcfdb1f ("intel_th: Make the switch allocate its subdevices") brings in new subdevice addition/removal logic that's broken for "host mode": the SWITCH device has no children to begin with, which is not handled in the code. This results in a null dereference bug later down the path. This patch fixes the subdevice removal code to handle host mode correctly. Signed-off-by: Alexander Shishkin Fixes: a753bfcfdb1f ("intel_th: Make the switch allocate its subdevices") CC: stable@vger.kernel.org # v4.14+ Signed-off-by: Greg Kroah-Hartman

intel_th: Allow forcing host mode through drvdata

2018-03-28T15:47:21Z

Some devices can only operate in host mode, so we need means of communicating this to the core driver on per-device basis. This adds a flag to drvdata to signal host-only capability to the core. Signed-off-by: Alexander Shishkin

intel_th: Pick up irq number from resources

2018-03-28T15:47:21Z

Platform devices pass their IRQs around as resources, so as a convenience for the glue layer code, allow them pass the IRQ to the core driver in the resources array. Signed-off-by: Alexander Shishkin

intel_th: Don't touch switch routing in host mode

2018-03-28T15:47:20Z

When the Trace Hub is operating in Host Debugger mode, it is up to the debugger to configure master routing even for the software sources. Do not do this in the driver in this case. Signed-off-by: Alexander Shishkin

intel_th: Use correct method of finding hub

2018-03-28T15:47:19Z

Since commit 8edc514b01e9 ("intel_th: Make SOURCE devices children of the root device") the hub is not the parent of SOURCE devices any more, so the new helper function should be used for that instead of always using the parent. The intel_th_set_output() path, however, still uses the old logic, leading to the hub driver structure being aliased with something else, like struct pci_driver or struct acpi_driver, and an incorrect call to an address inferred from that, potentially resulting in a crash. Fixes: 8edc514b01e9 ("intel_th: Make SOURCE devices children of the root device") Signed-off-by: Alexander Shishkin

intel_th: Add SPDX GPL-2.0 header to replace GPLv2 boilerplate

2018-03-28T15:47:19Z

This adds SPDX GPL-2.0 header to the Trace Hub driver and removes the GPLv2 boilerplate text. Signed-off-by: Alexander Shishkin

intel_th: Perform time resync on capture start

2017-08-25T15:48:00Z

On some devices (TH 2.x devices at the moment), the internal time counter is initially not synchronized to the global crystal clock, so the time stamps it produces will not be useful. In this case, the driver needs to force the time counter resync. This applies the workaround to relevant devices. Signed-off-by: Alexander Shishkin

intel_th: Add global activate/deactivate callbacks for the glue layers

2017-08-25T15:48:00Z

A glue layer may want to install its own hooks into trace capture start and stop paths to apply workarounds. This adds optional callbacks. Signed-off-by: Alexander Shishkin

intel_th: pci: Use drvdata for quirks

2017-08-25T15:47:59Z

Allow attaching miscellaneous quirk information to devices as drvdata. Signed-off-by: Alexander Shishkin