Chris Wilson 03e4e0a9e0 dma-buf/fence: Fix lock inversion within dma-fence-array ...

Ages ago Rob Clark noted,

"Currently with fence-array, we have a potential deadlock situation.  If
we fence_add_callback() on an array-fence, the array-fence's lock is
acquired first, and in it's ->enable_signaling() callback, it will install
cbs on it's array-member fences, so the array-member's lock is acquired
second.

But in the signal path, the array-member's lock is acquired first, and
the array-fence's lock acquired second."

Rob proposed either extensive changes to dma-fence to unnest the
fence-array signaling, or to defer the signaling onto a workqueue. This
is a more refined version of the later, that should keep the latency
of the fence signaling to a minimum by using an irq-work, which is
executed asap.

Reported-by: Rob Clark <robdclark@gmail.com>
Suggested-by: Rob Clark <robdclark@gmail.com>
References: 1476635975-21981-1-git-send-email-robdclark@gmail.com
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Rob Clark <robdclark@gmail.com>
Cc: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Sumit Semwal <sumit.semwal@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20171114162719.30958-1-chris@chris-wilson.co.uk

2017-11-21 22:37:05 +05:30

..

power

PM / QoS: Fix device resume latency PM QoS

2017-10-24 15:20:45 +02:00

regmap

regmap: regmap-w1: Fix build troubles

2017-07-17 15:53:00 +01:00

test

…

arch_topology.c

base: arch_topology: fix section mismatch build warnings

2017-09-18 17:05:07 +02:00

attribute_container.c

…

base.h

driver core: make device_{add|remove}_groups() public

2017-07-22 11:59:23 +02:00

bus.c

driver core: bus: Fix a potential double free

2017-08-31 18:57:30 +02:00

cacheinfo.c

…

class.c

…

component.c

…

container.c

…

core.c

Do not disable driver and bus shutdown hook when class shutdown hook is set.

2017-08-28 18:02:46 +02:00

cpu.c

PM / QoS: Fix device resume latency PM QoS

2017-10-24 15:20:45 +02:00

dd.c

initcall_debug: add deferred probe times

2017-08-03 17:48:49 -07:00

devcoredump.c

…

devres.c

…

devtmpfs.c

…

dma-coherent.c

dma-coherent: fix rmem_dma_device_init regression

2017-09-17 08:20:02 -07:00

dma-contiguous.c

…

dma-mapping.c

dma-coherent: remove the DMA_MEMORY_MAP and DMA_MEMORY_IO flags

2017-09-01 11:59:17 +02:00

driver.c

…

firmware_class.c

Revert "firmware: add sanity check on shutdown/suspend"

2017-09-10 21:19:06 -07:00

firmware.c

…

hypervisor.c

…

init.c

…

isa.c

…

Kconfig

dma-buf/fence: Fix lock inversion within dma-fence-array

2017-11-21 22:37:05 +05:30

Makefile

…

map.c

…

memory.c

mm, memory_hotplug: remove zone restrictions

2017-09-06 17:27:25 -07:00

module.c

…

node.c

mm: only display online cpus of the numa node

2017-10-13 16:18:32 -07:00

pinctrl.c

…

platform-msi.c

…

platform.c

driver core: platform: Don't read past the end of "driver_override" buffer

2017-09-18 17:16:47 +02:00

property.c

ACPI: properties: Align return codes of __acpi_node_get_property_reference()

2017-10-11 21:15:44 +02:00

soc.c

…

syscore.c

…

topology.c

base: topology: constify attribute_group structures.

2017-08-28 17:49:22 +02:00

transport_class.c

…