Skip to content

L
linux
Switch branch/tag
  1. 28 Jul, 2021 10 commits
  3. 27 Jul, 2021 1 commit
    • Vinod Koul's avatar
      Merge tag 'bus_remove_return_void-5.15' into next · c28d5d56
      Vinod Koul authored 
      This helps to remove conflict on idxd driver

 Conflicts:
	drivers/dma/idxd/sysfs.c
	drivers/dma/idxd/bus.c

Greg says:

Bus: Make remove callback return void tag

Tag for other trees/branches to pull from in order to have a stable
place to build off of if they want to add new busses for 5.15.
Signed-off-by: default avatarVinod Koul <vkoul@kernel.org>
      c28d5d56
  5. 21 Jul, 2021 25 commits
  7. 20 Jul, 2021 4 commits
    • Dave Jiang's avatar
      dmaengine: idxd: fix sequence for pci driver remove() and shutdown() · 49c4959f
      Dave Jiang authored 
      ->shutdown() call should only be responsible for quiescing the device.
Currently it is doing PCI device tear down. This causes issue when things
like MMIO mapping is removed while idxd_unregister_devices() will trigger
removal of idxd device sub-driver and still initiates MMIO writes to the
device. Another issue is with the unregistering of idxd 'struct device',
the memory context gets freed. So the teardown calls are accessing freed
memory and can cause kernel oops. Move all the teardown bits that doesn't
belong in shutdown to ->remove() call. Move unregistering of the idxd
conf_dev 'struct device' to after doing all the teardown to free all
the memory that's no longer needed.

Fixes: 47c16ac2 ("dmaengine: idxd: fix idxd conf_dev 'struct device' lifetime")
Signed-off-by: default avatarDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/162629983901.395844.17964803190905549615.stgit@djiang5-desk3.ch.intel.comSigned-off-by: default avatarVinod Koul <vkoul@kernel.org>
      49c4959f
    • Dave Jiang's avatar
      dmaengine: idxd: fix submission race window · 6b4b87f2
      Dave Jiang authored 
      Konstantin observed that when descriptors are submitted, the descriptor is
added to the pending list after the submission. This creates a race window
with the slight possibility that the descriptor can complete before it
gets added to the pending list and this window would cause the completion
handler to miss processing the descriptor.

To address the issue, the addition of the descriptor to the pending list
must be done before it gets submitted to the hardware. However, submitting
to swq with ENQCMDS instruction can cause a failure with the condition of
either wq is full or wq is not "active".

With the descriptor allocation being the gate to the wq capacity, it is not
possible to hit a retry with ENQCMDS submission to the swq. The only
possible failure can happen is when wq is no longer "active" due to hw
error and therefore we are moving towards taking down the portal. Given
this is a rare condition and there's no longer concern over I/O
performance, the driver can walk the completion lists in order to retrieve
and abort the descriptor.

The error path will set the descriptor to aborted status. It will take the
work list lock to prevent further processing of worklist. It will do a
delete_all on the pending llist to retrieve all descriptors on the pending
llist. The delete_all action does not require a lock. It will walk through
the acquired llist to find the aborted descriptor while add all remaining
descriptors to the work list since it holds the lock. If it does not find
the aborted descriptor on the llist, it will walk through the work
list. And if it still does not find the descriptor, then it means the
interrupt handler has removed the desc from the llist but is pending on
the work list lock and will process it once the error path releases the
lock.

Fixes: eb15e715 ("dmaengine: idxd: add interrupt handle request and release support")
Reported-by: default avatarKonstantin Ananyev <konstantin.ananyev@intel.com>
Signed-off-by: default avatarDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/162628855747.360485.10101925573082466530.stgit@djiang5-desk3.ch.intel.comSigned-off-by: default avatarVinod Koul <vkoul@kernel.org>
      6b4b87f2
    • Dave Jiang's avatar
      dmaengine: idxd: fix sequence for pci driver remove() and shutdown() · 7eb25da1
      Dave Jiang authored 
      ->shutdown() call should only be responsible for quiescing the device.
Currently it is doing PCI device tear down. This causes issue when things
like MMIO mapping is removed while idxd_unregister_devices() will trigger
removal of idxd device sub-driver and still initiates MMIO writes to the
device. Another issue is with the unregistering of idxd 'struct device',
the memory context gets freed. So the teardown calls are accessing freed
memory and can cause kernel oops. Move all the teardown bits that doesn't
belong in shutdown to ->remove() call. Move unregistering of the idxd
conf_dev 'struct device' to after doing all the teardown to free all
the memory that's no longer needed.

Fixes: 47c16ac2 ("dmaengine: idxd: fix idxd conf_dev 'struct device' lifetime")
Signed-off-by: default avatarDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/162629983901.395844.17964803190905549615.stgit@djiang5-desk3.ch.intel.comSigned-off-by: default avatarVinod Koul <vkoul@kernel.org>
      7eb25da1
    • Dave Jiang's avatar
      dmaengine: idxd: fix desc->vector that isn't being updated · 8ba89a3c
      Dave Jiang authored 
      Missing update for desc->vector when the wq vector gets updated. This
causes the desc->vector to always be at 0.

Fixes: da435aed ("dmaengine: idxd: fix array index when int_handles are being used")
Signed-off-by: default avatarDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/162628784374.353761.4736602409627820431.stgit@djiang5-desk3.ch.intel.comSigned-off-by: default avatarVinod Koul <vkoul@kernel.org>
      8ba89a3c
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7