mbox series

[v2,0/7] dmaengine: idxd: Add interrupt handle revoke support

Message ID 163528412413.3925689.7831987824972063153.stgit@djiang5-desk3.ch.intel.com (mailing list archive)
Headers show
Series dmaengine: idxd: Add interrupt handle revoke support | expand

Message

Dave Jiang Oct. 26, 2021, 9:35 p.m. UTC
v2:
- Fix comment header typo (Vinod)
- Declare and init struct w/o memset (Vinod)

The series adds support to refresh the interrupt handles when they become
invalid. Typically this happens during a VM live migration where a VM moves
from one machine to another. The driver will receive an interrupt to
indicate that interrupt handles need to be changed. The driver blocks the
current submissions and acquires new interrupt handles. All submissions
will be held off until the handle is refreshed. Already submitted
descriptor
will error with status of "incorrect interrupt handle" and be resubmitted
by the
driver.

---

Dave Jiang (7):
      dmaengine: idxd: rework descriptor free path on failure
      dmaengine: idxd: int handle management refactoring
      dmaengine: idxd: move interrupt handle assignment
      dmaengine: idxd: add helper for per interrupt handle drain
      dmaengine: idxd: create locked version of idxd_quiesce() call
      dmaengine: idxd: handle invalid interrupt handle descriptors
      dmaengine: idxd: handle interrupt handle revoked event


 drivers/dma/idxd/device.c    |  24 +++-
 drivers/dma/idxd/dma.c       |  18 ++-
 drivers/dma/idxd/idxd.h      |  13 +-
 drivers/dma/idxd/init.c      |  87 +++++++-------
 drivers/dma/idxd/irq.c       | 226 +++++++++++++++++++++++++++++++++++
 drivers/dma/idxd/registers.h |   1 +
 drivers/dma/idxd/submit.c    |  26 ++--
 drivers/dma/idxd/sysfs.c     |   1 -
 8 files changed, 331 insertions(+), 65 deletions(-)

--

Comments

Vinod Koul Nov. 22, 2021, 5:51 a.m. UTC | #1
On 26-10-21, 14:35, Dave Jiang wrote:
> v2:
> - Fix comment header typo (Vinod)
> - Declare and init struct w/o memset (Vinod)
> 
> The series adds support to refresh the interrupt handles when they become
> invalid. Typically this happens during a VM live migration where a VM moves
> from one machine to another. The driver will receive an interrupt to
> indicate that interrupt handles need to be changed. The driver blocks the
> current submissions and acquires new interrupt handles. All submissions
> will be held off until the handle is refreshed. Already submitted
> descriptor
> will error with status of "incorrect interrupt handle" and be resubmitted
> by the
> driver.

Applied, thanks