[PATCH v5 0/2] hw/block/nvme: dulbe and dsm support

From: Klaus Jensen <its@irrelevant.dk>
To: qemu-devel@nongnu.org
Cc: Kevin Wolf <kwolf@redhat.com>,
	qemu-block@nongnu.org, Klaus Jensen <k.jensen@samsung.com>,
	Max Reitz <mreitz@redhat.com>, Keith Busch <kbusch@kernel.org>,
	Klaus Jensen <its@irrelevant.dk>
Subject: [PATCH v5 0/2] hw/block/nvme: dulbe and dsm support
Date: Thu, 22 Oct 2020 20:49:57 +0200	[thread overview]
Message-ID: <20201022184959.240505-1-its@irrelevant.dk> (raw)

From: Klaus Jensen <k.jensen@samsung.com>

This adds support for the Deallocated or Unwritten Logical Block error
recovery feature as well as the Dataset Management command.

I wanted to add support for the NPDG and NPDA fields such that the host
could get a hint on how many blocks to request deallocation of for the
deallocation to actually happen, but I cannot find a realiable way to
get the actual block size of the underlying device. If it is an image on
a file system we could typically use the host page size, but if it is a
raw device, we might have 512 byte sectors that we can issue discards
on. And QEMU doesn't seem to provide this without root privileges at
least.

See the two patches for some gotchas.

I also integrated this into my zoned proposal. I'll spare you the v4, nobody
cares anyway. But I put it in my repo[1] for posterity.

  [1]: https://irrelevant.dk/g/pci-nvme.git/tag/?h=zoned-v4.

v5:
  - Restore status code from callback (Keith)

v4:
  - Removed mixed declaration and code (Keith)
  - Set NPDG and NPDA and account for the blockdev cluster size.

Klaus Jensen (2):
  hw/block/nvme: add dulbe support
  hw/block/nvme: add the dataset management command

 hw/block/nvme-ns.h    |   4 +
 hw/block/nvme.h       |   2 +
 include/block/nvme.h  |  12 ++-
 hw/block/nvme-ns.c    |  40 +++++++--
 hw/block/nvme.c       | 184 +++++++++++++++++++++++++++++++++++++++++-
 hw/block/trace-events |   4 +
 6 files changed, 237 insertions(+), 9 deletions(-)

-- 
2.28.0