Partially revert "hw/block/nvme: drain namespaces on sq deletion"
commite76fb260ca8fc2420a4ce792324af0544628b331
authorKlaus Jensen <k.jensen@samsung.com>
Thu, 17 Jun 2021 19:06:57 +0000 (17 21:06 +0200)
committerKlaus Jensen <k.jensen@samsung.com>
Tue, 29 Jun 2021 05:16:25 +0000 (29 07:16 +0200)
tree5d87526c81bfdfa11def29f4cf2e69b689b3e0ed
parent3bcf26d3d6192e350b165c681aff13f6697f30fb
Partially revert "hw/block/nvme: drain namespaces on sq deletion"

This partially reverts commit 98f84f5a4eca5c03e32fff20f246d9b4b96d6422.

Since all "multi aio" commands are now reimplemented to properly track
the nested aiocbs, we can revert the "hack" that was introduced to make
sure all requests we're properly drained upon sq deletion.

The revert is partial since we keep the assert that no outstanding
requests remain on the submission queue after the explicit cancellation.

Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
hw/nvme/ctrl.c