Otherwise duplicated slot can be used from different blk-mq hw queue.
The worsen thing is that V3's actual max queue depth is (4096 - 96), but
this patch claims that the device can support (4096 - 96) * 32 command
slots
To be clear about the hw, the hw supports max 4096 command tags and has 16
Is 4096 the max allowed host-wide command tags? Or per-queue's max commands
tags?
4096 is max allowed host wide, in range [0, 4096), and tags are not per
queue. HW delivery queues can be configured to be as large as desired.
And on another point I saw mentioned, the device supports multiple
submission and multiple completion queues. They are symmetrical, and any
command will always complete on the same queue index which it was submitted.
If it is per-queue's max command tags, V3 should be real MQ hardware,
otherwise it is still host wide. Otherwise, Hannes's patch can't work
because tag from different hw queue can be same.
hw queues. The hw queue depth is configurable by software, and we configure
it at 4096 per queue - same as max command tags - this is to support
possibility of all commands using the same queue simultaneously.
Suppose 4096 is the host-wide command tags:
As Hannes's patch changed to allow each blk-mq hw queue to accept 4096 commands,
there will be big contention in driver, given now the actual .can_queue becomes
4096 * 32, and how to avoid the same tag from different hw queue?
The documentation needs to be clarifed for this, to avoid future confusion.
Thanks,
Ming
.
