smartpqi: Change alignment for dma tags

[john.hall_microchip.com]

Summary

Problem: Under certain I/O conditions, a program doing large block disk reads can cause a controller to crash.

Root Cause: The SCSI read request and destination address in the BDMA descriptor is incorrect, causing the BDMA engine in the controller to assert.

Fix: Change the alignment for creating bus_dma_tags in the driver from PAGE_SIZE (4k) to 1, which allows the controller to manage it's own address range for BDMA transactions.

Risk: Medium

Exposure: This reverts a change first made to support NVMe drives on
Excalibur. At that time a 4k alignment was necessary. This no longer seems to be the case.

PR: 259541
Reported by: Ka Ho Ng <[e-mail]>

[editor_callfortesting.org]

I have testesd 14.0-BETA4 amd64 with and without this patch. dd'ing 100GB of /dev/random to a file on the SATA-booted drive works with both. This was definitely a problem with previous releases.

I will try this with ZFS, which appears to have been problematic with a previous 14.0-CURRENT snapshot.

[editor_callfortesting.org]

I have tested this with root-on-ZFS using a memstick installer build from my patched /usr/src

The system appears to be stable with both patched and unpatched 14.0-BETA4.

Could this have been resolved elsewhere?

[scott.benesh_microsemi.com]

Thanks for looking at the code. We are able to see sporadic problems when running "diskinfo -i /dev/daX" doing large asynchronous random reads (1024 kbytes) . The patch resolves the problem.