[PATCH v20 01/12] block: Introduce queue limits and sysfs for copy-offload support

Nitesh Shetty nj.shetty at samsung.com
Thu May 23 00:05:12 PDT 2024


On 22/05/24 10:49AM, Bart Van Assche wrote:
>On 5/21/24 07:25, Nitesh Shetty wrote:
>>On 20/05/24 03:42PM, Bart Van Assche wrote:
>>>On 5/20/24 03:20, Nitesh Shetty wrote:
>>>>+    if (max_copy_bytes & (queue_logical_block_size(q) - 1))
>>>>+        return -EINVAL;
>>>
>>>Wouldn't it be more user-friendly if this check would be left out? Does any code
>>>depend on max_copy_bytes being a multiple of the logical block size?
>>>
>>In block layer, we use max_copy_bytes to split larger copy into
>>device supported copy size.
>>Simple copy spec requires length to be logical block size aligned.
>>Hence this check.
>
>Will blkdev_copy_sanity_check() reject invalid copy requests even if this
>check is left out?
>
Yes, you are right. We have checks both places.
We will remove checks in one of the places.

Thank you,
Nitesh Shetty


More information about the Linux-nvme mailing list