[RFC 0/3] Btrfs checksum offload

[Kanchan Joshi]

On 1/30/2025 6:23 PM, Christoph Hellwig wrote:
>> Difference is that FS does not have to attach any PI for offload.
>>
>> Offload is about the Host doing as little as possible, and the closest
>> we get there is by setting PRACT bit.
> But that doesn't actually work.  The file system needs to be able
> to verify the checksum for failing over to other mirrors, repair,
> etc.

Right. That sounds like reusing the existing code on detecting 
checksum-specific failure. So maybe that can be handled iff this gets 
any far.

> Also if you trust the device to get things right you do not
> need to use PI at all - SSDs or hard drives that support PI generally
> use PI internally anyway, and PRACT just means you treat a format
> with PI like one without.  In other words - no need for an offload
> here, you might as well just trust the device if you're not doing
> end to end protection.

Agree that device maybe implementing internal E2E, but that's not a 
contract to be honored. Host can't trust until device says it explicitly.

Since Btrfs already has 'nodatasum' mount option, I assume there are 
deployments that prefer to optimize for checksum. I thought Btrfs will 
be more comfortable to give up (its own checksum tree) and exercise this 
more often if it certainly knows that device is also capable.