Re: [PATCH 2/2] hw/block/nvme: add write uncorrectable command

[Klaus Jensen <its@irrelevant.dk>]

[-- Attachment #1: Type: text/plain, Size: 1491 bytes --]

On Feb 12 00:37, Keith Busch wrote:
> On Thu, Feb 11, 2021 at 09:43:05AM +0100, Klaus Jensen wrote:
> > On Feb 11 12:37, Keith Busch wrote:
> > 
> > > Is there a use case with a real qemu guest wanting this?
> > 
> > Like for the extended metadata case (which also does not have a lot of
> > "public" exposure, but definitely have enterprise use), our main
> > motivation here was to ease testing for compliance suites and frameworks
> > like SPDK. 
> 
> I'm okay with the metadata patches.
> 
> > I'm honestly have no clue what so ever what a real world use
> > of Write Uncorrectable would be. It's been in the spec since 1.0, so
> > there must have been some reason, Is it just to align with SCSI WRITE
> > LONG? I'm not SCSI expert at all, but from what I can read it looks like
> > that was also intended as a feature for testing read error conditions.
> 
> I don't think it's for testing purposes.
> 
> If you need to send a burst of non-atomic writes (ex: writing a RAID
> stripe), a power failure can result in an inconsistent state where you
> don't know at a block level which ones have old data or new data. If you
> Write Uncorrectable first, you can never read old data, and thus have no
> "write hole".
> 
> Journalling solves this better, and I'm not aware of any real
> implementation relying on uncorrectable.

Right, thanks! I'm aware of the RAID write hole issue, but I did not
consider Write Uncorrectable as a possible means to solve it.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]