On 10/11/23 21:07, charlesfdotz@xxxxxxxxxxxx wrote:
Oct 11, 2023, 00:33 by dlemoal@xxxxxxxxxx:
On 10/10/23 23:31, charlesfdotz@xxxxxxxxxxxx wrote:
As I've said in all of my emails this is regarding drive-managed SMR disks that do no support trim. Most of these aren't detectable by the host because if there were then people wouldn't buy them as they're not fit for purpose. Western digital was sued over this and lost.
https://arstechnica.com/gadgets/2020/05/western-digital-gets-sued-for-sneaking-smr-disks-into-its-nas-channel/
https://www.tomshardware.com/news/wd-red-smr-lawsuit-pays-out-pennies-in-settlement-damages
There is no need to name names or point fingers here. Kernel open source
development is vendor agnostic and while my employer is indeed Western Digital,
I do not talk as a representative of WD products but as a kernel developer. So
calm down please if you want to keep the discussion open.
You can opine that this unlikely but we have reports (which I've already linked) that this works on some SMR drives. It's also pretty ironic someone from western digital of all places is saying drives wouldn't lie about how they function so they must be writing zeros when told to.
Again, the finger pointing here is totally inappropriate. My points are all in
good faith and I am talking about the drives I know of.
In any case, the key-word in your statement is "some drives". Given that you are
talking about something that is not standardized and so cannot be safely
generalized, there is nothing we can do. It is a hard NO from me to replace the
lack of trim support with writing zeroes.
Sincerely,
Chuck
--
Damien Le Moal
Western Digital Research
Presumably drives of the same model will work the say way so when I say "some drives" we're talking hundreds of
thousands or millions if even a few models function this way. In my initial email I included a link where this
seems to help on some models so I think it's very possible this could've made quite a lot of drives much more
usable. I was requesting a separate device mapper layer with the thought that someone with one of these drives
would test it and if zeroing sectors helped then they could add a "dm-trimemulation" layer on top of it. That's
not replacing anything only adding an option for users.
Oh, sure, they could. Problem here is that we don't have any specifics
on why this 'dm-trimemulation' layer helps in this situation.
What we do know, though, is that it works exactly _against_ how SMR
drives work internally. On SMR drives writing zeroes is about the
worst-case scenario, as it'll fill the disk with meaningless data but
_not_ reset zones such that you can write to it. IE essentially making
the drive unusable.
Understandably that we're a little concerned here.
Even more so as this is a firmware emulation issue on the drive, so
it _could_ be fixed by the drive manufacturer in updating the firmware
to actually support TRIM.
But sure, if you found that there are drives for this such an emulation
might help, by all means, send patches. I'd be happily reviewing them.
Cheers,
Hannes
--
Dr. Hannes Reinecke Kernel Storage Architect
hare@xxxxxxx +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), Geschäftsführer: Ivo Totev, Andrew
Myers, Andrew McDonald, Martje Boudien Moerman
