News for hackers. Stuff that matters.

A comment on Linux Kernel Documentation: NAND Error-Correction Code.

Re:Linux Kernel Documentation: NAND Error-Correction Code by pedrocr on Wednesday November 22, 2023 @11:13AM (#38381132)
Is there an NVMe manufacturer that sells their product as just naked flash? The first rule of hardware is that hardware providers don't know how to write software. It would be nice if we could get rid of the proprietary and buggy embedded operating systems running our storage as an opaque abstraction layer. Things like lying about flush that was recently on HN would stop being a problem.
[ Parent | on HN ]
Re:Linux Kernel Documentation: NAND Error-Correction Code by hlandau on Wednesday November 22, 2023 @01:02PM (#38382665)
These are called "open-channel SSDs", which seem to currently mainly be consigned to research purposes AFAICT. But some do exist.

https://en.wikipedia.org/wiki/Open-Channel_SSD

[ Parent | on HN ]
Re:Linux Kernel Documentation: NAND Error-Correction Code by cesarb on Wednesday November 22, 2023 @12:28PM (#38382163)
While not going as far as you want, one movement in that direction is NVMe Zoned Namespaces (https://www.anandtech.com/show/15959/nvme-zoned-namespaces-e...).
[ Parent | on HN ]
Re:Linux Kernel Documentation: NAND Error-Correction Code by loeg on Wednesday November 22, 2023 @12:46PM (#38382430)
Flexible Data Placement is another interesting development in the NVMe space.
[ Parent | on HN ]
Re:Linux Kernel Documentation: NAND Error-Correction Code by BitBanging on Wednesday November 22, 2023 @12:47PM (#38382439)
The NAND flash manufacturers usually have no software involvement unless they are in the business of also selling SSD controller SoCs. They will typically sell FDKs/SDKs that are heaping piles of garbage written overseas for their SoCs that provide minimum functionality for interfacing with all of the components.

I wouldn't really expect controller-less NAND flash to ever meaningfully compete with what we have now. There is a lot more specialized hardware in a typical SoC than just a microprocessor and NAND flash chips

[ Parent | on HN ]
Re:Linux Kernel Documentation: NAND Error-Correction Code by acquacow on Wednesday November 22, 2023 @11:50AM (#38381647)
Well, by definition, naked flash wouldn't be NVMe... so no.
[ Parent | on HN ]
Re:Linux Kernel Documentation: NAND Error-Correction Code by FirmwareBurner on Wednesday November 22, 2023 @11:53AM (#38381687)
It would just be flash
[ Parent | on HN ]