On 12/19/24 3:10 PM, Damien Le Moal wrote:
On 12/20/24 02:12, Bart Van Assche wrote:
On 12/18/24 10:00 PM, Christoph Hellwig wrote:
If you rely on order. If you are doing O_APPEND-style I/O on zonefs or
using a real file systems it's perfectly fine.
Using zone append operations defeats two of the advantages of zoned
storage. One of the advantages of zoned storage is that filesystems have
control over the layout of files on flash memory with regular writes.
Zone append still allows that: you can still chose the zone you write to.
(replying to an email of two weeks ago)
Zone append does not allow to control the LBA if multiple writes are
outstanding for the same zone. We have use cases for which controlling
the file layout is important. These use cases are loading large files
and applications as quickly as possible. At least for UFS devices
sequential read performance improves if files are laid out sequentially
on the storage medium.
It is not the first time that the desire to lay out files sequentially
on UFS devices comes up. Others have worked on this before. See e.g.
Jiaming Li "[RESEND PATCH 0/4] Implement File-Based optimization
functionality", November 2022
(https://lore.kernel.org/linux-scsi/20221102053058.21021-1-lijiaming3@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx/).
File-Based Optimization
(FBO) is a mechanism for requesting UFS devices to organize logical
blocks sequentially, e.g. logical blocks of a single file. Note: my
employer has not been involved in the standardization of FBO and has no
plans to implement FBO support. We prefer zoned storage.
Bart.
