On 12/12/14 01:42, Akira Hayakawa wrote:
1. Writeboost shouldn't split the bio into 4KB chunks. No. It is necessary. I know WALB (https://github.com/starpos/walb) logs data without splitting but the data structure becomes complicated. If you read my code carefully, you will notice that splitting helps the design simplicity and performance.
This is the first time I see someone claiming that reducing the request size improves performance. I don't know any SSD model for which splitting requests improves performance.
Additionally, since bio's are split by dm-writeboost, this makes me wonder how atomic writes will ever be supported ? Atomic writes are being standardized by the T10 SCSI committee. I don't think the Linux block layer already supports atomic writes today but I expect support for atomic writes to be added to the block layer sooner or later. See e.g. http://www.t10.org/doc13.htm / SBC-4 SPC-5 Atomic writes and reads for the latest draft specification.
Bart. -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
