|
From: Nikolaus R. <Nik...@ra...> - 2017-02-16 01:16:04
|
On Feb 12 2017, Yue Li <yli-wUU9E3n5/m56...@pu...> wrote:
> I noticed that in my testing, the writing size has always been set to
> 131072 bytes during each write (unless the actual data size is smaller
> than that), and during the reading, some times it is 131072, but
> sometimes it can be changed to 65536 and other sizes.
>
> I'm wondering are these write and read sizes controlled by user
> application (in my test i used dd), or it can actually be controlled by
> FUSE?
Depends. read() size is controlled by the kernel, unless you use
direct_io. write() size is controlled by userspace, unless you use
writeback_cache. Your filesystem can limit the maximum size of requests
that the kernel will send.
Best,
-Nikolaus
--
GPG encrypted emails preferred. Key id: 0xD113FCAC3C4E599F
Fingerprint: ED31 791B 2C5C 1613 AF38 8B8A D113 FCAC 3C4E 599F
»Time flies like an arrow, fruit flies like a Banana.«
|