I don't know why it would pause after 512k bytes, but it is an extremely
foolish design to provide only 50us of buffering. Bulk transfers are not
guaranteed to get any bandwidth. Even with an isochronous pipe, with its
guaranteed bandwidth, there is no guarantee that you could meet that
requirement. In the worst case, you need to provide almost 200us of
buffering.
I think so too.
But I must run such situation.
I search information about this,
but I cann't find.
It seems that
lower USB driver separates our one request to two requests per 512KB.
I want to change the size(512KB).
No, actually the lower USB driver chops up your request into requests no
larger than your bulk packet size, probably 512 bytes. No one chops it
into pieces of 512k.
--
Tim Roberts, timr@xxxxxxxxx
Providenza & Boekelheide, Inc.
.
Re: RFC: I/O bandwidth controller ... The implementation of an I/O scheduling algorithm is to a certain extent ...Proportional bandwidth scheduling: each process/group of processes ... Please note that the only thing we can do is to guarantee minimum...request is issued to the drive, look at the time the request is served, ... (Linux-Kernel)
Re: RFC: I/O bandwidth controller ... The implementation of an I/O scheduling algorithm is to a certain extent ...Proportional bandwidth scheduling: each process/group of processes ... Please note that the only thing we can do is to guarantee minimum...request is issued to the drive, look at the time the request is served, ... (Linux-Kernel)
Re: RFC: I/O bandwidth controller ... things like allocating I/O bandwidth on a per-device basis. ... The whole subsystem is created at compile time, but controller data... Please note that the only thing we can do is to guarantee minimum... the minimum bandwidth is more important than limitation of maximum bandwidth ... (Linux-Kernel)
Re: RFC: I/O bandwidth controller ... The implementation of an I/O scheduling algorithm is to a certain extent ...Proportional bandwidth scheduling: each process/group of processes ... Please note that the only thing we can do is to guarantee minimum... Of course, IO throttling is important, but, personally, I think guaranteeing the minimum bandwidth is more important than limitation of maximum bandwidth to satisfy the requirement in real business sites. ... (Linux-Kernel)
Re: RFC: I/O bandwidth controller ... The implementation of an I/O scheduling algorithm is to a certain extent ...Proportional bandwidth scheduling: each process/group of processes ... Please note that the only thing we can do is to guarantee minimum... Of course, IO throttling is important, but, personally, I think guaranteeing the minimum bandwidth is more important than limitation of maximum bandwidth to satisfy the requirement in real business sites. ... (Linux-Kernel)
