Am Donnerstag, den 01.10.2015, 16:13 +0800 schrieb Peter Chen:

On Wed, Sep 30, 2015 at 4:52 PM, Lucas Stach [off-list ref] wrote:

quoted

Am Mittwoch, den 30.09.2015, 10:17 +0800 schrieb Peter Chen:

quoted

It can improve the USB performance when choosing larger
burst size at some systems (bus size is larger), there is
no side effect if this burst size is larger than bus size.

Just for the record, as it seems too late to stop the train now:

I am _NOT_ happy that the extended binding has been accepted into the
USB tree despite outstanding review comments from me not being addressed
or even answered.

I sent patch for review at Aug 7th, queued this patches at Aug 13th
[1], and sent pull
request at Aug 14th, I am curious why you did not give comments
between this period utill
I complained one ethernet patch breaks i.mx ethernet function for v4.2
tree [2] at Aug 14th.

What do the USB changes have to do with any ethernet changes?

The comments I'm referring to are from Aug 14th
http://lists.infradead.org/pipermail/linux-arm-kernel/2015-August/364396.html

So you are telling me that giving reviewers not even a week before
sending the pull request is appropriate? And that sending the pull
despite there being open review comments is ok? Think again.

I will consider your comments, and send patch for improving it if possible.

As I said it may be too late now. By sending the pull request including
the binding change you forced everyone to now either accept the (in my
opinion not really perfect) binding, or hurry to change it in the
current RC phase, something which I'm not comfortable with. The damage
has been done.

[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2015-August/364352.html
[2] http://marc.info/?l=linux-netdev&m=143954500409636&w=2

-- 
Pengutronix e.K.             | Lucas Stach                 |
Industrial Linux Solutions   | http://www.pengutronix.de/  |