[PATCH 0/2] usb: core: return -EAGAIN on usb_* functions reentrance

Peter Mamonov pmamonov at gmail.com
Thu Oct 15 05:38:56 PDT 2015


On Wed, 14 Oct 2015 16:09:08 +0200
Sascha Hauer <s.hauer at pengutronix.de> wrote:

> Hi Peter,
> 
> On Wed, Oct 14, 2015 at 01:15:47PM +0300, Peter Mamonov wrote:
> > The second patch of this patch set adds reentrance detection to 
> > the the following functions: usb_submit_int_msg, usb_control_msg and
> > usb_bulk_msg. These functions will return -EAGAIN if reentered. 
> > Other public usb_* functions call one of these three functions.
> > This patch enables one to use regular *delay() functions
> > throughout the ehci-hcd.c.
> > 
> > The first patch adds check for -EAGAIN return value to the usb
> > keyboard driver. The driver will skip keyboard poll upon getting
> > this error value from either usb_submit_int_msg() or
> > usb_get_report() functions.
> > 
> > Peter Mamonov (2):
> > 	input: usb_kbd: skip poll on -EAGAIN error
> > 	usb: core: return -EAGAIN on usb_* functions reentrance
> 
> Applied on -next. I believe we no longer need mdelay_non_interruptible
> in the ehci-hcd driver, right? Also we don't have to detect
> re-entrancy in the ehci driver anymore. I have applied some patches
> removing that to -next. Could you check the result?

Yes.
Yes.
I've tested the current "next" branch - the usb stuff works fine.

> 
> Sascha
> 
> 




More information about the barebox mailing list