Re: am335x: USB DMA IRQ issue

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi,

On Sat, Feb 28, 2015 at 10:39:49PM +0100, Yegor Yefremov wrote:
> On Tue, Feb 24, 2015 at 5:36 PM, Felipe Balbi <balbi@xxxxxx> wrote:
> > Hi,
> >
> > On Thu, Feb 05, 2015 at 12:19:23PM +0100, Yegor Yefremov wrote:
> >> I have a problem with my am335x based board and USB.
> >>
> >> Kernel: Linux version 3.18.1 (YegorYefremov@development1) (gcc version
> >> 4.9.2 (Buildroot 2015.02-git-00797-gf1b07c0) ) #1 SMP Thu Jan 15
> >> 15:31:27 CET 2015
> >>
> >> I took two devices and equipped them with USB WLAN cards: Bus 001
> >> Device 003: ID 148f:5370 Ralink Technology, Corp. RT5370 Wireless
> >> Adapter. One device as AP and another as client. Client makes
> >> permanent ping to AP. And from time to time I start nuttcp session.
> >> After 2-3 days I get following errors:
> >
> > 2-3 days ? oh man... :-)
> >
> > Here's one thing to try. Can you do some variable size pingtest ?
> > Something like below should do ?
> >
> > random()
> > {
> >         size=$(dd if=/dev/urandom count=1 2>/dev/null|cksum|
> >         cut -f1 -d" "|tr '-' '\0')
> >         size=$(expr $size % 60000)
> > }
> >
> > num=$1
> >
> > if [ -z $num ]
> > then
> >         num=10000
> > fi
> >
> > while ! ifconfig usb0 >/dev/null 2>&1; do
> >         printf "waiting for usb0\n"
> >         sleep 1;
> > done
> >
> > ifconfig usb0 192.168.2.14
> >
> > for i in $(seq 1 $num); do
> >         random
> >         printf "%d: \t pinging with size %-27d" $i $size
> >         if ! ping -c 6 192.168.2.15 -s $size -q -i 0 >/dev/null 2>&1; then
> >                 printf "FAILED\n"
> >                 break
> >         fi
> >         printf "PASSED\n"
> > done
> 
> Still haven't tried your test, but have seen this interesting
> patch/proposal http://e2e.ti.com/support/arm/sitara_arm/f/791/t/404743.
> Seems to be somehow related.

patch makes sense. If you can make sure it really solves your problem
then send it as a proper patch on top of my testing/next, I'd be glad.

At some point, I'll just split dma callback into dedicated rx and tx
callbacks to avoid confusion in the future.

-- 
balbi

Attachment: signature.asc
Description: Digital signature


[Index of Archives]     [Linux Arm (vger)]     [ARM Kernel]     [ARM MSM]     [Linux Tegra]     [Linux WPAN Networking]     [Linux Wireless Networking]     [Maemo Users]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux