On Fri, Dec 12, 2008 at 1:09 AM, Paul Walmsley <paul@xxxxxxxxx> wrote: > Hello Grazvydas, > > On Fri, 12 Dec 2008, Grazvydas Ignotas wrote: > >> On Thu, Dec 11, 2008 at 11:07 PM, Paul Walmsley <paul@xxxxxxxxxxx> wrote: >> > Gravyzdas, can you send us a commit ID that results in successful device >> > enumeration (after reverting >> > http://marc.info/?l=linux-omap&m=122453174024860&w=2 ?) Sounds like it >> > was working for you in late October with that revert. >> >> There is no need to - EHCI works (almost) fine on pandora with current HEAD. > > Might you be willing to try a build with the first two patches > from this series posted earlier today with the subject: > > [PATCH 0/3 NOT FOR MERGING] ehci-omap clock rewrite test patches > > and see if things still almost work? If so, then we can get the nasty > direct CM writes out of that driver. Yes, with those it works exactly as without them (at least I haven't noticed any difference). >> I'm also getting suspend/resume related problems if I plug a hub >> without any devices connected (29sec), and then a device (52sec): > > If you disable USB autosuspend with no devices connected via > > echo -1 > /sys/module/usbcore/parameters/autosuspend > > and then repeat the same test as above, do your devices stay alive? Yep, this seems to resolve that issue, I can plug the devices any way I want with this. Thanks for showing this. On Fri, Dec 12, 2008 at 7:59 PM, Steve Sakoman <sakoman@xxxxxxxxx> wrote: > On Thu, Dec 11, 2008 at 12:41 PM, Grazvydas Ignotas <notasas@xxxxxxxxx> wrote: > >> If your host uses PHY mode, have you hacked EXT_PHY_RESET_GPIO* in >> drivers/usb/host/ehci-omap.c? BTW, if this stuff is really needed, it >> should go to platform_data.. > > What PHY chip are you using on Pandora? > > From the schematic it appears that Gumstix is using USB3320. > > I'm really puzzled by the difference in behavior :-( Unfortunately I don't know, will need to ask out hardware designer. The chip is so small that I can't see any markings on it. Does your PHY chip heat? There was actually a similar problem with rev2 pandora boards, where it would not detect devices and that tiny PHY chip would heat really badly. It later turned out to be a bad batch of PHY chips, so after switching to another and reflowing the boards USB would work fine (almost, with issues above). -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
