From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Warren Subject: Re: [RFC 1/5] usb: chipidea: Add support for Tegra20/30/114/124 Date: Thu, 26 May 2016 15:22:23 -0600 Message-ID: <5747690F.20707@wwwdotorg.org> References: <20160526154005.11558-1-thierry.reding@gmail.com> <20160526154005.11558-2-thierry.reding@gmail.com> <574767D8.8020203@wwwdotorg.org> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <574767D8.8020203-3lzwWm7+Weoh9ZMKESR00Q@public.gmane.org> Sender: linux-tegra-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Thierry Reding Cc: Peter Chen , Greg Kroah-Hartman , Alexandre Courbot , Jon Hunter , linux-usb-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-tegra-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-tegra@vger.kernel.org On 05/26/2016 03:17 PM, Stephen Warren wrote: > On 05/26/2016 09:40 AM, Thierry Reding wrote: >> From: Thierry Reding >> >> All of these Tegra SoC generations have a ChipIdea UDC IP block that can >> be used for device mode communication with a host. Implement rudimentary >> support that doesn't allow switching between host and device modes. > > Are you sure this is correct for Tegra20? I ask because for the /host/ > mode driver, there's a "has_hostpc" flag which is set to false for > Tegra20 and true for all other SoCs. In the U-Boot device mode driver > (if not in the kernel driver; I didn't check), there's a concept of "has > hostpc" too. I might expect that flag to be set the same way for both > drivers. That said, I /think/ the host and device HW are unrelated, so > it's possible has_hostpc might be set differently for them. > Unfortunately, we haven't enabled the device mode driver for any Tegra20 > system in U-Boot so I can't tell whether we should enable has_hostpc for > Tegra20's device mode driver. > > Still, if this code works then I guess it's likely correct... On the other hand, it looks like the kernel device mode driver might auto-detect this; in core.c:hw_device_init(), I see: reg = hw_read(ci, CAP_HCCPARAMS, HCCPARAMS_LEN) >> __ffs(HCCPARAMS_LEN); ci->hw_bank.lpm = reg; ... and in host.c:host_start(), I see: ehci->has_hostpc = ci->hw_bank.lpm;