From mboxrd@z Thu Jan 1 00:00:00 1970 From: marex@denx.de (Marek Vasut) Date: Wed, 9 May 2012 10:43:39 +0200 Subject: [RFC PATCH 00/11 V7] MXS: Add i.MX28 USB Host driver In-Reply-To: References: <1335837363-12376-1-git-send-email-marex@denx.de> <201205031527.16149.marex@denx.de> Message-ID: <201205091043.39787.marex@denx.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Dear Chen Peter-B29397, > > > > Still use IRQF_SHARED for the usb interrupt, but: > > > > * The IRQ handler in ehci-mxs, that's set in hc_driver.irq would be > > > > always > > > > returning IRQ_NONE. > > > > * The IRQ handler in imx-otg would call the callback registered via > > > > imx_otg_set_irq_handler() only if the port was actually in host mode > > > > > > The interrupt which is registered by hcd core is usb_hcd_irq, it will > > > > call > > > > > ehci_irq. The hc_driver.irq is included in usb_hcd_irq. > > > > Hm, that's correct. I'll dive in and see if there's any other way around. > > I > > think I saw IRQF_DISABLED somewhere there, but that might be just some > > phantom. > > Either way, is this direction I'm taking the correct one (have irq > > handler in > > imx-otg and send IRQs to gadget/host drivers from there)? > > Sorry for late as I can't see this email at my inbox. That's fine :) > If you can handle hcd irq well, I think it is correct direction. Good, any other problems you noticed? Thanks! Best regards, Marek Vasut