From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ey-out-2122.google.com (ey-out-2122.google.com [74.125.78.24]) by ozlabs.org (Postfix) with ESMTP id D18E1DE0D9 for ; Thu, 15 Jan 2009 06:29:19 +1100 (EST) Received: by ey-out-2122.google.com with SMTP id d26so87921eyd.15 for ; Wed, 14 Jan 2009 11:29:17 -0800 (PST) Sender: Peter Korsgaard To: Timur Tabi Subject: Re: [PATCH] fsldma: print correct IRQ on mpc83xx References: <1231944284-18296-1-git-send-email-jacmet@sunsite.dk> <496E0E85.7060801@freescale.com> <87iqohc19b.fsf@macbook.be.48ers.dk> <496E1317.9070804@freescale.com> <87eiz5c074.fsf@macbook.be.48ers.dk> <496E15E8.2060106@freescale.com> From: Peter Korsgaard Date: Wed, 14 Jan 2009 20:29:15 +0100 In-Reply-To: <496E15E8.2060106@freescale.com> (Timur Tabi's message of "Wed\, 14 Jan 2009 10\:42\:16 -0600") Message-ID: <87ab9tbsac.fsf@macbook.be.48ers.dk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linuxppc-dev@ozlabs.org, dan.j.williams@intel.com List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , >>>>> "Timur" == Timur Tabi writes: Timur> Peter Korsgaard wrote: >> Documentation/powerpc/dts-bindings/fsl/dma.txt and >> The NO_IRQ check in fsldma.c:fsl_dma_chan_probe() >> >> And it makes sense, there's no per-channel DMAC interrupts on mpc83xx. Timur> But the device trees do have IRQs in the channels: Timur> arch/powerpc/boot/dts/mpc836x_mds.dts: Timur> So I don't see what bug you're trying to fix. If you're Timur> saying that the interrupts should not be specific DMA Timur> channels, then the proper fix is to have the DMA driver pick Timur> up the interrupts from the DMA controller's node when Timur> necessary. Hacking up a printk doesn't fix anything. Ok, let me try again. SOME device trees indeed have an interrupt property per channel, but the bindings (Documentation/powerpc/dts-bindings/fsl/dma.txt) and the probe code in fsldma.c doesn't require it. The only minor issues we have is the cosmetic problem of the driver printing NO_IRQ (E.G. this patch), and the driver calling free_irq(NO_IRQ) (E.G. the other patch I sent today) in the release path. -- Bye, Peter Korsgaard