Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754796AbcKYODd convert rfc822-to-8bit (ORCPT ); Fri, 25 Nov 2016 09:03:33 -0500 Received: from unicorn.mansr.com ([81.2.72.234]:42350 "EHLO unicorn.mansr.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754768AbcKYODX (ORCPT ); Fri, 25 Nov 2016 09:03:23 -0500 From: =?iso-8859-1?Q?M=E5ns_Rullg=E5rd?= To: Russell King - ARM Linux Cc: Vinod Koul , Mason , Lars-Peter Clausen , Dave Jiang , Arnd Bergmann , Mark Brown , Linus Walleij , Bartlomiej Zolnierkiewicz , LKML , Laurent Pinchart , dmaengine@vger.kernel.org, Dan Williams , Jon Mason , Lee Jones , Maxime Ripard , Linux ARM Subject: Re: Tearing down DMA transfer setup after DMA client has finished References: <58356EA8.2010806@free.fr> <20161125045549.GC2698@localhost> <20161125124528.GJ14217@n2100.armlinux.org.uk> <20161125133436.GK14217@n2100.armlinux.org.uk> <20161125135830.GL14217@n2100.armlinux.org.uk> Date: Fri, 25 Nov 2016 14:03:20 +0000 In-Reply-To: <20161125135830.GL14217@n2100.armlinux.org.uk> (Russell King's message of "Fri, 25 Nov 2016 13:58:30 +0000") Message-ID: User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/24.5 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1404 Lines: 28 Russell King - ARM Linux writes: > On Fri, Nov 25, 2016 at 01:50:35PM +0000, M?ns Rullg?rd wrote: >> Russell King - ARM Linux writes: >> > It would be unfair to augment the API and add the burden on everyone >> > for the new API when 99.999% of the world doesn't require it. >> >> I don't think making this particular dma driver wait for the descriptor >> callback to return before reusing a channel quite amounts to a horrid >> hack. It certainly wouldn't burden anyone other than the poor drivers >> for devices connected to it, all of which are specific to Sigma AFAIK. > > Except when you stop to think that delaying in a tasklet is exactly > the same as randomly delaying in an interrupt handler - the tasklet > runs on the return path back to the parent context of an interrupt > handler. Even if you sleep in the tasklet, you're sleeping on behalf > of the currently executing thread - if it's a RT thread, you effectively > destroy the RT-ness of the thread. Let's hope no one cares about RT > performance on that hardware... That's why I suggested to do this only if the needed delay is known to be no more than a few bus cycles. The completion callback is currently the only post-transfer interaction we have between the dma and device drivers. To handle an arbitrarily long delay, some new interface will be required. -- M?ns Rullg?rd