From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lee Jones Subject: Re: [PATCH 4/7] mfd: add MFD based flexcard driver Date: Wed, 14 Aug 2013 10:45:28 +0100 Message-ID: <20130814094528.GC4046@lee--X1> References: <1376384922-8519-1-git-send-email-b.spranger@linutronix.de> <1376384922-8519-6-git-send-email-b.spranger@linutronix.de> <20130813095508.GA5109@lee--X1> <20130814101247.3ad43494@mitra.spranger.biz> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: netdev@vger.kernel.org, Alexander Frank , Sebastian Andrzej Siewior , Samuel Ortiz , Holger Dengler To: Benedikt Spranger Return-path: Received: from mail-ea0-f173.google.com ([209.85.215.173]:50194 "EHLO mail-ea0-f173.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752567Ab3HNJpf (ORCPT ); Wed, 14 Aug 2013 05:45:35 -0400 Received: by mail-ea0-f173.google.com with SMTP id g10so4694936eak.4 for ; Wed, 14 Aug 2013 02:45:34 -0700 (PDT) Content-Disposition: inline In-Reply-To: <20130814101247.3ad43494@mitra.spranger.biz> Sender: netdev-owner@vger.kernel.org List-ID: > > Wow, this is a huge patch which is going to be an insanely long, > > laborious and complex task to review it all. > This is a card with a "insane" amount of features: >=20 > 1) a Interrupt controller > 2) a DMA engine > 3) up to 8 CAN devices > 4) up to 4 FlexRay devices > 5) a dozen different trigger sources > 6) timer > 7) timestamp counter Right. Even more reason for the the patch to be broken up into many. > > Can you split this up into many patches? > Not sure how to split this up into a smaller and workable series of p= atches. > Maybe rip out the header and the DMA part into a separate patch. I was thinking more along the lines of: Patch 1: Basic compilation support - Core driver Kconfig entry - Introduction of the core driver - Basic probe(): memory allocation, register as MFD device - Basic remove(): unregister as MFD device =20 Patch 2: Basic PCI support - pci_enable_device() - pci_request_regions() - Memory mapping =20 Patch 3: - fc_init_dma() =20 Patch 4: - fc_init_uio() =20 Patch 5: - fc_init_event() =20 Patch 6: - fc_init_clk() =20 Patch 7: - fc_init_bus() =20 Patch 8 - 8+n: Start adding sysfs entries =20 Patch 8+n+1: Basic IRQ support =20 Etc etc etc Remember, the smaller the patches, the more chance that they'll be accepted. I already see quite a few issues with this patch. Breaking it up will allow for more thorough review and finer grain feedback. Issues I see after first glance: - Lack of Device Tree support - Lack of documentation for new sysfs entries - Different IRQs with same .name property - Memory allocations and mapping should be managed (devm_*) --=20 Lee Jones Linaro ST-Ericsson Landing Team Lead Linaro.org =E2=94=82 Open source software for ARM SoCs =46ollow Linaro: Facebook | Twitter | Blog