From mboxrd@z Thu Jan 1 00:00:00 1970 From: paul.gortmaker@windriver.com (Paul Gortmaker) Date: Tue, 28 Feb 2012 18:52:45 -0500 Subject: [PATCH] ARM: mx51: Fix build error due to missing include of In-Reply-To: <20120228132108.GB3617@n2100.arm.linux.org.uk> References: <1330431110-22020-1-git-send-email-fabio.estevam@freescale.com> <20120228125336.GA3617@n2100.arm.linux.org.uk> <20120228132108.GB3617@n2100.arm.linux.org.uk> Message-ID: <20120228235244.GA3778@windriver.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org [Re: [PATCH] ARM: mx51: Fix build error due to missing include of ] On 28/02/2012 (Tue 13:21) Russell King - ARM Linux wrote: [...] > The answer is not to start scattering lots of .c files with linux/bug.h > includes, but to fix linux/kernel.h. Hi Russell, I understand your concerns, and had I found myself "buggering" lots of .c files, I would have backed away and said to myself, "OK, this really isn't going to work - we really can't do this." But to put some metrics to it, after building for about 10 different arch and all their defconfigs, I only had to change about twenty[1] of the approximately sixteen thousand .c files in the kernel. That is about 0.125% -- which is one of the things that made me think "Hey, we can really do this cleanup!" Including this file, there were only two .c files in all of arch/arm that were using bug somehow. So I think having kernel.h not require bug.h is completely do-able. But in the end, I'm not welded to the idea. I think it is a win, but I'm not going to kick up a big stink if Linus or Andrew say they don't think it is of value, and would rather not rock the boat. Thanks, Paul. -- [1] https://lkml.org/lkml/2012/1/26/521