From: Harvey Harrison Date: Thu, 16 Oct 2008 00:32:51 +0000 Subject: Re: [PATCH] sparc64: Annotate pointers in PeeCeeI.c Message-Id: <1224117171.5815.17.camel@brick> List-Id: References: <1224116119.5815.11.camel@brick> <20081015.172351.96225000.davem@davemloft.net> In-Reply-To: <20081015.172351.96225000.davem@davemloft.net> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: David Miller Cc: sparclinux@vger.kernel.org, linux-kernel@vger.kernel.org, akpm@linux-foundation.org On Wed, 2008-10-15 at 17:23 -0700, David Miller wrote: > From: Harvey Harrison > Date: Wed, 15 Oct 2008 17:15:19 -0700 > > > This has no functional changes, but annotates the code to make > > the endianness more clear. In addition, removes some of the only > > users of cpu_to_le[16|32]p in the kernel. > > > > Signed-off-by: Harvey Harrison > > Remind me what the problem is with cpu_to_{le,be}*()? > Nothing inherently wrong with them, other than it seems that nearly every user (not this one) would be better served using the by-value versions. > The interface names define a direction, which in this case is > "CPU endianness to BE/LE endianness". And that is exactly > what is happening in the out*() routines. > > This aids comprehension of the code and is quite useful IMHO. Yes, and even from that angle I think my patch is more instructive to understanding the direction, compare for example in the same file outsw versus outsl which this patch changes. Harvey