From: Junio C Hamano <gitster@pobox.com>
Date: Mon, 23 Jul 2012 05:42:23 GMT
Subject: Re: [PATCH lt/block-sha1 0/2 v3] block-sha1: avoid pointer conversion that violates alignment constraints
Message-ID: <7vboj7nidc.fsf@alter.siamese.dyndns.org>
In-Reply-To: <20120723052605.GA13728@burratino>

Jonathan Nieder <jrnieder@gmail.com> writes:

> Junio C Hamano wrote:
>> Jonathan Nieder <jrnieder@gmail.com> writes:
>
>>> The big-endian part was just my idiocy, sorry.
>>
>> Hrm, do we want an update log message for 1/2 then?
>
> Hm, I thought all the crazy had been eliminated already.
>
> *looks again*
>
> I guess "using a single 32-bit load" makes it sound like it's using a
> big-endian load instead of a load followed by twiddling in registers.
>
> Simplest fix would be to drop the phrase "by using a single 32-bit
> load", leaving "... and gcc takes full advantage, resulting in a whole
> bunch of unaligned access traps."

I was just being stupid, misread the original code and did not
realize that the generic code quoted in your log message:

	#define get_be32(p) ( \
        	... byte-at-a-time implementation )

is not limited to big endian boxes [*1*].  So I think the message is
fine as-is.


[Footnote]

*1* In other words, the ntohl(*(uint *)(p)) is used only on selected
little endian boxes, but that does not mean the generic code was
big-endian only.  Little endian boxes that are not listed in #if
block do use the generic byte-at-a-time macro.