This is the mail archive of the
libc-ports@sources.redhat.com
mailing list for the libc-ports project.
Re: [ARM] architecture specific subdirectories & optimised memchr [V3]
- From: David Gilbert <david dot gilbert at linaro dot org>
- To: Roland McGrath <roland at hack dot frob dot com>
- Cc: libc-ports at sourceware dot org, patches at linaro dot org, "Joseph S. Myers" <joseph at codesourcery dot com>
- Date: Mon, 3 Oct 2011 09:48:52 +0100
- Subject: Re: [ARM] architecture specific subdirectories & optimised memchr [V3]
- References: <20110715181101.GA20980@davesworkthinkpad> <Pine.LNX.4.64.1108021437200.16898@digraph.polyomino.org.uk> <CA+1XiSfT7Ceky_bi1SzkcKbVFQh8WsKRcFKp39hcDRVS3PmexQ@mail.gmail.com> <Pine.LNX.4.64.1108041949260.31451@digraph.polyomino.org.uk> <20110928175913.GA11067@davesworkthinkpad> <20110928181810.EBCDB2C09B@topped-with-meat.com>
On 28 September 2011 19:18, Roland McGrath <roland@hack.frob.com> wrote:
> The usual way to do this sort of thing is by using --with-cpu at configure
> time.
See discussion http://cygwin.com/ml/libc-ports/2011-08/msg00005.html
The short story is that --with-cpu sets $submachine not $machine; in this case
we're using $machine to pick up ARM variants that have incompatible
extra instructions.
We could use $submachine to pick further subvariants - e.g. machine=armv7
and submachine=cortex-a9 for stuff which just has differences in optimisation.
My first version used $submachine rather $machine to let you do this
but was persuaded
that it was the wrong way!
Dave