Date: Wed, 20 Feb 2013 14:30:53 -0500 From: John Baldwin <jhb@freebsd.org> To: "John-Mark Gurney" <jmg@funkthat.com> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r247012 - in head/contrib/binutils: gas/config opcodes Message-ID: <201302201430.53615.jhb@freebsd.org> In-Reply-To: <20130220184200.GN55866@funkthat.com> References: <201302192135.r1JLZH27096984@svn.freebsd.org> <201302200809.23854.jhb@freebsd.org> <20130220184200.GN55866@funkthat.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday, February 20, 2013 1:42:00 pm John-Mark Gurney wrote: > John Baldwin wrote this message on Wed, Feb 20, 2013 at 08:09 -0500: > > On Tuesday, February 19, 2013 4:35:17 pm John-Mark Gurney wrote: > > > Author: jmg > > > Date: Tue Feb 19 21:35:17 2013 > > > New Revision: 247012 > > > URL: http://svnweb.freebsd.org/changeset/base/247012 > > > > > > Log: > > > add support for AES and PCLMULQDQ instructions to binutils... > > > > > > Thanks to Mike Belopuhov for the pointer to the OpenBSD patch, though > > > OpenBSD's gcc is very different that it only helped w/ where to modify, > > > not how... Thanks to jhb for some early reviews... > > > > > > Reviewed by: imp, kib > > > MFC after: 1 month > > > > Nice! Sorry I wasn't able to review this in more detail. :( Can you also add > > support for these instructions to ddb's disassembler? > > Considering that ddb doesn't appear to support xmm registers, that'll > be a bit of work... even simple instructions such as pxor aren't there > yet... So, it'd be more like adding all of the SSE instructions to db > than just adding the AES instructions... If I had time, I'd do it, but > I don't right now.. Ahh, that's fair. The ones I added recently did not use XMM registers so they weren't as tricky. > Also, I just happen to be looking at the declaration in > amd64/amd64/db_disasm.c of: > static const char * const db_reg[2][4][16] = { > > shouldn't we change that to: > static const char const db_reg[2][4][16][6] = { > > That would save a level of indirection, and also all those pointers > associated... I estimate that it would save about 1k of space on > amd64... it might be a bit less, but at least 512 bytes... > > Just a thought... On arm or mips I'd say yes. On amd64 I doubt it would be noticable, and the first version is arguably slightly more readable. I don't really care one way or another though. -- John Baldwin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201302201430.53615.jhb>