Date: Sat, 03 Jun 2006 20:35:50 +0000 From: "Mikhail T." <mi@aldan.algebra.com> To: John Baldwin <jhb@freebsd.org> Cc: Mikhail Teterin <mi+mx@aldan.algebra.com>, freebsd-current@freebsd.org Subject: Re: Assembler optimizations for libz Message-ID: <200606032035.52075@Misha> In-Reply-To: <200606031009.04640.jhb@freebsd.org> References: <200606022121.20535.mi%2Bmx@aldan.algebra.com> <200606031009.04640.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
=D3=D5=C2=CF=D4=C1 03 =DE=C5=D2=D7=C5=CE=D8 2006 14:09, John Baldwin, =F7= =C9 =CE=C1=D0=C9=D3=C1=CC=C9: =3D Can you expand on the problem you have with inffast.S? There exist both inffast.c and inffast.S. Although the latter is in a directory, that is explicitly listed in the VPATH, the former is picked up. I think, it should be the other way around -- the current directory, that's always in the VPATH implicitly, should be consulted last... =46or this reason, I had to rename inffast.S to _inffast.S. =3D Also, it might read better if the flow of the Makefile is [...] =3D That is, merge the two i386 sections into one section. I'm quite confident, that I will NOT be able to pick the right color for this structure. :-) I'll leave the perfection of the Makefile to whoever commits it -- what's in it now works good enough to demonstrate the benefit of the optimization... Whoever gets to it, though, should add the amd64 implementation too (posted in this thread). It is just the compression for now, I'm still working on the inffast.S... Thanks! Yours, -mi http://www.freebsd.org/cgi/query-pr.cgi?pr=3Dkern/96393
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200606032035.52075>