Date: Tue, 4 Jan 2005 00:04:21 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 491 ************************************************** Monday 03 January 2005 Number 491 ************************************************** Subjects for today 1 Re: fork() FXSAVE and Pentium Pro : Stefan.Neis at t-online.de 2 Re: fork() FXSAVE and Pentium Pro : lsunley at mb.sympatico.ca 3 SSES-alpha + Innotek libc : Dave Yeo" 4 Re: fork() FXSAVE and Pentium Pro : Dave Yeo" 5 Re: fork() FXSAVE and Pentium Pro : lsunley at mb.sympatico.ca 6 Re: SSES-alpha + Innotek libc : John Poltorak 7 Re: fork() FXSAVE and Pentium Pro : Stefan.Neis at t-online.de 8 cmovbe : John Poltorak **= Email 1 ==========================** Date: Sun, 2 Jan 2005 19:25:45 +0100 From: Stefan.Neis at t-online.de Subject: Re: fork() FXSAVE and Pentium Pro Dave Yeo wrote: > On Thu, 2004-30-12 at 18:45 +0100, Christian Hennecke wrote: > > >the Pentium Pro CPU also SIGILL's with fork(). > > > > Oh well. Guess what the CPUs of my two machines are - an AMD K6-III > > and > > a Pentium Pro. I guess the K6-III haves the same as the K6-2 in that > > regard, or doesn't it? > > > > FXSAVE was introduced with the Pentium II and a half, AMD introduced it > with the K7 Just for completeness, what you say about AMD apparently is not quite true. **= Email 2 ==========================** Date: Sun, 02 Jan 2005 12:58:14 -0500 From: lsunley at mb.sympatico.ca Subject: Re: fork() FXSAVE and Pentium Pro In <1ClAQE-2A6MlM0 at fwd04.sul.t-online.com>, on 01/02/05 at 07:25 PM, Stefan.Neis at t-online.de said: >Dave Yeo wrote: >> On Thu, 2004-30-12 at 18:45 +0100, Christian Hennecke wrote: >> > >the Pentium Pro CPU also SIGILL's with fork(). >> > >> > Oh well. Guess what the CPUs of my two machines are - an AMD K6-III >> > and >> > a Pentium Pro. I guess the K6-III haves the same as the K6-2 in that >> > regard, or doesn't it? >> > >> >> FXSAVE was introduced with the Pentium II and a half, AMD introduced it >> with the K7 >Just for completeness, what you say about AMD apparently is not quite >true. and... the complete information is??? -- ----------------------------------------------------------- lsunley at mb.sympatico.ca ----------------------------------------------------------- **= Email 3 ==========================** Date: Sun, 02 Jan 2005 11:36:50 -0800 From: "Dave Yeo" Subject: SSES-alpha + Innotek libc Interesting thing, I was playing around with star-1.5 (built with Innotek libc), tried to create an archive and star complained about not knowing who I was and couldn't find me in passwrd and refused to run. Ran whoami (from SSES in a innotek enviroment) and it complained , said I was L31 or somesuch. Updated SSES to sses-alpha-0.30 and everything was happy. Whoami reported I was root and star happily (well at least until it hit the 2GB limit) created an archive Interesting interplay between the two systems and if anyone is using libc+sses make sure you update to the latest Dave **= Email 4 ==========================** Date: Sun, 02 Jan 2005 13:24:52 -0800 From: "Dave Yeo" Subject: Re: fork() FXSAVE and Pentium Pro On Sun, 2 Jan 2005 19:25:45 +0100, Stefan.Neis at t-online.de wrote: >> FXSAVE was introduced with the Pentium II and a half, AMD introduced it >> with the K7 > >Just for completeness, what you say about AMD apparently is not quite true. That could well be. I had a hard time googling the info, mostly getting mailing lists about the Linux Kernel. When did AMD introduce the FXSAVE instruction? Also I'll add that the FXSAVE doesn't seem to be finallized on Intel until the PIII. While on the subject could someone try this on a platform that does support FXSAVE and report? (http://www.tbcnet.com/~clive/fxsave.zip) Most likely have to open a DOS window first. Dave **= Email 5 ==========================** Date: Sun, 02 Jan 2005 15:59:35 -0500 From: lsunley at mb.sympatico.ca Subject: Re: fork() FXSAVE and Pentium Pro In <20050102212453.4125FC34EE at generation.lgisp.net>, on 01/02/05 at 01:24 PM, "Dave Yeo" said: >On Sun, 2 Jan 2005 19:25:45 +0100, Stefan.Neis at t-online.de wrote: >>> FXSAVE was introduced with the Pentium II and a half, AMD introduced it >>> with the K7 >> >>Just for completeness, what you say about AMD apparently is not quite true. >That could well be. I had a hard time googling the info, mostly getting >mailing lists about the Linux Kernel. When did AMD introduce the FXSAVE >instruction? >Also I'll add that the FXSAVE doesn't seem to be finallized on Intel >until the PIII. While on the subject could someone try this on a platform >that does support FXSAVE and report? >(http://www.tbcnet.com/~clive/fxsave.zip) Most likely have to open a DOS >window first. Dave Cute program, nut you have to make stop exiting at light speed so you can see the output.... -- ----------------------------------------------------------- lsunley at mb.sympatico.ca ----------------------------------------------------------- **= Email 6 ==========================** Date: Mon, 3 Jan 2005 11:22:12 +0000 From: John Poltorak Subject: Re: SSES-alpha + Innotek libc On Sun, Jan 02, 2005 at 11:36:50AM -0800, Dave Yeo wrote: > Interesting thing, I was playing around with star-1.5 (built with Innotek libc), tried to create > an archive and star complained about not knowing who I was and couldn't find me in > passwrd and refused to run. Ran whoami (from SSES in a innotek enviroment) and it > complained , said I was L31 or somesuch. Updated SSES to sses-alpha-0.30 and > everything was happy. Whoami reported I was root and star happily (well at least until it > hit the 2GB limit) created an archive > Interesting interplay between the two systems and if anyone is using libc+sses make > sure you update to the latest I don't use SSES at the moment but it looks as though it should be a requirement for anyone setting up a Unix-like environment under OS/2, especially if it comes with PASSWD management programs which can be used by any program which requires them. > Dave > -- John **= Email 7 ==========================** Date: Mon, 3 Jan 2005 12:30:29 +0100 From: Stefan.Neis at t-online.de Subject: Re: fork() FXSAVE and Pentium Pro Hi, > >> FXSAVE was introduced with the Pentium II and a half, AMD introduced it > >> with the K7 > > >Just for completeness, what you say about AMD apparently is not quite > >true. > > and... > > the complete information is??? No idea - I don't know what CPU's actually are going to work. I'm just observing that the test program doesn't work on any of the boxes I have available, not even on the Athlon. :-( Truth might be similar to the "Pentium II and a half" above, maybe it's not working with older Athlon versions and it does work with newer ones? Regards, Stefan **= Email 8 ==========================** Date: Mon, 3 Jan 2005 12:04:11 +0000 From: John Poltorak Subject: cmovbe While trying to build POVRAY I get this error:- make all-recursive make[1]: Entering directory `G:/ux2bs/workdir/povray-3.6.1' Making all in libraries make[2]: Entering directory `G:/ux2bs/workdir/povray-3.6.1/libraries' Making all in zlib make[3]: Entering directory `G:/ux2bs/workdir/povray-3.6.1/libraries/zlib' make all-am make[4]: Entering directory `G:/ux2bs/workdir/povray-3.6.1/libraries/zlib' gcc -DHAVE_CONFIG_H -I. -I. -I. -I/usr/X11R6/include -pipe -O3 -march=i686 - mcpu=i686 -malign-double -c `test -f 'adler32.c' || echo './'`adler32.c {standard input}: Assembler messages: {standard input}:28: Error: no such 386 instruction: `cmovbe' make[4]: *** [adler32.o] Error 1 make[4]: Leaving directory `G:/ux2bs/workdir/povray-3.6.1/libraries/zlib' make[3]: *** [all] Error 2 make[3]: Leaving directory `G:/ux2bs/workdir/povray-3.6.1/libraries/zlib' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `G:/ux2bs/workdir/povray-3.6.1/libraries' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `G:/ux2bs/workdir/povray-3.6.1' make: *** [all] Error 2 Any ideas? It looks like an error building zlib. I thought zlib could be built on OS/2 without any problems... -- John