Date: Sat, 24 May 1997 13:07:07 +0900 From: KATO Takenori <kato@eclogite.eps.nagoya-u.ac.jp> To: smurfen@ludd.luth.se Cc: freebsd-bugs@hub.freebsd.org, kato@eclogite.eps.nagoya-u.ac.jp Subject: Re: kern/3611: Internal CPU cache on CyrixiInstead DX2 does not work Message-ID: <199705240407.NAA06741@gneiss.eps.nagoya-u.ac.jp> In-Reply-To: Your message of "Fri, 23 May 1997 15:21:12 %2B0200 (MET DST)" References: <Pine.SUN.3.95.970523145947.9036A-100000@brother.ludd.luth.se>
next in thread | previous in thread | raw e-mail | index | archive | help
From: Ola Persson <smurfen@ludd.luth.se> Subject: Re: kern/3611: Internal CPU cache on CyrixiInstead DX2 does not work Date: Fri, 23 May 1997 15:21:12 +0200 (MET DST) > Ok, I send it to you only, if it should be sent to -bugs, just fwd it. I Cc to freebsd-bugs. > With Internal cache set to write-back (works) > > CCR1=90 > CCR2=0 > CCR3=0CR0=8005003b > > With internal cache set to write-thru (makes system crash) > > CCR1=90 > CCR2=16 > CCR3=0CR0=a005003b Firstly, I think former is write-through case, because latter one shows that (1) NW (not write-through) is set and (2) WB interface pin is enabled. If former is write-through, CCR1, CCR2and CR0 seem to be normal. > I tried to compile a program just to see if the error was still there and > it immediately put me in DDB. Maybe it is after such a panic you would > like me to run show cyrixreg ? Hmm I just thought about that now... I'll > Postpone this mail and do it right away.... > > First error while compiling didn't put me in ddb. > Operands given don't match any known 386 instruction > > continuing..... > > Second error while compiling the same program as above > (../../kern/tty_pty.c) > cc: Internal compiler error: Program cc1 got fatal signal 11 > and it dumped core. > > continuing.... > > Third error while compiling ../../kern/vfs_init.c > > Operands given don't match any known 386 instruction > > Damn computer refuses to panic just cuz I want it to. Anyway I guess this > proves that there is an error. When I get it to dump into ddb I'll write > down the result of a show cyrixreg. There are two reason for strange signal 11 on Cyrix CPU: (1) Motherboard and memory problem. (c.f. http://www.bitwizard.nl/sig11/) (2) CPU bug. Page fault address in CR2 register is sometimes (i.e. not always) collapsed by interruption and trap_pfault cannot get correct address (CR2 is zeroed by interruption). This can be fixed by converting of exception 14 handler from task gate to interrupt gate. Fix is only in PC-98 tree now. ---- KATO Takenori <kato@eclogite.eps.nagoya-u.ac.jp> Dept. Earth Planet. Sci., Nagoya Univ., Nagoya, 464-01, Japan PGP public key: finger kato@eclogite.eps.nagoya-u.ac.jp ------------------- Powered by FreeBSD(98) -------------------
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199705240407.NAA06741>