Date: Wed, 05 Apr 2006 20:06:08 +0900 From: Ganbold <ganbold@micom.mng.net> To: Peter Jeremy <peterjeremy@optushome.com.au> Cc: freebsd-current@freebsd.org Subject: Re: CPU class not configured problem in CURRENT-solved, boot problem remains Message-ID: <4433A4A0.5060300@micom.mng.net> In-Reply-To: <20060405102528.GE699@turion.vk2pj.dyndns.org> References: <4431D4E5.3080507@micom.mng.net> <20060404090837.GC683@turion.vk2pj.dyndns.org> <44323B37.4040605@micom.mng.net> <20060404101223.GH683@turion.vk2pj.dyndns.org> <44336331.40909@micom.mng.net> <20060405075841.GB699@turion.vk2pj.dyndns.org> <443381F9.3010504@micom.mng.net> <20060405090326.GD699@turion.vk2pj.dyndns.org> <44338D08.3070000@micom.mng.net> <20060405102528.GE699@turion.vk2pj.dyndns.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Peter Jeremy wrote: > On Wed, 2006-Apr-05 18:25:28 +0900, Ganbold wrote: > >> db> x/lu cpu >> cpu: 16 >> > > That's correct for CPU_P4 but my reading of the code suggests that it > should be initialised by identify_cpu to CPU_686 (7). > > >> db> x/lu cpu_class >> cpu_class: 4 >> > > And that's correct for CPUCLASS_686. > > It's beginning to look like the "switch (cpu_class)" statement doesn't > have the I686_CPU code compiled in. Do you want to double check that > your kernel config really does have "cpu I686_CPU" and that you are > running a kernel compiled from that config. > Sorry for the noise, it was my big fault, I was using I486_CPU instead of I686_CPU in the config and never thought this problem might happen because of incorrect CPU definition in the config. Thanks a lot Peter, for pointing me this out. Only one problem still remains, it still can not boot without pressing "Enter" key at the boot prompt (- sign waits forever). I created /boot.config with -Dh, now it shows boot.config: -Dh and waits for user to press "Enter". FYI, it boots fine with Ubuntu Linux. Please advise me, where should I look in the code for solution? Is it under /usr/src/sys/boot/ or somewhere else? thanks a lot, Ganbold > If your i386 assembler is up to it, you might like to disassemble > printcpuinfo() using gdb and confirm that the above switch statement > looks correct. Unfortunately, it's a massive function so finding the > relevant code is not trivial - though the extensive use of global > variables makes it a lot easier than it could be. > > The relevant code in a -current kernel that I built last night looks > like the following. This covers the switch statement and the printf() > statements that immediately preceded and follow it. This was built > using CPUTYPE=athlon-xp COPTFLAGS="-O -pipe" - the relative offsets > may be different if you use different flags. If you find the address > of cpu_class (0xc0799440 in the following), there are only a couple of > references to it and only one is followed by those cmp statements. > > The critical code is the 'cmp $0x4,%eax' - this is the > 'case CPUCLASS_686:' test. > > 0xc0673400 <printcpuinfo+4432>: movl $0xc0799460,0x4(%esp) > 0xc0673408 <printcpuinfo+4440>: movl $0xc06e3416,(%esp) > 0xc067340f <printcpuinfo+4447>: call 0xc05243e0 <printf> > 0xc0673414 <printcpuinfo+4452>: mov 0xc0799440,%eax > 0xc0673419 <printcpuinfo+4457>: cmp $0x1,%eax > 0xc067341c <printcpuinfo+4460>: je 0xc067344b <printcpuinfo+4507> > 0xc067341e <printcpuinfo+4462>: cmp $0x1,%eax > 0xc0673421 <printcpuinfo+4465>: jg 0xc0673430 <printcpuinfo+4480> > 0xc0673423 <printcpuinfo+4467>: test %eax,%eax > 0xc0673425 <printcpuinfo+4469>: je 0xc067343a <printcpuinfo+4490> > 0xc0673427 <printcpuinfo+4471>: jmp 0xc0673526 <printcpuinfo+4726> > 0xc067342c <printcpuinfo+4476>: lea 0x0(%esi),%esi > 0xc0673430 <printcpuinfo+4480>: cmp $0x4,%eax > 0xc0673433 <printcpuinfo+4483>: je 0xc067345c <printcpuinfo+4524> > 0xc0673435 <printcpuinfo+4485>: jmp 0xc0673526 <printcpuinfo+4726> > 0xc067343a <printcpuinfo+4490>: movl $0xc06cdbea,(%esp) > 0xc0673441 <printcpuinfo+4497>: call 0xc05243e0 <printf> > 0xc0673446 <printcpuinfo+4502>: jmp 0xc0673532 <printcpuinfo+4738> > 0xc067344b <printcpuinfo+4507>: movl $0xc06e2d76,(%esp) > 0xc0673452 <printcpuinfo+4514>: call 0xc05243e0 <printf> > 0xc0673457 <printcpuinfo+4519>: jmp 0xc0673532 <printcpuinfo+4738> > 0xc067345c <printcpuinfo+4524>: mov 0xc079b960,%ebx > 0xc0673462 <printcpuinfo+4530>: mov 0xc079b964,%esi > 0xc0673468 <printcpuinfo+4536>: mov %ebx,%eax > 0xc067346a <printcpuinfo+4538>: mov %esi,%edx > 0xc067346c <printcpuinfo+4540>: add $0x1388,%eax > 0xc0673471 <printcpuinfo+4545>: adc $0x0,%edx > 0xc0673474 <printcpuinfo+4548>: movl $0xf4240,0x8(%esp) > 0xc067347c <printcpuinfo+4556>: movl $0x0,0xc(%esp) > 0xc0673484 <printcpuinfo+4564>: mov %eax,(%esp) > 0xc0673487 <printcpuinfo+4567>: mov %edx,0x4(%esp) > 0xc067348b <printcpuinfo+4571>: call 0xc068bca0 <__udivdi3> > 0xc0673490 <printcpuinfo+4576>: mov %eax,0xc07994e0 > 0xc0673495 <printcpuinfo+4581>: add $0x1387,%ebx > 0xc067349b <printcpuinfo+4587>: adc $0x0,%esi > 0xc067349e <printcpuinfo+4590>: movl $0x2710,0x8(%esp) > 0xc06734a6 <printcpuinfo+4598>: movl $0x0,0xc(%esp) > 0xc06734ae <printcpuinfo+4606>: mov %ebx,(%esp) > 0xc06734b1 <printcpuinfo+4609>: mov %esi,0x4(%esp) > 0xc06734b5 <printcpuinfo+4613>: call 0xc068bca0 <__udivdi3> > 0xc06734ba <printcpuinfo+4618>: mov %eax,%ecx > 0xc06734bc <printcpuinfo+4620>: mov $0x51eb851f,%edx > 0xc06734c1 <printcpuinfo+4625>: mov %ecx,%eax > 0xc06734c3 <printcpuinfo+4627>: mul %edx > 0xc06734c5 <printcpuinfo+4629>: shr $0x5,%edx > 0xc06734c8 <printcpuinfo+4632>: imul $0x64,%edx,%edx > 0xc06734cb <printcpuinfo+4635>: mov %ecx,%eax > 0xc06734cd <printcpuinfo+4637>: sub %edx,%eax > 0xc06734cf <printcpuinfo+4639>: mov %eax,0xc(%esp) > 0xc06734d3 <printcpuinfo+4643>: mov 0x8(%esp),%edx > 0xc06734d7 <printcpuinfo+4647>: mov 0xc(%esp),%ecx > 0xc06734db <printcpuinfo+4651>: mov %edx,0xffffffdc(%ebp) > 0xc06734de <printcpuinfo+4654>: mov %ecx,0xffffffe0(%ebp) > 0xc06734e1 <printcpuinfo+4657>: movl $0xf4240,0x8(%esp) > 0xc06734e9 <printcpuinfo+4665>: movl $0x0,0xc(%esp) > 0xc06734f1 <printcpuinfo+4673>: mov %ebx,(%esp) > 0xc06734f4 <printcpuinfo+4676>: mov %esi,0x4(%esp) > 0xc06734f8 <printcpuinfo+4680>: call 0xc068b610 <__divdi3> > 0xc06734fd <printcpuinfo+4685>: mov 0xffffffe0(%ebp),%ebx > 0xc0673500 <printcpuinfo+4688>: mov %ebx,0xc(%esp) > 0xc0673504 <printcpuinfo+4692>: mov %eax,0x4(%esp) > 0xc0673508 <printcpuinfo+4696>: mov %edx,0x8(%esp) > 0xc067350c <printcpuinfo+4700>: movl $0xc06e341b,(%esp) > 0xc0673513 <printcpuinfo+4707>: call 0xc05243e0 <printf> > 0xc0673518 <printcpuinfo+4712>: movl $0xc06e3171,(%esp) > 0xc067351f <printcpuinfo+4719>: call 0xc05243e0 <printf> > 0xc0673524 <printcpuinfo+4724>: jmp 0xc0673532 <printcpuinfo+4738> > 0xc0673526 <printcpuinfo+4726>: movl $0xc06e33dc,(%esp) > 0xc067352d <printcpuinfo+4733>: call 0xc05243e0 <printf> > 0xc0673532 <printcpuinfo+4738>: movl $0xc06e3429,(%esp) > 0xc0673539 <printcpuinfo+4745>: call 0xc05243e0 <printf> > > > >> db> >> >> Ganbold >> >> >> _______________________________________________ >> freebsd-current@freebsd.org mailing list >> http://lists.freebsd.org/mailman/listinfo/freebsd-current >> To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org" >> > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4433A4A0.5060300>