Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 5 Mar 2007 18:28:30 +0000 (GMT)
From:      Gavin Atkinson <gavin.atkinson@ury.york.ac.uk>
To:        Kostik Belousov <kostikbel@gmail.com>
Cc:        Peter Losher <plosher@plosh.net>, freebsd-stable@freebsd.org
Subject:   Re: 'panic: bad pte' error on 6.2-RELEASE (amd64)
Message-ID:  <20070305182522.P21692@ury.york.ac.uk>
In-Reply-To: <20070305100237.GI10453@deviant.kiev.zoral.com.ua>
References:  <45EBCF4C.9000901@plosh.net> <20070305100237.GI10453@deviant.kiev.zoral.com.ua>

next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 5 Mar 2007, Kostik Belousov wrote:

> On Mon, Mar 05, 2007 at 12:05:32AM -0800, Peter Losher wrote:
>> We recently updated one of our dual Opteron systems (w/ 4GB RAM) from
>> 5.5 to 6.2 (amd64 wipe and reinstalled) and about once a week, it panics
>> with the below message:
>>
>> -=-
>> TPTE at 0xffff8000040028a0  IS ZERO @ VA 800514000
>> panic: bad pte
>> cpuid = 2
>> KDB: stack backtrace:
>> panic() at 0xffffffff803fdd03 = panic+0x253
>> pmap_remove_pages() at 0xffffffff806072a3 = pmap_remove_pages+0x283
>> exec_new_vmspace() at 0xffffffff803e18e6 = exec_new_vmspace+0x216
>> exec_elf64_imgact() at 0xffffffff803cbb73 = exec_elf64_imgact+0x273
>> kern_execve() at 0xffffffff803e2107 = kern_execve+0x457
>> execve() at 0xffffffff803e2bed = execve+0x5d
>> syscall() at 0xffffffff8060d141 = syscall+0x4d1
>> Xfast_syscall() at 0xffffffff805f8128 = Xfast_syscall+0xa8
>> --- syscall (59, FreeBSD ELF64, execve), rip = 0x80069838c, rsp =
>> 0x7fffffffe7c8, rbp = 0x7fffffffecd0 ---
>> Uptime: 4d16h55m46s
>> -=-
>>
>> I do have a dump, and can make that available if need be.  Has anyone
>> encountered this recently and can shed any light on what might be
>> causing this?
>
> Did rev. 1.516.2.9 of sys/amd64/amd64/pmap.c changed (or even fixed) the
> problem ? (This is the same patch I already sent you).

Do you know if this patch is likely to fix the other "bad pte" panics that 
have been seen (usually during process exit())?  e.g.

http://docs.freebsd.org/cgi/mid.cgi?1148482556.35287.18.camel
http://lists.freebsd.org/pipermail/freebsd-current/2004-August/034909.html

I have a couple of amd64 machines I'm reluctant to make live until this is 
sorted...  I'll cvsup in the next day or two to see, although it wasn't 
consistantly reproduceable.

Thanks,

Gavin



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070305182522.P21692>