trap SIGBUS in aiowrite

wangzhonnew

signal: 10 -- SIGBUS
stack trace:
<StackTrace>
##### Object: /usr/lib/sparcv9/libaio.so.1
assfail(0x10000,0xffffffff72702000,0x0,0xffffffff77e08000,0x0,0xffffffff7b1e975 9) + 0x258
assfail(0x10000,0x0,0x3,0xffffffff77d02650,0x3,0xffffffff7fffd66c) + 0x170
SUNW_1.1(0xffffffff77d02650,0x0,0xffffffff7b1e9759,0xffffffff77d008b8,0x3,0x0) + 0x2544
aiowrite(0x1,0xffffffff77e08d10,0x2000,0x2e0000,0x0,0xffffffff73019008) + 0xe4
aiowrite(0x3,0x224348000,0x2000,0x2e0000,0x0,0xffffffff73019008) + 0x20

problem cannot be reproduced, from the bottom aiowrite, all parameters passed in are making sense to me... but i\'m not sure why it leads to assfail, could any of you give some idea what is the prototype of assfail? i couldn\'t find one with 6 parameters (the one i can find is with 3 parameters), and what can lead to assfail with signal SIGBUS

I\'m also curious why aiowrite looks recursivly call himself, i looked OpenSolaris and didn\'t find such behavior, is it true that this part has been changed in opensolaris or something wrong happened in my trap tool? or is it caused by 2 aio server running simutanously....

Operating system is: Solaris 64BIT
OSS reports operating system level as: Generic_118558-39.5.9.

Thanks!!

wangzhonnew

我在想是不是有可能是由于swap空间不够，造成内存的alignment出现错误？但是这还是很难解释……有没有谁能帮一下忙？