OK, I've played around with my kernel trying to figure out what causes this. The one thing I know is that it disappears if I leave pcm out of my kernel, but I don't think pcm is the culprit. Since the address at fault is constant, I've added code to mtrash_[cd]tor() which calls db_print_backtrace() every time we hit that address. I also added code to module_regiser_init() to identify the module being processed:
> module_register_init: registering 'isa/esscontrol' > mtrash_ctor(0xfffffc00007aab80, 64, 0) > mtrash_ctor(0xfffffc00007a8ba0, 32, 0) > mtrash_ctor(0xfffffc00007b6000, 8192, 0) > db_print_backtrace() at db_print_backtrace+0x18 > mtrash_ctor() at mtrash_ctor+0x68 > uma_zalloc_arg() at uma_zalloc_arg+0x160 > malloc() at malloc+0x94 > kobj_class_compile() at kobj_class_compile+0x2c > devclass_add_driver() at devclass_add_driver+0x64 > driver_module_handler() at driver_module_handler+0xb4 > module_register_init() at module_register_init+0xc0 > mi_startup() at mi_startup+0x144 > locorestart() at locorestart+0x64 > --- root of call graph --- > Memory modified after free 0xfffffc00007b6000(8184) > panic: Most recently used by none > > panic > Stopped at Debugger+0x38: zapnot v0,#0xf,v0 <v0=0x6> The problem is that this doesn't really tell me what I want (i.e. who held that block before it was allocated to isa/esscontrol). All I know about that is: mtrash_dtor(0xfffffc00007b6000, 8192, 0) because it seems that db_print_backtrace() is a nop at that point. Any suggestions as to how I can figure out who used that block of memory before it was allocated to the ess driver? DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
