Re: Panic when loading if_ath

Jille Timmermans wrote:
> Sam Leffler schreef:
>   
>> Jille Timmermans wrote:
>>     
>>> Hello list,
>>>
>>> A panic with if_ath (Atheros 2413)
>>>
>>> root /sys/modules/ath_rate_amrr# make
>>> root /sys/modules/ath_rate_amrr# make install
>>> root ~# kldload if_ath
>>> warning: KLD '/boot/kernel/if_ath.ko' is newer than the linker.hints file
>>> ath0: <Atheros 2413> mem 0xc9000000-0xc900ffff irq 18 at device 1.0 on
>>> pci5
>>> ath0: [ITHREAD]
>>>
>>> Fatal trap 12: page fault while in kernel mode
>>> cpuid = 0; apic id = 00
>>> fault virtual address = 0x0
>>> fault code = supervisor read, page not present
>>> instruction pointer = 0x20:0x0
>>> stack pointer = 0x28:0xc43cc748
>>> frame pointer = 0x28:0xc43cc75c
>>> code segment = base 0x0, limit 0xfffff, type 0x1b
>>>              = DPL 0, pres 1, def32 1, gran 1
>>> processor eflags = interrupt enabled, resume, IOPL = 0
>>> current_process = 993 (kldload)
>>> trap number = 12
>>> panic: page fault
>>> cpuid = 0
>>>
>>> [dump]
>>> [reboot]
>>> WARNING: /tmp was not properly dismounted
>>> WARNING: /usr was not properly dismounted
>>> WARNING: /var was not properly dismounted
>>>
>>> Fatal trap 12: page fault while in kernel mode
>>> cpuid = 0; apic id = 00
>>> fault virtual address = 0x5bbfe499
>>> fault code = supervisor read, page not present
>>> instruction pointer  = 0x20:0xc0684c37
>>> stack pointer = 0x28:0xc4364a5c
>>> frame pointer = 0x28:0xc4364c28
>>> code segment = base 0x0, limit 0xfffff, type 0x1b
>>>              = DPL 0, pres 1, def32 1, gran 1
>>> processor eflags = interrupt enabled, resume, IOPL = 0
>>> current process = 136 (ifconfig)
>>> trap number = 12
>>> panic: page fault
>>> cpuid = 0
>>>
>>> [a few crashes later, when it comes to mind to get a backtrace]
>>> root ~# kldload ath_rate
>>> root ~# kldload if_ath
>>> ath0: <Atheros 2413> mem 0xc9000000-0xc900ffff irq 18 at device 1.0 on
>>> pci5
>>> ath0: [ITHREAD]
>>> [panic]
>>> uart_z8530_class(c4ac5000, 0, 1, 1, c4390000, ...) at 0
>>> ar5212Attach(1a, c48b0000, 1, c4389000, c437c8ec, ...) at
>>> ar5212Attach+0x211
>>> ath_hal_attach(1a, c48b0000, 1, c4389000, c437c8ec, ...) at
>>> ath_hal_attach+0x56
>>> ath_attach(1a, c48b0000, 3, c437c940, ffffffff, ...) at ath_attach+0x9e
>>> ath_pci_attach(c46bb200, c4823854, c09fcf80, c099dc60, 80000000, ...) at
>>> ath_pci_attach+0x332
>>> device_attach([snip]) at device_attach+0x36f
>>> device_probe_and_attach([snip) at ...+0x43
>>> pci_driver_added([snip]) at +0x104
>>> devclass_add_driver([snip]) at +0xe8
>>> driver_module_handler([snip]) at +0x79
>>> module_register_init()
>>> linker_load_module()
>>> kern_kldload()
>>>
>>> (Please don't tell me you want any of that snips.)
>>>
>>> Any extra information I can provide to you ?
>>> I am willing to help debugging / crashing.
>>>
>>>   
>>>       
>> Try building the driver into the kernel.  Also use sample and not amrr.
>>     
> root ~# cd /sys/modules/ath_rate_sample
> root /sys/modules/ath_rate_sample# make
> root /sys/modules/ath_rate_sample# make install
> root /sys/modules/ath_rate_sample# kldload if_ath
> link_elf: symbol ath_hal_computetxtime undefined
> KLD if_ath.ko: depends on ath_rate - not available
> kldload: can't load if_ath: No such file or directory
> root /sys/modules/ath_rate_sample# kldload ath_rate
> link_elf: symbol ath_hal_computetxtime undefined
> kldload: can't load ath_rate: No such file or directory
>
> I'll try compiling it in.
>
>   
I know module building is broken; I posted recently to just build things 
into the kernel.  I've got a possible solution in p4 in the sam_vap 
branch if you want to try it.  I removed the ath_rate_* modules and just 
smooshed the code into the ath module so we go from 
(ath+ath_hal+ath_rate) to just (ath).

Right now I want to see if your problem is in the module mess, 
ath_rate_amrr, or something else.  I tested 2413 w/ the driver built 
into the kernel so I'm guessing it's amrr which I may just nuke entirely 
(along with onoe).

    Sam