Re: [PATCH] rtc: ac100: Fix multiple race conditions

[Quentin Schulz]

Hi Alexandre,

On 04/12/2017 15:05, Alexandre Belloni wrote:
> The probe function is not allowed to fail after registering the RTC because
> the following may happen:
> 
> CPU0:                                CPU1:
> sys_load_module()
>  do_init_module()
>   do_one_initcall()
>    cmos_do_probe()
>     rtc_device_register()
>      __register_chrdev()
>      cdev->owner = struct module*
>                                      open("/dev/rtc0")
>     rtc_device_unregister()
>   module_put()
>   free_module()
>    module_free(mod->module_core)
>    /* struct module *module is now
>       freed */
>                                       chrdev_open()
>                                        spin_lock(cdev_lock)
>                                        cdev_get()
>                                         try_module_get()
>                                          module_is_live()
>                                          /* dereferences already
>                                             freed struct module* */
> 
> Also, the interrupt handler: ac100_rtc_irq() is dereferencing chip->rtc but
> this may still be NULL when it is called, resulting in:
> Unable to handle kernel NULL pointer dereference at virtual address 00000194
> pgd = (ptrval)
> [00000194] *pgd=00000000
> Internal error: Oops: 5 [#1] SMP ARM
> Modules linked in:
> CPU: 0 PID: 72 Comm: irq/71-ac100-rt Not tainted 4.15.0-rc1-next-20171201-dirty #120
> Hardware name: Allwinner sun8i Family
> task: (ptrval) task.stack: (ptrval)
> PC is at mutex_lock+0x14/0x3c
> LR is at ac100_rtc_irq+0x38/0xc8
> pc : [<c06543a4>]    lr : [<c04d9a2c>]    psr: 60000053
> sp : ee9c9f28  ip : 00000000  fp : ee9adfdc
> r10: 00000000  r9 : c0a04c48  r8 : c015ed18
> r7 : ee9bd600  r6 : ee9c9f28  r5 : ee9af590  r4 : c0a04c48
> r3 : ef3cb3c0  r2 : 00000000  r1 : ee9af590  r0 : 00000194
> Flags: nZCv  IRQs on  FIQs off  Mode SVC_32  ISA ARM  Segment none
> Control: 10c5387d  Table: 4000406a  DAC: 00000051
> Process irq/71-ac100-rt (pid: 72, stack limit = 0x(ptrval))
> Stack: (0xee9c9f28 to 0xee9ca000)
> 9f20:                   00000000 7c2fd1be c015ed18 ee9adf40 ee9c0400 ee9c0400
> 9f40: ee9adf40 c015ed34 ee9c8000 ee9adf64 ee9c0400 c015f040 ee9adf80 00000000
> 9f60: c015ee24 7c2fd1be ee9adfc0 ee9adf80 00000000 ee9c8000 ee9adf40 c015eef4
> 9f80: ef1eba34 c0138f14 ee9c8000 ee9adf80 c0138df4 00000000 00000000 00000000
> 9fa0: 00000000 00000000 00000000 c01010e8 00000000 00000000 00000000 00000000
> 9fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> 9fe0: 00000000 00000000 00000000 00000000 00000013 00000000 ffffffff ffffffff
> [<c06543a4>] (mutex_lock) from [<c04d9a2c>] (ac100_rtc_irq+0x38/0xc8)
> [<c04d9a2c>] (ac100_rtc_irq) from [<c015ed34>] (irq_thread_fn+0x1c/0x54)
> [<c015ed34>] (irq_thread_fn) from [<c015f040>] (irq_thread+0x14c/0x214)
> [<c015f040>] (irq_thread) from [<c0138f14>] (kthread+0x120/0x150)
> [<c0138f14>] (kthread) from [<c01010e8>] (ret_from_fork+0x14/0x2c)
> 
> Solve both issues by moving to
> devm_rtc_allocate_device()/rtc_register_device()
> 
> Signed-off-by: Alexandre Belloni <alexandre.belloni@xxxxxxxxxxxxxxxxxx>

Tested-by: Quentin Schulz <quentin.schulz@xxxxxxxxxxxxxxxxxx>

Thanks,
Quentin
-- 
Quentin Schulz, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com