Re: [Linux kernel bug] WARNING in static_key_slow_inc_cpuslocked

[Thomas Gleixner]

On Sun, Jun 09 2024 at 10:25, Steven Rostedt wrote:
> On Sun, 09 Jun 2024 16:06:05 +0200
> Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
>> CPU 0                           	CPU 1
>> 
>> kernfs_fop_write_iter()			kernfs_fop_write_iter()
>>   set_attr_rdpmc()		  	  set_attr_rdpmc()
>>     arch_jump_label_transform_queue()       arch_jump_label_transform_queue()
>>      mutex_lock(text_mutex)                   mutex_lock(text_mutex)
>>      __jump_label_patch()
>>      text_poke_queue()
>>      mutex_unlokc(text_mutex)
>>                                               __jump_label_patch()
>> 
>> CPU 1 sees the original text and not the expected because CPU 0 did not
>> yet invoke arch_jump_label_transform_apply().
>> 
>> So clearly set_attr_rdpmc() lacks serialization, no?
>> 
> Hmm, but should jump labels fail when that happens? Or should it catch
> it, and not cause a BUG?

Well the bug is there to detect inconsistency and that clearly works :)

But I clearly can't read, because the jump label operations are
serialized via jump_label_mutex. Hrm...