Re: [PATCH] devcoredump : Serialize devcd_del work

[Thomas Gleixner]

On Fri, Apr 22 2022 at 15:53, Johannes Berg wrote:
> On Fri, 2022-04-22 at 15:41 +0200, Johannes Berg wrote:
>> On Tue, 2022-04-19 at 15:57 +0530, Mukesh Ojha wrote:
>> >    INIT_DELAYED_WORK
>> >    schedule_delayed_work
>> > 
>> 
>> Wouldn't it be easier to simply schedule this before adding it to sysfs
>> and sending the uevent?

Only if it's guaranteed that the timer will not expire before
add_device() succeeds. It's likely, but there is virt....

> Hm. I think that would solve this problem, but not all of the problems
> here ...
>
> Even with your change, I believe it's still racy wrt. disabled_store(),
> since that flushes the work but devcd_data_write() remains reachable
> (and might in fact be waiting for the mutex after your change), so I
> think we need an additional flag somewhere (in addition to the mutex) to
> serialize all of these things against each other.

Plus there is disabled_store() which iterates over the devices and
reschedules the work. How is that supposed to be protected against a
concurrent devcd_del()?

Why needs disabled_store() to do that at all?

Thanks,

        tglx