Your fix for most_usb.c::hdm_disconnect

Your fix for most_usb.c::hdm_disconnect

[Oliver Neukum]

Hi,

I am going through syzbot's list of open issues.
It look like like the issue you posted a patch for

https://lore.kernel.org/all/20240730035745.110637-1-cam.alvarez.i@gmail.com/T/

is still open. Can something be done? Has the patch been lost?
Is there a remaining issue?

	Regards
		Oliver

Re: Your fix for most_usb.c::hdm_disconnect

[Greg Kroah-Hartman]

On Wed, Jan 22, 2025 at 02:44:59PM +0100, Oliver Neukum wrote:
> Hi,
> 
> I am going through syzbot's list of open issues.
> It look like like the issue you posted a patch for
> 
> https://lore.kernel.org/all/20240730035745.110637-1-cam.alvarez.i@gmail.com/T/
> 
> is still open. Can something be done? Has the patch been lost?
> Is there a remaining issue?

The subject line was totally wrong which is why it was probably ignored.

Also, are you sure it is right?  Grabbing a reference in a disconnect
function feels very wrong...

thanks,

greg k-h

Re: Your fix for most_usb.c::hdm_disconnect

[Camila Alvarez Inostroza]

On Wed, 22 Jan 2025, Greg Kroah-Hartman wrote:

> On Wed, Jan 22, 2025 at 02:44:59PM +0100, Oliver Neukum wrote:
>> Hi,
>>
>> I am going through syzbot's list of open issues.
>> It look like like the issue you posted a patch for
>>
>> https://lore.kernel.org/all/20240730035745.110637-1-cam.alvarez.i@gmail.com/T/
>>
>> is still open. Can something be done? Has the patch been lost?
>> Is there a remaining issue?
>
> The subject line was totally wrong which is why it was probably ignored.
>
> Also, are you sure it is right?  Grabbing a reference in a disconnect
> function feels very wrong...
>
Yes, I think so as well. I don't think my patch solves the issue at all.

thanks,
Camila
> thanks,
>
> greg k-h
>

Re: Your fix for most_usb.c::hdm_disconnect

[Oliver Neukum]

On 22.01.25 14:59, Camila Alvarez Inostroza wrote:
> 
> 
> On Wed, 22 Jan 2025, Greg Kroah-Hartman wrote:
> 
>> On Wed, Jan 22, 2025 at 02:44:59PM +0100, Oliver Neukum wrote:
>>> Hi,
>>>
>>> I am going through syzbot's list of open issues.
>>> It look like like the issue you posted a patch for
>>>
>>> https://lore.kernel.org/all/20240730035745.110637-1-cam.alvarez.i@gmail.com/T/
>>>
>>> is still open. Can something be done? Has the patch been lost?
>>> Is there a remaining issue?
>>
>> The subject line was totally wrong which is why it was probably ignored.
>>
>> Also, are you sure it is right?  Grabbing a reference in a disconnect
>> function feels very wrong...
>>
> Yes, I think so as well. I don't think my patch solves the issue at all.

Hi Camila,

thank you for the clarification. Frankly I don't understand the refcounting
logic in this drive. Let's add the author.

Hi Christian,

there is an open issue with the musb driver. May I ask with respect to disconect
handling where the get_device calls, which musb_disconnect() balances with put_device()
are?

	Regards
		Oliver