Re: linux-next: Tree for Apr 10 (lib/test_printf.ko)

[Randy Dunlap]

On 4/15/20 11:55 AM, Steven Rostedt wrote:
> On Tue, Apr 14, 2020 at 12:26:29PM -0700, Brendan Higgins wrote:
>> On Sat, Apr 11, 2020 at 11:22 PM Randy Dunlap <rdunlap@xxxxxxxxxxxxx> wrote:
>>>
>>> On 4/11/20 10:36 PM, Naresh Kamboju wrote:
>>>> FYI,
>>>>
>>>> Is this problem related to,
>>>>
>>>> Regression reported on Linux next 5.6.0-rc4-next-20200305 on x86_64,
>>>> i386, arm and arm64. The steps to reproduce is running kselftests lib
>>>> printf.sh test case.
>>>> Which is doing modprobe operations.
>>>>
>>>> BUG: kernel NULL pointer dereference, address: 00 - ida_free+0x76/0x140
>>>>
>>>> https://lore.kernel.org/linux-kselftest/CAFd5g46Bwd8HS9-xjHLh_rB59Nfw8iAnM6aFe0QPcveewDUT6g@xxxxxxxxxxxxxx/T/
>>>>
>>>
>>> Looks similar. Lots of fwnode, software_node, ida stuff there.
>>
>> Sorry for the late reply, I was out.
>>
>> Yeah, I am pretty sure it is the same. Heikki proposed a fix that I am
>> going to try.
> 
> My test suite just tripped over this bug. Is this the patch that you think
> fixes it?
> 
> https://lore.kernel.org/linux-kselftest/20200414081513.GD2828150@xxxxxxxxxxxxxxxxx

Yes, it is, but GregKH and Rafael Wysocki don't seem to like that patch
and are suggesting some changes in lib/kobject.c (only pseudocode,
no patch yet).

> I'll add it to see if I can continue my testing.

See the thread
[PATCH v1] kobject: make sure parent is not released before children

Here is Rafael's suggestion:
https://lore.kernel.org/linux-kselftest/CAJZ5v0hNemTDVa_S-FfVMbrKjM-RWYoHh88asnUvTNxZinY2cw@xxxxxxxxxxxxxx/

-- 
~Randy