Re: Fwd: w_scan hangs on 6.3.7 and does not react on kill -9

[Bagas Sanjaya]

On 6/12/23 20:52, Bagas Sanjaya wrote:
> Hi,
> 
> I notice a regression report on Bugzilla [1]. Quoting from it:
> 
>> w_scan tool on kernel 6.3.7 hangs so badly that it cannot be killed by kill -9 
>> I tried also w_scan_cpp 20230604 but it also fails I have 2040:8268 Hauppauge soloHD device as reported by lsusb. During reboot of the OS it prints a lot of kernel errors but it is after syslog is killed I guess and in syslog messages I do not see anything. On default Debian kernel 5.10.0-23 this problem does not exists.
> 
> 
> See Bugzilla for the full thread and attached dmesg and kernel config.
> 
> Lukasz: On what hardware you have this regression? Also, it is really
> helpful if you can perform bisection (as outlined
> in Documentation/admin-guide/bug-bisect.html) to find the culprit,
> when developers can't figure it out by inspecting the code alone.
> Last but not least, please also try latest mainline (currently v6.4-rc6).
> 
> Anyway, I'm adding it to regzbot:
> 
> #regzbot introduced: v5.10..v6.3 https://bugzilla.kernel.org/show_bug.cgi?id=217540
> #regzbot title: w_scan zombie (unkillable) on kernel v6.3
> 

Another reporter on Bugzilla with similar regression as this one had
bisected the culprit, so:

Hyunwoo Kim: It looks like this regression is caused by a backported
commit of yours. Would you like to take a look on it?

#regzbot introduced: 8994830135b38b

Thanks.

-- 
An old man doll... just what I always wanted! - Clara