Re: Kernel hangs in cifs_reconnect

[Paulo Alcantara <pc@cjr.nz>]

Aurélien Aptel <aaptel@suse.com> writes:

> Martijn de Gouw <martijn.de.gouw@prodrive-technologies.com> writes:
>> Our Linux VMs reports call traces about processes being stuck.
>> I've attached the full dmesg of one of the call traces below.
>>
>> The machine is running kernel 5.3.1 SMP. All mounts are mounted via the
>> dfs shares on our domain controller and have the following options in fstab:
>> nohandlecache,multiuser,sec=krb5,noperm,user=xxxx,file_mode=0600,dir_mode=0700,vers=3.0
>
> It looks like its DFS related. The DFS cache code takes the reconnect
> mutex and crashes with no chance to give back the mutex, making all
> other process hang while waiting for it.

Yeah, makes sense.

Martijn,

Could you please provide us with some debug logs with the following:

	# echo 'module cifs +p' > /sys/kernel/debug/dynamic_debug/control
	# echo 'file fs/cifs/* +p' > /sys/kernel/debug/dynamic_debug/control
	# echo 1 > /proc/fs/cifs/cifsFYI
	# echo 1 > /sys/module/dns_resolver/parameters/debug

Besides, if you could also enable KASAN, that would be great.

Thanks,
Paulo