Re: [PATCH v3 31/38] x86/resctrl: resctrl_exit() teardown resctrl but leave the mount point
From: James Morse
Date: Thu Jul 04 2024 - 12:41:52 EST
Hi Reinette,
On 28/06/2024 17:53, Reinette Chatre wrote:
> On 6/14/24 8:00 AM, James Morse wrote:
>> resctrl_exit() was intended for use when the 'resctrl' module was unloaded.
>> resctrl can't be built as a module, and the kernfs helpers are not exported
>> so this is unlikely to change. MPAM has an error interrupt which indicates
>> the MPAM driver has gone haywire. Should this occur tasks could run with
>> the wrong control values, leading to bad performance for impoartant tasks.
>
> impoartant -> important
>
>> The MPAM driver needs a way to tell resctrl that no further configuration
>> should be attempted.
>>
>> Using resctrl_exit() for this leaves the system in a funny state as
>> resctrl is still mounted, but cannot be un-mounted because the sysfs
>> directory that is typically used has been removed. Dave Martin suggests
>> this may cause systemd trouble in the future as not all filesystems
>> can be unmounted.
>>
>> Add calls to remove all the files and directories in resctrl, and
>> remove the sysfs_remove_mount_point() call that leaves the system
>> in a funny state. When triggered, this causes all the resctrl files
>> to disappear. resctrl can be unmounted, but not mounted again.
> I am not familiar with these flows so I would like to confirm ...
> In this scenario the resctrl filesystem will be unregistered, are
> you saying that it is possible to unmount a filesystem after it has
> been unregistered?
Counter-intuitively: yes.
The rules are described in fs/filesystems.c: We can access the members of the struct
file_system_type if the list lock is held, or a reference is held to the module. This is
how /proc/mounts is able to print the filesystem name from struct file_system_type without
taking the lock - it holds a reference to any module to prevent the structure from being
freed. Because resctrl can't be built as a module, we can say there is always a reference
held, and we can never free struct file_system_type.
Thanks,
James