Re: [PATCH RFC v2 3/7] cgroup: Add interface to allow drivers to lookup process cgroup membership

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Quoting Matt Roper (2018-02-01 19:53:11)
> Drivers will need to save/restore the appropriate cgroup data for the process
> submitting a driver request.  Add an interface for drivers to determine the
> cgroup for the userspace process submitting a driver request.
> 
> Cc: Tejun Heo <tj@xxxxxxxxxx>
> Cc: cgroups@xxxxxxxxxxxxxxx
> Signed-off-by: Matt Roper <matthew.d.roper@xxxxxxxxx>
> ---
>  include/linux/cgroup.h        |  6 ++++++
>  kernel/cgroup/cgroup_driver.c | 24 ++++++++++++++++++++++++
>  2 files changed, 30 insertions(+)
> 
> diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
> index 0ba1374122c7..05ebfb97bcde 100644
> --- a/include/linux/cgroup.h
> +++ b/include/linux/cgroup.h
> @@ -842,6 +842,7 @@ void cgroup_driver_release(struct cgroup_driver *drv);
>  struct cgroup_driver_data * cgroup_driver_get_data(struct cgroup_driver *drv,
>                                                    struct cgroup *cgrp,
>                                                    bool *is_new);
> +struct cgroup *cgroup_for_driver_process(struct pid *pid);
>  
>  #else /* !CONFIG_CGROUP_DRIVER */
>  
> @@ -857,6 +858,11 @@ cgroup_driver_get_data(struct cgroup_driver *drv,
>  {
>         return ERR_PTR(-EINVAL);
>  }
> +static inline struct cgroup *
> +cgroup_for_driver_process(struct pid *pid)
> +{
> +       return NULL;
> +}
>  
>  #endif /* !CONFIG_CGROUP_DRIVER */
>  
> diff --git a/kernel/cgroup/cgroup_driver.c b/kernel/cgroup/cgroup_driver.c
> index 0d893395dc7b..4f870cbb9212 100644
> --- a/kernel/cgroup/cgroup_driver.c
> +++ b/kernel/cgroup/cgroup_driver.c
> @@ -128,3 +128,27 @@ cgroup_driver_get_data(struct cgroup_driver *drv,
>         return data;
>  }
>  EXPORT_SYMBOL(cgroup_driver_get_data);
> +
> +/**
> + * cgroup_for_driver_process - return the cgroup for a process
> + * @pid: process to lookup cgroup for
> + *
> + * Returns the cgroup from the v2 hierarchy that a process belongs to.
> + * This function is intended to be called from drivers and will obtain
> + * the necessary cgroup locks.
> + *
> + * RETURNS:
> + * Process' cgroup in the default (v2) hierarchy
> + */
> +struct cgroup *
> +cgroup_for_driver_process(struct pid *pid)
> +{
> +       struct task_struct *task = pid_task(pid, PIDTYPE_PID);
> +
> +       mutex_lock(&cgroup_mutex);
> +       spin_lock_irq(&css_set_lock);
> +       task_cgroup_from_root(task, &cgrp_dfl_root);
> +       spin_unlock_irq(&css_set_lock);
> +       mutex_unlock(&cgroup_mutex);

Gut feeling says use task = get_pid_task(); if (!task) return NULL and
put_task_struct(task); for safety.

Shouldn't there be a return here?
> +}
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx




[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]
  Powered by Linux