Re: [PATCH v17 03/15] mm/damon: Implement region based sampling

[SeongJae Park]

On Mon, 6 Jul 2020 13:53:10 +0200 SeongJae Park <sjp...@amazon.com> wrote:

> From: SeongJae Park <sjp...@amazon.de>
> 
> This commit implements DAMON's target address space independent high
> level logics for basic access check and region based sampling.  This
> doesn't work alone, but needs the target address space specific low
> level pritimives implementation for the monitoring target address ranges
> construction and the access check, though.  A reference implementation
> of those will be provided by a later commit.  Nevertheless, users can
> implement and use their own versions for their specific use cases.
>
[...]
> +/**
> + * damon_start() - Starts monitoring with given context.
> + * @ctx:     monitoring context
> + *
> + * Return: 0 on success, negative error code otherwise.
> + */
> +int damon_start(struct damon_ctx *ctx)
> +{
> +     int err = -EBUSY;
> +
> +     mutex_lock(&ctx->kdamond_lock);
> +     if (!ctx->kdamond) {
> +             err = 0;
> +             ctx->kdamond_stop = false;
> +             ctx->kdamond = kthread_run(kdamond_fn, ctx, "kdamond");

Oops, this means 'kdamond_fn' could see the unset '->kdamond'.  I will use
'kthread_create()' and 'wake_up_process()' in the next spin.

> +             if (IS_ERR(ctx->kdamond))
> +                     err = PTR_ERR(ctx->kdamond);
> +     }
> +     mutex_unlock(&ctx->kdamond_lock);
> +
> +     return err;
> +}

So, the change would be something like below:

--- a/mm/damon.c
+++ b/mm/damon.c
@@ -1464,9 +1464,11 @@ int damon_start(struct damon_ctx *ctx)
        if (!ctx->kdamond) {
                err = 0;
                ctx->kdamond_stop = false;
-               ctx->kdamond = kthread_run(kdamond_fn, ctx, "kdamond");
+               ctx->kdamond = kthread_create(kdamond_fn, ctx, "kdamond");
                if (IS_ERR(ctx->kdamond))
                        err = PTR_ERR(ctx->kdamond);
+               else
+                       wake_up_process(ctx->kdamond);
        }
        mutex_unlock(&ctx->kdamond_lock);


Thanks,
SeongJae Park