Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Andi Kleen <ak@linux.intel.com>
To:     Mike Kravetz <mike.kravetz@oracle.com>
Cc:     Waiman Long <longman@redhat.com>,
        Matthew Wilcox <willy@infradead.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        Michal Hocko <mhocko@kernel.org>
Subject: Re: [PATCH v2] hugetlbfs: Disable softIRQ when taking hugetlb_lock
References: <20191211194615.18502-1-longman@redhat.com>
        <4fbc39a9-2c9c-4c2c-2b13-a548afe6083c@oracle.com>
        <32d2d4f2-83b9-2e40-05e2-71cd07e01b80@redhat.com>
        <0fcce71f-bc20-0ea3-b075-46592c8d533d@oracle.com>
        <20191212060650.ftqq27ftutxpc5hq@linux-p48b>
Date:   Thu, 12 Dec 2019 13:32:39 -0800
In-Reply-To: <20191212060650.ftqq27ftutxpc5hq@linux-p48b> (Davidlohr Bueso's
        message of "Wed, 11 Dec 2019 22:06:50 -0800")
Message-ID: <871rt99qmg.fsf@linux.intel.com>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.2 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Davidlohr Bueso <dave@stgolabs.net> writes:
> +void free_huge_page(struct page *page)
> +{
> +	struct hugetlb_free_page_work work;
> +
> +	work.page = page;
> +	INIT_WORK_ONSTACK(&work.work, free_huge_page_workfn);
> +	queue_work(hugetlb_free_page_wq, &work.work);
> +
> +	/*
> +	 * Wait until free_huge_page is done.
> +	 */
> +	flush_work(&work.work);
> +	destroy_work_on_stack(&work.work);

Does flushing really work in softirq context?

Anyways, waiting seems inefficient over fire'n'forget

You'll need a per cpu pre allocated work item and a queue.
Then take a lock on the the queue and link the page into
it and trigger the work item if it's not already pending.

And add a in_interrupt() check of course.


-Andi