From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [PATCH v2] ksm: React on changing "sleep_millisecs" parameter faster References: <20181211100346.GE2342@uranus.lan> <154452399396.4921.3418365102240528000.stgit@localhost.localdomain> <20181211111300.GF2342@uranus.lan> From: Kirill Tkhai Message-ID: <26fb047f-9164-96ae-a6cd-7a7efa41d43b@virtuozzo.com> Date: Tue, 11 Dec 2018 15:22:42 +0300 MIME-Version: 1.0 In-Reply-To: <20181211111300.GF2342@uranus.lan> Content-Type: text/plain; charset=windows-1252 Content-Language: en-US Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org To: Cyrill Gorcunov Cc: akpm@linux-foundation.org, mhocko@suse.com, linux-mm@kvack.org, linux-kernel@vger.kernel.org, gorcunov@virtuozzo.com List-ID: On 11.12.2018 14:13, Cyrill Gorcunov wrote: > On Tue, Dec 11, 2018 at 01:26:59PM +0300, Kirill Tkhai wrote: >> ksm thread unconditionally sleeps in ksm_scan_thread() >> after each iteration: >> >> schedule_timeout_interruptible( >> msecs_to_jiffies(ksm_thread_sleep_millisecs)) >> >> The timeout is configured in /sys/kernel/mm/ksm/sleep_millisecs. >> >> In case of user writes a big value by a mistake, and the thread >> enters into schedule_timeout_interruptible(), it's not possible >> to cancel the sleep by writing a new smaler value; the thread >> is just sleeping till timeout expires. >> >> The patch fixes the problem by waking the thread each time >> after the value is updated. >> >> This also may be useful for debug purposes; and also for userspace >> daemons, which change sleep_millisecs value in dependence of >> system load. >> >> Signed-off-by: Kirill Tkhai >> >> v2: Use wait_event_interruptible_timeout() instead of unconditional >> schedule_timeout(). > ... >> @@ -2844,7 +2849,10 @@ static ssize_t sleep_millisecs_store(struct kobject *kobj, >> if (err || msecs > UINT_MAX) >> return -EINVAL; >> >> + mutex_lock(&ksm_thread_mutex); >> ksm_thread_sleep_millisecs = msecs; >> + mutex_unlock(&ksm_thread_mutex); >> + wake_up_interruptible(&ksm_iter_wait); > > Btw, just thought -- if we start using this mutex here don't we > open a window for force attack on the thread self execution, > iow if there gonna be a million of writers do we have a guarantee > thread ksm_scan_thread will grab the mutex earlier than writers > (or somewhere inbetween)? This file is permitted for global root only. I don't think there is a problem. If someone wants to make ksm helpless, a person may just write a big "sleep_millisecs" value. KSM thread won't be executed almost all the time in this case. Kirill