From mboxrd@z Thu Jan 1 00:00:00 1970 From: Johannes Weiner Subject: Re: [RFC PATCH 1/7] cgroup: rstat: only disable interrupts for the percpu lock Date: Thu, 23 Mar 2023 14:19:25 -0400 Message-ID: <20230323181925.GA742010@cmpxchg.org> References: <20230323173343.GF739026@cmpxchg.org> Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Return-path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=cmpxchg-org.20210112.gappssmtp.com; s=20210112; t=1679595566; h=in-reply-to:content-transfer-encoding:content-disposition :mime-version:references:message-id:subject:cc:to:from:date:from:to :cc:subject:date:message-id:reply-to; bh=4UdIOBBpU5M34Q+dnxHktyTShJiu1SyYtXU+zhX+XRQ=; b=iSci0p8JFRPXqHPEil3M1UZ/AptTFLyiH4nOONlx8jlszjb2lxcuRkMBw7UZXMYuIb 9VHiSPLwsHCGWKxBshnY3dxP8I+VjQfUKrOS8QxSItXdSt2WijBjz2KEim5klCxbUPhR gr9geE8sgGsa6YxvHsaCLYHtCOhtZ6Cj5uxcAZqUHM/Oo9TpSV6LTpCjwajd0gTTy2/o iGV78mzHEJs46abGoP8FjHVkdxrPvpdUU0X6sI4iRzWXH+1PxvpLLlNb42LXVJdJuGZu 4FeSqA110YAy7/7r+UTZpX1+gN+1sKQK2Ig4sSlYBcZ1rci0KNN0Hu5Nnn+hpheYoXSX jdGw== Content-Disposition: inline In-Reply-To: List-ID: Content-Type: text/plain; charset="utf-8" To: Yosry Ahmed Cc: Shakeel Butt , Tejun Heo , Josef Bacik , Jens Axboe , Zefan Li , Michal Hocko , Roman Gushchin , Muchun Song , Andrew Morton , Vasily Averin , cgroups@vger.kernel.org, linux-block@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, bpf@vger.kernel.org On Thu, Mar 23, 2023 at 11:09:30AM -0700, Yosry Ahmed wrote: > On Thu, Mar 23, 2023 at 10:33 AM Johannes Weiner wrote: > > > > On Thu, Mar 23, 2023 at 09:17:33AM -0700, Yosry Ahmed wrote: > > > On Thu, Mar 23, 2023 at 9:10 AM Shakeel Butt wrote: > > > > > > > > On Thu, Mar 23, 2023 at 8:46 AM Shakeel Butt wrote: > > > > > > > > > > On Thu, Mar 23, 2023 at 8:43 AM Yosry Ahmed wrote: > > > > > > > > > > > > On Thu, Mar 23, 2023 at 8:40 AM Shakeel Butt wrote: > > > > > > > > > > > > > > On Thu, Mar 23, 2023 at 6:36 AM Yosry Ahmed wrote: > > > > > > > > > > > > > > > [...] > > > > > > > > > > > > > > > > > > > > > 2. Are we really calling rstat flush in irq context? > > > > > > > > > > > > > > > > > > > > I think it is possible through the charge/uncharge path: > > > > > > > > > > memcg_check_events()->mem_cgroup_threshold()->mem_cgroup_usage(). I > > > > > > > > > > added the protection against flushing in an interrupt context for > > > > > > > > > > future callers as well, as it may cause a deadlock if we don't disable > > > > > > > > > > interrupts when acquiring cgroup_rstat_lock. > > > > > > > > > > > > > > > > > > > > > 3. The mem_cgroup_flush_stats() call in mem_cgroup_usage() is only > > > > > > > > > > > done for root memcg. Why is mem_cgroup_threshold() interested in root > > > > > > > > > > > memcg usage? Why not ignore root memcg in mem_cgroup_threshold() ? > > > > > > > > > > > > > > > > > > > > I am not sure, but the code looks like event notifications may be set > > > > > > > > > > up on root memcg, which is why we need to check thresholds. > > > > > > > > > > > > > > > > > > This is something we should deprecate as root memcg's usage is ill defined. > > > > > > > > > > > > > > > > Right, but I think this would be orthogonal to this patch series. > > > > > > > > > > > > > > > > > > > > > > I don't think we can make cgroup_rstat_lock a non-irq-disabling lock > > > > > > > without either breaking a link between mem_cgroup_threshold and > > > > > > > cgroup_rstat_lock or make mem_cgroup_threshold work without disabling > > > > > > > irqs. > > > > > > > > > > > > > > So, this patch can not be applied before either of those two tasks are > > > > > > > done (and we may find more such scenarios). > > > > > > > > > > > > > > > > > > Could you elaborate why? > > > > > > > > > > > > My understanding is that with an in_task() check to make sure we only > > > > > > acquire cgroup_rstat_lock from non-irq context it should be fine to > > > > > > acquire cgroup_rstat_lock without disabling interrupts. > > > > > > > > > > From mem_cgroup_threshold() code path, cgroup_rstat_lock will be taken > > > > > with irq disabled while other code paths will take cgroup_rstat_lock > > > > > with irq enabled. This is a potential deadlock hazard unless > > > > > cgroup_rstat_lock is always taken with irq disabled. > > > > > > > > Oh you are making sure it is not taken in the irq context through > > > > should_skip_flush(). Hmm seems like a hack. Normally it is recommended > > > > to actually remove all such users instead of silently > > > > ignoring/bypassing the functionality. > > > > +1 > > > > It shouldn't silently skip the requested operation, rather it > > shouldn't be requested from an incompatible context. > > > > > > So, how about removing mem_cgroup_flush_stats() from > > > > mem_cgroup_usage(). It will break the known chain which is taking > > > > cgroup_rstat_lock with irq disabled and you can add > > > > WARN_ON_ONCE(!in_task()). > > > > > > This changes the behavior in a more obvious way because: > > > 1. The memcg_check_events()->mem_cgroup_threshold()->mem_cgroup_usage() > > > path is also exercised in a lot of paths outside irq context, this > > > will change the behavior for any event thresholds on the root memcg. > > > With proposed skipped flushing in irq context we only change the > > > behavior in a small subset of cases. > > > > Can you do > > > > /* Note: stale usage data when called from irq context!! */ > > if (in_task()) > > mem_cgroup_flush_stats() > > > > directly in the callsite? Maybe even include the whole callchain in > > the comment that's currently broken and needs fixing/removing. > > Yeah, we can do that in mem_cgroup_usage(), which is the only context > that I am aware of that may flush from irq context. We can also add > WARN_ON_ONCE(!in_task()) in the rstat core flushing code to catch any > other code paths that we are not aware of -- which may result in a > deadlock, but hopefully if there is a violation it will be caught soon > enough. That sounds good to me, thanks!