Received: by 2002:a25:1985:0:0:0:0:0 with SMTP id 127csp6214ybz; Thu, 30 Apr 2020 15:22:29 -0700 (PDT) X-Google-Smtp-Source: APiQypJrJhd/l9Jf+NzpdR8le5MW0Nv7KtcQl/D9Ge0LXEaRTfPl4T8N/54g42iPXTRJ78mEUK2t X-Received: by 2002:aa7:cdcb:: with SMTP id h11mr1083033edw.264.1588285349297; Thu, 30 Apr 2020 15:22:29 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1588285349; cv=none; d=google.com; s=arc-20160816; b=lkHeh3jOhCuXFiWxyTixARvd/vNBExpnZkJF9dUBxiNoMmw5xIYCAP2vq4dqXW4OH7 JMwJhMdP0wdiKl7UHVMhTNM9JtLF78FNGPraACTmackBzszCPYob9RGGsWhKOYVe+AXA e5aZ50fr7IOEOZo76kvBIxDxexMWLbAFv9+R1S3L2ZU2Pe2n3lMDi8Ukq7jC/Mmwih12 8n4W8PvfCe3J3THjoL8/Za18KcipWtf7k2gJMs9nvotVziLoP/PwTELcn+wXtkpwjk1N nc6WW1Y+c0Kz1Nik9o5jXjcTy0+wFXTHa1CDNRLt5Rcu//RSAOCmSmASyZED9p+km90J /vSQ== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:sender:dkim-signature:content-language :content-transfer-encoding:in-reply-to:mime-version:user-agent:date :message-id:references:cc:to:from:subject; bh=w/+Cr9VQxl0rdOKMwo55fWqb+uL6Teyc7xqvUUJB8nk=; b=l2tGN++W/wSb+2jmSHSep4ksA+clrbxrCpMo1awoUglGvw/q4B3R5PG4jAGQwYb3Tc xfbcZYYCCW8jmye2rh1sfz44wesBjC6TmZPQQ47DivnhetRY6Jr4lHdJ4RRO0KYpiGws 679puXM6cO8FCNTB0mM6EHTeagnlWHmPpGw4Bwr59Yzuw+Qpp3f6mAa4/3UY1aH4r9zV zT8QHlOQvtJp2AKX/T2pgErM5jnhZ3aYS6G6HtKzr46hHTwCM1yIqey3woqNor+DO49d DY36oeJPqro+RSU2ffK7qNmvazIHi6irirneObekLsGG+oOriciWs6CZVmrLsosyB0t+ kN1Q== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@nvidia.com header.s=n1 header.b=jlD8S9bk; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=nvidia.com Return-Path: Received: from vger.kernel.org (vger.kernel.org. [23.128.96.18]) by mx.google.com with ESMTP id n7si582913ejy.248.2020.04.30.15.22.06; Thu, 30 Apr 2020 15:22:29 -0700 (PDT) Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18; Authentication-Results: mx.google.com; dkim=pass header.i=@nvidia.com header.s=n1 header.b=jlD8S9bk; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=nvidia.com Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727114AbgD3WUg (ORCPT + 99 others); Thu, 30 Apr 2020 18:20:36 -0400 Received: from hqnvemgate26.nvidia.com ([216.228.121.65]:18046 "EHLO hqnvemgate26.nvidia.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726447AbgD3WUg (ORCPT ); Thu, 30 Apr 2020 18:20:36 -0400 Received: from hqpgpgate101.nvidia.com (Not Verified[216.228.121.13]) by hqnvemgate26.nvidia.com (using TLS: TLSv1.2, DES-CBC3-SHA) id ; Thu, 30 Apr 2020 15:20:23 -0700 Received: from hqmail.nvidia.com ([172.20.161.6]) by hqpgpgate101.nvidia.com (PGP Universal service); Thu, 30 Apr 2020 15:20:36 -0700 X-PGP-Universal: processed; by hqpgpgate101.nvidia.com on Thu, 30 Apr 2020 15:20:36 -0700 Received: from DRHQMAIL107.nvidia.com (10.27.9.16) by HQMAIL107.nvidia.com (172.20.187.13) with Microsoft SMTP Server (TLS) id 15.0.1473.3; Thu, 30 Apr 2020 22:20:35 +0000 Received: from [10.2.165.152] (10.124.1.5) by DRHQMAIL107.nvidia.com (10.27.9.16) with Microsoft SMTP Server (TLS) id 15.0.1473.3; Thu, 30 Apr 2020 22:20:29 +0000 Subject: Re: [RFC PATCH v11 6/9] media: tegra: Add Tegra210 Video input driver From: Sowjanya Komatineni To: Dmitry Osipenko , , , , , , CC: , , , , , References: <1588197606-32124-1-git-send-email-skomatineni@nvidia.com> <1588197606-32124-7-git-send-email-skomatineni@nvidia.com> <4da289e6-036f-853b-beb4-379d6462adb0@gmail.com>

<7d31d24f-f353-7e82-3ff9-cdba8b773d1e@nvidia.com> <06a4a067-8d54-4322-b2a6-14e82eaeda29@nvidia.com> <47873bbd-cf90-4595-5a99-7e9113327ecc@nvidia.com> <71532440-f455-cc24-74f7-9ccad5947099@gmail.com> <298187f6-2425-4813-1ae1-f256c179623e@nvidia.com> <9c942bc9-703e-3bbb-eeab-f37e69dc1ded@nvidia.com> <668d9b65-9590-cc97-41c3-2c1a5cfbbe61@nvidia.com> <289d9c92-383f-3257-de7b-46179724285a@nvidia.com> <9aa64f21-7b23-7228-b5eb-d2ff092682ad@nvidia.com> <668cc4a0-2c81-0d87-b801-9fbf64e19137@nvidia.com> Message-ID: Date: Thu, 30 Apr 2020 15:19:00 -0700 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.7.0 MIME-Version: 1.0 In-Reply-To: <668cc4a0-2c81-0d87-b801-9fbf64e19137@nvidia.com> X-Originating-IP: [10.124.1.5] X-ClientProxiedBy: HQMAIL111.nvidia.com (172.20.187.18) To DRHQMAIL107.nvidia.com (10.27.9.16) Content-Type: text/plain; charset="utf-8"; format=flowed Content-Transfer-Encoding: quoted-printable Content-Language: en-US DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=nvidia.com; s=n1; t=1588285223; bh=w/+Cr9VQxl0rdOKMwo55fWqb+uL6Teyc7xqvUUJB8nk=; h=X-PGP-Universal:Subject:From:To:CC:References:Message-ID:Date: User-Agent:MIME-Version:In-Reply-To:X-Originating-IP: X-ClientProxiedBy:Content-Type:Content-Transfer-Encoding: Content-Language; b=jlD8S9bkc07IvIAs1vyLks6zla/YWiKcJ8Yuoo66G+CA5cz/XHfFgYErNo4uNE2rc RhWUQ7NVMNqsh/zMhTD955wDejKvGl457NoT2sN67uhDaUgn1zXphjmXWE5zjEDCzi Js5QVCqLuPFCNtezAfAhl9p0Jol4AZghcBolAYiSHoOFm+yqgS3LOWuuEhnj58EL1A draEopun7850Ttse1ZN2+HsnRiS+dbNJvtvpzIx1+oWsO5TfoxWqNdZGGpTyMY5PCV eVnJGq/XYVg+1sPg40JdpohJTNLq5I039KZFWXuBq8WHUtbmFhk9Yc2ITXYEMh1XcA hYcrkuviLyW9w== Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 4/30/20 3:16 PM, Sowjanya Komatineni wrote: > > On 4/30/20 2:53 PM, Sowjanya Komatineni wrote: >> >> On 4/30/20 2:37 PM, Sowjanya Komatineni wrote: >>> >>> On 4/30/20 2:26 PM, Sowjanya Komatineni wrote: >>>> >>>> On 4/30/20 2:17 PM, Dmitry Osipenko wrote: >>>>> 30.04.2020 23:02, Sowjanya Komatineni =D0=BF=D0=B8=D1=88=D0=B5=D1=82: >>>>>> On 4/30/20 12:53 PM, Sowjanya Komatineni wrote: >>>>>>> On 4/30/20 12:46 PM, Sowjanya Komatineni wrote: >>>>>>>> On 4/30/20 12:33 PM, Dmitry Osipenko wrote: >>>>>>>>> 30.04.2020 22:09, Sowjanya Komatineni =D0=BF=D0=B8=D1=88=D0=B5=D1= =82: >>>>>>>>>> On 4/30/20 11:18 AM, Sowjanya Komatineni wrote: >>>>>>>>>>> On 4/30/20 10:06 AM, Sowjanya Komatineni wrote: >>>>>>>>>>>> On 4/30/20 9:29 AM, Sowjanya Komatineni wrote: >>>>>>>>>>>>> On 4/30/20 9:04 AM, Sowjanya Komatineni wrote: >>>>>>>>>>>>>> On 4/30/20 7:13 AM, Dmitry Osipenko wrote: >>>>>>>>>>>>>>> 30.04.2020 17:02, Dmitry Osipenko =D0=BF=D0=B8=D1=88=D0=B5= =D1=82: >>>>>>>>>>>>>>>> 30.04.2020 16:56, Dmitry Osipenko =D0=BF=D0=B8=D1=88=D0=B5= =D1=82: >>>>>>>>>>>>>>>>> 30.04.2020 01:00, Sowjanya Komatineni =D0=BF=D0=B8=D1=88= =D0=B5=D1=82: >>>>>>>>>>>>>>>>>> +static int chan_capture_kthread_finish(void *data) >>>>>>>>>>>>>>>>>> +{ >>>>>>>>>>>>>>>>>> +=C2=A0=C2=A0=C2=A0 struct tegra_vi_channel *chan =3D da= ta; >>>>>>>>>>>>>>>>>> +=C2=A0=C2=A0=C2=A0 struct tegra_channel_buffer *buf; >>>>>>>>>>>>>>>>>> + >>>>>>>>>>>>>>>>>> +=C2=A0=C2=A0=C2=A0 set_freezable(); >>>>>>>>>>>>>>>>>> + >>>>>>>>>>>>>>>>>> +=C2=A0=C2=A0=C2=A0 while (1) { >>>>>>>>>>>>>>>>>> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 try_to_freez= e(); >>>>>>>>>>>>>>>>> I guess it won't be great to freeze in the middle of a=20 >>>>>>>>>>>>>>>>> capture >>>>>>>>>>>>>>>>> process, so: >>>>>>>>>>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if= (list_empty(&chan->done)) >>>>>>>>>>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0 try_to_freeze(); >>>>>>>>>>>>>>>> And here should be some locking protection in order not=20 >>>>>>>>>>>>>>>> race >>>>>>>>>>>>>>>> with >>>>>>>>>>>>>>>> the >>>>>>>>>>>>>>>> chan_capture_kthread_start because kthread_finish could=20 >>>>>>>>>>>>>>>> freeze >>>>>>>>>>>>>>>> before >>>>>>>>>>>>>>>> kthread_start. >>>>>>>>>>>>>>> Or maybe both start / finish threads should simply be=20 >>>>>>>>>>>>>>> allowed to >>>>>>>>>>>>>>> freeze >>>>>>>>>>>>>>> only when both capture and done lists are empty. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> if (list_empty(&chan->capture) && >>>>>>>>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 list_empty(&chan->done= )) >>>>>>>>>>>>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0try_to_freeze(); >>>>>>>>>>>>>> good to freeze when not in middle of the frame capture=20 >>>>>>>>>>>>>> but why >>>>>>>>>>>>>> should we not allow freeze in between captures? >>>>>>>>>>>>>> >>>>>>>>>>>>>> Other drivers do allow freeze in between frame captures. >>>>>>>>>>>>>> >>>>>>>>>>>>>> I guess we can freeze before dequeue for capture and in=20 >>>>>>>>>>>>>> finish >>>>>>>>>>>>>> thread we can freeze after capture done. This also don't=20 >>>>>>>>>>>>>> need to >>>>>>>>>>>>>> check for list_empty with freeze to allow between frame=20 >>>>>>>>>>>>>> captures. >>>>>>>>>>>>>> >>>>>>>>>>>>> Also if we add check for both lists empty, freeze is not=20 >>>>>>>>>>>>> allowed as >>>>>>>>>>>>> long as streaming is going on and in case of continuous=20 >>>>>>>>>>>>> streaming >>>>>>>>>>>>> freeze will never happen. >>>>>>>>>>> To allow freeze b/w frames (but not in middle of a frame), >>>>>>>>>>> >>>>>>>>>>> for capture_start thread, probably we can do unconditional >>>>>>>>>>> try_to_freeze() >>>>>>>>> Is it possible to use wait_event_freezable()? >>>>>>>>> >>>>>>>>> https://www.kernel.org/doc/Documentation/power/freezing-of-tasks.= txt=20 >>>>>>>>> >>>>>>>>> >>>>>>>>> Will the wait_event_interruptible() be woken up when system=20 >>>>>>>>> freezes? >>>>>>>> Based on wait_event_freezable implementation, looks like it=20 >>>>>>>> similar >>>>>>>> to wait_event_interruptible + try_to_free() as it does >>>>>>>> freezable_schedule unlike schedule with wait_event_interruptible. >>>>>>>> >>>>>>>> So using this for capture_start may be ok to allow freeze before >>>>>>>> start of frame. But can't use for capture_finish as this is=20 >>>>>>>> same as >>>>>>>> wait_event_interruptible followed by unconditional try_to_freeze. >>>>>>>> >>>>>>>>>>> for capture_finish thread, at end of capture done we can do >>>>>>>>>>> try_to_freeze() only when done list is empty >>>>>>>>> This doesn't prevent situation where the done-list is empty=20 >>>>>>>>> and the >>>>>>>>> "finish" thread freezes, in the same time the "start" thread=20 >>>>>>>>> issues new >>>>>>>>> capture and then freezes too. >>>>>>>>> >>>>>>>>> 1. "start" thread issues capture >>>>>>>>> >>>>>>>>> 2. "finish" thread wakes and waits for the capture to complete >>>>>>>>> >>>>>>>>> 3. "start" thread begins another capture, waits for FRAME_START >>>>>>>>> >>>>>>>>> 4. system freezing activates >>>>>>>>> >>>>>>>>> 5. "finish" thread completes the capture and freezes because=20 >>>>>>>>> done-list >>>>>>>>> is empty >>>>>>>>> >>>>>>>>> 6. "start" thread gets FRAME_START, issues another capture and=20 >>>>>>>>> freezes >>>>>>>> This will not happen as we allow double buffering done list=20 >>>>>>>> will not >>>>>>>> be empty till stream stop happens >>>>>>>> >>>>>>>> There will always be 1 outstanding frame in done list >>>>>>> Correction, there will always be 1 outstanding buffer except=20 >>>>>>> beginning >>>>>>> during beginning of stream. >>>>>>> >>>>>>> Except during beginning frames, done list will not be empty for all >>>>>>> subsequent streaming process >>>>>> Also to be clear, hardware sees next frame start event prior to=20 >>>>>> previous >>>>>> frame mw_ack event as mw_ack event happens after frame end. So once >>>>>> initial buffer got queued to done list to finish processes, while >>>>>> waiting for mw_ack next frame start happens and pushes next=20 >>>>>> buffer to >>>>>> done list. >>>>> What about this variant: >>>>> >>>>> 1. "start" thread wakes up to start capture >>>>> >>>>> 2. system freezing activates >>>>> >>>>> 3. "finish" thread wakes up and freezes >>>> >>>> finish thread will wake up only when done list is not=20 >>>> empty/kthread_stop/wake even from capture start thread. >>>> >>>> Also when I said will allow try_to_free when done list is empty I=20 >>>> meant to have this at end of capture_done() in finish thread >>>> >>>>> >>>>> 4. "start" thread issues capture and freezes >>>>> >>>>> And again, I assume that system's freezing should wake >>>>> wait_event_interruptible(), otherwise it won't be possible to freeze >>>>> idling threads at all and freezing should fail (IIUC). >>>> >>>> Based on kernel doc on freezing, looks like when we mark thread as=20 >>>> freezable, freeze state happens when we explicitly call try_to_freeze. >>>> >>>> I don't think its other way where freeze causes=20 >>>> wait_event_interruptible to wake up. >> >> Based on my understanding when we mark thread as freezable, >> >> with wait_event_freezable() - after wait event, it invokes=20 >> try_to_freeze(). So frozen state enters unconditionally with this. >> >> with wait_event_interruptible - we do try_to_freeze when its safe to=20 >> enter frozen state. >> >> https://www.kernel.org/doc/Documentation/power/freezing-of-tasks.txt >> > Sorry correction. When system tries to freeze tasks looks like it will=20 > sending signal to thread and wake up happens when signal is sent to=20 > thread and freezable thread should invoke try_to_free when its safe to=20 > free freeze_task() sends fake signal https://elixir.bootlin.com/linux/v5.7-rc2/source/kernel/freezer.c#L115 > >> >>>> >>>>> And in this case synchronization between start/finish threads=20 >>>>> should be >>>>> needed in regards to freezing. >>>> >>>> Was thinking to have counter to track outstanding frame w.r.t=20 >>>> single shot issue b/w start and finish and allow to freeze only=20 >>>> when no outstanding frames in process. >>>> >>>> This will make sure freeze will not happen when any buffers are in=20 >>>> progress >>>> >>>>> Note that this could be a wrong assumption, I'm not closely familiar >>>>> with how freezer works. >>> >>> kthread_start can unconditionally allow try_to_freeze before start=20 >>> of frame capture >>> >>> We can compute captures inflight w.r.t single shot issued during=20 >>> capture start and finished frames by kthread_finish and allow=20 >>> kthread_finish to freeze only when captures inflight is 0. >>> >>> This allows freeze to happen b/w frames but not in middle of frame will have caps inflight check in v12 to allow freeze finish thread only=20 when no captures are in progress >>> >>>