From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 056FECEDD93 for ; Wed, 9 Oct 2024 13:39:26 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 8267910E190; Wed, 9 Oct 2024 13:39:25 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=intel.com header.i=@intel.com header.b="dRhu1JXS"; dkim-atps=neutral Received: from mgamail.intel.com (mgamail.intel.com [198.175.65.19]) by gabe.freedesktop.org (Postfix) with ESMTPS id 446B410E190 for ; Wed, 9 Oct 2024 13:39:24 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1728481164; x=1760017164; h=message-id:subject:from:to:cc:date:in-reply-to: references:content-transfer-encoding:mime-version; bh=pomC987U5cmPsqGF90kzsyEtu40UHaXy8+GvrxYOIr8=; b=dRhu1JXSr4tMHcqs1zL7DLDpU4ZLRO+m2XPdp78sQm6YMt/coc8oTWi0 reZj1jt+88E7FPILRE5Q5L1s7opnMytPBxKdPOLfEjCO9sG0a9yURHEFB uJkN7NEt2U7cmGA5efGQmsNiPTc2IvE+FC3qjpHyVnZpY6Gf3srd/5LKy cmheKyI4XDqYoDCETyYwpP0wbnK8ESgZBwXMQg63z4BbLdTLioEWjWmzo JzCkgTg3UV9Sb0TaVJwa4exK0ZZLLsrzNjTUnreldVKHmiUOi0vewKbhu DRlaakFx1mKfiJCpu+ufikgQa/n8kGbJWfl25mka5nvuDFaKzmZ+0TFDm g==; X-CSE-ConnectionGUID: E55LoqLBSuWhAbjYbEQVmA== X-CSE-MsgGUID: HhQRVsbtTDuM30CCoxBafw== X-IronPort-AV: E=McAfee;i="6700,10204,11220"; a="27674368" X-IronPort-AV: E=Sophos;i="6.11,189,1725346800"; d="scan'208";a="27674368" Received: from fmviesa009.fm.intel.com ([10.60.135.149]) by orvoesa111.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 09 Oct 2024 06:39:24 -0700 X-CSE-ConnectionGUID: 5ErrRuXnTP2LZDXLqC8HEA== X-CSE-MsgGUID: hNmEwNEgQaua9sHZI/nHog== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.11,189,1725346800"; d="scan'208";a="76262102" Received: from oandoniu-mobl3.ger.corp.intel.com (HELO [10.245.245.243]) ([10.245.245.243]) by fmviesa009-auth.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 09 Oct 2024 06:39:22 -0700 Message-ID: Subject: Re: [PATCH 4/7] drm/ttm: move LRU walk defines into new internal header From: Thomas =?ISO-8859-1?Q?Hellstr=F6m?= To: Christian =?ISO-8859-1?Q?K=F6nig?= , Simona Vetter Cc: Matthew Brost , dri-devel@lists.freedesktop.org Date: Wed, 09 Oct 2024 15:39:20 +0200 In-Reply-To: <2bca2a15-029e-4d8a-9eb4-3dc54f42798e@amd.com> References: <440bb9a5-54b8-46ef-b6db-50110af5c02a@amd.com> <5a2f24bce352b65a1fb6e933c406b3ab1efa33e3.camel@linux.intel.com> <4d4c532a-ff35-4172-9b71-93f5d130711b@amd.com> <3afe3ab2-4a58-49a9-acd7-c989980c68f2@amd.com> <4c634e5c1bd9907f315d8b3535ebb6154819d5ea.camel@linux.intel.com> <2bca2a15-029e-4d8a-9eb4-3dc54f42798e@amd.com> Autocrypt: addr=thomas.hellstrom@linux.intel.com; prefer-encrypt=mutual; keydata=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 Organization: Intel Sweden AB, Registration Number: 556189-6027 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable User-Agent: Evolution 3.50.4 (3.50.4-1.fc39) MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" On Mon, 2024-10-07 at 11:08 +0200, Christian K=C3=B6nig wrote: > Hi Thomas, >=20 > I'm on sick leave, but I will try to answer questions and share some=20 > thoughts on this to unblock you. >=20 > Am 18.09.24 um 14:57 schrieb Thomas Hellstr=C3=B6m: > > Sima, Christian > >=20 > > I've updated the shrinker series now with a guarded for_each macro > > instead: > >=20 > > https://patchwork.freedesktop.org/patch/614514/?series=3D131815&rev=3D9 >=20 > Yeah that looks like a big step in the right direction. >=20 > > (Note I forgot to remove the export of the previous LRU walker). > >=20 > > =C2=A0 so the midlayer argument is now not an issue anymore. The > > cleanup.h > > guard provides some additional protection against drivers exiting > > the > > LRU loop early. > >=20 > > So remaining is the question whether the driver is allowed to > > discard a > > suggested bo to shrink from TTM. > >=20 > > Arguments for: > >=20 > > 1) Not allowing that would require teaching TTM about purgeable > > objects. >=20 > I think that is actually not correct. TTM already knows about > purgeable=20 > objects. >=20 > E.g. when TTM asks the driver what to do with evicted objects it is=20 > perfectly valid to return an empty placement list indicating that the > backing store can just be thrown away. >=20 > We use this for things like temporary buffers for example. >=20 > That this doesn't apply to swapping looks like a design bug in the=20 > driver/TTM interface to me. Yes we can do that with TTM, but for shrinking there's no point in trying to shrink when there is no swap-space left, other than purgeable object. The number of shrinkable objects returned in shrinker::count needs to reflect that, and *then* only those objects should be selected for shrinking. If we were to announce that to TTM using a callback, we're actually back to version 1 of this series which was rejected by you exactly since it was using callbacks a year or so ago???? >=20 > > 2) Devices who need the blitter during shrinking would want to punt > > runtime_pm_get() to kswapd to avoid sleeping direct reclaim. >=20 > I think the outcome of the discussion is that runtime PM should never > be=20 > mixed into TTM swapping. >=20 > You can power up blocks to enable a HW blitter for swapping, but this > then can't be driven by the runtime PM framework. Still that power-on might be sleeping, so what's the difference using runtime-pm or not? Why should the driver implement yet another power interface, just because TTM refuses to publish a sane LRU walk interface? >=20 > > 3) If those devices end up blitting (LNL) to be able to shrink, > > they > > would want to punt waiting for the fence to signal to kswapd to > > avoid > > waiting in direct reclaim. >=20 > Mhm, what do you mean with that? When we fired the blitter from direct reclaim, we get a fence. If we wait for it in direct reclaim we will be sleeping waiting for gpu, which is bad form. We'd like return a failure so the object is retried when idle, or from kswapd. >=20 >=20 > > 4) It looks like we need to resort to folio_trylock in the shmem > > backup > > backend when shrinking is called for gfp_t =3D GFP_NOFS. A failing > > trylock will require a new bo. >=20 > Why would a folio trylock succeed with one BO and not another? Good point. We'd fail anyway but would perhaps need to call SHRINK_STOP.. >=20 > And why would that trylock something the device specific driver > should=20 > handle? It happens in the TTM shrinker helper called from the driver in the spirit of demidlayering. >=20 > > Arguments against: > > None really. I thought the idea of demidlayering would be to allow > > the > > driver more freedom. >=20 > Well that is a huge argument against it. Giving drivers more freedom > is=20 > absolutely not something which turned out to be valuable in the past. So then what's the point of demidlayering? >=20 > Instead we should put device drivers in a very strict corset of=20 > validated approaches to do things. >=20 > Background is that in my experience driver developers are perfectly=20 > willing to do unclean approaches which only work in like 99% of all=20 > cases just to get a bit more performance or simpler driver > implementation. >=20 > Those approaches are not legal and in my opinion as subsystem > maintainer=20 > I think we need to be more strict and push back much harder on stuff=20 > like that. Still, historically that has made developers abandon common components for driver-specific solutions.=20 And the question is still not answered. Exactly *why* can't the driver fail and continue traversing the LRU, because all our argumentation revolves around this and you have yet to provide an objective reason why.=20 /Thomas >=20 > Regards, > Christian. >=20 > >=20 > > So any feedback appreciated. If that is found acceptable we can > > proceed > > with reviewing this patch and also with the shrinker series. > >=20 > > Thanks, > > Thomas > >=20 > >=20 > > On Mon, 2024-09-02 at 13:07 +0200, Daniel Vetter wrote: > > > On Wed, Aug 28, 2024 at 02:20:34PM +0200, Christian K=C3=B6nig wrote: > > > > Am 27.08.24 um 19:53 schrieb Daniel Vetter: > > > > > On Tue, Aug 27, 2024 at 06:52:13PM +0200, Daniel Vetter > > > > > wrote: > > > > > > On Thu, Aug 22, 2024 at 03:19:29PM +0200, Christian K=C3=B6nig > > > > > > wrote: > > > > > > > Completely agree that this is complicated, but I still > > > > > > > don't > > > > > > > see the need > > > > > > > for it. > > > > > > >=20 > > > > > > > Drivers just need to use pm_runtime_get_if_in_use() > > > > > > > inside > > > > > > > the shrinker and > > > > > > > postpone all hw activity until resume. > > > > > > Not good enough, at least long term I think. Also > > > > > > postponing hw > > > > > > activity > > > > > > to resume doesn't solve the deadlock issue, if you still > > > > > > need > > > > > > to grab ttm > > > > > > locks on resume. > > > > > Pondered this specific aspect some more, and I think you > > > > > still > > > > > have a race > > > > > here (even if you avoid the deadlock): If the condiditional > > > > > rpm_get call > > > > > fails there's no guarantee that the device will > > > > > suspend/resume > > > > > and clean > > > > > up the GART mapping. > > > > Well I think we have a major disconnect here. When the device > > > > is > > > > powered > > > > down there is no GART mapping to clean up any more. > > > >=20 > > > > In other words GART is a table in local memory (VRAM) when the > > > > device is > > > > powered down this table is completely destroyed. Any BO which > > > > was > > > > mapped > > > > inside this table is now not mapped any more. > > > >=20 > > > > So when the shrinker wants to evict a BO which is marked as > > > > mapped > > > > to GART > > > > and the device is powered down we just skip the GART unmapping > > > > part > > > > because > > > > that has already implicitly happened during power down. > > > >=20 > > > > Before mapping any BO into the GART again we power the GPU up > > > > through the > > > > runtime PM calls. And while powering it up again the GART is > > > > restored. > > > My point is that you can't tell whether the device will power > > > down or > > > not, > > > you can only tell whether there's a chance it might be powering > > > down > > > and > > > so you can't get at the rpm reference without deadlock issues. > > >=20 > > > > > The race gets a bit smaller if you use > > > > > pm_runtime_get_if_active(), but even then you might catch it > > > > > right when > > > > > resume almost finished. > > > > What race are you talking about? > > > >=20 > > > > The worst thing which could happen is that we restore a GART > > > > entry > > > > which > > > > isn't needed any more, but that is pretty much irrelevant since > > > > we > > > > only > > > > clear them to avoid some hw bugs. > > > The race I'm seeing is where you thought the GART entry is not > > > issue, > > > tossed an object, but the device didn't suspend, so might still > > > use > > > it. > > >=20 > > > I guess if we're clearly separating the sw allocation of the > > > TTM_TT > > > with > > > the physical entries in the GART that should all work, but feels > > > a > > > bit > > > tricky. The race I've seen is essentially these two getting out > > > of > > > sync. > > >=20 > > > So maybe it was me who's stuck. > > >=20 > > > What I wonder is whether it works in practice, since on the > > > restore > > > side > > > you need to get some locks to figure out which gart mappings > > > exist > > > and > > > need restoring. And that's the same locks as the shrinker needs > > > to > > > figure > > > out whether it might need to reap a gart mapping. > > >=20 > > > Or do you just copy the gart entries over and restore them > > > exactly > > > as-is, > > > so that there's no shared locks? > > >=20 > > > > > That means we'll have ttm bo hanging around with GART > > > > > allocations/mappings > > > > > which aren't actually valid anymore (since they might escape > > > > > the > > > > > cleanup > > > > > upon resume due to the race). That doesn't feel like a solid > > > > > design > > > > > either. > > > > I'm most likely missing something, but I'm really scratching my > > > > head where > > > > you see a problem here. > > > I guess one issue is that at least traditionally, igfx drivers > > > have > > > nested > > > runtime pm within dma_resv lock. And dgpu drivers the other way > > > round. > > > Which is a bit awkward if you're trying for common code. > > >=20 > > > Cheers, Sima >=20