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 X-Spam-Level: X-Spam-Status: No, score=-15.8 required=3.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id C9BD6C433DB for ; Thu, 14 Jan 2021 22:41:20 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 5420023977 for ; Thu, 14 Jan 2021 22:41:20 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 5420023977 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=konsulko.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 75BB66B00A0; Thu, 14 Jan 2021 17:41:19 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 70DB78D00F0; Thu, 14 Jan 2021 17:41:19 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 621196B00A5; Thu, 14 Jan 2021 17:41:19 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0051.hostedemail.com [216.40.44.51]) by kanga.kvack.org (Postfix) with ESMTP id 4CF116B00A0 for ; Thu, 14 Jan 2021 17:41:19 -0500 (EST) Received: from smtpin05.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay02.hostedemail.com (Postfix) with ESMTP id 0C5C3363D for ; Thu, 14 Jan 2021 22:41:19 +0000 (UTC) X-FDA: 77705853078.05.end35_321469d2752a Received: from filter.hostedemail.com (10.5.16.251.rfc1918.com [10.5.16.251]) by smtpin05.hostedemail.com (Postfix) with ESMTP id E2DF2180222AB for ; Thu, 14 Jan 2021 22:41:18 +0000 (UTC) X-HE-Tag: end35_321469d2752a X-Filterd-Recvd-Size: 9376 Received: from mail-lf1-f49.google.com (mail-lf1-f49.google.com [209.85.167.49]) by imf34.hostedemail.com (Postfix) with ESMTP for ; Thu, 14 Jan 2021 22:41:18 +0000 (UTC) Received: by mail-lf1-f49.google.com with SMTP id a12so10401899lfl.6 for ; Thu, 14 Jan 2021 14:41:17 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=konsulko.com; s=google; h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc; bh=SY1HN+r4epF6M/QDgSX4NKwjn1lj53VyzVpAPsBhiYs=; b=XSjrSL8Gupx9YISpttSDKVkQYm2Sntg3vxDYLLfxOH1NQG3/nLxXhI59kItf1RwWgO jcgEkOL42AG0hv7iadcmVGoObFwvhWi4g+YWlB66CWpz5bompkKLrETzVBsYMqfeYIL9 jwhTYLKHMewbRCcL2QqvO5tlp0VBFkdAT2BeI= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=SY1HN+r4epF6M/QDgSX4NKwjn1lj53VyzVpAPsBhiYs=; b=eYAmW6IFVmHXzQpspkiM/aiNJJacggKAK/vLrPRFtL88YGrAckAEgYYrpctpmHmS2U DgLvdVfHx+ECVZxTPIAPGUbrySNDFfJpIpDIaPLddYFFFGgfnyjbd++HqfQfKWo8hTLG q+hVc4k/T/68TGym59Jr8a7/Sd7XNHJBF+dl70U6gOtrulzBKq6+KTgXoIodtFa6+5gg WMTiz3aeyjhD5rBmHh5kDnVNuSxDE/HsoHBJCqrEAb1ISpOH+ulXdgMNyhk4e157JcjU zDjPLBjFejsFE5daeIX8ELKgdt4sVc9punIDyntEpFBANkmVgmjzDa8r0ITpzZBt+MPU YI6w== X-Gm-Message-State: AOAM530x8K/5cwQfvcgnbqqFpqgr8lgfsSgh+jtQnMMfOmsSb+i0P7aS 7CxaGy8D/EGc2GL6nE4VflYoagWihPK7hXoX0l4FYw== X-Google-Smtp-Source: ABdhPJx3+H/+Euoe8gkJqOKxvPaNv5gDtFdyN5Ts1iCXt4FodOdihIagrubcspUl2MsLbfib6KOuJQru6FczV+B/Bw4= X-Received: by 2002:a05:6512:3e7:: with SMTP id n7mr4011278lfq.585.1610664076541; Thu, 14 Jan 2021 14:41:16 -0800 (PST) MIME-Version: 1.0 References: <1608894171-54174-1-git-send-email-tiantao6@hisilicon.com> <1608894171-54174-2-git-send-email-tiantao6@hisilicon.com> In-Reply-To: From: Vitaly Wool Date: Thu, 14 Jan 2021 23:41:05 +0100 Message-ID: Subject: Re: [RFC mm/zswap 1/2] mm/zswap: add the flag can_sleep_mapped To: Shakeel Butt Cc: Minchan Kim , Tian Tao , Seth Jennings , Dan Streetman , Andrew Morton , Barry Song , Linux-MM , Sergey Senozhatsky Content-Type: text/plain; charset="UTF-8" X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: On Thu, Jan 14, 2021 at 10:09 PM Shakeel Butt wrote: > > On Thu, Jan 14, 2021 at 11:53 AM Vitaly Wool wrote: > > > > On Thu, Jan 14, 2021 at 8:21 PM Shakeel Butt wrote: > > > > > > On Thu, Jan 14, 2021 at 11:05 AM Vitaly Wool wrote: > > > > > > > > On Thu, Jan 14, 2021 at 7:56 PM Minchan Kim wrote: > > > > > > > > > > On Thu, Jan 14, 2021 at 07:40:50PM +0100, Vitaly Wool wrote: > > > > > > On Thu, Jan 14, 2021 at 7:29 PM Minchan Kim wrote: > > > > > > > > > > > > > > On Fri, Dec 25, 2020 at 07:02:50PM +0800, Tian Tao wrote: > > > > > > > > add a flag to zpool, named is "can_sleep_mapped", and have it set true > > > > > > > > for zbud/z3fold, set false for zsmalloc. Then zswap could go the current > > > > > > > > path if the flag is true; and if it's false, copy data from src to a > > > > > > > > temporary buffer, then unmap the handle, take the mutex, process the > > > > > > > > buffer instead of src to avoid sleeping function called from atomic > > > > > > > > context. > > > > > > > > > > > > > > > > Signed-off-by: Tian Tao > > > > > > > > --- > > > > > > > > include/linux/zpool.h | 3 +++ > > > > > > > > mm/zpool.c | 13 +++++++++++++ > > > > > > > > mm/zswap.c | 50 +++++++++++++++++++++++++++++++++++++++++++++----- > > > > > > > > 3 files changed, 61 insertions(+), 5 deletions(-) > > > > > > > > > > > > > > > > diff --git a/include/linux/zpool.h b/include/linux/zpool.h > > > > > > > > index 51bf430..e899701 100644 > > > > > > > > --- a/include/linux/zpool.h > > > > > > > > +++ b/include/linux/zpool.h > > > > > > > > @@ -73,6 +73,7 @@ u64 zpool_get_total_size(struct zpool *pool); > > > > > > > > * @malloc: allocate mem from a pool. > > > > > > > > * @free: free mem from a pool. > > > > > > > > * @shrink: shrink the pool. > > > > > > > > + * @sleep_mapped: whether zpool driver can sleep during map. > > > > > > > > > > > > > > I don't think it's a good idea. It just breaks zpool abstraction > > > > > > > in that it exposes internal implementation to user to avoid issue > > > > > > > zswap recently introduced. It also conflicts zpool_map_handle's > > > > > > > semantic. > > > > > > > > > > > > > > Rather than introducing another break in zpool due to the new > > > > > > > zswap feature recenlty introduced, zswap could introduce > > > > > > > CONFIG_ZSWAP_HW_COMPRESSOR. Once it's configured, zsmalloc could > > > > > > > be disabled. And with disabling CONFIG_ZSWAP_HW_COMPRESSOR, zswap > > > > > > > doesn't need to make any bounce buffer copy so that no existing > > > > > > > zsmalloc user will see performance regression. > > > > > > > > > > > > I believe it won't help that much -- the new compressor API presumes > > > > > > that the caller may sleep during compression and that will be an > > > > > > accident waiting to happen as long as we use it and keep the handle > > > > > > mapped in zsmalloc case. > > > > > > > > > > > > Or maybe I interpreted you wrong and you are suggesting re-introducing > > > > > > calls to the old API under this #ifdef, is that the case? > > > > > > > > > > Yub. zswap could abstract that part under #ifdef to keep old behavior. > > > > > > > > We can reconsider this option when zsmalloc implements reclaim > > > > callback. So far it's obviously too much a mess for a reason so weak. > > > > > > > > > > Sorry I don't understand the link between zsmalloc implementing shrink > > > callback and this patch. > > > > There is none. There is a link between taking all the burden to revive > > zsmalloc for zswap at the cost of extra zswap complexity and zsmalloc > > not being fully zswap compatible. > > > > The ultimate zswap goal is to cache hottest swapped-out pages in a > > compressed format. zsmalloc doesn't implement reclaim callback, and > > therefore zswap can *not* fulfill its goal since old pages are there > > to stay, and it can't accept new hotter ones. So, let me make it > > clear: zswap/zsmalloc combo is a legacy corner case. > > > > This is the first time I am hearing that zswap/zsmalloc combo is a > legacy configuration. We use zswap in production and intentionally > size the zswap pool to never have to go to the backing device. So > absence of reclaim callback is not an issue for us. Please let me know > if the zswap/zsmalloc combo is officially on its way to deprecation. No, zswap/zsmalloc combo not on the way to deprecation. I generally would not advise on using it but your particular case does make sense (although using frontswap/zswap without a backing device *is* a corner case). > > > This patch is adding an overhead for all > > > zswap+zsmalloc users irrespective of availability of hardware. If we > > > want to add support for new hardware, please add without impacting the > > > current users. > > > > No, it's not like that. zswap+zsmalloc combination is currently > > already broken > > By broken do you mean the missing reclaim callback? No. I mean deadlocks in -rt kernel / scheduling while atomic bugs in the mainline. Missing reclaim callback goes into "not fully compatible" section in my world. > > and this patch implements a way to work that around. > > The users are already impacted and that is of course a problem. > > Are you talking about rt users or was there some other report? I don't think the issue is specific to -rt series. There may (and will) still be "scheduling while atomic" bugs occurring, just not as often. > > The workaround is not perfect but looks good enough for me. > > I would like a see page fault perf experiment for the non-hardware case. I second that. @tiantao (H), would it be possible to provide one? Also, correct me if I'm wrong but from what I recall, the acomp API does one redundant copy less than the old comp API so zsmalloc should be back to square one even with the buffer implemented in this patch. The other backends should do a little better though but if so, it's the upside of not taking too many spinlocks. Best regards, Vitaly