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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id C29CDEB64DD for ; Wed, 19 Jul 2023 02:35:23 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229766AbjGSCfW (ORCPT ); Tue, 18 Jul 2023 22:35:22 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:54802 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229512AbjGSCfV (ORCPT ); Tue, 18 Jul 2023 22:35:21 -0400 Received: from out-6.mta1.migadu.com (out-6.mta1.migadu.com [95.215.58.6]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 8BA3D1BC3 for ; Tue, 18 Jul 2023 19:35:18 -0700 (PDT) Content-Type: text/plain; charset=us-ascii DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.dev; s=key1; t=1689734116; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=E9ZAEbBiO24j7Uknj5PMKtSmu00bsu4fHz6ctS0uYi8=; b=n0asVgRiKnEBpyy9M1+lERgGLufw19A9QodAMrALzkWRaj5U6PtL8Oxgm592EvQRn6p6RZ gpiOyhAV0RHQCG9XzBocXXqaQbUbtEJQyf4VWcZFbMYhYpw1Ns6TrWF5/urtuidNDtZkSh sZnvHKgALLeThB1MKDjpaXG8D1MM93c= MIME-Version: 1.0 Subject: Re: [PATCH v2 1/2] hugetlb: Do not clear hugetlb dtor until allocating vmemmap X-Report-Abuse: Please report any abuse attempt to abuse@migadu.com and include these headers. From: Muchun Song In-Reply-To: <20230718004942.113174-2-mike.kravetz@oracle.com> Date: Wed, 19 Jul 2023 10:34:36 +0800 Cc: Linux Memory Management List , LKML , Jiaqi Yan , Naoya Horiguchi , Muchun Song , Miaohe Lin , Axel Rasmussen , James Houghton , Michal Hocko , Andrew Morton , stable@vger.kernel.org Content-Transfer-Encoding: quoted-printable Message-Id: <4015AB29-76AD-4CA7-89C4-5916E2F8670B@linux.dev> References: <20230718004942.113174-1-mike.kravetz@oracle.com> <20230718004942.113174-2-mike.kravetz@oracle.com> To: Mike Kravetz X-Migadu-Flow: FLOW_OUT Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > On Jul 18, 2023, at 08:49, Mike Kravetz = wrote: >=20 > Freeing a hugetlb page and releasing base pages back to the underlying > allocator such as buddy or cma is performed in two steps: > - remove_hugetlb_folio() is called to remove the folio from hugetlb > lists, get a ref on the page and remove hugetlb destructor. This > all must be done under the hugetlb lock. After this call, the page > can be treated as a normal compound page or a collection of base > size pages. > - update_and_free_hugetlb_folio() is called to allocate vmemmap if > needed and the free routine of the underlying allocator is called > on the resulting page. We can not hold the hugetlb lock here. >=20 > One issue with this scheme is that a memory error could occur between > these two steps. In this case, the memory error handling code treats > the old hugetlb page as a normal compound page or collection of base > pages. It will then try to SetPageHWPoison(page) on the page with an > error. If the page with error is a tail page without vmemmap, a write > error will occur when trying to set the flag. >=20 > Address this issue by modifying remove_hugetlb_folio() and > update_and_free_hugetlb_folio() such that the hugetlb destructor is = not > cleared until after allocating vmemmap. Since clearing the destructor > requires holding the hugetlb lock, the clearing is done in > remove_hugetlb_folio() if the vmemmap is present. This saves a > lock/unlock cycle. Otherwise, destructor is cleared in > update_and_free_hugetlb_folio() after allocating vmemmap. >=20 > Note that this will leave hugetlb pages in a state where they are = marked > free (by hugetlb specific page flag) and have a ref count. This is = not > a normal state. The only code that would notice is the memory error > code, and it is set up to retry in such a case. >=20 > A subsequent patch will create a routine to do bulk processing of > vmemmap allocation. This will eliminate a lock/unlock cycle for each > hugetlb page in the case where we are freeing a large number of pages. >=20 > Fixes: ad2fa3717b74 ("mm: hugetlb: alloc the vmemmap pages associated = with each HugeTLB page") > Cc: > Signed-off-by: Mike Kravetz Reviewed-by: Muchun Song Thanks.