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 9523EC433EF for ; Wed, 24 Nov 2021 14:03:48 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1354247AbhKXOG5 (ORCPT ); Wed, 24 Nov 2021 09:06:57 -0500 Received: from mail.kernel.org ([198.145.29.99]:50822 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1353751AbhKXOCg (ORCPT ); Wed, 24 Nov 2021 09:02:36 -0500 Received: by mail.kernel.org (Postfix) with ESMTPSA id D728B61A81; Wed, 24 Nov 2021 13:10:04 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1637759405; bh=cX5CyYvbcEYag1cmPgNea++N9v20C836Mt7qe6OHTUQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=wQiGqSTMTOEk7OkqnjMiw3xcoy9kpT2vpWqRUdrFUlbOVjnvlt/Wk+6J075rmoWll XkIqEwSJ9knPgSPc4jPwN0JpOHirzLqRQr/FGBzAWm7OBKx57s57I9BDxBftgjfP3X xH81yZkTOTfUrqje+O394fR262ldWCDjALrzLPZY= From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Mina Almasry , James Houghton , Mike Kravetz , Wei Xu , Andrew Morton , Linus Torvalds Subject: [PATCH 5.15 197/279] hugetlb, userfaultfd: fix reservation restore on userfaultfd error Date: Wed, 24 Nov 2021 12:58:04 +0100 Message-Id: <20211124115725.541561611@linuxfoundation.org> X-Mailer: git-send-email 2.34.0 In-Reply-To: <20211124115718.776172708@linuxfoundation.org> References: <20211124115718.776172708@linuxfoundation.org> User-Agent: quilt/0.66 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Mina Almasry commit cc30042df6fcc82ea18acf0dace831503e60a0b7 upstream. Currently in the is_continue case in hugetlb_mcopy_atomic_pte(), if we bail out using "goto out_release_unlock;" in the cases where idx >= size, or !huge_pte_none(), the code will detect that new_pagecache_page == false, and so call restore_reserve_on_error(). In this case I see restore_reserve_on_error() delete the reservation, and the following call to remove_inode_hugepages() will increment h->resv_hugepages causing a 100% reproducible leak. We should treat the is_continue case similar to adding a page into the pagecache and set new_pagecache_page to true, to indicate that there is no reservation to restore on the error path, and we need not call restore_reserve_on_error(). Rename new_pagecache_page to page_in_pagecache to make that clear. Link: https://lkml.kernel.org/r/20211117193825.378528-1-almasrymina@google.com Fixes: c7b1850dfb41 ("hugetlb: don't pass page cache pages to restore_reserve_on_error") Signed-off-by: Mina Almasry Reported-by: James Houghton Reviewed-by: Mike Kravetz Cc: Wei Xu Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman --- mm/hugetlb.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -5236,13 +5236,14 @@ int hugetlb_mcopy_atomic_pte(struct mm_s int ret = -ENOMEM; struct page *page; int writable; - bool new_pagecache_page = false; + bool page_in_pagecache = false; if (is_continue) { ret = -EFAULT; page = find_lock_page(mapping, idx); if (!page) goto out; + page_in_pagecache = true; } else if (!*pagep) { /* If a page already exists, then it's UFFDIO_COPY for * a non-missing case. Return -EEXIST. @@ -5330,7 +5331,7 @@ int hugetlb_mcopy_atomic_pte(struct mm_s ret = huge_add_to_page_cache(page, mapping, idx); if (ret) goto out_release_nounlock; - new_pagecache_page = true; + page_in_pagecache = true; } ptl = huge_pte_lockptr(h, dst_mm, dst_pte); @@ -5394,7 +5395,7 @@ out_release_unlock: if (vm_shared || is_continue) unlock_page(page); out_release_nounlock: - if (!new_pagecache_page) + if (!page_in_pagecache) restore_reserve_on_error(h, dst_vma, dst_addr, page); put_page(page); goto out;