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=-7.0 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS 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 053A5C169C4 for ; Fri, 8 Feb 2019 06:31:36 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id BC91B2147C for ; Fri, 8 Feb 2019 06:31:35 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726911AbfBHGaN (ORCPT ); Fri, 8 Feb 2019 01:30:13 -0500 Received: from mx2.suse.de ([195.135.220.15]:49296 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1726004AbfBHGaM (ORCPT ); Fri, 8 Feb 2019 01:30:12 -0500 X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx1.suse.de (Postfix) with ESMTP id A4AA4AF1F; Fri, 8 Feb 2019 06:30:10 +0000 (UTC) Subject: Re: [PATCH 1/2] btrfs: reserve space for inheriting properties To: Josef Bacik , linux-btrfs@vger.kernel.org, kernel-team@fb.com References: <20190207165426.15866-1-josef@toxicpanda.com> <20190207165426.15866-2-josef@toxicpanda.com> From: Nikolay Borisov Openpgp: preference=signencrypt Autocrypt: addr=nborisov@suse.com; prefer-encrypt=mutual; keydata= mQINBFiKBz4BEADNHZmqwhuN6EAzXj9SpPpH/nSSP8YgfwoOqwrP+JR4pIqRK0AWWeWCSwmZ T7g+RbfPFlmQp+EwFWOtABXlKC54zgSf+uulGwx5JAUFVUIRBmnHOYi/lUiE0yhpnb1KCA7f u/W+DkwGerXqhhe9TvQoGwgCKNfzFPZoM+gZrm+kWv03QLUCr210n4cwaCPJ0Nr9Z3c582xc bCUVbsjt7BN0CFa2BByulrx5xD9sDAYIqfLCcZetAqsTRGxM7LD0kh5WlKzOeAXj5r8DOrU2 GdZS33uKZI/kZJZVytSmZpswDsKhnGzRN1BANGP8sC+WD4eRXajOmNh2HL4P+meO1TlM3GLl EQd2shHFY0qjEo7wxKZI1RyZZ5AgJnSmehrPCyuIyVY210CbMaIKHUIsTqRgY5GaNME24w7h TyyVCy2qAM8fLJ4Vw5bycM/u5xfWm7gyTb9V1TkZ3o1MTrEsrcqFiRrBY94Rs0oQkZvunqia c+NprYSaOG1Cta14o94eMH271Kka/reEwSZkC7T+o9hZ4zi2CcLcY0DXj0qdId7vUKSJjEep c++s8ncFekh1MPhkOgNj8pk17OAESanmDwksmzh1j12lgA5lTFPrJeRNu6/isC2zyZhTwMWs k3LkcTa8ZXxh0RfWAqgx/ogKPk4ZxOXQEZetkEyTFghbRH2BIwARAQABtCNOaWtvbGF5IEJv cmlzb3YgPG5ib3Jpc292QHN1c2UuY29tPokCOAQTAQIAIgUCWIo48QIbAwYLCQgHAwIGFQgC CQoLBBYCAwECHgECF4AACgkQcb6CRuU/KFc0eg/9GLD3wTQz9iZHMFbjiqTCitD7B6dTLV1C ddZVlC8Hm/TophPts1bWZORAmYIihHHI1EIF19+bfIr46pvfTu0yFrJDLOADMDH+Ufzsfy2v HSqqWV/nOSWGXzh8bgg/ncLwrIdEwBQBN9SDS6aqsglagvwFD91UCg/TshLlRxD5BOnuzfzI Leyx2c6YmH7Oa1R4MX9Jo79SaKwdHt2yRN3SochVtxCyafDlZsE/efp21pMiaK1HoCOZTBp5 VzrIP85GATh18pN7YR9CuPxxN0V6IzT7IlhS4Jgj0NXh6vi1DlmKspr+FOevu4RVXqqcNTSS E2rycB2v6cttH21UUdu/0FtMBKh+rv8+yD49FxMYnTi1jwVzr208vDdRU2v7Ij/TxYt/v4O8 V+jNRKy5Fevca/1xroQBICXsNoFLr10X5IjmhAhqIH8Atpz/89ItS3+HWuE4BHB6RRLM0gy8 T7rN6ja+KegOGikp/VTwBlszhvfLhyoyjXI44Tf3oLSFM+8+qG3B7MNBHOt60CQlMkq0fGXd mm4xENl/SSeHsiomdveeq7cNGpHi6i6ntZK33XJLwvyf00PD7tip/GUj0Dic/ZUsoPSTF/mG EpuQiUZs8X2xjK/AS/l3wa4Kz2tlcOKSKpIpna7V1+CMNkNzaCOlbv7QwprAerKYywPCoOSC 7P25Ag0EWIoHPgEQAMiUqvRBZNvPvki34O/dcTodvLSyOmK/MMBDrzN8Cnk302XfnGlW/YAQ csMWISKKSpStc6tmD+2Y0z9WjyRqFr3EGfH1RXSv9Z1vmfPzU42jsdZn667UxrRcVQXUgoKg QYx055Q2FdUeaZSaivoIBD9WtJq/66UPXRRr4H/+Y5FaUZx+gWNGmBT6a0S/GQnHb9g3nonD jmDKGw+YO4P6aEMxyy3k9PstaoiyBXnzQASzdOi39BgWQuZfIQjN0aW+Dm8kOAfT5i/yk59h VV6v3NLHBjHVw9kHli3jwvsizIX9X2W8tb1SefaVxqvqO1132AO8V9CbE1DcVT8fzICvGi42 FoV/k0QOGwq+LmLf0t04Q0csEl+h69ZcqeBSQcIMm/Ir+NorfCr6HjrB6lW7giBkQl6hhomn l1mtDP6MTdbyYzEiBFcwQD4terc7S/8ELRRybWQHQp7sxQM/Lnuhs77MgY/e6c5AVWnMKd/z MKm4ru7A8+8gdHeydrRQSWDaVbfy3Hup0Ia76J9FaolnjB8YLUOJPdhI2vbvNCQ2ipxw3Y3c KhVIpGYqwdvFIiz0Fej7wnJICIrpJs/+XLQHyqcmERn3s/iWwBpeogrx2Lf8AGezqnv9woq7 OSoWlwXDJiUdaqPEB/HmGfqoRRN20jx+OOvuaBMPAPb+aKJyle8zABEBAAGJAh8EGAECAAkF AliKBz4CGwwACgkQcb6CRuU/KFdacg/+M3V3Ti9JYZEiIyVhqs+yHb6NMI1R0kkAmzsGQ1jU zSQUz9AVMR6T7v2fIETTT/f5Oout0+Hi9cY8uLpk8CWno9V9eR/B7Ifs2pAA8lh2nW43FFwp IDiSuDbH6oTLmiGCB206IvSuaQCp1fed8U6yuqGFcnf0ZpJm/sILG2ECdFK9RYnMIaeqlNQm iZicBY2lmlYFBEaMXHoy+K7nbOuizPWdUKoKHq+tmZ3iA+qL5s6Qlm4trH28/fPpFuOmgP8P K+7LpYLNSl1oQUr+WlqilPAuLcCo5Vdl7M7VFLMq4xxY/dY99aZx0ZJQYFx0w/6UkbDdFLzN upT7NIN68lZRucImffiWyN7CjH23X3Tni8bS9ubo7OON68NbPz1YIaYaHmnVQCjDyDXkQoKC R82Vf9mf5slj0Vlpf+/Wpsv/TH8X32ajva37oEQTkWNMsDxyw3aPSps6MaMafcN7k60y2Wk/ TCiLsRHFfMHFY6/lq/c0ZdOsGjgpIK0G0z6et9YU6MaPuKwNY4kBdjPNBwHreucrQVUdqRRm RcxmGC6ohvpqVGfhT48ZPZKZEWM+tZky0mO7bhZYxMXyVjBn4EoNTsXy1et9Y1dU3HVJ8fod 5UqrNrzIQFbdeM0/JqSLrtlTcXKJ7cYFa9ZM2AP7UIN9n1UWxq+OPY9YMOewVfYtL8M= Message-ID: <12a76243-ae38-974b-957f-c50955d53934@suse.com> Date: Fri, 8 Feb 2019 08:30:08 +0200 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.2.1 MIME-Version: 1.0 In-Reply-To: <20190207165426.15866-2-josef@toxicpanda.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 8bit Sender: linux-btrfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-btrfs@vger.kernel.org On 7.02.19 г. 18:54 ч., Josef Bacik wrote: > We've been seeing errors on our build servers related to failing to > inherit inode properties. This is because we do not pre-reserve space > for them, instead trying to reserve space with NO_FLUSH at inheritance > time. NO_FLUSH can transiently fail, but we'll still complain. It's Put one or two sentences describing the implications of BTRFS_RESERVE_NO_FLUSH. I.e that it NO_FLUSH won't try to flush current metadata reservation which could result in the said transient failure. Just to give a bit more context for someone who might be reading the commit message in the future. > just an extra credit, so simply add that to the places that call > btrfs_new_inode and call it good enough. > > Signed-off-by: Josef Bacik > --- > fs/btrfs/inode.c | 78 ++++++++++++++++++++++---------------------------------- > fs/btrfs/ioctl.c | 27 ++++++++++++-------- > 2 files changed, 46 insertions(+), 59 deletions(-) > > diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c > index 6126de9b8b9c..0da4a9d6d9fe 100644 > --- a/fs/btrfs/inode.c > +++ b/fs/btrfs/inode.c > @@ -59,6 +59,14 @@ struct btrfs_dio_data { > int overwrite; > }; > > +/* > + * 2 for inode item and ref > + * 2 for dir items > + * 1 for xattr if selinux is on > + * 1 for inherited properties > + */ > +#define BTRFS_NEW_INODE_ITEMS 6 Rather than having scattered defines I'd much prefer if we have an enum and over time add all distinct reservations to that enum and change btrfs_start_transaction's interface to take this enum. It will be much more descriptive than having scattered defines. > + > static const struct inode_operations btrfs_dir_inode_operations; > static const struct inode_operations btrfs_symlink_inode_operations; > static const struct inode_operations btrfs_dir_ro_inode_operations; > @@ -6479,12 +6487,7 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, > u64 objectid; > u64 index = 0; > > - /* > - * 2 for inode item and ref > - * 2 for dir items > - * 1 for xattr if selinux is on > - */ > - trans = btrfs_start_transaction(root, 5); > + trans = btrfs_start_transaction(root, BTRFS_NEW_INODE_ITEMS); > if (IS_ERR(trans)) > return PTR_ERR(trans); > > @@ -6543,12 +6546,7 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, > u64 objectid; > u64 index = 0; > > - /* > - * 2 for inode item and ref > - * 2 for dir items > - * 1 for xattr if selinux is on > - */ > - trans = btrfs_start_transaction(root, 5); > + trans = btrfs_start_transaction(root, BTRFS_NEW_INODE_ITEMS); > if (IS_ERR(trans)) > return PTR_ERR(trans); > > @@ -6695,12 +6693,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) > u64 objectid = 0; > u64 index = 0; > > - /* > - * 2 items for inode and ref > - * 2 items for dir items > - * 1 for xattr if selinux is on > - */ > - trans = btrfs_start_transaction(root, 5); > + trans = btrfs_start_transaction(root, BTRFS_NEW_INODE_ITEMS); > if (IS_ERR(trans)) > return PTR_ERR(trans); > > @@ -9428,14 +9421,11 @@ static int btrfs_rename_exchange(struct inode *old_dir, > down_read(&fs_info->subvol_sem); > > /* > - * We want to reserve the absolute worst case amount of items. So if > - * both inodes are subvols and we need to unlink them then that would > - * require 4 item modifications, but if they are both normal inodes it > - * would require 5 item modifications, so we'll assume their normal > - * inodes. So 5 * 2 is 10, plus 2 for the new links, so 12 total items > - * should cover the worst case number of items we'll modify. > + * The same math from btrfs_rename applies here, except we need an extra > + * 2 items for the new links. > */ > - trans = btrfs_start_transaction(root, 12); > + trans = btrfs_start_transaction(root, > + (BTRFS_NEW_INODE_ITEMS << 1) + 2); > if (IS_ERR(trans)) { > ret = PTR_ERR(trans); > goto out_notrans; > @@ -9768,19 +9758,19 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, > if (old_ino == BTRFS_FIRST_FREE_OBJECTID) > down_read(&fs_info->subvol_sem); > /* > - * We want to reserve the absolute worst case amount of items. So if > - * both inodes are subvols and we need to unlink them then that would > - * require 4 item modifications, but if they are both normal inodes it > - * would require 5 item modifications, so we'll assume they are normal > - * inodes. So 5 * 2 is 10, plus 1 for the new link, so 11 total items > - * should cover the worst case number of items we'll modify. > - * If our rename has the whiteout flag, we need more 5 units for the > - * new inode (1 inode item, 1 inode ref, 2 dir items and 1 xattr item > - * when selinux is enabled). > + * We want to reserve the absolute worst case amount of items. Subvol > + * inodes don't have an inode item to worry about and don't have a > + * selinux attr, so we use the BTRFS_NEW_INODE_ITEMS counter for how > + * much it costs per inode to modify. Worse case we'll have to mess > + * with 2 inodes, so 2 x BTRFS_NEW_INODE_ITEMS, and then we need an > + * extra reservation for the new link. > + * > + * If our rename has the whiteout flag we need a full new inode which > + * means another set of BTRFS_NEW_INODE_ITEMS. > */ > - trans_num_items = 11; > + trans_num_items = (BTRFS_NEW_INODE_ITEMS << 1) + 1; > if (flags & RENAME_WHITEOUT) > - trans_num_items += 5; > + trans_num_items += BTRFS_NEW_INODE_ITEMS; > trans = btrfs_start_transaction(root, trans_num_items); > if (IS_ERR(trans)) { > ret = PTR_ERR(trans); > @@ -10149,14 +10139,8 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, > if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(fs_info)) > return -ENAMETOOLONG; > > - /* > - * 2 items for inode item and ref > - * 2 items for dir items > - * 1 item for updating parent inode item > - * 1 item for the inline extent item > - * 1 item for xattr if selinux is on > - */ > - trans = btrfs_start_transaction(root, 7); > + /* 1 item for the inline extent item */ > + trans = btrfs_start_transaction(root, BTRFS_NEW_INODE_ITEMS + 1); > if (IS_ERR(trans)) > return PTR_ERR(trans); > > @@ -10427,10 +10411,8 @@ static int btrfs_tmpfile(struct inode *dir, struct dentry *dentry, umode_t mode) > u64 index; > int ret = 0; > > - /* > - * 5 units required for adding orphan entry > - */ > - trans = btrfs_start_transaction(root, 5); > + /* 1 unit required for adding orphan entry */ > + trans = btrfs_start_transaction(root, BTRFS_NEW_INODE_ITEMS + 1); > if (IS_ERR(trans)) > return PTR_ERR(trans); > > diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c > index f38a659c918c..21f8ab2d8570 100644 > --- a/fs/btrfs/ioctl.c > +++ b/fs/btrfs/ioctl.c > @@ -83,6 +83,17 @@ struct btrfs_ioctl_send_args_32 { > struct btrfs_ioctl_send_args_32) > #endif > > +/* > + * 1 - parent dir inode > + * 2 - dir entries > + * 1 - root item > + * 2 - root ref/backref > + * 1 - root of snapshot > + * 1 - UUID item > + * 1 - properties > + */ > +#define BTRFS_NEW_ROOT_ITEMS 9 Even if you choose to have defines currently, I insist on them being grouped in one place e.g. ctree.h > + > static int btrfs_clone(struct inode *src, struct inode *inode, > u64 off, u64 olen, u64 olen_aligned, u64 destoff, > int no_time_update); > @@ -596,7 +607,8 @@ static noinline int create_subvol(struct inode *dir, > * The same as the snapshot creation, please see the comment > * of create_snapshot(). > */ > - ret = btrfs_subvolume_reserve_metadata(root, &block_rsv, 8, false); > + ret = btrfs_subvolume_reserve_metadata(root, &block_rsv, > + BTRFS_NEW_ROOT_ITEMS, false); > if (ret) > goto fail_free; > > @@ -804,17 +816,10 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir, > > btrfs_init_block_rsv(&pending_snapshot->block_rsv, > BTRFS_BLOCK_RSV_TEMP); > - /* > - * 1 - parent dir inode > - * 2 - dir entries > - * 1 - root item > - * 2 - root ref/backref > - * 1 - root of snapshot > - * 1 - UUID item > - */ > + > ret = btrfs_subvolume_reserve_metadata(BTRFS_I(dir)->root, > - &pending_snapshot->block_rsv, 8, > - false); > + &pending_snapshot->block_rsv, > + BTRFS_NEW_ROOT_ITEMS, false); > if (ret) > goto dec_and_free; > >