From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mout.gmx.net ([212.227.15.19]:54292 "EHLO mout.gmx.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932090AbdLTFl1 (ORCPT ); Wed, 20 Dec 2017 00:41:27 -0500 Subject: Re: [PATCH v2 06/17] btrfs-progs: lowmem check: introduce force_cow_in_new_chunk() To: Su Yue , linux-btrfs@vger.kernel.org References: <20171220045731.19343-1-suy.fnst@cn.fujitsu.com> <20171220045731.19343-7-suy.fnst@cn.fujitsu.com> From: Qu Wenruo Message-ID: <37268331-348d-fe29-db97-e289bb2dea1c@gmx.com> Date: Wed, 20 Dec 2017 13:41:03 +0800 MIME-Version: 1.0 In-Reply-To: <20171220045731.19343-7-suy.fnst@cn.fujitsu.com> Content-Type: multipart/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; boundary="a5v4ETLA4wFJfkGClhyWPTJy2LRCjX25j" Sender: linux-btrfs-owner@vger.kernel.org List-ID: This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --a5v4ETLA4wFJfkGClhyWPTJy2LRCjX25j Content-Type: multipart/mixed; boundary="VXpHVfUEjB3IuBIg7Pf6Lnwy5At0NGqRA"; protected-headers="v1" From: Qu Wenruo To: Su Yue , linux-btrfs@vger.kernel.org Message-ID: <37268331-348d-fe29-db97-e289bb2dea1c@gmx.com> Subject: Re: [PATCH v2 06/17] btrfs-progs: lowmem check: introduce force_cow_in_new_chunk() References: <20171220045731.19343-1-suy.fnst@cn.fujitsu.com> <20171220045731.19343-7-suy.fnst@cn.fujitsu.com> In-Reply-To: <20171220045731.19343-7-suy.fnst@cn.fujitsu.com> --VXpHVfUEjB3IuBIg7Pf6Lnwy5At0NGqRA Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: quoted-printable On 2017=E5=B9=B412=E6=9C=8820=E6=97=A5 12:57, Su Yue wrote: > Introduce create_chunk_and_block_block_group() to allocate new chunk > and corresponding block group. >=20 > The new function force_cow_in_new_chunk() first allocates new chunk > and records its start. > Then it modifies all metadata block groups cached and full. > Finally it marks the new block group uncached and unfree. > In the next CoW, extents states will be updated automatically by > cache_block_group(). >=20 > Suggested-by: Qu Wenruo > Signed-off-by: Su Yue > --- > cmds-check.c | 80 ++++++++++++++++++++++++++++++++++++++++++++++++++++= ++++++++ > 1 file changed, 80 insertions(+) >=20 > diff --git a/cmds-check.c b/cmds-check.c > index d98d4bda7357..311c8a9f45e8 100644 > --- a/cmds-check.c > +++ b/cmds-check.c > @@ -10911,6 +10911,86 @@ out: > return ret; > } > =20 > +static int create_chunk_and_block_group(struct btrfs_fs_info *fs_info,= > + u64 flags, u64 *start, u64 *nbytes) > +{ > + struct btrfs_trans_handle *trans; > + struct btrfs_root *root =3D fs_info->extent_root; > + int ret; > + > + if ((flags & BTRFS_BLOCK_GROUP_TYPE_MASK) =3D=3D 0) > + return -EINVAL; > + > + trans =3D btrfs_start_transaction(root, 1); > + if (IS_ERR(trans)) { > + ret =3D PTR_ERR(trans); > + error("error starting transaction %s", strerror(-ret)); > + return ret; > + } > + ret =3D btrfs_alloc_chunk(trans, fs_info, start, nbytes, flags); > + if (ret) { > + error("fail to allocate new chunk %s", strerror(-ret)); > + goto out; > + } > + ret =3D btrfs_make_block_group(trans, fs_info, 0, flags, > + BTRFS_FIRST_CHUNK_TREE_OBJECTID, *start, *nbytes); > + if (ret) { > + error("fail to make block group for chunk %llu %llu %s", > + *start, *nbytes, strerror(-ret)); > + goto out; > + } > +out: > + btrfs_commit_transaction(trans, root); > + return ret; > +} > + > +static int force_cow_in_new_chunk(struct btrfs_fs_info *fs_info) > +{ > + struct btrfs_block_group_cache *bg; > + u64 start; > + u64 nbytes; > + u64 alloc_profile; > + u64 flags; > + int ret; > + > + alloc_profile =3D (fs_info->avail_metadata_alloc_bits & > + fs_info->metadata_alloc_profile); > + flags =3D BTRFS_BLOCK_GROUP_METADATA | alloc_profile; > + if (btrfs_fs_incompat(fs_info, MIXED_GROUPS)) > + flags |=3D BTRFS_BLOCK_GROUP_DATA; > + > + ret =3D create_chunk_and_block_group(fs_info, flags, &start, &nbytes)= ; Why bother allocating a new chunk by yourself? Just mark all block groups full and that's all. Any later tree block allocation like btrfs_search_slot() with @cow =3D 1 will trigger chunk allocation automatically. Thanks, Qu > + if (!ret) > + printf("Create new chunk %llu %llu\n", start, nbytes); > + else > + goto err; > + > + flags =3D BTRFS_BLOCK_GROUP_METADATA; > + /* Mark all metadata block groups cached and full in free space*/ > + ret =3D modify_block_groups_cache(fs_info, flags, 1); > + if (ret) > + goto clear_bg_cache; > + > + bg =3D btrfs_lookup_block_group(fs_info, start); > + if (!bg) { > + ret =3D -ENOENT; > + error("fail to look up block group %llu %llu", start, nbytes); > + goto clear_bg_cache; > + } > + > + /* Clear block group cache just allocated */ > + ret =3D modify_block_group_cache(fs_info, bg, 0); > + if (ret) > + goto clear_bg_cache; > + > + return 0; > + > +clear_bg_cache: > + modify_block_groups_cache(fs_info, flags, 0); > +err: > + return ret; > +} > + > static int check_extent_refs(struct btrfs_root *root, > struct cache_tree *extent_cache) > { >=20 --VXpHVfUEjB3IuBIg7Pf6Lnwy5At0NGqRA-- --a5v4ETLA4wFJfkGClhyWPTJy2LRCjX25j Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- iQFLBAEBCAA1FiEELd9y5aWlW6idqkLhwj2R86El/qgFAlo59/AXHHF1d2VucnVv LmJ0cmZzQGdteC5jb20ACgkQwj2R86El/qip7gf/XS2rkC9ED4hDAyWSXFwBt8pQ gdcQA+i3x4I3udXkJmOVrvCBTOkiIOf9CiOJAoc1Fi8mxfdb8T6iqnzRqHKlpMXJ GbPsJ1G/JJZw4uMnJtWXOgf1yhaggYGlMxad54MWyLtIwq36dNAnOD8+xYTAfd+U WX/ahIM5B/QdCP6SASs/dlnOLjXNgGiBU10d7wieyYUnX9Teau5jDIHGnO1qSvsh zk2MxI2N9B5FfJ6cRNja5oTEKyWhBhn4ooymEb1NldT4NCCYo3R+q3m98AUwAgiJ CPnGC1KExPu3Ei+px7ZTRZUiZgVlt5lO8j0ZwkEKb4oX9hGNXkdgWgJEe0gtEQ== =iHUr -----END PGP SIGNATURE----- --a5v4ETLA4wFJfkGClhyWPTJy2LRCjX25j--