From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from userp1040.oracle.com ([156.151.31.81]:47948 "EHLO userp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S966433AbdEOTSs (ORCPT ); Mon, 15 May 2017 15:18:48 -0400 Date: Mon, 15 May 2017 12:18:09 -0700 From: "Darrick J. Wong" Subject: Re: [PATCH 1/9] xfs_io: support the new getfsmap ioctl Message-ID: <20170515191809.GO4519@birch.djwong.org> References: <149417257252.24656.2629280196308754994.stgit@birch.djwong.org> <149417258268.24656.3441330396895207773.stgit@birch.djwong.org> <212fe5da-d589-1614-0980-ab583db1df5d@sandeen.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <212fe5da-d589-1614-0980-ab583db1df5d@sandeen.net> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: Eric Sandeen Cc: sandeen@redhat.com, linux-xfs@vger.kernel.org On Mon, May 08, 2017 at 04:01:11PM -0500, Eric Sandeen wrote: > On 5/7/17 10:56 AM, Darrick J. Wong wrote: > > From: Darrick J. Wong > > > > Signed-off-by: Darrick J. Wong > > --- > > io/Makefile | 4 > > io/copy_file_range.c | 2 > > io/encrypt.c | 1 > > io/fsmap.c | 559 ++++++++++++++++++++++++++++++++++++++++++++++++++ > > io/init.c | 8 + > > io/io.h | 14 + > > io/open.c | 21 ++ > > io/pwrite.c | 2 > > io/reflink.c | 4 > > io/sendfile.c | 2 > > man/man8/xfs_io.8 | 47 ++++ > > 11 files changed, 651 insertions(+), 13 deletions(-) > > create mode 100644 io/fsmap.c > > > > > > diff --git a/io/Makefile b/io/Makefile > > index 435ccff..8d3a30e 100644 > > --- a/io/Makefile > > +++ b/io/Makefile > > @@ -99,6 +99,10 @@ ifeq ($(HAVE_MREMAP),yes) > > LCFLAGS += -DHAVE_MREMAP > > endif > > > > +ifeq ($(HAVE_GETFSMAP),yes) > > +CFILES += fsmap.c > > +endif > > + > > default: depend $(LTCOMMAND) > > > > include $(BUILDRULES) > > diff --git a/io/copy_file_range.c b/io/copy_file_range.c > > index 249c649..d1dfc5a 100644 > > --- a/io/copy_file_range.c > > +++ b/io/copy_file_range.c > > @@ -121,7 +121,7 @@ copy_range_f(int argc, char **argv) > > if (optind != argc - 1) > > return command_usage(©_range_cmd); > > > > - fd = openfile(argv[optind], NULL, IO_READONLY, 0); > > + fd = openfile(argv[optind], NULL, IO_READONLY, 0, NULL); > > if (fd < 0) > > return 0; > > > > diff --git a/io/encrypt.c b/io/encrypt.c > > index d844c5e..26ab97c 100644 > > --- a/io/encrypt.c > > +++ b/io/encrypt.c > > @@ -20,6 +20,7 @@ > > #include "platform_defs.h" > > #include "command.h" > > #include "init.h" > > +#include "path.h" > > #include "io.h" > > > > #ifndef ARRAY_SIZE > > diff --git a/io/fsmap.c b/io/fsmap.c > > new file mode 100644 > > index 0000000..4128fae > > --- /dev/null > > +++ b/io/fsmap.c > > @@ -0,0 +1,559 @@ > > +/* > > + * Copyright (C) 2017 Oracle. All Rights Reserved. > > + * > > + * Author: Darrick J. Wong > > + * > > + * This program is free software; you can redistribute it and/or > > + * modify it under the terms of the GNU General Public License > > + * as published by the Free Software Foundation; either version 2 > > + * of the License, or (at your option) any later version. > > + * > > + * This program is distributed in the hope that it would be useful, > > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the > > + * GNU General Public License for more details. > > + * > > + * You should have received a copy of the GNU General Public License > > + * along with this program; if not, write the Free Software Foundation, > > + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. > > + */ > > +#include "platform_defs.h" > > +#include "command.h" > > +#include "init.h" > > +#include "path.h" > > +#include "io.h" > > +#include "input.h" > > + > > +static cmdinfo_t fsmap_cmd; > > +static dev_t xfs_data_dev; > > + > > +static void > > +fsmap_help(void) > > +{ > > + printf(_( > > +"\n" > > +" prints the block mapping for an XFS filesystem" > > Did you want an \n" on the above line too? Seems a bit odd to have it only on the next line. > (one \n is fine too, but not like this) :) > > > +"\n" > > +" Example:\n" > > +" 'fsmap -dlrv [-n nr] [startoff] [endoff]' - tabular format verbose map, including unwritten extents\n" > > This doesn't match the short help: > > + fsmap_cmd.args = _("[-v] [-n nx] [start] [end]"); > > or the manpage: > > +.BI "fsmap [ \-v ] [ \-n " nx " ] [ " start " ] [ " end " ] > > but getopt sez: > > > + while ((c = getopt(argc, argv, "dln:rv")) != EOF) { > > so I guess the short help & manpage need updates. > > Also, I don't think the "Example:" above is valid: > > if (dflag + lflag + rflag > 1) > return command_usage(&fsmap_cmd); > > so the help needs to indicate that exactly 1 of -d, -l, or -r is required. > > > +"\n" > > +" fsmap prints the map of disk blocks used by the whole filesystem.\n" > > +" The map lists each extent used by the file > > > > "the file?" what file? > > , as well as regions in the\n" > > +" filesystem that do not have any corresponding blocks (free space).\n" > > +" By default, each line of the listing takes the following form:\n" > > +" extent: [startoffset..endoffset] owner startblock..endblock\n" > > say what "owner" means in this case? > > > +" All the file offsets and disk blocks are in units of 512-byte blocks.\n" > > +" -d -- query only the data device.\n" > > +" -l -- query only the log device.\n" > > +" -r -- query only the realtime device.\n" > > +" -n -- query n extents.\n" > > +" -v -- Verbose information, specify ag info. Show flags legend on 2nd -v\n" > > +"\n")); > > +} > > + > > +static int > > +numlen( > > + off64_t val) > > +{ > > + off64_t tmp; > > + int len; > > + > > + for (len = 0, tmp = val; tmp > 0; tmp = tmp/10) > > + len++; > > + return (len == 0 ? 1 : len); > > +} > > Hm, copy #3, do we still not have anywhere to put common stuff like this? > > > + > > +#define OWNER_BUF_SZ 32 > > +static const char * > > +special_owner( > > + __int64_t owner, > > + char *buf) > > +{ > > + switch (owner) { > > + case XFS_FMR_OWN_FREE: > > + return _("free space"); > > + case XFS_FMR_OWN_UNKNOWN: > > + return _("unknown"); > > + case XFS_FMR_OWN_FS: > > + return _("static fs metadata"); > > + case XFS_FMR_OWN_LOG: > > + return _("journalling log"); > > + case XFS_FMR_OWN_AG: > > + return _("per-AG metadata"); > > + case XFS_FMR_OWN_INOBT: > > + return _("inode btree"); > > + case XFS_FMR_OWN_INODES: > > + return _("inodes"); > > + case XFS_FMR_OWN_REFC: > > + return _("refcount btree"); > > + case XFS_FMR_OWN_COW: > > + return _("cow reservation"); > > + case XFS_FMR_OWN_DEFECTIVE: > > + return _("defective"); > > + default: > > + snprintf(buf, OWNER_BUF_SZ, _("special %u:%u"), > > + FMR_OWNER_TYPE(owner), FMR_OWNER_CODE(owner)); > > + return buf; > > + } > > +} > > + > > +static void > > +dump_map( > > + unsigned long long *nr, > > + struct fsmap_head *head) > > +{ > > + unsigned long long i; > > + struct fsmap *p; > > + char owner[OWNER_BUF_SZ]; > > + > > + for (i = 0, p = head->fmh_recs; i < head->fmh_entries; i++, p++) { > > + printf("\t%llu: %u:%u [%lld..%lld]: ", i + (*nr), > > + major(p->fmr_device), minor(p->fmr_device), > > + (long long)BTOBBT(p->fmr_physical), > > + (long long)BTOBBT(p->fmr_physical + p->fmr_length - 1)); > > + if (p->fmr_flags & FMR_OF_SPECIAL_OWNER) > > + printf("%s", special_owner(p->fmr_owner, owner)); > > + else if (p->fmr_flags & FMR_OF_EXTENT_MAP) > > + printf(_("inode %lld extent map"), > > + (long long) p->fmr_owner); > > + else > > + printf(_("inode %lld %lld..%lld"), > > + (long long)p->fmr_owner, > > + (long long)BTOBBT(p->fmr_offset), > > + (long long)BTOBBT(p->fmr_offset + p->fmr_length - 1)); > > + printf(_(" %lld blocks\n"), > > + (long long)BTOBBT(p->fmr_length)); > > + } > > + > > + (*nr) += head->fmh_entries; > > +} > > + > > +/* > > + * Verbose mode displays: > > + * extent: major:minor [startblock..endblock]: startoffset..endoffset \ > > + * ag# (agoffset..agendoffset) totalbbs flags > > + */ > > +#define MINRANGE_WIDTH 16 > > +#define MINAG_WIDTH 2 > > +#define MINTOT_WIDTH 5 > > +#define NFLG 7 /* count of flags */ > > +#define FLG_NULL 00000000 /* Null flag */ > > +#define FLG_SHARED 01000000 /* shared extent */ > > +#define FLG_ATTR_FORK 00100000 /* attribute fork */ > > +#define FLG_PRE 00010000 /* Unwritten extent */ > > +#define FLG_BSU 00001000 /* Not on begin of stripe unit */ > > +#define FLG_ESU 00000100 /* Not on end of stripe unit */ > > +#define FLG_BSW 00000010 /* Not on begin of stripe width */ > > +#define FLG_ESW 00000001 /* Not on end of stripe width */ > > These really couldn't share w/ io/bmap.c? :( > (could FLG_ATTR_FORK go to the end so that it matches bmap until then?) Well yes, we could flip them (I already did); afaict the only change that needs to happen is an update to xfs/274. > speakinawhich, I wonder how much of this is copied from bmap, and what could > be shared? > > I'm going to go off & look at that a bit, may have further comments. :) Do you have further comments? --D > > > > diff --git a/man/man8/xfs_io.8 b/man/man8/xfs_io.8 > > index 29a036c..2c956b8 100644 > > --- a/man/man8/xfs_io.8 > > +++ b/man/man8/xfs_io.8 > > @@ -301,6 +301,53 @@ ioctl. Options behave as described in the > > .BR xfs_bmap (8) > > manual page. > > .TP > > +.BI "fsmap [ \-v ] [ \-n " nx " ] [ " start " ] [ " end " ] > > +Prints the mapping of disk blocks used by an XFS filesystem. > > FOREIGN_OK too right? > > -Eric > > > The map > > +lists each extent used by files, allocation group metadata, > > +journalling logs, and static filesystem metadata, as well as any > > +regions that are unused. Each line of the listings takes the > > +following form: > > +.PP > > +.RS > > +.IR extent ": " major ":" minor " [" startblock .. endblock "]: " owner " " startoffset .. endoffset " " length > > +.PP > > +Static filesystem metadata, allocation group metadata, btrees, > > +journalling logs, and free space are marked by replacing the > > +.IR startoffset .. endoffset > > +with the appropriate marker. All blocks, offsets, and lengths are specified > > +in units of 512-byte blocks, no matter what the filesystem's block size is. > > +.BI "The optional " start " and " end " arguments can be used to constrain > > +the output to a particular range of disk blocks. > > +.RE > > +.RS 1.0i > > +.PD 0 > > +.TP > > +.BI \-n " num_extents" > > +If this option is given, > > +.B xfs_fsmap > > +obtains the extent list of the file in groups of > > +.I num_extents > > +extents. In the absence of > > +.BR \-n ", " xfs_fsmap > > +queries the system for the number of extents in the filesystem and uses that > > +value to compute the group size. > > +.TP > > +.B \-v > > +Shows verbose information. When this flag is specified, additional AG > > +specific information is appended to each line in the following form: > > +.IP > > +.RS 1.2i > > +.IR agno " (" startagblock .. endagblock ") " nblocks " " flags > > +.RE > > +.IP > > +A second > > +.B \-v > > +option will print out the > > +.I flags > > +legend. > > +.RE > > +.PD > > +.TP > > .BI "extsize [ \-R | \-D ] [ " value " ]" > > Display and/or modify the preferred extent size used when allocating > > space for the currently open file. If the > > > > -- > > To unsubscribe from this list: send the line "unsubscribe linux-xfs" in > > the body of a message to majordomo@vger.kernel.org > > More majordomo info at http://vger.kernel.org/majordomo-info.html > >