From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Eric Sandeen <sandeen@sandeen.net>
Cc: sandeen@redhat.com, linux-xfs@vger.kernel.org
Subject: Re: [PATCH 03/27] xfs_scrub: set up command line argument parsing
Date: Thu, 11 Jan 2018 18:03:07 -0800 [thread overview]
Message-ID: <20180112020307.GU5602@magnolia> (raw)
In-Reply-To: <e9e65539-8811-af76-ebf6-0d75cc261401@sandeen.net>
On Thu, Jan 11, 2018 at 07:30:11PM -0600, Eric Sandeen wrote:
> On 1/5/18 7:51 PM, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> >
> > Parse command line options in order to set up the context in which we
> > will scrub the filesystem.
> >
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> > scrub/common.h | 8 ++
> > scrub/xfs_scrub.c | 207 +++++++++++++++++++++++++++++++++++++++++++++++++++++
> > scrub/xfs_scrub.h | 34 +++++++++
> > 3 files changed, 249 insertions(+)
> >
> >
> > diff --git a/scrub/common.h b/scrub/common.h
> > index f620620..15a59bd 100644
> > --- a/scrub/common.h
> > +++ b/scrub/common.h
> > @@ -48,4 +48,12 @@ void __record_preen(struct scrub_ctx *ctx, const char *descr, const char *file,
> > #define str_info(ctx, str, ...) __str_info(ctx, str, __FILE__, __LINE__, __VA_ARGS__)
> > #define dbg_printf(fmt, ...) {if (debug > 1) {printf(fmt, __VA_ARGS__);}}
> >
> > +/* Is this debug tweak enabled? */
> > +static inline bool
> > +debug_tweak_on(
> > + const char *name)
> > +{
> > + return debug && getenv(name) != NULL;
>
> since it's debug anyway, I wonder if printing
> "FOO_BAR_TWEAK is on" would be useful here.
>
> > +}
> > +
> > #endif /* XFS_SCRUB_COMMON_H_ */
> > diff --git a/scrub/xfs_scrub.c b/scrub/xfs_scrub.c
> > index 10116a8..9db3b41 100644
> > --- a/scrub/xfs_scrub.c
> > +++ b/scrub/xfs_scrub.c
> > @@ -20,7 +20,12 @@
> > #include <stdio.h>
> > #include <pthread.h>
> > #include <stdbool.h>
> > +#include <stdlib.h>
> > +#include "platform_defs.h"
> > +#include "xfs.h"
> > +#include "input.h"
> > #include "xfs_scrub.h"
> > +#include "common.h"
> >
> > /*
> > * XFS Online Metadata Scrub (and Repair)
> > @@ -107,11 +112,213 @@ unsigned int debug;
> > /* Should we dump core if errors happen? */
> > bool dumpcore;
> >
> > +/* Display resource usage at the end of each phase? */
> > +bool display_rusage;
> > +
> > +/* Background mode; higher values insert more pauses between scrub calls. */
> > +unsigned int bg_mode;
> > +
> > +/* Maximum number of processors available to us. */
> > +int nproc;
> > +
> > +/* Number of threads we're allowed to use. */
> > +unsigned int nr_threads;
> > +
> > +/* Verbosity; higher values print more information. */
> > +bool verbose;
> > +
> > +/* Should we scrub the data blocks? */
> > +bool scrub_data;
> > +
> > +/* Size of a memory page. */
> > +long page_size;
> > +
> > +static void __attribute__((noreturn))
> > +usage(void)
> > +{
> > + fprintf(stderr, _("Usage: %s [OPTIONS] mountpoint\n"), progname);
> > + fprintf(stderr, _("-a:\tStop after this many errors are found.\n"));
> > + fprintf(stderr, _("-b:\tBackground mode.\n"));
> > + fprintf(stderr, _("-e:\tWhat to do if errors are found.\n"));
> > + fprintf(stderr, _("-m:\tPath to /etc/mtab.\n"));
> > + fprintf(stderr, _("-n:\tDry run. Do not modify anything.\n"));
> > + fprintf(stderr, _("-T:\tDisplay timing/usage information.\n"));
> > + fprintf(stderr, _("-v:\tVerbose output.\n"));
> > + fprintf(stderr, _("-V:\tPrint version.\n"));
> > + fprintf(stderr, _("-x:\tScrub file data too.\n"));
> > + fprintf(stderr, _("-y:\tRepair all errors.\n"));
> > +
> > + exit(16);
> > +}
> > +
> > int
> > main(
> > int argc,
> > char **argv)
> > {
> > + int c;
> > + char *mtab = NULL;
> > + char *repairstr = "";
> > + struct scrub_ctx ctx = {0};
> > + unsigned long long total_errors;
> > + bool moveon = true;
> > + static bool injected;
> > + int ret = 0;
> > +
> > fprintf(stderr, "XXX: This program is not complete!\n");
> > return 4;
> > +
> > + progname = basename(argv[0]);
> > + setlocale(LC_ALL, "");
> > + bindtextdomain(PACKAGE, LOCALEDIR);
> > + textdomain(PACKAGE);
> > +
> > + pthread_mutex_init(&ctx.lock, NULL);
> > + ctx.mode = SCRUB_MODE_DEFAULT;
> > + ctx.error_action = ERRORS_CONTINUE;
> > + while ((c = getopt(argc, argv, "a:bde:m:nTvxVy")) != EOF) {
> > + switch (c) {
> > + case 'a':
> > + ctx.max_errors = cvt_u64(optarg, 10);
> > + if (errno) {
> > + perror(optarg);
> > + usage();
> > + }
> > + break;
> > + case 'b':
> > + nr_threads = 1;
> > + bg_mode++;
> > + break;
> > + case 'd':
> > + debug++;
> > + dumpcore = true;
> > + break;
> > + case 'e':
> > + if (!strcmp("continue", optarg))
> > + ctx.error_action = ERRORS_CONTINUE;
> > + else if (!strcmp("shutdown", optarg))
> > + ctx.error_action = ERRORS_SHUTDOWN;
> > + else
> > + usage();
>
> Nothing tells me what I did wrong here,
>
> # scrub/xfs_scrub -e make_it_so /mnt/test
> Usage: xfs_scrub [OPTIONS] mountpoint
> -a: Stop after this many errors are found.
> -b: Background mode.
> -C: Print progress information to this fd.
> -e: What to do if errors are found.
> ...
>
> I told it what to do... what's wrong?
Unknown error behavior "$optarg". ?
> > + break;
> > + case 'm':
> > + mtab = optarg;
> > + break;
> > + case 'n':
> > + if (ctx.mode != SCRUB_MODE_DEFAULT) {
> > + fprintf(stderr,
> > +_("Only one of the options -n or -y may be specified.\n"));
> > + return 1;
> > + }
> > + ctx.mode = SCRUB_MODE_DRY_RUN;
> > + break;
> > + case 'T':
> > + display_rusage = true;
> > + break;
> > + case 'v':
> > + verbose = true;
> > + break;
> > + case 'V':
> > + fprintf(stdout, _("%s version %s\n"), progname,
> > + VERSION);
> > + fflush(stdout);
> > + exit(0);
> > + case 'x':
> > + scrub_data = true;
> > + break;
> > + case 'y':
> > + if (ctx.mode != SCRUB_MODE_DEFAULT) {
> > + fprintf(stderr,
> > +_("Only one of the options -n or -y may be specified.\n"));
> > + return 1;
> > + }
> > + ctx.mode = SCRUB_MODE_REPAIR;
> > + break;
> > + case '?':
>
> '?' isn't in the getopt string ...
The getopt manpage says it returns '?' for an unknown parameter, so I
provide the specific case here so that nobody can accidentally add a
second (case '?') statement.
IOWs, it's a defensive move.
> # scrub/xfs_scrub ?
> xfs_scrub: could not stat: ?: No such file or directory
>
>
> > + /* fall through */
> > + default:
> > + usage();
> > + }
> > + }
> > +
> > + /* Override thread count if debugger */
> > + if (debug_tweak_on("XFS_SCRUB_THREADS")) {
>
> can you document all these tweaks somewhere near the top in a comment?
/*
* Known debug tweaks (pass -d and set the environment variable):
* XFS_SCRUB_FORCE_ERROR -- pretend all metadata is corrupt
* XFS_SCRUB_FORCE_REPAIR -- repair all metadata even if it's ok
* XFS_SCRUB_NO_KERNEL -- pretend there is no kernel ioctl
* XFS_SCRUB_NO_SCSI_VERIFY -- disable SCSI VERIFY (if present)
* XFS_SCRUB_PHASE -- run only this scrub phase
* XFS_SCRUB_THREADS -- start exactly this number of threads
*/
> > + unsigned int x;
> > +
> > + x = cvt_u32(getenv("XFS_SCRUB_THREADS"), 10);
> > + if (errno) {
> > + perror("nr_threads");
> > + usage();
> > + }
> > + nr_threads = x;
> > + }
> > +
> > + if (optind != argc - 1)
> > + usage();
> > +
> > + ctx.mntpoint = strdup(argv[optind]);
> > +
> > + /*
> > + * If the user did not specify an explicit mount table, try to use
> > + * /proc/mounts if it is available, else /etc/mtab. We prefer
> > + * /proc/mounts because it is kernel controlled, while /etc/mtab
> > + * may contain garbage that userspace tools like pam_mounts wrote
> > + * into it.
> > + */
> > + if (!mtab) {
> > + if (access(_PATH_PROC_MOUNTS, R_OK) == 0)
> > + mtab = _PATH_PROC_MOUNTS;
> > + else
> > + mtab = _PATH_MOUNTED;
> > + }
> > +
> > + /* How many CPUs? */
> > + nproc = sysconf(_SC_NPROCESSORS_ONLN);
> > + if (nproc < 1)
> > + nproc = 1;
> > +
> > + /* Set up a page-aligned buffer for read verification. */
> > + page_size = sysconf(_SC_PAGESIZE);
> > + if (page_size < 0) {
> > + str_errno(&ctx, ctx.mntpoint);
> > + goto out;
> > + }
> > +
> > + if (debug_tweak_on("XFS_SCRUB_FORCE_REPAIR") && !injected) {
> > + ctx.mode = SCRUB_MODE_REPAIR;
> > + injected = true;
> > + }
>
> what is "injected" used for? How could it already be set?.
Not needed here, will remove.
> > +
> > + if (xfs_scrub_excessive_errors(&ctx))
> > + str_info(&ctx, ctx.mntpoint, _("Too many errors; aborting."));
>
> wait wut? oh right, you'll add $DO_STUFF above this in later patches ;)
>
>
> Rest looks ok
Ok.
--D
>
> -Eric
> --
> 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
next prev parent reply other threads:[~2018-01-12 2:03 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-01-06 1:51 [PATCH v11 00/27] xfsprogs: online scrub/repair support Darrick J. Wong
2018-01-06 1:51 ` [PATCH 01/27] xfs_scrub: create online filesystem scrub program Darrick J. Wong
2018-01-12 0:16 ` Eric Sandeen
2018-01-12 1:08 ` Darrick J. Wong
2018-01-12 1:07 ` Eric Sandeen
2018-01-12 1:10 ` Darrick J. Wong
2018-01-06 1:51 ` [PATCH 02/27] xfs_scrub: common error handling Darrick J. Wong
2018-01-12 1:15 ` Eric Sandeen
2018-01-12 1:23 ` Darrick J. Wong
2018-01-06 1:51 ` [PATCH 03/27] xfs_scrub: set up command line argument parsing Darrick J. Wong
2018-01-11 23:39 ` Eric Sandeen
2018-01-12 1:53 ` Darrick J. Wong
2018-01-12 1:30 ` Eric Sandeen
2018-01-12 2:03 ` Darrick J. Wong [this message]
2018-01-06 1:51 ` [PATCH 04/27] xfs_scrub: dispatch the various phases of the scrub program Darrick J. Wong
2018-01-06 1:51 ` [PATCH 05/27] xfs_scrub: figure out how many threads we're going to need Darrick J. Wong
2018-01-06 1:52 ` [PATCH 06/27] xfs_scrub: create an abstraction for a block device Darrick J. Wong
2018-01-11 23:24 ` Eric Sandeen
2018-01-11 23:59 ` Darrick J. Wong
2018-01-12 0:04 ` Eric Sandeen
2018-01-12 1:27 ` Darrick J. Wong
2018-01-06 1:52 ` [PATCH 07/27] xfs_scrub: find XFS filesystem geometry Darrick J. Wong
2018-01-06 1:52 ` [PATCH 08/27] xfs_scrub: add inode iteration functions Darrick J. Wong
2018-01-06 1:52 ` [PATCH 09/27] xfs_scrub: add space map " Darrick J. Wong
2018-01-06 1:52 ` [PATCH 10/27] xfs_scrub: add file " Darrick J. Wong
2018-01-11 23:19 ` Eric Sandeen
2018-01-12 0:24 ` Darrick J. Wong
2018-01-06 1:52 ` [PATCH 11/27] xfs_scrub: filesystem counter collection functions Darrick J. Wong
2018-01-06 1:52 ` [PATCH 12/27] xfs_scrub: wrap the scrub ioctl Darrick J. Wong
2018-01-11 23:12 ` Eric Sandeen
2018-01-12 0:28 ` Darrick J. Wong
2018-01-06 1:52 ` [PATCH 13/27] xfs_scrub: scan filesystem and AG metadata Darrick J. Wong
2018-01-06 1:52 ` [PATCH 14/27] xfs_scrub: thread-safe stats counter Darrick J. Wong
2018-01-06 1:53 ` [PATCH 15/27] xfs_scrub: scan inodes Darrick J. Wong
2018-01-06 1:53 ` [PATCH 16/27] xfs_scrub: check directory connectivity Darrick J. Wong
2018-01-06 1:53 ` [PATCH 17/27] xfs_scrub: warn about suspicious characters in directory/xattr names Darrick J. Wong
2018-01-06 1:53 ` [PATCH 18/27] xfs_scrub: warn about normalized Unicode name collisions Darrick J. Wong
2018-01-16 23:52 ` Eric Sandeen
2018-01-16 23:57 ` Eric Sandeen
2018-01-16 23:59 ` Darrick J. Wong
2018-01-06 1:53 ` [PATCH 19/27] xfs_scrub: create a bitmap data structure Darrick J. Wong
2018-01-06 1:53 ` [PATCH 20/27] xfs_scrub: create infrastructure to read verify data blocks Darrick J. Wong
2018-01-06 1:53 ` [PATCH 21/27] xfs_scrub: scrub file " Darrick J. Wong
2018-01-11 23:25 ` Eric Sandeen
2018-01-12 0:29 ` Darrick J. Wong
2018-01-06 1:53 ` [PATCH 22/27] xfs_scrub: optionally use SCSI READ VERIFY commands to scrub data blocks on disk Darrick J. Wong
2018-01-06 1:53 ` [PATCH 23/27] xfs_scrub: check summary counters Darrick J. Wong
2018-01-06 1:54 ` [PATCH 24/27] xfs_scrub: fstrim the free areas if there are no errors on the filesystem Darrick J. Wong
2018-01-16 22:07 ` Eric Sandeen
2018-01-16 22:23 ` Darrick J. Wong
2018-01-06 1:54 ` [PATCH 25/27] xfs_scrub: progress indicator Darrick J. Wong
2018-01-11 23:27 ` Eric Sandeen
2018-01-12 0:32 ` Darrick J. Wong
2018-01-06 1:54 ` [PATCH 26/27] xfs_scrub: create a script to scrub all xfs filesystems Darrick J. Wong
2018-01-06 1:54 ` [PATCH 27/27] xfs_scrub: integrate services with systemd Darrick J. Wong
2018-01-06 3:50 ` [PATCH 07/27] xfs_scrub: find XFS filesystem geometry Darrick J. Wong
2018-01-12 4:17 ` [PATCH v11 00/27] xfsprogs: online scrub/repair support Eric Sandeen
2018-01-17 1:31 ` Darrick J. Wong
2018-01-16 19:21 ` [PATCH 28/27] xfs_scrub: wire up repair ioctl Darrick J. Wong
2018-01-16 19:21 ` [PATCH 29/27] xfs_scrub: schedule and manage repairs to the filesystem Darrick J. Wong
-- strict thread matches above, loose matches on Subject: below --
2017-11-17 21:00 [PATCH v10 00/27] xfsprogs-4.15: online scrub/repair support Darrick J. Wong
2017-11-17 21:00 ` [PATCH 03/27] xfs_scrub: set up command line argument parsing Darrick J. Wong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180112020307.GU5602@magnolia \
--to=darrick.wong@oracle.com \
--cc=linux-xfs@vger.kernel.org \
--cc=sandeen@redhat.com \
--cc=sandeen@sandeen.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).