From: Brian Foster <bfoster@redhat.com>
To: Eric Sandeen <sandeen@redhat.com>
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH 01/13] xfs_db: Allow writes of corrupted data by optionally skipping write verifiers
Date: Thu, 19 Mar 2015 11:07:06 -0400 [thread overview]
Message-ID: <20150319150706.GA11669@laptop.bfoster> (raw)
In-Reply-To: <1426624395-8258-2-git-send-email-sandeen@redhat.com>
On Tue, Mar 17, 2015 at 03:33:03PM -0500, Eric Sandeen wrote:
> Being able to write corrupt data is handy if we wish to test
> repair against specific types of corruptions.
>
> Add a "-c" option to the write command which allows this.
>
> Note that this also skips CRC updates; it's not currently possible
> to write invalid data with a valid CRC; CRC recalculation is
> intertwined with validation.
>
> Signed-off-by: Eric Sandeen <sandeen@redhat.com>
> ---
> db/io.c | 7 +++++++
> db/io.h | 1 +
> db/write.c | 35 +++++++++++++++++++++++++++++++----
> man/man8/xfs_db.8 | 8 +++++++-
> 4 files changed, 46 insertions(+), 5 deletions(-)
>
> diff --git a/db/io.c b/db/io.c
> index 7f1b76a..c5898f1 100644
> --- a/db/io.c
> +++ b/db/io.c
> @@ -457,6 +457,13 @@ write_cur_bbs(void)
> }
>
> void
> +xfs_dummy_verify(
> + struct xfs_buf *bp)
> +{
> + return;
> +}
> +
> +void
> write_cur(void)
> {
> if (iocur_sp < 0) {
> diff --git a/db/io.h b/db/io.h
> index 71082e6..31d96b4 100644
> --- a/db/io.h
> +++ b/db/io.h
> @@ -63,6 +63,7 @@ extern void set_cur(const struct typ *t, __int64_t d, int c, int ring_add,
> bbmap_t *bbmap);
> extern void ring_add(void);
> extern void set_iocur_type(const struct typ *t);
> +extern void xfs_dummy_verify(struct xfs_buf *bp);
>
> /*
> * returns -1 for unchecked, 0 for bad and 1 for good
> diff --git a/db/write.c b/db/write.c
> index a0f14f4..4511859 100644
> --- a/db/write.c
> +++ b/db/write.c
> @@ -38,7 +38,7 @@ static int write_f(int argc, char **argv);
> static void write_help(void);
>
> static const cmdinfo_t write_cmd =
> - { "write", NULL, write_f, 0, -1, 0, N_("[field or value]..."),
> + { "write", NULL, write_f, 0, -1, 0, N_("[-c] [field or value]..."),
> N_("write value to disk"), write_help };
>
> void
> @@ -79,6 +79,7 @@ write_help(void)
> " String mode: 'write \"This_is_a_filename\" - write null terminated string.\n"
> "\n"
> " In data mode type 'write' by itself for a list of specific commands.\n\n"
> +" Specifying the -c option will allow writes of invalid (corrupt) data.\n\n"
> ));
>
> }
> @@ -90,6 +91,10 @@ write_f(
> {
> pfunc_t pf;
> extern char *progname;
> + int c;
> + int corrupt = 0; /* Allow write of corrupt data; skip verification */
> + struct xfs_buf_ops nowrite_ops;
> + const struct xfs_buf_ops *stashed_ops = NULL;
Still has whitespace damage on the line above and the if (argc) thing in
the hunk below. Can we kill those? This looks fine to me otherwise:
Reviewed-by: Brian Foster <bfoster@redhat.com>
>
> if (x.isreadonly & LIBXFS_ISREADONLY) {
> dbprintf(_("%s started in read only mode, writing disabled\n"),
> @@ -109,12 +114,34 @@ write_f(
> return 0;
> }
>
> - /* move past the "write" command */
> - argc--;
> - argv++;
> + if (argc) while ((c = getopt(argc, argv, "c")) != EOF) {
> + switch (c) {
> + case 'c':
> + corrupt = 1;
> + break;
> + default:
> + dbprintf(_("bad option for write command\n"));
> + return 0;
> + }
> + }
> +
> + argc -= optind;
> + argv += optind;
> +
> + if (corrupt) {
> + /* Temporarily remove write verifier to write bad data */
> + stashed_ops = iocur_top->bp->b_ops;
> + nowrite_ops.verify_read = stashed_ops->verify_read;
> + nowrite_ops.verify_write = xfs_dummy_verify;
> + iocur_top->bp->b_ops = &nowrite_ops;
> + dbprintf(_("Allowing write of corrupted data\n"));
> + }
>
> (*pf)(DB_WRITE, cur_typ->fields, argc, argv);
>
> + if (stashed_ops)
> + iocur_top->bp->b_ops = stashed_ops;
> +
> return 0;
> }
>
> diff --git a/man/man8/xfs_db.8 b/man/man8/xfs_db.8
> index 4d8d4ff..d527230 100644
> --- a/man/man8/xfs_db.8
> +++ b/man/man8/xfs_db.8
> @@ -711,7 +711,7 @@ and
> bits respectively, and their string equivalent reported
> (but no modifications are made).
> .TP
> -.BI "write [" "field value" "] ..."
> +.BI "write [\-c] [" "field value" "] ..."
> Write a value to disk.
> Specific fields can be set in structures (struct mode),
> or a block can be set to data values (data mode),
> @@ -729,6 +729,12 @@ contents of the block can be shifted or rotated left or right, or filled
> with a sequence, a constant value, or a random value. In this mode
> .B write
> with no arguments gives more information on the allowed commands.
> +.RS 1.0i
> +.TP 0.4i
> +.B \-c
> +Skip write verifiers and CRC recalculation; allows invalid data to be written
> +to disk.
> +.RE
> .SH TYPES
> This section gives the fields in each structure type and their meanings.
> Note that some types of block cover multiple actual structures,
> --
> 1.7.1
>
> _______________________________________________
> xfs mailing list
> xfs@oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
next prev parent reply other threads:[~2015-03-19 15:07 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-03-17 20:33 [PATCH 00/13] xfsprogs: roll-up of previously sent patches Eric Sandeen
2015-03-17 20:33 ` [PATCH 01/13] xfs_db: Allow writes of corrupted data by optionally skipping write verifiers Eric Sandeen
2015-03-19 15:07 ` Brian Foster [this message]
2015-03-23 20:00 ` [PATCH 01/13 V2] " Eric Sandeen
2015-03-24 12:07 ` Brian Foster
2015-03-17 20:33 ` [PATCH 02/13] xfs_db: fix inode CRC validity state, and warn on read if invalid Eric Sandeen
2015-03-19 15:07 ` Brian Foster
2015-03-21 2:26 ` Eric Sandeen
2015-03-21 14:18 ` Brian Foster
2015-03-23 20:11 ` [PATCH 02/13 V2] " Eric Sandeen
2015-03-24 12:07 ` Brian Foster
2015-03-17 20:33 ` [PATCH 03/13] xfs_db: add crc manipulation commands Eric Sandeen
2015-03-19 15:07 ` Brian Foster
2015-03-21 2:30 ` Eric Sandeen
2015-03-21 14:18 ` Brian Foster
2015-03-23 20:01 ` [PATCH 03/13 DROP] " Eric Sandeen
2015-03-17 20:33 ` [PATCH 04/13] xfs_db: nlink fields are valid for di_version == 3, too Eric Sandeen
2015-03-17 20:33 ` [PATCH 05/13] xfs_repair: dirty inode in process_sf_dir2 if we change namelen Eric Sandeen
2015-03-17 20:33 ` [PATCH 06/13] xfs_repair: remove impossible tests in process_sf_dir2 Eric Sandeen
2015-03-17 20:33 ` [PATCH 07/13] xfs_repair: collapse 2 cases " Eric Sandeen
2015-03-17 20:33 ` [PATCH 08/13] xfs_repair: remove last-entry hack " Eric Sandeen
2015-03-17 20:33 ` [PATCH 09/13] libxfs: remove ASSERT on ftype read from disk Eric Sandeen
2015-03-19 16:46 ` Brian Foster
2015-03-19 17:27 ` Eric Sandeen
2015-03-23 20:13 ` PATCH 09/13 V2] " Eric Sandeen
2015-03-24 12:07 ` Brian Foster
2015-03-17 20:33 ` [PATCH 10/13] xfs_repair: clear need_root_dotdot if we rebuild the root dir Eric Sandeen
2015-03-17 20:33 ` [PATCH 11/13] xfs_repair: set *parent if process_dir2_data() fixes root inode Eric Sandeen
2015-03-17 20:33 ` [PATCH 12/13] xfs_repair: don't clear . or .. in process_dir2_data Eric Sandeen
2015-03-19 16:47 ` Brian Foster
2015-03-19 17:29 ` Eric Sandeen
2015-03-19 17:54 ` Brian Foster
2015-03-19 17:59 ` Eric Sandeen
2015-03-19 18:07 ` Brian Foster
2015-03-23 20:17 ` [PATCH 12/13 V2] " Eric Sandeen
2015-03-24 12:07 ` Brian Foster
2015-03-17 20:33 ` [PATCH 13/13] xfs_repair: validate & fix inode CRCs Eric Sandeen
2015-03-19 16:47 ` Brian Foster
2015-03-23 20:19 ` [PATCH 13/13 V2] " Eric Sandeen
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=20150319150706.GA11669@laptop.bfoster \
--to=bfoster@redhat.com \
--cc=sandeen@redhat.com \
--cc=xfs@oss.sgi.com \
/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