From: Stefan Behrens <sbehrens@giantdisaster.de>
To: "Lauri Võsandi" <lauri.vosandi@gmail.com>, linux-btrfs@vger.kernel.org
Subject: Re: [PATCH] btrfs-progs: receive explicit parent support
Date: Mon, 27 Apr 2015 09:28:18 +0200 [thread overview]
Message-ID: <553DE512.7080606@giantdisaster.de> (raw)
In-Reply-To: <1430043165-20641-1-git-send-email-lauri.vosandi@gmail.com>
On Sun, 26 Apr 2015 12:12:45 +0200, Lauri Võsandi wrote:
> This patch adds command-line flag -p to btrfs receive
> which makes it possible to disable automatic parent
> search for incremental snapshots and use explicitly
> specified path instead.
>
> Signed-off-by: Lauri Võsandi <lauri.vosandi@gmail.com>
> ---
> cmds-receive.c | 34 ++++++++++++++++++++++++++--------
> 1 file changed, 26 insertions(+), 8 deletions(-)
>
> diff --git a/cmds-receive.c b/cmds-receive.c
> index b7cf3f9..391d281 100644
> --- a/cmds-receive.c
> +++ b/cmds-receive.c
> @@ -61,6 +61,7 @@ struct btrfs_receive
> char *root_path;
> char *dest_dir_path; /* relative to root_path */
> char *full_subvol_path;
> + char *explicit_parent_path;
> int dest_dir_chroot;
>
> struct subvol_info *cur_subvol;
> @@ -220,20 +221,32 @@ static int process_snapshot(const char *path, const u8 *uuid, u64 ctransid,
> fprintf(stderr, "receiving snapshot %s uuid=%s, "
> "ctransid=%llu ", path, uuid_str,
> r->cur_subvol->stransid);
> - uuid_unparse(parent_uuid, uuid_str);
> - fprintf(stderr, "parent_uuid=%s, parent_ctransid=%llu\n",
> - uuid_str, parent_ctransid);
> }
>
> memset(&args_v2, 0, sizeof(args_v2));
> strncpy_null(args_v2.name, path);
>
> - parent_subvol = subvol_uuid_search(&r->sus, 0, parent_uuid,
> - parent_ctransid, NULL, subvol_search_by_received_uuid);
> - if (!parent_subvol) {
> + if (r->explicit_parent_path) {
> + if (g_verbose) {
> + fprintf(stderr, "using explicit parent %s\n",
> + r->explicit_parent_path);
> + }
> + parent_subvol = subvol_uuid_search(&r->sus, 0, NULL,
> + 0, r->explicit_parent_path, subvol_search_by_path);
This won't work if you receive more than one snapshot, each one derived
from the previous one ("btrfs send -e snap1 snap2 snap3 snap4"). You
would always use the first one as the parent, not the predecessor.
That's implemented differently for the -p option in
git://git.kernel.org/pub/scm/linux/kernel/git/arne/far-progs.git
> + } else {
> + if (g_verbose) {
> + uuid_unparse(parent_uuid, uuid_str);
> + fprintf(stderr, "parent_uuid=%s, parent_ctransid=%llu\n",
> + uuid_str, parent_ctransid);
> + }
> parent_subvol = subvol_uuid_search(&r->sus, 0, parent_uuid,
> - parent_ctransid, NULL, subvol_search_by_uuid);
> + parent_ctransid, NULL, subvol_search_by_received_uuid);
> + if (!parent_subvol) {
> + parent_subvol = subvol_uuid_search(&r->sus, 0, parent_uuid,
> + parent_ctransid, NULL, subvol_search_by_uuid);
> + }
This used to be a search for the received_uuid only. Why is this code
changed like this in the branch that is executed when -p is not
specified, in a patch that has the goal to add -p with new functionality
if -p is specified?
> }
> +
> if (!parent_subvol) {
> ret = -ENOENT;
> fprintf(stderr, "ERROR: could not find parent subvolume\n");
> @@ -962,11 +975,14 @@ int cmd_receive(int argc, char **argv)
> { NULL, 0, NULL, 0 }
> };
>
> - c = getopt_long(argc, argv, "Cevf:", long_opts, NULL);
> + c = getopt_long(argc, argv, "Cevf:p:", long_opts, NULL);
> if (c < 0)
> break;
>
> switch (c) {
> + case 'p':
> + r.explicit_parent_path = optarg;
> + break;
> case 'v':
> g_verbose++;
> break;
> @@ -1028,6 +1044,8 @@ const char * const cmd_receive_usage[] = {
> " in the data stream. Without this option,",
> " the receiver terminates only if an error",
> " is recognized or on EOF.",
> + "-p <subvol> Disables the automatic searching for parents",
> + " if incremental streams are received.",
> "-C|--chroot confine the process to <mount> using chroot",
> "--max-errors <N> Terminate as soon as N errors happened while",
> " processing commands from the send stream.",
>
next prev parent reply other threads:[~2015-04-27 7:28 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-04-26 10:12 [PATCH] btrfs-progs: receive explicit parent support Lauri Võsandi
2015-04-26 11:19 ` Hugo Mills
2015-04-26 11:37 ` lauri
2015-05-14 17:35 ` Hugo Mills
2015-05-21 18:02 ` lauri
2015-05-07 19:06 ` lauri
2015-04-27 7:28 ` Stefan Behrens [this message]
2015-04-27 9:23 ` lauri
2015-04-27 19:29 ` Stefan Behrens
2015-05-07 17:50 ` [PATCH v2] This patch adds command-line flag -p to btrfs receive which makes it possible to disable automatic parent search for incremental snapshots and use explicitly specified path instead. Works also with multiple incremental snapshots Lauri Võsandi
2015-05-07 18:53 ` [PATCH v2] btrfs-progs: receive explicit parent support Lauri Võsandi
2015-05-12 16:17 ` David Sterba
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=553DE512.7080606@giantdisaster.de \
--to=sbehrens@giantdisaster.de \
--cc=lauri.vosandi@gmail.com \
--cc=linux-btrfs@vger.kernel.org \
/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).