All of lore.kernel.org
 help / color / mirror / Atom feed
From: Benjamin Marzinski <bmarzins@redhat.com>
To: Martin Wilck <martin.wilck@suse.com>
Cc: Christophe Varoqui <christophe.varoqui@opensvc.com>,
	dm-devel@lists.linux.dev
Subject: Re: [PATCH 23/44] libmultipath: improve dm_get_wwid() return value logic
Date: Wed, 10 Jul 2024 19:34:56 -0400	[thread overview]
Message-ID: <Zo8aoNebaxrgKlzo@redhat.com> (raw)
In-Reply-To: <20240709213935.177028-24-mwilck@suse.com>

On Tue, Jul 09, 2024 at 11:39:14PM +0200, Martin Wilck wrote:
> Make dm_get_wwid() return different status codes for non-existing maps,
> maps that exists but are not multipath maps, and generic error case,
> and handle these return codes appropriately in callers.
> 
> Signed-off-by: Martin Wilck <mwilck@suse.com>
> ---
>  libmultipath/alias.c     |  5 +++--
>  libmultipath/configure.c | 23 +++++++++++------------
>  libmultipath/devmapper.c | 21 ++++++++++++++++-----
>  libmultipath/wwids.c     |  2 +-
>  tests/alias.c            | 10 +++++-----
>  5 files changed, 36 insertions(+), 25 deletions(-)
> 
> diff --git a/libmultipath/alias.c b/libmultipath/alias.c
> index 10e58a7..2ab9499 100644
> --- a/libmultipath/alias.c
> +++ b/libmultipath/alias.c
> @@ -408,13 +408,14 @@ static bool alias_already_taken(const char *alias, const char *map_wwid)
>  {
>  
>  	char wwid[WWID_SIZE];
> +	int rc = dm_get_wwid(alias, wwid, sizeof(wwid));
>  
>  	/* If the map doesn't exist, it's fine */
> -	if (dm_get_wwid(alias, wwid, sizeof(wwid)) != 0)

We used to assume the alias was not taken if dm_get_wwid() failed with
an error. Now we assume that the alias is taken. I think we should
continue to assume we can use the alias if we get DM_ERR.

> +	if (rc == DMP_NOT_FOUND)
>  		return false;
>  
>  	/* If both the name and the wwid match, it's fine.*/
> -	if (strncmp(map_wwid, wwid, sizeof(wwid)) == 0)
> +	if (rc == DMP_OK && strncmp(map_wwid, wwid, sizeof(wwid)) == 0)
>  		return false;
>  
>  	condlog(3, "%s: alias '%s' already taken, reselecting alias",
> diff --git a/libmultipath/configure.c b/libmultipath/configure.c
> index 666d4e8..565ea5c 100644
> --- a/libmultipath/configure.c
> +++ b/libmultipath/configure.c
> @@ -845,18 +845,17 @@ int domap(struct multipath *mpp, char *params, int is_daemon)
>  
>  	if (mpp->action == ACT_CREATE && dm_map_present(mpp->alias)) {
>  		char wwid[WWID_SIZE];
> +		int rc = dm_get_wwid(mpp->alias, wwid, sizeof(wwid));
>  
> -		if (dm_get_wwid(mpp->alias, wwid, sizeof(wwid)) == 0) {
> -			if (!strncmp(mpp->wwid, wwid, sizeof(wwid))) {
> -				condlog(3, "%s: map already present",
> -					mpp->alias);
> -				mpp->action = ACT_RELOAD;
> -			} else {
> -				condlog(0, "%s: map \"%s\" already present with WWID %s, skipping",
> -					mpp->wwid, mpp->alias, wwid);
> -				condlog(0, "please check alias settings in config and bindings file");
> -				mpp->action = ACT_REJECT;
> -			}
> +		if (rc == DMP_OK && !strncmp(mpp->wwid, wwid, sizeof(wwid))) {
> +			condlog(3, "%s: map already present",
> +				mpp->alias);
> +			mpp->action = ACT_RELOAD;

If we return DMP_NO_MATCH, we'll tell the user that the device is already
present will a WWID of "". This is the same as before, but perhaps it
would be better to tell the user that the alias is in use by a
non-multipath device.

> +		} else if (rc == DMP_OK || rc == DMP_NO_MATCH) {
> +			condlog(1, "%s: map \"%s\" already present with WWID \"%s\", skipping\n"
> +				   "please check alias settings in config and bindings file",
> +				mpp->wwid, mpp->alias, wwid);
> +			mpp->action = ACT_REJECT;
>  		}
>  	}
>  
> @@ -1320,7 +1319,7 @@ static int _get_refwwid(enum mpath_cmds cmd, const char *dev,
>  		break;
>  
>  	case DEV_DEVMAP:
> -		if (((dm_get_wwid(dev, tmpwwid, WWID_SIZE)) == 0)
> +		if (((dm_get_wwid(dev, tmpwwid, WWID_SIZE)) == DMP_OK)
>  		    && (strlen(tmpwwid)))
>  			refwwid = tmpwwid;
>  
> diff --git a/libmultipath/devmapper.c b/libmultipath/devmapper.c
> index 94ef369..1eebcb5 100644
> --- a/libmultipath/devmapper.c
> +++ b/libmultipath/devmapper.c
> @@ -829,19 +829,30 @@ int dm_get_map(const char *name, unsigned long long *size, char **outparams)
>  	}
>  }
>  
> +/**
> + * dm_get_wwid(): return WWID for a multipath map
> + * @returns:
> + *    DMP_OK if successful
> + *    DMP_NOT_FOUND if the map doesn't exist
> + *    DMP_NO_MATCH if the map exists but is not a multipath map
> + *    DMP_ERR for other errors
> + */
>  int dm_get_wwid(const char *name, char *uuid, int uuid_len)
>  {
>  	char tmp[DM_UUID_LEN];
> +	int rc = dm_get_dm_uuid(name, tmp);
>  
> -	if (dm_get_dm_uuid(name, tmp) != DMP_OK)
> -		return 1;
> +	if (rc != DMP_OK)
> +		return rc;
>  
>  	if (!strncmp(tmp, UUID_PREFIX, UUID_PREFIX_LEN))
>  		strlcpy(uuid, tmp + UUID_PREFIX_LEN, uuid_len);
> -	else
> +	else {

This seems a little inconsistent with the change in our returns.
Before, if we found a device but it wasn't multipath device,
dm_get_wwid() would exit with 0 but set uuid to an empty string. Now we
return DMP_NO_MATCH in this case but we still clear uuid. It seems like
perhaps we should either set uuid to an empty string for all cases
except DM_OK or we should not touch it for all cases except DM_OK.
Alternatively, perhaps I'm over-thinking this.

-Ben 

>  		uuid[0] = '\0';
> +		return DMP_NO_MATCH;
> +	}
>  
> -	return 0;
> +	return DMP_OK;
>  }
>  
>  static int is_mpath_part(const char *part_name, const char *map_name)
> @@ -1377,7 +1388,7 @@ struct multipath *dm_get_multipath(const char *name)
>  	if (dm_get_map(name, &mpp->size, NULL) != DMP_OK)
>  		goto out;
>  
> -	if (dm_get_wwid(name, mpp->wwid, WWID_SIZE) != 0)
> +	if (dm_get_wwid(name, mpp->wwid, WWID_SIZE) != DMP_OK)
>  		condlog(2, "%s: failed to get uuid for %s", __func__, name);
>  	if (dm_get_info(name, &mpp->dmi) != 0)
>  		condlog(2, "%s: failed to get info for %s", __func__, name);
> diff --git a/libmultipath/wwids.c b/libmultipath/wwids.c
> index 7a4cb74..aac18c0 100644
> --- a/libmultipath/wwids.c
> +++ b/libmultipath/wwids.c
> @@ -295,7 +295,7 @@ should_multipath(struct path *pp1, vector pathvec, vector mpvec)
>  		struct multipath *mp = find_mp_by_wwid(mpvec, pp1->wwid);
>  
>  		if (mp != NULL &&
> -		    dm_get_wwid(mp->alias, tmp_wwid, WWID_SIZE) == 0 &&
> +		    dm_get_wwid(mp->alias, tmp_wwid, WWID_SIZE) == DMP_OK &&
>  		    !strncmp(tmp_wwid, pp1->wwid, WWID_SIZE)) {
>  			condlog(3, "wwid %s is already multipathed, keeping it",
>  				pp1->wwid);
> diff --git a/tests/alias.c b/tests/alias.c
> index 1f78656..a95b308 100644
> --- a/tests/alias.c
> +++ b/tests/alias.c
> @@ -84,7 +84,7 @@ int __wrap_dm_get_wwid(const char *name, char *uuid, int uuid_len)
>  	check_expected(uuid_len);
>  	assert_non_null(uuid);
>  	ret = mock_type(int);
> -	if (ret == 0)
> +	if (ret == DMP_OK)
>  		strcpy(uuid, mock_ptr_type(char *));
>  	return ret;
>  }
> @@ -438,14 +438,14 @@ static void mock_unused_alias(const char *alias)
>  {
>  	expect_string(__wrap_dm_get_wwid, name, alias);
>  	expect_value(__wrap_dm_get_wwid, uuid_len, WWID_SIZE);
> -	will_return(__wrap_dm_get_wwid, 1);
> +	will_return(__wrap_dm_get_wwid, DMP_NOT_FOUND);
>  }
>  
>  static void mock_self_alias(const char *alias, const char *wwid)
>  {
>  	expect_string(__wrap_dm_get_wwid, name, alias);
>  	expect_value(__wrap_dm_get_wwid, uuid_len, WWID_SIZE);
> -	will_return(__wrap_dm_get_wwid, 0);
> +	will_return(__wrap_dm_get_wwid, DMP_OK);
>  	will_return(__wrap_dm_get_wwid, wwid);
>  }
>  
> @@ -471,14 +471,14 @@ static void mock_self_alias(const char *alias, const char *wwid)
>  	do {								\
>  		expect_string(__wrap_dm_get_wwid, name, alias);		\
>  		expect_value(__wrap_dm_get_wwid, uuid_len, WWID_SIZE);	\
> -		will_return(__wrap_dm_get_wwid, 1);			\
> +		will_return(__wrap_dm_get_wwid, DMP_NOT_FOUND);		\
>  	} while (0)
>  
>  #define mock_used_alias(alias, wwid)					\
>  	do {								\
>  		expect_string(__wrap_dm_get_wwid, name, alias);		\
>  		expect_value(__wrap_dm_get_wwid, uuid_len, WWID_SIZE);	\
> -		will_return(__wrap_dm_get_wwid, 0);			\
> +		will_return(__wrap_dm_get_wwid, DMP_OK);		\
>  		will_return(__wrap_dm_get_wwid, "WWID_USED");		\
>  		expect_condlog(3, USED_STR(alias, wwid));		\
>  	} while(0)
> -- 
> 2.45.2


  reply	other threads:[~2024-07-10 23:35 UTC|newest]

Thread overview: 75+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-07-09 21:38 [PATCH 00/44] multipath-tools: devmapper API refactored Martin Wilck
2024-07-09 21:38 ` [PATCH 01/44] multipath-tools CI: more fixes for arm/v7 Martin Wilck
2024-07-10  7:06   ` Martin Wilck
2024-07-09 21:38 ` [PATCH 02/44] multipath-tools CI: fix dmevents test for Debian Sid, arm/v7 Martin Wilck
2024-07-09 21:38 ` [PATCH 03/44] create-config.mk: use printf instead of /bin/echo Martin Wilck
2024-07-09 21:38 ` [PATCH 04/44] multipathd.service.in: use @BINDIR@ instead of /sbin Martin Wilck
2024-07-09 21:38 ` [PATCH 05/44] Makefile.inc: replace @BINDIR@ with $(TGTDIR)/$(bindir) Martin Wilck
2024-07-09 21:38 ` [PATCH 06/44] kpartx.rules: use @BINDIR@ to locate kpartx Martin Wilck
2024-07-10 16:30   ` Benjamin Marzinski
2024-07-10 20:15     ` Martin Wilck
2024-07-09 21:38 ` [PATCH 07/44] multipath-tools: Remove hard-coded paths to executables Martin Wilck
2024-07-09 21:38 ` [PATCH 08/44] multipath-tools: compile_commands.json fixes Martin Wilck
2024-07-09 21:39 ` [PATCH 09/44] multipath-tools: .gitignore: ignore o.wrap files for CI helpers Martin Wilck
2024-07-09 21:39 ` [PATCH 10/44] libmultipath: remove unused includes in devmapper.h Martin Wilck
2024-07-09 21:39 ` [PATCH 11/44] libmultipath: use DM_DEVICE_INFO in dm_mapname() Martin Wilck
2024-07-09 21:39 ` [PATCH 12/44] multipath-tools: don't call dm_task_no_open_count() Martin Wilck
2024-07-09 21:39 ` [PATCH 13/44] libmpathutil: export cleanup_udev_device() Martin Wilck
2024-07-09 21:39 ` [PATCH 14/44] libmpathutil: add cleanup_vector() Martin Wilck
2024-07-09 21:39 ` [PATCH 15/44] libmultipath: add cleanup helpers for struct multipath Martin Wilck
2024-07-09 21:39 ` [PATCH 16/44] libmultipath: add cleanup_dm_task(), and use it in devmapper.c Martin Wilck
2024-07-10 20:12   ` Benjamin Marzinski
2024-07-10 20:18     ` Martin Wilck
2024-07-09 21:39 ` [PATCH 17/44] libmultipath: add libmp_mapinfo() Martin Wilck
2024-07-10 22:53   ` Benjamin Marzinski
2024-07-11 11:00     ` Martin Wilck
2024-07-09 21:39 ` [PATCH 18/44] libmultipath tests: add tests for libmp_mapinfo() Martin Wilck
2024-07-09 21:39 ` [PATCH 19/44] libmultipath: implement dm_get_info() and dm_map_present() with new API Martin Wilck
2024-07-09 21:39 ` [PATCH 20/44] libmultipath: remove dm_get_prefixed_uuid() Martin Wilck
2024-07-09 21:39 ` [PATCH 21/44] libmultipath: is_mpath_part(): improve parsing Martin Wilck
2024-07-10 22:54   ` Benjamin Marzinski
2024-07-11 11:08     ` Martin Wilck
2024-07-11 17:01       ` Benjamin Marzinski
2024-07-11 17:41         ` Martin Wilck
2024-07-09 21:39 ` [PATCH 22/44] libmultipath: rename dm_get_uuid() -> dm_get_wwid() Martin Wilck
2024-07-09 21:39 ` [PATCH 23/44] libmultipath: improve dm_get_wwid() return value logic Martin Wilck
2024-07-10 23:34   ` Benjamin Marzinski [this message]
2024-07-11 12:25     ` Martin Wilck
2024-07-11 16:38       ` Benjamin Marzinski
2024-07-11 22:00       ` Martin Wilck
2024-07-09 21:39 ` [PATCH 24/44] libmultipath: reimplement dm_map_name() with new API Martin Wilck
2024-07-10 23:41   ` Benjamin Marzinski
2024-07-09 21:39 ` [PATCH 25/44] libmultipath: reimplement dm_map_present_by_uuid() Martin Wilck
2024-07-09 21:39 ` [PATCH 26/44] libmultipath: reimplement dm_get_opencount() with new API Martin Wilck
2024-07-09 21:39 ` [PATCH 27/44] libmpathpersist: skip redundant dm_map_present() call Martin Wilck
2024-07-09 21:39 ` [PATCH 28/44] libmultipath: implement dm_is_mpath() with new API Martin Wilck
2024-07-11  0:21   ` Benjamin Marzinski
2024-07-11 12:32     ` Martin Wilck
2024-07-09 21:39 ` [PATCH 29/44] libmultipath: implement dm_get_multipath() " Martin Wilck
2024-07-09 21:39 ` [PATCH 30/44] libmultipath: use libmp_mapinfo() in _dm_flush_map() Martin Wilck
2024-07-09 21:39 ` [PATCH 31/44] libmultipath: add is_mpath_uuid() helper Martin Wilck
2024-07-11  3:38   ` Benjamin Marzinski
2024-07-11 12:56     ` Martin Wilck
2024-07-11 18:39       ` Benjamin Marzinski
2024-07-11 18:59         ` Martin Wilck
2024-07-09 21:39 ` [PATCH 32/44] libmultipath: add is_mpath_part_uuid() helper Martin Wilck
2024-07-09 21:39 ` [PATCH 33/44] libmultipath: add dmp_errstr() helper Martin Wilck
2024-07-09 21:39 ` [PATCH 34/44] libmultipath: use libmp_mapinfo() in do_foreach_partmaps() Martin Wilck
2024-07-09 21:39 ` [PATCH 35/44] libmultipath: use libmp_pathinfo() in update_multipath_table() Martin Wilck
2024-07-11  5:16   ` Benjamin Marzinski
2024-07-11 15:29     ` Martin Wilck
2024-07-09 21:39 ` [PATCH 36/44] libmultipath: update mpp->dmi " Martin Wilck
2024-07-09 21:39 ` [PATCH 37/44] libmultipath: drop extra call to dm_map_present() in domap() Martin Wilck
2024-07-09 21:39 ` [PATCH 38/44] libmultipath: split off update_multipath_table__() Martin Wilck
2024-07-09 21:39 ` [PATCH 39/44] multipath: implement check_usable_paths() with libmp_pathinfo() Martin Wilck
2024-07-11  5:38   ` Benjamin Marzinski
2024-07-09 21:39 ` [PATCH 40/44] multipathd: implement add_map_without_path() with libmp_mapinfo() Martin Wilck
2024-07-11  6:11   ` Benjamin Marzinski
2024-07-09 21:39 ` [PATCH 41/44] libmultipath: simplify dm_get_maps() Martin Wilck
2024-07-09 21:39 ` [PATCH 42/44] llibmultipath: fix return code check for dm_is_suspended() Martin Wilck
2024-07-11  6:27   ` Benjamin Marzinski
2024-07-11 15:35     ` Martin Wilck
2024-07-09 21:39 ` [PATCH 43/44] libmpathpersist: use libmp_mapinfo() in get_mpvec() Martin Wilck
2024-07-11  6:43   ` Benjamin Marzinski
2024-07-09 21:39 ` [PATCH 44/44] libmpathpersist: use mpp->alias in do_mpath_persistent_reserve_out() Martin Wilck
2024-07-11  6:55 ` [PATCH 00/44] multipath-tools: devmapper API refactored Benjamin Marzinski

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=Zo8aoNebaxrgKlzo@redhat.com \
    --to=bmarzins@redhat.com \
    --cc=christophe.varoqui@opensvc.com \
    --cc=dm-devel@lists.linux.dev \
    --cc=martin.wilck@suse.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.