rpms/device-mapper-multipath
7
0
Fork 0
Code Issues Pull Requests Releases Activity

import device-mapper-multipath-0.8.4-11.el8

Browse Source
This commit is contained in:
CentOS Sources 2021-04-06 12:27:46 +00:00 committed by Andrew Lukoshko
commit d35fa1930e
67 changed files with 8870 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.device-mapper-multipath.metadata Normal file
View File

@ -0,0 +1 @@
b52c2be340449664f0a122070838f6d8edd42e4a SOURCES/multipath-tools-0.8.4.tgz

1
.gitignore vendored Normal file
View File

@ -0,0 +1 @@
SOURCES/multipath-tools-0.8.4.tgz

42
SOURCES/0001-libmultipath-assign-variable-to-make-gcc-happy.patch Normal file
View File

@ -0,0 +1,42 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue, 17 Mar 2020 17:28:24 -0500
Subject: [PATCH] libmultipath: assign variable to make gcc happy
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
There is nothing wrong with is_queueing not being set at the start
of __set_no_path_retry(), it will always get set before it is accessed,
but gcc 8.2.1 is failing with
structs_vec.c: In function ‘__set_no_path_retry’:
structs_vec.c:339:7: error: ‘is_queueing’ may be used uninitialized in
this function [-Werror=maybe-uninitialized]
bool is_queueing;
^~~~~~~~~~~
so, assign a value to make it happy.
Reviewed-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/structs_vec.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libmultipath/structs_vec.c b/libmultipath/structs_vec.c
index 3dbbaa0f..077f2e42 100644
--- a/libmultipath/structs_vec.c
+++ b/libmultipath/structs_vec.c
@@ -336,7 +336,7 @@ static void leave_recovery_mode(struct multipath *mpp)
void __set_no_path_retry(struct multipath *mpp, bool check_features)
{
- bool is_queueing;
+ bool is_queueing = false; /* assign a value to make gcc happy */
check_features = check_features && mpp->features != NULL;
if (check_features)
--
2.17.2

66
SOURCES/0002-libmutipath-don-t-close-fd-on-dm_lib_release.patch Normal file
View File

@ -0,0 +1,66 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Sat, 21 Mar 2020 23:49:59 -0500
Subject: [PATCH] libmutipath: don't close fd on dm_lib_release
If dm_hold_control_open() isn't set, when dm_lib_release() is called, it
will close the control fd. The control fd will get re-opened on the next
dm_task_run() call, but if there is a dm_task_run() call already
in progress in another thread, it can fail. Since many of the
device-mapper callouts happen with the vecs lock held, this wasn't too
noticeable, but there is code that calls dm_task_run() without the
vecs lock held, notably the dmevent waiter code.
Since, as Martin pointed out, dm_hold_control_open() hasn't always
existed in libdevmapper, check if it's supported on compilation,
and update the version requirements if so.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/Makefile | 4 ++++
libmultipath/devmapper.c | 7 ++++++-
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/libmultipath/Makefile b/libmultipath/Makefile
index e5651e49..ad690a49 100644
--- a/libmultipath/Makefile
+++ b/libmultipath/Makefile
@@ -36,6 +36,10 @@ ifneq ($(call check_func,dm_task_deferred_remove,/usr/include/libdevmapper.h),0)
CFLAGS += -DLIBDM_API_DEFERRED
endif
+ifneq ($(call check_func,dm_hold_control_dev,/usr/include/libdevmapper.h),0)
+ CFLAGS += -DLIBDM_API_HOLD_CONTROL
+endif
+
OBJS = memory.o parser.o vector.o devmapper.o callout.o \
hwtable.o blacklist.o util.o dmparser.o config.o \
structs.o discovery.o propsel.o dict.o \
diff --git a/libmultipath/devmapper.c b/libmultipath/devmapper.c
index bed8ddc6..13a1cf53 100644
--- a/libmultipath/devmapper.c
+++ b/libmultipath/devmapper.c
@@ -108,7 +108,9 @@ dm_lib_prereq (void)
{
char version[64];
int v[3];
-#if defined(LIBDM_API_DEFERRED)
+#if defined(LIBDM_API_HOLD_CONTROL)
+ int minv[3] = {1, 2, 111};
+#elif defined(LIBDM_API_DEFERRED)
int minv[3] = {1, 2, 89};
#elif defined(DM_SUBSYSTEM_UDEV_FLAG0)
int minv[3] = {1, 2, 82};
@@ -254,6 +256,9 @@ void libmp_dm_init(void)
memcpy(conf->version, version, sizeof(version));
put_multipath_config(conf);
dm_init(verbosity);
+#ifdef LIBDM_API_HOLD_CONTROL
+ dm_hold_control_dev(1);
+#endif
dm_udev_set_sync_support(libmp_dm_udev_sync);
}
--
2.17.2

64
SOURCES/0003-libmultipath-allow-force-reload-with-no-active-paths.patch Normal file
View File

@ -0,0 +1,64 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 19 Mar 2020 22:17:51 -0500
Subject: [PATCH] libmultipath: allow force reload with no active paths
If the partition information has changed on multipath devices (say,
because it was updated on another node that has access to the same
storage), users expect that running "multipathd reconfigure" will update
that. However, if the checkers for the multipath device are pending for
too long when the the device is reconfigured, multipathd will give up
waiting for them, and refuse to reload the device, since there are no
active paths. This means that no kpartx update will be triggered.
Multipath is fully capable of reloading a multipath device that has no
active paths. This has been possible for years. If multipath is supposed
to reload the device, it should do so, even if there are no active paths.
Generally, when multipath is force reloaded, kpartx will be updated.
However when a device is reloaded with no paths, the udev rules won't
run kpartx. But they also weren't running kpartx when the first valid
path appeared, even though the dm activation rules get run in this case.
This changes 11-dm-mpath.rules to run kpartx when a device goes from no
usable paths to having usable paths.
Reviewed-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/configure.c | 6 ------
multipath/11-dm-mpath.rules | 2 +-
2 files changed, 1 insertion(+), 7 deletions(-)
diff --git a/libmultipath/configure.c b/libmultipath/configure.c
index c95848a0..96c79610 100644
--- a/libmultipath/configure.c
+++ b/libmultipath/configure.c
@@ -710,12 +710,6 @@ select_action (struct multipath * mpp, vector curmp, int force_reload)
return;
}
- if (pathcount(mpp, PATH_UP) == 0) {
- mpp->action = ACT_IMPOSSIBLE;
- condlog(3, "%s: set ACT_IMPOSSIBLE (no usable path)",
- mpp->alias);
- return;
- }
if (force_reload) {
mpp->force_udev_reload = 1;
mpp->action = ACT_RELOAD;
diff --git a/multipath/11-dm-mpath.rules b/multipath/11-dm-mpath.rules
index 07320a14..cd522e8c 100644
--- a/multipath/11-dm-mpath.rules
+++ b/multipath/11-dm-mpath.rules
@@ -75,7 +75,7 @@ ENV{MPATH_DEVICE_READY}=="0", ENV{DM_UDEV_DISABLE_OTHER_RULES_FLAG}="1"
ENV{MPATH_DEVICE_READY}!="0", ENV{.MPATH_DEVICE_READY_OLD}=="0",\
ENV{DM_UDEV_DISABLE_OTHER_RULES_FLAG}="$env{DM_DISABLE_OTHER_RULES_FLAG_OLD}",\
ENV{DM_DISABLE_OTHER_RULES_FLAG_OLD}="",\
- ENV{DM_ACTIVATION}="1"
+ ENV{DM_ACTIVATION}="1", ENV{MPATH_UNCHANGED}="0"
# The code to check multipath state ends here. We need to set
# properties and symlinks regardless whether the map is usable or
--
2.17.2

31
SOURCES/0004-libmpathpersist-depend-on-libmultipath.patch Normal file
View File

@ -0,0 +1,31 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Christian Hesse <mail@eworm.de>
Date: Wed, 6 May 2020 09:35:47 +0200
Subject: [PATCH] libmpathpersist: depend on libmultipath
Without this the build fails with:
/usr/bin/ld: cannot find -lmultipath
Signed-off-by: Christian Hesse <mail@eworm.de>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
Makefile | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/Makefile b/Makefile
index 1dee3680..ba1d73ba 100644
--- a/Makefile
+++ b/Makefile
@@ -28,7 +28,7 @@ all: $(BUILDDIRS)
$(BUILDDIRS):
$(MAKE) -C $@
-multipath multipathd mpathpersist: libmultipath
+libmpathpersist multipath multipathd mpathpersist: libmultipath
mpathpersist: libmpathpersist
$(BUILDDIRS.clean):
--
2.17.2

34
SOURCES/0005-multipath-tools-Makefile-more-dependency-fixes-for-p.patch Normal file
View File

@ -0,0 +1,34 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Martin Wilck <mwilck@suse.com>
Date: Tue, 12 May 2020 00:39:21 +0200
Subject: [PATCH] multipath-tools: Makefile: more dependency fixes for parallel
build
Extend the late fixes from Christian.
Cc: Christian Hesse <mail@eworm.de>
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
Makefile | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/Makefile b/Makefile
index ba1d73ba..fec3b73b 100644
--- a/Makefile
+++ b/Makefile
@@ -28,8 +28,9 @@ all: $(BUILDDIRS)
$(BUILDDIRS):
$(MAKE) -C $@
-libmpathpersist multipath multipathd mpathpersist: libmultipath
-mpathpersist: libmpathpersist
+libmultipath libdmmp: libmpathcmd
+libmpathpersist multipath multipathd: libmultipath
+mpathpersist multipathd: libmpathpersist
$(BUILDDIRS.clean):
$(MAKE) -C ${@:.clean=} clean
--
2.17.2

33
SOURCES/0006-multipath-tools-Makefile.inc-set-Wno-error-clobbered.patch Normal file
View File

@ -0,0 +1,33 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Martin Wilck <mwilck@suse.com>
Date: Tue, 12 May 2020 00:39:24 +0200
Subject: [PATCH] multipath-tools: Makefile.inc: set -Wno-error=clobbered
We need to ignore -Wclobbered because gcc has trouble dealing with glibc's
implementation of pthread_cleanup_push().
For some variants of gcc, -Wno-clobbered alone isn't enough if -Werror is also
set. Compilation with -Wno-error=clobbered works, though.
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
Makefile.inc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/Makefile.inc b/Makefile.inc
index d4d1e0dd..9060ac9b 100644
--- a/Makefile.inc
+++ b/Makefile.inc
@@ -91,7 +91,7 @@ TEST_CC_OPTION = $(shell \
STACKPROT := $(call TEST_CC_OPTION,-fstack-protector-strong,-fstack-protector)
ERROR_DISCARDED_QUALIFIERS := $(call TEST_CC_OPTION,-Werror=discarded-qualifiers,)
-WNOCLOBBERED := $(call TEST_CC_OPTION,-Wno-clobbered,)
+WNOCLOBBERED := $(call TEST_CC_OPTION,-Wno-clobbered -Wno-error=clobbered,)
OPTFLAGS = -O2 -g -pipe -Werror -Wall -Wextra -Wformat=2 -Werror=implicit-int \
-Werror=implicit-function-declaration -Werror=format-security \
--
2.17.2

91
SOURCES/0007-libmultipath-discovery.c-use-z-qualifier-for-size_t.patch Normal file
View File

@ -0,0 +1,91 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Martin Wilck <mwilck@suse.com>
Date: Tue, 12 May 2020 00:39:25 +0200
Subject: [PATCH] libmultipath: discovery.c: use %z qualifier for size_t
Otherwise compilation for 32bit targets spits out warnings.
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/discovery.c | 16 ++++++++--------
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/libmultipath/discovery.c b/libmultipath/discovery.c
index ee3290cd..ffec5162 100644
--- a/libmultipath/discovery.c
+++ b/libmultipath/discovery.c
@@ -986,7 +986,7 @@ parse_vpd_pg80(const unsigned char *in, char *out, size_t out_len)
}
if (len >= out_len) {
- condlog(2, "vpd pg80 overflow, %lu/%lu bytes required",
+ condlog(2, "vpd pg80 overflow, %zu/%zu bytes required",
len + 1, out_len);
len = out_len - 1;
}
@@ -1087,7 +1087,7 @@ parse_vpd_pg83(const unsigned char *in, size_t in_len,
len = sprintf(out, "%d", vpd_type);
if (2 * vpd_len >= out_len - len) {
- condlog(1, "%s: WWID overflow, type %d, %lu/%lu bytes required",
+ condlog(1, "%s: WWID overflow, type %d, %zu/%zu bytes required",
__func__, vpd_type,
2 * vpd_len + len + 1, out_len);
vpd_len = (out_len - len - 1) / 2;
@@ -1096,7 +1096,7 @@ parse_vpd_pg83(const unsigned char *in, size_t in_len,
len += sprintf(out + len,
"%02x", vpd[i]);
} else if (vpd_type == 0x8 && vpd_len < 4) {
- condlog(1, "%s: VPD length %lu too small for designator type 8",
+ condlog(1, "%s: VPD length %zu too small for designator type 8",
__func__, vpd_len);
return -EINVAL;
} else if (vpd_type == 0x8) {
@@ -1112,7 +1112,7 @@ parse_vpd_pg83(const unsigned char *in, size_t in_len,
while (len > 2 && vpd[len - 2] == '\0')
--len;
if (len > out_len - 1) {
- condlog(1, "%s: WWID overflow, type 8/%c, %lu/%lu bytes required",
+ condlog(1, "%s: WWID overflow, type 8/%c, %zu/%zu bytes required",
__func__, out[0], len + 1, out_len);
len = out_len - 1;
}
@@ -1136,7 +1136,7 @@ parse_vpd_pg83(const unsigned char *in, size_t in_len,
while ((p = memchr(vpd, ' ', vpd_len))) {
p_len = p - vpd;
if (len + p_len > out_len - 1) {
- condlog(1, "%s: WWID overflow, type 1, %lu/%lu bytes required",
+ condlog(1, "%s: WWID overflow, type 1, %zu/%zu bytes required",
__func__, len + p_len, out_len);
p_len = out_len - len - 1;
}
@@ -1162,7 +1162,7 @@ parse_vpd_pg83(const unsigned char *in, size_t in_len,
p_len = vpd_len;
if (p_len > 0 && len < out_len - 1) {
if (len + p_len > out_len - 1) {
- condlog(1, "%s: WWID overflow, type 1, %lu/%lu bytes required",
+ condlog(1, "%s: WWID overflow, type 1, %zu/%zu bytes required",
__func__, len + p_len + 1, out_len);
p_len = out_len - len - 1;
}
@@ -1186,14 +1186,14 @@ parse_vpd_c0_hp3par(const unsigned char *in, size_t in_len,
memset(out, 0x0, out_len);
if (in_len <= 4 || (in[4] > 3 && in_len < 44)) {
- condlog(3, "HP/3PAR vendor specific VPD page length too short: %lu", in_len);
+ condlog(3, "HP/3PAR vendor specific VPD page length too short: %zu", in_len);
return -EINVAL;
}
if (in[4] <= 3) /* revision must be > 3 to have Vomlume Name */
return -ENODATA;
len = get_unaligned_be32(&in[40]);
if (len > out_len || len + 44 > in_len) {
- condlog(3, "HP/3PAR vendor specific Volume name too long: %lu",
+ condlog(3, "HP/3PAR vendor specific Volume name too long: %zu",
len);
return -EINVAL;
}
--
2.17.2

185
SOURCES/0008-libmultipath-eliminate-more-signed-unsigned-comparis.patch Normal file
View File

@ -0,0 +1,185 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Martin Wilck <mwilck@suse.com>
Date: Tue, 12 May 2020 00:39:26 +0200
Subject: [PATCH] libmultipath: eliminate more signed/unsigned comparisons
Fix some more compiler warnings about signed/unsigned comparison.
I've observed these only on 32bit builds, therefore they went unnoticed
before.
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/print.c | 12 ++++++------
libmultipath/prioritizers/alua_spc3.h | 2 +-
multipathd/cli_handlers.c | 20 ++++++++++----------
multipathd/main.c | 2 +-
4 files changed, 18 insertions(+), 18 deletions(-)
diff --git a/libmultipath/print.c b/libmultipath/print.c
index b944ef32..298b3764 100644
--- a/libmultipath/print.c
+++ b/libmultipath/print.c
@@ -1958,25 +1958,25 @@ char *snprint_config(const struct config *conf, int *len,
}
c = reply + snprint_defaults(conf, reply, maxlen);
- if ((c - reply) == maxlen)
+ if (c == reply + maxlen)
continue;
c += snprint_blacklist(conf, c, reply + maxlen - c);
- if ((c - reply) == maxlen)
+ if (c == reply + maxlen)
continue;
c += snprint_blacklist_except(conf, c, reply + maxlen - c);
- if ((c - reply) == maxlen)
+ if (c == reply + maxlen)
continue;
c += snprint_hwtable(conf, c, reply + maxlen - c,
hwtable ? hwtable : conf->hwtable);
- if ((c - reply) == maxlen)
+ if (c == reply + maxlen)
continue;
c += snprint_overrides(conf, c, reply + maxlen - c,
conf->overrides);
- if ((c - reply) == maxlen)
+ if (c == reply + maxlen)
continue;
if (VECTOR_SIZE(conf->mptable) > 0 ||
@@ -1984,7 +1984,7 @@ char *snprint_config(const struct config *conf, int *len,
c += snprint_mptable(conf, c, reply + maxlen - c,
mpvec);
- if ((c - reply) < maxlen) {
+ if (c < reply + maxlen) {
if (len)
*len = c - reply;
return reply;
diff --git a/libmultipath/prioritizers/alua_spc3.h b/libmultipath/prioritizers/alua_spc3.h
index 18b495ef..7ba2cf4c 100644
--- a/libmultipath/prioritizers/alua_spc3.h
+++ b/libmultipath/prioritizers/alua_spc3.h
@@ -284,7 +284,7 @@ struct rtpg_data {
#define RTPG_FOR_EACH_PORT_GROUP(p, g) \
for( \
g = &(p->data[0]); \
- (((char *) g) - ((char *) p)) < get_unaligned_be32(p->length); \
+ ((char *) g) < ((char *) p) + get_unaligned_be32(p->length); \
g = (struct rtpg_tpg_dscr *) ( \
((char *) g) + \
sizeof(struct rtpg_tpg_dscr) + \
diff --git a/multipathd/cli_handlers.c b/multipathd/cli_handlers.c
index 7d878c88..31c3d9fd 100644
--- a/multipathd/cli_handlers.c
+++ b/multipathd/cli_handlers.c
@@ -66,7 +66,7 @@ show_paths (char ** r, int * len, struct vectors * vecs, char * style,
c += snprint_foreign_paths(c, reply + maxlen - c,
style, pretty);
- again = ((c - reply) == (maxlen - 1));
+ again = (c == reply + maxlen - 1);
REALLOC_REPLY(reply, again, maxlen);
}
@@ -102,7 +102,7 @@ show_path (char ** r, int * len, struct vectors * vecs, struct path *pp,
c += snprint_path(c, reply + maxlen - c, style, pp, 0);
- again = ((c - reply) == (maxlen - 1));
+ again = (c == reply + maxlen - 1);
REALLOC_REPLY(reply, again, maxlen);
}
@@ -131,7 +131,7 @@ show_map_topology (char ** r, int * len, struct multipath * mpp,
c = reply;
c += snprint_multipath_topology(c, reply + maxlen - c, mpp, 2);
- again = ((c - reply) == (maxlen - 1));
+ again = (c == reply + maxlen - 1);
REALLOC_REPLY(reply, again, maxlen);
}
@@ -171,7 +171,7 @@ show_maps_topology (char ** r, int * len, struct vectors * vecs)
}
c += snprint_foreign_topology(c, reply + maxlen - c, 2);
- again = ((c - reply) == (maxlen - 1));
+ again = (c == reply + maxlen - 1);
REALLOC_REPLY(reply, again, maxlen);
}
@@ -209,7 +209,7 @@ show_maps_json (char ** r, int * len, struct vectors * vecs)
c = reply;
c += snprint_multipath_topology_json(c, maxlen, vecs);
- again = ((c - reply) == maxlen);
+ again = (c == reply + maxlen);
REALLOC_REPLY(reply, again, maxlen);
}
@@ -238,7 +238,7 @@ show_map_json (char ** r, int * len, struct multipath * mpp,
c = reply;
c += snprint_multipath_map_json(c, maxlen, mpp);
- again = ((c - reply) == maxlen);
+ again = (c == reply + maxlen);
REALLOC_REPLY(reply, again, maxlen);
}
@@ -487,7 +487,7 @@ show_map (char ** r, int *len, struct multipath * mpp, char * style,
c += snprint_multipath(c, reply + maxlen - c, style,
mpp, pretty);
- again = ((c - reply) == (maxlen - 1));
+ again = (c == reply + maxlen - 1);
REALLOC_REPLY(reply, again, maxlen);
}
@@ -533,7 +533,7 @@ show_maps (char ** r, int *len, struct vectors * vecs, char * style,
}
c += snprint_foreign_multipaths(c, reply + maxlen - c,
style, pretty);
- again = ((c - reply) == (maxlen - 1));
+ again = (c == reply + maxlen - 1);
REALLOC_REPLY(reply, again, maxlen);
}
@@ -1297,7 +1297,7 @@ show_blacklist (char ** r, int * len)
c = reply;
c += snprint_blacklist_report(conf, c, maxlen);
- again = ((c - reply) == maxlen);
+ again = (c == reply + maxlen);
REALLOC_REPLY(reply, again, maxlen);
}
pthread_cleanup_pop(1);
@@ -1339,7 +1339,7 @@ show_devices (char ** r, int * len, struct vectors *vecs)
c = reply;
c += snprint_devices(conf, c, maxlen, vecs);
- again = ((c - reply) == maxlen);
+ again = (c == reply + maxlen);
REALLOC_REPLY(reply, again, maxlen);
}
pthread_cleanup_pop(1);
diff --git a/multipathd/main.c b/multipathd/main.c
index 8baf9abe..6b7db2c0 100644
--- a/multipathd/main.c
+++ b/multipathd/main.c
@@ -2374,7 +2374,7 @@ checkerloop (void *ap)
conf = get_multipath_config();
max_checkint = conf->max_checkint;
put_multipath_config(conf);
- if (diff_time.tv_sec > max_checkint)
+ if (diff_time.tv_sec > (time_t)max_checkint)
condlog(1, "path checkers took longer "
"than %lu seconds, consider "
"increasing max_polling_interval",
--
2.17.2

49
SOURCES/0009-libmultipath-set_uint-fix-parsing-for-32bit.patch Normal file
View File

@ -0,0 +1,49 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Martin Wilck <mwilck@suse.com>
Date: Tue, 12 May 2020 00:39:27 +0200
Subject: [PATCH] libmultipath: set_uint: fix parsing for 32bit
On architectures where sizeof(long) == sizeof(int), the code wouldn't
work as intended. Use strtoul instead. As strtoul happily parses
negative numbers as input, require the number to begin with a digit.
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/dict.c | 11 +++++++----
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/libmultipath/dict.c b/libmultipath/dict.c
index 3e25e74f..0e9ea387 100644
--- a/libmultipath/dict.c
+++ b/libmultipath/dict.c
@@ -60,19 +60,22 @@ static int
set_uint(vector strvec, void *ptr)
{
unsigned int *uint_ptr = (unsigned int *)ptr;
- char *buff, *eptr;
- long res;
+ char *buff, *eptr, *p;
+ unsigned long res;
int rc;
buff = set_value(strvec);
if (!buff)
return 1;
- res = strtol(buff, &eptr, 10);
+ p = buff;
+ while (isspace(*p))
+ p++;
+ res = strtoul(p, &eptr, 10);
if (eptr > buff)
while (isspace(*eptr))
eptr++;
- if (*buff == '\0' || *eptr != '\0' || res < 0 || res > UINT_MAX) {
+ if (*buff == '\0' || *eptr != '\0' || !isdigit(*p) || res > UINT_MAX) {
condlog(1, "%s: invalid value for %s: \"%s\"",
__func__, (char*)VECTOR_SLOT(strvec, 0), buff);
rc = 1;
--
2.17.2

33
SOURCES/0010-multipath-tools-Makefile-add-install-dependency.patch Normal file
View File

@ -0,0 +1,33 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Martin Wilck <mwilck@suse.com>
Date: Tue, 12 May 2020 22:38:22 +0200
Subject: [PATCH] multipath-tools Makefile: add install dependency
$(libdir) must exist before running "make install" on prioritizer, checker,
and foreign libraries.
Cc: Christian Hesse <mail@eworm.de>
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
Makefile | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/Makefile b/Makefile
index fec3b73b..8bcaba66 100644
--- a/Makefile
+++ b/Makefile
@@ -32,6 +32,10 @@ libmultipath libdmmp: libmpathcmd
libmpathpersist multipath multipathd: libmultipath
mpathpersist multipathd: libmpathpersist
+libmultipath/checkers.install \
+ libmultipath/prioritizers.install \
+ libmultipath/foreign.install: libmultipath.install
+
$(BUILDDIRS.clean):
$(MAKE) -C ${@:.clean=} clean
--
2.17.2

66
SOURCES/0011-RH-fixup-udev-rules-for-redhat.patch Normal file
View File

@ -0,0 +1,66 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 13 Apr 2017 07:22:23 -0500
Subject: [PATCH] RH: fixup udev rules for redhat
The multipath rules need to run after scsi_id is run. This means moving
them after 60-persistent-storage.rules for redhat. Redhat also uses a
different naming scheme for partitions than SuSE.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
Makefile.inc | 2 +-
kpartx/kpartx.rules | 2 +-
multipath/Makefile | 4 ++--
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/Makefile.inc b/Makefile.inc
index 9060ac9b..034752d9 100644
--- a/Makefile.inc
+++ b/Makefile.inc
@@ -53,7 +53,7 @@ endif
prefix =
exec_prefix = $(prefix)
usr_prefix = $(prefix)
-bindir = $(exec_prefix)/sbin
+bindir = $(exec_prefix)/usr/sbin
libudevdir = $(prefix)/$(SYSTEMDPATH)/udev
udevrulesdir = $(libudevdir)/rules.d
multipathdir = $(TOPDIR)/libmultipath
diff --git a/kpartx/kpartx.rules b/kpartx/kpartx.rules
index 8f990494..8a3a1718 100644
--- a/kpartx/kpartx.rules
+++ b/kpartx/kpartx.rules
@@ -32,6 +32,6 @@ LABEL="mpath_kpartx_end"
GOTO="kpartx_end"
LABEL="run_kpartx"
-RUN+="/sbin/kpartx -un -p -part /dev/$name"
+RUN+="/sbin/kpartx -un /dev/$name"
LABEL="kpartx_end"
diff --git a/multipath/Makefile b/multipath/Makefile
index 0828a8f7..b9bbb3cf 100644
--- a/multipath/Makefile
+++ b/multipath/Makefile
@@ -24,7 +24,7 @@ install:
$(INSTALL_PROGRAM) -m 755 $(EXEC) $(DESTDIR)$(bindir)/
$(INSTALL_PROGRAM) -d $(DESTDIR)$(udevrulesdir)
$(INSTALL_PROGRAM) -m 644 11-dm-mpath.rules $(DESTDIR)$(udevrulesdir)
- $(INSTALL_PROGRAM) -m 644 $(EXEC).rules $(DESTDIR)$(libudevdir)/rules.d/56-multipath.rules
+ $(INSTALL_PROGRAM) -m 644 $(EXEC).rules $(DESTDIR)$(libudevdir)/rules.d/62-multipath.rules
$(INSTALL_PROGRAM) -d $(DESTDIR)$(man8dir)
$(INSTALL_PROGRAM) -m 644 $(EXEC).8.gz $(DESTDIR)$(man8dir)
$(INSTALL_PROGRAM) -d $(DESTDIR)$(man5dir)
@@ -33,7 +33,7 @@ install:
uninstall:
$(RM) $(DESTDIR)$(bindir)/$(EXEC)
$(RM) $(DESTDIR)$(udevrulesdir)/11-dm-mpath.rules
- $(RM) $(DESTDIR)$(libudevdir)/rules.d/56-multipath.rules
+ $(RM) $(DESTDIR)$(libudevdir)/rules.d/62-multipath.rules
$(RM) $(DESTDIR)$(man8dir)/$(EXEC).8.gz
$(RM) $(DESTDIR)$(man5dir)/$(EXEC).conf.5.gz
--
2.17.2

105
SOURCES/0012-RH-Remove-the-property-blacklist-exception-builtin.patch Normal file
View File

@ -0,0 +1,105 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Wed, 2 Jul 2014 12:49:53 -0500
Subject: [PATCH] RH: Remove the property blacklist exception builtin
Multipath set the default property blacklist exceptions to
(ID_SCSI_VPD|ID_WWN). This has the effect of blacklisting some internal
devices. These devices may never have multiple paths, but it is nice
to be able to set multipath up on them all the same. This patch simply
removes the default, and makes it so that if no property
blacklist_exception is given, then devices aren't failed for not matching
it.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/blacklist.c | 9 ++-------
multipath/multipath.conf.5 | 11 ++++++-----
tests/blacklist.c | 6 ++----
3 files changed, 10 insertions(+), 16 deletions(-)
diff --git a/libmultipath/blacklist.c b/libmultipath/blacklist.c
index 00e8dbdb..d9691b17 100644
--- a/libmultipath/blacklist.c
+++ b/libmultipath/blacklist.c
@@ -204,12 +204,6 @@ setup_default_blist (struct config * conf)
if (store_ble(conf->blist_devnode, str, ORIGIN_DEFAULT))
return 1;
- str = STRDUP("(SCSI_IDENT_|ID_WWN)");
- if (!str)
- return 1;
- if (store_ble(conf->elist_property, str, ORIGIN_DEFAULT))
- return 1;
-
vector_foreach_slot (conf->hwtable, hwe, i) {
if (hwe->bl_product) {
if (find_blacklist_device(conf->blist_device,
@@ -411,7 +405,8 @@ filter_property(struct config *conf, struct udev_device *udev, int lvl,
*uid_attribute != '\0';
bool uid_attr_seen = false;
- r = MATCH_PROPERTY_BLIST_MISSING;
+ if (VECTOR_SIZE(conf->elist_property))
+ r = MATCH_PROPERTY_BLIST_MISSING;
udev_list_entry_foreach(list_entry,
udev_device_get_properties_list_entry(udev)) {
diff --git a/multipath/multipath.conf.5 b/multipath/multipath.conf.5
index 05a5e8ff..3455b1cc 100644
--- a/multipath/multipath.conf.5
+++ b/multipath/multipath.conf.5
@@ -1286,9 +1286,14 @@ keywords. Both are regular expressions. For a full description of these keywords
Regular expression for an udev property. All
devices that have matching udev properties will be excluded/included.
The handling of the \fIproperty\fR keyword is special,
-because devices \fBmust\fR have at least one whitelisted udev property;
+because if a property blacklist_exception is set, devices \fBmust\fR have at
+least one whitelisted udev property;
otherwise they're treated as blacklisted, and the message
"\fIblacklisted, udev property missing\fR" is displayed in the logs.
+For example, setting the property blacklist_exception to
+\fB(SCSI_IDENT_|ID_WWN)\fR, will cause well-behaved SCSI devices and devices
+that provide a WWN (World Wide Number) to be included, and all others to be
+excluded. This works to exclude most non-multipathable devices.
.
.RS
.PP
@@ -1299,10 +1304,6 @@ Blacklisting by missing properties is only applied to devices which do have the
property specified by \fIuid_attribute\fR (e.g. \fIID_SERIAL\fR)
set. Previously, it was applied to every device, possibly causing devices to be
blacklisted because of temporary I/O error conditions.
-.PP
-The default \fIblacklist exception\fR is: \fB(SCSI_IDENT_|ID_WWN)\fR, causing
-well-behaved SCSI devices and devices that provide a WWN (World Wide Number)
-to be included, and all others to be excluded.
.RE
.TP
.B protocol
diff --git a/tests/blacklist.c b/tests/blacklist.c
index 6e7c1864..cc8a9a4a 100644
--- a/tests/blacklist.c
+++ b/tests/blacklist.c
@@ -271,7 +271,7 @@ static void test_property_missing(void **state)
conf.blist_property = blist_property_wwn;
expect_condlog(3, "sdb: blacklisted, udev property missing\n");
assert_int_equal(filter_property(&conf, &udev, 3, "ID_SERIAL"),
- MATCH_PROPERTY_BLIST_MISSING);
+ MATCH_NOTHING);
assert_int_equal(filter_property(&conf, &udev, 3, "ID_BLAH"),
MATCH_NOTHING);
assert_int_equal(filter_property(&conf, &udev, 3, ""),
@@ -363,9 +363,7 @@ static void test_filter_path_missing1(void **state)
conf.blist_device = blist_device_foo_bar;
conf.blist_protocol = blist_protocol_fcp;
conf.blist_wwid = blist_wwid_xyzzy;
- expect_condlog(3, "sdb: blacklisted, udev property missing\n");
- assert_int_equal(filter_path(&conf, &miss1_pp),
- MATCH_PROPERTY_BLIST_MISSING);
+ assert_int_equal(filter_path(&conf, &miss1_pp), MATCH_NOTHING);
}
/* This one matches the property whitelist, to test the other missing
--
2.17.2

106
SOURCES/0013-RH-don-t-start-without-a-config-file.patch Normal file
View File

@ -0,0 +1,106 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Wed, 15 Oct 2014 10:39:30 -0500
Subject: [PATCH] RH: don't start without a config file
If /etc/multipath.conf doesn't exist, don't start multipathd and blacklist
all devices when running multipath. A completely blank configuration file
is almost never what users want. Also, people may have the multipath
packages installed but don't want to use them. This patch provides a
simple way to disable multipath. Simply removing or renaming
/etc/multipath.conf will keep multipath from doing anything.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/config.c | 15 +++++++++++++++
libmultipath/config.h | 1 +
multipath/multipath.rules | 1 +
multipathd/multipathd.8 | 2 ++
multipathd/multipathd.service | 1 +
5 files changed, 20 insertions(+)
diff --git a/libmultipath/config.c b/libmultipath/config.c
index b4d87689..b36778b0 100644
--- a/libmultipath/config.c
+++ b/libmultipath/config.c
@@ -26,6 +26,7 @@
#include "devmapper.h"
#include "mpath_cmd.h"
#include "propsel.h"
+#include "version.h"
static int
hwe_strmatch (const struct hwentry *hwe1, const struct hwentry *hwe2)
@@ -778,6 +779,20 @@ load_config (char * file)
goto out;
}
factorize_hwtable(conf->hwtable, builtin_hwtable_size, file);
+ } else {
+ condlog(0, "/etc/multipath.conf does not exist, blacklisting all devices.");
+ if (conf->blist_devnode == NULL) {
+ conf->blist_devnode = vector_alloc();
+ if (!conf->blist_devnode) {
+ condlog(0, "cannot allocate blacklist\n");
+ goto out;
+ }
+ }
+ if (store_ble(conf->blist_devnode, strdup(".*"),
+ ORIGIN_NO_CONFIG)) {
+ condlog(0, "cannot store default no-config blacklist\n");
+ goto out;
+ }
}
conf->processed_main_config = 1;
diff --git a/libmultipath/config.h b/libmultipath/config.h
index ceecff2d..3368d8c9 100644
--- a/libmultipath/config.h
+++ b/libmultipath/config.h
@@ -9,6 +9,7 @@
#define ORIGIN_DEFAULT 0
#define ORIGIN_CONFIG 1
+#define ORIGIN_NO_CONFIG 2
/*
* In kernel, fast_io_fail == 0 means immediate failure on rport delete.
diff --git a/multipath/multipath.rules b/multipath/multipath.rules
index 9df11a95..0486bf70 100644
--- a/multipath/multipath.rules
+++ b/multipath/multipath.rules
@@ -9,6 +9,7 @@ IMPORT{cmdline}="nompath"
ENV{nompath}=="?*", GOTO="end_mpath"
IMPORT{cmdline}="multipath"
ENV{multipath}=="off", GOTO="end_mpath"
+TEST!="/etc/multipath.conf", GOTO="end_mpath"
ENV{DEVTYPE}!="partition", GOTO="test_dev"
IMPORT{parent}="DM_MULTIPATH_DEVICE_PATH"
diff --git a/multipathd/multipathd.8 b/multipathd/multipathd.8
index 048a838d..8bd47a80 100644
--- a/multipathd/multipathd.8
+++ b/multipathd/multipathd.8
@@ -39,6 +39,8 @@ map regains its maximum performance and redundancy.
This daemon executes the external \fBmultipath\fR tool when events occur.
In turn, the multipath tool signals the multipathd daemon when it is done with
devmap reconfiguration, so that it can refresh its failed path list.
+
+In this Linux distribution, multipathd does not run unless a /etc/multipath.conf file exists.
.
.
.\" ----------------------------------------------------------------------------
diff --git a/multipathd/multipathd.service b/multipathd/multipathd.service
index ba24983e..17434cef 100644
--- a/multipathd/multipathd.service
+++ b/multipathd/multipathd.service
@@ -4,6 +4,7 @@ Wants=systemd-udev-trigger.service systemd-udev-settle.service
Before=iscsi.service iscsid.service lvm2-activation-early.service
Before=local-fs-pre.target blk-availability.service
After=multipathd.socket systemd-udev-trigger.service systemd-udev-settle.service
+ConditionPathExists=/etc/multipath.conf
DefaultDependencies=no
Conflicts=shutdown.target
ConditionKernelCommandLine=!nompath
--
2.17.2

64
SOURCES/0014-RH-use-rpm-optflags-if-present.patch Normal file
View File

@ -0,0 +1,64 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Wed, 19 Apr 2017 06:10:01 -0500
Subject: [PATCH] RH: use rpm optflags if present
Use the passed in optflags when compiling as an RPM, and keep the
default flags as close as possible to the current fedora flags, while
still being generic.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
Makefile.inc | 29 +++++++++++++++++++++--------
1 file changed, 21 insertions(+), 8 deletions(-)
diff --git a/Makefile.inc b/Makefile.inc
index 034752d9..c2abd301 100644
--- a/Makefile.inc
+++ b/Makefile.inc
@@ -89,16 +89,29 @@ TEST_CC_OPTION = $(shell \
echo "$(2)"; \
fi)
-STACKPROT := $(call TEST_CC_OPTION,-fstack-protector-strong,-fstack-protector)
ERROR_DISCARDED_QUALIFIERS := $(call TEST_CC_OPTION,-Werror=discarded-qualifiers,)
WNOCLOBBERED := $(call TEST_CC_OPTION,-Wno-clobbered -Wno-error=clobbered,)
+ifndef RPM_OPT_FLAGS
+ STACKPROT := $(call TEST_CC_OPTION,-fstack-protector-strong,-fstack-protector)
+ OPTFLAGS = -O2 -g -pipe -Wall -Werror=format-security \
+ -Wp,-D_FORTIFY_SOURCE=2 -fexceptions \
+ $(STACKPROT) -grecord-gcc-switches \
+ -fasynchronous-unwind-tables
+ ifeq ($(shell test -f /usr/lib/rpm/redhat/redhat-hardened-cc1 && echo 1),1)
+ OPTFLAGS += -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1
+ endif
+ ifeq ($(shell test -f /usr/lib/rpm/redhat/redhat-annobin-cc1 && echo 1),1)
+ OPTFLAGS += -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1
+ endif
+else
+ OPTFLAGS = $(RPM_OPT_FLAGS)
+endif
+OPTFLAGS += -Werror -Wextra -Wstrict-prototypes -Wformat=2 \
+ -Werror=implicit-int -Werror=implicit-function-declaration \
+ $(WNOCLOBBERED) \
+ -Werror=cast-qual $(ERROR_DISCARDED_QUALIFIERS) \
+ --param=ssp-buffer-size=4
-OPTFLAGS = -O2 -g -pipe -Werror -Wall -Wextra -Wformat=2 -Werror=implicit-int \
- -Werror=implicit-function-declaration -Werror=format-security \
- $(WNOCLOBBERED) \
- -Werror=cast-qual $(ERROR_DISCARDED_QUALIFIERS) \
- $(STACKPROT) --param=ssp-buffer-size=4
-CPPFLAGS := -Wp,-D_FORTIFY_SOURCE=2
CFLAGS := $(OPTFLAGS) -DBIN_DIR=\"$(bindir)\" -DLIB_STRING=\"${LIB}\" -DRUN_DIR=\"${RUN}\" \
-MMD -MP $(CFLAGS)
BIN_CFLAGS = -fPIE -DPIE
@@ -135,4 +148,4 @@ check_file = $(shell \
%.o: %.c
@echo building $@ because of $?
- $(CC) $(CFLAGS) $(CPPFLAGS) -c -o $@ $<
+ $(CC) $(CFLAGS) -c -o $@ $<
--
2.17.2

774
SOURCES/0015-RH-add-mpathconf.patch Normal file
View File

@ -0,0 +1,774 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 16 Oct 2014 15:49:01 -0500
Subject: [PATCH] RH: add mpathconf
mpathconf is a program (largely based on lvmcomf) to help users
configure /etc/multipath.conf and enable or disable multipathing. It
has a couple of built-in options that can be set directly from the
command line. But, mostly it is used to get a multipath.conf file
with the OS defaults, and to enable and disable multipathing via
a single command.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/config.c | 2 +
multipath/Makefile | 5 +
multipath/mpathconf | 555 ++++++++++++++++++++++++++++++++++++++++++
multipath/mpathconf.8 | 135 ++++++++++
4 files changed, 697 insertions(+)
create mode 100644 multipath/mpathconf
create mode 100644 multipath/mpathconf.8
diff --git a/libmultipath/config.c b/libmultipath/config.c
index b36778b0..26f8e050 100644
--- a/libmultipath/config.c
+++ b/libmultipath/config.c
@@ -781,6 +781,8 @@ load_config (char * file)
factorize_hwtable(conf->hwtable, builtin_hwtable_size, file);
} else {
condlog(0, "/etc/multipath.conf does not exist, blacklisting all devices.");
+ condlog(0, "You can run \"/sbin/mpathconf --enable\" to create");
+ condlog(0, "/etc/multipath.conf. See man mpathconf(8) for more details");
if (conf->blist_devnode == NULL) {
conf->blist_devnode = vector_alloc();
if (!conf->blist_devnode) {
diff --git a/multipath/Makefile b/multipath/Makefile
index b9bbb3cf..e720c7f6 100644
--- a/multipath/Makefile
+++ b/multipath/Makefile
@@ -18,10 +18,12 @@ $(EXEC): $(OBJS) $(multipathdir)/libmultipath.so $(mpathcmddir)/libmpathcmd.so
$(CC) $(CFLAGS) $(OBJS) -o $(EXEC) $(LDFLAGS) $(LIBDEPS)
$(GZIP) $(EXEC).8 > $(EXEC).8.gz
$(GZIP) $(EXEC).conf.5 > $(EXEC).conf.5.gz
+ $(GZIP) mpathconf.8 > mpathconf.8.gz
install:
$(INSTALL_PROGRAM) -d $(DESTDIR)$(bindir)
$(INSTALL_PROGRAM) -m 755 $(EXEC) $(DESTDIR)$(bindir)/
+ $(INSTALL_PROGRAM) -m 755 mpathconf $(DESTDIR)$(bindir)/
$(INSTALL_PROGRAM) -d $(DESTDIR)$(udevrulesdir)
$(INSTALL_PROGRAM) -m 644 11-dm-mpath.rules $(DESTDIR)$(udevrulesdir)
$(INSTALL_PROGRAM) -m 644 $(EXEC).rules $(DESTDIR)$(libudevdir)/rules.d/62-multipath.rules
@@ -29,13 +31,16 @@ install:
$(INSTALL_PROGRAM) -m 644 $(EXEC).8.gz $(DESTDIR)$(man8dir)
$(INSTALL_PROGRAM) -d $(DESTDIR)$(man5dir)
$(INSTALL_PROGRAM) -m 644 $(EXEC).conf.5.gz $(DESTDIR)$(man5dir)
+ $(INSTALL_PROGRAM) -m 644 mpathconf.8.gz $(DESTDIR)$(man8dir)
uninstall:
$(RM) $(DESTDIR)$(bindir)/$(EXEC)
$(RM) $(DESTDIR)$(udevrulesdir)/11-dm-mpath.rules
$(RM) $(DESTDIR)$(libudevdir)/rules.d/62-multipath.rules
+ $(RM) $(DESTDIR)$(bindir)/mpathconf
$(RM) $(DESTDIR)$(man8dir)/$(EXEC).8.gz
$(RM) $(DESTDIR)$(man5dir)/$(EXEC).conf.5.gz
+ $(RM) $(DESTDIR)$(man8dir)/mpathconf.8.gz
clean: dep_clean
$(RM) core *.o $(EXEC) *.gz
diff --git a/multipath/mpathconf b/multipath/mpathconf
new file mode 100644
index 00000000..f34003c9
--- /dev/null
+++ b/multipath/mpathconf
@@ -0,0 +1,555 @@
+#!/bin/bash
+#
+# Copyright (C) 2010 Red Hat, Inc. All rights reserved.
+#
+# This file is part of the device-mapper-multipath package.
+#
+# This copyrighted material is made available to anyone wishing to use,
+# modify, copy, or redistribute it subject to the terms and conditions
+# of the GNU General Public License v.2.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+
+#
+# Simple editting of /etc/multipath.conf
+# This program was largely ripped off from lvmconf
+#
+
+unset ENABLE FIND FRIENDLY PROPERTY FOREIGN MODULE MULTIPATHD HAVE_DISABLE HAVE_WWID_DISABLE HAVE_FIND HAVE_BLACKLIST HAVE_EXCEPTIONS HAVE_DEFAULTS HAVE_FRIENDLY HAVE_PROPERTY HAVE_FOREIGN HAVE_MULTIPATHD HAVE_MODULE HAVE_OUTFILE SHOW_STATUS CHANGED_CONFIG WWID_LIST
+
+DEFAULT_CONFIG="# device-mapper-multipath configuration file
+
+# For a complete list of the default configuration values, run either:
+# # multipath -t
+# or
+# # multipathd show config
+
+# For a list of configuration options with descriptions, see the
+# multipath.conf man page.
+
+defaults {
+ user_friendly_names yes
+ find_multipaths yes
+ enable_foreign \"^$\"
+}
+
+blacklist_exceptions {
+ property \"(SCSI_IDENT_|ID_WWN)\"
+}"
+
+CONFIGFILE="/etc/multipath.conf"
+OUTPUTFILE="/etc/multipath.conf"
+MULTIPATHDIR="/etc/multipath"
+TMPFILE="/etc/multipath/.multipath.conf.tmp"
+WWIDS=0
+
+function usage
+{
+ echo "usage: $0 <command>"
+ echo ""
+ echo "Commands:"
+ echo "Enable: --enable "
+ echo "Disable: --disable"
+ echo "Only allow certain wwids (instead of enable): --allow <WWID>"
+ echo "Set user_friendly_names (Default y): --user_friendly_names <y|n>"
+ echo "Set find_multipaths (Default y): --find_multipaths <y|n>"
+ echo "Set default property blacklist (Default y): --property_blacklist <y|n>"
+ echo "Set enable_foreign to show foreign devices (Default n): --enable_foreign <y|n>"
+ echo "Load the dm-multipath modules on enable (Default y): --with_module <y|n>"
+ echo "start/stop/reload multipathd (Default n): --with_multipathd <y|n>"
+ echo "select output file (Default /etc/multipath.conf): --outfile <FILE>"
+ echo ""
+}
+
+function add_wwid
+{
+ INDEX=0
+ while [ "$INDEX" -lt "$WWIDS" ] ; do
+ if [ "$1" = "${WWID_LIST[$INDEX]}" ] ; then
+ return
+ fi
+ ((INDEX++))
+ done
+ WWID_LIST[$WWIDS]="$1"
+ ((WWIDS++))
+}
+
+function get_dm_deps
+{
+ shift 3
+ while [ -n "$1" -a -n "$2" ]; do
+ MAJOR=$(echo $1 | tr -d '(,')
+ MINOR=$(echo $2 | tr -d ')')
+ UUID=`dmsetup info -c --noheadings -o uuid -j $MAJOR -m $MINOR 2> /dev/null`
+ if [ -n "$UUID" ] ; then
+ set_dm_wwid $UUID
+ fi
+ shift 2
+ done
+}
+
+function set_dm_wwid
+{
+ if [[ "$1" =~ ^part[[:digit:]]+-mpath- ]] ; then
+ add_wwid "${1##part*-mpath-}"
+ elif [[ "$1" =~ ^mpath- ]] ; then
+ add_wwid "${1##mpath-}"
+ else
+ get_dm_deps `dmsetup deps -u $1`
+ fi
+}
+
+function set_wwid
+{
+ UUID=""
+ if [[ "$1" =~ ^[[:digit:]]+:[[:digit:]]+$ ]] ; then
+ MAJOR=${1%%:*}
+ MINOR=${1##*:}
+ UUID=`dmsetup info -c --noheadings -o uuid -j $MAJOR -m $MINOR 2> /dev/null`
+ else
+ UUID=`dmsetup info -c --noheadings -o uuid $1 2> /dev/null`
+ fi
+ if [ -n "$UUID" ] ; then
+ set_dm_wwid $UUID
+ else
+ add_wwid "$1"
+ fi
+}
+
+function parse_args
+{
+ while [ -n "$1" ]; do
+ case $1 in
+ --enable)
+ ENABLE=1
+ shift
+ ;;
+ --disable)
+ ENABLE=0
+ shift
+ ;;
+ --allow)
+ ENABLE=2
+ if [ -n "$2" ]; then
+ set_wwid $2
+ shift 2
+ else
+ usage
+ exit 1
+ fi
+ ;;
+ --user_friendly_names)
+ if [ -n "$2" ]; then
+ FRIENDLY=$2
+ shift 2
+ else
+ usage
+ exit 1
+ fi
+ ;;
+ --find_multipaths)
+ if [ -n "$2" ]; then
+ FIND=$2
+ shift 2
+ else
+ usage
+ exit 1
+ fi
+ ;;
+ --property_blacklist)
+ if [ -n "$2" ]; then
+ PROPERTY=$2
+ shift 2
+ else
+ usage
+ exit 1
+ fi
+ ;;
+ --enable_foreign)
+ if [ -n "$2" ]; then
+ FOREIGN=$2
+ shift 2
+ else
+ usage
+ exit 1
+ fi
+ ;;
+ --with_module)
+ if [ -n "$2" ]; then
+ MODULE=$2
+ shift 2
+ else
+ usage
+ exit 1
+ fi
+ ;;
+ --with_multipathd)
+ if [ -n "$2" ]; then
+ MULTIPATHD=$2
+ shift 2
+ else
+ usage
+ exit 1
+ fi
+ ;;
+ --outfile)
+ if [ -n "$2" ]; then
+ OUTPUTFILE=$2
+ HAVE_OUTFILE=1
+ shift 2
+ else
+ usage
+ exit 1
+ fi
+ ;;
+ *)
+ usage
+ exit
+ esac
+ done
+}
+
+function validate_args
+{
+ if [ "$ENABLE" = "0" ] && [ -n "$FRIENDLY" -o -n "$FIND" -o -n "$PROPERTY" -o -n "$MODULE" ]; then
+ echo "ignoring extra parameters on disable"
+ FRIENDLY=""
+ FIND=""
+ PROPERTY=""
+ MODULE=""
+ fi
+ if [ -n "$FRIENDLY" ] && [ "$FRIENDLY" != "y" -a "$FRIENDLY" != "n" ]; then
+ echo "--user_friendly_names must be either 'y' or 'n'"
+ exit 1
+ fi
+ if [ -n "$FIND" ] && [ "$FIND" != "y" -a "$FIND" != "n" ]; then
+ echo "--find_multipaths must be either 'y' or 'n'"
+ exit 1
+ fi
+ if [ -n "$PROPERTY" ] && [ "$PROPERTY" != "y" -a "$PROPERTY" != "n" ]; then
+ echo "--property_blacklist must be either 'y' or 'n'"
+ exit 1
+ fi
+ if [ -n "$FOREIGN" ] && [ "$FOREIGN" != "y" -a "$FOREIGN" != "n" ]; then
+ echo "--enable_foreign must be either 'y' or 'n'"
+ exit 1
+ fi
+ if [ -z "$ENABLE" -a -z "$FIND" -a -z "$FRIENDLY" -a -z "$PROPERTY" ]; then
+ SHOW_STATUS=1
+ fi
+ if [ -n "$MODULE" ] && [ "$MODULE" != "y" -a "$MODULE" != "n" ]; then
+ echo "--with_module must be either 'y' or 'n'"
+ exit 1
+ fi
+ if [ -n "$MULTIPATHD" ] && [ "$MULTIPATHD" != "y" -a "$MULTIPATHD" != "n" ]; then
+ echo "--with_multipathd must be either 'y' or 'n'"
+ exit 1
+ fi
+ if [ "$ENABLE" = 2 -a -z "$HAVE_OUTFILE" ]; then
+ echo "Because --allow makes changes that cannot be automatically reversed,"
+ echo "you must set --outfile when you set --allow"
+ exit 1
+ fi
+}
+
+function add_blacklist_exceptions
+{
+ INDEX=0
+ while [ "$INDEX" -lt "$WWIDS" ] ; do
+ sed -i '/^blacklist_exceptions[[:space:]]*{/ a\
+ wwid '"\"${WWID_LIST[$INDEX]}\""'
+' $TMPFILE
+ ((INDEX++))
+ done
+}
+
+umask 0077
+
+parse_args "$@"
+
+validate_args
+
+if [ ! -d "$MULTIPATHDIR" ]; then
+ echo "/etc/multipath/ does not exist. failing"
+ exit 1
+fi
+
+rm $TMPFILE 2> /dev/null
+echo "$DEFAULT_CONFIG" > $TMPFILE
+if [ -f "$CONFIGFILE" ]; then
+ cp $CONFIGFILE $TMPFILE
+fi
+
+if grep -q "^blacklist[[:space:]]*{" $TMPFILE ; then
+ HAVE_BLACKLIST=1
+fi
+
+if grep -q "^blacklist_exceptions[[:space:]]*{" $TMPFILE ; then
+ HAVE_EXCEPTIONS=1
+fi
+
+if grep -q "^defaults[[:space:]]*{" $TMPFILE ; then
+ HAVE_DEFAULTS=1
+fi
+
+if [ -z "$MODULE" -o "$MODULE" = "y" ]; then
+ if lsmod | grep -q "dm_multipath" ; then
+ HAVE_MODULE=1
+ else
+ HAVE_MODULE=0
+ fi
+fi
+
+if [ "$MULTIPATHD" = "y" ]; then
+ if /bin/systemctl status multipathd.service > /dev/null 2>&1 ; then
+ HAVE_MULTIPATHD=1
+ else
+ HAVE_MULTIPATHD=0
+ fi
+fi
+
+if [ "$HAVE_BLACKLIST" = "1" ]; then
+ if sed -n '/^blacklist[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*devnode \"\.\?\*\"" ; then
+ HAVE_DISABLE=1
+ elif sed -n '/^blacklist[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*#[[:space:]]*devnode \"\.\?\*\"" ; then
+ HAVE_DISABLE=0
+ fi
+fi
+
+if [ "$HAVE_BLACKLIST" = "1" ]; then
+ if sed -n '/^blacklist[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*wwid \"\.\?\*\"" ; then
+ HAVE_WWID_DISABLE=1
+ elif sed -n '/^blacklist[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*#[[:space:]]*wwid \"\.\?\*\"" ; then
+ HAVE_WWID_DISABLE=0
+ fi
+fi
+
+if [ "$HAVE_DEFAULTS" = "1" ]; then
+ if sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*find_multipaths[[:space:]]*\(yes\|1\)" ; then
+ HAVE_FIND=1
+ elif sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*find_multipaths[[:space:]]*\(no\|0\)" ; then
+ HAVE_FIND=0
+ fi
+ if sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*user_friendly_names[[:space:]]*\(yes\|1\)" ; then
+ HAVE_FRIENDLY=1
+ elif sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*user_friendly_names[[:space:]]*\(no\|0\)" ; then
+ HAVE_FRIENDLY=0
+ fi
+ if sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*#[[:space:]]*enable_foreign" ; then
+ HAVE_FOREIGN=0
+ elif sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*enable_foreign[[:space:]]*\"\^\$\"" ; then
+ HAVE_FOREIGN=1
+ elif sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*enable_foreign" ; then
+ HAVE_FOREIGN=2
+ fi
+fi
+
+if [ "$HAVE_EXCEPTIONS" = "1" ]; then
+ if sed -n '/^blacklist_exceptions[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*property[[:space:]]*\"(SCSI_IDENT_|ID_WWN)\"" ; then
+ HAVE_PROPERTY=1
+ elif sed -n '/^blacklist_exceptions[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*#[[:space:]]*property[[:space:]]*\"(SCSI_IDENT_|ID_WWN)\"" ; then
+ HAVE_PROPERTY=0
+ fi
+fi
+
+if [ -n "$SHOW_STATUS" ]; then
+ if [ -z "$HAVE_DISABLE" -o "$HAVE_DISABLE" = 0 ]; then
+ echo "multipath is enabled"
+ else
+ echo "multipath is disabled"
+ fi
+ if [ -z "$HAVE_FIND" -o "$HAVE_FIND" = 0 ]; then
+ echo "find_multipaths is disabled"
+ else
+ echo "find_multipaths is enabled"
+ fi
+ if [ -z "$HAVE_FRIENDLY" -o "$HAVE_FRIENDLY" = 0 ]; then
+ echo "user_friendly_names is disabled"
+ else
+ echo "user_friendly_names is enabled"
+ fi
+ if [ -z "$HAVE_PROPERTY" -o "$HAVE_PROPERTY" = 0 ]; then
+ echo "default property blacklist is disabled"
+ else
+ echo "default property blacklist is enabled"
+ fi
+ if [ -z "$HAVE_FOREIGN" -o "$HAVE_FOREIGN" = 0 ]; then
+ echo "enable_foreign is not set (all foreign multipath devices will be shown)"
+ elif [ "$HAVE_FOREIGN" = 1 ]; then
+ echo "enable_foreign is set (no foreign multipath devices will be shown)"
+ else
+ echo "enable_foreign is set (foreign multipath devices may not be shown)"
+ fi
+ if [ -n "$HAVE_MODULE" ]; then
+ if [ "$HAVE_MODULE" = 1 ]; then
+ echo "dm_multipath module is loaded"
+ else
+ echo "dm_multipath module is not loaded"
+ fi
+ fi
+ if [ -z "$HAVE_MULTIPATHD" ]; then
+ if /bin/systemctl status multipathd.service > /dev/null 2>&1 ; then
+ HAVE_MULTIPATHD=1
+ else
+ HAVE_MULTIPATHD=0
+ fi
+ fi
+ if [ "$HAVE_MULTIPATHD" = 1 ]; then
+ echo "multipathd is running"
+ else
+ echo "multipathd is not running"
+ fi
+ exit 0
+fi
+
+if [ -z "$HAVE_BLACKLIST" ]; then
+ cat >> $TMPFILE <<- _EOF_
+
+blacklist {
+}
+_EOF_
+fi
+
+if [ -z "$HAVE_DEFAULTS" ]; then
+ cat >> $TMPFILE <<- _EOF_
+
+defaults {
+}
+_EOF_
+fi
+
+if [ "$ENABLE" = 2 ]; then
+ if [ "$HAVE_DISABLE" = 1 ]; then
+ sed -i '/^blacklist[[:space:]]*{/,/^}/ s/^[[:space:]]*devnode \"\.\?\*\"/# devnode ".*"/' $TMPFILE
+ fi
+ if [ -z "$HAVE_WWID_DISABLE" ]; then
+ sed -i '/^blacklist[[:space:]]*{/ a\
+ wwid ".*"
+' $TMPFILE
+ elif [ "$HAVE_WWID_DISABLE" = 0 ]; then
+ sed -i '/^blacklist[[:space:]]*{/,/^}/ s/^[[:space:]]*#[[:space:]]*wwid \"\.\?\*\"/ wwid ".*"/' $TMPFILE
+ fi
+ if [ "$HAVE_EXCEPTIONS" = 1 ]; then
+ sed -i '/^blacklist_exceptions[[:space:]]*{/,/^}/ {/^[[:space:]]*wwid/ d}' $TMPFILE
+ else
+ cat >> $TMPFILE <<- _EOF_
+
+blacklist_exceptions {
+}
+_EOF_
+ fi
+ add_blacklist_exceptions
+elif [ "$ENABLE" = 1 ]; then
+ if [ "$HAVE_DISABLE" = 1 ]; then
+ sed -i '/^blacklist[[:space:]]*{/,/^}/ s/^[[:space:]]*devnode \"\.\?\*\"/# devnode ".*"/' $TMPFILE
+ fi
+elif [ "$ENABLE" = 0 ]; then
+ if [ -z "$HAVE_DISABLE" ]; then
+ sed -i '/^blacklist[[:space:]]*{/ a\
+ devnode ".*"
+' $TMPFILE
+ elif [ "$HAVE_DISABLE" = 0 ]; then
+ sed -i '/^blacklist[[:space:]]*{/,/^}/ s/^[[:space:]]*#[[:space:]]*devnode \"\.\?\*\"/ devnode ".*"/' $TMPFILE
+ fi
+fi
+
+if [ "$FIND" = "n" ]; then
+ if [ "$HAVE_FIND" = 1 ]; then
+ sed -i '/^defaults[[:space:]]*{/,/^}/ s/^[[:space:]]*find_multipaths[[:space:]]*\(yes\|1\)/ find_multipaths no/' $TMPFILE
+ CHANGED_CONFIG=1
+ fi
+elif [ "$FIND" = "y" ]; then
+ if [ -z "$HAVE_FIND" ]; then
+ sed -i '/^defaults[[:space:]]*{/ a\
+ find_multipaths yes
+' $TMPFILE
+ CHANGED_CONFIG=1
+ elif [ "$HAVE_FIND" = 0 ]; then
+ sed -i '/^defaults[[:space:]]*{/,/^}/ s/^[[:space:]]*find_multipaths[[:space:]]*\(no\|0\)/ find_multipaths yes/' $TMPFILE
+ CHANGED_CONFIG=1
+ fi
+fi
+
+if [ "$FRIENDLY" = "n" ]; then
+ if [ "$HAVE_FRIENDLY" = 1 ]; then
+ sed -i '/^defaults[[:space:]]*{/,/^}/ s/^[[:space:]]*user_friendly_names[[:space:]]*\(yes\|1\)/ user_friendly_names no/' $TMPFILE
+ CHANGED_CONFIG=1
+ fi
+elif [ "$FRIENDLY" = "y" ]; then
+ if [ -z "$HAVE_FRIENDLY" ]; then
+ sed -i '/^defaults[[:space:]]*{/ a\
+ user_friendly_names yes
+' $TMPFILE
+ CHANGED_CONFIG=1
+ elif [ "$HAVE_FRIENDLY" = 0 ]; then
+ sed -i '/^defaults[[:space:]]*{/,/^}/ s/^[[:space:]]*user_friendly_names[[:space:]]*\(no\|0\)/ user_friendly_names yes/' $TMPFILE
+ CHANGED_CONFIG=1
+ fi
+fi
+
+if [ "$PROPERTY" = "n" ]; then
+ if [ "$HAVE_PROPERTY" = 1 ]; then
+ sed -i '/^blacklist_exceptions[[:space:]]*{/,/^}/ s/^[[:space:]]*property[[:space:]]*\"(SCSI_IDENT_|ID_WWN)\"/# property \"(SCSI_IDENT_|ID_WWN)\"/' $TMPFILE
+ CHANGED_CONFIG=1
+ fi
+elif [ "$PROPERTY" = "y" ]; then
+ if [ -z "$HAVE_PROPERTY" ]; then
+ sed -i '/^blacklist_exceptions[[:space:]]*{/ a\
+ property "(SCSI_IDENT_|ID_WWN)"
+' $TMPFILE
+ CHANGED_CONFIG=1
+ elif [ "$HAVE_PROPERTY" = 0 ]; then
+ sed -i '/^blacklist_exceptions[[:space:]]*{/,/^}/ s/^[[:space:]]*#[[:space:]]*property[[:space:]]*\"(SCSI_IDENT_|ID_WWN)\"/ property \"(SCSI_IDENT_|ID_WWN)\"/' $TMPFILE
+ CHANGED_CONFIG=1
+ fi
+fi
+
+if [ "$FOREIGN" = "y" ]; then
+ if [ "$HAVE_FOREIGN" = 1 -o "$HAVE_FOREIGN" = 2 ]; then
+ sed -i '/^defaults[[:space:]]*{/,/^}/ s/^[[:space:]]*enable_foreign/# enable_foreign/' $TMPFILE
+ CHANGED_CONFIG=1
+ fi
+elif [ "$FOREIGN" = "n" ]; then
+ if [ -z "$HAVE_FOREIGN" ]; then
+ sed -i '/^defaults[[:space:]]*{/ a\
+ enable_foreign "^$"
+' $TMPFILE
+ CHANGED_CONFIG=1
+ elif [ "$HAVE_FOREIGN" = 0 -o "$HAVE_FOREIGN" = 2 ]; then
+ sed -i '/^defaults[[:space:]]*{/,/^}/ s/^[[:space:]]*#\?[[:space:]]*enable_foreign.*$/ enable_foreign "^$"/' $TMPFILE
+ CHANGED_CONFIG=1
+ fi
+fi
+
+if [ -f "$OUTPUTFILE" ]; then
+ cp $OUTPUTFILE $OUTPUTFILE.old
+ if [ $? != 0 ]; then
+ echo "failed to backup old config file, $OUTPUTFILE not updated"
+ exit 1
+ fi
+fi
+
+cp $TMPFILE $OUTPUTFILE
+if [ $? != 0 ]; then
+ echo "failed to copy new config file into place, check $OUTPUTFILE is still OK"
+ exit 1
+fi
+
+rm -f $TMPFILE
+
+if [ "$ENABLE" = 1 ]; then
+ if [ "$HAVE_MODULE" = 0 ]; then
+ modprobe dm_multipath
+ fi
+ if [ "$HAVE_MULTIPATHD" = 0 ]; then
+ systemctl start multipathd.service
+ fi
+elif [ "$ENABLE" = 0 ]; then
+ if [ "$HAVE_MULTIPATHD" = 1 ]; then
+ systemctl stop multipathd.service
+ fi
+elif [ -n "$CHANGED_CONFIG" -a "$HAVE_MULTIPATHD" = 1 ]; then
+ systemctl reload multipathd.service
+fi
diff --git a/multipath/mpathconf.8 b/multipath/mpathconf.8
new file mode 100644
index 00000000..b82961d6
--- /dev/null
+++ b/multipath/mpathconf.8
@@ -0,0 +1,135 @@
+.TH MPATHCONF 8 "June 2010" "" "Linux Administrator's Manual"
+.SH NAME
+mpathconf - A tool for configuring device-mapper-multipath
+.SH SYNOPSIS
+.B mpathconf
+.RB [\| commands \|]
+.RB [\| options \|]
+.SH DESCRIPTION
+.B mpathconf
+is a utility that creates or modifies
+.B /etc/multipath.conf.
+It can enable or disable multipathing and configure some common options.
+.B mpathconf
+can also load the
+.B dm_multipath
+module, start and stop the
+.B multipathd
+daemon, and configure the
+.B multipathd
+service to start automatically or not. If
+.B mpathconf
+is called with no commands, it will display the current configuration, but
+will not create of modify
+.B /etc/multipath.conf
+
+The default options for mpathconf are
+.B --with_module
+The
+.B --with_multipathd
+option is not set by default. Enabling multipathing will load the
+.B dm_multipath
+module but it will not immediately start it. This is so
+that users can manually edit their config file if necessary, before starting
+.B multipathd.
+
+If
+.B /etc/multipath.conf
+already exists, mpathconf will edit it. If it does not exist, mpathconf will
+create a default file with
+.B user_friendly_names
+and
+.B find_multipaths
+set. To disable these, use the
+.B --user_friendly_names n
+and
+.B --find_multipaths n
+options
+.SH COMMANDS
+.TP
+.B --enable
+Removes any line that blacklists all device nodes from the
+.B /etc/multipath.conf
+blacklist section. Also, creates
+.B /etc/multipath.conf
+if it doesn't exist.
+.TP
+.B --disable
+Adds a line that blacklists all device nodes to the
+.B /etc/multipath.conf
+blacklist section. If no blacklist section exists, it will create one.
+.TP
+.B --allow \fB<device>\fP
+Modifies the \fB/etc/multipath/conf\fP blacklist to blacklist all
+wwids and the blacklist_exceptions to whitelist \fB<device>\fP. \fB<device>\fP
+can be in the form of MAJOR:MINOR, a wwid, or the name of a device-mapper
+device, either a multipath device, or any device on stacked on top of one or
+more multipath devices. This command can be used multiple times to allow
+multiple devices. \fBNOTE:\fP This action will create a configuration file that
+mpathconf will not be able to revert back to its previous state. Because
+of this, \fB--outfile\fP is required when using \fB--allow\fP.
+.TP
+.B --user_friendly_names \fP { \fBy\fP | \fBn\fP }
+If set to \fBy\fP, this adds the line
+.B user_friendly_names yes
+to the
+.B /etc/multipath.conf
+defaults section. If set to \fBn\fP, this removes the line, if present. This
+command can be used along with any other command.
+.TP
+.B --find_multipaths\fP { \fBy\fP | \fBn\fP }
+If set to \fBy\fP, this adds the line
+.B find_multipaths yes
+to the
+.B /etc/multipath.conf
+defaults section. If set to \fBn\fP, this removes the line, if present. This
+command can be used along with any other command.
+.TP
+.B --property_blacklist \fP { \fBy\fP | \fBn\fP }
+If set to \fBy\fP, this adds the line
+.B property "(SCSI_IDENT_|ID_WWN)"
+to the
+.B /etc/multipath.conf
+blacklist_exceptions section. If set to \fBn\fP, this removes the line, if
+present. This command can be used along with any other command.
+.TP
+.B --enable_foreign\fP { \fBy\fP | \fBn\fP }
+If set to \fBn\fP, this adds the line
+.B enable_foreign "^$"
+to the
+.B /etc/multipath.conf
+defaults section. if set to \fBy\fP, this removes the line, if present. This
+command can be used along with any other command.
+.TP
+.B --outfile \fB<filename>\fP
+Write the resulting multipath configuration to \fB<filename>\fP instead of
+\fB/etc/multipath.conf\fP.
+.SH OPTIONS
+.TP
+.B --with_module\fP { \fBy\fP | \fBn\fP }
+If set to \fBy\fP, this runs
+.B modprobe dm_multipath
+to install the multipath modules. This option only works with the
+.B --enable
+command. This option is set to \fBy\fP by default.
+.TP
+.B --with_multipathd { \fBy\fP | \fBn\fP }
+If set to \fBy\fP, this runs
+.B service multipathd start
+to start the multipathd daemon on \fB--enable\fP,
+.B service multipathd stop
+to stop the multipathd daemon on \fB--disable\fP, and
+.B service multipathd reload
+to reconfigure multipathd on \fB--user_frindly_names\fP and
+\fB--find_multipaths\fP.
+This option is set to \fBn\fP by default.
+.SH FILES
+.BR /etc/multipath.conf
+.SH "SEE ALSO"
+.BR multipath.conf (5),
+.BR modprobe (8),
+.BR multipath (8),
+.BR multipathd (8),
+.BR service (8),
+.SH AUTHOR
+Benjamin Marzinski <bmarzins@redhat.com>
--
2.17.2

169
SOURCES/0016-RH-add-wwids-from-kernel-cmdline-mpath.wwids-with-A.patch Normal file
View File

@ -0,0 +1,169 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Fri, 17 Oct 2014 11:20:34 -0500
Subject: [PATCH] RH: add wwids from kernel cmdline mpath.wwids with -A
This patch adds another option to multipath, "-A", which reads
/proc/cmdline for mpath.wwid=<WWID> options, and adds any wwids it finds
to /etc/multipath/wwids. While this isn't usually important during
normal operation, since these wwids should already be added, it can be
helpful during installation, to make sure that multipath can claim
devices as its own, before LVM or something else makes use of them. The
patch also execs "/sbin/multipath -A" before running multipathd in
multipathd.service
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/wwids.c | 44 +++++++++++++++++++++++++++++++++++
libmultipath/wwids.h | 1 +
multipath/main.c | 10 ++++++--
multipath/multipath.8 | 7 +++++-
multipathd/multipathd.service | 1 +
5 files changed, 60 insertions(+), 3 deletions(-)
diff --git a/libmultipath/wwids.c b/libmultipath/wwids.c
index 28a2150d..fab6fc8f 100644
--- a/libmultipath/wwids.c
+++ b/libmultipath/wwids.c
@@ -454,3 +454,47 @@ int op ## _wwid(const char *wwid) \
declare_failed_wwid_op(is_failed, false)
declare_failed_wwid_op(mark_failed, true)
declare_failed_wwid_op(unmark_failed, true)
+
+int remember_cmdline_wwid(void)
+{
+ FILE *f = NULL;
+ char buf[LINE_MAX], *next, *ptr;
+ int ret = 0;
+
+ f = fopen("/proc/cmdline", "re");
+ if (!f) {
+ condlog(0, "can't open /proc/cmdline : %s", strerror(errno));
+ return -1;
+ }
+
+ if (!fgets(buf, sizeof(buf), f)) {
+ if (ferror(f))
+ condlog(0, "read of /proc/cmdline failed : %s",
+ strerror(errno));
+ else
+ condlog(0, "couldn't read /proc/cmdline");
+ fclose(f);
+ return -1;
+ }
+ fclose(f);
+ next = buf;
+ while((ptr = strstr(next, "mpath.wwid="))) {
+ ptr += 11;
+ next = strpbrk(ptr, " \t\n");
+ if (next) {
+ *next = '\0';
+ next++;
+ }
+ if (strlen(ptr)) {
+ if (remember_wwid(ptr) != 0)
+ ret = -1;
+ }
+ else {
+ condlog(0, "empty mpath.wwid kernel command line option");
+ ret = -1;
+ }
+ if (!next)
+ break;
+ }
+ return ret;
+}
diff --git a/libmultipath/wwids.h b/libmultipath/wwids.h
index 0c6ee54d..e32a0b0e 100644
--- a/libmultipath/wwids.h
+++ b/libmultipath/wwids.h
@@ -17,6 +17,7 @@ int remember_wwid(char *wwid);
int check_wwids_file(char *wwid, int write_wwid);
int remove_wwid(char *wwid);
int replace_wwids(vector mp);
+int remember_cmdline_wwid(void);
enum {
WWID_IS_NOT_FAILED = 0,
diff --git a/multipath/main.c b/multipath/main.c
index cf9d2a28..78822ee1 100644
--- a/multipath/main.c
+++ b/multipath/main.c
@@ -138,7 +138,7 @@ usage (char * progname)
fprintf (stderr, " %s [-v level] [-R retries] -F\n", progname);
fprintf (stderr, " %s [-v level] [-l|-ll] [device]\n", progname);
fprintf (stderr, " %s [-v level] [-a|-w] device\n", progname);
- fprintf (stderr, " %s [-v level] -W\n", progname);
+ fprintf (stderr, " %s [-v level] [-A|-W]\n", progname);
fprintf (stderr, " %s [-v level] [-i] [-c|-C] device\n", progname);
fprintf (stderr, " %s [-v level] [-i] [-u|-U]\n", progname);
fprintf (stderr, " %s [-h|-t|-T]\n", progname);
@@ -151,6 +151,8 @@ usage (char * progname)
" -f flush a multipath device map\n"
" -F flush all multipath device maps\n"
" -a add a device wwid to the wwids file\n"
+ " -A add devices from kernel command line mpath.wwids\n"
+ " parameters to wwids file\n"
" -c check if a device should be a path in a multipath device\n"
" -C check if a multipath device has usable paths\n"
" -q allow queue_if_no_path when multipathd is not running\n"
@@ -907,7 +909,7 @@ main (int argc, char *argv[])
multipath_conf = conf;
conf->retrigger_tries = 0;
conf->force_sync = 1;
- while ((arg = getopt(argc, argv, ":adcChl::FfM:v:p:b:BrR:itTquUwW")) != EOF ) {
+ while ((arg = getopt(argc, argv, ":aAdcChl::FfM:v:p:b:BrR:itTquUwW")) != EOF ) {
switch(arg) {
case 1: printf("optarg : %s\n",optarg);
break;
@@ -977,6 +979,10 @@ main (int argc, char *argv[])
case 'T':
cmd = CMD_DUMP_CONFIG;
break;
+ case 'A':
+ if (remember_cmdline_wwid() != 0)
+ exit(RTVL_FAIL);
+ exit(RTVL_OK);
case 'h':
usage(argv[0]);
exit(RTVL_OK);
diff --git a/multipath/multipath.8 b/multipath/multipath.8
index 9cdd05a3..8befc45a 100644
--- a/multipath/multipath.8
+++ b/multipath/multipath.8
@@ -63,7 +63,7 @@ multipath \- Device mapper target autoconfig.
.B multipath
.RB [\| \-v\ \c
.IR level \|]
-.B -W
+.RB [\| \-A | \-W \|]
.
.LP
.B multipath
@@ -145,6 +145,11 @@ device mapper, path checkers ...).
Add the WWID for the specified device to the WWIDs file.
.
.TP
+.B \-A
+Add the WWIDs from any kernel command line \fImpath.wwid\fR parameters to the
+WWIDs file.
+.
+.TP
.B \-w
Remove the WWID for the specified device from the WWIDs file.
.
diff --git a/multipathd/multipathd.service b/multipathd/multipathd.service
index 17434cef..0fbcc46b 100644
--- a/multipathd/multipathd.service
+++ b/multipathd/multipathd.service
@@ -15,6 +15,7 @@ Type=notify
NotifyAccess=main
LimitCORE=infinity
ExecStartPre=-/sbin/modprobe -a scsi_dh_alua scsi_dh_emc scsi_dh_rdac dm-multipath
+ExecStartPre=-/sbin/multipath -A
ExecStart=/sbin/multipathd -d -s
ExecReload=/sbin/multipathd reconfigure
TasksMax=infinity
--
2.17.2

121
SOURCES/0017-RH-warn-on-invalid-regex-instead-of-failing.patch Normal file
View File

@ -0,0 +1,121 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 6 Nov 2017 21:39:28 -0600
Subject: [PATCH] RH: warn on invalid regex instead of failing
multipath.conf used to allow "*" as a match everything regular expression,
instead of requiring ".*". Instead of erroring when the old style
regular expressions are used, it should print a warning and convert
them.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/dict.c | 27 +++++++++++++++++++++------
libmultipath/parser.c | 13 +++++++++++++
libmultipath/parser.h | 1 +
3 files changed, 35 insertions(+), 6 deletions(-)
diff --git a/libmultipath/dict.c b/libmultipath/dict.c
index 0e9ea387..184d4b22 100644
--- a/libmultipath/dict.c
+++ b/libmultipath/dict.c
@@ -103,6 +103,21 @@ set_str(vector strvec, void *ptr)
return 0;
}
+static int
+set_regex(vector strvec, void *ptr)
+{
+ char **str_ptr = (char **)ptr;
+
+ if (*str_ptr)
+ FREE(*str_ptr);
+ *str_ptr = set_regex_value(strvec);
+
+ if (!*str_ptr)
+ return 1;
+
+ return 0;
+}
+
static int
set_yes_no(vector strvec, void *ptr)
{
@@ -1504,7 +1519,7 @@ ble_ ## option ## _handler (struct config *conf, vector strvec) \
if (!conf->option) \
return 1; \
\
- buff = set_value(strvec); \
+ buff = set_regex_value(strvec); \
if (!buff) \
return 1; \
\
@@ -1520,7 +1535,7 @@ ble_ ## option ## _ ## name ## _handler (struct config *conf, vector strvec) \
if (!conf->option) \
return 1; \
\
- buff = set_value(strvec); \
+ buff = set_regex_value(strvec); \
if (!buff) \
return 1; \
\
@@ -1623,16 +1638,16 @@ device_handler(struct config *conf, vector strvec)
return 0;
}
-declare_hw_handler(vendor, set_str)
+declare_hw_handler(vendor, set_regex)
declare_hw_snprint(vendor, print_str)
-declare_hw_handler(product, set_str)
+declare_hw_handler(product, set_regex)
declare_hw_snprint(product, print_str)
-declare_hw_handler(revision, set_str)
+declare_hw_handler(revision, set_regex)
declare_hw_snprint(revision, print_str)
-declare_hw_handler(bl_product, set_str)
+declare_hw_handler(bl_product, set_regex)
declare_hw_snprint(bl_product, print_str)
declare_hw_handler(hwhandler, set_str)
diff --git a/libmultipath/parser.c b/libmultipath/parser.c
index d478b177..a184511b 100644
--- a/libmultipath/parser.c
+++ b/libmultipath/parser.c
@@ -382,6 +382,19 @@ oom:
return NULL;
}
+void *
+set_regex_value(vector strvec)
+{
+ char *buff = set_value(strvec);
+
+ if (buff && strcmp("*", buff) == 0) {
+ condlog(0, "Invalid regular expression \"*\" in multipath.conf. Using \".*\"");
+ FREE(buff);
+ return strdup(".*");
+ }
+ return buff;
+}
+
/* non-recursive configuration stream handler */
static int kw_level = 0;
diff --git a/libmultipath/parser.h b/libmultipath/parser.h
index 62906e98..b7917052 100644
--- a/libmultipath/parser.h
+++ b/libmultipath/parser.h
@@ -77,6 +77,7 @@ extern void dump_keywords(vector keydump, int level);
extern void free_keywords(vector keywords);
extern vector alloc_strvec(char *string);
extern void *set_value(vector strvec);
+extern void *set_regex_value(vector strvec);
extern int process_file(struct config *conf, char *conf_file);
extern struct keyword * find_keyword(vector keywords, vector v, char * name);
int snprint_keyword(char *buff, int len, char *fmt, struct keyword *kw,
--
2.17.2

29
SOURCES/0018-RH-reset-default-find_mutipaths-value-to-off.patch Normal file
View File

@ -0,0 +1,29 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 7 Jun 2018 17:43:52 -0500
Subject: [PATCH] RH: reset default find_mutipaths value to off
Upstream has changed to default find_multipaths to "strict". For now
Redhat will retain the previous default of "off".
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/defaults.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libmultipath/defaults.h b/libmultipath/defaults.h
index e5ee6afe..52fe05b9 100644
--- a/libmultipath/defaults.h
+++ b/libmultipath/defaults.h
@@ -22,7 +22,7 @@
#define DEFAULT_NO_PATH_RETRY NO_PATH_RETRY_UNDEF
#define DEFAULT_VERBOSITY 2
#define DEFAULT_REASSIGN_MAPS 0
-#define DEFAULT_FIND_MULTIPATHS FIND_MULTIPATHS_STRICT
+#define DEFAULT_FIND_MULTIPATHS FIND_MULTIPATHS_OFF
#define DEFAULT_FAST_IO_FAIL 5
#define DEFAULT_DEV_LOSS_TMO 600
#define DEFAULT_RETAIN_HWHANDLER RETAIN_HWHANDLER_ON
--
2.17.2

26
SOURCES/0019-RH-Fix-nvme-compilation-warning.patch Normal file
View File

@ -0,0 +1,26 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Fri, 25 Jan 2019 14:54:56 -0600
Subject: [PATCH] RH: Fix nvme compilation warning
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/nvme/argconfig.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libmultipath/nvme/argconfig.h b/libmultipath/nvme/argconfig.h
index adb192b6..bfd10ef8 100644
--- a/libmultipath/nvme/argconfig.h
+++ b/libmultipath/nvme/argconfig.h
@@ -76,7 +76,7 @@ struct argconfig_commandline_options {
extern "C" {
#endif
-typedef void argconfig_help_func();
+typedef void argconfig_help_func(void);
void argconfig_append_usage(const char *str);
void argconfig_print_help(const char *program_desc,
const struct argconfig_commandline_options *options);
--
2.17.2

87
SOURCES/0020-RH-attempt-to-get-ANA-info-via-sysfs-first.patch Normal file
View File

@ -0,0 +1,87 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 11 Apr 2019 13:25:42 -0500
Subject: [PATCH] RH: attempt to get ANA info via sysfs first
When the ANA prioritizer is run, first see if the "ana_state" sysfs file
exists, and if it does, try to read the state from there. If that fails,
fallback to using an ioctl.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/prioritizers/ana.c | 31 +++++++++++++++++++++++++++++--
1 file changed, 29 insertions(+), 2 deletions(-)
diff --git a/libmultipath/prioritizers/ana.c b/libmultipath/prioritizers/ana.c
index b5c7873d..e139360c 100644
--- a/libmultipath/prioritizers/ana.c
+++ b/libmultipath/prioritizers/ana.c
@@ -24,6 +24,7 @@
#include "prio.h"
#include "util.h"
#include "structs.h"
+#include "sysfs.h"
enum {
ANA_ERR_GETCTRL_FAILED = 1,
@@ -36,6 +37,7 @@ enum {
ANA_ERR_GETNS_FAILED,
ANA_ERR_NO_MEMORY,
ANA_ERR_NO_INFORMATION,
+ ANA_ERR_INVALID_STATE,
};
static const char *ana_errmsg[] = {
@@ -49,6 +51,7 @@ static const char *ana_errmsg[] = {
[ANA_ERR_GETNS_FAILED] = "couldn't get namespace info",
[ANA_ERR_NO_MEMORY] = "out of memory",
[ANA_ERR_NO_INFORMATION] = "invalid fd",
+ [ANA_ERR_INVALID_STATE] = "invalid state",
};
static const char *anas_string[] = {
@@ -107,6 +110,27 @@ static int get_ana_state(__u32 nsid, __u32 anagrpid, void *ana_log,
return -ANA_ERR_GETANAS_NOTFOUND;
}
+static int get_ana_info_sysfs(struct path *pp)
+{
+ char state[32];
+
+ if (!pp->udev || sysfs_attr_get_value(pp->udev, "ana_state", state,
+ sizeof(state)) <= 0)
+ return -ANA_ERR_NO_INFORMATION;
+
+ if (strcmp(state, "optimized") == 0)
+ return NVME_ANA_OPTIMIZED;
+ if (strcmp(state, "non-optimized") == 0)
+ return NVME_ANA_NONOPTIMIZED;
+ if (strcmp(state, "inaccessible") == 0)
+ return NVME_ANA_INACCESSIBLE;
+ if (strcmp(state, "persistent-loss") == 0)
+ return NVME_ANA_PERSISTENT_LOSS;
+ if (strcmp(state, "change") == 0)
+ return NVME_ANA_CHANGE;
+ return -ANA_ERR_INVALID_STATE;
+}
+
static int get_ana_info(struct path * pp)
{
int rc;
@@ -210,8 +234,11 @@ int getprio(struct path *pp, __attribute__((unused)) char *args,
if (pp->fd < 0)
rc = -ANA_ERR_NO_INFORMATION;
- else
- rc = get_ana_info(pp);
+ else {
+ rc = get_ana_info_sysfs(pp);
+ if (rc < 0)
+ rc = get_ana_info(pp);
+ }
switch (rc) {
case NVME_ANA_OPTIMIZED:
--
2.17.2

139
SOURCES/0021-libmultipath-remove-_blacklist_exceptions-functions.patch Normal file
View File

@ -0,0 +1,139 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 8 Jun 2020 14:27:51 -0500
Subject: [PATCH] libmultipath: remove _blacklist_exceptions functions
_blacklist_exceptions() and _blacklist_exceptions_device() are exactly
the same as _blacklist() and _blacklist_device(), so remove them, and
give the remaining functions to a more general name.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/blacklist.c | 62 ++++++++++------------------------------
1 file changed, 15 insertions(+), 47 deletions(-)
diff --git a/libmultipath/blacklist.c b/libmultipath/blacklist.c
index d9691b17..04d3adb9 100644
--- a/libmultipath/blacklist.c
+++ b/libmultipath/blacklist.c
@@ -101,21 +101,8 @@ int set_ble_device(vector blist, char * vendor, char * product, int origin)
return 0;
}
-int
-_blacklist_exceptions (vector elist, const char * str)
-{
- int i;
- struct blentry * ele;
-
- vector_foreach_slot (elist, ele, i) {
- if (!regexec(&ele->regex, str, 0, NULL, 0))
- return 1;
- }
- return 0;
-}
-
-int
-_blacklist (vector blist, const char * str)
+static int
+match_reglist (vector blist, const char * str)
{
int i;
struct blentry * ble;
@@ -127,28 +114,9 @@ _blacklist (vector blist, const char * str)
return 0;
}
-int
-_blacklist_exceptions_device(const struct _vector *elist, const char * vendor,
- const char * product)
-{
- int i;
- struct blentry_device * ble;
-
- vector_foreach_slot (elist, ble, i) {
- if (!ble->vendor && !ble->product)
- continue;
- if ((!ble->vendor ||
- !regexec(&ble->vendor_reg, vendor, 0, NULL, 0)) &&
- (!ble->product ||
- !regexec(&ble->product_reg, product, 0, NULL, 0)))
- return 1;
- }
- return 0;
-}
-
-int
-_blacklist_device (const struct _vector *blist, const char * vendor,
- const char * product)
+static int
+match_reglist_device (const struct _vector *blist, const char * vendor,
+ const char * product)
{
int i;
struct blentry_device * ble;
@@ -294,9 +262,9 @@ filter_device (vector blist, vector elist, char * vendor, char * product,
int r = MATCH_NOTHING;
if (vendor && product) {
- if (_blacklist_exceptions_device(elist, vendor, product))
+ if (match_reglist_device(elist, vendor, product))
r = MATCH_DEVICE_BLIST_EXCEPT;
- else if (_blacklist_device(blist, vendor, product))
+ else if (match_reglist_device(blist, vendor, product))
r = MATCH_DEVICE_BLIST;
}
@@ -310,9 +278,9 @@ filter_devnode (vector blist, vector elist, char * dev)
int r = MATCH_NOTHING;
if (dev) {
- if (_blacklist_exceptions(elist, dev))
+ if (match_reglist(elist, dev))
r = MATCH_DEVNODE_BLIST_EXCEPT;
- else if (_blacklist(blist, dev))
+ else if (match_reglist(blist, dev))
r = MATCH_DEVNODE_BLIST;
}
@@ -326,9 +294,9 @@ filter_wwid (vector blist, vector elist, char * wwid, char * dev)
int r = MATCH_NOTHING;
if (wwid) {
- if (_blacklist_exceptions(elist, wwid))
+ if (match_reglist(elist, wwid))
r = MATCH_WWID_BLIST_EXCEPT;
- else if (_blacklist(blist, wwid))
+ else if (match_reglist(blist, wwid))
r = MATCH_WWID_BLIST;
}
@@ -345,9 +313,9 @@ filter_protocol(vector blist, vector elist, struct path * pp)
if (pp) {
snprint_path_protocol(buf, sizeof(buf), pp);
- if (_blacklist_exceptions(elist, buf))
+ if (match_reglist(elist, buf))
r = MATCH_PROTOCOL_BLIST_EXCEPT;
- else if (_blacklist(blist, buf))
+ else if (match_reglist(blist, buf))
r = MATCH_PROTOCOL_BLIST;
}
@@ -417,11 +385,11 @@ filter_property(struct config *conf, struct udev_device *udev, int lvl,
if (check_missing_prop && !strcmp(env, uid_attribute))
uid_attr_seen = true;
- if (_blacklist_exceptions(conf->elist_property, env)) {
+ if (match_reglist(conf->elist_property, env)) {
r = MATCH_PROPERTY_BLIST_EXCEPT;
break;
}
- if (_blacklist(conf->blist_property, env)) {
+ if (match_reglist(conf->blist_property, env)) {
r = MATCH_PROPERTY_BLIST;
break;
}
--
2.17.2

95
SOURCES/0022-libmultipath-fix-parser-issue-with-comments-in-strin.patch Normal file
View File

@ -0,0 +1,95 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 8 Jun 2020 20:23:56 -0500
Subject: [PATCH] libmultipath: fix parser issue with comments in strings
If a quoted string starts with '#' or '!', the parser will stop
parsing the line, thinking that it's a comment. It should only
be checking for comments outside of quoted strings. Fixed this and
added unit tests to verify it.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/parser.c | 4 +++-
tests/parser.c | 42 ++++++++++++++++++++++++++++++++++++++++++
2 files changed, 45 insertions(+), 1 deletion(-)
diff --git a/libmultipath/parser.c b/libmultipath/parser.c
index a184511b..a7285a35 100644
--- a/libmultipath/parser.c
+++ b/libmultipath/parser.c
@@ -300,8 +300,10 @@ alloc_strvec(char *string)
(isspace((int) *cp) || !isascii((int) *cp)))
&& *cp != '\0')
cp++;
- if (*cp == '\0' || *cp == '!' || *cp == '#')
+ if (*cp == '\0' ||
+ (!in_string && (*cp == '!' || *cp == '#'))) {
return strvec;
+ }
}
out:
vector_free(strvec);
diff --git a/tests/parser.c b/tests/parser.c
index 29859dac..5772391e 100644
--- a/tests/parser.c
+++ b/tests/parser.c
@@ -440,6 +440,46 @@ static void test18(void **state)
free_strvec(v);
}
+static void test19(void **state)
+{
+#define QUOTED19 "!value"
+ vector v = alloc_strvec("key \"" QUOTED19 "\"");
+ char *val;
+
+ assert_int_equal(VECTOR_SIZE(v), 4);
+ assert_string_equal(VECTOR_SLOT(v, 0), "key");
+ assert_true(is_quote(VECTOR_SLOT(v, 1)));
+ assert_string_equal(VECTOR_SLOT(v, 2), QUOTED19);
+ assert_true(is_quote(VECTOR_SLOT(v, 3)));
+ assert_int_equal(validate_config_strvec(v, test_file), 0);
+
+ val = set_value(v);
+ assert_string_equal(val, QUOTED19);
+
+ free(val);
+ free_strvec(v);
+}
+
+static void test20(void **state)
+{
+#define QUOTED20 "#value"
+ vector v = alloc_strvec("key \"" QUOTED20 "\"");
+ char *val;
+
+ assert_int_equal(VECTOR_SIZE(v), 4);
+ assert_string_equal(VECTOR_SLOT(v, 0), "key");
+ assert_true(is_quote(VECTOR_SLOT(v, 1)));
+ assert_string_equal(VECTOR_SLOT(v, 2), QUOTED20);
+ assert_true(is_quote(VECTOR_SLOT(v, 3)));
+ assert_int_equal(validate_config_strvec(v, test_file), 0);
+
+ val = set_value(v);
+ assert_string_equal(val, QUOTED20);
+
+ free(val);
+ free_strvec(v);
+}
+
int test_config_parser(void)
{
const struct CMUnitTest tests[] = {
@@ -461,6 +501,8 @@ int test_config_parser(void)
cmocka_unit_test(test16),
cmocka_unit_test(test17),
cmocka_unit_test(test18),
+ cmocka_unit_test(test19),
+ cmocka_unit_test(test20),
};
return cmocka_run_group_tests(tests, setup, teardown);
}
--
2.17.2

435
SOURCES/0023-libmultipath-invert-regexes-that-start-with-exclamat.patch Normal file
View File

@ -0,0 +1,435 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 8 Jun 2020 13:40:16 -0500
Subject: [PATCH] libmultipath: invert regexes that start with exclamation
point
The number of devices that multipath needs to blacklist keeps growing,
and the udev rules already have
KERNEL!="sd*|dasd*|nvme*", GOTO="end_mpath"
so they only work correctly with these device types. Instead of
individually blacklisting every type of device that can't be
multipathed, multipath's default blacklist should work like the udev
rule, and blacklist all devices that aren't scsi, dasd, or nvme.
Unfortunately, the c regex library doesn't support negative lookahead.
Instead, multipath should treat "!" at the beginning of
blacklist/exceptions regexes as inverse matching the rest of the regex.
If users need to match a literal '!' as the first character of their
regex, they can use "\!" instead. This allows multipath to change the
default devnode blacklist regex to "!^(sd[a-z]|dasd[a-z]|nvme[0-9])".
Extra tests have been added to the blacklist unit tests to verify the
inverse matching code and the new default blacklist.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/blacklist.c | 41 +++++++++-----
libmultipath/blacklist.h | 3 +
multipath/multipath.conf.5 | 17 ++++--
tests/blacklist.c | 110 +++++++++++++++++++++++++++++++++++++
tests/test-lib.c | 2 +-
5 files changed, 155 insertions(+), 18 deletions(-)
diff --git a/libmultipath/blacklist.c b/libmultipath/blacklist.c
index 04d3adb9..0c58aa32 100644
--- a/libmultipath/blacklist.c
+++ b/libmultipath/blacklist.c
@@ -15,9 +15,24 @@
#include "structs_vec.h"
#include "print.h"
+char *check_invert(char *str, bool *invert)
+{
+ if (str[0] == '!') {
+ *invert = true;
+ return str + 1;
+ }
+ if (str[0] == '\\' && str[1] == '!') {
+ *invert = false;
+ return str + 1;
+ }
+ *invert = false;
+ return str;
+}
+
int store_ble(vector blist, char * str, int origin)
{
struct blentry * ble;
+ char *regex_str;
if (!str)
return 0;
@@ -30,7 +45,8 @@ int store_ble(vector blist, char * str, int origin)
if (!ble)
goto out;
- if (regcomp(&ble->regex, str, REG_EXTENDED|REG_NOSUB))
+ regex_str = check_invert(str, &ble->invert);
+ if (regcomp(&ble->regex, regex_str, REG_EXTENDED|REG_NOSUB))
goto out1;
if (!vector_alloc_slot(blist))
@@ -66,6 +82,7 @@ int alloc_ble_device(vector blist)
int set_ble_device(vector blist, char * vendor, char * product, int origin)
{
struct blentry_device * ble;
+ char *regex_str;
if (!blist)
return 1;
@@ -76,7 +93,8 @@ int set_ble_device(vector blist, char * vendor, char * product, int origin)
return 1;
if (vendor) {
- if (regcomp(&ble->vendor_reg, vendor,
+ regex_str = check_invert(vendor, &ble->vendor_invert);
+ if (regcomp(&ble->vendor_reg, regex_str,
REG_EXTENDED|REG_NOSUB)) {
FREE(vendor);
if (product)
@@ -86,7 +104,8 @@ int set_ble_device(vector blist, char * vendor, char * product, int origin)
ble->vendor = vendor;
}
if (product) {
- if (regcomp(&ble->product_reg, product,
+ regex_str = check_invert(product, &ble->product_invert);
+ if (regcomp(&ble->product_reg, regex_str,
REG_EXTENDED|REG_NOSUB)) {
FREE(product);
if (vendor) {
@@ -108,7 +127,7 @@ match_reglist (vector blist, const char * str)
struct blentry * ble;
vector_foreach_slot (blist, ble, i) {
- if (!regexec(&ble->regex, str, 0, NULL, 0))
+ if (!!regexec(&ble->regex, str, 0, NULL, 0) == ble->invert)
return 1;
}
return 0;
@@ -125,9 +144,11 @@ match_reglist_device (const struct _vector *blist, const char * vendor,
if (!ble->vendor && !ble->product)
continue;
if ((!ble->vendor ||
- !regexec(&ble->vendor_reg, vendor, 0, NULL, 0)) &&
+ !!regexec(&ble->vendor_reg, vendor, 0, NULL, 0) ==
+ ble->vendor_invert) &&
(!ble->product ||
- !regexec(&ble->product_reg, product, 0, NULL, 0)))
+ !!regexec(&ble->product_reg, product, 0, NULL, 0) ==
+ ble->product_invert))
return 1;
}
return 0;
@@ -160,13 +181,7 @@ setup_default_blist (struct config * conf)
char * str;
int i;
- str = STRDUP("^(ram|zram|raw|loop|fd|md|dm-|sr|scd|st|dcssblk)[0-9]");
- if (!str)
- return 1;
- if (store_ble(conf->blist_devnode, str, ORIGIN_DEFAULT))
- return 1;
-
- str = STRDUP("^(td|hd|vd)[a-z]");
+ str = STRDUP("!^(sd[a-z]|dasd[a-z]|nvme[0-9])");
if (!str)
return 1;
if (store_ble(conf->blist_devnode, str, ORIGIN_DEFAULT))
diff --git a/libmultipath/blacklist.h b/libmultipath/blacklist.h
index 2d721f60..4305857d 100644
--- a/libmultipath/blacklist.h
+++ b/libmultipath/blacklist.h
@@ -20,6 +20,7 @@
struct blentry {
char * str;
regex_t regex;
+ bool invert;
int origin;
};
@@ -28,6 +29,8 @@ struct blentry_device {
char * product;
regex_t vendor_reg;
regex_t product_reg;
+ bool vendor_invert;
+ bool product_invert;
int origin;
};
diff --git a/multipath/multipath.conf.5 b/multipath/multipath.conf.5
index 3455b1cc..6dc26f10 100644
--- a/multipath/multipath.conf.5
+++ b/multipath/multipath.conf.5
@@ -1248,6 +1248,16 @@ being handled by multipath-tools.
.LP
.
.
+In the \fIblacklist\fR and \fIblacklist_exceptions\fR sections, starting a
+quoted value with an exclamation mark \fB"!"\fR will invert the matching
+of the rest of the regular expression. For instance, \fB"!^sd[a-z]"\fR will
+match all values that do not start with \fB"sd[a-z]"\fR. The exclamation mark
+can be escaped \fB"\\!"\fR to match a literal \fB!\fR at the start of a
+regular expression. \fBNote:\fR The exclamation mark must be inside quotes,
+otherwise it will be treated as starting a comment.
+.LP
+.
+.
The \fIblacklist_exceptions\fR section is used to revert the actions of the
\fIblacklist\fR section. This allows one to selectively include ("whitelist") devices which
would normally be excluded via the \fIblacklist\fR section. A common usage is
@@ -1264,10 +1274,9 @@ unless explicitly stated.
Regular expression matching the device nodes to be excluded/included.
.RS
.PP
-The default \fIblacklist\fR consists of the regular expressions
-"^(ram|zram|raw|loop|fd|md|dm-|sr|scd|st|dcssblk)[0-9]" and
-"^(td|hd|vd)[a-z]". This causes virtual devices, non-disk devices, and some other
-device types to be excluded from multipath handling by default.
+The default \fIblacklist\fR consists of the regular expression
+\fB"!^(sd[a-z]|dasd[a-z]|nvme[0-9])"\fR. This causes all device types other
+than scsi, dasd, and nvme to be excluded from multipath handling by default.
.RE
.TP
.B wwid
diff --git a/tests/blacklist.c b/tests/blacklist.c
index cc8a9a4a..d20e97af 100644
--- a/tests/blacklist.c
+++ b/tests/blacklist.c
@@ -60,20 +60,46 @@ __wrap_udev_list_entry_get_name(struct udev_list_entry *list_entry)
return *(const char **)list_entry;
}
+vector elist_property_default;
+vector blist_devnode_default;
vector blist_devnode_sdb;
+vector blist_devnode_sdb_inv;
vector blist_all;
vector blist_device_foo_bar;
+vector blist_device_foo_inv_bar;
+vector blist_device_foo_bar_inv;
vector blist_device_all;
vector blist_wwid_xyzzy;
+vector blist_wwid_xyzzy_inv;
vector blist_protocol_fcp;
+vector blist_protocol_fcp_inv;
vector blist_property_wwn;
+vector blist_property_wwn_inv;
static int setup(void **state)
{
+ struct config conf;
+
+ memset(&conf, 0, sizeof(conf));
+ conf.blist_devnode = vector_alloc();
+ if (!conf.blist_devnode)
+ return -1;
+ conf.elist_property = vector_alloc();
+ if (!conf.elist_property)
+ return -1;
+ if (setup_default_blist(&conf) != 0)
+ return -1;
+ elist_property_default = conf.elist_property;
+ blist_devnode_default = conf.blist_devnode;
+
blist_devnode_sdb = vector_alloc();
if (!blist_devnode_sdb ||
store_ble(blist_devnode_sdb, strdup("sdb"), ORIGIN_CONFIG))
return -1;
+ blist_devnode_sdb_inv = vector_alloc();
+ if (!blist_devnode_sdb_inv ||
+ store_ble(blist_devnode_sdb_inv, strdup("!sdb"), ORIGIN_CONFIG))
+ return -1;
blist_all = vector_alloc();
if (!blist_all || store_ble(blist_all, strdup(".*"), ORIGIN_CONFIG))
@@ -84,6 +110,18 @@ static int setup(void **state)
set_ble_device(blist_device_foo_bar, strdup("foo"), strdup("bar"),
ORIGIN_CONFIG))
return -1;
+ blist_device_foo_inv_bar = vector_alloc();
+ if (!blist_device_foo_inv_bar ||
+ alloc_ble_device(blist_device_foo_inv_bar) ||
+ set_ble_device(blist_device_foo_inv_bar, strdup("!foo"),
+ strdup("bar"), ORIGIN_CONFIG))
+ return -1;
+ blist_device_foo_bar_inv = vector_alloc();
+ if (!blist_device_foo_bar_inv ||
+ alloc_ble_device(blist_device_foo_bar_inv) ||
+ set_ble_device(blist_device_foo_bar_inv, strdup("foo"),
+ strdup("!bar"), ORIGIN_CONFIG))
+ return -1;
blist_device_all = vector_alloc();
if (!blist_device_all || alloc_ble_device(blist_device_all) ||
@@ -95,29 +133,50 @@ static int setup(void **state)
if (!blist_wwid_xyzzy ||
store_ble(blist_wwid_xyzzy, strdup("xyzzy"), ORIGIN_CONFIG))
return -1;
+ blist_wwid_xyzzy_inv = vector_alloc();
+ if (!blist_wwid_xyzzy_inv ||
+ store_ble(blist_wwid_xyzzy_inv, strdup("!xyzzy"), ORIGIN_CONFIG))
+ return -1;
blist_protocol_fcp = vector_alloc();
if (!blist_protocol_fcp ||
store_ble(blist_protocol_fcp, strdup("scsi:fcp"), ORIGIN_CONFIG))
return -1;
+ blist_protocol_fcp_inv = vector_alloc();
+ if (!blist_protocol_fcp_inv ||
+ store_ble(blist_protocol_fcp_inv, strdup("!scsi:fcp"),
+ ORIGIN_CONFIG))
+ return -1;
blist_property_wwn = vector_alloc();
if (!blist_property_wwn ||
store_ble(blist_property_wwn, strdup("ID_WWN"), ORIGIN_CONFIG))
return -1;
+ blist_property_wwn_inv = vector_alloc();
+ if (!blist_property_wwn_inv ||
+ store_ble(blist_property_wwn_inv, strdup("!ID_WWN"), ORIGIN_CONFIG))
+ return -1;
return 0;
}
static int teardown(void **state)
{
+ free_blacklist(elist_property_default);
+ free_blacklist(blist_devnode_default);
free_blacklist(blist_devnode_sdb);
+ free_blacklist(blist_devnode_sdb_inv);
free_blacklist(blist_all);
free_blacklist_device(blist_device_foo_bar);
+ free_blacklist_device(blist_device_foo_inv_bar);
+ free_blacklist_device(blist_device_foo_bar_inv);
free_blacklist_device(blist_device_all);
free_blacklist(blist_wwid_xyzzy);
+ free_blacklist(blist_wwid_xyzzy_inv);
free_blacklist(blist_protocol_fcp);
+ free_blacklist(blist_protocol_fcp_inv);
free_blacklist(blist_property_wwn);
+ free_blacklist(blist_property_wwn_inv);
return 0;
}
@@ -141,6 +200,11 @@ static void test_devnode_blacklist(void **state)
expect_condlog(3, "sdb: device node name blacklisted\n");
assert_int_equal(filter_devnode(blist_devnode_sdb, NULL, "sdb"),
MATCH_DEVNODE_BLIST);
+ assert_int_equal(filter_devnode(blist_devnode_sdb_inv, NULL, "sdb"),
+ MATCH_NOTHING);
+ expect_condlog(3, "sdc: device node name blacklisted\n");
+ assert_int_equal(filter_devnode(blist_devnode_sdb_inv, NULL, "sdc"),
+ MATCH_DEVNODE_BLIST);
}
static void test_devnode_whitelist(void **state)
@@ -159,12 +223,39 @@ static void test_devnode_missing(void **state)
MATCH_NOTHING);
}
+static void test_devnode_default(void **state)
+{
+ assert_int_equal(filter_devnode(blist_devnode_default, NULL, "sdaa"),
+ MATCH_NOTHING);
+ assert_int_equal(filter_devnode(blist_devnode_default, NULL, "nvme0n1"),
+ MATCH_NOTHING);
+ assert_int_equal(filter_devnode(blist_devnode_default, NULL, "dasda"),
+ MATCH_NOTHING);
+ expect_condlog(3, "hda: device node name blacklisted\n");
+ assert_int_equal(filter_devnode(blist_devnode_default, NULL, "hda"),
+ MATCH_DEVNODE_BLIST);
+}
+
static void test_device_blacklist(void **state)
{
expect_condlog(3, "sdb: (foo:bar) vendor/product blacklisted\n");
assert_int_equal(filter_device(blist_device_foo_bar, NULL, "foo",
"bar", "sdb"),
MATCH_DEVICE_BLIST);
+ assert_int_equal(filter_device(blist_device_foo_inv_bar, NULL, "foo",
+ "bar", "sdb"),
+ MATCH_NOTHING);
+ assert_int_equal(filter_device(blist_device_foo_bar_inv, NULL, "foo",
+ "bar", "sdb"),
+ MATCH_NOTHING);
+ expect_condlog(3, "sdb: (baz:bar) vendor/product blacklisted\n");
+ assert_int_equal(filter_device(blist_device_foo_inv_bar, NULL, "baz",
+ "bar", "sdb"),
+ MATCH_DEVICE_BLIST);
+ expect_condlog(3, "sdb: (foo:baz) vendor/product blacklisted\n");
+ assert_int_equal(filter_device(blist_device_foo_bar_inv, NULL, "foo",
+ "baz", "sdb"),
+ MATCH_DEVICE_BLIST);
}
static void test_device_whitelist(void **state)
@@ -191,6 +282,11 @@ static void test_wwid_blacklist(void **state)
expect_condlog(3, "sdb: wwid xyzzy blacklisted\n");
assert_int_equal(filter_wwid(blist_wwid_xyzzy, NULL, "xyzzy", "sdb"),
MATCH_WWID_BLIST);
+ assert_int_equal(filter_wwid(blist_wwid_xyzzy_inv, NULL, "xyzzy",
+ "sdb"), MATCH_NOTHING);
+ expect_condlog(3, "sdb: wwid plugh blacklisted\n");
+ assert_int_equal(filter_wwid(blist_wwid_xyzzy_inv, NULL, "plugh",
+ "sdb"), MATCH_WWID_BLIST);
}
static void test_wwid_whitelist(void **state)
@@ -218,6 +314,12 @@ static void test_protocol_blacklist(void **state)
expect_condlog(3, "sdb: protocol scsi:fcp blacklisted\n");
assert_int_equal(filter_protocol(blist_protocol_fcp, NULL, &pp),
MATCH_PROTOCOL_BLIST);
+ assert_int_equal(filter_protocol(blist_protocol_fcp_inv, NULL, &pp),
+ MATCH_NOTHING);
+ pp.sg_id.proto_id = SCSI_PROTOCOL_ATA;
+ expect_condlog(3, "sdb: protocol scsi:ata blacklisted\n");
+ assert_int_equal(filter_protocol(blist_protocol_fcp_inv, NULL, &pp),
+ MATCH_PROTOCOL_BLIST);
}
static void test_protocol_whitelist(void **state)
@@ -245,10 +347,17 @@ static void test_protocol_missing(void **state)
static void test_property_blacklist(void **state)
{
static struct udev_device udev = { "sdb", { "ID_FOO", "ID_WWN", "ID_BAR", NULL } };
+ static struct udev_device udev_inv = { "sdb", { "ID_WWN", NULL } };
conf.blist_property = blist_property_wwn;
expect_condlog(3, "sdb: udev property ID_WWN blacklisted\n");
assert_int_equal(filter_property(&conf, &udev, 3, "ID_SERIAL"),
MATCH_PROPERTY_BLIST);
+ conf.blist_property = blist_property_wwn_inv;
+ expect_condlog(3, "sdb: udev property ID_FOO blacklisted\n");
+ assert_int_equal(filter_property(&conf, &udev, 3, "ID_SERIAL"),
+ MATCH_PROPERTY_BLIST);
+ assert_int_equal(filter_property(&conf, &udev_inv, 3, "ID_SERIAL"),
+ MATCH_NOTHING);
}
/* the property check works different in that you check all the property
@@ -482,6 +591,7 @@ int test_blacklist(void)
cmocka_unit_test(test_devnode_blacklist),
cmocka_unit_test(test_devnode_whitelist),
cmocka_unit_test(test_devnode_missing),
+ cmocka_unit_test(test_devnode_default),
cmocka_unit_test(test_device_blacklist),
cmocka_unit_test(test_device_whitelist),
cmocka_unit_test(test_device_missing),
diff --git a/tests/test-lib.c b/tests/test-lib.c
index 59275163..08ff2d8d 100644
--- a/tests/test-lib.c
+++ b/tests/test-lib.c
@@ -15,7 +15,7 @@
#include "test-lib.h"
const int default_mask = (DI_SYSFS|DI_BLACKLIST|DI_WWID|DI_CHECKER|DI_PRIO);
-const char default_devnode[] = "sdTEST";
+const char default_devnode[] = "sdxTEST";
const char default_wwid[] = "TEST-WWID";
/* default_wwid should be a substring of default_wwid_1! */
const char default_wwid_1[] = "TEST-WWID-1";
--
2.17.2

322
SOURCES/0024-libmultipath-make-dm_get_map-status-return-codes-sym.patch Normal file
View File

@ -0,0 +1,322 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue, 23 Jun 2020 22:17:31 -0500
Subject: [PATCH] libmultipath: make dm_get_map/status return codes symbolic
dm_get_map() and dm_get_status() now use symbolic return codes. They
also differentiate between failing to get information from device-mapper
and not finding the requested device. These symboilc return codes are
also used by update_multipath_* functions.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/devmapper.c | 51 +++++++++++++++++++++++++-------------
libmultipath/devmapper.h | 6 +++++
libmultipath/structs_vec.c | 45 +++++++++++++++++++--------------
multipathd/main.c | 12 ++++-----
4 files changed, 72 insertions(+), 42 deletions(-)
diff --git a/libmultipath/devmapper.c b/libmultipath/devmapper.c
index 13a1cf53..f6204e5f 100644
--- a/libmultipath/devmapper.c
+++ b/libmultipath/devmapper.c
@@ -525,36 +525,43 @@ int dm_map_present(const char * str)
int dm_get_map(const char *name, unsigned long long *size, char *outparams)
{
- int r = 1;
+ int r = DMP_ERR;
struct dm_task *dmt;
uint64_t start, length;
char *target_type = NULL;
char *params = NULL;
if (!(dmt = libmp_dm_task_create(DM_DEVICE_TABLE)))
- return 1;
+ return r;
if (!dm_task_set_name(dmt, name))
goto out;
dm_task_no_open_count(dmt);
- if (!dm_task_run(dmt))
+ errno = 0;
+ if (!dm_task_run(dmt)) {
+ if (dm_task_get_errno(dmt) == ENXIO)
+ r = DMP_NOT_FOUND;
goto out;
+ }
+ r = DMP_NOT_FOUND;
/* Fetch 1st target */
- dm_get_next_target(dmt, NULL, &start, &length,
- &target_type, &params);
+ if (dm_get_next_target(dmt, NULL, &start, &length,
+ &target_type, &params) != NULL)
+ /* more than one target */
+ goto out;
if (size)
*size = length;
if (!outparams) {
- r = 0;
+ r = DMP_OK;
goto out;
}
if (snprintf(outparams, PARAMS_SIZE, "%s", params) <= PARAMS_SIZE)
- r = 0;
+ r = DMP_OK;
out:
dm_task_destroy(dmt);
return r;
@@ -628,35 +635,45 @@ is_mpath_part(const char *part_name, const char *map_name)
int dm_get_status(const char *name, char *outstatus)
{
- int r = 1;
+ int r = DMP_ERR;
struct dm_task *dmt;
uint64_t start, length;
char *target_type = NULL;
char *status = NULL;
if (!(dmt = libmp_dm_task_create(DM_DEVICE_STATUS)))
- return 1;
+ return r;
if (!dm_task_set_name(dmt, name))
goto out;
dm_task_no_open_count(dmt);
- if (!dm_task_run(dmt))
+ errno = 0;
+ if (!dm_task_run(dmt)) {
+ if (dm_task_get_errno(dmt) == ENXIO)
+ r = DMP_NOT_FOUND;
goto out;
+ }
+ r = DMP_NOT_FOUND;
/* Fetch 1st target */
- dm_get_next_target(dmt, NULL, &start, &length,
- &target_type, &status);
+ if (dm_get_next_target(dmt, NULL, &start, &length,
+ &target_type, &status) != NULL)
+ goto out;
+
+ if (!target_type || strcmp(target_type, TGT_MPATH) != 0)
+ goto out;
+
if (!status) {
condlog(2, "get null status.");
goto out;
}
if (snprintf(outstatus, PARAMS_SIZE, "%s", status) <= PARAMS_SIZE)
- r = 0;
+ r = DMP_OK;
out:
- if (r)
+ if (r != DMP_OK)
condlog(0, "%s: error getting map status string", name);
dm_task_destroy(dmt);
@@ -866,7 +883,7 @@ int _dm_flush_map (const char * mapname, int need_sync, int deferred_remove,
return 1;
if (need_suspend &&
- !dm_get_map(mapname, &mapsize, params) &&
+ dm_get_map(mapname, &mapsize, params) == DMP_OK &&
strstr(params, "queue_if_no_path")) {
if (!dm_queue_if_no_path(mapname, 0))
queue_if_no_path = 1;
@@ -1075,7 +1092,7 @@ struct multipath *dm_get_multipath(const char *name)
if (!mpp->alias)
goto out;
- if (dm_get_map(name, &mpp->size, NULL))
+ if (dm_get_map(name, &mpp->size, NULL) != DMP_OK)
goto out;
dm_get_uuid(name, mpp->wwid, WWID_SIZE);
@@ -1259,7 +1276,7 @@ do_foreach_partmaps (const char * mapname,
/*
* and we can fetch the map table from the kernel
*/
- !dm_get_map(names->name, &size, &params[0]) &&
+ dm_get_map(names->name, &size, &params[0]) == DMP_OK &&
/*
* and the table maps over the multipath map
diff --git a/libmultipath/devmapper.h b/libmultipath/devmapper.h
index 7557a86b..adb55000 100644
--- a/libmultipath/devmapper.h
+++ b/libmultipath/devmapper.h
@@ -27,6 +27,12 @@
#define UUID_PREFIX "mpath-"
#define UUID_PREFIX_LEN (sizeof(UUID_PREFIX) - 1)
+enum {
+ DMP_ERR,
+ DMP_OK,
+ DMP_NOT_FOUND,
+};
+
void dm_init(int verbosity);
void libmp_dm_init(void);
void libmp_udev_set_sync_support(int on);
diff --git a/libmultipath/structs_vec.c b/libmultipath/structs_vec.c
index 077f2e42..8137ea21 100644
--- a/libmultipath/structs_vec.c
+++ b/libmultipath/structs_vec.c
@@ -196,43 +196,47 @@ extract_hwe_from_path(struct multipath * mpp)
int
update_multipath_table (struct multipath *mpp, vector pathvec, int is_daemon)
{
+ int r = DMP_ERR;
char params[PARAMS_SIZE] = {0};
if (!mpp)
- return 1;
+ return r;
- if (dm_get_map(mpp->alias, &mpp->size, params)) {
- condlog(3, "%s: cannot get map", mpp->alias);
- return 1;
+ r = dm_get_map(mpp->alias, &mpp->size, params);
+ if (r != DMP_OK) {
+ condlog(3, "%s: %s", mpp->alias, (r == DMP_ERR)? "error getting table" : "map not present");
+ return r;
}
if (disassemble_map(pathvec, params, mpp, is_daemon)) {
condlog(3, "%s: cannot disassemble map", mpp->alias);
- return 1;
+ return DMP_ERR;
}
- return 0;
+ return DMP_OK;
}
int
update_multipath_status (struct multipath *mpp)
{
+ int r = DMP_ERR;
char status[PARAMS_SIZE] = {0};
if (!mpp)
- return 1;
+ return r;
- if (dm_get_status(mpp->alias, status)) {
- condlog(3, "%s: cannot get status", mpp->alias);
- return 1;
+ r = dm_get_status(mpp->alias, status);
+ if (r != DMP_OK) {
+ condlog(3, "%s: %s", mpp->alias, (r == DMP_ERR)? "error getting status" : "map not present");
+ return r;
}
if (disassemble_status(status, mpp)) {
condlog(3, "%s: cannot disassemble status", mpp->alias);
- return 1;
+ return DMP_ERR;
}
- return 0;
+ return DMP_OK;
}
void sync_paths(struct multipath *mpp, vector pathvec)
@@ -264,10 +268,10 @@ int
update_multipath_strings(struct multipath *mpp, vector pathvec, int is_daemon)
{
struct pathgroup *pgp;
- int i;
+ int i, r = DMP_ERR;
if (!mpp)
- return 1;
+ return r;
update_mpp_paths(mpp, pathvec);
condlog(4, "%s: %s", mpp->alias, __FUNCTION__);
@@ -276,18 +280,21 @@ update_multipath_strings(struct multipath *mpp, vector pathvec, int is_daemon)
free_pgvec(mpp->pg, KEEP_PATHS);
mpp->pg = NULL;
- if (update_multipath_table(mpp, pathvec, is_daemon))
- return 1;
+ r = update_multipath_table(mpp, pathvec, is_daemon);
+ if (r != DMP_OK)
+ return r;
+
sync_paths(mpp, pathvec);
- if (update_multipath_status(mpp))
- return 1;
+ r = update_multipath_status(mpp);
+ if (r != DMP_OK)
+ return r;
vector_foreach_slot(mpp->pg, pgp, i)
if (pgp->paths)
path_group_prio_update(pgp);
- return 0;
+ return DMP_OK;
}
static void enter_recovery_mode(struct multipath *mpp)
diff --git a/multipathd/main.c b/multipathd/main.c
index 6b7db2c0..e3427d3d 100644
--- a/multipathd/main.c
+++ b/multipathd/main.c
@@ -414,7 +414,7 @@ int __setup_multipath(struct vectors *vecs, struct multipath *mpp,
goto out;
}
- if (update_multipath_strings(mpp, vecs->pathvec, 1)) {
+ if (update_multipath_strings(mpp, vecs->pathvec, 1) != DMP_OK) {
condlog(0, "%s: failed to setup multipath", mpp->alias);
goto out;
}
@@ -553,9 +553,9 @@ add_map_without_path (struct vectors *vecs, const char *alias)
mpp->mpe = find_mpe(conf->mptable, mpp->wwid);
put_multipath_config(conf);
- if (update_multipath_table(mpp, vecs->pathvec, 1))
+ if (update_multipath_table(mpp, vecs->pathvec, 1) != DMP_OK)
goto out;
- if (update_multipath_status(mpp))
+ if (update_multipath_status(mpp) != DMP_OK)
goto out;
if (!vector_alloc_slot(vecs->mpvec))
@@ -1346,8 +1346,8 @@ map_discovery (struct vectors * vecs)
return 1;
vector_foreach_slot (vecs->mpvec, mpp, i)
- if (update_multipath_table(mpp, vecs->pathvec, 1) ||
- update_multipath_status(mpp)) {
+ if (update_multipath_table(mpp, vecs->pathvec, 1) != DMP_OK ||
+ update_multipath_status(mpp) != DMP_OK) {
remove_map(mpp, vecs, 1);
i--;
}
@@ -2087,7 +2087,7 @@ check_path (struct vectors * vecs, struct path * pp, unsigned int ticks)
/*
* Synchronize with kernel state
*/
- if (update_multipath_strings(pp->mpp, vecs->pathvec, 1)) {
+ if (update_multipath_strings(pp->mpp, vecs->pathvec, 1) != DMP_OK) {
condlog(1, "%s: Could not synchronize with kernel state",
pp->dev);
pp->dmstate = PSTATE_UNDEF;
--
2.17.2

116
SOURCES/0025-multipathd-fix-check_path-errors-with-removed-map.patch Normal file
View File

@ -0,0 +1,116 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 11 Jun 2020 15:41:18 -0500
Subject: [PATCH] multipathd: fix check_path errors with removed map
If a multipath device is removed during, or immediately before the call
to check_path(), multipathd can behave incorrectly. A missing multpath
device will cause update_multipath_strings() to fail, setting
pp->dmstate to PSTATE_UNDEF. If the path is up, this state will cause
reinstate_path() to be called, which will also fail. This will trigger
a reload, restoring the recently removed device.
If update_multipath_strings() fails because there is no multipath
device, check_path should just quit, since the remove dmevent and uevent
are likely already queued up. Also, I don't see any reason to reload the
multipath device if reinstate fails. This code was added by
fac68d7a99ef17d496079538a5c6836acd7911ab, which clamined that reinstate
could fail if the path was disabled. Looking through the current kernel
code, I can't see any reason why a reinstate would fail, where a reload
would help. If the path was missing from the multipath device,
update_multipath_strings() would already catch that, and quit
check_path() early, which make more sense to me than reloading does.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
multipathd/main.c | 44 +++++++++++++++++++-------------------------
1 file changed, 19 insertions(+), 25 deletions(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index e3427d3d..1d9ce7f7 100644
--- a/multipathd/main.c
+++ b/multipathd/main.c
@@ -1611,22 +1611,18 @@ fail_path (struct path * pp, int del_active)
/*
* caller must have locked the path list before calling that function
*/
-static int
+static void
reinstate_path (struct path * pp)
{
- int ret = 0;
-
if (!pp->mpp)
- return 0;
+ return;
- if (dm_reinstate_path(pp->mpp->alias, pp->dev_t)) {
+ if (dm_reinstate_path(pp->mpp->alias, pp->dev_t))
condlog(0, "%s: reinstate failed", pp->dev_t);
- ret = 1;
- } else {
+ else {
condlog(2, "%s: reinstated", pp->dev_t);
update_queue_mode_add_path(pp->mpp);
}
- return ret;
}
static void
@@ -2087,9 +2083,16 @@ check_path (struct vectors * vecs, struct path * pp, unsigned int ticks)
/*
* Synchronize with kernel state
*/
- if (update_multipath_strings(pp->mpp, vecs->pathvec, 1) != DMP_OK) {
- condlog(1, "%s: Could not synchronize with kernel state",
- pp->dev);
+ ret = update_multipath_strings(pp->mpp, vecs->pathvec, 1);
+ if (ret != DMP_OK) {
+ if (ret == DMP_NOT_FOUND) {
+ /* multipath device missing. Likely removed */
+ condlog(1, "%s: multipath device '%s' not found",
+ pp->dev, pp->mpp->alias);
+ return 0;
+ } else
+ condlog(1, "%s: Couldn't synchronize with kernel state",
+ pp->dev);
pp->dmstate = PSTATE_UNDEF;
}
/* if update_multipath_strings orphaned the path, quit early */
@@ -2179,12 +2182,8 @@ check_path (struct vectors * vecs, struct path * pp, unsigned int ticks)
/*
* reinstate this path
*/
- if (!disable_reinstate && reinstate_path(pp)) {
- condlog(3, "%s: reload map", pp->dev);
- ev_add_path(pp, vecs, 1);
- pp->tick = 1;
- return 0;
- }
+ if (!disable_reinstate)
+ reinstate_path(pp);
new_path_up = 1;
if (oldchkrstate != PATH_UP && oldchkrstate != PATH_GHOST)
@@ -2200,15 +2199,10 @@ check_path (struct vectors * vecs, struct path * pp, unsigned int ticks)
else if (newstate == PATH_UP || newstate == PATH_GHOST) {
if ((pp->dmstate == PSTATE_FAILED ||
pp->dmstate == PSTATE_UNDEF) &&
- !disable_reinstate) {
+ !disable_reinstate)
/* Clear IO errors */
- if (reinstate_path(pp)) {
- condlog(3, "%s: reload map", pp->dev);
- ev_add_path(pp, vecs, 1);
- pp->tick = 1;
- return 0;
- }
- } else {
+ reinstate_path(pp);
+ else {
LOG_MSG(4, verbosity, pp);
if (pp->checkint != max_checkint) {
/*
--
2.17.2

45
SOURCES/0026-libmultipath-make-dm_flush_maps-only-return-0-on-suc.patch Normal file
View File

@ -0,0 +1,45 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Wed, 17 Jun 2020 13:31:37 -0500
Subject: [PATCH] libmultipath: make dm_flush_maps only return 0 on success
dm_flush_maps() returned both 0 and 1 on error, depending on which part
of the function it was in, but the caller was always treating 0 as a
success. Make dm_flush_maps() always return 1 on error and 0 on success.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/devmapper.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/libmultipath/devmapper.c b/libmultipath/devmapper.c
index f6204e5f..cda83ce4 100644
--- a/libmultipath/devmapper.c
+++ b/libmultipath/devmapper.c
@@ -953,13 +953,13 @@ dm_flush_map_nopaths(const char * mapname, int deferred_remove)
int dm_flush_maps (int retries)
{
- int r = 0;
+ int r = 1;
struct dm_task *dmt;
struct dm_names *names;
unsigned next = 0;
if (!(dmt = libmp_dm_task_create (DM_DEVICE_LIST)))
- return 0;
+ return r;
dm_task_no_open_count(dmt);
@@ -972,6 +972,7 @@ int dm_flush_maps (int retries)
if (!names->dev)
goto out;
+ r = 0;
do {
r |= dm_suspend_and_flush_map(names->name, retries);
next = names->next;
--
2.17.2

160
SOURCES/0027-multipathd-add-del-maps-multipathd-command.patch Normal file
View File

@ -0,0 +1,160 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 15 Jun 2020 17:00:54 -0500
Subject: [PATCH] multipathd: add "del maps" multipathd command
This will flush all multipath devices.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/devmapper.c | 7 +++++--
libmultipath/devmapper.h | 2 +-
multipath/main.c | 2 +-
multipathd/cli.c | 1 +
multipathd/cli_handlers.c | 19 +++++++++++++++++++
multipathd/cli_handlers.h | 1 +
multipathd/main.c | 3 ++-
multipathd/main.h | 1 +
8 files changed, 31 insertions(+), 5 deletions(-)
diff --git a/libmultipath/devmapper.c b/libmultipath/devmapper.c
index cda83ce4..7f98bf9d 100644
--- a/libmultipath/devmapper.c
+++ b/libmultipath/devmapper.c
@@ -951,7 +951,7 @@ dm_flush_map_nopaths(const char * mapname, int deferred_remove)
#endif
-int dm_flush_maps (int retries)
+int dm_flush_maps (int need_suspend, int retries)
{
int r = 1;
struct dm_task *dmt;
@@ -974,7 +974,10 @@ int dm_flush_maps (int retries)
r = 0;
do {
- r |= dm_suspend_and_flush_map(names->name, retries);
+ if (need_suspend)
+ r |= dm_suspend_and_flush_map(names->name, retries);
+ else
+ r |= dm_flush_map(names->name);
next = names->next;
names = (void *) names + next;
} while (next);
diff --git a/libmultipath/devmapper.h b/libmultipath/devmapper.h
index adb55000..7e8812ad 100644
--- a/libmultipath/devmapper.h
+++ b/libmultipath/devmapper.h
@@ -55,7 +55,7 @@ int dm_flush_map_nopaths(const char * mapname, int deferred_remove);
#define dm_suspend_and_flush_map(mapname, retries) \
_dm_flush_map(mapname, 1, 0, 1, retries)
int dm_cancel_deferred_remove(struct multipath *mpp);
-int dm_flush_maps (int retries);
+int dm_flush_maps (int need_suspend, int retries);
int dm_fail_path(const char * mapname, char * path);
int dm_reinstate_path(const char * mapname, char * path);
int dm_queue_if_no_path(const char *mapname, int enable);
diff --git a/multipath/main.c b/multipath/main.c
index 78822ee1..7ab3102f 100644
--- a/multipath/main.c
+++ b/multipath/main.c
@@ -1127,7 +1127,7 @@ main (int argc, char *argv[])
goto out;
}
else if (conf->remove == FLUSH_ALL) {
- r = dm_flush_maps(retries) ? RTVL_FAIL : RTVL_OK;
+ r = dm_flush_maps(1, retries) ? RTVL_FAIL : RTVL_OK;
goto out;
}
while ((r = configure(conf, cmd, dev_type, dev)) == RTVL_RETRY)
diff --git a/multipathd/cli.c b/multipathd/cli.c
index 800c0fbe..bdc9fb10 100644
--- a/multipathd/cli.c
+++ b/multipathd/cli.c
@@ -568,6 +568,7 @@ cli_init (void) {
add_handler(DEL+PATH, NULL);
add_handler(ADD+MAP, NULL);
add_handler(DEL+MAP, NULL);
+ add_handler(DEL+MAPS, NULL);
add_handler(SWITCH+MAP+GROUP, NULL);
add_handler(RECONFIGURE, NULL);
add_handler(SUSPEND+MAP, NULL);
diff --git a/multipathd/cli_handlers.c b/multipathd/cli_handlers.c
index 31c3d9fd..782bb003 100644
--- a/multipathd/cli_handlers.c
+++ b/multipathd/cli_handlers.c
@@ -852,6 +852,25 @@ cli_del_map (void * v, char ** reply, int * len, void * data)
return rc;
}
+int
+cli_del_maps (void *v, char **reply, int *len, void *data)
+{
+ struct vectors * vecs = (struct vectors *)data;
+ struct multipath *mpp;
+ int i, ret = 0;
+
+ condlog(2, "remove maps (operator)");
+ vector_foreach_slot(vecs->mpvec, mpp, i) {
+ if (flush_map(mpp, vecs, 0))
+ ret++;
+ else
+ i--;
+ }
+ /* flush any multipath maps that aren't currently known by multipathd */
+ ret |= dm_flush_maps(0, 0);
+ return ret;
+}
+
int
cli_reload(void *v, char **reply, int *len, void *data)
{
diff --git a/multipathd/cli_handlers.h b/multipathd/cli_handlers.h
index 0f451064..6f57b429 100644
--- a/multipathd/cli_handlers.h
+++ b/multipathd/cli_handlers.h
@@ -26,6 +26,7 @@ int cli_add_path (void * v, char ** reply, int * len, void * data);
int cli_del_path (void * v, char ** reply, int * len, void * data);
int cli_add_map (void * v, char ** reply, int * len, void * data);
int cli_del_map (void * v, char ** reply, int * len, void * data);
+int cli_del_maps (void * v, char ** reply, int * len, void * data);
int cli_switch_group(void * v, char ** reply, int * len, void * data);
int cli_reconfigure(void * v, char ** reply, int * len, void * data);
int cli_resize(void * v, char ** reply, int * len, void * data);
diff --git a/multipathd/main.c b/multipathd/main.c
index 1d9ce7f7..1d0579e9 100644
--- a/multipathd/main.c
+++ b/multipathd/main.c
@@ -631,7 +631,7 @@ sync_maps_state(vector mpvec)
sync_map_state(mpp);
}
-static int
+int
flush_map(struct multipath * mpp, struct vectors * vecs, int nopaths)
{
int r;
@@ -1551,6 +1551,7 @@ uxlsnrloop (void * ap)
set_handler_callback(DEL+PATH, cli_del_path);
set_handler_callback(ADD+MAP, cli_add_map);
set_handler_callback(DEL+MAP, cli_del_map);
+ set_handler_callback(DEL+MAPS, cli_del_maps);
set_handler_callback(SWITCH+MAP+GROUP, cli_switch_group);
set_unlocked_handler_callback(RECONFIGURE, cli_reconfigure);
set_handler_callback(SUSPEND+MAP, cli_suspend);
diff --git a/multipathd/main.h b/multipathd/main.h
index 7bb8463f..5dff17e5 100644
--- a/multipathd/main.h
+++ b/multipathd/main.h
@@ -28,6 +28,7 @@ int ev_add_path (struct path *, struct vectors *, int);
int ev_remove_path (struct path *, struct vectors *, int);
int ev_add_map (char *, const char *, struct vectors *);
int ev_remove_map (char *, char *, int, struct vectors *);
+int flush_map(struct multipath *, struct vectors *, int);
int set_config_state(enum daemon_status);
void * mpath_alloc_prin_response(int prin_sa);
int prin_do_scsi_ioctl(char *, int rq_servact, struct prin_resp * resp,
--
2.17.2

103
SOURCES/0028-multipath-make-flushing-maps-work-like-other-command.patch Normal file
View File

@ -0,0 +1,103 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 15 Jun 2020 23:54:29 -0500
Subject: [PATCH] multipath: make flushing maps work like other commands
The config structure doesn't need a special variable just for removes.
Multipath can just use the cmd variable, like it does for the other
commands.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/config.h | 3 ++-
libmultipath/configure.h | 3 ---
multipath/main.c | 20 ++++++++++----------
3 files changed, 12 insertions(+), 14 deletions(-)
diff --git a/libmultipath/config.h b/libmultipath/config.h
index 3368d8c9..4042eba6 100644
--- a/libmultipath/config.h
+++ b/libmultipath/config.h
@@ -39,6 +39,8 @@ enum mpath_cmds {
CMD_ADD_WWID,
CMD_USABLE_PATHS,
CMD_DUMP_CONFIG,
+ CMD_FLUSH_ONE,
+ CMD_FLUSH_ALL,
};
enum force_reload_types {
@@ -143,7 +145,6 @@ struct config {
unsigned int max_checkint;
bool use_watchdog;
int pgfailback;
- int remove;
int rr_weight;
int no_path_retry;
int user_friendly_names;
diff --git a/libmultipath/configure.h b/libmultipath/configure.h
index d7509000..0e33bf40 100644
--- a/libmultipath/configure.h
+++ b/libmultipath/configure.h
@@ -45,9 +45,6 @@ enum {
CP_RETRY,
};
-#define FLUSH_ONE 1
-#define FLUSH_ALL 2
-
struct vectors;
int setup_map (struct multipath * mpp, char * params, int params_size,
diff --git a/multipath/main.c b/multipath/main.c
index 7ab3102f..a2080029 100644
--- a/multipath/main.c
+++ b/multipath/main.c
@@ -942,10 +942,10 @@ main (int argc, char *argv[])
cmd = CMD_DRY_RUN;
break;
case 'f':
- conf->remove = FLUSH_ONE;
+ cmd = CMD_FLUSH_ONE;
break;
case 'F':
- conf->remove = FLUSH_ALL;
+ cmd = CMD_FLUSH_ALL;
break;
case 'l':
if (optarg && !strncmp(optarg, "l", 1))
@@ -1084,6 +1084,10 @@ main (int argc, char *argv[])
condlog(0, "the -w option requires a device");
goto out;
}
+ if (cmd == CMD_FLUSH_ONE && dev_type != DEV_DEVMAP) {
+ condlog(0, "the -f option requires a map name to remove");
+ goto out;
+ }
switch(delegate_to_multipathd(cmd, dev, dev_type, conf)) {
case DELEGATE_OK:
@@ -1117,16 +1121,12 @@ main (int argc, char *argv[])
}
if (retries < 0)
retries = conf->remove_retries;
- if (conf->remove == FLUSH_ONE) {
- if (dev_type == DEV_DEVMAP) {
- r = dm_suspend_and_flush_map(dev, retries) ?
- RTVL_FAIL : RTVL_OK;
- } else
- condlog(0, "must provide a map name to remove");
-
+ if (cmd == CMD_FLUSH_ONE) {
+ r = dm_suspend_and_flush_map(dev, retries) ?
+ RTVL_FAIL : RTVL_OK;
goto out;
}
- else if (conf->remove == FLUSH_ALL) {
+ else if (cmd == CMD_FLUSH_ALL) {
r = dm_flush_maps(1, retries) ? RTVL_FAIL : RTVL_OK;
goto out;
}
--
2.17.2

63
SOURCES/0029-multipath-delegate-flushing-maps-to-multipathd.patch Normal file
View File

@ -0,0 +1,63 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue, 16 Jun 2020 16:25:34 -0500
Subject: [PATCH] multipath: delegate flushing maps to multipathd
Since there can be problems with removing maps outside of multipathd,
multipath should attempt to delegate this command to multipathd.
However, multipathd doesn't attempt to suspend the device, in order
to avoid potential hangs. If delegating to multipathd fails, multipath
should try the remove itself.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
multipath/main.c | 14 ++++++++++++++
multipath/multipath.8 | 4 ++--
2 files changed, 16 insertions(+), 2 deletions(-)
diff --git a/multipath/main.c b/multipath/main.c
index a2080029..612c6815 100644
--- a/multipath/main.c
+++ b/multipath/main.c
@@ -828,6 +828,20 @@ int delegate_to_multipathd(enum mpath_cmds cmd,
if (cmd == CMD_CREATE && conf->force_reload == FORCE_RELOAD_YES) {
p += snprintf(p, n, "reconfigure");
}
+ else if (cmd == CMD_FLUSH_ONE && dev && dev_type == DEV_DEVMAP) {
+ p += snprintf(p, n, "del map %s", dev);
+ /* multipathd doesn't try as hard, to avoid potentially
+ * hanging. If it fails, retry with the regular multipath
+ * command */
+ r = NOT_DELEGATED;
+ }
+ else if (cmd == CMD_FLUSH_ALL) {
+ p += snprintf(p, n, "del maps");
+ /* multipathd doesn't try as hard, to avoid potentially
+ * hanging. If it fails, retry with the regular multipath
+ * command */
+ r = NOT_DELEGATED;
+ }
/* Add other translations here */
if (strlen(command) == 0)
diff --git a/multipath/multipath.8 b/multipath/multipath.8
index 8befc45a..47a33f9b 100644
--- a/multipath/multipath.8
+++ b/multipath/multipath.8
@@ -125,11 +125,11 @@ the system.
Other operation modes are chosen by using one of the following command line switches:
.TP
.B \-f
-Flush (remove) a multipath device map specified as parameter, if unused.
+Flush (remove) a multipath device map specified as parameter, if unused. This operation is delegated to the multipathd daemon if it's running.
.
.TP
.B \-F
-Flush (remove) all unused multipath device maps.
+Flush (remove) all unused multipath device maps. This operation is delegated to the multipathd daemon if it's running.
.
.TP
.B \-l
--
2.17.2

62
SOURCES/0030-multipath-add-option-to-skip-multipathd-delegation.patch Normal file
View File

@ -0,0 +1,62 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue, 16 Jun 2020 17:36:33 -0500
Subject: [PATCH] multipath: add option to skip multipathd delegation
Add the -D option to allow users to skip delegating commands to
multipathd.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/config.h | 1 +
multipath/main.c | 8 +++++++-
2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/libmultipath/config.h b/libmultipath/config.h
index 4042eba6..160867cd 100644
--- a/libmultipath/config.h
+++ b/libmultipath/config.h
@@ -191,6 +191,7 @@ struct config {
int ghost_delay;
int find_multipaths_timeout;
int marginal_pathgroups;
+ int skip_delegate;
unsigned int version[3];
unsigned int sequence_nr;
diff --git a/multipath/main.c b/multipath/main.c
index 612c6815..3c3d2398 100644
--- a/multipath/main.c
+++ b/multipath/main.c
@@ -825,6 +825,9 @@ int delegate_to_multipathd(enum mpath_cmds cmd,
*p = '\0';
n = sizeof(command);
+ if (conf->skip_delegate)
+ return NOT_DELEGATED;
+
if (cmd == CMD_CREATE && conf->force_reload == FORCE_RELOAD_YES) {
p += snprintf(p, n, "reconfigure");
}
@@ -923,7 +926,7 @@ main (int argc, char *argv[])
multipath_conf = conf;
conf->retrigger_tries = 0;
conf->force_sync = 1;
- while ((arg = getopt(argc, argv, ":aAdcChl::FfM:v:p:b:BrR:itTquUwW")) != EOF ) {
+ while ((arg = getopt(argc, argv, ":aAdDcChl::FfM:v:p:b:BrR:itTquUwW")) != EOF ) {
switch(arg) {
case 1: printf("optarg : %s\n",optarg);
break;
@@ -955,6 +958,9 @@ main (int argc, char *argv[])
if (cmd == CMD_CREATE)
cmd = CMD_DRY_RUN;
break;
+ case 'D':
+ conf->skip_delegate = 1;
+ break;
case 'f':
cmd = CMD_FLUSH_ONE;
break;
--
2.17.2

43
SOURCES/0031-libmultipath-fix-sysfs-dev_loss_tmo-parsing.patch Normal file
View File

@ -0,0 +1,43 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 25 Jun 2020 20:46:08 -0500
Subject: [PATCH] libmultipath: fix sysfs dev_loss_tmo parsing
dev_loss_tmo is a u32 value. However the kernel sysfs code prints it as
a signed integer. This means that if dev_loss_tmo is above INT_MAX, the
sysfs value will be a negative number. Parsing this was causing
sysfs_set_rport_tmo() to fail.
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/discovery.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/libmultipath/discovery.c b/libmultipath/discovery.c
index ffec5162..83a41a4a 100644
--- a/libmultipath/discovery.c
+++ b/libmultipath/discovery.c
@@ -583,7 +583,7 @@ sysfs_set_rport_tmo(struct multipath *mpp, struct path *pp)
struct udev_device *rport_dev = NULL;
char value[16], *eptr;
char rport_id[32];
- unsigned long long tmo = 0;
+ unsigned int tmo;
int ret;
sprintf(rport_id, "rport-%d:%d-%d",
@@ -607,8 +607,8 @@ sysfs_set_rport_tmo(struct multipath *mpp, struct path *pp)
"error %d", rport_id, -ret);
goto out;
}
- tmo = strtoull(value, &eptr, 0);
- if (value == eptr || tmo == ULLONG_MAX) {
+ tmo = strtoul(value, &eptr, 0);
+ if (value == eptr) {
condlog(0, "%s: Cannot parse dev_loss_tmo "
"attribute '%s'", rport_id, value);
goto out;
--
2.17.2

267
SOURCES/0032-kpartx-read-devices-with-direct-IO.patch Normal file
View File

@ -0,0 +1,267 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Fri, 26 Jun 2020 20:06:24 -0500
Subject: [PATCH] kpartx: read devices with direct IO
If kpartx is used on top of shared storage, and a device has its
partition table changed on one machine, and then kpartx is run on
another, it may not see the new data, because the cache still contains
the old data, and there is nothing to tell the machine running kpartx to
invalidate it. To solve this, kpartx should read the devices using
direct io.
One issue with how this code has been updated is that the original code
for getblock() always read 1024 bytes. The new code reads a logical
sector size chunk of the device, and returns a pointer to the 512 byte
sector that the caller asked for, within that (possibly larger) chunk.
This means that if the logical sector size is 512, then the code is now
only reading 512 bytes. Looking through the code for the various
partition types, I can't see a case where more than 512 bytes is needed
and getblock() is used. If anyone has a reason why this code should be
reading 1024 bytes at minmum, I can certainly change this. But when I
looked, I couldn't find a case where reading 512 bytes would cause a
problem.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
kpartx/dasd.c | 7 ++++---
kpartx/gpt.c | 22 +++++++++----------
kpartx/kpartx.c | 56 +++++++++++++++++++++++++++++++++++++++----------
kpartx/kpartx.h | 2 ++
4 files changed, 61 insertions(+), 26 deletions(-)
diff --git a/kpartx/dasd.c b/kpartx/dasd.c
index 14b9d3aa..f0398645 100644
--- a/kpartx/dasd.c
+++ b/kpartx/dasd.c
@@ -22,6 +22,7 @@
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
+#define _GNU_SOURCE
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
@@ -117,13 +118,13 @@ read_dasd_pt(int fd, __attribute__((unused)) struct slice all,
sprintf(pathname, "/dev/.kpartx-node-%u-%u",
(unsigned int)major(dev), (unsigned int)minor(dev));
- if ((fd_dasd = open(pathname, O_RDONLY)) == -1) {
+ if ((fd_dasd = open(pathname, O_RDONLY | O_DIRECT)) == -1) {
/* Devicenode does not exist. Try to create one */
if (mknod(pathname, 0600 | S_IFBLK, dev) == -1) {
/* Couldn't create a device node */
return -1;
}
- fd_dasd = open(pathname, O_RDONLY);
+ fd_dasd = open(pathname, O_RDONLY | O_DIRECT);
/*
* The file will vanish when the last process (we)
* has ceased to access it.
@@ -175,7 +176,7 @@ read_dasd_pt(int fd, __attribute__((unused)) struct slice all,
* Get volume label, extract name and type.
*/
- if (!(data = (unsigned char *)malloc(blocksize)))
+ if (aligned_malloc((void **)&data, blocksize, NULL))
goto out;
diff --git a/kpartx/gpt.c b/kpartx/gpt.c
index 785b34ea..f7fefb70 100644
--- a/kpartx/gpt.c
+++ b/kpartx/gpt.c
@@ -243,8 +243,7 @@ alloc_read_gpt_entries(int fd, gpt_header * gpt)
if (!count) return NULL;
- pte = (gpt_entry *)malloc(count);
- if (!pte)
+ if (aligned_malloc((void **)&pte, get_sector_size(fd), &count))
return NULL;
memset(pte, 0, count);
@@ -269,12 +268,11 @@ static gpt_header *
alloc_read_gpt_header(int fd, uint64_t lba)
{
gpt_header *gpt;
- gpt = (gpt_header *)
- malloc(sizeof (gpt_header));
- if (!gpt)
+ size_t size = sizeof (gpt_header);
+ if (aligned_malloc((void **)&gpt, get_sector_size(fd), &size))
return NULL;
- memset(gpt, 0, sizeof (*gpt));
- if (!read_lba(fd, lba, gpt, sizeof (gpt_header))) {
+ memset(gpt, 0, size);
+ if (!read_lba(fd, lba, gpt, size)) {
free(gpt);
return NULL;
}
@@ -498,6 +496,7 @@ find_valid_gpt(int fd, gpt_header ** gpt, gpt_entry ** ptes)
gpt_header *pgpt = NULL, *agpt = NULL;
gpt_entry *pptes = NULL, *aptes = NULL;
legacy_mbr *legacymbr = NULL;
+ size_t size = sizeof(legacy_mbr);
uint64_t lastlba;
if (!gpt || !ptes)
return 0;
@@ -526,11 +525,10 @@ find_valid_gpt(int fd, gpt_header ** gpt, gpt_entry ** ptes)
}
/* This will be added to the EFI Spec. per Intel after v1.02. */
- legacymbr = malloc(sizeof (*legacymbr));
- if (legacymbr) {
- memset(legacymbr, 0, sizeof (*legacymbr));
- read_lba(fd, 0, (uint8_t *) legacymbr,
- sizeof (*legacymbr));
+ if (aligned_malloc((void **)&legacymbr, get_sector_size(fd),
+ &size) == 0) {
+ memset(legacymbr, 0, size);
+ read_lba(fd, 0, (uint8_t *) legacymbr, size);
good_pmbr = is_pmbr_valid(legacymbr);
free(legacymbr);
legacymbr=NULL;
diff --git a/kpartx/kpartx.c b/kpartx/kpartx.c
index d3620c5c..c24ad6d9 100644
--- a/kpartx/kpartx.c
+++ b/kpartx/kpartx.c
@@ -19,6 +19,7 @@
* cva, 2002-10-26
*/
+#define _GNU_SOURCE
#include <stdio.h>
#include <fcntl.h>
#include <errno.h>
@@ -41,7 +42,6 @@
#define SIZE(a) (sizeof(a)/sizeof((a)[0]))
-#define READ_SIZE 1024
#define MAXTYPES 64
#define MAXSLICES 256
#define DM_TARGET "linear"
@@ -388,7 +388,7 @@ main(int argc, char **argv){
set_delimiter(mapname, delim);
}
- fd = open(device, O_RDONLY);
+ fd = open(device, O_RDONLY | O_DIRECT);
if (fd == -1) {
perror(device);
@@ -690,9 +690,9 @@ xmalloc (size_t size) {
*/
static int
-sseek(int fd, unsigned int secnr) {
+sseek(int fd, unsigned int secnr, int secsz) {
off64_t in, out;
- in = ((off64_t) secnr << 9);
+ in = ((off64_t) secnr * secsz);
out = 1;
if ((out = lseek64(fd, in, SEEK_SET)) != in)
@@ -703,6 +703,31 @@ sseek(int fd, unsigned int secnr) {
return 0;
}
+int
+aligned_malloc(void **mem_p, size_t align, size_t *size_p)
+{
+ static size_t pgsize = 0;
+ size_t size;
+ int err;
+
+ if (!mem_p || !align || (size_p && !*size_p))
+ return EINVAL;
+
+ if (!pgsize)
+ pgsize = getpagesize();
+
+ if (size_p)
+ size = ((*size_p + align - 1) / align) * align;
+ else
+ size = pgsize;
+
+ err = posix_memalign(mem_p, pgsize, size);
+ if (!err && size_p)
+ *size_p = size;
+ return err;
+}
+
+/* always in sector size blocks */
static
struct block {
unsigned int secnr;
@@ -710,30 +735,39 @@ struct block {
struct block *next;
} *blockhead;
+/* blknr is always in 512 byte blocks */
char *
-getblock (int fd, unsigned int secnr) {
+getblock (int fd, unsigned int blknr) {
+ unsigned int secsz = get_sector_size(fd);
+ unsigned int blks_per_sec = secsz / 512;
+ unsigned int secnr = blknr / blks_per_sec;
+ unsigned int blk_off = (blknr % blks_per_sec) * 512;
struct block *bp;
for (bp = blockhead; bp; bp = bp->next)
if (bp->secnr == secnr)
- return bp->block;
+ return bp->block + blk_off;
- if (sseek(fd, secnr))
+ if (sseek(fd, secnr, secsz))
return NULL;
bp = xmalloc(sizeof(struct block));
bp->secnr = secnr;
bp->next = blockhead;
blockhead = bp;
- bp->block = (char *) xmalloc(READ_SIZE);
+ if (aligned_malloc((void **)&bp->block, secsz, NULL)) {
+ fprintf(stderr, "aligned_malloc failed\n");
+ exit(1);
+ }
- if (read(fd, bp->block, READ_SIZE) != READ_SIZE) {
+ if (read(fd, bp->block, secsz) != secsz) {
fprintf(stderr, "read error, sector %d\n", secnr);
- bp->block = NULL;
+ blockhead = bp->next;
+ return NULL;
}
- return bp->block;
+ return bp->block + blk_off;
}
int
diff --git a/kpartx/kpartx.h b/kpartx/kpartx.h
index 67edeb82..727632c1 100644
--- a/kpartx/kpartx.h
+++ b/kpartx/kpartx.h
@@ -1,6 +1,7 @@
#ifndef _KPARTX_H
#define _KPARTX_H
+#include <stddef.h>
#include <stdint.h>
#include <sys/ioctl.h>
@@ -61,6 +62,7 @@ extern ptreader read_mac_pt;
extern ptreader read_sun_pt;
extern ptreader read_ps3_pt;
+int aligned_malloc(void **mem_p, size_t align, size_t *size_p);
char *getblock(int fd, unsigned int secnr);
static inline unsigned int
--
2.17.2

52
SOURCES/0033-kpartx-handle-alternate-bsd-disklabel-location.patch Normal file
View File

@ -0,0 +1,52 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue, 30 Jun 2020 10:49:59 -0500
Subject: [PATCH] kpartx: handle alternate bsd disklabel location
bsd disk labels can either be at the start of the second sector, or 64
bytes into the first sector, but kpartx only handled the first case.
However the second case is what parted creates, and what the linux
kernel partition code expects. kpartx should handle both cases.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
kpartx/bsd.c | 16 ++++++++++++++--
1 file changed, 14 insertions(+), 2 deletions(-)
diff --git a/kpartx/bsd.c b/kpartx/bsd.c
index 0e661fbc..950b0f92 100644
--- a/kpartx/bsd.c
+++ b/kpartx/bsd.c
@@ -1,6 +1,7 @@
#include "kpartx.h"
#include <stdio.h>
+#define BSD_LABEL_OFFSET 64
#define BSD_DISKMAGIC (0x82564557UL) /* The disk magic number */
#define XBSD_MAXPARTITIONS 16
#define BSD_FS_UNUSED 0
@@ -60,8 +61,19 @@ read_bsd_pt(int fd, struct slice all, struct slice *sp, unsigned int ns) {
return -1;
l = (struct bsd_disklabel *) bp;
- if (l->d_magic != BSD_DISKMAGIC)
- return -1;
+ if (l->d_magic != BSD_DISKMAGIC) {
+ /*
+ * BSD disklabels can also start 64 bytes offset from the
+ * start of the first sector
+ */
+ bp = getblock(fd, offset);
+ if (bp == NULL)
+ return -1;
+
+ l = (struct bsd_disklabel *)(bp + 64);
+ if (l->d_magic != BSD_DISKMAGIC)
+ return -1;
+ }
max_partitions = 16;
if (l->d_npartitions < max_partitions)
--
2.17.2

61
SOURCES/0034-libmultipath-fix-checker-detection-for-nvme-devices.patch Normal file
View File

@ -0,0 +1,61 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue, 30 Jun 2020 13:59:13 -0500
Subject: [PATCH] libmultipath: fix checker detection for nvme devices
In order to fix hwhandler autodetection, commit 8794a776 made
detect_alua() differentiate between failures to detect whether alua was
supported, and successfully detecting that it was not supported.
However, this causes nvme devices to get the TUR checker assigned to
them. This is because there is nothing in detect_alua() to make it only
work on scsi devices, and select_checker wasn't updated to handle
detect_alua() failing without setting pp->tpgs to TPGS_NONE.
detect_alua() should automatically set pp->tpgs to TPGS_NONE and exit on
non-scsi devices. Also, select_checker() should not assume that a
devices is ALUA, simply because if failed to detect if alua was
supported.
Fixes: 8794a776 "libmultipath: fix ALUA autodetection when paths are
down"
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/discovery.c | 6 ++++++
libmultipath/propsel.c | 4 +++-
2 files changed, 9 insertions(+), 1 deletion(-)
diff --git a/libmultipath/discovery.c b/libmultipath/discovery.c
index 83a41a4a..aa5942c3 100644
--- a/libmultipath/discovery.c
+++ b/libmultipath/discovery.c
@@ -887,6 +887,12 @@ detect_alua(struct path * pp)
int tpgs;
unsigned int timeout;
+
+ if (pp->bus != SYSFS_BUS_SCSI) {
+ pp->tpgs = TPGS_NONE;
+ return;
+ }
+
if (sysfs_get_timeout(pp, &timeout) <= 0)
timeout = DEF_TIMEOUT;
diff --git a/libmultipath/propsel.c b/libmultipath/propsel.c
index 897e48ca..d362beb4 100644
--- a/libmultipath/propsel.c
+++ b/libmultipath/propsel.c
@@ -521,7 +521,9 @@ int select_checker(struct config *conf, struct path *pp)
if (check_rdac(pp)) {
ckr_name = RDAC;
goto out;
- } else if (path_get_tpgs(pp) != TPGS_NONE) {
+ }
+ path_get_tpgs(pp);
+ if (pp->tpgs != TPGS_NONE && pp->tpgs != TPGS_UNDEF) {
ckr_name = TUR;
goto out;
}
--
2.17.2

30
SOURCES/0035-Makefile.inc-trim-extra-information-from-systemd-ver.patch Normal file
View File

@ -0,0 +1,30 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 6 Jul 2020 13:21:12 -0500
Subject: [PATCH] Makefile.inc: trim extra information from systemd version
Some systemd versions print extra information in the
"pkg-config --modversion" output, which confuses make. Trim this
off.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
Makefile.inc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/Makefile.inc b/Makefile.inc
index c2abd301..220009e0 100644
--- a/Makefile.inc
+++ b/Makefile.inc
@@ -37,7 +37,7 @@ endif
ifndef SYSTEMD
ifeq ($(shell pkg-config --modversion libsystemd >/dev/null 2>&1 && echo 1), 1)
- SYSTEMD = $(shell pkg-config --modversion libsystemd)
+ SYSTEMD = $(shell pkg-config --modversion libsystemd | awk '{print $$1}')
else
ifeq ($(shell systemctl --version >/dev/null 2>&1 && echo 1), 1)
SYSTEMD = $(shell systemctl --version 2> /dev/null | \
--
2.17.2

26
SOURCES/0036-kpartx-fix-Wsign-compare-error.patch Normal file
View File

@ -0,0 +1,26 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 6 Jul 2020 17:28:46 -0500
Subject: [PATCH] kpartx: fix -Wsign-compare error
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
kpartx/kpartx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kpartx/kpartx.c b/kpartx/kpartx.c
index c24ad6d9..653ce0c8 100644
--- a/kpartx/kpartx.c
+++ b/kpartx/kpartx.c
@@ -738,7 +738,7 @@ struct block {
/* blknr is always in 512 byte blocks */
char *
getblock (int fd, unsigned int blknr) {
- unsigned int secsz = get_sector_size(fd);
+ int secsz = get_sector_size(fd);
unsigned int blks_per_sec = secsz / 512;
unsigned int secnr = blknr / blks_per_sec;
unsigned int blk_off = (blknr % blks_per_sec) * 512;
--
2.17.2

86
SOURCES/0037-libmultipath-count-pending-paths-as-active-on-loads.patch Normal file
View File

@ -0,0 +1,86 @@
From 7159242be31dbb3f25aa67920462107bc2bc5fe0 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 9 Jul 2020 18:20:18 -0500
Subject: [PATCH] libmultipath: count pending paths as active on loads
When multipath loads a table, it signals to udev if there are no active
paths. Multipath wasn't counting pending paths as active. This meant
that if all the paths were pending, udev would treat the device as not
ready, and not run kpartx on it. Even if the pending paths later
because active and were reinstated, the kernel would not send a new
uevent, because from its point of view, they were always up.
The alternative would be to continue to treat them as failed in the udev
rules, but then also tell the kernel that they were down, so that it
would trigger a uevent when they were reinstated. However, this could
lead to newly created multipath devices failing IO, simply because the
path checkers hadn't returned yet. Having udev assume that the the
device is up, like the kernel does, seems like the safer option.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/devmapper.c | 3 ++-
libmultipath/structs.c | 20 ++++++++++++++++++++
libmultipath/structs.h | 1 +
3 files changed, 23 insertions(+), 1 deletion(-)
diff --git a/libmultipath/devmapper.c b/libmultipath/devmapper.c
index 7f98bf9d..91ff0b3d 100644
--- a/libmultipath/devmapper.c
+++ b/libmultipath/devmapper.c
@@ -408,7 +408,8 @@ static uint16_t build_udev_flags(const struct multipath *mpp, int reload)
/* DM_UDEV_DISABLE_LIBRARY_FALLBACK is added in dm_addmap */
return (mpp->skip_kpartx == SKIP_KPARTX_ON ?
MPATH_UDEV_NO_KPARTX_FLAG : 0) |
- ((count_active_paths(mpp) == 0 || mpp->ghost_delay_tick > 0) ?
+ ((count_active_pending_paths(mpp) == 0 ||
+ mpp->ghost_delay_tick > 0) ?
MPATH_UDEV_NO_PATHS_FLAG : 0) |
(reload && !mpp->force_udev_reload ?
MPATH_UDEV_RELOAD_FLAG : 0);
diff --git a/libmultipath/structs.c b/libmultipath/structs.c
index 2dd378c4..dda9884c 100644
--- a/libmultipath/structs.c
+++ b/libmultipath/structs.c
@@ -500,6 +500,26 @@ int count_active_paths(const struct multipath *mpp)
return count;
}
+int count_active_pending_paths(const struct multipath *mpp)
+{
+ struct pathgroup *pgp;
+ struct path *pp;
+ int count = 0;
+ int i, j;
+
+ if (!mpp->pg)
+ return 0;
+
+ vector_foreach_slot (mpp->pg, pgp, i) {
+ vector_foreach_slot (pgp->paths, pp, j) {
+ if (pp->state == PATH_UP || pp->state == PATH_GHOST ||
+ pp->state == PATH_PENDING)
+ count++;
+ }
+ }
+ return count;
+}
+
int pathcmp(const struct pathgroup *pgp, const struct pathgroup *cpgp)
{
int i, j;
diff --git a/libmultipath/structs.h b/libmultipath/structs.h
index 9bd39eb1..8e78b8c0 100644
--- a/libmultipath/structs.h
+++ b/libmultipath/structs.h
@@ -465,6 +465,7 @@ struct path * first_path (const struct multipath *mpp);
int pathcountgr (const struct pathgroup *, int);
int pathcount (const struct multipath *, int);
int count_active_paths(const struct multipath *);
+int count_active_pending_paths(const struct multipath *);
int pathcmp (const struct pathgroup *, const struct pathgroup *);
int add_feature (char **, const char *);
int remove_feature (char **, const char *);
--
2.17.2

58
SOURCES/0038-multipath-deal-with-failures-flushing-maps.patch Normal file
View File

@ -0,0 +1,58 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 13 Jul 2020 15:41:15 -0500
Subject: [PATCH] multipath: deal with failures flushing maps
dm_flush_maps() was failing if there were no device-mapper devices at
all, instead of returning success, since there is nothing to do.
delegate_to_multipathd() was returning success, even if the multipathd
command failed. Also, if the command was set to fail with NOT_DELEGATED,
it shouldn't print any errors, since multipath will try to issue to
command itself.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/devmapper.c | 2 +-
multipath/main.c | 9 ++++++---
2 files changed, 7 insertions(+), 4 deletions(-)
diff --git a/libmultipath/devmapper.c b/libmultipath/devmapper.c
index 91ff0b3d..3f70e576 100644
--- a/libmultipath/devmapper.c
+++ b/libmultipath/devmapper.c
@@ -970,10 +970,10 @@ int dm_flush_maps (int need_suspend, int retries)
if (!(names = dm_task_get_names (dmt)))
goto out;
+ r = 0;
if (!names->dev)
goto out;
- r = 0;
do {
if (need_suspend)
r |= dm_suspend_and_flush_map(names->name, retries);
diff --git a/multipath/main.c b/multipath/main.c
index 3c3d2398..607cada2 100644
--- a/multipath/main.c
+++ b/multipath/main.c
@@ -869,9 +869,12 @@ int delegate_to_multipathd(enum mpath_cmds cmd,
goto out;
}
- if (reply != NULL && *reply != '\0' && strcmp(reply, "ok\n"))
- printf("%s", reply);
- r = DELEGATE_OK;
+ if (reply != NULL && *reply != '\0') {
+ if (strcmp(reply, "fail\n"))
+ r = DELEGATE_OK;
+ if (r != NOT_DELEGATED && strcmp(reply, "ok\n"))
+ printf("%s", reply);
+ }
out:
FREE(reply);
--
2.17.2

54
SOURCES/0039-libmultipath-factor-out-code-to-get-vpd-page-data.patch Normal file
View File

@ -0,0 +1,54 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue, 3 Nov 2020 14:27:58 -0600
Subject: [PATCH] libmultipath: factor out code to get vpd page data
A future patch will reuse the code to get the vpd page data, so factor
it out from get_vpd_sgio().
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/discovery.c | 19 +++++++++++++++----
1 file changed, 15 insertions(+), 4 deletions(-)
diff --git a/libmultipath/discovery.c b/libmultipath/discovery.c
index aa5942c3..eb1e735d 100644
--- a/libmultipath/discovery.c
+++ b/libmultipath/discovery.c
@@ -1239,11 +1239,10 @@ get_vpd_sysfs (struct udev_device *parent, int pg, char * str, int maxlen)
return len;
}
-int
-get_vpd_sgio (int fd, int pg, int vend_id, char * str, int maxlen)
+static int
+fetch_vpd_page(int fd, int pg, unsigned char *buff)
{
- int len, buff_len;
- unsigned char buff[4096];
+ int buff_len;
memset(buff, 0x0, 4096);
if (sgio_get_vpd(buff, 4096, fd, pg) < 0) {
@@ -1264,6 +1263,18 @@ get_vpd_sgio (int fd, int pg, int vend_id, char * str, int maxlen)
condlog(3, "vpd pg%02x page truncated", pg);
buff_len = 4096;
}
+ return buff_len;
+}
+
+int
+get_vpd_sgio (int fd, int pg, int vend_id, char * str, int maxlen)
+{
+ int len, buff_len;
+ unsigned char buff[4096];
+
+ buff_len = fetch_vpd_page(fd, pg, buff);
+ if (buff_len < 0)
+ return buff_len;
if (pg == 0x80)
len = parse_vpd_pg80(buff, str, maxlen);
else if (pg == 0x83)
--
2.17.2

98
SOURCES/0040-libmultipath-limit-reading-0xc9-vpd-page.patch Normal file
View File

@ -0,0 +1,98 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Wed, 7 Oct 2020 21:43:02 -0500
Subject: [PATCH] libmultipath: limit reading 0xc9 vpd page
Only rdac arrays support 0xC9 vpd page inquiries. All other arrays will
return a failure. Only do the rdac inquiry when detecting array
capabilities if the array's path checker is explicitly set to rdac, or
the path checker is not set, and the array reports that it supports vpd
page 0xC9 in the Supported VPD Pages (0x00) vpd page.
Multipath was doing the check if either the path checker was set to
rdac, or no path checker was set. This means that for almost all
non-rdac arrays, multipath was issuing a bad inquiry. This was annoying
users.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/discovery.c | 25 +++++++++++++++++++++++++
libmultipath/discovery.h | 1 +
libmultipath/propsel.c | 10 ++++++----
3 files changed, 32 insertions(+), 4 deletions(-)
diff --git a/libmultipath/discovery.c b/libmultipath/discovery.c
index eb1e735d..01aadba9 100644
--- a/libmultipath/discovery.c
+++ b/libmultipath/discovery.c
@@ -1266,6 +1266,31 @@ fetch_vpd_page(int fd, int pg, unsigned char *buff)
return buff_len;
}
+/* heavily based on sg_inq.c from sg3_utils */
+bool
+is_vpd_page_supported(int fd, int pg)
+{
+ int i, len, buff_len;
+ unsigned char buff[4096];
+
+ buff_len = fetch_vpd_page(fd, 0x00, buff);
+ if (buff_len < 0)
+ return false;
+ if (buff_len < 4) {
+ condlog(3, "VPD page 00h too short");
+ return false;
+ }
+
+ len = buff[3] + 4;
+ if (len > buff_len)
+ condlog(3, "vpd page 00h trucated, expected %d, have %d",
+ len, buff_len);
+ for (i = 4; i < len; ++i)
+ if (buff[i] == pg)
+ return true;
+ return false;
+}
+
int
get_vpd_sgio (int fd, int pg, int vend_id, char * str, int maxlen)
{
diff --git a/libmultipath/discovery.h b/libmultipath/discovery.h
index 6444887d..d3193daf 100644
--- a/libmultipath/discovery.h
+++ b/libmultipath/discovery.h
@@ -56,6 +56,7 @@ int sysfs_get_asymmetric_access_state(struct path *pp,
char *buff, int buflen);
int get_uid(struct path * pp, int path_state, struct udev_device *udev,
int allow_fallback);
+bool is_vpd_page_supported(int fd, int pg);
/*
* discovery bitmask
diff --git a/libmultipath/propsel.c b/libmultipath/propsel.c
index d362beb4..d7febec6 100644
--- a/libmultipath/propsel.c
+++ b/libmultipath/propsel.c
@@ -496,13 +496,15 @@ check_rdac(struct path * pp)
{
int len;
char buff[44];
- const char *checker_name;
+ const char *checker_name = NULL;
if (pp->bus != SYSFS_BUS_SCSI)
return 0;
- /* Avoid ioctl if this is likely not an RDAC array */
- if (__do_set_from_hwe(checker_name, pp, checker_name) &&
- strcmp(checker_name, RDAC))
+ /* Avoid checking 0xc9 if this is likely not an RDAC array */
+ if (!__do_set_from_hwe(checker_name, pp, checker_name) &&
+ !is_vpd_page_supported(pp->fd, 0xC9))
+ return 0;
+ if (checker_name && strcmp(checker_name, RDAC))
return 0;
len = get_vpd_sgio(pp->fd, 0xC9, 0, buff, 44);
if (len <= 0)
--
2.17.2

44
SOURCES/0041-libmultipath-add-device-to-hwtable.c.patch Normal file
View File

@ -0,0 +1,44 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Steve Schremmer <steve.schremmer@netapp.com>
Date: Mon, 6 Jul 2020 20:22:35 +0000
Subject: [PATCH] libmultipath: add device to hwtable.c
Add FUJITSU ETERNUS_AHB defaults.
Signed-off-by: Steve Schremmer <steve.schremmer@netapp.com>
Reviewed-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/hwtable.c | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/libmultipath/hwtable.c b/libmultipath/hwtable.c
index d1fcfdb3..d680bdfc 100644
--- a/libmultipath/hwtable.c
+++ b/libmultipath/hwtable.c
@@ -428,6 +428,22 @@ static struct hwentry default_hw[] = {
.pgpolicy = MULTIBUS,
.no_path_retry = 10,
},
+ {
+ /*
+ * ETERNUS AB/HB
+ * Maintainer: NetApp RDAC team <ng-eseries-upstream-maintainers@netapp.com>
+ */
+ .vendor = "FUJITSU",
+ .product = "ETERNUS_AHB",
+ .bl_product = "Universal Xport",
+ .pgpolicy = GROUP_BY_PRIO,
+ .checker_name = RDAC,
+ .features = "2 pg_init_retries 50",
+ .hwhandler = "1 rdac",
+ .prio_name = PRIO_RDAC,
+ .pgfailback = -FAILBACK_IMMEDIATE,
+ .no_path_retry = 30,
+ },
/*
* Hitachi Vantara
*
--
2.17.2

159
SOURCES/0042-libmultipath-move-fast_io_fail-defines-to-structs.h.patch Normal file
View File

@ -0,0 +1,159 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 12 Oct 2020 16:06:11 -0500
Subject: [PATCH] libmultipath: move fast_io_fail defines to structs.h
Since fast_io_fail is part of the multipath struct, its symbolic values
belong in structs.h. Also, make it an instance of a general enum, which
will be used again in future patches, and change the set/print functions
which use it to use the general enum instead.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/config.h | 8 --------
libmultipath/dict.c | 30 +++++++++++++++---------------
libmultipath/dict.h | 2 +-
libmultipath/propsel.c | 2 +-
libmultipath/structs.h | 17 +++++++++++++++++
5 files changed, 34 insertions(+), 25 deletions(-)
diff --git a/libmultipath/config.h b/libmultipath/config.h
index 160867cd..f38c7639 100644
--- a/libmultipath/config.h
+++ b/libmultipath/config.h
@@ -11,14 +11,6 @@
#define ORIGIN_CONFIG 1
#define ORIGIN_NO_CONFIG 2
-/*
- * In kernel, fast_io_fail == 0 means immediate failure on rport delete.
- * OTOH '0' means not-configured in various places in multipath-tools.
- */
-#define MP_FAST_IO_FAIL_UNSET (0)
-#define MP_FAST_IO_FAIL_OFF (-1)
-#define MP_FAST_IO_FAIL_ZERO (-2)
-
enum devtypes {
DEV_NONE,
DEV_DEVT,
diff --git a/libmultipath/dict.c b/libmultipath/dict.c
index 184d4b22..ce8e1cda 100644
--- a/libmultipath/dict.c
+++ b/libmultipath/dict.c
@@ -834,7 +834,7 @@ declare_mp_attr_handler(gid, set_gid)
declare_mp_attr_snprint(gid, print_gid)
static int
-set_fast_io_fail(vector strvec, void *ptr)
+set_undef_off_zero(vector strvec, void *ptr)
{
char * buff;
int *int_ptr = (int *)ptr;
@@ -844,36 +844,36 @@ set_fast_io_fail(vector strvec, void *ptr)
return 1;
if (strcmp(buff, "off") == 0)
- *int_ptr = MP_FAST_IO_FAIL_OFF;
+ *int_ptr = UOZ_OFF;
else if (sscanf(buff, "%d", int_ptr) != 1 ||
- *int_ptr < MP_FAST_IO_FAIL_ZERO)
- *int_ptr = MP_FAST_IO_FAIL_UNSET;
+ *int_ptr < UOZ_ZERO)
+ *int_ptr = UOZ_UNDEF;
else if (*int_ptr == 0)
- *int_ptr = MP_FAST_IO_FAIL_ZERO;
+ *int_ptr = UOZ_ZERO;
FREE(buff);
return 0;
}
int
-print_fast_io_fail(char * buff, int len, long v)
+print_undef_off_zero(char * buff, int len, long v)
{
- if (v == MP_FAST_IO_FAIL_UNSET)
+ if (v == UOZ_UNDEF)
return 0;
- if (v == MP_FAST_IO_FAIL_OFF)
+ if (v == UOZ_OFF)
return snprintf(buff, len, "\"off\"");
- if (v == MP_FAST_IO_FAIL_ZERO)
+ if (v == UOZ_ZERO)
return snprintf(buff, len, "0");
return snprintf(buff, len, "%ld", v);
}
-declare_def_handler(fast_io_fail, set_fast_io_fail)
-declare_def_snprint_defint(fast_io_fail, print_fast_io_fail,
+declare_def_handler(fast_io_fail, set_undef_off_zero)
+declare_def_snprint_defint(fast_io_fail, print_undef_off_zero,
DEFAULT_FAST_IO_FAIL)
-declare_ovr_handler(fast_io_fail, set_fast_io_fail)
-declare_ovr_snprint(fast_io_fail, print_fast_io_fail)
-declare_hw_handler(fast_io_fail, set_fast_io_fail)
-declare_hw_snprint(fast_io_fail, print_fast_io_fail)
+declare_ovr_handler(fast_io_fail, set_undef_off_zero)
+declare_ovr_snprint(fast_io_fail, print_undef_off_zero)
+declare_hw_handler(fast_io_fail, set_undef_off_zero)
+declare_hw_snprint(fast_io_fail, print_undef_off_zero)
static int
set_dev_loss(vector strvec, void *ptr)
diff --git a/libmultipath/dict.h b/libmultipath/dict.h
index a40ac66f..a917e1ca 100644
--- a/libmultipath/dict.h
+++ b/libmultipath/dict.h
@@ -13,7 +13,7 @@ int print_rr_weight(char *buff, int len, long v);
int print_pgfailback(char *buff, int len, long v);
int print_pgpolicy(char *buff, int len, long v);
int print_no_path_retry(char *buff, int len, long v);
-int print_fast_io_fail(char *buff, int len, long v);
+int print_undef_off_zero(char *buff, int len, long v);
int print_dev_loss(char *buff, int len, unsigned long v);
int print_reservation_key(char * buff, int len, struct be64 key, uint8_t
flags, int source);
diff --git a/libmultipath/propsel.c b/libmultipath/propsel.c
index d7febec6..725db2b1 100644
--- a/libmultipath/propsel.c
+++ b/libmultipath/propsel.c
@@ -755,7 +755,7 @@ int select_fast_io_fail(struct config *conf, struct multipath *mp)
mp_set_conf(fast_io_fail);
mp_set_default(fast_io_fail, DEFAULT_FAST_IO_FAIL);
out:
- print_fast_io_fail(buff, 12, mp->fast_io_fail);
+ print_undef_off_zero(buff, 12, mp->fast_io_fail);
condlog(3, "%s: fast_io_fail_tmo = %s %s", mp->alias, buff, origin);
return 0;
}
diff --git a/libmultipath/structs.h b/libmultipath/structs.h
index 8e78b8c0..29209984 100644
--- a/libmultipath/structs.h
+++ b/libmultipath/structs.h
@@ -229,6 +229,23 @@ enum vpd_vendor_ids {
VPD_VP_ARRAY_SIZE, /* This must remain the last entry */
};
+/*
+ * Multipath treats 0 as undefined for optional config parameters.
+ * Use this for cases where 0 is a valid option for systems multipath
+ * is communicating with
+ */
+enum undefined_off_zero {
+ UOZ_UNDEF = 0,
+ UOZ_OFF = -1,
+ UOZ_ZERO = -2,
+};
+
+enum fast_io_fail_states {
+ MP_FAST_IO_FAIL_UNSET = UOZ_UNDEF,
+ MP_FAST_IO_FAIL_OFF = UOZ_OFF,
+ MP_FAST_IO_FAIL_ZERO = UOZ_ZERO,
+};
+
struct vpd_vendor_page {
int pg;
const char *name;
--
2.17.2

295
SOURCES/0043-libmultipath-add-eh_deadline-multipath.conf-paramete.patch Normal file
View File

@ -0,0 +1,295 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Wed, 14 Oct 2020 18:38:20 -0500
Subject: [PATCH] libmultipath: add eh_deadline multipath.conf parameter
There are times a fc rport is never lost, meaning that fast_io_fail_tmo
and dev_loss_tmo never trigger, but scsi commands still hang. This can
cause problems in cases where users have string timing requirements, and
the easiest way to solve these issues is to set eh_deadline. Since it's
already possible to set fast_io_fail_tmo and dev_loss_tmo from
multipath.conf, and have multipath take care of setting it correctly for
the scsi devices in sysfs, it makes sense to allow users to set
eh_deadline here as well.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/config.c | 2 ++
libmultipath/config.h | 2 ++
libmultipath/configure.c | 1 +
libmultipath/dict.c | 10 +++++++
libmultipath/discovery.c | 58 +++++++++++++++++++++++++++++++++-----
libmultipath/propsel.c | 17 +++++++++++
libmultipath/propsel.h | 1 +
libmultipath/structs.h | 7 +++++
multipath/multipath.conf.5 | 16 +++++++++++
9 files changed, 107 insertions(+), 7 deletions(-)
diff --git a/libmultipath/config.c b/libmultipath/config.c
index 26f8e050..a71db2d0 100644
--- a/libmultipath/config.c
+++ b/libmultipath/config.c
@@ -359,6 +359,7 @@ merge_hwe (struct hwentry * dst, struct hwentry * src)
merge_num(flush_on_last_del);
merge_num(fast_io_fail);
merge_num(dev_loss);
+ merge_num(eh_deadline);
merge_num(user_friendly_names);
merge_num(retain_hwhandler);
merge_num(detect_prio);
@@ -514,6 +515,7 @@ store_hwe (vector hwtable, struct hwentry * dhwe)
hwe->flush_on_last_del = dhwe->flush_on_last_del;
hwe->fast_io_fail = dhwe->fast_io_fail;
hwe->dev_loss = dhwe->dev_loss;
+ hwe->eh_deadline = dhwe->eh_deadline;
hwe->user_friendly_names = dhwe->user_friendly_names;
hwe->retain_hwhandler = dhwe->retain_hwhandler;
hwe->detect_prio = dhwe->detect_prio;
diff --git a/libmultipath/config.h b/libmultipath/config.h
index f38c7639..a22c1b4e 100644
--- a/libmultipath/config.h
+++ b/libmultipath/config.h
@@ -64,6 +64,7 @@ struct hwentry {
int flush_on_last_del;
int fast_io_fail;
unsigned int dev_loss;
+ int eh_deadline;
int user_friendly_names;
int retain_hwhandler;
int detect_prio;
@@ -149,6 +150,7 @@ struct config {
int attribute_flags;
int fast_io_fail;
unsigned int dev_loss;
+ int eh_deadline;
int log_checker_err;
int allow_queueing;
int find_multipaths;
diff --git a/libmultipath/configure.c b/libmultipath/configure.c
index 96c79610..b7113291 100644
--- a/libmultipath/configure.c
+++ b/libmultipath/configure.c
@@ -340,6 +340,7 @@ int setup_map(struct multipath *mpp, char *params, int params_size,
select_gid(conf, mpp);
select_fast_io_fail(conf, mpp);
select_dev_loss(conf, mpp);
+ select_eh_deadline(conf, mpp);
select_reservation_key(conf, mpp);
select_deferred_remove(conf, mpp);
select_marginal_path_err_sample_time(conf, mpp);
diff --git a/libmultipath/dict.c b/libmultipath/dict.c
index ce8e1cda..8fd91d8c 100644
--- a/libmultipath/dict.c
+++ b/libmultipath/dict.c
@@ -911,6 +911,13 @@ declare_ovr_snprint(dev_loss, print_dev_loss)
declare_hw_handler(dev_loss, set_dev_loss)
declare_hw_snprint(dev_loss, print_dev_loss)
+declare_def_handler(eh_deadline, set_undef_off_zero)
+declare_def_snprint(eh_deadline, print_undef_off_zero)
+declare_ovr_handler(eh_deadline, set_undef_off_zero)
+declare_ovr_snprint(eh_deadline, print_undef_off_zero)
+declare_hw_handler(eh_deadline, set_undef_off_zero)
+declare_hw_snprint(eh_deadline, print_undef_off_zero)
+
static int
set_pgpolicy(vector strvec, void *ptr)
{
@@ -1776,6 +1783,7 @@ init_keywords(vector keywords)
install_keyword("gid", &def_gid_handler, &snprint_def_gid);
install_keyword("fast_io_fail_tmo", &def_fast_io_fail_handler, &snprint_def_fast_io_fail);
install_keyword("dev_loss_tmo", &def_dev_loss_handler, &snprint_def_dev_loss);
+ install_keyword("eh_deadline", &def_eh_deadline_handler, &snprint_def_eh_deadline);
install_keyword("bindings_file", &def_bindings_file_handler, &snprint_def_bindings_file);
install_keyword("wwids_file", &def_wwids_file_handler, &snprint_def_wwids_file);
install_keyword("prkeys_file", &def_prkeys_file_handler, &snprint_def_prkeys_file);
@@ -1885,6 +1893,7 @@ init_keywords(vector keywords)
install_keyword("flush_on_last_del", &hw_flush_on_last_del_handler, &snprint_hw_flush_on_last_del);
install_keyword("fast_io_fail_tmo", &hw_fast_io_fail_handler, &snprint_hw_fast_io_fail);
install_keyword("dev_loss_tmo", &hw_dev_loss_handler, &snprint_hw_dev_loss);
+ install_keyword("eh_deadline", &hw_eh_deadline_handler, &snprint_hw_eh_deadline);
install_keyword("user_friendly_names", &hw_user_friendly_names_handler, &snprint_hw_user_friendly_names);
install_keyword("retain_attached_hw_handler", &hw_retain_hwhandler_handler, &snprint_hw_retain_hwhandler);
install_keyword("detect_prio", &hw_detect_prio_handler, &snprint_hw_detect_prio);
@@ -1925,6 +1934,7 @@ init_keywords(vector keywords)
install_keyword("flush_on_last_del", &ovr_flush_on_last_del_handler, &snprint_ovr_flush_on_last_del);
install_keyword("fast_io_fail_tmo", &ovr_fast_io_fail_handler, &snprint_ovr_fast_io_fail);
install_keyword("dev_loss_tmo", &ovr_dev_loss_handler, &snprint_ovr_dev_loss);
+ install_keyword("eh_deadline", &ovr_eh_deadline_handler, &snprint_ovr_eh_deadline);
install_keyword("user_friendly_names", &ovr_user_friendly_names_handler, &snprint_ovr_user_friendly_names);
install_keyword("retain_attached_hw_handler", &ovr_retain_hwhandler_handler, &snprint_ovr_retain_hwhandler);
install_keyword("detect_prio", &ovr_detect_prio_handler, &snprint_ovr_detect_prio);
diff --git a/libmultipath/discovery.c b/libmultipath/discovery.c
index 01aadba9..a328aafa 100644
--- a/libmultipath/discovery.c
+++ b/libmultipath/discovery.c
@@ -577,6 +577,42 @@ sysfs_get_asymmetric_access_state(struct path *pp, char *buff, int buflen)
return !!preferred;
}
+static int
+sysfs_set_eh_deadline(struct multipath *mpp, struct path *pp)
+{
+ struct udev_device *hostdev;
+ char host_name[HOST_NAME_LEN], value[16];
+ int ret;
+
+ if (mpp->eh_deadline == EH_DEADLINE_UNSET)
+ return 0;
+
+ sprintf(host_name, "host%d", pp->sg_id.host_no);
+ hostdev = udev_device_new_from_subsystem_sysname(udev,
+ "scsi_host", host_name);
+ if (!hostdev)
+ return 1;
+
+ if (mpp->eh_deadline == EH_DEADLINE_OFF)
+ sprintf(value, "off");
+ else if (mpp->eh_deadline == EH_DEADLINE_ZERO)
+ sprintf(value, "0");
+ else
+ snprintf(value, 16, "%u", mpp->eh_deadline);
+
+ ret = sysfs_attr_set_value(hostdev, "eh_deadline",
+ value, strlen(value));
+ /*
+ * not all scsi drivers support setting eh_deadline, so failing
+ * is totally reasonable
+ */
+ if (ret <= 0)
+ condlog(4, "%s: failed to set eh_deadline to %s, error %d", udev_device_get_sysname(hostdev), value, -ret);
+
+ udev_device_unref(hostdev);
+ return (ret <= 0);
+}
+
static void
sysfs_set_rport_tmo(struct multipath *mpp, struct path *pp)
{
@@ -787,16 +823,24 @@ sysfs_set_scsi_tmo (struct multipath *mpp, unsigned int checkint)
mpp->alias, mpp->fast_io_fail);
mpp->fast_io_fail = MP_FAST_IO_FAIL_OFF;
}
- if (!mpp->dev_loss && mpp->fast_io_fail == MP_FAST_IO_FAIL_UNSET)
+ if (!mpp->dev_loss && mpp->fast_io_fail == MP_FAST_IO_FAIL_UNSET &&
+ mpp->eh_deadline == EH_DEADLINE_UNSET)
return 0;
vector_foreach_slot(mpp->paths, pp, i) {
- if (pp->sg_id.proto_id == SCSI_PROTOCOL_FCP)
- sysfs_set_rport_tmo(mpp, pp);
- if (pp->sg_id.proto_id == SCSI_PROTOCOL_ISCSI)
- sysfs_set_session_tmo(mpp, pp);
- if (pp->sg_id.proto_id == SCSI_PROTOCOL_SAS)
- sysfs_set_nexus_loss_tmo(mpp, pp);
+ if (pp->bus != SYSFS_BUS_SCSI)
+ continue;
+
+ if (mpp->dev_loss ||
+ mpp->fast_io_fail != MP_FAST_IO_FAIL_UNSET) {
+ if (pp->sg_id.proto_id == SCSI_PROTOCOL_FCP)
+ sysfs_set_rport_tmo(mpp, pp);
+ else if (pp->sg_id.proto_id == SCSI_PROTOCOL_ISCSI)
+ sysfs_set_session_tmo(mpp, pp);
+ else if (pp->sg_id.proto_id == SCSI_PROTOCOL_SAS)
+ sysfs_set_nexus_loss_tmo(mpp, pp);
+ }
+ sysfs_set_eh_deadline(mpp, pp);
}
return 0;
}
diff --git a/libmultipath/propsel.c b/libmultipath/propsel.c
index 725db2b1..1150cfe8 100644
--- a/libmultipath/propsel.c
+++ b/libmultipath/propsel.c
@@ -776,6 +776,23 @@ out:
return 0;
}
+int select_eh_deadline(struct config *conf, struct multipath *mp)
+{
+ const char *origin;
+ char buff[12];
+
+ mp_set_ovr(eh_deadline);
+ mp_set_hwe(eh_deadline);
+ mp_set_conf(eh_deadline);
+ mp->eh_deadline = EH_DEADLINE_UNSET;
+ /* not changing sysfs in default cause, so don't print anything */
+ return 0;
+out:
+ print_undef_off_zero(buff, 12, mp->eh_deadline);
+ condlog(3, "%s: eh_deadline = %s %s", mp->alias, buff, origin);
+ return 0;
+}
+
int select_flush_on_last_del(struct config *conf, struct multipath *mp)
{
const char *origin;
diff --git a/libmultipath/propsel.h b/libmultipath/propsel.h
index 3d6edd8a..a68bacf0 100644
--- a/libmultipath/propsel.h
+++ b/libmultipath/propsel.h
@@ -17,6 +17,7 @@ int select_uid(struct config *conf, struct multipath *mp);
int select_gid(struct config *conf, struct multipath *mp);
int select_fast_io_fail(struct config *conf, struct multipath *mp);
int select_dev_loss(struct config *conf, struct multipath *mp);
+int select_eh_deadline(struct config *conf, struct multipath *mp);
int select_reservation_key(struct config *conf, struct multipath *mp);
int select_retain_hwhandler (struct config *conf, struct multipath * mp);
int select_detect_prio(struct config *conf, struct path * pp);
diff --git a/libmultipath/structs.h b/libmultipath/structs.h
index 29209984..65542dea 100644
--- a/libmultipath/structs.h
+++ b/libmultipath/structs.h
@@ -246,6 +246,12 @@ enum fast_io_fail_states {
MP_FAST_IO_FAIL_ZERO = UOZ_ZERO,
};
+enum eh_deadline_states {
+ EH_DEADLINE_UNSET = UOZ_UNDEF,
+ EH_DEADLINE_OFF = UOZ_OFF,
+ EH_DEADLINE_ZERO = UOZ_ZERO,
+};
+
struct vpd_vendor_page {
int pg;
const char *name;
@@ -366,6 +372,7 @@ struct multipath {
int ghost_delay;
int ghost_delay_tick;
unsigned int dev_loss;
+ int eh_deadline;
uid_t uid;
gid_t gid;
mode_t mode;
diff --git a/multipath/multipath.conf.5 b/multipath/multipath.conf.5
index 6dc26f10..60954574 100644
--- a/multipath/multipath.conf.5
+++ b/multipath/multipath.conf.5
@@ -700,6 +700,22 @@ The default is: \fB600\fR
.
.
.TP
+.B eh_deadline
+Specify the maximum number of seconds the SCSI layer will spend doing error
+handling when scsi devices fail. After this timeout the scsi layer will perform
+a full HBA reset. Setting this may be necessary in cases where the rport is
+never lost, so \fIfast_io_fail_tmo\fR and \fIdev_loss_tmo\fR will never
+trigger, but (frequently do to load) scsi commands still hang. \fBNote:\fR when
+the scsi error handler performs the HBA reset, all target paths on that HBA
+will be affected. eh_deadline should only be set in cases where all targets on
+the affected HBAs are multipathed.
+.RS
+.TP
+The default is: \fB<unset>\fR
+.RE
+.
+.
+.TP
.B bindings_file
The full pathname of the binding file to be used when the user_friendly_names
option is set.
--
2.17.2

89
SOURCES/0044-libmultipath-don-t-dlclose-tur-checker-DSO.patch Normal file
View File

@ -0,0 +1,89 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Fri, 23 Oct 2020 11:38:24 -0500
Subject: [PATCH] libmultipath: don't dlclose tur checker DSO
The multipathd tur checker thread is designed to be able to finish at
any time, even after the tur checker itself has been freed. The
multipathd shutdown code makes sure all the checkers have been freed
before freeing the checker_class and calling dlclose() to unload the
DSO, but this doesn't guarantee that the checker threads have finished.
If one hasn't, the DSO will get unloaded while the thread still running
code from it, causing a segfault. Unfortunately, it's not possible to be
sure that all tur checker threads have ended during shutdown, without
making them joinable.
However, since libmultipath will never be reinitialized after it has
been uninitialzed, not dlclosing the tur checker DSO once a thread is
started has minimal cost (keeping the DSO code around until the program
exits, which usually happens right after freeing the checkers).
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/checkers.c | 10 +++++++++-
libmultipath/checkers.h | 1 +
libmultipath/checkers/tur.c | 1 +
3 files changed, 11 insertions(+), 1 deletion(-)
diff --git a/libmultipath/checkers.c b/libmultipath/checkers.c
index 8d2be8a9..6359e5d8 100644
--- a/libmultipath/checkers.c
+++ b/libmultipath/checkers.c
@@ -21,6 +21,7 @@ struct checker_class {
void (*reset)(void); /* to reset the global variables */
const char **msgtable;
short msgtable_size;
+ int keep_dso;
};
char *checker_state_names[] = {
@@ -69,7 +70,7 @@ void free_checker_class(struct checker_class *c)
list_del(&c->node);
if (c->reset)
c->reset();
- if (c->handle) {
+ if (c->handle && !c->keep_dso) {
if (dlclose(c->handle) != 0) {
condlog(0, "Cannot unload checker %s: %s",
c->name, dlerror());
@@ -192,6 +193,13 @@ out:
return NULL;
}
+void checker_keep_dso(struct checker * c)
+{
+ if (!c || !c->cls)
+ return;
+ c->cls->keep_dso = 1;
+}
+
void checker_set_fd (struct checker * c, int fd)
{
if (!c)
diff --git a/libmultipath/checkers.h b/libmultipath/checkers.h
index b458118d..f183cff9 100644
--- a/libmultipath/checkers.h
+++ b/libmultipath/checkers.h
@@ -145,6 +145,7 @@ void checker_reset (struct checker *);
void checker_set_sync (struct checker *);
void checker_set_async (struct checker *);
void checker_set_fd (struct checker *, int);
+void checker_keep_dso(struct checker *c);
void checker_enable (struct checker *);
void checker_disable (struct checker *);
int checker_check (struct checker *, int);
diff --git a/libmultipath/checkers/tur.c b/libmultipath/checkers/tur.c
index e886fcf8..fd58d62a 100644
--- a/libmultipath/checkers/tur.c
+++ b/libmultipath/checkers/tur.c
@@ -394,6 +394,7 @@ int libcheck_check(struct checker * c)
uatomic_set(&ct->running, 1);
tur_set_async_timeout(c);
setup_thread_attr(&attr, 32 * 1024, 1);
+ checker_keep_dso(c);
r = pthread_create(&ct->thread, &attr, tur_thread, ct);
pthread_attr_destroy(&attr);
if (r) {
--
2.17.2

41
SOURCES/0045-libmultipath-warn-about-missing-braces-at-end-of-mul.patch Normal file
View File

@ -0,0 +1,41 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 16 Nov 2020 13:52:09 -0600
Subject: [PATCH] libmultipath: warn about missing braces at end of
multipath.conf
Multipath doesn't warn when multipath.conf is missing closing braces at
the end of the file. This has confused people about the correct config
file syntax, so add a warning.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/parser.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/libmultipath/parser.c b/libmultipath/parser.c
index a7285a35..48b54e87 100644
--- a/libmultipath/parser.c
+++ b/libmultipath/parser.c
@@ -544,7 +544,7 @@ process_stream(struct config *conf, FILE *stream, vector keywords, char *file)
if (!strcmp(str, EOB)) {
if (kw_level > 0) {
free_strvec(strvec);
- break;
+ goto out;
}
condlog(0, "unmatched '%s' at line %d of %s",
EOB, line_nr, file);
@@ -583,7 +583,8 @@ process_stream(struct config *conf, FILE *stream, vector keywords, char *file)
free_strvec(strvec);
}
-
+ if (kw_level == 1)
+ condlog(1, "missing '%s' at end of %s", EOB, file);
out:
FREE(buf);
free_uniques(uniques);
--
2.17.2

36
SOURCES/0046-libmultipath-ignore-multipaths-sections-without-wwid.patch Normal file
View File

@ -0,0 +1,36 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 16 Nov 2020 16:38:21 -0600
Subject: [PATCH] libmultipath: ignore multipaths sections without wwid option
"multipathd show config local" was crashing in find_mp_by_wwid() if
the multipath configuration included a multipaths section that did
not set a wwid option. There is no reason to keep a mpentry that
didn't set its wwid. Remove it in merge_mptable().
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/config.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/libmultipath/config.c b/libmultipath/config.c
index a71db2d0..dc81c994 100644
--- a/libmultipath/config.c
+++ b/libmultipath/config.c
@@ -444,6 +444,13 @@ void merge_mptable(vector mptable)
int i, j;
vector_foreach_slot(mptable, mp1, i) {
+ /* drop invalid multipath configs */
+ if (!mp1->wwid) {
+ condlog(0, "multipaths config section missing wwid");
+ vector_del_slot(mptable, i--);
+ free_mpe(mp1);
+ continue;
+ }
j = i + 1;
vector_foreach_slot_after(mptable, mp2, j) {
if (strcmp(mp1->wwid, mp2->wwid))
--
2.17.2

25
SOURCES/0047-tests-fix-missing-priority-blacklist-test.patch Normal file
View File

@ -0,0 +1,25 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 16 Nov 2020 23:21:41 -0600
Subject: [PATCH] tests: fix missing priority blacklist test
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
tests/blacklist.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/tests/blacklist.c b/tests/blacklist.c
index d20e97af..401600d9 100644
--- a/tests/blacklist.c
+++ b/tests/blacklist.c
@@ -378,7 +378,6 @@ static void test_property_missing(void **state)
{
static struct udev_device udev = { "sdb", { "ID_FOO", "ID_BAZ", "ID_BAR", "ID_SERIAL", NULL } };
conf.blist_property = blist_property_wwn;
- expect_condlog(3, "sdb: blacklisted, udev property missing\n");
assert_int_equal(filter_property(&conf, &udev, 3, "ID_SERIAL"),
MATCH_NOTHING);
assert_int_equal(filter_property(&conf, &udev, 3, "ID_BLAH"),
--
2.17.2

33
SOURCES/0048-mpathpersist-Fix-Register-and-Ignore-with-0x00-SARK.patch Normal file
View File

@ -0,0 +1,33 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 5 Nov 2020 09:15:43 -0600
Subject: [PATCH] mpathpersist: Fix Register and Ignore with 0x00 SARK
When the Register and Ignore command is run with sg_persist, if a 0x00
Service Action Reservation Key is given or the --param-sark option is
not used at all, sg_persist will clear the registration. mpathpersist
will fail with an error. This patch fixes mpathpersist to work like
sg_persist in this case.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmpathpersist/mpath_persist.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/libmpathpersist/mpath_persist.c b/libmpathpersist/mpath_persist.c
index 3da7a6cf..aa196008 100644
--- a/libmpathpersist/mpath_persist.c
+++ b/libmpathpersist/mpath_persist.c
@@ -321,7 +321,8 @@ int __mpath_persistent_reserve_out ( int fd, int rq_servact, int rq_scope,
}
if (memcmp(paramp->key, &mpp->reservation_key, 8) &&
- memcmp(paramp->sa_key, &mpp->reservation_key, 8)) {
+ memcmp(paramp->sa_key, &mpp->reservation_key, 8) &&
+ (prkey || rq_servact != MPATH_PROUT_REG_IGN_SA)) {
condlog(0, "%s: configured reservation key doesn't match: 0x%" PRIx64, alias, get_be64(mpp->reservation_key));
ret = MPATH_PR_SYNTAX_ERROR;
goto out1;
--
2.17.2

38
SOURCES/0049-mpathpersist-update-prkeys-file-on-changing-registra.patch Normal file
View File

@ -0,0 +1,38 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 23 Nov 2020 20:45:50 -0600
Subject: [PATCH] mpathpersist: update prkeys file on changing registrations
When the "reservation_key" option is set to "file" and Register command
is run with both the current Reservation Key and a new Service Action
Reservation Key, mpathpersist will change the registration, but will not
update the prkeys file. This means that future paths that come online
will not be able to register, since multipathd is still using the old
reservation key. Fix this.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmpathpersist/mpath_persist.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/libmpathpersist/mpath_persist.c b/libmpathpersist/mpath_persist.c
index aa196008..a01dfb0b 100644
--- a/libmpathpersist/mpath_persist.c
+++ b/libmpathpersist/mpath_persist.c
@@ -307,9 +307,10 @@ int __mpath_persistent_reserve_out ( int fd, int rq_servact, int rq_scope,
memcpy(&prkey, paramp->sa_key, 8);
if (mpp->prkey_source == PRKEY_SOURCE_FILE && prkey &&
- ((!get_be64(mpp->reservation_key) &&
- rq_servact == MPATH_PROUT_REG_SA) ||
- rq_servact == MPATH_PROUT_REG_IGN_SA)) {
+ (rq_servact == MPATH_PROUT_REG_IGN_SA ||
+ (rq_servact == MPATH_PROUT_REG_SA &&
+ (!get_be64(mpp->reservation_key) ||
+ memcmp(paramp->key, &mpp->reservation_key, 8) == 0)))) {
memcpy(&mpp->reservation_key, paramp->sa_key, 8);
if (update_prkey_flags(alias, get_be64(mpp->reservation_key),
paramp->sa_flags)) {
--
2.17.2

37
SOURCES/0050-multipathd-Fix-multipathd-stopping-on-shutdown.patch Normal file
View File

@ -0,0 +1,37 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 14 Dec 2020 14:16:29 -0600
Subject: [PATCH] multipathd: Fix multipathd stopping on shutdown
According to man "systemd.special"
"shutdown.target: ... Services that shall be terminated on system
shutdown shall add Conflicts= and Before= dependencies to this unit for
their service unit, which is implicitly done when
DefaultDependencies=yes is set (the default)."
multipathd.service sets DefaultDependencies=no and includes the
Conflits= dependency, but not the Before= one. This can cause multipathd
to continue running past when it is supposed to during shutdown.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
multipathd/multipathd.service | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/multipathd/multipathd.service b/multipathd/multipathd.service
index 0fbcc46b..bc8fa07a 100644
--- a/multipathd/multipathd.service
+++ b/multipathd/multipathd.service
@@ -2,7 +2,7 @@
Description=Device-Mapper Multipath Device Controller
Wants=systemd-udev-trigger.service systemd-udev-settle.service
Before=iscsi.service iscsid.service lvm2-activation-early.service
-Before=local-fs-pre.target blk-availability.service
+Before=local-fs-pre.target blk-availability.service shutdown.target
After=multipathd.socket systemd-udev-trigger.service systemd-udev-settle.service
ConditionPathExists=/etc/multipath.conf
DefaultDependencies=no
--
2.17.2

36
SOURCES/0051-multipath.conf.5-Improve-checker_timeout-description.patch Normal file
View File

@ -0,0 +1,36 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue, 15 Dec 2020 12:47:14 -0600
Subject: [PATCH] multipath.conf.5: Improve checker_timeout description
I was asked to explain how checker_timeout works for checkers like
directio, that don't issue scsi commands with an explicit timeout
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
multipath/multipath.conf.5 | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/multipath/multipath.conf.5 b/multipath/multipath.conf.5
index 60954574..a5686090 100644
--- a/multipath/multipath.conf.5
+++ b/multipath/multipath.conf.5
@@ -634,8 +634,13 @@ The default is: \fBno\fR
.
.TP
.B checker_timeout
-Specify the timeout to use for path checkers and prioritizers that issue SCSI
-commands with an explicit timeout, in seconds.
+Specify the timeout to use for path checkers and prioritizers, in seconds.
+Only prioritizers that issue scsi commands use checker_timeout. Checkers
+that support an asynchronous mode (\fItur\fR and \fIdirectio\fR), will
+return shortly after being called by multipathd, regardless of whether the
+storage array responds. If the storage array hasn't responded, mulitpathd will
+check for a response every second, until \fIchecker_timeout\fR seconds have
+elapsed.
.RS
.TP
The default is: in \fB/sys/block/sd<x>/device/timeout\fR
--
2.17.2

34
SOURCES/0052-libmultipath-check-for-null-wwid-before-strcmp.patch Normal file
View File

@ -0,0 +1,34 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 4 Jan 2021 21:59:54 -0600
Subject: [PATCH] libmultipath: check for null wwid before strcmp
Commit 749aabd0 (libmultipath: ignore multipaths sections without wwid
option) removed all mpentries with a NULL wwid, but didn't stop strcmp()
from being run on them in merge_mptable(). The result of strcmp() with
a NULL parameter is undefined, so fix that.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Reviewed-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/config.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libmultipath/config.c b/libmultipath/config.c
index dc81c994..dd645f17 100644
--- a/libmultipath/config.c
+++ b/libmultipath/config.c
@@ -453,7 +453,7 @@ void merge_mptable(vector mptable)
}
j = i + 1;
vector_foreach_slot_after(mptable, mp2, j) {
- if (strcmp(mp1->wwid, mp2->wwid))
+ if (!mp2->wwid || strcmp(mp1->wwid, mp2->wwid))
continue;
condlog(1, "%s: duplicate multipath config section for %s",
__func__, mp1->wwid);
--
2.17.2

27
SOURCES/0053-libmultipath-make-find_err_path_by_dev-static.patch Normal file
View File

@ -0,0 +1,27 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 14 Jan 2021 20:20:22 -0600
Subject: [PATCH] libmultipath: make find_err_path_by_dev() static
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Reviewed-by: Martin Wilck <mwilck@suse.com>
---
libmultipath/io_err_stat.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libmultipath/io_err_stat.c b/libmultipath/io_err_stat.c
index 1b9cd6c0..449760a0 100644
--- a/libmultipath/io_err_stat.c
+++ b/libmultipath/io_err_stat.c
@@ -89,7 +89,7 @@ static void rcu_unregister(__attribute__((unused)) void *param)
rcu_unregister_thread();
}
-struct io_err_stat_path *find_err_path_by_dev(vector pathvec, char *dev)
+static struct io_err_stat_path *find_err_path_by_dev(vector pathvec, char *dev)
{
int i;
struct io_err_stat_path *pp;
--
2.17.2

291
SOURCES/0054-multipathd-avoid-io_err_stat-crash-during-shutdown.patch Normal file
View File

@ -0,0 +1,291 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 14 Jan 2021 20:20:23 -0600
Subject: [PATCH] multipathd: avoid io_err_stat crash during shutdown
The checker thread is reponsible for enqueueing paths for the
io_err_stat thread to check. During shutdown, the io_err_stat thread is
shut down and cleaned up before the checker thread. There is no code
to make sure that the checker thread isn't accessing the io_err_stat
pathvec or its mutex while they are being freed, which can lead to
memory corruption crashes.
To solve this, get rid of the io_err_stat_pathvec structure, and
statically define the mutex. This means that the mutex is always valid
to access, and the io_err_stat pathvec can only be accessed while
holding it. If the io_err_stat thread has already been cleaned up
when the checker tries to access the pathvec, it will be NULL, and the
checker will simply fail to enqueue the path.
This change also fixes a bug in free_io_err_pathvec(), which previously
only attempted to free the pathvec if it was not set, instead of when it
was set.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Reviewed-by: Martin Wilck <mwilck@suse.com>
---
libmultipath/io_err_stat.c | 112 +++++++++++++++----------------------
libmultipath/util.c | 5 ++
libmultipath/util.h | 1 +
3 files changed, 50 insertions(+), 68 deletions(-)
diff --git a/libmultipath/io_err_stat.c b/libmultipath/io_err_stat.c
index 449760a0..f6c564f0 100644
--- a/libmultipath/io_err_stat.c
+++ b/libmultipath/io_err_stat.c
@@ -34,6 +34,7 @@
#include "lock.h"
#include "time-util.h"
#include "io_err_stat.h"
+#include "util.h"
#define IOTIMEOUT_SEC 60
#define TIMEOUT_NO_IO_NSEC 10000000 /*10ms = 10000000ns*/
@@ -46,12 +47,6 @@
#define io_err_stat_log(prio, fmt, args...) \
condlog(prio, "io error statistic: " fmt, ##args)
-
-struct io_err_stat_pathvec {
- pthread_mutex_t mutex;
- vector pathvec;
-};
-
struct dio_ctx {
struct timespec io_starttime;
unsigned int blksize;
@@ -76,9 +71,10 @@ pthread_attr_t io_err_stat_attr;
static pthread_mutex_t io_err_thread_lock = PTHREAD_MUTEX_INITIALIZER;
static pthread_cond_t io_err_thread_cond = PTHREAD_COND_INITIALIZER;
+static pthread_mutex_t io_err_pathvec_lock = PTHREAD_MUTEX_INITIALIZER;
static int io_err_thread_running = 0;
-static struct io_err_stat_pathvec *paths;
+static vector io_err_pathvec;
struct vectors *vecs;
io_context_t ioctx;
@@ -208,46 +204,23 @@ static void free_io_err_stat_path(struct io_err_stat_path *p)
FREE(p);
}
-static struct io_err_stat_pathvec *alloc_pathvec(void)
-{
- struct io_err_stat_pathvec *p;
- int r;
-
- p = (struct io_err_stat_pathvec *)MALLOC(sizeof(*p));
- if (!p)
- return NULL;
- p->pathvec = vector_alloc();
- if (!p->pathvec)
- goto out_free_struct_pathvec;
- r = pthread_mutex_init(&p->mutex, NULL);
- if (r)
- goto out_free_member_pathvec;
-
- return p;
-
-out_free_member_pathvec:
- vector_free(p->pathvec);
-out_free_struct_pathvec:
- FREE(p);
- return NULL;
-}
-
-static void free_io_err_pathvec(struct io_err_stat_pathvec *p)
+static void free_io_err_pathvec(void)
{
struct io_err_stat_path *path;
int i;
- if (!p)
- return;
- pthread_mutex_destroy(&p->mutex);
- if (!p->pathvec) {
- vector_foreach_slot(p->pathvec, path, i) {
- destroy_directio_ctx(path);
- free_io_err_stat_path(path);
- }
- vector_free(p->pathvec);
+ pthread_mutex_lock(&io_err_pathvec_lock);
+ pthread_cleanup_push(cleanup_mutex, &io_err_pathvec_lock);
+ if (!io_err_pathvec)
+ goto out;
+ vector_foreach_slot(io_err_pathvec, path, i) {
+ destroy_directio_ctx(path);
+ free_io_err_stat_path(path);
}
- FREE(p);
+ vector_free(io_err_pathvec);
+ io_err_pathvec = NULL;
+out:
+ pthread_cleanup_pop(1);
}
/*
@@ -259,13 +232,13 @@ static int enqueue_io_err_stat_by_path(struct path *path)
{
struct io_err_stat_path *p;
- pthread_mutex_lock(&paths->mutex);
- p = find_err_path_by_dev(paths->pathvec, path->dev);
+ pthread_mutex_lock(&io_err_pathvec_lock);
+ p = find_err_path_by_dev(io_err_pathvec, path->dev);
if (p) {
- pthread_mutex_unlock(&paths->mutex);
+ pthread_mutex_unlock(&io_err_pathvec_lock);
return 0;
}
- pthread_mutex_unlock(&paths->mutex);
+ pthread_mutex_unlock(&io_err_pathvec_lock);
p = alloc_io_err_stat_path();
if (!p)
@@ -277,18 +250,18 @@ static int enqueue_io_err_stat_by_path(struct path *path)
if (setup_directio_ctx(p))
goto free_ioerr_path;
- pthread_mutex_lock(&paths->mutex);
- if (!vector_alloc_slot(paths->pathvec))
+ pthread_mutex_lock(&io_err_pathvec_lock);
+ if (!vector_alloc_slot(io_err_pathvec))
goto unlock_destroy;
- vector_set_slot(paths->pathvec, p);
- pthread_mutex_unlock(&paths->mutex);
+ vector_set_slot(io_err_pathvec, p);
+ pthread_mutex_unlock(&io_err_pathvec_lock);
io_err_stat_log(2, "%s: enqueue path %s to check",
path->mpp->alias, path->dev);
return 0;
unlock_destroy:
- pthread_mutex_unlock(&paths->mutex);
+ pthread_mutex_unlock(&io_err_pathvec_lock);
destroy_directio_ctx(p);
free_ioerr_path:
free_io_err_stat_path(p);
@@ -421,9 +394,9 @@ static int delete_io_err_stat_by_addr(struct io_err_stat_path *p)
{
int i;
- i = find_slot(paths->pathvec, p);
+ i = find_slot(io_err_pathvec, p);
if (i != -1)
- vector_del_slot(paths->pathvec, i);
+ vector_del_slot(io_err_pathvec, i);
destroy_directio_ctx(p);
free_io_err_stat_path(p);
@@ -594,7 +567,7 @@ static void poll_async_io_timeout(void)
if (clock_gettime(CLOCK_MONOTONIC, &curr_time) != 0)
return;
- vector_foreach_slot(paths->pathvec, pp, i) {
+ vector_foreach_slot(io_err_pathvec, pp, i) {
for (j = 0; j < CONCUR_NR_EVENT; j++) {
rc = try_to_cancel_timeout_io(pp->dio_ctx_array + j,
&curr_time, pp->devname);
@@ -640,7 +613,7 @@ static void handle_async_io_done_event(struct io_event *io_evt)
int rc = PATH_UNCHECKED;
int i, j;
- vector_foreach_slot(paths->pathvec, pp, i) {
+ vector_foreach_slot(io_err_pathvec, pp, i) {
for (j = 0; j < CONCUR_NR_EVENT; j++) {
ct = pp->dio_ctx_array + j;
if (&ct->io == io_evt->obj) {
@@ -674,19 +647,14 @@ static void service_paths(void)
struct io_err_stat_path *pp;
int i;
- pthread_mutex_lock(&paths->mutex);
- vector_foreach_slot(paths->pathvec, pp, i) {
+ pthread_mutex_lock(&io_err_pathvec_lock);
+ vector_foreach_slot(io_err_pathvec, pp, i) {
send_batch_async_ios(pp);
process_async_ios_event(TIMEOUT_NO_IO_NSEC, pp->devname);
poll_async_io_timeout();
poll_io_err_stat(vecs, pp);
}
- pthread_mutex_unlock(&paths->mutex);
-}
-
-static void cleanup_unlock(void *arg)
-{
- pthread_mutex_unlock((pthread_mutex_t*) arg);
+ pthread_mutex_unlock(&io_err_pathvec_lock);
}
static void cleanup_exited(__attribute__((unused)) void *arg)
@@ -744,12 +712,17 @@ int start_io_err_stat_thread(void *data)
io_err_stat_log(4, "io_setup failed");
return 1;
}
- paths = alloc_pathvec();
- if (!paths)
+
+ pthread_mutex_lock(&io_err_pathvec_lock);
+ io_err_pathvec = vector_alloc();
+ if (!io_err_pathvec) {
+ pthread_mutex_unlock(&io_err_pathvec_lock);
goto destroy_ctx;
+ }
+ pthread_mutex_unlock(&io_err_pathvec_lock);
pthread_mutex_lock(&io_err_thread_lock);
- pthread_cleanup_push(cleanup_unlock, &io_err_thread_lock);
+ pthread_cleanup_push(cleanup_mutex, &io_err_thread_lock);
ret = pthread_create(&io_err_stat_thr, &io_err_stat_attr,
io_err_stat_loop, data);
@@ -769,7 +742,10 @@ int start_io_err_stat_thread(void *data)
return 0;
out_free:
- free_io_err_pathvec(paths);
+ pthread_mutex_lock(&io_err_pathvec_lock);
+ vector_free(io_err_pathvec);
+ io_err_pathvec = NULL;
+ pthread_mutex_unlock(&io_err_pathvec_lock);
destroy_ctx:
io_destroy(ioctx);
io_err_stat_log(0, "failed to start io_error statistic thread");
@@ -785,6 +761,6 @@ void stop_io_err_stat_thread(void)
pthread_cancel(io_err_stat_thr);
pthread_join(io_err_stat_thr, NULL);
- free_io_err_pathvec(paths);
+ free_io_err_pathvec();
io_destroy(ioctx);
}
diff --git a/libmultipath/util.c b/libmultipath/util.c
index 51c38c87..dd30a46e 100644
--- a/libmultipath/util.c
+++ b/libmultipath/util.c
@@ -469,3 +469,8 @@ void close_fd(void *arg)
{
close((long)arg);
}
+
+void cleanup_mutex(void *arg)
+{
+ pthread_mutex_unlock(arg);
+}
diff --git a/libmultipath/util.h b/libmultipath/util.h
index 56bd78c7..ce277680 100644
--- a/libmultipath/util.h
+++ b/libmultipath/util.h
@@ -44,6 +44,7 @@ void set_max_fds(rlim_t max_fds);
pthread_cleanup_push(((void (*)(void *))&f), (arg))
void close_fd(void *arg);
+void cleanup_mutex(void *arg);
struct scandir_result {
struct dirent **di;
--
2.17.2

146
SOURCES/0055-multipathd-avoid-io_err_stat-ABBA-deadlock.patch Normal file
View File

@ -0,0 +1,146 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 14 Jan 2021 20:20:24 -0600
Subject: [PATCH] multipathd: avoid io_err_stat ABBA deadlock
When the checker thread enqueues paths for the io_err_stat thread to
check, it calls enqueue_io_err_stat_by_path() with the vecs lock held.
start_io_err_stat_thread() is also called with the vecs lock held.
These two functions both lock io_err_pathvec_lock. When the io_err_stat
thread updates the paths in vecs->pathvec in poll_io_err_stat(), it has
the io_err_pathvec_lock held, and then locks the vecs lock. This can
cause an ABBA deadlock.
To solve this, service_paths() no longer updates the paths in
vecs->pathvec with the io_err_pathvec_lock held. It does this by moving
the io_err_stat_path from io_err_pathvec to a local vector when it needs
to update the path. After releasing the io_err_pathvec_lock, it goes
through this temporary vector, updates the paths with the vecs lock
held, and then frees everything.
This change fixes a bug in service_paths() where elements were being
deleted from io_err_pathvec, without the index being decremented,
causing the loop to skip elements. Also, service_paths() could be
cancelled while holding the io_err_pathvec_lock, so it should have a
cleanup handler.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Reviewed-by: Martin Wilck <mwilck@suse.com>
---
libmultipath/io_err_stat.c | 56 ++++++++++++++++++++++----------------
1 file changed, 32 insertions(+), 24 deletions(-)
diff --git a/libmultipath/io_err_stat.c b/libmultipath/io_err_stat.c
index f6c564f0..63ee2e07 100644
--- a/libmultipath/io_err_stat.c
+++ b/libmultipath/io_err_stat.c
@@ -390,20 +390,6 @@ recover:
return 0;
}
-static int delete_io_err_stat_by_addr(struct io_err_stat_path *p)
-{
- int i;
-
- i = find_slot(io_err_pathvec, p);
- if (i != -1)
- vector_del_slot(io_err_pathvec, i);
-
- destroy_directio_ctx(p);
- free_io_err_stat_path(p);
-
- return 0;
-}
-
static void account_async_io_state(struct io_err_stat_path *pp, int rc)
{
switch (rc) {
@@ -420,17 +406,26 @@ static void account_async_io_state(struct io_err_stat_path *pp, int rc)
}
}
-static int poll_io_err_stat(struct vectors *vecs, struct io_err_stat_path *pp)
+static int io_err_stat_time_up(struct io_err_stat_path *pp)
{
struct timespec currtime, difftime;
- struct path *path;
- double err_rate;
if (clock_gettime(CLOCK_MONOTONIC, &currtime) != 0)
- return 1;
+ return 0;
timespecsub(&currtime, &pp->start_time, &difftime);
if (difftime.tv_sec < pp->total_time)
return 0;
+ return 1;
+}
+
+static void end_io_err_stat(struct io_err_stat_path *pp)
+{
+ struct timespec currtime;
+ struct path *path;
+ double err_rate;
+
+ if (clock_gettime(CLOCK_MONOTONIC, &currtime) != 0)
+ currtime = pp->start_time;
io_err_stat_log(4, "%s: check end", pp->devname);
@@ -469,10 +464,6 @@ static int poll_io_err_stat(struct vectors *vecs, struct io_err_stat_path *pp)
pp->devname);
}
lock_cleanup_pop(vecs->lock);
-
- delete_io_err_stat_by_addr(pp);
-
- return 0;
}
static int send_each_async_io(struct dio_ctx *ct, int fd, char *dev)
@@ -632,6 +623,7 @@ static void process_async_ios_event(int timeout_nsecs, char *dev)
struct timespec timeout = { .tv_nsec = timeout_nsecs };
errno = 0;
+ pthread_testcancel();
n = io_getevents(ioctx, 1L, CONCUR_NR_EVENT, events, &timeout);
if (n < 0) {
io_err_stat_log(3, "%s: async io events returned %d (errno=%s)",
@@ -644,17 +636,33 @@ static void process_async_ios_event(int timeout_nsecs, char *dev)
static void service_paths(void)
{
+ struct _vector _pathvec = {0};
+ /* avoid gcc warnings that &_pathvec will never be NULL in vector ops */
+ struct _vector * const tmp_pathvec = &_pathvec;
struct io_err_stat_path *pp;
int i;
pthread_mutex_lock(&io_err_pathvec_lock);
+ pthread_cleanup_push(cleanup_mutex, &io_err_pathvec_lock);
vector_foreach_slot(io_err_pathvec, pp, i) {
send_batch_async_ios(pp);
process_async_ios_event(TIMEOUT_NO_IO_NSEC, pp->devname);
poll_async_io_timeout();
- poll_io_err_stat(vecs, pp);
+ if (io_err_stat_time_up(pp)) {
+ if (!vector_alloc_slot(tmp_pathvec))
+ continue;
+ vector_del_slot(io_err_pathvec, i--);
+ vector_set_slot(tmp_pathvec, pp);
+ }
}
- pthread_mutex_unlock(&io_err_pathvec_lock);
+ pthread_cleanup_pop(1);
+ vector_foreach_slot_backwards(tmp_pathvec, pp, i) {
+ end_io_err_stat(pp);
+ vector_del_slot(tmp_pathvec, i);
+ destroy_directio_ctx(pp);
+ free_io_err_stat_path(pp);
+ }
+ vector_reset(tmp_pathvec);
}
static void cleanup_exited(__attribute__((unused)) void *arg)
--
2.17.2

111
SOURCES/0056-multipathd-use-get_monotonic_time-in-io_err_stat-cod.patch Normal file
View File

@ -0,0 +1,111 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 14 Jan 2021 20:20:25 -0600
Subject: [PATCH] multipathd: use get_monotonic_time() in io_err_stat code
Instead of calling clock_gettime(), and dealing with failure
conditions, just call get_monotonic_time().
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Reviewed-by: Martin Wilck <mwilck@suse.com>
---
libmultipath/io_err_stat.c | 34 +++++++++++-----------------------
1 file changed, 11 insertions(+), 23 deletions(-)
diff --git a/libmultipath/io_err_stat.c b/libmultipath/io_err_stat.c
index 63ee2e07..3389d693 100644
--- a/libmultipath/io_err_stat.c
+++ b/libmultipath/io_err_stat.c
@@ -305,8 +305,7 @@ int io_err_stat_handle_pathfail(struct path *path)
* the repeated count threshold and time frame, we assume a path
* which fails at least twice within 60 seconds is flaky.
*/
- if (clock_gettime(CLOCK_MONOTONIC, &curr_time) != 0)
- return 1;
+ get_monotonic_time(&curr_time);
if (path->io_err_pathfail_cnt == 0) {
path->io_err_pathfail_cnt++;
path->io_err_pathfail_starttime = curr_time.tv_sec;
@@ -362,9 +361,9 @@ int need_io_err_check(struct path *pp)
}
if (pp->io_err_pathfail_cnt != PATH_IO_ERR_WAITING_TO_CHECK)
return 1;
- if (clock_gettime(CLOCK_MONOTONIC, &curr_time) != 0 ||
- (curr_time.tv_sec - pp->io_err_dis_reinstate_time) >
- pp->mpp->marginal_path_err_recheck_gap_time) {
+ get_monotonic_time(&curr_time);
+ if ((curr_time.tv_sec - pp->io_err_dis_reinstate_time) >
+ pp->mpp->marginal_path_err_recheck_gap_time) {
io_err_stat_log(4, "%s: reschedule checking after %d seconds",
pp->dev,
pp->mpp->marginal_path_err_recheck_gap_time);
@@ -410,8 +409,7 @@ static int io_err_stat_time_up(struct io_err_stat_path *pp)
{
struct timespec currtime, difftime;
- if (clock_gettime(CLOCK_MONOTONIC, &currtime) != 0)
- return 0;
+ get_monotonic_time(&currtime);
timespecsub(&currtime, &pp->start_time, &difftime);
if (difftime.tv_sec < pp->total_time)
return 0;
@@ -424,8 +422,7 @@ static void end_io_err_stat(struct io_err_stat_path *pp)
struct path *path;
double err_rate;
- if (clock_gettime(CLOCK_MONOTONIC, &currtime) != 0)
- currtime = pp->start_time;
+ get_monotonic_time(&currtime);
io_err_stat_log(4, "%s: check end", pp->devname);
@@ -474,11 +471,7 @@ static int send_each_async_io(struct dio_ctx *ct, int fd, char *dev)
ct->io_starttime.tv_sec == 0) {
struct iocb *ios[1] = { &ct->io };
- if (clock_gettime(CLOCK_MONOTONIC, &ct->io_starttime) != 0) {
- ct->io_starttime.tv_sec = 0;
- ct->io_starttime.tv_nsec = 0;
- return rc;
- }
+ get_monotonic_time(&ct->io_starttime);
io_prep_pread(&ct->io, fd, ct->buf, ct->blksize, 0);
if (io_submit(ioctx, 1, ios) != 1) {
io_err_stat_log(5, "%s: io_submit error %i",
@@ -497,8 +490,7 @@ static void send_batch_async_ios(struct io_err_stat_path *pp)
struct dio_ctx *ct;
struct timespec currtime, difftime;
- if (clock_gettime(CLOCK_MONOTONIC, &currtime) != 0)
- return;
+ get_monotonic_time(&currtime);
/*
* Give a free time for all IO to complete or timeout
*/
@@ -513,11 +505,8 @@ static void send_batch_async_ios(struct io_err_stat_path *pp)
if (!send_each_async_io(ct, pp->fd, pp->devname))
pp->io_nr++;
}
- if (pp->start_time.tv_sec == 0 && pp->start_time.tv_nsec == 0 &&
- clock_gettime(CLOCK_MONOTONIC, &pp->start_time)) {
- pp->start_time.tv_sec = 0;
- pp->start_time.tv_nsec = 0;
- }
+ if (pp->start_time.tv_sec == 0 && pp->start_time.tv_nsec == 0)
+ get_monotonic_time(&pp->start_time);
}
static int try_to_cancel_timeout_io(struct dio_ctx *ct, struct timespec *t,
@@ -556,8 +545,7 @@ static void poll_async_io_timeout(void)
int rc = PATH_UNCHECKED;
int i, j;
- if (clock_gettime(CLOCK_MONOTONIC, &curr_time) != 0)
- return;
+ get_monotonic_time(&curr_time);
vector_foreach_slot(io_err_pathvec, pp, i) {
for (j = 0; j < CONCUR_NR_EVENT; j++) {
rc = try_to_cancel_timeout_io(pp->dio_ctx_array + j,
--
2.17.2

101
SOURCES/0057-multipathd-combine-free_io_err_stat_path-and-destroy.patch Normal file
View File

@ -0,0 +1,101 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Thu, 14 Jan 2021 20:20:26 -0600
Subject: [PATCH] multipathd: combine free_io_err_stat_path and
destroy_directio_ctx
destroy_directio_ctx() is only called from free_io_err_stat_path(), and
free_io_err_stat_path() is very short, so combine them.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Reviewed-by: Martin Wilck <mwilck@suse.com>
---
libmultipath/io_err_stat.c | 24 ++++++++++--------------
1 file changed, 10 insertions(+), 14 deletions(-)
diff --git a/libmultipath/io_err_stat.c b/libmultipath/io_err_stat.c
index 3389d693..bf1d3910 100644
--- a/libmultipath/io_err_stat.c
+++ b/libmultipath/io_err_stat.c
@@ -163,12 +163,15 @@ fail_close:
return 1;
}
-static void destroy_directio_ctx(struct io_err_stat_path *p)
+static void free_io_err_stat_path(struct io_err_stat_path *p)
{
int i;
- if (!p || !p->dio_ctx_array)
+ if (!p)
return;
+ if (!p->dio_ctx_array)
+ goto free_path;
+
cancel_inflight_io(p);
for (i = 0; i < CONCUR_NR_EVENT; i++)
@@ -177,6 +180,8 @@ static void destroy_directio_ctx(struct io_err_stat_path *p)
if (p->fd > 0)
close(p->fd);
+free_path:
+ FREE(p);
}
static struct io_err_stat_path *alloc_io_err_stat_path(void)
@@ -199,11 +204,6 @@ static struct io_err_stat_path *alloc_io_err_stat_path(void)
return p;
}
-static void free_io_err_stat_path(struct io_err_stat_path *p)
-{
- FREE(p);
-}
-
static void free_io_err_pathvec(void)
{
struct io_err_stat_path *path;
@@ -213,10 +213,8 @@ static void free_io_err_pathvec(void)
pthread_cleanup_push(cleanup_mutex, &io_err_pathvec_lock);
if (!io_err_pathvec)
goto out;
- vector_foreach_slot(io_err_pathvec, path, i) {
- destroy_directio_ctx(path);
+ vector_foreach_slot(io_err_pathvec, path, i)
free_io_err_stat_path(path);
- }
vector_free(io_err_pathvec);
io_err_pathvec = NULL;
out:
@@ -252,7 +250,7 @@ static int enqueue_io_err_stat_by_path(struct path *path)
goto free_ioerr_path;
pthread_mutex_lock(&io_err_pathvec_lock);
if (!vector_alloc_slot(io_err_pathvec))
- goto unlock_destroy;
+ goto unlock_pathvec;
vector_set_slot(io_err_pathvec, p);
pthread_mutex_unlock(&io_err_pathvec_lock);
@@ -260,9 +258,8 @@ static int enqueue_io_err_stat_by_path(struct path *path)
path->mpp->alias, path->dev);
return 0;
-unlock_destroy:
+unlock_pathvec:
pthread_mutex_unlock(&io_err_pathvec_lock);
- destroy_directio_ctx(p);
free_ioerr_path:
free_io_err_stat_path(p);
@@ -647,7 +644,6 @@ static void service_paths(void)
vector_foreach_slot_backwards(tmp_pathvec, pp, i) {
end_io_err_stat(pp);
vector_del_slot(tmp_pathvec, i);
- destroy_directio_ctx(pp);
free_io_err_stat_path(pp);
}
vector_reset(tmp_pathvec);
--
2.17.2

110
SOURCES/0058-multipathd-cleanup-logging-for-marginal-paths.patch Normal file
View File

@ -0,0 +1,110 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 18 Jan 2021 22:46:04 -0600
Subject: [PATCH] multipathd: cleanup logging for marginal paths
io_err_stat logged at level 2 whenever it enqueued a path to check,
which could happen multiple times while a path was marginal. On the
other hand if marginal_pathgroups wasn't set, multipathd didn't log when
paths were set to marginal. Now io_err_stat only logs at level 2 when
something unexpected happens, but multipathd will always log when a
path switches its marginal state.
This patch also fixes an issue where paths in the delayed state could
get set to the pending state if they could not be checked in time.
Aside from going against the idea the paths should not be set to pending
if they already have a valid state, this caused multipathd to log a
message whenever the path state switched to from delayed to pending and
then back.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Reviewed-by: Martin Wilck <mwilck@suse.com>
---
libmultipath/io_err_stat.c | 7 +++----
multipathd/main.c | 20 +++++++++++++-------
2 files changed, 16 insertions(+), 11 deletions(-)
diff --git a/libmultipath/io_err_stat.c b/libmultipath/io_err_stat.c
index bf1d3910..ee711f7f 100644
--- a/libmultipath/io_err_stat.c
+++ b/libmultipath/io_err_stat.c
@@ -254,7 +254,7 @@ static int enqueue_io_err_stat_by_path(struct path *path)
vector_set_slot(io_err_pathvec, p);
pthread_mutex_unlock(&io_err_pathvec_lock);
- io_err_stat_log(2, "%s: enqueue path %s to check",
+ io_err_stat_log(3, "%s: enqueue path %s to check",
path->mpp->alias, path->dev);
return 0;
@@ -353,7 +353,7 @@ int need_io_err_check(struct path *pp)
if (uatomic_read(&io_err_thread_running) == 0)
return 0;
if (count_active_paths(pp->mpp) <= 0) {
- io_err_stat_log(2, "%s: recover path early", pp->dev);
+ io_err_stat_log(2, "%s: no paths. recovering early", pp->dev);
goto recover;
}
if (pp->io_err_pathfail_cnt != PATH_IO_ERR_WAITING_TO_CHECK)
@@ -371,8 +371,7 @@ int need_io_err_check(struct path *pp)
* Or else, return 1 to set path state to PATH_SHAKY
*/
if (r == 1) {
- io_err_stat_log(3, "%s: enqueue fails, to recover",
- pp->dev);
+ io_err_stat_log(2, "%s: enqueue failed. recovering early", pp->dev);
goto recover;
} else
pp->io_err_pathfail_cnt = PATH_IO_ERR_IN_CHECKING;
diff --git a/multipathd/main.c b/multipathd/main.c
index 1d0579e9..cc1aeea2 100644
--- a/multipathd/main.c
+++ b/multipathd/main.c
@@ -2041,8 +2041,8 @@ check_path (struct vectors * vecs, struct path * pp, unsigned int ticks)
pathinfo(pp, conf, 0);
pthread_cleanup_pop(1);
return 1;
- } else if ((newstate != PATH_UP && newstate != PATH_GHOST) &&
- (pp->state == PATH_DELAYED)) {
+ } else if ((newstate != PATH_UP && newstate != PATH_GHOST &&
+ newstate != PATH_PENDING) && (pp->state == PATH_DELAYED)) {
/* If path state become failed again cancel path delay state */
pp->state = newstate;
return 1;
@@ -2104,8 +2104,9 @@ check_path (struct vectors * vecs, struct path * pp, unsigned int ticks)
if ((newstate == PATH_UP || newstate == PATH_GHOST) &&
(san_path_check_enabled(pp->mpp) ||
marginal_path_check_enabled(pp->mpp))) {
- int was_marginal = pp->marginal;
if (should_skip_path(pp)) {
+ if (!pp->marginal && pp->state != PATH_DELAYED)
+ condlog(2, "%s: path is now marginal", pp->dev);
if (!marginal_pathgroups) {
if (marginal_path_check_enabled(pp->mpp))
/* to reschedule as soon as possible,
@@ -2115,13 +2116,18 @@ check_path (struct vectors * vecs, struct path * pp, unsigned int ticks)
pp->state = PATH_DELAYED;
return 1;
}
- if (!was_marginal) {
+ if (!pp->marginal) {
pp->marginal = 1;
marginal_changed = 1;
}
- } else if (marginal_pathgroups && was_marginal) {
- pp->marginal = 0;
- marginal_changed = 1;
+ } else {
+ if (pp->marginal || pp->state == PATH_DELAYED)
+ condlog(2, "%s: path is no longer marginal",
+ pp->dev);
+ if (marginal_pathgroups && pp->marginal) {
+ pp->marginal = 0;
+ marginal_changed = 1;
+ }
}
}
--
2.17.2

277
SOURCES/0059-libmpathpersist-fix-thread-safety-of-default-functio.patch Normal file
View File

@ -0,0 +1,277 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Mon, 25 Jan 2021 23:31:04 -0600
Subject: [PATCH] libmpathpersist: fix thread safety of default functions
commit a839e39e ("libmpathpersist: factor out initialization and
teardown") made mpath_presistent_reserve_{in,out} use share variables
for curmp and pathvec. There are users of this library that call these
functions in a multi-threaded process, and this change causes their
application to crash. config and udev are also shared variables, but
libmpathpersist doesn't write to the config in
mpath_presistent_reserve_{in,out}, and looking into the libudev code, I
don't see any place where libmpathpersist uses the udev object in a way
that isn't thread-safe.
This patch makes mpath_presistent_reserve_{in,out} go back to using
local variables for curmp and pathvec, so that multiple threads won't
be operating on these variables at the same time.
Fixes: a839e39e ("libmpathpersist: factor out initialization and teardown")
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Reviewed-by: Martin Wilck <mwilck@suse.com>
---
libmpathpersist/mpath_persist.c | 116 +++++++++++++++++++++-----------
libmpathpersist/mpath_persist.h | 24 +++++--
2 files changed, 94 insertions(+), 46 deletions(-)
diff --git a/libmpathpersist/mpath_persist.c b/libmpathpersist/mpath_persist.c
index a01dfb0b..07a5f17f 100644
--- a/libmpathpersist/mpath_persist.c
+++ b/libmpathpersist/mpath_persist.c
@@ -147,72 +147,60 @@ mpath_prin_activepath (struct multipath *mpp, int rq_servact,
return ret;
}
-int mpath_persistent_reserve_in (int fd, int rq_servact,
- struct prin_resp *resp, int noisy, int verbose)
-{
- int ret = mpath_persistent_reserve_init_vecs(verbose);
-
- if (ret != MPATH_PR_SUCCESS)
- return ret;
- ret = __mpath_persistent_reserve_in(fd, rq_servact, resp, noisy);
- mpath_persistent_reserve_free_vecs();
- return ret;
-}
-
-int mpath_persistent_reserve_out ( int fd, int rq_servact, int rq_scope,
- unsigned int rq_type, struct prout_param_descriptor *paramp, int noisy, int verbose)
-{
- int ret = mpath_persistent_reserve_init_vecs(verbose);
-
- if (ret != MPATH_PR_SUCCESS)
- return ret;
- ret = __mpath_persistent_reserve_out(fd, rq_servact, rq_scope, rq_type,
- paramp, noisy);
- mpath_persistent_reserve_free_vecs();
- return ret;
-}
-
static vector curmp;
static vector pathvec;
-void mpath_persistent_reserve_free_vecs(void)
+static void __mpath_persistent_reserve_free_vecs(vector curmp, vector pathvec)
{
free_multipathvec(curmp, KEEP_PATHS);
free_pathvec(pathvec, FREE_PATHS);
+}
+
+void mpath_persistent_reserve_free_vecs(void)
+{
+ __mpath_persistent_reserve_free_vecs(curmp, pathvec);
curmp = pathvec = NULL;
}
-int mpath_persistent_reserve_init_vecs(int verbose)
+static int __mpath_persistent_reserve_init_vecs(vector *curmp_p,
+ vector *pathvec_p, int verbose)
{
struct config *conf = get_multipath_config();
conf->verbosity = verbose;
put_multipath_config(conf);
- if (curmp)
+ if (*curmp_p)
return MPATH_PR_SUCCESS;
/*
* allocate core vectors to store paths and multipaths
*/
- curmp = vector_alloc ();
- pathvec = vector_alloc ();
+ *curmp_p = vector_alloc ();
+ *pathvec_p = vector_alloc ();
- if (!curmp || !pathvec){
+ if (!*curmp_p || !*pathvec_p){
condlog (0, "vector allocation failed.");
goto err;
}
- if (dm_get_maps(curmp))
+ if (dm_get_maps(*curmp_p))
goto err;
return MPATH_PR_SUCCESS;
err:
- mpath_persistent_reserve_free_vecs();
+ __mpath_persistent_reserve_free_vecs(*curmp_p, *pathvec_p);
+ *curmp_p = *pathvec_p = NULL;
return MPATH_PR_DMMP_ERROR;
}
-static int mpath_get_map(int fd, char **palias, struct multipath **pmpp)
+int mpath_persistent_reserve_init_vecs(int verbose)
+{
+ return __mpath_persistent_reserve_init_vecs(&curmp, &pathvec, verbose);
+}
+
+static int mpath_get_map(vector curmp, vector pathvec, int fd, char **palias,
+ struct multipath **pmpp)
{
int ret = MPATH_PR_DMMP_ERROR;
struct stat info;
@@ -272,13 +260,13 @@ out:
return ret;
}
-int __mpath_persistent_reserve_in (int fd, int rq_servact,
- struct prin_resp *resp, int noisy)
+static int do_mpath_persistent_reserve_in (vector curmp, vector pathvec,
+ int fd, int rq_servact, struct prin_resp *resp, int noisy)
{
struct multipath *mpp;
int ret;
- ret = mpath_get_map(fd, NULL, &mpp);
+ ret = mpath_get_map(curmp, pathvec, fd, NULL, &mpp);
if (ret != MPATH_PR_SUCCESS)
return ret;
@@ -287,8 +275,17 @@ int __mpath_persistent_reserve_in (int fd, int rq_servact,
return ret;
}
-int __mpath_persistent_reserve_out ( int fd, int rq_servact, int rq_scope,
- unsigned int rq_type, struct prout_param_descriptor *paramp, int noisy)
+
+int __mpath_persistent_reserve_in (int fd, int rq_servact,
+ struct prin_resp *resp, int noisy)
+{
+ return do_mpath_persistent_reserve_in(curmp, pathvec, fd, rq_servact,
+ resp, noisy);
+}
+
+static int do_mpath_persistent_reserve_out(vector curmp, vector pathvec, int fd,
+ int rq_servact, int rq_scope, unsigned int rq_type,
+ struct prout_param_descriptor *paramp, int noisy)
{
struct multipath *mpp;
char *alias;
@@ -296,7 +293,7 @@ int __mpath_persistent_reserve_out ( int fd, int rq_servact, int rq_scope,
uint64_t prkey;
struct config *conf;
- ret = mpath_get_map(fd, &alias, &mpp);
+ ret = mpath_get_map(curmp, pathvec, fd, &alias, &mpp);
if (ret != MPATH_PR_SUCCESS)
return ret;
@@ -366,6 +363,45 @@ out1:
return ret;
}
+
+int __mpath_persistent_reserve_out ( int fd, int rq_servact, int rq_scope,
+ unsigned int rq_type, struct prout_param_descriptor *paramp, int noisy)
+{
+ return do_mpath_persistent_reserve_out(curmp, pathvec, fd, rq_servact,
+ rq_scope, rq_type, paramp,
+ noisy);
+}
+
+int mpath_persistent_reserve_in (int fd, int rq_servact,
+ struct prin_resp *resp, int noisy, int verbose)
+{
+ vector curmp = NULL, pathvec;
+ int ret = __mpath_persistent_reserve_init_vecs(&curmp, &pathvec,
+ verbose);
+
+ if (ret != MPATH_PR_SUCCESS)
+ return ret;
+ ret = do_mpath_persistent_reserve_in(curmp, pathvec, fd, rq_servact,
+ resp, noisy);
+ __mpath_persistent_reserve_free_vecs(curmp, pathvec);
+ return ret;
+}
+
+int mpath_persistent_reserve_out ( int fd, int rq_servact, int rq_scope,
+ unsigned int rq_type, struct prout_param_descriptor *paramp, int noisy, int verbose)
+{
+ vector curmp = NULL, pathvec;
+ int ret = __mpath_persistent_reserve_init_vecs(&curmp, &pathvec,
+ verbose);
+
+ if (ret != MPATH_PR_SUCCESS)
+ return ret;
+ ret = do_mpath_persistent_reserve_out(curmp, pathvec, fd, rq_servact,
+ rq_scope, rq_type, paramp, noisy);
+ __mpath_persistent_reserve_free_vecs(curmp, pathvec);
+ return ret;
+}
+
int
get_mpvec (vector curmp, vector pathvec, char * refwwid)
{
diff --git a/libmpathpersist/mpath_persist.h b/libmpathpersist/mpath_persist.h
index 7cf4faf9..0e4e0e53 100644
--- a/libmpathpersist/mpath_persist.h
+++ b/libmpathpersist/mpath_persist.h
@@ -215,9 +215,13 @@ extern int mpath_persistent_reserve_in (int fd, int rq_servact, struct prin_resp
/*
* DESCRIPTION :
- * This function is like mpath_persistent_reserve_in(), except that it doesn't call
- * mpath_persistent_reserve_init_vecs() and mpath_persistent_reserve_free_vecs()
- * before and after the actual PR call.
+ * This function is like mpath_persistent_reserve_in(), except that it
+ * requires mpath_persistent_reserve_init_vecs() to be called before the
+ * PR call to set up internal variables. These must later be cleanup up
+ * by calling mpath_persistent_reserve_free_vecs().
+ *
+ * RESTRICTIONS:
+ * This function uses static internal variables, and is not thread-safe.
*/
extern int __mpath_persistent_reserve_in(int fd, int rq_servact,
struct prin_resp *resp, int noisy);
@@ -249,9 +253,13 @@ extern int mpath_persistent_reserve_out ( int fd, int rq_servact, int rq_scope,
int verbose);
/*
* DESCRIPTION :
- * This function is like mpath_persistent_reserve_out(), except that it doesn't call
- * mpath_persistent_reserve_init_vecs() and mpath_persistent_reserve_free_vecs()
- * before and after the actual PR call.
+ * This function is like mpath_persistent_reserve_out(), except that it
+ * requires mpath_persistent_reserve_init_vecs() to be called before the
+ * PR call to set up internal variables. These must later be cleanup up
+ * by calling mpath_persistent_reserve_free_vecs().
+ *
+ * RESTRICTIONS:
+ * This function uses static internal variables, and is not thread-safe.
*/
extern int __mpath_persistent_reserve_out( int fd, int rq_servact, int rq_scope,
unsigned int rq_type, struct prout_param_descriptor *paramp,
@@ -265,6 +273,7 @@ extern int __mpath_persistent_reserve_out( int fd, int rq_servact, int rq_scope,
* @verbose: Set verbosity level. Input argument. value:0 to 3. 0->disabled, 3->Max verbose
*
* RESTRICTIONS:
+ * This function uses static internal variables, and is not thread-safe.
*
* RETURNS: MPATH_PR_SUCCESS if successful else returns any of the status specified
* above in RETURN_STATUS.
@@ -275,6 +284,9 @@ int mpath_persistent_reserve_init_vecs(int verbose);
* DESCRIPTION :
* This function frees data structures allocated by
* mpath_persistent_reserve_init_vecs().
+ *
+ * RESTRICTIONS:
+ * This function uses static internal variables, and is not thread-safe.
*/
void mpath_persistent_reserve_free_vecs(void);
--
2.17.2

61
SOURCES/0060-kpartx-free-loop-device-after-listing-partitions.patch Normal file
View File

@ -0,0 +1,61 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue, 9 Feb 2021 17:16:04 -0600
Subject: [PATCH] kpartx: free loop device after listing partitions
If "kpartx -l" is run on a file that doesn't already have a loop device
associated with it, it will create a loop device to run the command.
Starting with da59d15c6 ("Fix loopback file with kpartx -av"), it will
not free the loop device when exitting. This is because it checks if the
the file it stat()ed is a regular file, before freeing the loop device.
However, after da59d15c6, stat() is rerun on the loop device itself, so
the check fails. There is no need to check this, if loopcreated is
true, then the file will be a kpartx created loop device, and should be
freed.
Also, keep kpartx from printing that the loop device has been removed
at normal verbosity.
Fixes: da59d15c6 ("Fix loopback file with kpartx -av")
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
kpartx/kpartx.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/kpartx/kpartx.c b/kpartx/kpartx.c
index 653ce0c8..a337a07b 100644
--- a/kpartx/kpartx.c
+++ b/kpartx/kpartx.c
@@ -407,7 +407,7 @@ main(int argc, char **argv){
fprintf(stderr, "can't del loop : %s\n",
loopdev);
r = 1;
- } else
+ } else if (verbose)
fprintf(stderr, "loop deleted : %s\n", loopdev);
}
goto end;
@@ -649,16 +649,17 @@ main(int argc, char **argv){
if (n > 0)
break;
}
- if (what == LIST && loopcreated && S_ISREG (buf.st_mode)) {
+ if (what == LIST && loopcreated) {
if (fd != -1)
close(fd);
if (del_loop(device)) {
if (verbose)
- printf("can't del loop : %s\n",
+ fprintf(stderr, "can't del loop : %s\n",
device);
exit(1);
}
- printf("loop deleted : %s\n", device);
+ if (verbose)
+ fprintf(stderr, "loop deleted : %s\n", device);
}
end:
--
2.17.2

134
SOURCES/0061-RH-fix-find_multipaths-in-mpathconf.patch Normal file
View File

@ -0,0 +1,134 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Wed, 10 Feb 2021 15:42:42 -0600
Subject: [PATCH] RH: fix find_multipaths in mpathconf
mpathconf wasn't correctly dealing with the new rhel-8 values for
find_multipaths
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
multipath/mpathconf | 38 +++++++++++++++++++-------------------
multipath/mpathconf.8 | 14 +++++++-------
2 files changed, 26 insertions(+), 26 deletions(-)
diff --git a/multipath/mpathconf b/multipath/mpathconf
index f34003c9..2f4f3eaf 100644
--- a/multipath/mpathconf
+++ b/multipath/mpathconf
@@ -54,7 +54,7 @@ function usage
echo "Disable: --disable"
echo "Only allow certain wwids (instead of enable): --allow <WWID>"
echo "Set user_friendly_names (Default y): --user_friendly_names <y|n>"
- echo "Set find_multipaths (Default y): --find_multipaths <y|n>"
+ echo "Set find_multipaths (Default y): --find_multipaths <yes|no|strict|greedy|smart>"
echo "Set default property blacklist (Default y): --property_blacklist <y|n>"
echo "Set enable_foreign to show foreign devices (Default n): --enable_foreign <y|n>"
echo "Load the dm-multipath modules on enable (Default y): --with_module <y|n>"
@@ -224,8 +224,12 @@ function validate_args
echo "--user_friendly_names must be either 'y' or 'n'"
exit 1
fi
- if [ -n "$FIND" ] && [ "$FIND" != "y" -a "$FIND" != "n" ]; then
- echo "--find_multipaths must be either 'y' or 'n'"
+ if [ "$FIND" = "y" ]; then
+ FIND="yes"
+ elif [ "$FIND" = "n" ]; then
+ FIND="no"
+ elif [ -n "$FIND" ] && [ "$FIND" != "yes" -a "$FIND" != "no" -a "$FIND" != "strict" -a "$FIND" != "greedy" -a "$FIND" != "smart" ]; then
+ echo "--find_multipaths must be one of 'yes' 'no' 'strict' 'greedy' or 'smart'"
exit 1
fi
if [ -n "$PROPERTY" ] && [ "$PROPERTY" != "y" -a "$PROPERTY" != "n" ]; then
@@ -327,10 +331,11 @@ if [ "$HAVE_BLACKLIST" = "1" ]; then
fi
if [ "$HAVE_DEFAULTS" = "1" ]; then
- if sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*find_multipaths[[:space:]]*\(yes\|1\)" ; then
- HAVE_FIND=1
- elif sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*find_multipaths[[:space:]]*\(no\|0\)" ; then
- HAVE_FIND=0
+ HAVE_FIND=`sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | sed -n 's/^[[:blank:]]*find_multipaths[[:blank:]]*\([^[:blank:]]*\).*$/\1/p' | sed -n 1p`
+ if [ "$HAVE_FIND" = "1" ]; then
+ HAVE_FIND="yes"
+ elif [ "$HAVE_FIND" = "0" ]; then
+ HAVE_FIND="no"
fi
if sed -n '/^defaults[[:space:]]*{/,/^}/ p' $TMPFILE | grep -q "^[[:space:]]*user_friendly_names[[:space:]]*\(yes\|1\)" ; then
HAVE_FRIENDLY=1
@@ -360,10 +365,10 @@ if [ -n "$SHOW_STATUS" ]; then
else
echo "multipath is disabled"
fi
- if [ -z "$HAVE_FIND" -o "$HAVE_FIND" = 0 ]; then
- echo "find_multipaths is disabled"
+ if [ -z "$HAVE_FIND" ]; then
+ echo "find_multipaths is no"
else
- echo "find_multipaths is enabled"
+ echo "find_multipaths is $HAVE_FIND"
fi
if [ -z "$HAVE_FRIENDLY" -o "$HAVE_FRIENDLY" = 0 ]; then
echo "user_friendly_names is disabled"
@@ -455,19 +460,14 @@ elif [ "$ENABLE" = 0 ]; then
fi
fi
-if [ "$FIND" = "n" ]; then
- if [ "$HAVE_FIND" = 1 ]; then
- sed -i '/^defaults[[:space:]]*{/,/^}/ s/^[[:space:]]*find_multipaths[[:space:]]*\(yes\|1\)/ find_multipaths no/' $TMPFILE
- CHANGED_CONFIG=1
- fi
-elif [ "$FIND" = "y" ]; then
+if [ -n "$FIND" ]; then
if [ -z "$HAVE_FIND" ]; then
sed -i '/^defaults[[:space:]]*{/ a\
- find_multipaths yes
+ find_multipaths '"$FIND"'
' $TMPFILE
CHANGED_CONFIG=1
- elif [ "$HAVE_FIND" = 0 ]; then
- sed -i '/^defaults[[:space:]]*{/,/^}/ s/^[[:space:]]*find_multipaths[[:space:]]*\(no\|0\)/ find_multipaths yes/' $TMPFILE
+ elif [ "$FIND" != "$HAVE_FIND" ]; then
+ sed -i '/^defaults[[:space:]]*{/,/^}/ s/^[[:blank:]]*find_multipaths[[:blank:]]*[^[:blank:]]*/ find_multipaths '"$FIND"'/' $TMPFILE
CHANGED_CONFIG=1
fi
fi
diff --git a/multipath/mpathconf.8 b/multipath/mpathconf.8
index b82961d6..83515eb4 100644
--- a/multipath/mpathconf.8
+++ b/multipath/mpathconf.8
@@ -38,9 +38,9 @@ If
already exists, mpathconf will edit it. If it does not exist, mpathconf will
create a default file with
.B user_friendly_names
-and
+set and
.B find_multipaths
-set. To disable these, use the
+set to \fByes\fP. To disable these, use the
.B --user_friendly_names n
and
.B --find_multipaths n
@@ -77,13 +77,13 @@ to the
defaults section. If set to \fBn\fP, this removes the line, if present. This
command can be used along with any other command.
.TP
-.B --find_multipaths\fP { \fBy\fP | \fBn\fP }
-If set to \fBy\fP, this adds the line
-.B find_multipaths yes
+.B --find_multipaths\fP { \fByes\fP | \fBno\fP | \fBstrict\fP | \fBgreedy\fP | \fBsmart\fP }
+If set to \fB<value>\fP, this adds the line
+.B find_multipaths <value>
to the
.B /etc/multipath.conf
-defaults section. If set to \fBn\fP, this removes the line, if present. This
-command can be used along with any other command.
+defaults section. This command can be used along with any other command.
+\fBy\fP and \fBn\fP can be used instead of \fByes\fP and \fBno\fP.
.TP
.B --property_blacklist \fP { \fBy\fP | \fBn\fP }
If set to \fBy\fP, this adds the line
--
2.17.2

45
SOURCES/0062-libmultipath-select_action-don-t-drop-map-if-alias-c.patch Normal file
View File

@ -0,0 +1,45 @@
From 7a7b96246b84ccf533a6f4dc0424830792fdb96a Mon Sep 17 00:00:00 2001
From: Martin Wilck <mwilck@suse.com>
Date: Fri, 3 Jul 2020 15:17:09 +0200
Subject: [PATCH] libmultipath: select_action(): don't drop map if alias
clashes
If for a given map, if we find that the requested alias is already
used by a map with different WWID, while the map's own WWID is
not used yet, give up the alias and use the WWID instead. This
is safer than trying to destroy the existing map, which is likely
to fail.
This allows us to make use const for the "curmp" parameter.
Reviewed-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/configure.c | 11 ++++++-----
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/libmultipath/configure.c b/libmultipath/configure.c
index b7113291..2e8f34f9 100644
--- a/libmultipath/configure.c
+++ b/libmultipath/configure.c
@@ -690,12 +690,13 @@ select_action (struct multipath * mpp, vector curmp, int force_reload)
}
if (!cmpp) {
- condlog(2, "%s: remove (wwid changed)", mpp->alias);
- dm_flush_map(mpp->alias);
- strlcpy(cmpp_by_name->wwid, mpp->wwid, WWID_SIZE);
- drop_multipath(curmp, cmpp_by_name->wwid, KEEP_PATHS);
+ condlog(1, "%s: can't use alias \"%s\" used by %s, falling back to WWID",
+ mpp->wwid, mpp->alias, cmpp_by_name->wwid);
+ /* We can do this because wwid wasn't found */
+ free(mpp->alias);
+ mpp->alias = strdup(mpp->wwid);
mpp->action = ACT_CREATE;
- condlog(3, "%s: set ACT_CREATE (map wwid change)",
+ condlog(3, "%s: set ACT_CREATE (map does not exist, name changed)",
mpp->alias);
return;
}
--
2.17.2

231
SOURCES/0063-libmultipath-check-if-user_friendly_name-is-in-use.patch Normal file
View File

@ -0,0 +1,231 @@
From e714eb26fddc8768a8de279d1de3ffedab35929e Mon Sep 17 00:00:00 2001
From: Benjamin Marzinski <bmarzins@redhat.com>
Date: Fri, 5 Mar 2021 21:40:57 -0600
Subject: [PATCH] libmultipath: check if user_friendly_name is in use
If there are multipath devices that have user_friendly_names but do not
have their bindings in the bindings_file, get_user_friendly_alias() can
currently give out those names again. This can result in an incorrect
entry in the bindings file, and a device that gets created with a WWID
alias instead of a user_friendly_name. This situation can happen after
the pivot root, if a multipath device is created in the initramfs. If
this device doesn't have a binding in the regular filesystem
bindings_file and a new multipath device is created before it can add
its binding, the new device can steal that user_friendly_name during
multipathd's initial configure.
To solve this, get_user_friendly_alias() now calls lookup_binding() with
a new paramter, telling it to check if the id it found is already in use
by a diffent device. If so, lookup_binding() will continue to check open
ids, until it finds one that it not currently in use by a dm device.
---
libmultipath/alias.c | 48 +++++++++++++++++++++++++++++++++++++++++---
tests/alias.c | 22 ++++++++++----------
2 files changed, 56 insertions(+), 14 deletions(-)
diff --git a/libmultipath/alias.c b/libmultipath/alias.c
index 14401cae..01f737f4 100644
--- a/libmultipath/alias.c
+++ b/libmultipath/alias.c
@@ -17,6 +17,7 @@
#include "vector.h"
#include "checkers.h"
#include "structs.h"
+#include "devmapper.h"
/*
@@ -104,6 +105,28 @@ scan_devname(const char *alias, const char *prefix)
return n;
}
+static int
+id_already_taken(int id, const char *prefix, const char *map_wwid)
+{
+ char alias[LINE_MAX];
+
+ if (format_devname(alias, id, LINE_MAX, prefix) < 0)
+ return 0;
+
+ if (dm_map_present(alias)) {
+ char wwid[WWID_SIZE];
+
+ /* If both the name and the wwid match, then it's fine.*/
+ if (dm_get_uuid(alias, wwid, sizeof(wwid)) == 0 &&
+ strncmp(map_wwid, wwid, sizeof(wwid)) == 0)
+ return 0;
+ condlog(3, "%s: alias '%s' already taken, but not in bindings file. reselecting alias", map_wwid, alias);
+ return 1;
+ }
+ return 0;
+}
+
+
/*
* Returns: 0 if matching entry in WWIDs file found
* -1 if an error occurs
@@ -113,7 +136,7 @@ scan_devname(const char *alias, const char *prefix)
*/
static int
lookup_binding(FILE *f, const char *map_wwid, char **map_alias,
- const char *prefix)
+ const char *prefix, int check_if_taken)
{
char buf[LINE_MAX];
unsigned int line_nr = 0;
@@ -168,12 +191,31 @@ lookup_binding(FILE *f, const char *map_wwid, char **map_alias,
return 0;
}
}
+ if (!prefix && check_if_taken)
+ id = -1;
if (id >= smallest_bigger_id) {
if (biggest_id < INT_MAX)
id = biggest_id + 1;
else
id = -1;
}
+ if (id > 0 && check_if_taken) {
+ while(id_already_taken(id, prefix, map_wwid)) {
+ if (id == INT_MAX) {
+ id = -1;
+ break;
+ }
+ id++;
+ if (id == smallest_bigger_id) {
+ if (biggest_id == INT_MAX) {
+ id = -1;
+ break;
+ }
+ if (biggest_id >= smallest_bigger_id)
+ id = biggest_id + 1;
+ }
+ }
+ }
if (id < 0) {
condlog(0, "no more available user_friendly_names");
return -1;
@@ -316,7 +358,7 @@ use_existing_alias (const char *wwid, const char *file, const char *alias_old,
goto out;
}
- id = lookup_binding(f, wwid, &alias, NULL);
+ id = lookup_binding(f, wwid, &alias, NULL, 0);
if (alias) {
condlog(3, "Use existing binding [%s] for WWID [%s]",
alias, wwid);
@@ -373,7 +415,7 @@ get_user_friendly_alias(const char *wwid, const char *file, const char *prefix,
return NULL;
}
- id = lookup_binding(f, wwid, &alias, prefix);
+ id = lookup_binding(f, wwid, &alias, prefix, 1);
if (id < 0) {
fclose(f);
return NULL;
diff --git a/tests/alias.c b/tests/alias.c
index 30414db0..ab1a9325 100644
--- a/tests/alias.c
+++ b/tests/alias.c
@@ -356,7 +356,7 @@ static void lb_empty(void **state)
will_return(__wrap_fgets, NULL);
expect_condlog(3, "No matching wwid [WWID0] in bindings file.\n");
- rc = lookup_binding(NULL, "WWID0", &alias, NULL);
+ rc = lookup_binding(NULL, "WWID0", &alias, NULL, 0);
assert_int_equal(rc, 1);
assert_ptr_equal(alias, NULL);
}
@@ -369,7 +369,7 @@ static void lb_match_a(void **state)
will_return(__wrap_fgets, "MPATHa WWID0\n");
expect_condlog(3, "Found matching wwid [WWID0] in bindings file."
" Setting alias to MPATHa\n");
- rc = lookup_binding(NULL, "WWID0", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID0", &alias, "MPATH", 0);
assert_int_equal(rc, 0);
assert_ptr_not_equal(alias, NULL);
assert_string_equal(alias, "MPATHa");
@@ -384,7 +384,7 @@ static void lb_nomatch_a(void **state)
will_return(__wrap_fgets, "MPATHa WWID0\n");
will_return(__wrap_fgets, NULL);
expect_condlog(3, "No matching wwid [WWID1] in bindings file.\n");
- rc = lookup_binding(NULL, "WWID1", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID1", &alias, "MPATH", 0);
assert_int_equal(rc, 2);
assert_ptr_equal(alias, NULL);
}
@@ -398,7 +398,7 @@ static void lb_match_c(void **state)
will_return(__wrap_fgets, "MPATHc WWID1\n");
expect_condlog(3, "Found matching wwid [WWID1] in bindings file."
" Setting alias to MPATHc\n");
- rc = lookup_binding(NULL, "WWID1", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID1", &alias, "MPATH", 0);
assert_int_equal(rc, 0);
assert_ptr_not_equal(alias, NULL);
assert_string_equal(alias, "MPATHc");
@@ -414,7 +414,7 @@ static void lb_nomatch_a_c(void **state)
will_return(__wrap_fgets, "MPATHc WWID1\n");
will_return(__wrap_fgets, NULL);
expect_condlog(3, "No matching wwid [WWID2] in bindings file.\n");
- rc = lookup_binding(NULL, "WWID2", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID2", &alias, "MPATH", 0);
assert_int_equal(rc, 2);
assert_ptr_equal(alias, NULL);
}
@@ -428,7 +428,7 @@ static void lb_nomatch_c_a(void **state)
will_return(__wrap_fgets, "MPATHa WWID0\n");
will_return(__wrap_fgets, NULL);
expect_condlog(3, "No matching wwid [WWID2] in bindings file.\n");
- rc = lookup_binding(NULL, "WWID2", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID2", &alias, "MPATH", 0);
assert_int_equal(rc, 2);
assert_ptr_equal(alias, NULL);
}
@@ -443,7 +443,7 @@ static void lb_nomatch_a_b(void **state)
will_return(__wrap_fgets, "MPATHb WWID1\n");
will_return(__wrap_fgets, NULL);
expect_condlog(3, "No matching wwid [WWID2] in bindings file.\n");
- rc = lookup_binding(NULL, "WWID2", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID2", &alias, "MPATH", 0);
assert_int_equal(rc, 3);
assert_ptr_equal(alias, NULL);
}
@@ -459,7 +459,7 @@ static void lb_nomatch_a_b_bad(void **state)
will_return(__wrap_fgets, NULL);
expect_condlog(3, "Ignoring malformed line 3 in bindings file\n");
expect_condlog(3, "No matching wwid [WWID2] in bindings file.\n");
- rc = lookup_binding(NULL, "WWID2", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID2", &alias, "MPATH", 0);
assert_int_equal(rc, 3);
assert_ptr_equal(alias, NULL);
}
@@ -474,7 +474,7 @@ static void lb_nomatch_b_a(void **state)
will_return(__wrap_fgets, "MPATHa WWID0\n");
will_return(__wrap_fgets, NULL);
expect_condlog(3, "No matching wwid [WWID2] in bindings file.\n");
- rc = lookup_binding(NULL, "WWID2", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID2", &alias, "MPATH", 0);
assert_int_equal(rc, 27);
assert_ptr_equal(alias, NULL);
}
@@ -490,7 +490,7 @@ static void lb_nomatch_int_max(void **state)
will_return(__wrap_fgets, "MPATHa WWID0\n");
will_return(__wrap_fgets, NULL);
expect_condlog(0, "no more available user_friendly_names\n");
- rc = lookup_binding(NULL, "WWID2", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID2", &alias, "MPATH", 0);
assert_int_equal(rc, -1);
assert_ptr_equal(alias, NULL);
}
@@ -505,7 +505,7 @@ static void lb_nomatch_int_max_m1(void **state)
will_return(__wrap_fgets, "MPATHa WWID0\n");
will_return(__wrap_fgets, NULL);
expect_condlog(3, "No matching wwid [WWID2] in bindings file.\n");
- rc = lookup_binding(NULL, "WWID2", &alias, "MPATH");
+ rc = lookup_binding(NULL, "WWID2", &alias, "MPATH", 0);
assert_int_equal(rc, INT_MAX);
assert_ptr_equal(alias, NULL);
}
--
2.17.2

93
SOURCES/multipath.conf Normal file
View File

@ -0,0 +1,93 @@
# This is a basic configuration file with some examples, for device mapper
# multipath.
#
# For a complete list of the default configuration values, run either
# multipath -t
# or
# multipathd show config
#
# For a list of configuration options with descriptions, see the multipath.conf
# man page
## By default, devices with vendor = "IBM" and product = "S/390.*" are
## blacklisted. To enable mulitpathing on these devies, uncomment the
## following lines.
#blacklist_exceptions {
# device {
# vendor "IBM"
# product "S/390.*"
# }
#}
## Use user friendly names, instead of using WWIDs as names.
defaults {
user_friendly_names yes
find_multipaths yes
}
##
## Here is an example of how to configure some standard options.
##
#
#defaults {
# udev_dir /dev
# polling_interval 10
# selector "round-robin 0"
# path_grouping_policy multibus
# prio alua
# path_checker readsector0
# rr_min_io 100
# max_fds 8192
# rr_weight priorities
# failback immediate
# no_path_retry fail
# user_friendly_names yes
#}
##
## The wwid line in the following blacklist section is shown as an example
## of how to blacklist devices by wwid. The 2 devnode lines are the
## compiled in default blacklist. If you want to blacklist entire types
## of devices, such as all scsi devices, you should use a devnode line.
## However, if you want to blacklist specific devices, you should use
## a wwid line. Since there is no guarantee that a specific device will
## not change names on reboot (from /dev/sda to /dev/sdb for example)
## devnode lines are not recommended for blacklisting specific devices.
##
#blacklist {
# wwid 26353900f02796769
# devnode "^(ram|raw|loop|fd|md|dm-|sr|scd|st)[0-9]*"
# devnode "^hd[a-z]"
#}
#multipaths {
# multipath {
# wwid 3600508b4000156d700012000000b0000
# alias yellow
# path_grouping_policy multibus
# path_checker readsector0
# path_selector "round-robin 0"
# failback manual
# rr_weight priorities
# no_path_retry 5
# }
# multipath {
# wwid 1DEC_____321816758474
# alias red
# }
#}
#devices {
# device {
# vendor "COMPAQ "
# product "HSV110 (C)COMPAQ"
# path_grouping_policy multibus
# path_checker readsector0
# path_selector "round-robin 0"
# hardware_handler "0"
# failback 15
# rr_weight priorities
# no_path_retry queue
# }
# device {
# vendor "COMPAQ "
# product "MSA1000 "
# path_grouping_policy multibus
# }
#}

1884
SPECS/device-mapper-multipath.spec Normal file
View File

File diff suppressed because it is too large Load Diff