Re: [infiniband-diags] [UPDATED PATCH] [3/3] support --load-cache in iblinkinfo and ibqueryerrors
Hey Sasha,
Here's an updated patch with the cleanup changes as we discussed.
Al
On Sat, 2010-01-16 at 16:28 +0200, Sasha Khapyorsky wrote:
On 10:23 Fri 15 Jan , Al Chu wrote:
Hi Sasha,
This adds the --load-cache options to iblinkinfo and ibqueryerrors.
Al
--
Albert Chu
ch...@llnl.gov
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory
From: Albert Chu ch...@llnl.gov
Date: Thu, 10 Dec 2009 11:22:50 -0800
Subject: [PATCH] support --load-cache in iblinkinfo and ibqueryerrors
Signed-off-by: Albert Chu ch...@llnl.gov
---
infiniband-diags/man/iblinkinfo.8| 11 ++-
infiniband-diags/man/ibqueryerrors.8 | 10 ++-
infiniband-diags/src/iblinkinfo.c| 52
+---
infiniband-diags/src/ibqueryerrors.c | 53
++---
4 files changed, 99 insertions(+), 27 deletions(-)
diff --git a/infiniband-diags/man/iblinkinfo.8
b/infiniband-diags/man/iblinkinfo.8
index 0f53b00..f184edf 100644
--- a/infiniband-diags/man/iblinkinfo.8
+++ b/infiniband-diags/man/iblinkinfo.8
@@ -6,7 +6,7 @@ iblinkinfo \- report link info for all links in the fabric
.SH SYNOPSIS
.B iblinkinfo
[-hcdl -C ca_name -P ca_port -v lt,hoq,vlstall -S guid
--D direct_route]
+-D direct_route \-\-load\-cache filename]
.SH DESCRIPTION
.PP
@@ -42,7 +42,14 @@ Print port capabilities (enabled and supported values)
\fB\-P ca_port\fRuse the specified ca_port for the search.
.TP
\fB\-R\fR (This option is obsolete and does nothing)
-
+.TP
+\fB\-\-load\-cache\fR filename
+Load and use the cached ibnetdiscover data stored in the specified
+filename. May be useful for outputting and learning about other
+fabrics or a previous state of a fabric. Cannot be used if user
+specifies a directo route path. See
+.B ibnetdiscover
+for information on caching ibnetdiscover output.
.SH AUTHOR
.TP
diff --git a/infiniband-diags/man/ibqueryerrors.8
b/infiniband-diags/man/ibqueryerrors.8
index 83a2b5a..56a0d67 100644
--- a/infiniband-diags/man/ibqueryerrors.8
+++ b/infiniband-diags/man/ibqueryerrors.8
@@ -6,7 +6,7 @@ ibqueryerrors \- query and report non-zero IB port counters
.SH SYNOPSIS
.B ibqueryerrors
[-s err1,err2,... -c -r -C ca_name -P ca_port -G node_guid
--D direct_route -d -k -K]
+-D direct_route -d -k -K \-\-load\-cache filename]
.SH DESCRIPTION
.PP
@@ -60,6 +60,14 @@ specified the data counters will be cleared without any
printed output.
.TP
\fB\-\-details\fR include transmit discard details
.TP
+\fB\-\-load\-cache\fR filename
+Load and use the cached ibnetdiscover data stored in the specified
+filename. May be useful for outputting and learning about other
+fabrics or a previous state of a fabric. Cannot be used if user
+specifies a directo route path. See
+.B ibnetdiscover
+for information on caching ibnetdiscover output.
+.TP
\fB\-R\fR (This option is obsolete and does nothing)
.SH COMMON OPTIONS
diff --git a/infiniband-diags/src/iblinkinfo.c
b/infiniband-diags/src/iblinkinfo.c
index 21b31bb..10e3ad5 100644
--- a/infiniband-diags/src/iblinkinfo.c
+++ b/infiniband-diags/src/iblinkinfo.c
@@ -55,6 +55,7 @@
static char *node_name_map_file = NULL;
static nn_map_t *node_name_map = NULL;
+static char *load_cache_file = NULL;
static uint64_t guid = 0;
static char *guid_str = NULL;
@@ -230,6 +231,9 @@ static int process_opt(void *context, int ch, char
*optarg)
case 1:
node_name_map_file = strdup(optarg);
break;
+ case 2:
+ load_cache_file = strdup(optarg);
+ break;
case 'S':
guid_str = optarg;
guid = (uint64_t) strtoull(guid_str, 0, 0);
@@ -291,6 +295,7 @@ int main(int argc, char **argv)
print additional switch settings (PktLifeTime, HoqLife,
VLStallCount)},
{portguids, 'g', 0, NULL,
print port guids instead of node guids},
+ {load-cache, 2, 1, file, filename of ibnetdiscover cache
to load},
{GNDN, 'R', 0, NULL,
(This option is obsolete and does nothing)},
{0}
@@ -317,6 +322,11 @@ int main(int argc, char **argv)
mad_rpc_set_timeout(ibmad_port, ibd_timeout);
node_name_map = open_node_name_map(node_name_map_file);
+
+ if (dr_path load_cache_file) {
+ fprintf(stderr, Cannot specify cache and direct route path\n);
+ exit(1);
+ }
Why is this limitation needed really?
if (dr_path) {
/* only scan part of the fabric */
@@ -334,19 +344,37 @@ int main(int argc, char **argv)
guid_str);
}
- if (resolved = 0)
- if ((fabric = ibnd_discover_fabric(ibmad_port, port_id,
-
Re: [infiniband-diags] [UPDATED PATCH] [3/3] support --load-cache in iblinkinfo and ibqueryerrors
On 10:23 Fri 15 Jan , Al Chu wrote:
Hi Sasha,
This adds the --load-cache options to iblinkinfo and ibqueryerrors.
Al
--
Albert Chu
ch...@llnl.gov
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory
From: Albert Chu ch...@llnl.gov
Date: Thu, 10 Dec 2009 11:22:50 -0800
Subject: [PATCH] support --load-cache in iblinkinfo and ibqueryerrors
Signed-off-by: Albert Chu ch...@llnl.gov
---
infiniband-diags/man/iblinkinfo.8| 11 ++-
infiniband-diags/man/ibqueryerrors.8 | 10 ++-
infiniband-diags/src/iblinkinfo.c| 52 +---
infiniband-diags/src/ibqueryerrors.c | 53 ++---
4 files changed, 99 insertions(+), 27 deletions(-)
diff --git a/infiniband-diags/man/iblinkinfo.8
b/infiniband-diags/man/iblinkinfo.8
index 0f53b00..f184edf 100644
--- a/infiniband-diags/man/iblinkinfo.8
+++ b/infiniband-diags/man/iblinkinfo.8
@@ -6,7 +6,7 @@ iblinkinfo \- report link info for all links in the fabric
.SH SYNOPSIS
.B iblinkinfo
[-hcdl -C ca_name -P ca_port -v lt,hoq,vlstall -S guid
--D direct_route]
+-D direct_route \-\-load\-cache filename]
.SH DESCRIPTION
.PP
@@ -42,7 +42,14 @@ Print port capabilities (enabled and supported values)
\fB\-P ca_port\fRuse the specified ca_port for the search.
.TP
\fB\-R\fR (This option is obsolete and does nothing)
-
+.TP
+\fB\-\-load\-cache\fR filename
+Load and use the cached ibnetdiscover data stored in the specified
+filename. May be useful for outputting and learning about other
+fabrics or a previous state of a fabric. Cannot be used if user
+specifies a directo route path. See
+.B ibnetdiscover
+for information on caching ibnetdiscover output.
.SH AUTHOR
.TP
diff --git a/infiniband-diags/man/ibqueryerrors.8
b/infiniband-diags/man/ibqueryerrors.8
index 83a2b5a..56a0d67 100644
--- a/infiniband-diags/man/ibqueryerrors.8
+++ b/infiniband-diags/man/ibqueryerrors.8
@@ -6,7 +6,7 @@ ibqueryerrors \- query and report non-zero IB port counters
.SH SYNOPSIS
.B ibqueryerrors
[-s err1,err2,... -c -r -C ca_name -P ca_port -G node_guid
--D direct_route -d -k -K]
+-D direct_route -d -k -K \-\-load\-cache filename]
.SH DESCRIPTION
.PP
@@ -60,6 +60,14 @@ specified the data counters will be cleared without any
printed output.
.TP
\fB\-\-details\fR include transmit discard details
.TP
+\fB\-\-load\-cache\fR filename
+Load and use the cached ibnetdiscover data stored in the specified
+filename. May be useful for outputting and learning about other
+fabrics or a previous state of a fabric. Cannot be used if user
+specifies a directo route path. See
+.B ibnetdiscover
+for information on caching ibnetdiscover output.
+.TP
\fB\-R\fR (This option is obsolete and does nothing)
.SH COMMON OPTIONS
diff --git a/infiniband-diags/src/iblinkinfo.c
b/infiniband-diags/src/iblinkinfo.c
index 21b31bb..10e3ad5 100644
--- a/infiniband-diags/src/iblinkinfo.c
+++ b/infiniband-diags/src/iblinkinfo.c
@@ -55,6 +55,7 @@
static char *node_name_map_file = NULL;
static nn_map_t *node_name_map = NULL;
+static char *load_cache_file = NULL;
static uint64_t guid = 0;
static char *guid_str = NULL;
@@ -230,6 +231,9 @@ static int process_opt(void *context, int ch, char
*optarg)
case 1:
node_name_map_file = strdup(optarg);
break;
+ case 2:
+ load_cache_file = strdup(optarg);
+ break;
case 'S':
guid_str = optarg;
guid = (uint64_t) strtoull(guid_str, 0, 0);
@@ -291,6 +295,7 @@ int main(int argc, char **argv)
print additional switch settings (PktLifeTime, HoqLife,
VLStallCount)},
{portguids, 'g', 0, NULL,
print port guids instead of node guids},
+ {load-cache, 2, 1, file, filename of ibnetdiscover cache
to load},
{GNDN, 'R', 0, NULL,
(This option is obsolete and does nothing)},
{0}
@@ -317,6 +322,11 @@ int main(int argc, char **argv)
mad_rpc_set_timeout(ibmad_port, ibd_timeout);
node_name_map = open_node_name_map(node_name_map_file);
+
+ if (dr_path load_cache_file) {
+ fprintf(stderr, Cannot specify cache and direct route path\n);
+ exit(1);
+ }
Why is this limitation needed really?
if (dr_path) {
/* only scan part of the fabric */
@@ -334,19 +344,37 @@ int main(int argc, char **argv)
guid_str);
}
- if (resolved = 0)
- if ((fabric = ibnd_discover_fabric(ibmad_port, port_id,
-hops)) == NULL)
- IBWARN
- (Single node discover failed; attempting full
scan\n);
-
- if (!fabric)
- if ((fabric =
-
Re: [infiniband-diags] [UPDATED PATCH] [3/3] support --load-cache in iblinkinfo and ibqueryerrors
Hey Sasha,
answers inlined below
On Sat, 2010-01-16 at 16:28 +0200, Sasha Khapyorsky wrote:
On 10:23 Fri 15 Jan , Al Chu wrote:
Hi Sasha,
This adds the --load-cache options to iblinkinfo and ibqueryerrors.
Al
--
Albert Chu
ch...@llnl.gov
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory
From: Albert Chu ch...@llnl.gov
Date: Thu, 10 Dec 2009 11:22:50 -0800
Subject: [PATCH] support --load-cache in iblinkinfo and ibqueryerrors
Signed-off-by: Albert Chu ch...@llnl.gov
---
infiniband-diags/man/iblinkinfo.8| 11 ++-
infiniband-diags/man/ibqueryerrors.8 | 10 ++-
infiniband-diags/src/iblinkinfo.c| 52
+---
infiniband-diags/src/ibqueryerrors.c | 53
++---
4 files changed, 99 insertions(+), 27 deletions(-)
diff --git a/infiniband-diags/man/iblinkinfo.8
b/infiniband-diags/man/iblinkinfo.8
index 0f53b00..f184edf 100644
--- a/infiniband-diags/man/iblinkinfo.8
+++ b/infiniband-diags/man/iblinkinfo.8
@@ -6,7 +6,7 @@ iblinkinfo \- report link info for all links in the fabric
.SH SYNOPSIS
.B iblinkinfo
[-hcdl -C ca_name -P ca_port -v lt,hoq,vlstall -S guid
--D direct_route]
+-D direct_route \-\-load\-cache filename]
.SH DESCRIPTION
.PP
@@ -42,7 +42,14 @@ Print port capabilities (enabled and supported values)
\fB\-P ca_port\fRuse the specified ca_port for the search.
.TP
\fB\-R\fR (This option is obsolete and does nothing)
-
+.TP
+\fB\-\-load\-cache\fR filename
+Load and use the cached ibnetdiscover data stored in the specified
+filename. May be useful for outputting and learning about other
+fabrics or a previous state of a fabric. Cannot be used if user
+specifies a directo route path. See
+.B ibnetdiscover
+for information on caching ibnetdiscover output.
.SH AUTHOR
.TP
diff --git a/infiniband-diags/man/ibqueryerrors.8
b/infiniband-diags/man/ibqueryerrors.8
index 83a2b5a..56a0d67 100644
--- a/infiniband-diags/man/ibqueryerrors.8
+++ b/infiniband-diags/man/ibqueryerrors.8
@@ -6,7 +6,7 @@ ibqueryerrors \- query and report non-zero IB port counters
.SH SYNOPSIS
.B ibqueryerrors
[-s err1,err2,... -c -r -C ca_name -P ca_port -G node_guid
--D direct_route -d -k -K]
+-D direct_route -d -k -K \-\-load\-cache filename]
.SH DESCRIPTION
.PP
@@ -60,6 +60,14 @@ specified the data counters will be cleared without any
printed output.
.TP
\fB\-\-details\fR include transmit discard details
.TP
+\fB\-\-load\-cache\fR filename
+Load and use the cached ibnetdiscover data stored in the specified
+filename. May be useful for outputting and learning about other
+fabrics or a previous state of a fabric. Cannot be used if user
+specifies a directo route path. See
+.B ibnetdiscover
+for information on caching ibnetdiscover output.
+.TP
\fB\-R\fR (This option is obsolete and does nothing)
.SH COMMON OPTIONS
diff --git a/infiniband-diags/src/iblinkinfo.c
b/infiniband-diags/src/iblinkinfo.c
index 21b31bb..10e3ad5 100644
--- a/infiniband-diags/src/iblinkinfo.c
+++ b/infiniband-diags/src/iblinkinfo.c
@@ -55,6 +55,7 @@
static char *node_name_map_file = NULL;
static nn_map_t *node_name_map = NULL;
+static char *load_cache_file = NULL;
static uint64_t guid = 0;
static char *guid_str = NULL;
@@ -230,6 +231,9 @@ static int process_opt(void *context, int ch, char
*optarg)
case 1:
node_name_map_file = strdup(optarg);
break;
+ case 2:
+ load_cache_file = strdup(optarg);
+ break;
case 'S':
guid_str = optarg;
guid = (uint64_t) strtoull(guid_str, 0, 0);
@@ -291,6 +295,7 @@ int main(int argc, char **argv)
print additional switch settings (PktLifeTime, HoqLife,
VLStallCount)},
{portguids, 'g', 0, NULL,
print port guids instead of node guids},
+ {load-cache, 2, 1, file, filename of ibnetdiscover cache
to load},
{GNDN, 'R', 0, NULL,
(This option is obsolete and does nothing)},
{0}
@@ -317,6 +322,11 @@ int main(int argc, char **argv)
mad_rpc_set_timeout(ibmad_port, ibd_timeout);
node_name_map = open_node_name_map(node_name_map_file);
+
+ if (dr_path load_cache_file) {
+ fprintf(stderr, Cannot specify cache and direct route path\n);
+ exit(1);
+ }
Why is this limitation needed really?
I spoke to Ira about it awhile ago. I think what we decided was that
while technically you can do a DR path, b/c you can load a cache from
anywhere in the cluster, you won't know if the DR path is legal or
correct at point A vs point B. In contrast, the node_guid input is
valid anywhere you are on the cluster.
I suppose we
