* [PATCH 0/2] Fix for machines without numa
@ 2021-02-22 22:07 John Kacur
2021-02-22 22:07 ` [PATCH 1/2] rt-tests: Don't assume numa is available at runtime John Kacur
` (2 more replies)
0 siblings, 3 replies; 5+ messages in thread
From: John Kacur @ 2021-02-22 22:07 UTC (permalink / raw)
To: RT, Christian Eggers
Cc: Daniel Wagner, Kurt Kanzenbach, Clark Williams,
Sebastian Andrzej Siewior, John Kacur
Trying to return the code to a state where numa is required at compile
time but not at runtime
- The first patch should allow cyclictest to run on a machine
without numa.
- The second patch is a clean-up of some old code
The code is pushed in git to branch unstable/devel/no-numa-runtime
so that people can test before I integrate the code back in the main
branch
Thanks
John
John Kacur (2):
rt-tests: Don't assume numa is available at runtime
rt-tests: remove rt_numa_bitmask_count in rt_numa.h
src/cyclictest/cyclictest.c | 8 ++------
src/cyclictest/rt_numa.h | 12 ------------
src/lib/rt-numa.c | 17 +++++++++++------
src/signaltest/signaltest.c | 8 ++------
4 files changed, 15 insertions(+), 30 deletions(-)
--
2.26.2
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 1/2] rt-tests: Don't assume numa is available at runtime
2021-02-22 22:07 [PATCH 0/2] Fix for machines without numa John Kacur
@ 2021-02-22 22:07 ` John Kacur
2021-02-22 22:07 ` [PATCH 2/2] rt-tests: remove rt_numa_bitmask_count in rt_numa.h John Kacur
2021-02-24 9:25 ` [PATCH 0/2] Fix for machines without numa Kurt Kanzenbach
2 siblings, 0 replies; 5+ messages in thread
From: John Kacur @ 2021-02-22 22:07 UTC (permalink / raw)
To: RT, Christian Eggers
Cc: Daniel Wagner, Kurt Kanzenbach, Clark Williams,
Sebastian Andrzej Siewior, John Kacur
- Rework numa_initialize a bit to return the status of numa
- Don't fail if numa is not available after the call to numa_initialize
Signed-off-by: John Kacur <jkacur@redhat.com>
---
src/cyclictest/cyclictest.c | 8 ++------
src/lib/rt-numa.c | 17 +++++++++++------
src/signaltest/signaltest.c | 8 ++------
3 files changed, 15 insertions(+), 18 deletions(-)
diff --git a/src/cyclictest/cyclictest.c b/src/cyclictest/cyclictest.c
index 3e31937f7088..157047837259 100644
--- a/src/cyclictest/cyclictest.c
+++ b/src/cyclictest/cyclictest.c
@@ -1018,9 +1018,7 @@ static void process_options(int argc, char *argv[], int max_cpus)
/* smp sets AFFINITY_USEALL in OPT_SMP */
if (smp)
break;
- if (numa_initialize())
- fatal("Couldn't initialize libnuma");
- numa = 1;
+ numa = numa_initialize();
if (optarg) {
parse_cpumask(optarg, max_cpus, &affinity_mask);
setaffinity = AFFINITY_SPECIFIED;
@@ -1204,9 +1202,7 @@ static void process_options(int argc, char *argv[], int max_cpus)
/* if smp wasn't requested, test for numa automatically */
if (!smp) {
- if (numa_initialize())
- fatal("Couldn't initialize libnuma");
- numa = 1;
+ numa = numa_initialize();
if (setaffinity == AFFINITY_UNSPECIFIED)
setaffinity = AFFINITY_USEALL;
}
diff --git a/src/lib/rt-numa.c b/src/lib/rt-numa.c
index dbeaef696876..babcc634d57e 100644
--- a/src/lib/rt-numa.c
+++ b/src/lib/rt-numa.c
@@ -13,19 +13,24 @@
#include "rt-error.h"
#include "rt-numa.h"
-/* numa_available() must be called before any other calls to the numa library */
+/*
+ * numa_available() must be called before any other calls to the numa library
+ * returns 0 if numa is available, or 1 if numa is not available
+ */
int numa_initialize(void)
{
- static int is_initialized;
+ static int is_initialized; // Only call numa_available once
+ static int numa;
if (is_initialized == 1)
- return 0;
+ return numa;
- if (numa_available() == -1)
- return -1;
+ if (numa_available() != -1)
+ numa = 1;
is_initialized = 1;
- return 0;
+
+ return numa;
}
int get_available_cpus(struct bitmask *cpumask)
diff --git a/src/signaltest/signaltest.c b/src/signaltest/signaltest.c
index 4f8e7caea2c1..b1a7e1db8302 100644
--- a/src/signaltest/signaltest.c
+++ b/src/signaltest/signaltest.c
@@ -253,9 +253,7 @@ static void process_options(int argc, char *argv[], unsigned int max_cpus)
/* smp sets AFFINITY_USEALL in OPT_SMP */
if (smp)
break;
- if (numa_initialize())
- fatal("Couldn't initialize libnuma");
- numa = 1;
+ numa = numa_initialize();
if (optarg) {
parse_cpumask(optarg, max_cpus, &affinity_mask);
setaffinity = AFFINITY_SPECIFIED;
@@ -339,9 +337,7 @@ static void process_options(int argc, char *argv[], unsigned int max_cpus)
/* if smp wasn't requested, test for numa automatically */
if (!smp) {
- if (numa_initialize())
- fatal("Couldn't initialize libnuma");
- numa = 1;
+ numa = numa_initialize();
if (setaffinity == AFFINITY_UNSPECIFIED)
setaffinity = AFFINITY_USEALL;
}
--
2.26.2
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH 2/2] rt-tests: remove rt_numa_bitmask_count in rt_numa.h
2021-02-22 22:07 [PATCH 0/2] Fix for machines without numa John Kacur
2021-02-22 22:07 ` [PATCH 1/2] rt-tests: Don't assume numa is available at runtime John Kacur
@ 2021-02-22 22:07 ` John Kacur
2021-02-24 9:25 ` [PATCH 0/2] Fix for machines without numa Kurt Kanzenbach
2 siblings, 0 replies; 5+ messages in thread
From: John Kacur @ 2021-02-22 22:07 UTC (permalink / raw)
To: RT, Christian Eggers
Cc: Daniel Wagner, Kurt Kanzenbach, Clark Williams,
Sebastian Andrzej Siewior, John Kacur
Remove rt_numa_bitmask_count as it is unused.
The code uses numa_bitmask_weight() from numa instead
Signed-off-by: John Kacur <jkacur@redhat.com>
---
src/cyclictest/rt_numa.h | 12 ------------
1 file changed, 12 deletions(-)
diff --git a/src/cyclictest/rt_numa.h b/src/cyclictest/rt_numa.h
index 4cbd979c21e8..03f060e144ce 100644
--- a/src/cyclictest/rt_numa.h
+++ b/src/cyclictest/rt_numa.h
@@ -83,16 +83,4 @@ static inline void rt_bitmask_free(struct bitmask *mask)
numa_bitmask_free(mask);
}
-/** Returns number of bits set in mask. */
-static inline unsigned int rt_numa_bitmask_count(const struct bitmask *mask)
-{
- unsigned int num_bits = 0, i;
- for (i = 0; i < mask->size; i++) {
- if (rt_numa_bitmask_isbitset(mask, i))
- num_bits++;
- }
- /* Could stash this instead of recomputing every time. */
- return num_bits;
-}
-
#endif /* _RT_NUMA_H */
--
2.26.2
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH 0/2] Fix for machines without numa
2021-02-22 22:07 [PATCH 0/2] Fix for machines without numa John Kacur
2021-02-22 22:07 ` [PATCH 1/2] rt-tests: Don't assume numa is available at runtime John Kacur
2021-02-22 22:07 ` [PATCH 2/2] rt-tests: remove rt_numa_bitmask_count in rt_numa.h John Kacur
@ 2021-02-24 9:25 ` Kurt Kanzenbach
2021-02-25 3:36 ` John Kacur
2 siblings, 1 reply; 5+ messages in thread
From: Kurt Kanzenbach @ 2021-02-24 9:25 UTC (permalink / raw)
To: John Kacur, RT, Christian Eggers
Cc: Daniel Wagner, Clark Williams, Sebastian Andrzej Siewior, John Kacur
[-- Attachment #1: Type: text/plain, Size: 532 bytes --]
Hi John,
On Mon Feb 22 2021, John Kacur wrote:
> Trying to return the code to a state where numa is required at compile
> time but not at runtime
>
> - The first patch should allow cyclictest to run on a machine
> without numa.
>
> - The second patch is a clean-up of some old code
>
> The code is pushed in git to branch unstable/devel/no-numa-runtime
> so that people can test before I integrate the code back in the main
> branch
arm32 seems to work fine.
Tested-by: Kurt Kanzenbach <kurt@linutronix.de> # arm32
Thanks,
Kurt
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 0/2] Fix for machines without numa
2021-02-24 9:25 ` [PATCH 0/2] Fix for machines without numa Kurt Kanzenbach
@ 2021-02-25 3:36 ` John Kacur
0 siblings, 0 replies; 5+ messages in thread
From: John Kacur @ 2021-02-25 3:36 UTC (permalink / raw)
To: Kurt Kanzenbach
Cc: RT, Christian Eggers, Daniel Wagner, Clark Williams,
Sebastian Andrzej Siewior
On Wed, 24 Feb 2021, Kurt Kanzenbach wrote:
> Hi John,
>
> On Mon Feb 22 2021, John Kacur wrote:
> > Trying to return the code to a state where numa is required at compile
> > time but not at runtime
> >
> > - The first patch should allow cyclictest to run on a machine
> > without numa.
> >
> > - The second patch is a clean-up of some old code
> >
> > The code is pushed in git to branch unstable/devel/no-numa-runtime
> > so that people can test before I integrate the code back in the main
> > branch
>
> arm32 seems to work fine.
>
> Tested-by: Kurt Kanzenbach <kurt@linutronix.de> # arm32
>
> Thanks,
> Kurt
>
Thanks, I added your tested by and merged all the changes into
unstable/devel/latest
John
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2021-02-25 3:37 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-02-22 22:07 [PATCH 0/2] Fix for machines without numa John Kacur
2021-02-22 22:07 ` [PATCH 1/2] rt-tests: Don't assume numa is available at runtime John Kacur
2021-02-22 22:07 ` [PATCH 2/2] rt-tests: remove rt_numa_bitmask_count in rt_numa.h John Kacur
2021-02-24 9:25 ` [PATCH 0/2] Fix for machines without numa Kurt Kanzenbach
2021-02-25 3:36 ` John Kacur
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).