* [PATCH 0/3] perf bench numa: fix few minor issues
@ 2020-06-29 15:50 Alexander Gordeev
2020-06-29 15:50 ` [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption Alexander Gordeev
` (2 more replies)
0 siblings, 3 replies; 7+ messages in thread
From: Alexander Gordeev @ 2020-06-29 15:50 UTC (permalink / raw)
To: linux-kernel; +Cc: Alexander Gordeev, Arnaldo Carvalho de Melo, Ingo Molnar
Hi all,
Please, find couple of fixes to the tool. Note, I did not bother
fixing leaked memory allocated with numa_allocate_cpumask().
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Alexander Gordeev (3):
perf bench numa: fix incorrect NUMA toplogy assumption
perf bench numa: fix benchmark names
perf bench numa: fix wrong benchmark configuration
tools/perf/bench/numa.c | 72 ++++++++++++++++++++++++-------------------------
1 file changed, 35 insertions(+), 37 deletions(-)
--
1.8.3.1
^ permalink raw reply [flat|nested] 7+ messages in thread* [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption 2020-06-29 15:50 [PATCH 0/3] perf bench numa: fix few minor issues Alexander Gordeev @ 2020-06-29 15:50 ` Alexander Gordeev 2020-06-30 16:18 ` Alexander Gordeev 2020-06-30 16:18 ` [PATCH v2 " Alexander Gordeev 2020-06-29 15:50 ` [PATCH 2/3] perf bench numa: fix benchmark names Alexander Gordeev 2020-06-29 15:50 ` [PATCH 3/3] perf bench numa: fix wrong benchmark configuration Alexander Gordeev 2 siblings, 2 replies; 7+ messages in thread From: Alexander Gordeev @ 2020-06-29 15:50 UTC (permalink / raw) To: linux-kernel; +Cc: Alexander Gordeev, Arnaldo Carvalho de Melo, Ingo Molnar The current code assumes that CPUs are evenly spread among NUMA nodes. That is generally incorrect and leads to failure on systems that have different NUMA topology. Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Ingo Molnar <mingo@redhat.com> Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> --- tools/perf/bench/numa.c | 27 +++++++++++++-------------- 1 file changed, 13 insertions(+), 14 deletions(-) diff --git a/tools/perf/bench/numa.c b/tools/perf/bench/numa.c index 5797253..5497c74 100644 --- a/tools/perf/bench/numa.c +++ b/tools/perf/bench/numa.c @@ -247,12 +247,13 @@ static int is_node_present(int node) */ static bool node_has_cpus(int node) { - struct bitmask *cpu = numa_allocate_cpumask(); + struct bitmask *cpumask = numa_allocate_cpumask(); unsigned int i; - if (cpu && !numa_node_to_cpus(node, cpu)) { - for (i = 0; i < cpu->size; i++) { - if (numa_bitmask_isbitset(cpu, i)) + BUG_ON(cpumask); + if (!numa_node_to_cpus(node, cpumask)) { + for (i = 0; i < cpumask->size; i++) { + if (numa_bitmask_isbitset(cpumask, i)) return true; } } @@ -288,14 +289,10 @@ static cpu_set_t bind_to_cpu(int target_cpu) static cpu_set_t bind_to_node(int target_node) { - int cpus_per_node = g->p.nr_cpus / nr_numa_nodes(); cpu_set_t orig_mask, mask; int cpu; int ret; - BUG_ON(cpus_per_node * nr_numa_nodes() != g->p.nr_cpus); - BUG_ON(!cpus_per_node); - ret = sched_getaffinity(0, sizeof(orig_mask), &orig_mask); BUG_ON(ret); @@ -305,13 +302,15 @@ static cpu_set_t bind_to_node(int target_node) for (cpu = 0; cpu < g->p.nr_cpus; cpu++) CPU_SET(cpu, &mask); } else { - int cpu_start = (target_node + 0) * cpus_per_node; - int cpu_stop = (target_node + 1) * cpus_per_node; - - BUG_ON(cpu_stop > g->p.nr_cpus); + struct bitmask *cpumask = numa_allocate_cpumask(); - for (cpu = cpu_start; cpu < cpu_stop; cpu++) - CPU_SET(cpu, &mask); + BUG_ON(!cpumask); + if (!numa_node_to_cpus(target_node, cpumask)) { + for (cpu = 0; cpu < (int)cpumask->size; cpu++) { + if (numa_bitmask_isbitset(cpumask, cpu)) + CPU_SET(cpu, &mask); + } + } } ret = sched_setaffinity(0, sizeof(mask), &mask); -- 1.8.3.1 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption 2020-06-29 15:50 ` [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption Alexander Gordeev @ 2020-06-30 16:18 ` Alexander Gordeev 2020-06-30 16:18 ` [PATCH v2 " Alexander Gordeev 1 sibling, 0 replies; 7+ messages in thread From: Alexander Gordeev @ 2020-06-30 16:18 UTC (permalink / raw) To: linux-kernel; +Cc: Arnaldo Carvalho de Melo, Ingo Molnar On Mon, Jun 29, 2020 at 05:50:07PM +0200, Alexander Gordeev wrote: > The current code assumes that CPUs are evenly spread among > NUMA nodes. That is generally incorrect and leads to failure > on systems that have different NUMA topology. > > Cc: Arnaldo Carvalho de Melo <acme@kernel.org> > Cc: Ingo Molnar <mingo@redhat.com> > Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> > --- > tools/perf/bench/numa.c | 27 +++++++++++++-------------- > 1 file changed, 13 insertions(+), 14 deletions(-) > > diff --git a/tools/perf/bench/numa.c b/tools/perf/bench/numa.c > index 5797253..5497c74 100644 > --- a/tools/perf/bench/numa.c > +++ b/tools/perf/bench/numa.c > @@ -247,12 +247,13 @@ static int is_node_present(int node) > */ > static bool node_has_cpus(int node) > { > - struct bitmask *cpu = numa_allocate_cpumask(); > + struct bitmask *cpumask = numa_allocate_cpumask(); > unsigned int i; > > - if (cpu && !numa_node_to_cpus(node, cpu)) { > - for (i = 0; i < cpu->size; i++) { > - if (numa_bitmask_isbitset(cpu, i)) > + BUG_ON(cpumask); self-nack - should have been !cpumask, will send the fixed version > + if (!numa_node_to_cpus(node, cpumask)) { > + for (i = 0; i < cpumask->size; i++) { > + if (numa_bitmask_isbitset(cpumask, i)) > return true; > } > } > @@ -288,14 +289,10 @@ static cpu_set_t bind_to_cpu(int target_cpu) > > static cpu_set_t bind_to_node(int target_node) > { > - int cpus_per_node = g->p.nr_cpus / nr_numa_nodes(); > cpu_set_t orig_mask, mask; > int cpu; > int ret; > > - BUG_ON(cpus_per_node * nr_numa_nodes() != g->p.nr_cpus); > - BUG_ON(!cpus_per_node); > - > ret = sched_getaffinity(0, sizeof(orig_mask), &orig_mask); > BUG_ON(ret); > > @@ -305,13 +302,15 @@ static cpu_set_t bind_to_node(int target_node) > for (cpu = 0; cpu < g->p.nr_cpus; cpu++) > CPU_SET(cpu, &mask); > } else { > - int cpu_start = (target_node + 0) * cpus_per_node; > - int cpu_stop = (target_node + 1) * cpus_per_node; > - > - BUG_ON(cpu_stop > g->p.nr_cpus); > + struct bitmask *cpumask = numa_allocate_cpumask(); > > - for (cpu = cpu_start; cpu < cpu_stop; cpu++) > - CPU_SET(cpu, &mask); > + BUG_ON(!cpumask); > + if (!numa_node_to_cpus(target_node, cpumask)) { > + for (cpu = 0; cpu < (int)cpumask->size; cpu++) { > + if (numa_bitmask_isbitset(cpumask, cpu)) > + CPU_SET(cpu, &mask); > + } > + } > } > > ret = sched_setaffinity(0, sizeof(mask), &mask); > -- > 1.8.3.1 > ^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH v2 1/3] perf bench numa: fix incorrect NUMA toplogy assumption 2020-06-29 15:50 ` [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption Alexander Gordeev 2020-06-30 16:18 ` Alexander Gordeev @ 2020-06-30 16:18 ` Alexander Gordeev 1 sibling, 0 replies; 7+ messages in thread From: Alexander Gordeev @ 2020-06-30 16:18 UTC (permalink / raw) To: linux-kernel; +Cc: Arnaldo Carvalho de Melo, Ingo Molnar The current code assumes that CPUs are evenly spread among NUMA nodes. That is generally incorrect and leads to failure on systems that have different NUMA topology. Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Ingo Molnar <mingo@redhat.com> Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> --- tools/perf/bench/numa.c | 27 +++++++++++++-------------- 1 file changed, 13 insertions(+), 14 deletions(-) diff --git a/tools/perf/bench/numa.c b/tools/perf/bench/numa.c index 5797253..23e224e 100644 --- a/tools/perf/bench/numa.c +++ b/tools/perf/bench/numa.c @@ -247,12 +247,13 @@ static int is_node_present(int node) */ static bool node_has_cpus(int node) { - struct bitmask *cpu = numa_allocate_cpumask(); + struct bitmask *cpumask = numa_allocate_cpumask(); unsigned int i; - if (cpu && !numa_node_to_cpus(node, cpu)) { - for (i = 0; i < cpu->size; i++) { - if (numa_bitmask_isbitset(cpu, i)) + BUG_ON(!cpumask); + if (!numa_node_to_cpus(node, cpumask)) { + for (i = 0; i < cpumask->size; i++) { + if (numa_bitmask_isbitset(cpumask, i)) return true; } } @@ -288,14 +289,10 @@ static cpu_set_t bind_to_cpu(int target_cpu) static cpu_set_t bind_to_node(int target_node) { - int cpus_per_node = g->p.nr_cpus / nr_numa_nodes(); cpu_set_t orig_mask, mask; int cpu; int ret; - BUG_ON(cpus_per_node * nr_numa_nodes() != g->p.nr_cpus); - BUG_ON(!cpus_per_node); - ret = sched_getaffinity(0, sizeof(orig_mask), &orig_mask); BUG_ON(ret); @@ -305,13 +302,15 @@ static cpu_set_t bind_to_node(int target_node) for (cpu = 0; cpu < g->p.nr_cpus; cpu++) CPU_SET(cpu, &mask); } else { - int cpu_start = (target_node + 0) * cpus_per_node; - int cpu_stop = (target_node + 1) * cpus_per_node; - - BUG_ON(cpu_stop > g->p.nr_cpus); + struct bitmask *cpumask = numa_allocate_cpumask(); - for (cpu = cpu_start; cpu < cpu_stop; cpu++) - CPU_SET(cpu, &mask); + BUG_ON(!cpumask); + if (!numa_node_to_cpus(target_node, cpumask)) { + for (cpu = 0; cpu < (int)cpumask->size; cpu++) { + if (numa_bitmask_isbitset(cpumask, cpu)) + CPU_SET(cpu, &mask); + } + } } ret = sched_setaffinity(0, sizeof(mask), &mask); -- 1.8.3.1 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 2/3] perf bench numa: fix benchmark names 2020-06-29 15:50 [PATCH 0/3] perf bench numa: fix few minor issues Alexander Gordeev 2020-06-29 15:50 ` [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption Alexander Gordeev @ 2020-06-29 15:50 ` Alexander Gordeev 2020-06-29 15:50 ` [PATCH 3/3] perf bench numa: fix wrong benchmark configuration Alexander Gordeev 2 siblings, 0 replies; 7+ messages in thread From: Alexander Gordeev @ 2020-06-29 15:50 UTC (permalink / raw) To: linux-kernel; +Cc: Alexander Gordeev, Arnaldo Carvalho de Melo, Ingo Molnar Several benchmarks have names that do not correspond to their configuration. Fix that and also some whitespace inconsistencies. Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Ingo Molnar <mingo@redhat.com> Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> --- tools/perf/bench/numa.c | 42 +++++++++++++++++++++--------------------- 1 file changed, 21 insertions(+), 21 deletions(-) diff --git a/tools/perf/bench/numa.c b/tools/perf/bench/numa.c index 5497c74..7965a55 100644 --- a/tools/perf/bench/numa.c +++ b/tools/perf/bench/numa.c @@ -812,12 +812,12 @@ static u64 do_work(u8 *__data, long bytes, int nr, int nr_max, int loop, u64 val } } } else if (!g->p.data_backwards || (nr + loop) & 1) { + /* Process data forwards: */ d0 = data + off; d = data + off + 1; d1 = data + words; - /* Process data forwards: */ for (;;) { if (unlikely(d >= d1)) d = data; @@ -835,7 +835,6 @@ static u64 do_work(u8 *__data, long bytes, int nr, int nr_max, int loop, u64 val d = data + off - 1; d1 = data + words; - /* Process data forwards: */ for (;;) { if (unlikely(d < data)) d = data + words-1; @@ -1732,12 +1731,12 @@ static int run_bench_numa(const char *name, const char **argv) */ static const char *tests[][MAX_ARGS] = { /* Basic single-stream NUMA bandwidth measurements: */ - { "RAM-bw-local,", "mem", "-p", "1", "-t", "1", "-P", "1024", + { "RAM-bw-local,", "mem", "-p", "1", "-t", "1", "-P", "1024", "-C" , "0", "-M", "0", OPT_BW_RAM }, { "RAM-bw-local-NOTHP,", "mem", "-p", "1", "-t", "1", "-P", "1024", "-C" , "0", "-M", "0", OPT_BW_RAM_NOTHP }, - { "RAM-bw-remote,", "mem", "-p", "1", "-t", "1", "-P", "1024", + { "RAM-bw-remote,", "mem", "-p", "1", "-t", "1", "-P", "1024", "-C" , "0", "-M", "1", OPT_BW_RAM }, /* 2-stream NUMA bandwidth measurements: */ @@ -1779,27 +1778,28 @@ static int run_bench_numa(const char *name, const char **argv) "mem", "-p", "8", "-t", "1", "-P", " 512", OPT_BW_NOTHP }, { "16x1-bw-process,", "mem", "-p", "16", "-t", "1", "-P", "256", OPT_BW }, - { " 4x1-bw-thread,", "mem", "-p", "1", "-t", "4", "-T", "256", OPT_BW }, - { " 8x1-bw-thread,", "mem", "-p", "1", "-t", "8", "-T", "256", OPT_BW }, - { "16x1-bw-thread,", "mem", "-p", "1", "-t", "16", "-T", "128", OPT_BW }, - { "32x1-bw-thread,", "mem", "-p", "1", "-t", "32", "-T", "64", OPT_BW }, + { " 1x4-bw-thread,", "mem", "-p", "1", "-t", "4", "-T", "256", OPT_BW }, + { " 1x8-bw-thread,", "mem", "-p", "1", "-t", "8", "-T", "256", OPT_BW }, + { "1x16-bw-thread,", "mem", "-p", "1", "-t", "16", "-T", "128", OPT_BW }, + { "1x32-bw-thread,", "mem", "-p", "1", "-t", "32", "-T", "64", OPT_BW }, - { " 2x3-bw-thread,", "mem", "-p", "2", "-t", "3", "-P", "512", OPT_BW }, - { " 4x4-bw-thread,", "mem", "-p", "4", "-t", "4", "-P", "512", OPT_BW }, - { " 4x6-bw-thread,", "mem", "-p", "4", "-t", "6", "-P", "512", OPT_BW }, - { " 4x8-bw-thread,", "mem", "-p", "4", "-t", "8", "-P", "512", OPT_BW }, - { " 4x8-bw-thread-NOTHP,", + { " 2x3-bw-process,", "mem", "-p", "2", "-t", "3", "-P", "512", OPT_BW }, + { " 4x4-bw-process,", "mem", "-p", "4", "-t", "4", "-P", "512", OPT_BW }, + { " 4x6-bw-process,", "mem", "-p", "4", "-t", "6", "-P", "512", OPT_BW }, + { " 4x8-bw-process,", "mem", "-p", "4", "-t", "8", "-P", "512", OPT_BW }, + { " 4x8-bw-process-NOTHP,", "mem", "-p", "4", "-t", "8", "-P", "512", OPT_BW_NOTHP }, - { " 3x3-bw-thread,", "mem", "-p", "3", "-t", "3", "-P", "512", OPT_BW }, - { " 5x5-bw-thread,", "mem", "-p", "5", "-t", "5", "-P", "512", OPT_BW }, + { " 3x3-bw-process,", "mem", "-p", "3", "-t", "3", "-P", "512", OPT_BW }, + { " 5x5-bw-process,", "mem", "-p", "5", "-t", "5", "-P", "512", OPT_BW }, - { "2x16-bw-thread,", "mem", "-p", "2", "-t", "16", "-P", "512", OPT_BW }, - { "1x32-bw-thread,", "mem", "-p", "1", "-t", "32", "-P", "2048", OPT_BW }, + { "2x16-bw-process,", "mem", "-p", "2", "-t", "16", "-P", "512", OPT_BW }, + { "1x32-bw-process,", "mem", "-p", "1", "-t", "32", "-P", "2048", OPT_BW }, - { "numa02-bw,", "mem", "-p", "1", "-t", "32", "-T", "32", OPT_BW }, - { "numa02-bw-NOTHP,", "mem", "-p", "1", "-t", "32", "-T", "32", OPT_BW_NOTHP }, - { "numa01-bw-thread,", "mem", "-p", "2", "-t", "16", "-T", "192", OPT_BW }, - { "numa01-bw-thread-NOTHP,", + { "1x32-bw-thread,", "mem", "-p", "1", "-t", "32", "-T", "32", OPT_BW }, + { "1x32-bw-thread-NOTHP,", + "mem", "-p", "1", "-t", "32", "-T", "32", OPT_BW_NOTHP }, + { "2x16-bw-thread,", "mem", "-p", "2", "-t", "16", "-T", "192", OPT_BW }, + { "2x16-bw-thread-NOTHP,", "mem", "-p", "2", "-t", "16", "-T", "192", OPT_BW_NOTHP }, }; -- 1.8.3.1 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 3/3] perf bench numa: fix wrong benchmark configuration 2020-06-29 15:50 [PATCH 0/3] perf bench numa: fix few minor issues Alexander Gordeev 2020-06-29 15:50 ` [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption Alexander Gordeev 2020-06-29 15:50 ` [PATCH 2/3] perf bench numa: fix benchmark names Alexander Gordeev @ 2020-06-29 15:50 ` Alexander Gordeev 2 siblings, 0 replies; 7+ messages in thread From: Alexander Gordeev @ 2020-06-29 15:50 UTC (permalink / raw) To: linux-kernel; +Cc: Alexander Gordeev, Arnaldo Carvalho de Melo, Ingo Molnar This update fixes a wrong number of processes for one benchmark and re-groups two others in a more unified way. Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Ingo Molnar <mingo@redhat.com> Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> --- tools/perf/bench/numa.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/tools/perf/bench/numa.c b/tools/perf/bench/numa.c index 7965a55..30fa6ef 100644 --- a/tools/perf/bench/numa.c +++ b/tools/perf/bench/numa.c @@ -1753,7 +1753,7 @@ static int run_bench_numa(const char *name, const char **argv) { " 1x3-convergence,", "mem", "-p", "1", "-t", "3", "-P", "512", OPT_CONV }, { " 1x4-convergence,", "mem", "-p", "1", "-t", "4", "-P", "512", OPT_CONV }, { " 1x6-convergence,", "mem", "-p", "1", "-t", "6", "-P", "1020", OPT_CONV }, - { " 2x3-convergence,", "mem", "-p", "3", "-t", "3", "-P", "1020", OPT_CONV }, + { " 2x3-convergence,", "mem", "-p", "2", "-t", "3", "-P", "1020", OPT_CONV }, { " 3x3-convergence,", "mem", "-p", "3", "-t", "3", "-P", "1020", OPT_CONV }, { " 4x4-convergence,", "mem", "-p", "4", "-t", "4", "-P", "512", OPT_CONV }, { " 4x4-convergence-NOTHP,", @@ -1782,6 +1782,8 @@ static int run_bench_numa(const char *name, const char **argv) { " 1x8-bw-thread,", "mem", "-p", "1", "-t", "8", "-T", "256", OPT_BW }, { "1x16-bw-thread,", "mem", "-p", "1", "-t", "16", "-T", "128", OPT_BW }, { "1x32-bw-thread,", "mem", "-p", "1", "-t", "32", "-T", "64", OPT_BW }, + { "1x32-bw-thread-NOTHP,", + "mem", "-p", "1", "-t", "32", "-T", "64", OPT_BW_NOTHP }, { " 2x3-bw-process,", "mem", "-p", "2", "-t", "3", "-P", "512", OPT_BW }, { " 4x4-bw-process,", "mem", "-p", "4", "-t", "4", "-P", "512", OPT_BW }, @@ -1795,9 +1797,6 @@ static int run_bench_numa(const char *name, const char **argv) { "2x16-bw-process,", "mem", "-p", "2", "-t", "16", "-P", "512", OPT_BW }, { "1x32-bw-process,", "mem", "-p", "1", "-t", "32", "-P", "2048", OPT_BW }, - { "1x32-bw-thread,", "mem", "-p", "1", "-t", "32", "-T", "32", OPT_BW }, - { "1x32-bw-thread-NOTHP,", - "mem", "-p", "1", "-t", "32", "-T", "32", OPT_BW_NOTHP }, { "2x16-bw-thread,", "mem", "-p", "2", "-t", "16", "-T", "192", OPT_BW }, { "2x16-bw-thread-NOTHP,", "mem", "-p", "2", "-t", "16", "-T", "192", OPT_BW_NOTHP }, -- 1.8.3.1 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 0/3] perf bench numa: fix few minor issues @ 2020-07-06 9:42 Alexander Gordeev 2020-07-06 9:42 ` [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption Alexander Gordeev 0 siblings, 1 reply; 7+ messages in thread From: Alexander Gordeev @ 2020-07-06 9:42 UTC (permalink / raw) To: linux-perf-users; +Cc: linux-kernel, Alexander Gordeev Hi all, Please, find couple of fixes to the tool. Note, I did not bother fixing leaked memory allocated with numa_allocate_cpumask(). Alexander Gordeev (3): perf bench numa: fix incorrect NUMA toplogy assumption perf bench numa: fix benchmark names perf bench numa: fix wrong benchmark configuration tools/perf/bench/numa.c | 72 ++++++++++++++++++++++++------------------------- 1 file changed, 35 insertions(+), 37 deletions(-) -- 1.8.3.1 ^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption 2020-07-06 9:42 [PATCH 0/3] perf bench numa: fix few minor issues Alexander Gordeev @ 2020-07-06 9:42 ` Alexander Gordeev 0 siblings, 0 replies; 7+ messages in thread From: Alexander Gordeev @ 2020-07-06 9:42 UTC (permalink / raw) To: linux-perf-users; +Cc: linux-kernel, Alexander Gordeev The current code assumes that CPUs are evenly spread among NUMA nodes. That is generally incorrect and leads to failure on systems that have different NUMA topology. Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> --- tools/perf/bench/numa.c | 27 +++++++++++++-------------- 1 file changed, 13 insertions(+), 14 deletions(-) diff --git a/tools/perf/bench/numa.c b/tools/perf/bench/numa.c index 5797253..23e224e 100644 --- a/tools/perf/bench/numa.c +++ b/tools/perf/bench/numa.c @@ -247,12 +247,13 @@ static int is_node_present(int node) */ static bool node_has_cpus(int node) { - struct bitmask *cpu = numa_allocate_cpumask(); + struct bitmask *cpumask = numa_allocate_cpumask(); unsigned int i; - if (cpu && !numa_node_to_cpus(node, cpu)) { - for (i = 0; i < cpu->size; i++) { - if (numa_bitmask_isbitset(cpu, i)) + BUG_ON(!cpumask); + if (!numa_node_to_cpus(node, cpumask)) { + for (i = 0; i < cpumask->size; i++) { + if (numa_bitmask_isbitset(cpumask, i)) return true; } } @@ -288,14 +289,10 @@ static cpu_set_t bind_to_cpu(int target_cpu) static cpu_set_t bind_to_node(int target_node) { - int cpus_per_node = g->p.nr_cpus / nr_numa_nodes(); cpu_set_t orig_mask, mask; int cpu; int ret; - BUG_ON(cpus_per_node * nr_numa_nodes() != g->p.nr_cpus); - BUG_ON(!cpus_per_node); - ret = sched_getaffinity(0, sizeof(orig_mask), &orig_mask); BUG_ON(ret); @@ -305,13 +302,15 @@ static cpu_set_t bind_to_node(int target_node) for (cpu = 0; cpu < g->p.nr_cpus; cpu++) CPU_SET(cpu, &mask); } else { - int cpu_start = (target_node + 0) * cpus_per_node; - int cpu_stop = (target_node + 1) * cpus_per_node; - - BUG_ON(cpu_stop > g->p.nr_cpus); + struct bitmask *cpumask = numa_allocate_cpumask(); - for (cpu = cpu_start; cpu < cpu_stop; cpu++) - CPU_SET(cpu, &mask); + BUG_ON(!cpumask); + if (!numa_node_to_cpus(target_node, cpumask)) { + for (cpu = 0; cpu < (int)cpumask->size; cpu++) { + if (numa_bitmask_isbitset(cpumask, cpu)) + CPU_SET(cpu, &mask); + } + } } ret = sched_setaffinity(0, sizeof(mask), &mask); -- 1.8.3.1 ^ permalink raw reply related [flat|nested] 7+ messages in thread
end of thread, other threads:[~2020-07-06 9:43 UTC | newest] Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2020-06-29 15:50 [PATCH 0/3] perf bench numa: fix few minor issues Alexander Gordeev 2020-06-29 15:50 ` [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption Alexander Gordeev 2020-06-30 16:18 ` Alexander Gordeev 2020-06-30 16:18 ` [PATCH v2 " Alexander Gordeev 2020-06-29 15:50 ` [PATCH 2/3] perf bench numa: fix benchmark names Alexander Gordeev 2020-06-29 15:50 ` [PATCH 3/3] perf bench numa: fix wrong benchmark configuration Alexander Gordeev 2020-07-06 9:42 [PATCH 0/3] perf bench numa: fix few minor issues Alexander Gordeev 2020-07-06 9:42 ` [PATCH 1/3] perf bench numa: fix incorrect NUMA toplogy assumption Alexander Gordeev
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.