From: Akihiko Odaki <akihiko.odaki@daynix.com>
To: unlisted-recipients:; (no To-header on input)
Cc: util-linux@vger.kernel.org, Akihiko Odaki <akihiko.odaki@daynix.com>
Subject: [PATCH] lscpu: Decouple cputype and topology
Date: Wed, 5 Apr 2023 15:12:38 +0900 [thread overview]
Message-ID: <20230405061238.11888-1-akihiko.odaki@daynix.com> (raw)
Before this change, lscpu_cputype held topology information. This
design is incompatible with heterogenous configurations where there are
several CPU types.
One problem is that logical IDs of e.g. clusters overlap across
different CPU types. For example, consider a
1-socket/2-cluster/1-core/1-thread system. One of the clusters has
"P-cores", and the other has "E-cores". P-cores and E-cores have
different CPU types. Before this change,
"lscpu -p=CPU,Core,Cluster,Socket" output something like the following
for the system:
0,0,0,0
1,0,0,0
Note that lscpu assigns the same core/cluster ID for the two CPUs
although they are actually in different cores and clusters.
To fix the inconsistency and ambiguity of such IDs, move the topology
information from lscpu_cputype to lscpu_cxt. For the earlier example,
the output will change as follows:
0,0,0,0
1,1,1,0
This also changes how the topology is described in the summary which
lscpu prints when it is executed with no arguments. Before this change,
the topology information was associated with CPU types, but the
topology information is shown in a separate section now.
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
---
sys-utils/lscpu-arm.c | 6 +--
sys-utils/lscpu-cputype.c | 1 -
sys-utils/lscpu-topology.c | 108 ++++++++-----------------------------
sys-utils/lscpu.c | 64 +++++++++++-----------
sys-utils/lscpu.h | 32 +++++------
5 files changed, 70 insertions(+), 141 deletions(-)
diff --git a/sys-utils/lscpu-arm.c b/sys-utils/lscpu-arm.c
index 86e5ea7b4..5ea534bd9 100644
--- a/sys-utils/lscpu-arm.c
+++ b/sys-utils/lscpu-arm.c
@@ -343,9 +343,6 @@ static void arm_decode(struct lscpu_cxt *cxt, struct lscpu_cputype *ct)
arm_ids_decode(ct);
arm_rXpY_decode(ct);
-
- if (!cxt->noalive && cxt->is_cluster)
- ct->nr_socket_on_cluster = get_number_of_physical_sockets_from_dmi();
}
static int is_cluster_arm(struct lscpu_cxt *cxt)
@@ -368,4 +365,7 @@ void lscpu_decode_arm(struct lscpu_cxt *cxt)
for (i = 0; i < cxt->ncputypes; i++)
arm_decode(cxt, cxt->cputypes[i]);
+
+ if (!cxt->noalive && cxt->is_cluster)
+ cxt->nr_socket_on_cluster = get_number_of_physical_sockets_from_dmi();
}
diff --git a/sys-utils/lscpu-cputype.c b/sys-utils/lscpu-cputype.c
index 31f4009bc..18790f35f 100644
--- a/sys-utils/lscpu-cputype.c
+++ b/sys-utils/lscpu-cputype.c
@@ -77,7 +77,6 @@ void lscpu_unref_cputype(struct lscpu_cputype *ct)
if (--ct->refcount <= 0) {
DBG(TYPE, ul_debugobj(ct, " freeing %s/%s", ct->vendor, ct->model));
- lscpu_cputype_free_topology(ct);
free(ct->vendor);
free(ct->bios_vendor);
free(ct->machinetype); /* s390 */
diff --git a/sys-utils/lscpu-topology.c b/sys-utils/lscpu-topology.c
index fe4da7cd5..379b2b584 100644
--- a/sys-utils/lscpu-topology.c
+++ b/sys-utils/lscpu-topology.c
@@ -41,14 +41,14 @@ static void free_cpuset_array(cpu_set_t **ary, int items)
free(ary);
}
-void lscpu_cputype_free_topology(struct lscpu_cputype *ct)
+void lscpu_free_topology(struct lscpu_cxt *cxt)
{
- if (!ct)
+ if (!cxt)
return;
- free_cpuset_array(ct->coremaps, ct->ncores);
- free_cpuset_array(ct->socketmaps, ct->nsockets);
- free_cpuset_array(ct->bookmaps, ct->nbooks);
- free_cpuset_array(ct->drawermaps, ct->ndrawers);
+ free_cpuset_array(cxt->coremaps, cxt->ncores);
+ free_cpuset_array(cxt->socketmaps, cxt->nsockets);
+ free_cpuset_array(cxt->bookmaps, cxt->nbooks);
+ free_cpuset_array(cxt->drawermaps, cxt->ndrawers);
}
void lscpu_free_caches(struct lscpu_cache *caches, size_t n)
@@ -89,26 +89,21 @@ void lscpu_sort_caches(struct lscpu_cache *caches, size_t n)
/* Read topology for specified type */
-static int cputype_read_topology(struct lscpu_cxt *cxt, struct lscpu_cputype *ct)
+static int cputype_read_topology(struct lscpu_cxt *cxt)
{
size_t i, npos;
struct path_cxt *sys;
- int nthreads = 0, sw_topo = 0;
- FILE *fd;
sys = cxt->syscpu; /* /sys/devices/system/cpu/ */
npos = cxt->npossibles; /* possible CPUs */
- DBG(TYPE, ul_debugobj(ct, "reading %s/%s/%s topology",
- ct->vendor ?: "", ct->model ?: "", ct->modelname ?:""));
-
for (i = 0; i < cxt->npossibles; i++) {
struct lscpu_cpu *cpu = cxt->cpus[i];
cpu_set_t *thread_siblings = NULL, *core_siblings = NULL;
cpu_set_t *book_siblings = NULL, *drawer_siblings = NULL;
- int num, n;
+ int num;
- if (!cpu || cpu->type != ct)
+ if (!cpu)
continue;
num = cpu->logical_id;
@@ -126,12 +121,6 @@ static int cputype_read_topology(struct lscpu_cxt *cxt, struct lscpu_cputype *ct
ul_path_readf_cpuset(sys, &drawer_siblings, cxt->maxcpus,
"cpu%d/topology/drawer_siblings", num);
- n = CPU_COUNT_S(cxt->setsize, thread_siblings);
- if (!n)
- n = 1;
- if (n > nthreads)
- nthreads = n;
-
/* Allocate arrays for topology maps.
*
* For each map we make sure that it can have up to ncpuspos
@@ -140,76 +129,26 @@ static int cputype_read_topology(struct lscpu_cxt *cxt, struct lscpu_cputype *ct
* E.g. completely virtualized architectures like s390 may
* have multiple sockets of different sizes.
*/
- if (!ct->coremaps)
- ct->coremaps = xcalloc(npos, sizeof(cpu_set_t *));
- if (!ct->socketmaps)
- ct->socketmaps = xcalloc(npos, sizeof(cpu_set_t *));
- if (!ct->bookmaps && book_siblings)
- ct->bookmaps = xcalloc(npos, sizeof(cpu_set_t *));
- if (!ct->drawermaps && drawer_siblings)
- ct->drawermaps = xcalloc(npos, sizeof(cpu_set_t *));
+ if (!cxt->coremaps)
+ cxt->coremaps = xcalloc(npos, sizeof(cpu_set_t *));
+ if (!cxt->socketmaps)
+ cxt->socketmaps = xcalloc(npos, sizeof(cpu_set_t *));
+ if (!cxt->bookmaps && book_siblings)
+ cxt->bookmaps = xcalloc(npos, sizeof(cpu_set_t *));
+ if (!cxt->drawermaps && drawer_siblings)
+ cxt->drawermaps = xcalloc(npos, sizeof(cpu_set_t *));
/* add to topology maps */
- add_cpuset_to_array(ct->coremaps, &ct->ncores, thread_siblings, cxt->setsize);
- add_cpuset_to_array(ct->socketmaps, &ct->nsockets, core_siblings, cxt->setsize);
+ add_cpuset_to_array(cxt->coremaps, &cxt->ncores, thread_siblings, cxt->setsize);
+ add_cpuset_to_array(cxt->socketmaps, &cxt->nsockets, core_siblings, cxt->setsize);
if (book_siblings)
- add_cpuset_to_array(ct->bookmaps, &ct->nbooks, book_siblings, cxt->setsize);
+ add_cpuset_to_array(cxt->bookmaps, &cxt->nbooks, book_siblings, cxt->setsize);
if (drawer_siblings)
- add_cpuset_to_array(ct->drawermaps, &ct->ndrawers, drawer_siblings, cxt->setsize);
-
- }
+ add_cpuset_to_array(cxt->drawermaps, &cxt->ndrawers, drawer_siblings, cxt->setsize);
- /* s390 detects its cpu topology via /proc/sysinfo, if present.
- * Using simply the cpu topology masks in sysfs will not give
- * usable results since everything is virtualized. E.g.
- * virtual core 0 may have only 1 cpu, but virtual core 2 may
- * five cpus.
- * If the cpu topology is not exported (e.g. 2nd level guest)
- * fall back to old calculation scheme.
- */
- if ((fd = ul_path_fopen(cxt->procfs, "r", "sysinfo"))) {
- int t0, t1;
- char buf[BUFSIZ];
-
- DBG(TYPE, ul_debugobj(ct, " reading sysinfo"));
-
- while (fgets(buf, sizeof(buf), fd) != NULL) {
- if (sscanf(buf, "CPU Topology SW: %d %d %zu %zu %zu %zu",
- &t0, &t1,
- &ct->ndrawers_per_system,
- &ct->nbooks_per_drawer,
- &ct->nsockets_per_book,
- &ct->ncores_per_socket) == 6) {
- sw_topo = 1;
- DBG(TYPE, ul_debugobj(ct, " using SW topology"));
- break;
- }
- }
- if (fd)
- fclose(fd);
}
- ct->nthreads_per_core = nthreads;
- if (ct->mtid) {
- uint64_t x;
- if (ul_strtou64(ct->mtid, &x, 10) == 0 && x <= ULONG_MAX)
- ct->nthreads_per_core = (size_t) x + 1;
- }
-
- if (!sw_topo) {
- ct->ncores_per_socket = ct->nsockets ? ct->ncores / ct->nsockets : 0;
- ct->nsockets_per_book = ct->nbooks ? ct->nsockets / ct->nbooks : 0;
- ct->nbooks_per_drawer = ct->ndrawers ? ct->nbooks / ct->ndrawers : 0;
- ct->ndrawers_per_system = ct->ndrawers;
- }
-
- DBG(TYPE, ul_debugobj(ct, " nthreads: %zu (per core)", ct->nthreads_per_core));
- DBG(TYPE, ul_debugobj(ct, " ncores: %zu (%zu per socket)", ct->ncores, ct->ncores_per_socket));
- DBG(TYPE, ul_debugobj(ct, " nsockets: %zu (%zu per books)", ct->nsockets, ct->nsockets_per_book));
- DBG(TYPE, ul_debugobj(ct, " nbooks: %zu (%zu per drawer)", ct->nbooks, ct->nbooks_per_drawer));
- DBG(TYPE, ul_debugobj(ct, " ndrawers: %zu (%zu per system)", ct->ndrawers, ct->ndrawers_per_system));
-
return 0;
}
@@ -622,11 +561,10 @@ float lsblk_cputype_get_scalmhz(struct lscpu_cxt *cxt, struct lscpu_cputype *ct)
int lscpu_read_topology(struct lscpu_cxt *cxt)
{
size_t i;
- int rc = 0;
+ int rc;
- for (i = 0; i < cxt->ncputypes; i++)
- rc += cputype_read_topology(cxt, cxt->cputypes[i]);
+ rc = cputype_read_topology(cxt);
for (i = 0; rc == 0 && i < cxt->npossibles; i++) {
struct lscpu_cpu *cpu = cxt->cpus[i];
diff --git a/sys-utils/lscpu.c b/sys-utils/lscpu.c
index 86fd0b968..9516112cc 100644
--- a/sys-utils/lscpu.c
+++ b/sys-utils/lscpu.c
@@ -285,6 +285,8 @@ static void lscpu_free_context(struct lscpu_cxt *cxt)
lscpu_free_caches(cxt->ecaches, cxt->necaches);
lscpu_free_caches(cxt->caches, cxt->ncaches);
+ lscpu_free_topology(cxt);
+
free(cxt);
}
@@ -329,8 +331,8 @@ static void get_cell_boolean(
#define fill_id(_cxt, _cpu, NAME, _buf, _bufsz) \
__fill_id(_cxt, (_cpu), \
(_cpu)-> NAME ## id, \
- (_cpu)->type-> NAME ## maps, \
- (_cpu)->type->n ## NAME ## s, \
+ (_cxt)-> NAME ## maps, \
+ (_cxt)->n ## NAME ## s, \
_buf, _bufsz)
static char *get_cell_data(
@@ -872,33 +874,6 @@ print_summary_cputype(struct lscpu_cxt *cxt,
if (ct->model || ct->revision)
add_summary_s(tb, sec, _("Model:"), ct->revision ? ct->revision : ct->model);
- add_summary_n(tb, sec, _("Thread(s) per core:"), ct->nthreads_per_core);
- if (cxt->is_cluster)
- add_summary_n(tb, sec, _("Core(s) per cluster:"), ct->ncores_per_socket);
- else
- add_summary_n(tb, sec, _("Core(s) per socket:"), ct->ncores_per_socket);
-
- if (ct->nbooks) {
- add_summary_n(tb, sec, _("Socket(s) per book:"), ct->nsockets_per_book);
- if (ct->ndrawers_per_system || ct->ndrawers) {
- add_summary_n(tb, sec, _("Book(s) per drawer:"), ct->nbooks_per_drawer);
- add_summary_n(tb, sec, _("Drawer(s):"), ct->ndrawers_per_system ?: ct->ndrawers);
- } else
- add_summary_n(tb, sec, _("Book(s):"), ct->nbooks_per_drawer ?: ct->nbooks);
- } else {
- if (cxt->is_cluster) {
- if (ct->nr_socket_on_cluster > 0)
- add_summary_n(tb, sec, _("Socket(s):"), ct->nr_socket_on_cluster);
- else
- add_summary_s(tb, sec, _("Socket(s):"), "-");
-
- add_summary_n(tb, sec, _("Cluster(s):"),
- ct->nsockets_per_book ?: ct->nsockets);
- } else
- add_summary_n(tb, sec, _("Socket(s):"),
- ct->nsockets_per_book ?: ct->nsockets);
- }
-
if (ct->stepping)
add_summary_s(tb, sec, _("Stepping:"), ct->stepping);
if (ct->freqboost >= 0)
@@ -1042,6 +1017,29 @@ static void print_summary(struct lscpu_cxt *cxt)
print_summary_cputype(cxt, cxt->cputypes[i], tb, sec);
sec = NULL;
+
+ /* Section: topology */
+ sec = add_summary_e(tb, NULL, _("Topology:"));
+ add_summary_n(tb, sec, _("Thread(s):"), cxt->npossibles);
+ add_summary_n(tb, sec, _("Core(s):"), cxt->ncores);
+
+ if (cxt->is_cluster) {
+ add_summary_n(tb, sec, _("Cluster(s):"), cxt->nsockets);
+
+ if (cxt->nr_socket_on_cluster > 0)
+ add_summary_n(tb, sec, _("Socket(s):"), cxt->nr_socket_on_cluster);
+ else
+ add_summary_s(tb, sec, _("Socket(s):"), "-");
+ } else
+ add_summary_n(tb, sec, _("Socket(s):"), cxt->nsockets);
+
+ if (cxt->nbooks) {
+ if (cxt->ndrawers)
+ add_summary_n(tb, sec, _("Drawer(s):"), cxt->ndrawers);
+
+ add_summary_n(tb, sec, _("Book(s):"), cxt->nbooks);
+ }
+
/* Section: vitualiazation */
if (cxt->virt) {
sec = add_summary_e(tb, NULL, _("Virtualization features:"));
@@ -1376,17 +1374,17 @@ int main(int argc, char *argv[])
columns[ncolumns++] = COL_CPU_CPU;
if (cxt->nnodes)
columns[ncolumns++] = COL_CPU_NODE;
- if (ct && ct->ndrawers)
+ if (cxt->ndrawers)
columns[ncolumns++] = COL_CPU_DRAWER;
- if (ct && ct->nbooks)
+ if (cxt->nbooks)
columns[ncolumns++] = COL_CPU_BOOK;
- if (ct && ct->nsockets) {
+ if (cxt->nsockets) {
if (cxt->is_cluster)
columns[ncolumns++] = COL_CPU_CLUSTER;
else
columns[ncolumns++] = COL_CPU_SOCKET;
}
- if (ct && ct->ncores)
+ if (cxt->ncores)
columns[ncolumns++] = COL_CPU_CORE;
if (cxt->ncaches)
columns[ncolumns++] = COL_CPU_CACHE;
diff --git a/sys-utils/lscpu.h b/sys-utils/lscpu.h
index 8cb0d630c..5f6617b76 100644
--- a/sys-utils/lscpu.h
+++ b/sys-utils/lscpu.h
@@ -78,31 +78,13 @@ struct lscpu_cputype {
size_t physchips; /* Physical chips */
size_t physcoresperchip; /* Physical cores per chip */
- size_t nthreads_per_core;
- size_t ncores_per_socket;
- size_t nsockets_per_book;
- size_t nbooks_per_drawer;
- size_t ndrawers_per_system;
-
char *dynamic_mhz; /* s390; copy from the first CPU */
char *static_mhz; /* s390; copy from the first CPU */
- /* siblings maps */
- size_t ncores;
- cpu_set_t **coremaps;
- size_t nsockets;
- cpu_set_t **socketmaps;
- size_t nbooks;
- cpu_set_t **bookmaps;
- size_t ndrawers;
- cpu_set_t **drawermaps;
-
unsigned int has_freq : 1,
has_configured : 1,
has_polarization : 1,
has_addresses : 1;
-
- size_t nr_socket_on_cluster; /* the number of sockets if the is_cluster is 1 */
};
/* dispatching modes */
@@ -247,6 +229,18 @@ struct lscpu_cxt {
bytes : 1;
int is_cluster; /* For aarch64 if the machine doesn't have ACPI PPTT */
+
+ /* siblings maps */
+ size_t ncores;
+ cpu_set_t **coremaps;
+ size_t nsockets;
+ cpu_set_t **socketmaps;
+ size_t nbooks;
+ cpu_set_t **bookmaps;
+ size_t ndrawers;
+ cpu_set_t **drawermaps;
+
+ size_t nr_socket_on_cluster; /* the number of sockets if the is_cluster is 1 */
};
#define is_cpu_online(_cxt, _cpu) \
@@ -277,7 +271,7 @@ struct lscpu_cache *lscpu_cpu_get_cache(struct lscpu_cxt *cxt,
struct lscpu_cpu *cpu, const char *name);
int lscpu_read_topology(struct lscpu_cxt *cxt);
-void lscpu_cputype_free_topology(struct lscpu_cputype *ct);
+void lscpu_free_topology(struct lscpu_cxt *cxt);
float lsblk_cputype_get_maxmhz(struct lscpu_cxt *cxt, struct lscpu_cputype *ct);
float lsblk_cputype_get_minmhz(struct lscpu_cxt *cxt, struct lscpu_cputype *ct);
--
2.40.0
next reply other threads:[~2023-04-05 6:12 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-04-05 6:12 Akihiko Odaki [this message]
2023-04-05 10:43 ` [PATCH] lscpu: Decouple cputype and topology Karel Zak
2023-04-05 15:32 ` Akihiko Odaki
2023-04-25 18:29 ` Akihiko Odaki
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230405061238.11888-1-akihiko.odaki@daynix.com \
--to=akihiko.odaki@daynix.com \
--cc=util-linux@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is 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 read-only IMAP folder(s) and NNTP newsgroup(s).