From a9e3760b0838299649c0d57cca44daaf40ba3c33 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 14 Sep 2026 22:12:34 -1000 Subject: sched_ext: Maintain an online cid mask in the scheduler arena Schedulers on the default cid mapping treat [0, nr_online_cids) as the online set and restart on hotplug. Schedulers that install their own mapping with scx_bpf_cid_override() have no way to learn which cids are online: the count no longer identifies members and the CPU-form cpumask is unusable from cid programs. This is an obvious hole in the cid API. Add scx_bpf_online_cmask(), a kernel-maintained cmask in the scheduler's arena, allocated alongside the per-CPU scratch masks and populated after the cid mapping is finalized and before ops.init(), for child schedulers too. The pointer stays valid through ops.exit() with no reference to take. It is the arena offset as a void pointer, the same form struct_ops arena arguments arrive in. The verifier types the void return as a scalar for the program's arena cast. The mask follows the SCX hotplug notifications: seeded from cpu_active_mask and updated before ops.cid_online/offline() runs, so it lags cpu_online_mask only inside a hotplug transition. Updates walk the scheduler list under the lock that also serializes unlinking. Reads are live, not atomic snapshots. Root initialization excludes hotplug. v2: Reworded the getter kerneldoc (Andrea Righi). Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi --- tools/testing/selftests/prctl/.gitignore | 6 ++ tools/testing/selftests/prctl/Makefile | 14 +++ tools/testing/selftests/prctl/config | 1 + .../prctl/disable-tsc-ctxt-sw-stress-test.c | 98 +++++++++++++++++++ .../prctl/disable-tsc-on-off-stress-test.c | 97 +++++++++++++++++++ tools/testing/selftests/prctl/disable-tsc-test.c | 96 +++++++++++++++++++ .../selftests/prctl/set-anon-vma-name-test.c | 104 +++++++++++++++++++++ tools/testing/selftests/prctl/set-process-name.c | 94 +++++++++++++++++++ 8 files changed, 510 insertions(+) create mode 100644 tools/testing/selftests/prctl/.gitignore create mode 100644 tools/testing/selftests/prctl/Makefile create mode 100644 tools/testing/selftests/prctl/config create mode 100644 tools/testing/selftests/prctl/disable-tsc-ctxt-sw-stress-test.c create mode 100644 tools/testing/selftests/prctl/disable-tsc-on-off-stress-test.c create mode 100644 tools/testing/selftests/prctl/disable-tsc-test.c create mode 100644 tools/testing/selftests/prctl/set-anon-vma-name-test.c create mode 100644 tools/testing/selftests/prctl/set-process-name.c (limited to 'tools/testing/selftests/prctl') diff --git a/tools/testing/selftests/prctl/.gitignore b/tools/testing/selftests/prctl/.gitignore new file mode 100644 index 000000000..05d5e3166 --- /dev/null +++ b/tools/testing/selftests/prctl/.gitignore @@ -0,0 +1,6 @@ +# SPDX-License-Identifier: GPL-2.0-only +disable-tsc-ctxt-sw-stress-test +disable-tsc-on-off-stress-test +disable-tsc-test +set-anon-vma-name-test +set-process-name diff --git a/tools/testing/selftests/prctl/Makefile b/tools/testing/selftests/prctl/Makefile new file mode 100644 index 000000000..e770e86fa --- /dev/null +++ b/tools/testing/selftests/prctl/Makefile @@ -0,0 +1,14 @@ +# SPDX-License-Identifier: GPL-2.0 +ifndef CROSS_COMPILE +ARCH ?= $(shell uname -m 2>/dev/null || echo not) +override ARCH := $(shell echo $(ARCH) | sed -e s/i.86/x86/ -e s/x86_64/x86/) + +ifeq ($(ARCH),x86) +TEST_PROGS := disable-tsc-ctxt-sw-stress-test disable-tsc-on-off-stress-test \ + disable-tsc-test set-anon-vma-name-test set-process-name +all: $(TEST_PROGS) + +include ../lib.mk + +endif +endif diff --git a/tools/testing/selftests/prctl/config b/tools/testing/selftests/prctl/config new file mode 100644 index 000000000..c6ed03c54 --- /dev/null +++ b/tools/testing/selftests/prctl/config @@ -0,0 +1 @@ +CONFIG_ANON_VMA_NAME=y diff --git a/tools/testing/selftests/prctl/disable-tsc-ctxt-sw-stress-test.c b/tools/testing/selftests/prctl/disable-tsc-ctxt-sw-stress-test.c new file mode 100644 index 000000000..6d1a5ee8e --- /dev/null +++ b/tools/testing/selftests/prctl/disable-tsc-ctxt-sw-stress-test.c @@ -0,0 +1,98 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Tests for prctl(PR_GET_TSC, ...) / prctl(PR_SET_TSC, ...) + * + * Tests if the control register is updated correctly + * at context switches + * + * Warning: this test will cause a very high load for a few seconds + * + */ + +#include +#include +#include +#include +#include +#include + + +#include +#include + +/* Get/set the process' ability to use the timestamp counter instruction */ +#ifndef PR_GET_TSC +#define PR_GET_TSC 25 +#define PR_SET_TSC 26 +# define PR_TSC_ENABLE 1 /* allow the use of the timestamp counter */ +# define PR_TSC_SIGSEGV 2 /* throw a SIGSEGV instead of reading the TSC */ +#endif + +static uint64_t rdtsc(void) +{ +uint32_t lo, hi; +/* We cannot use "=A", since this would use %rax on x86_64 */ +__asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi)); +return (uint64_t)hi << 32 | lo; +} + +static void sigsegv_expect(int sig) +{ + /* */ +} + +static void segvtask(void) +{ + if (prctl(PR_SET_TSC, PR_TSC_SIGSEGV) < 0) + { + perror("prctl"); + exit(0); + } + signal(SIGSEGV, sigsegv_expect); + alarm(10); + rdtsc(); + fprintf(stderr, "FATAL ERROR, rdtsc() succeeded while disabled\n"); + exit(0); +} + + +static void sigsegv_fail(int sig) +{ + fprintf(stderr, "FATAL ERROR, rdtsc() failed while enabled\n"); + exit(0); +} + +static void rdtsctask(void) +{ + if (prctl(PR_SET_TSC, PR_TSC_ENABLE) < 0) + { + perror("prctl"); + exit(0); + } + signal(SIGSEGV, sigsegv_fail); + alarm(10); + for(;;) rdtsc(); +} + + +int main(void) +{ + int n_tasks = 100, i; + + fprintf(stderr, "[No further output means we're all right]\n"); + + for (i=0; i +#include +#include +#include +#include +#include + + +#include +#include + +/* Get/set the process' ability to use the timestamp counter instruction */ +#ifndef PR_GET_TSC +#define PR_GET_TSC 25 +#define PR_SET_TSC 26 +# define PR_TSC_ENABLE 1 /* allow the use of the timestamp counter */ +# define PR_TSC_SIGSEGV 2 /* throw a SIGSEGV instead of reading the TSC */ +#endif + +/* snippet from wikipedia :-) */ + +static uint64_t rdtsc(void) +{ +uint32_t lo, hi; +/* We cannot use "=A", since this would use %rax on x86_64 */ +__asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi)); +return (uint64_t)hi << 32 | lo; +} + +int should_segv = 0; + +static void sigsegv_cb(int sig) +{ + if (!should_segv) + { + fprintf(stderr, "FATAL ERROR, rdtsc() failed while enabled\n"); + exit(0); + } + if (prctl(PR_SET_TSC, PR_TSC_ENABLE) < 0) + { + perror("prctl"); + exit(0); + } + should_segv = 0; + + rdtsc(); +} + +static void task(void) +{ + signal(SIGSEGV, sigsegv_cb); + alarm(10); + for(;;) + { + rdtsc(); + if (should_segv) + { + fprintf(stderr, "FATAL ERROR, rdtsc() succeeded while disabled\n"); + exit(0); + } + if (prctl(PR_SET_TSC, PR_TSC_SIGSEGV) < 0) + { + perror("prctl"); + exit(0); + } + should_segv = 1; + } +} + + +int main(void) +{ + int n_tasks = 100, i; + + fprintf(stderr, "[No further output means we're all right]\n"); + + for (i=0; i +#include +#include +#include +#include + + +#include +#include + +/* Get/set the process' ability to use the timestamp counter instruction */ +#ifndef PR_GET_TSC +#define PR_GET_TSC 25 +#define PR_SET_TSC 26 +# define PR_TSC_ENABLE 1 /* allow the use of the timestamp counter */ +# define PR_TSC_SIGSEGV 2 /* throw a SIGSEGV instead of reading the TSC */ +#endif + +const char *tsc_names[] = +{ + [0] = "[not set]", + [PR_TSC_ENABLE] = "PR_TSC_ENABLE", + [PR_TSC_SIGSEGV] = "PR_TSC_SIGSEGV", +}; + +static uint64_t rdtsc(void) +{ +uint32_t lo, hi; +/* We cannot use "=A", since this would use %rax on x86_64 */ +__asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi)); +return (uint64_t)hi << 32 | lo; +} + +static void sigsegv_cb(int sig) +{ + int tsc_val = 0; + + printf("[ SIG_SEGV ]\n"); + printf("prctl(PR_GET_TSC, &tsc_val); "); + fflush(stdout); + + if ( prctl(PR_GET_TSC, &tsc_val) == -1) + perror("prctl"); + + printf("tsc_val == %s\n", tsc_names[tsc_val]); + printf("prctl(PR_SET_TSC, PR_TSC_ENABLE)\n"); + fflush(stdout); + if ( prctl(PR_SET_TSC, PR_TSC_ENABLE) == -1) + perror("prctl"); + + printf("rdtsc() == "); +} + +int main(void) +{ + int tsc_val = 0; + + signal(SIGSEGV, sigsegv_cb); + + printf("rdtsc() == %llu\n", (unsigned long long)rdtsc()); + printf("prctl(PR_GET_TSC, &tsc_val); "); + fflush(stdout); + + if ( prctl(PR_GET_TSC, &tsc_val) == -1) + perror("prctl"); + + printf("tsc_val == %s\n", tsc_names[tsc_val]); + printf("rdtsc() == %llu\n", (unsigned long long)rdtsc()); + printf("prctl(PR_SET_TSC, PR_TSC_ENABLE)\n"); + fflush(stdout); + + if ( prctl(PR_SET_TSC, PR_TSC_ENABLE) == -1) + perror("prctl"); + + printf("rdtsc() == %llu\n", (unsigned long long)rdtsc()); + printf("prctl(PR_SET_TSC, PR_TSC_SIGSEGV)\n"); + fflush(stdout); + + if ( prctl(PR_SET_TSC, PR_TSC_SIGSEGV) == -1) + perror("prctl"); + + printf("rdtsc() == "); + fflush(stdout); + printf("%llu\n", (unsigned long long)rdtsc()); + fflush(stdout); + + exit(EXIT_SUCCESS); +} + diff --git a/tools/testing/selftests/prctl/set-anon-vma-name-test.c b/tools/testing/selftests/prctl/set-anon-vma-name-test.c new file mode 100644 index 000000000..ac6721b18 --- /dev/null +++ b/tools/testing/selftests/prctl/set-anon-vma-name-test.c @@ -0,0 +1,104 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * This test covers the anonymous VMA naming functionality through prctl calls + */ + +#include +#include +#include +#include +#include +#include + +#include "kselftest_harness.h" + +#define AREA_SIZE 1024 + +#define GOOD_NAME "goodname" +#define BAD_NAME "badname\1" + +#ifndef PR_SET_VMA +#define PR_SET_VMA 0x53564d41 +#define PR_SET_VMA_ANON_NAME 0 +#endif + + +int rename_vma(unsigned long addr, unsigned long size, char *name) +{ + int res; + + res = prctl(PR_SET_VMA, PR_SET_VMA_ANON_NAME, addr, size, name); + if (res < 0) + return -errno; + return res; +} + +int was_renaming_successful(char *target_name, unsigned long ptr) +{ + FILE *maps_file; + + char line_buf[512], name[128], mode[8]; + unsigned long start_addr, end_addr, offset; + unsigned int major_id, minor_id, node_id; + + char target_buf[128]; + int res = 0, sscanf_res; + + // The entry name in maps will be in format [anon:] + sprintf(target_buf, "[anon:%s]", target_name); + maps_file = fopen("/proc/self/maps", "r"); + if (!maps_file) { + printf("## /proc/self/maps file opening error\n"); + return 0; + } + + // Parse the maps file to find the entry we renamed + while (fgets(line_buf, sizeof(line_buf), maps_file)) { + sscanf_res = sscanf(line_buf, "%lx-%lx %7s %lx %u:%u %u %s", &start_addr, + &end_addr, mode, &offset, &major_id, + &minor_id, &node_id, name); + if (sscanf_res == EOF) { + res = 0; + printf("## EOF while parsing the maps file\n"); + break; + } + if (!strcmp(name, target_buf) && start_addr == ptr) { + res = 1; + break; + } + } + fclose(maps_file); + return res; +} + +FIXTURE(vma) { + void *ptr_anon, *ptr_not_anon; +}; + +FIXTURE_SETUP(vma) { + self->ptr_anon = mmap(NULL, AREA_SIZE, PROT_READ | PROT_WRITE, + MAP_PRIVATE | MAP_ANONYMOUS, 0, 0); + ASSERT_NE(self->ptr_anon, NULL); + self->ptr_not_anon = mmap(NULL, AREA_SIZE, PROT_READ | PROT_WRITE, + MAP_PRIVATE, 0, 0); + ASSERT_NE(self->ptr_not_anon, NULL); +} + +FIXTURE_TEARDOWN(vma) { + munmap(self->ptr_anon, AREA_SIZE); + munmap(self->ptr_not_anon, AREA_SIZE); +} + +TEST_F(vma, renaming) { + TH_LOG("Try to rename the VMA with correct parameters"); + EXPECT_GE(rename_vma((unsigned long)self->ptr_anon, AREA_SIZE, GOOD_NAME), 0); + EXPECT_TRUE(was_renaming_successful(GOOD_NAME, (unsigned long)self->ptr_anon)); + + TH_LOG("Try to pass invalid name (with non-printable character \\1) to rename the VMA"); + EXPECT_EQ(rename_vma((unsigned long)self->ptr_anon, AREA_SIZE, BAD_NAME), -EINVAL); + + TH_LOG("Try to rename non-anonymous VMA"); + EXPECT_EQ(rename_vma((unsigned long) self->ptr_not_anon, AREA_SIZE, GOOD_NAME), -EINVAL); +} + +TEST_HARNESS_MAIN diff --git a/tools/testing/selftests/prctl/set-process-name.c b/tools/testing/selftests/prctl/set-process-name.c new file mode 100644 index 000000000..3f7b146d3 --- /dev/null +++ b/tools/testing/selftests/prctl/set-process-name.c @@ -0,0 +1,94 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * This test covers the PR_SET_NAME functionality of prctl calls + */ + +#include +#include +#include + +#include "kselftest_harness.h" + +#define CHANGE_NAME "changename" +#define EMPTY_NAME "" +#define TASK_COMM_LEN 16 +#define MAX_PATH_LEN 50 + +int set_name(char *name) +{ + int res; + + res = prctl(PR_SET_NAME, name, NULL, NULL, NULL); + + if (res < 0) + return -errno; + return res; +} + +int check_is_name_correct(char *check_name) +{ + char name[TASK_COMM_LEN]; + int res; + + res = prctl(PR_GET_NAME, name, NULL, NULL, NULL); + + if (res < 0) + return -errno; + + return !strcmp(name, check_name); +} + +int check_null_pointer(char *check_name) +{ + char *name = NULL; + int res; + + res = prctl(PR_GET_NAME, name, NULL, NULL, NULL); + + return res; +} + +int check_name(void) +{ + + int pid; + + pid = getpid(); + FILE *fptr = NULL; + char path[MAX_PATH_LEN] = {}; + char name[TASK_COMM_LEN] = {}; + char output[TASK_COMM_LEN] = {}; + int j; + + j = snprintf(path, MAX_PATH_LEN, "/proc/self/task/%d/comm", pid); + fptr = fopen(path, "r"); + if (!fptr) + return -EIO; + + fscanf(fptr, "%s", output); + if (ferror(fptr)) + return -EIO; + + int res = prctl(PR_GET_NAME, name, NULL, NULL, NULL); + + if (res < 0) + return -errno; + + return !strcmp(output, name); +} + +TEST(rename_process) { + + EXPECT_GE(set_name(CHANGE_NAME), 0); + EXPECT_TRUE(check_is_name_correct(CHANGE_NAME)); + + EXPECT_GE(set_name(EMPTY_NAME), 0); + EXPECT_TRUE(check_is_name_correct(EMPTY_NAME)); + + EXPECT_GE(set_name(CHANGE_NAME), 0); + EXPECT_LT(check_null_pointer(CHANGE_NAME), 0); + + EXPECT_TRUE(check_name()); +} + +TEST_HARNESS_MAIN -- cgit v1.3.1