diff options
Diffstat (limited to 'tools/testing/selftests')
67 files changed, 2672 insertions, 335 deletions
diff --git a/tools/testing/selftests/Makefile b/tools/testing/selftests/Makefile index e03bc15ce731..9018f45d631d 100644 --- a/tools/testing/selftests/Makefile +++ b/tools/testing/selftests/Makefile @@ -32,6 +32,7 @@ TARGETS += lkdtm TARGETS += membarrier TARGETS += memfd TARGETS += memory-hotplug +TARGETS += mincore TARGETS += mount TARGETS += mqueue TARGETS += net diff --git a/tools/testing/selftests/cgroup/.gitignore b/tools/testing/selftests/cgroup/.gitignore index aa6de65b0838..84cfcabea838 100644 --- a/tools/testing/selftests/cgroup/.gitignore +++ b/tools/testing/selftests/cgroup/.gitignore @@ -2,3 +2,4 @@ test_memcontrol test_core test_freezer +test_kmem
\ No newline at end of file diff --git a/tools/testing/selftests/cgroup/Makefile b/tools/testing/selftests/cgroup/Makefile index 967f268fde74..f027d933595b 100644 --- a/tools/testing/selftests/cgroup/Makefile +++ b/tools/testing/selftests/cgroup/Makefile @@ -6,11 +6,13 @@ all: TEST_FILES := with_stress.sh TEST_PROGS := test_stress.sh TEST_GEN_PROGS = test_memcontrol +TEST_GEN_PROGS += test_kmem TEST_GEN_PROGS += test_core TEST_GEN_PROGS += test_freezer include ../lib.mk $(OUTPUT)/test_memcontrol: cgroup_util.c ../clone3/clone3_selftests.h +$(OUTPUT)/test_kmem: cgroup_util.c ../clone3/clone3_selftests.h $(OUTPUT)/test_core: cgroup_util.c ../clone3/clone3_selftests.h $(OUTPUT)/test_freezer: cgroup_util.c ../clone3/clone3_selftests.h diff --git a/tools/testing/selftests/cgroup/cgroup_util.c b/tools/testing/selftests/cgroup/cgroup_util.c index 8a637ca7d73a..05853b0b8831 100644 --- a/tools/testing/selftests/cgroup/cgroup_util.c +++ b/tools/testing/selftests/cgroup/cgroup_util.c @@ -106,7 +106,7 @@ int cg_read_strcmp(const char *cgroup, const char *control, /* Handle the case of comparing against empty string */ if (!expected) - size = 32; + return -1; else size = strlen(expected) + 1; diff --git a/tools/testing/selftests/cgroup/test_kmem.c b/tools/testing/selftests/cgroup/test_kmem.c new file mode 100644 index 000000000000..0941aa16157e --- /dev/null +++ b/tools/testing/selftests/cgroup/test_kmem.c @@ -0,0 +1,450 @@ +// SPDX-License-Identifier: GPL-2.0 +#define _GNU_SOURCE + +#include <linux/limits.h> +#include <fcntl.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <sys/stat.h> +#include <sys/types.h> +#include <unistd.h> +#include <sys/wait.h> +#include <errno.h> +#include <sys/sysinfo.h> +#include <pthread.h> + +#include "../kselftest.h" +#include "cgroup_util.h" + + +/* + * Memory cgroup charging and vmstat data aggregation is performed using + * percpu batches 32 pages big (look at MEMCG_CHARGE_BATCH). So the maximum + * discrepancy between charge and vmstat entries is number of cpus multiplied + * by 32 pages multiplied by 2. + */ +#define MAX_VMSTAT_ERROR (4096 * 32 * 2 * get_nprocs()) + + +static int alloc_dcache(const char *cgroup, void *arg) +{ + unsigned long i; + struct stat st; + char buf[128]; + + for (i = 0; i < (unsigned long)arg; i++) { + snprintf(buf, sizeof(buf), + "/something-non-existent-with-a-long-name-%64lu-%d", + i, getpid()); + stat(buf, &st); + } + + return 0; +} + +/* + * This test allocates 100000 of negative dentries with long names. + * Then it checks that "slab" in memory.stat is larger than 1M. + * Then it sets memory.high to 1M and checks that at least 1/2 + * of slab memory has been reclaimed. + */ +static int test_kmem_basic(const char *root) +{ + int ret = KSFT_FAIL; + char *cg = NULL; + long slab0, slab1, current; + + cg = cg_name(root, "kmem_basic_test"); + if (!cg) + goto cleanup; + + if (cg_create(cg)) + goto cleanup; + + if (cg_run(cg, alloc_dcache, (void *)100000)) + goto cleanup; + + slab0 = cg_read_key_long(cg, "memory.stat", "slab "); + if (slab0 < (1 << 20)) + goto cleanup; + + cg_write(cg, "memory.high", "1M"); + slab1 = cg_read_key_long(cg, "memory.stat", "slab "); + if (slab1 <= 0) + goto cleanup; + + current = cg_read_long(cg, "memory.current"); + if (current <= 0) + goto cleanup; + + if (slab1 < slab0 / 2 && current < slab0 / 2) + ret = KSFT_PASS; +cleanup: + cg_destroy(cg); + free(cg); + + return ret; +} + +static void *alloc_kmem_fn(void *arg) +{ + alloc_dcache(NULL, (void *)100); + return NULL; +} + +static int alloc_kmem_smp(const char *cgroup, void *arg) +{ + int nr_threads = 2 * get_nprocs(); + pthread_t *tinfo; + unsigned long i; + int ret = -1; + + tinfo = calloc(nr_threads, sizeof(pthread_t)); + if (tinfo == NULL) + return -1; + + for (i = 0; i < nr_threads; i++) { + if (pthread_create(&tinfo[i], NULL, &alloc_kmem_fn, + (void *)i)) { + free(tinfo); + return -1; + } + } + + for (i = 0; i < nr_threads; i++) { + ret = pthread_join(tinfo[i], NULL); + if (ret) + break; + } + + free(tinfo); + return ret; +} + +static int cg_run_in_subcgroups(const char *parent, + int (*fn)(const char *cgroup, void *arg), + void *arg, int times) +{ + char *child; + int i; + + for (i = 0; i < times; i++) { + child = cg_name_indexed(parent, "child", i); + if (!child) + return -1; + + if (cg_create(child)) { + cg_destroy(child); + free(child); + return -1; + } + + if (cg_run(child, fn, NULL)) { + cg_destroy(child); + free(child); + return -1; + } + + cg_destroy(child); + free(child); + } + + return 0; +} + +/* + * The test creates and destroys a large number of cgroups. In each cgroup it + * allocates some slab memory (mostly negative dentries) using 2 * NR_CPUS + * threads. Then it checks the sanity of numbers on the parent level: + * the total size of the cgroups should be roughly equal to + * anon + file + slab + kernel_stack. + */ +static int test_kmem_memcg_deletion(const char *root) +{ + long current, slab, anon, file, kernel_stack, sum; + int ret = KSFT_FAIL; + char *parent; + + parent = cg_name(root, "kmem_memcg_deletion_test"); + if (!parent) + goto cleanup; + + if (cg_create(parent)) + goto cleanup; + + if (cg_write(parent, "cgroup.subtree_control", "+memory")) + goto cleanup; + + if (cg_run_in_subcgroups(parent, alloc_kmem_smp, NULL, 100)) + goto cleanup; + + current = cg_read_long(parent, "memory.current"); + slab = cg_read_key_long(parent, "memory.stat", "slab "); + anon = cg_read_key_long(parent, "memory.stat", "anon "); + file = cg_read_key_long(parent, "memory.stat", "file "); + kernel_stack = cg_read_key_long(parent, "memory.stat", "kernel_stack "); + if (current < 0 || slab < 0 || anon < 0 || file < 0 || + kernel_stack < 0) + goto cleanup; + + sum = slab + anon + file + kernel_stack; + if (abs(sum - current) < MAX_VMSTAT_ERROR) { + ret = KSFT_PASS; + } else { + printf("memory.current = %ld\n", current); + printf("slab + anon + file + kernel_stack = %ld\n", sum); + printf("slab = %ld\n", slab); + printf("anon = %ld\n", anon); + printf("file = %ld\n", file); + printf("kernel_stack = %ld\n", kernel_stack); + } + +cleanup: + cg_destroy(parent); + free(parent); + + return ret; +} + +/* + * The test reads the entire /proc/kpagecgroup. If the operation went + * successfully (and the kernel didn't panic), the test is treated as passed. + */ +static int test_kmem_proc_kpagecgroup(const char *root) +{ + unsigned long buf[128]; + int ret = KSFT_FAIL; + ssize_t len; + int fd; + + fd = open("/proc/kpagecgroup", O_RDONLY); + if (fd < 0) + return ret; + + do { + len = read(fd, buf, sizeof(buf)); + } while (len > 0); + + if (len == 0) + ret = KSFT_PASS; + + close(fd); + return ret; +} + +static void *pthread_wait_fn(void *arg) +{ + sleep(100); + return NULL; +} + +static int spawn_1000_threads(const char *cgroup, void *arg) +{ + int nr_threads = 1000; + pthread_t *tinfo; + unsigned long i; + long stack; + int ret = -1; + + tinfo = calloc(nr_threads, sizeof(pthread_t)); + if (tinfo == NULL) + return -1; + + for (i = 0; i < nr_threads; i++) { + if (pthread_create(&tinfo[i], NULL, &pthread_wait_fn, + (void *)i)) { + free(tinfo); + return(-1); + } + } + + stack = cg_read_key_long(cgroup, "memory.stat", "kernel_stack "); + if (stack >= 4096 * 1000) + ret = 0; + + free(tinfo); + return ret; +} + +/* + * The test spawns a process, which spawns 1000 threads. Then it checks + * that memory.stat's kernel_stack is at least 1000 pages large. + */ +static int test_kmem_kernel_stacks(const char *root) +{ + int ret = KSFT_FAIL; + char *cg = NULL; + + cg = cg_name(root, "kmem_kernel_stacks_test"); + if (!cg) + goto cleanup; + + if (cg_create(cg)) + goto cleanup; + + if (cg_run(cg, spawn_1000_threads, NULL)) + goto cleanup; + + ret = KSFT_PASS; +cleanup: + cg_destroy(cg); + free(cg); + + return ret; +} + +/* + * This test sequentionally creates 30 child cgroups, allocates some + * kernel memory in each of them, and deletes them. Then it checks + * that the number of dying cgroups on the parent level is 0. + */ +static int test_kmem_dead_cgroups(const char *root) +{ + int ret = KSFT_FAIL; + char *parent; + long dead; + int i; + + parent = cg_name(root, "kmem_dead_cgroups_test"); + if (!parent) + goto cleanup; + + if (cg_create(parent)) + goto cleanup; + + if (cg_write(parent, "cgroup.subtree_control", "+memory")) + goto cleanup; + + if (cg_run_in_subcgroups(parent, alloc_dcache, (void *)100, 30)) + goto cleanup; + + for (i = 0; i < 5; i++) { + dead = cg_read_key_long(parent, "cgroup.stat", + "nr_dying_descendants "); + if (dead == 0) { + ret = KSFT_PASS; + break; + } + /* + * Reclaiming cgroups might take some time, + * let's wait a bit and repeat. + */ + sleep(1); + } + +cleanup: + cg_destroy(parent); + free(parent); + + return ret; +} + +/* + * This test creates a sub-tree with 1000 memory cgroups. + * Then it checks that the memory.current on the parent level + * is greater than 0 and approximates matches the percpu value + * from memory.stat. + */ +static int test_percpu_basic(const char *root) +{ + int ret = KSFT_FAIL; + char *parent, *child; + long current, percpu; + int i; + + parent = cg_name(root, "percpu_basic_test"); + if (!parent) + goto cleanup; + + if (cg_create(parent)) + goto cleanup; + + if (cg_write(parent, "cgroup.subtree_control", "+memory")) + goto cleanup; + + for (i = 0; i < 1000; i++) { + child = cg_name_indexed(parent, "child", i); + if (!child) + return -1; + + if (cg_create(child)) + goto cleanup_children; + + free(child); + } + + current = cg_read_long(parent, "memory.current"); + percpu = cg_read_key_long(parent, "memory.stat", "percpu "); + + if (current > 0 && percpu > 0 && abs(current - percpu) < + MAX_VMSTAT_ERROR) + ret = KSFT_PASS; + else + printf("memory.current %ld\npercpu %ld\n", + current, percpu); + +cleanup_children: + for (i = 0; i < 1000; i++) { + child = cg_name_indexed(parent, "child", i); + cg_destroy(child); + free(child); + } + +cleanup: + cg_destroy(parent); + free(parent); + + return ret; +} + +#define T(x) { x, #x } +struct kmem_test { + int (*fn)(const char *root); + const char *name; +} tests[] = { + T(test_kmem_basic), + T(test_kmem_memcg_deletion), + T(test_kmem_proc_kpagecgroup), + T(test_kmem_kernel_stacks), + T(test_kmem_dead_cgroups), + T(test_percpu_basic), +}; +#undef T + +int main(int argc, char **argv) +{ + char root[PATH_MAX]; + int i, ret = EXIT_SUCCESS; + + if (cg_find_unified_root(root, sizeof(root))) + ksft_exit_skip("cgroup v2 isn't mounted\n"); + + /* + * Check that memory controller is available: + * memory is listed in cgroup.controllers + */ + if (cg_read_strstr(root, "cgroup.controllers", "memory")) + ksft_exit_skip("memory controller isn't available\n"); + + if (cg_read_strstr(root, "cgroup.subtree_control", "memory")) + if (cg_write(root, "cgroup.subtree_control", "+memory")) + ksft_exit_skip("Failed to set memory controller\n"); + + for (i = 0; i < ARRAY_SIZE(tests); i++) { + switch (tests[i].fn(root)) { + case KSFT_PASS: + ksft_test_result_pass("%s\n", tests[i].name); + break; + case KSFT_SKIP: + ksft_test_result_skip("%s\n", tests[i].name); + break; + default: + ret = EXIT_FAILURE; + ksft_test_result_fail("%s\n", tests[i].name); + break; + } + } + + return ret; +} diff --git a/tools/testing/selftests/kmod/kmod.sh b/tools/testing/selftests/kmod/kmod.sh index ea2147248ebe..afd42387e8b2 100755 --- a/tools/testing/selftests/kmod/kmod.sh +++ b/tools/testing/selftests/kmod/kmod.sh @@ -343,7 +343,7 @@ kmod_test_0001_driver() kmod_defaults_driver config_num_threads 1 - printf '\000' >"$DIR"/config_test_driver + printf $NAME >"$DIR"/config_test_driver config_trigger ${FUNCNAME[0]} config_expect_result ${FUNCNAME[0]} MODULE_NOT_FOUND } @@ -354,7 +354,7 @@ kmod_test_0001_fs() kmod_defaults_fs config_num_threads 1 - printf '\000' >"$DIR"/config_test_fs + printf $NAME >"$DIR"/config_test_fs config_trigger ${FUNCNAME[0]} config_expect_result ${FUNCNAME[0]} -EINVAL } diff --git a/tools/testing/selftests/livepatch/README b/tools/testing/selftests/livepatch/README index 621d325425c2..0942dd5826f8 100644 --- a/tools/testing/selftests/livepatch/README +++ b/tools/testing/selftests/livepatch/README @@ -6,8 +6,8 @@ This is a small set of sanity tests for the kernel livepatching. The test suite loads and unloads several test kernel modules to verify livepatch behavior. Debug information is logged to the kernel's message -buffer and parsed for expected messages. (Note: the tests will clear -the message buffer between individual tests.) +buffer and parsed for expected messages. (Note: the tests will compare +the message buffer for only the duration of each individual test.) Config @@ -35,9 +35,9 @@ Adding tests ------------ See the common functions.sh file for the existing collection of utility -functions, most importantly setup_config() and check_result(). The -latter function greps the kernel's ring buffer for "livepatch:" and -"test_klp" strings, so tests be sure to include one of those strings for -result comparison. Other utility functions include general module -loading and livepatch loading helpers (waiting for patch transitions, -sysfs entries, etc.) +functions, most importantly setup_config(), start_test() and +check_result(). The latter function greps the kernel's ring buffer for +"livepatch:" and "test_klp" strings, so tests be sure to include one of +those strings for result comparison. Other utility functions include +general module loading and livepatch loading helpers (waiting for patch +transitions, sysfs entries, etc.) diff --git a/tools/testing/selftests/livepatch/functions.sh b/tools/testing/selftests/livepatch/functions.sh index 2aab9791791d..1aba83c87ad3 100644 --- a/tools/testing/selftests/livepatch/functions.sh +++ b/tools/testing/selftests/livepatch/functions.sh @@ -41,6 +41,17 @@ function die() { exit 1 } +# save existing dmesg so we can detect new content +function save_dmesg() { + SAVED_DMESG=$(mktemp --tmpdir -t klp-dmesg-XXXXXX) + dmesg > "$SAVED_DMESG" +} + +# cleanup temporary dmesg file from save_dmesg() +function cleanup_dmesg_file() { + rm -f "$SAVED_DMESG" +} + function push_config() { DYNAMIC_DEBUG=$(grep '^kernel/livepatch' /sys/kernel/debug/dynamic_debug/control | \ awk -F'[: ]' '{print "file " $1 " line " $2 " " $4}') @@ -64,10 +75,16 @@ function set_dynamic_debug() { } function set_ftrace_enabled() { - result=$(sysctl kernel.ftrace_enabled="$1" 2>&1 | paste --serial --delimiters=' ') + result=$(sysctl -q kernel.ftrace_enabled="$1" 2>&1 && \ + sysctl kernel.ftrace_enabled 2>&1) echo "livepatch: $result" > /dev/kmsg } +function cleanup() { + pop_config + cleanup_dmesg_file +} + # setup_config - save the current config and set a script exit trap that # restores the original config. Setup the dynamic debug # for verbose livepatching output and turn on @@ -77,7 +94,7 @@ function setup_config() { push_config set_dynamic_debug set_ftrace_enabled 1 - trap pop_config EXIT INT TERM HUP + trap cleanup EXIT INT TERM HUP } # loop_until(cmd) - loop a command until it is successful or $MAX_RETRIES, @@ -243,13 +260,28 @@ function set_pre_patch_ret { die "failed to set pre_patch_ret parameter for $mod module" } +function start_test { + local test="$1" + + save_dmesg + echo -n "TEST: $test ... " + log "===== TEST: $test =====" +} + # check_result() - verify dmesg output # TODO - better filter, out of order msgs, etc? function check_result { local expect="$*" local result - result=$(dmesg | grep -v 'tainting' | grep -e 'livepatch:' -e 'test_klp' | sed 's/^\[[ 0-9.]*\] //') + # Note: when comparing dmesg output, the kernel log timestamps + # help differentiate repeated testing runs. Remove them with a + # post-comparison sed filter. + + result=$(dmesg | comm -13 "$SAVED_DMESG" - | \ + grep -e 'livepatch:' -e 'test_klp' | \ + grep -v '\(tainting\|taints\) kernel' | \ + sed 's/^\[[ 0-9.]*\] //') if [[ "$expect" == "$result" ]] ; then echo "ok" @@ -257,4 +289,6 @@ function check_result { echo -e "not ok\n\n$(diff -upr --label expected --label result <(echo "$expect") <(echo "$result"))\n" die "livepatch kselftest(s) failed" fi + + cleanup_dmesg_file } diff --git a/tools/testing/selftests/livepatch/test-callbacks.sh b/tools/testing/selftests/livepatch/test-callbacks.sh index a35289b13c9c..90b26dbb2626 100755 --- a/tools/testing/selftests/livepatch/test-callbacks.sh +++ b/tools/testing/selftests/livepatch/test-callbacks.sh @@ -12,8 +12,6 @@ MOD_TARGET_BUSY=test_klp_callbacks_busy setup_config -# TEST: target module before livepatch -# # Test a combination of loading a kernel module and a livepatch that # patches a function in the first module. Load the target module # before the livepatch module. Unload them in the same order. @@ -28,8 +26,7 @@ setup_config # unpatching transition starts. klp_objects are reverted, post-patch # callbacks execute and the transition completes. -echo -n "TEST: target module before livepatch ... " -dmesg -C +start_test "target module before livepatch" load_mod $MOD_TARGET load_lp $MOD_LIVEPATCH @@ -63,8 +60,6 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete $MOD_TARGET: ${MOD_TARGET}_exit" -# TEST: module_coming notifier -# # This test is similar to the previous test, but (un)load the livepatch # module before the target kernel module. This tests the livepatch # core's module_coming handler. @@ -78,8 +73,7 @@ $MOD_TARGET: ${MOD_TARGET}_exit" # - On livepatch disable, all currently loaded klp_objects' (vmlinux and # $MOD_TARGET) pre/post-unpatch callbacks are executed. -echo -n "TEST: module_coming notifier ... " -dmesg -C +start_test "module_coming notifier" load_lp $MOD_LIVEPATCH load_mod $MOD_TARGET @@ -114,8 +108,6 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete $MOD_TARGET: ${MOD_TARGET}_exit" -# TEST: module_going notifier -# # Test loading the livepatch after a targeted kernel module, then unload # the kernel module before disabling the livepatch. This tests the # livepatch core's module_going handler. @@ -129,8 +121,7 @@ $MOD_TARGET: ${MOD_TARGET}_exit" # - When the livepatch is disabled, pre and post-unpatch callbacks are # run for the remaining klp_object, vmlinux. -echo -n "TEST: module_going notifier ... " -dmesg -C +start_test "module_going notifier" load_mod $MOD_TARGET load_lp $MOD_LIVEPATCH @@ -165,8 +156,6 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete % rmmod $MOD_LIVEPATCH" -# TEST: module_coming and module_going notifiers -# # This test is similar to the previous test, however the livepatch is # loaded first. This tests the livepatch core's module_coming and # module_going handlers. @@ -180,8 +169,7 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete # from the $MOD_TARGET klp_object. As such, only pre and # post-unpatch callbacks are executed when this occurs. -echo -n "TEST: module_coming and module_going notifiers ... " -dmesg -C +start_test "module_coming and module_going notifiers" load_lp $MOD_LIVEPATCH load_mod $MOD_TARGET @@ -217,8 +205,6 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete % rmmod $MOD_LIVEPATCH" -# TEST: target module not present -# # A simple test of loading a livepatch without one of its patch target # klp_objects ever loaded ($MOD_TARGET). # @@ -227,8 +213,7 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete # - As expected, only pre/post-(un)patch handlers are executed for # vmlinux. -echo -n "TEST: target module not present ... " -dmesg -C +start_test "target module not present" load_lp $MOD_LIVEPATCH disable_lp $MOD_LIVEPATCH @@ -252,8 +237,6 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete % rmmod $MOD_LIVEPATCH" -# TEST: pre-patch callback -ENODEV -# # Test a scenario where a vmlinux pre-patch callback returns a non-zero # status (ie, failure). # @@ -265,8 +248,7 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete # The result is that the insmod command refuses to load the livepatch # module. -echo -n "TEST: pre-patch callback -ENODEV ... " -dmesg -C +start_test "pre-patch callback -ENODEV" load_mod $MOD_TARGET load_failing_mod $MOD_LIVEPATCH pre_patch_ret=-19 @@ -288,8 +270,6 @@ modprobe: ERROR: could not insert '$MOD_LIVEPATCH': No such device $MOD_TARGET: ${MOD_TARGET}_exit" -# TEST: module_coming + pre-patch callback -ENODEV -# # Similar to the previous test, setup a livepatch such that its vmlinux # pre-patch callback returns success. However, when a targeted kernel # module is later loaded, have the livepatch return a failing status @@ -307,8 +287,7 @@ $MOD_TARGET: ${MOD_TARGET}_exit" # # - Pre/post-unpatch callbacks are run for the vmlinux klp_object. -echo -n "TEST: module_coming + pre-patch callback -ENODEV ... " -dmesg -C +start_test "module_coming + pre-patch callback -ENODEV" load_lp $MOD_LIVEPATCH set_pre_patch_ret $MOD_LIVEPATCH -19 @@ -341,8 +320,6 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete % rmmod $MOD_LIVEPATCH" -# TEST: multiple target modules -# # Test loading multiple targeted kernel modules. This test-case is # mainly for comparing with the next test-case. # @@ -353,12 +330,9 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete # module. Post-patch callbacks are executed and the transition # completes quickly. -echo -n "TEST: multiple target modules ... " -dmesg -C +start_test "multiple target modules" -load_mod $MOD_TARGET_BUSY sleep_secs=0 -# give $MOD_TARGET_BUSY::busymod_work_func() a chance to run -sleep 5 +load_mod $MOD_TARGET_BUSY block_transition=N load_lp $MOD_LIVEPATCH load_mod $MOD_TARGET unload_mod $MOD_TARGET @@ -366,9 +340,9 @@ disable_lp $MOD_LIVEPATCH unload_lp $MOD_LIVEPATCH unload_mod $MOD_TARGET_BUSY -check_result "% modprobe $MOD_TARGET_BUSY sleep_secs=0 +check_result "% modprobe $MOD_TARGET_BUSY block_transition=N $MOD_TARGET_BUSY: ${MOD_TARGET_BUSY}_init -$MOD_TARGET_BUSY: busymod_work_func, sleeping 0 seconds ... +$MOD_TARGET_BUSY: busymod_work_func enter $MOD_TARGET_BUSY: busymod_work_func exit % modprobe $MOD_LIVEPATCH livepatch: enabling patch '$MOD_LIVEPATCH' @@ -404,11 +378,8 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete $MOD_TARGET_BUSY: ${MOD_TARGET_BUSY}_exit" - -# TEST: busy target module -# # A similar test as the previous one, but force the "busy" kernel module -# to do longer work. +# to block the livepatch transition. # # The livepatching core will refuse to patch a task that is currently # executing a to-be-patched function -- the consistency model stalls the @@ -417,8 +388,7 @@ $MOD_TARGET_BUSY: ${MOD_TARGET_BUSY}_exit" # function for a long time. Meanwhile, load and unload other target # kernel modules while the livepatch transition is in progress. # -# - Load the "busy" kernel module, this time make it do 10 seconds worth -# of work. +# - Load the "busy" kernel module, this time make its work function loop # # - Meanwhile, the livepatch is loaded. Notice that the patch # transition does not complete as the targeted "busy" module is @@ -435,23 +405,25 @@ $MOD_TARGET_BUSY: ${MOD_TARGET_BUSY}_exit" # klp_object's post-patch callbacks executed, the remaining # klp_object's pre-unpatch callbacks are skipped. -echo -n "TEST: busy target module ... " -dmesg -C +start_test "busy target module" -load_mod $MOD_TARGET_BUSY sleep_secs=10 +load_mod $MOD_TARGET_BUSY block_transition=Y load_lp_nowait $MOD_LIVEPATCH -# Don't wait for transition, load $MOD_TARGET while the transition -# is still stalled in $MOD_TARGET_BUSY::busymod_work_func() -sleep 5 + +# Wait until the livepatch reports in-transition state, i.e. that it's +# stalled on $MOD_TARGET_BUSY::busymod_work_func() +loop_until 'grep -q '^1$' /sys/kernel/livepatch/$MOD_LIVEPATCH/transition' || + die "failed to stall transition" + load_mod $MOD_TARGET unload_mod $MOD_TARGET disable_lp $MOD_LIVEPATCH unload_lp $MOD_LIVEPATCH unload_mod $MOD_TARGET_BUSY -check_result "% modprobe $MOD_TARGET_BUSY sleep_secs=10 +check_result "% modprobe $MOD_TARGET_BUSY block_transition=Y $MOD_TARGET_BUSY: ${MOD_TARGET_BUSY}_init -$MOD_TARGET_BUSY: busymod_work_func, sleeping 10 seconds ... +$MOD_TARGET_BUSY: busymod_work_func enter % modprobe $MOD_LIVEPATCH livepatch: enabling patch '$MOD_LIVEPATCH' livepatch: '$MOD_LIVEPATCH': initializing patching transition @@ -479,8 +451,6 @@ $MOD_TARGET_BUSY: busymod_work_func exit $MOD_TARGET_BUSY: ${MOD_TARGET_BUSY}_exit" -# TEST: multiple livepatches -# # Test loading multiple livepatches. This test-case is mainly for comparing # with the next test-case. # @@ -488,8 +458,7 @@ $MOD_TARGET_BUSY: ${MOD_TARGET_BUSY}_exit" # execute as each patch progresses through its (un)patching # transition. -echo -n "TEST: multiple livepatches ... " -dmesg -C +start_test "multiple livepatches" load_lp $MOD_LIVEPATCH load_lp $MOD_LIVEPATCH2 @@ -532,8 +501,6 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete % rmmod $MOD_LIVEPATCH" -# TEST: atomic replace -# # Load multiple livepatches, but the second as an 'atomic-replace' # patch. When the latter loads, the original livepatch should be # disabled and *none* of its pre/post-unpatch callbacks executed. On @@ -548,8 +515,7 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete # - Once the atomic replace module is loaded, only its pre and post # unpatch callbacks are executed. -echo -n "TEST: atomic replace ... " -dmesg -C +start_test "atomic replace" load_lp $MOD_LIVEPATCH load_lp $MOD_LIVEPATCH2 replace=1 diff --git a/tools/testing/selftests/livepatch/test-ftrace.sh b/tools/testing/selftests/livepatch/test-ftrace.sh index e2a76887f40a..552e165512f4 100755 --- a/tools/testing/selftests/livepatch/test-ftrace.sh +++ b/tools/testing/selftests/livepatch/test-ftrace.sh @@ -9,13 +9,11 @@ MOD_LIVEPATCH=test_klp_livepatch setup_config -# TEST: livepatch interaction with ftrace_enabled sysctl # - turn ftrace_enabled OFF and verify livepatches can't load # - turn ftrace_enabled ON and verify livepatch can load # - verify that ftrace_enabled can't be turned OFF while a livepatch is loaded -echo -n "TEST: livepatch interaction with ftrace_enabled sysctl ... " -dmesg -C +start_test "livepatch interaction with ftrace_enabled sysctl" set_ftrace_enabled 0 load_failing_mod $MOD_LIVEPATCH @@ -53,7 +51,7 @@ livepatch: '$MOD_LIVEPATCH': initializing patching transition livepatch: '$MOD_LIVEPATCH': starting patching transition livepatch: '$MOD_LIVEPATCH': completing patching transition livepatch: '$MOD_LIVEPATCH': patching complete -livepatch: sysctl: setting key \"kernel.ftrace_enabled\": Device or resource busy kernel.ftrace_enabled = 0 +livepatch: sysctl: setting key \"kernel.ftrace_enabled\": Device or resource busy % echo 0 > /sys/kernel/livepatch/$MOD_LIVEPATCH/enabled livepatch: '$MOD_LIVEPATCH': initializing unpatching transition livepatch: '$MOD_LIVEPATCH': starting unpatching transition diff --git a/tools/testing/selftests/livepatch/test-livepatch.sh b/tools/testing/selftests/livepatch/test-livepatch.sh index 493e3df415a1..5fe79ac34be1 100755 --- a/tools/testing/selftests/livepatch/test-livepatch.sh +++ b/tools/testing/selftests/livepatch/test-livepatch.sh @@ -10,13 +10,11 @@ MOD_REPLACE=test_klp_atomic_replace setup_config -# TEST: basic function patching # - load a livepatch that modifies the output from /proc/cmdline and # verify correct behavior # - unload the livepatch and make sure the patch was removed -echo -n "TEST: basic function patching ... " -dmesg -C +start_test "basic function patching" load_lp $MOD_LIVEPATCH @@ -47,15 +45,13 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete % rmmod $MOD_LIVEPATCH" -# TEST: multiple livepatches # - load a livepatch that modifies the output from /proc/cmdline and # verify correct behavior # - load another livepatch and verify that both livepatches are active # - unload the second livepatch and verify that the first is still active # - unload the first livepatch and verify none are active -echo -n "TEST: multiple livepatches ... " -dmesg -C +start_test "multiple livepatches" load_lp $MOD_LIVEPATCH @@ -109,7 +105,6 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete % rmmod $MOD_LIVEPATCH" -# TEST: atomic replace livepatch # - load a livepatch that modifies the output from /proc/cmdline and # verify correct behavior # - load an atomic replace livepatch and verify that only the second is active @@ -117,8 +112,7 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete # is still active # - remove the atomic replace livepatch and verify that none are active -echo -n "TEST: atomic replace livepatch ... " -dmesg -C +start_test "atomic replace livepatch" load_lp $MOD_LIVEPATCH diff --git a/tools/testing/selftests/livepatch/test-shadow-vars.sh b/tools/testing/selftests/livepatch/test-shadow-vars.sh index 1aae73299114..e04cb354f56b 100755 --- a/tools/testing/selftests/livepatch/test-shadow-vars.sh +++ b/tools/testing/selftests/livepatch/test-shadow-vars.sh @@ -9,52 +9,71 @@ MOD_TEST=test_klp_shadow_vars setup_config -# TEST: basic shadow variable API # - load a module that exercises the shadow variable API -echo -n "TEST: basic shadow variable API ... " -dmesg -C +start_test "basic shadow variable API" load_mod $MOD_TEST unload_mod $MOD_TEST check_result "% modprobe $MOD_TEST -$MOD_TEST: klp_shadow_get(obj=PTR5, id=0x1234) = PTR0 +$MOD_TEST: klp_shadow_get(obj=PTR1, id=0x1234) = PTR0 $MOD_TEST: got expected NULL result -$MOD_TEST: shadow_ctor: PTR6 -> PTR1 -$MOD_TEST: klp_shadow_alloc(obj=PTR5, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR7, ctor_data=PTR1 = PTR6 -$MOD_TEST: shadow_ctor: PTR8 -> PTR2 -$MOD_TEST: klp_shadow_alloc(obj=PTR9, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR7, ctor_data=PTR2 = PTR8 -$MOD_TEST: shadow_ctor: PTR10 -> PTR3 -$MOD_TEST: klp_shadow_alloc(obj=PTR5, id=0x1235, size=8, gfp_flags=GFP_KERNEL), ctor=PTR7, ctor_data=PTR3 = PTR10 -$MOD_TEST: klp_shadow_get(obj=PTR5, id=0x1234) = PTR6 -$MOD_TEST: got expected PTR6 -> PTR1 result +$MOD_TEST: shadow_ctor: PTR3 -> PTR2 +$MOD_TEST: klp_shadow_get_or_alloc(obj=PTR1, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR4, ctor_data=PTR2 = PTR3 +$MOD_TEST: shadow_ctor: PTR6 -> PTR5 +$MOD_TEST: klp_shadow_alloc(obj=PTR1, id=0x1235, size=8, gfp_flags=GFP_KERNEL), ctor=PTR4, ctor_data=PTR5 = PTR6 +$MOD_TEST: shadow_ctor: PTR8 -> PTR7 +$MOD_TEST: klp_shadow_alloc(obj=PTR9, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR4, ctor_data=PTR7 = PTR8 +$MOD_TEST: shadow_ctor: PTR11 -> PTR10 +$MOD_TEST: klp_shadow_alloc(obj=PTR9, id=0x1235, size=8, gfp_flags=GFP_KERNEL), ctor=PTR4, ctor_data=PTR10 = PTR11 +$MOD_TEST: shadow_ctor: PTR13 -> PTR12 +$MOD_TEST: klp_shadow_get_or_alloc(obj=PTR14, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR4, ctor_data=PTR12 = PTR13 +$MOD_TEST: shadow_ctor: PTR16 -> PTR15 +$MOD_TEST: klp_shadow_alloc(obj=PTR14, id=0x1235, size=8, gfp_flags=GFP_KERNEL), ctor=PTR4, ctor_data=PTR15 = PTR16 +$MOD_TEST: klp_shadow_get(obj=PTR1, id=0x1234) = PTR3 +$MOD_TEST: got expected PTR3 -> PTR2 result +$MOD_TEST: klp_shadow_get(obj=PTR1, id=0x1235) = PTR6 +$MOD_TEST: got expected PTR6 -> PTR5 result $MOD_TEST: klp_shadow_get(obj=PTR9, id=0x1234) = PTR8 -$MOD_TEST: got expected PTR8 -> PTR2 result -$MOD_TEST: klp_shadow_get(obj=PTR5, id=0x1235) = PTR10 -$MOD_TEST: got expected PTR10 -> PTR3 result -$MOD_TEST: shadow_ctor: PTR11 -> PTR4 -$MOD_TEST: klp_shadow_get_or_alloc(obj=PTR12, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR7, ctor_data=PTR4 = PTR11 -$MOD_TEST: klp_shadow_get_or_alloc(obj=PTR12, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR7, ctor_data=PTR4 = PTR11 -$MOD_TEST: got expected PTR11 -> PTR4 result -$MOD_TEST: shadow_dtor(obj=PTR5, shadow_data=PTR6) -$MOD_TEST: klp_shadow_free(obj=PTR5, id=0x1234, dtor=PTR13) -$MOD_TEST: klp_shadow_get(obj=PTR5, id=0x1234) = PTR0 +$MOD_TEST: got expected PTR8 -> PTR7 result +$MOD_TEST: klp_shadow_get(obj=PTR9, id=0x1235) = PTR11 +$MOD_TEST: got expected PTR11 -> PTR10 result +$MOD_TEST: klp_shadow_get(obj=PTR14, id=0x1234) = PTR13 +$MOD_TEST: got expected PTR13 -> PTR12 result +$MOD_TEST: klp_shadow_get(obj=PTR14, id=0x1235) = PTR16 +$MOD_TEST: got expected PTR16 -> PTR15 result +$MOD_TEST: klp_shadow_get_or_alloc(obj=PTR1, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR4, ctor_data=PTR2 = PTR3 +$MOD_TEST: got expected PTR3 -> PTR2 result +$MOD_TEST: klp_shadow_get_or_alloc(obj=PTR9, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR4, ctor_data=PTR7 = PTR8 +$MOD_TEST: got expected PTR8 -> PTR7 result +$MOD_TEST: klp_shadow_get_or_alloc(obj=PTR14, id=0x1234, size=8, gfp_flags=GFP_KERNEL), ctor=PTR4, ctor_data=PTR12 = PTR13 +$MOD_TEST: got expected PTR13 -> PTR12 result +$MOD_TEST: shadow_dtor(obj=PTR1, shadow_data=PTR3) +$MOD_TEST: klp_shadow_free(obj=PTR1, id=0x1234, dtor=PTR17) +$MOD_TEST: klp_shadow_get(obj=PTR1, id=0x1234) = PTR0 $MOD_TEST: got expected NULL result $MOD_TEST: shadow_dtor(obj=PTR9, shadow_data=PTR8) -$MOD_TEST: klp_shadow_free(obj=PTR9, id=0x1234, dtor=PTR13) +$MOD_TEST: klp_shadow_free(obj=PTR9, id=0x1234, dtor=PTR17) $MOD_TEST: klp_shadow_get(obj=PTR9, id=0x1234) = PTR0 $MOD_TEST: got expected NULL result -$MOD_TEST: shadow_dtor(obj=PTR12, shadow_data=PTR11) -$MOD_TEST: klp_shadow_free(obj=PTR12, id=0x1234, dtor=PTR13) -$MOD_TEST: klp_shadow_get(obj=PTR12, id=0x1234) = PTR0 +$MOD_TEST: shadow_dtor(obj=PTR14, shadow_data=PTR13) +$MOD_TEST: klp_shadow_free(obj=PTR14, id=0x1234, dtor=PTR17) +$MOD_TEST: klp_shadow_get(obj=PTR14, id=0x1234) = PTR0 $MOD_TEST: got expected NULL result -$MOD_TEST: klp_shadow_get(obj=PTR5, id=0x1235) = PTR10 -$MOD_TEST: got expected PTR10 -> PTR3 result -$MOD_TEST: shadow_dtor(obj=PTR5, shadow_data=PTR10) -$MOD_TEST: klp_shadow_free_all(id=0x1235, dtor=PTR13) -$MOD_TEST: klp_shadow_get(obj=PTR5, id=0x1234) = PTR0 -$MOD_TEST: shadow_get() got expected NULL result -% rmmod test_klp_shadow_vars" +$MOD_TEST: klp_shadow_get(obj=PTR1, id=0x1235) = PTR6 +$MOD_TEST: got expected PTR6 -> PTR5 result +$MOD_TEST: klp_shadow_get(obj=PTR9, id=0x1235) = PTR11 +$MOD_TEST: got expected PTR11 -> PTR10 result +$MOD_TEST: klp_shadow_get(obj=PTR14, id=0x1235) = PTR16 +$MOD_TEST: got expected PTR16 -> PTR15 result +$MOD_TEST: klp_shadow_free_all(id=0x1235, dtor=PTR0) +$MOD_TEST: klp_shadow_get(obj=PTR1, id=0x1235) = PTR0 +$MOD_TEST: got expected NULL result +$MOD_TEST: klp_shadow_get(obj=PTR9, id=0x1235) = PTR0 +$MOD_TEST: got expected NULL result +$MOD_TEST: klp_shadow_get(obj=PTR14, id=0x1235) = PTR0 +$MOD_TEST: got expected NULL result +% rmmod $MOD_TEST" exit 0 diff --git a/tools/testing/selftests/livepatch/test-state.sh b/tools/testing/selftests/livepatch/test-state.sh index a08212708115..38656721c958 100755 --- a/tools/testing/selftests/livepatch/test-state.sh +++ b/tools/testing/selftests/livepatch/test-state.sh @@ -10,10 +10,10 @@ MOD_LIVEPATCH3=test_klp_state3 setup_config -# TEST: Loading and removing a module that modifies the system state -echo -n "TEST: system state modification ... " -dmesg -C +# Load and remove a module that modifies the system state + +start_test "system state modification" load_lp $MOD_LIVEPATCH disable_lp $MOD_LIVEPATCH @@ -41,10 +41,9 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete % rmmod $MOD_LIVEPATCH" -# TEST: Take over system state change by a cumulative patch +# Take over system state change by a cumulative patch -echo -n "TEST: taking over system state modification ... " -dmesg -C +start_test "taking over system state modification" load_lp $MOD_LIVEPATCH load_lp $MOD_LIVEPATCH2 @@ -85,10 +84,9 @@ livepatch: '$MOD_LIVEPATCH2': unpatching complete % rmmod $MOD_LIVEPATCH2" -# TEST: Take over system state change by a cumulative patch +# Take over system state change by a cumulative patch -echo -n "TEST: compatible cumulative livepatches ... " -dmesg -C +start_test "compatible cumulative livepatches" load_lp $MOD_LIVEPATCH2 load_lp $MOD_LIVEPATCH3 @@ -142,10 +140,9 @@ livepatch: '$MOD_LIVEPATCH2': unpatching complete % rmmod $MOD_LIVEPATCH3" -# TEST: Failure caused by incompatible cumulative livepatches +# Failure caused by incompatible cumulative livepatches -echo -n "TEST: incompatible cumulative livepatches ... " -dmesg -C +start_test "incompatible cumulative livepatches" load_lp $MOD_LIVEPATCH2 load_failing_mod $MOD_LIVEPATCH diff --git a/tools/testing/selftests/mincore/.gitignore b/tools/testing/selftests/mincore/.gitignore new file mode 100644 index 000000000000..15c4dfc2df00 --- /dev/null +++ b/tools/testing/selftests/mincore/.gitignore @@ -0,0 +1,2 @@ +# SPDX-License-Identifier: GPL-2.0+ +mincore_selftest diff --git a/tools/testing/selftests/mincore/Makefile b/tools/testing/selftests/mincore/Makefile new file mode 100644 index 000000000000..38c7db1e8926 --- /dev/null +++ b/tools/testing/selftests/mincore/Makefile @@ -0,0 +1,6 @@ +# SPDX-License-Identifier: GPL-2.0+ + +CFLAGS += -Wall + +TEST_GEN_PROGS := mincore_selftest +include ../lib.mk diff --git a/tools/testing/selftests/mincore/mincore_selftest.c b/tools/testing/selftests/mincore/mincore_selftest.c new file mode 100644 index 000000000000..5a1e85ff5d32 --- /dev/null +++ b/tools/testing/selftests/mincore/mincore_selftest.c @@ -0,0 +1,361 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * kselftest suite for mincore(). + * + * Copyright (C) 2020 Collabora, Ltd. + */ + +#define _GNU_SOURCE + +#include <stdio.h> +#include <errno.h> +#include <unistd.h> +#include <stdlib.h> +#include <sys/mman.h> +#include <string.h> +#include <fcntl.h> +#include <string.h> + +#include "../kselftest.h" +#include "../kselftest_harness.h" + +/* Default test file size: 4MB */ +#define MB (1UL << 20) +#define FILE_SIZE (4 * MB) + + +/* + * Tests the user interface. This test triggers most of the documented + * error conditions in mincore(). + */ +TEST(basic_interface) +{ + int retval; + int page_size; + unsigned char vec[1]; + char *addr; + + page_size = sysconf(_SC_PAGESIZE); + + /* Query a 0 byte sized range */ + retval = mincore(0, 0, vec); + EXPECT_EQ(0, retval); + + /* Addresses in the specified range are invalid or unmapped */ + errno = 0; + retval = mincore(NULL, page_size, vec); + EXPECT_EQ(-1, retval); + EXPECT_EQ(ENOMEM, errno); + + errno = 0; + addr = mmap(NULL, page_size, PROT_READ | PROT_WRITE, + MAP_SHARED | MAP_ANONYMOUS, -1, 0); + ASSERT_NE(MAP_FAILED, addr) { + TH_LOG("mmap error: %s", strerror(errno)); + } + + /* <addr> argument is not page-aligned */ + errno = 0; + retval = mincore(addr + 1, page_size, vec); + EXPECT_EQ(-1, retval); + EXPECT_EQ(EINVAL, errno); + + /* <length> argument is too large */ + errno = 0; + retval = mincore(addr, -1, vec); + EXPECT_EQ(-1, retval); + EXPECT_EQ(ENOMEM, errno); + + /* <vec> argument points to an illegal address */ + errno = 0; + retval = mincore(addr, page_size, NULL); + EXPECT_EQ(-1, retval); + EXPECT_EQ(EFAULT, errno); + munmap(addr, page_size); +} + + +/* + * Test mincore() behavior on a private anonymous page mapping. + * Check that the page is not loaded into memory right after the mapping + * but after accessing it (on-demand allocation). + * Then free the page and check that it's not memory-resident. + */ +TEST(check_anonymous_locked_pages) +{ + unsigned char vec[1]; + char *addr; + int retval; + int page_size; + + page_size = sysconf(_SC_PAGESIZE); + + /* Map one page and check it's not memory-resident */ + errno = 0; + addr = mmap(NULL, page_size, PROT_READ | PROT_WRITE, + MAP_PRIVATE | MAP_ANONYMOUS, -1, 0); + ASSERT_NE(MAP_FAILED, addr) { + TH_LOG("mmap error: %s", strerror(errno)); + } + retval = mincore(addr, page_size, vec); + ASSERT_EQ(0, retval); + ASSERT_EQ(0, vec[0]) { + TH_LOG("Page found in memory before use"); + } + + /* Touch the page and check again. It should now be in memory */ + addr[0] = 1; + mlock(addr, page_size); + retval = mincore(addr, page_size, vec); + ASSERT_EQ(0, retval); + ASSERT_EQ(1, vec[0]) { + TH_LOG("Page not found in memory after use"); + } + + /* + * It shouldn't be memory-resident after unlocking it and + * marking it as unneeded. + */ + munlock(addr, page_size); + madvise(addr, page_size, MADV_DONTNEED); + retval = mincore(addr, page_size, vec); + ASSERT_EQ(0, retval); + ASSERT_EQ(0, vec[0]) { + TH_LOG("Page in memory after being zapped"); + } + munmap(addr, page_size); +} + + +/* + * Check mincore() behavior on huge pages. + * This test will be skipped if the mapping fails (ie. if there are no + * huge pages available). + * + * Make sure the system has at least one free huge page, check + * "HugePages_Free" in /proc/meminfo. + * Increment /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages if + * needed. + */ +TEST(check_huge_pages) +{ + unsigned char vec[1]; + char *addr; + int retval; + int page_size; + + page_size = sysconf(_SC_PAGESIZE); + + errno = 0; + addr = mmap(NULL, page_size, PROT_READ | PROT_WRITE, + MAP_PRIVATE | MAP_ANONYMOUS | MAP_HUGETLB, + -1, 0); + if (addr == MAP_FAILED) { + if (errno == ENOMEM) + SKIP(return, "No huge pages available."); + else + TH_LOG("mmap error: %s", strerror(errno)); + } + retval = mincore(addr, page_size, vec); + ASSERT_EQ(0, retval); + ASSERT_EQ(0, vec[0]) { + TH_LOG("Page found in memory before use"); + } + + addr[0] = 1; + mlock(addr, page_size); + retval = mincore(addr, page_size, vec); + ASSERT_EQ(0, retval); + ASSERT_EQ(1, vec[0]) { + TH_LOG("Page not found in memory after use"); + } + + munlock(addr, page_size); + munmap(addr, page_size); +} + + +/* + * Test mincore() behavior on a file-backed page. + * No pages should be loaded into memory right after the mapping. Then, + * accessing any address in the mapping range should load the page + * containing the address and a number of subsequent pages (readahead). + * + * The actual readahead settings depend on the test environment, so we + * can't make a lot of assumptions about that. This test covers the most + * general cases. + */ +TEST(check_file_mmap) +{ + unsigned char *vec; + int vec_size; + char *addr; + int retval; + int page_size; + int fd; + int i; + int ra_pages = 0; + + page_size = sysconf(_SC_PAGESIZE); + vec_size = FILE_SIZE / page_size; + if (FILE_SIZE % page_size) + vec_size++; + + vec = calloc(vec_size, sizeof(unsigned char)); + ASSERT_NE(NULL, vec) { + TH_LOG("Can't allocate array"); + } + + errno = 0; + fd = open(".", O_TMPFILE | O_RDWR, 0600); + ASSERT_NE(-1, fd) { + TH_LOG("Can't create temporary file: %s", + strerror(errno)); + } + errno = 0; + retval = fallocate(fd, 0, 0, FILE_SIZE); + ASSERT_EQ(0, retval) { + TH_LOG("Error allocating space for the temporary file: %s", + strerror(errno)); + } + + /* + * Map the whole file, the pages shouldn't be fetched yet. + */ + errno = 0; + addr = mmap(NULL, FILE_SIZE, PROT_READ | PROT_WRITE, + MAP_SHARED, fd, 0); + ASSERT_NE(MAP_FAILED, addr) { + TH_LOG("mmap error: %s", strerror(errno)); + } + retval = mincore(addr, FILE_SIZE, vec); + ASSERT_EQ(0, retval); + for (i = 0; i < vec_size; i++) { + ASSERT_EQ(0, vec[i]) { + TH_LOG("Unexpected page in memory"); + } + } + + /* + * Touch a page in the middle of the mapping. We expect the next + * few pages (the readahead window) to be populated too. + */ + addr[FILE_SIZE / 2] = 1; + retval = mincore(addr, FILE_SIZE, vec); + ASSERT_EQ(0, retval); + ASSERT_EQ(1, vec[FILE_SIZE / 2 / page_size]) { + TH_LOG("Page not found in memory after use"); + } + + i = FILE_SIZE / 2 / page_size + 1; + while (i < vec_size && vec[i]) { + ra_pages++; + i++; + } + EXPECT_GT(ra_pages, 0) { + TH_LOG("No read-ahead pages found in memory"); + } + + EXPECT_LT(i, vec_size) { + TH_LOG("Read-ahead pages reached the end of the file"); + } + /* + * End of the readahead window. The rest of the pages shouldn't + * be in memory. + */ + if (i < vec_size) { + while (i < vec_size && !vec[i]) + i++; + EXPECT_EQ(vec_size, i) { + TH_LOG("Unexpected page in memory beyond readahead window"); + } + } + + munmap(addr, FILE_SIZE); + close(fd); + free(vec); +} + + +/* + * Test mincore() behavior on a page backed by a tmpfs file. This test + * performs the same steps as the previous one. However, we don't expect + * any readahead in this case. + */ +TEST(check_tmpfs_mmap) +{ + unsigned char *vec; + int vec_size; + char *addr; + int retval; + int page_size; + int fd; + int i; + int ra_pages = 0; + + page_size = sysconf(_SC_PAGESIZE); + vec_size = FILE_SIZE / page_size; + if (FILE_SIZE % page_size) + vec_size++; + + vec = calloc(vec_size, sizeof(unsigned char)); + ASSERT_NE(NULL, vec) { + TH_LOG("Can't allocate array"); + } + + errno = 0; + fd = open("/dev/shm", O_TMPFILE | O_RDWR, 0600); + ASSERT_NE(-1, fd) { + TH_LOG("Can't create temporary file: %s", + strerror(errno)); + } + errno = 0; + retval = fallocate(fd, 0, 0, FILE_SIZE); + ASSERT_EQ(0, retval) { + TH_LOG("Error allocating space for the temporary file: %s", + strerror(errno)); + } + + /* + * Map the whole file, the pages shouldn't be fetched yet. + */ + errno = 0; + addr = mmap(NULL, FILE_SIZE, PROT_READ | PROT_WRITE, + MAP_SHARED, fd, 0); + ASSERT_NE(MAP_FAILED, addr) { + TH_LOG("mmap error: %s", strerror(errno)); + } + retval = mincore(addr, FILE_SIZE, vec); + ASSERT_EQ(0, retval); + for (i = 0; i < vec_size; i++) { + ASSERT_EQ(0, vec[i]) { + TH_LOG("Unexpected page in memory"); + } + } + + /* + * Touch a page in the middle of the mapping. We expect only + * that page to be fetched into memory. + */ + addr[FILE_SIZE / 2] = 1; + retval = mincore(addr, FILE_SIZE, vec); + ASSERT_EQ(0, retval); + ASSERT_EQ(1, vec[FILE_SIZE / 2 / page_size]) { + TH_LOG("Page not found in memory after use"); + } + + i = FILE_SIZE / 2 / page_size + 1; + while (i < vec_size && vec[i]) { + ra_pages++; + i++; + } + ASSERT_EQ(ra_pages, 0) { + TH_LOG("Read-ahead pages found in memory"); + } + + munmap(addr, FILE_SIZE); + close(fd); + free(vec); +} + +TEST_HARNESS_MAIN diff --git a/tools/testing/selftests/powerpc/alignment/alignment_handler.c b/tools/testing/selftests/powerpc/alignment/alignment_handler.c index 0453c50c949c..55ef15184057 100644 --- a/tools/testing/selftests/powerpc/alignment/alignment_handler.c +++ b/tools/testing/selftests/powerpc/alignment/alignment_handler.c @@ -9,7 +9,17 @@ * This selftest exercises the powerpc alignment fault handler. * * We create two sets of source and destination buffers, one in regular memory, - * the other cache-inhibited (we use /dev/fb0 for this). + * the other cache-inhibited (by default we use /dev/fb0 for this, but an + * alterative path for cache-inhibited memory may be provided). + * + * One way to get cache-inhibited memory is to use the "mem" kernel parameter + * to limit the kernel to less memory than actually exists. Addresses above + * the limit may still be accessed but will be treated as cache-inhibited. For + * example, if there is actually 4GB of memory and the parameter "mem=3GB" is + * used, memory from address 0xC0000000 onwards is treated as cache-inhibited. + * To access this region /dev/mem is used. The kernel should be configured + * without CONFIG_STRICT_DEVMEM. In this case use: + * ./alignment_handler /dev/mem 0xc0000000 * * We initialise the source buffers, then use whichever set of load/store * instructions is under test to copy bytes from the source buffers to the @@ -48,11 +58,14 @@ #include <asm/cputable.h> #include "utils.h" +#include "instructions.h" int bufsize; int debug; int testing; volatile int gotsig; +char *cipath = "/dev/fb0"; +long cioffset; void sighandler(int sig, siginfo_t *info, void *ctx) { @@ -84,6 +97,17 @@ void sighandler(int sig, siginfo_t *info, void *ctx) } \ rc |= do_test(#name, test_##name) +#define TESTP(name, ld_op, st_op, ld_reg, st_reg) \ + void test_##name(char *s, char *d) \ + { \ + asm volatile( \ + ld_op(ld_reg, %0, 0, 0) \ + st_op(st_reg, %1, 0, 0) \ + :: "r"(s), "r"(d), "r"(0) \ + : "memory", "vs0", "vs32", "r31"); \ + } \ + rc |= do_test(#name, test_##name) + #define LOAD_VSX_XFORM_TEST(op) TEST(op, op, stxvd2x, XFORM, 32, 32) #define STORE_VSX_XFORM_TEST(op) TEST(op, lxvd2x, op, XFORM, 32, 32) #define LOAD_VSX_DFORM_TEST(op) TEST(op, op, stxv, DFORM, 32, 32) @@ -103,6 +127,17 @@ void sighandler(int sig, siginfo_t *info, void *ctx) #define LOAD_FLOAT_XFORM_TEST(op) TEST(op, op, stfdx, XFORM, 0, 0) #define STORE_FLOAT_XFORM_TEST(op) TEST(op, lfdx, op, XFORM, 0, 0) +#define LOAD_MLS_PREFIX_TEST(op) TESTP(op, op, PSTD, 31, 31) +#define STORE_MLS_PREFIX_TEST(op) TESTP(op, PLD, op, 31, 31) + +#define LOAD_8LS_PREFIX_TEST(op) TESTP(op, op, PSTD, 31, 31) +#define STORE_8LS_PREFIX_TEST(op) TESTP(op, PLD, op, 31, 31) + +#define LOAD_FLOAT_MLS_PREFIX_TEST(op) TESTP(op, op, PSTFD, 0, 0) +#define STORE_FLOAT_MLS_PREFIX_TEST(op) TESTP(op, PLFD, op, 0, 0) + +#define LOAD_VSX_8LS_PREFIX_TEST(op, tail) TESTP(op, op, PSTXV ## tail, 0, 32) +#define STORE_VSX_8LS_PREFIX_TEST(op, tail) TESTP(op, PLXV ## tail, op, 32, 0) /* FIXME: Unimplemented tests: */ // STORE_DFORM_TEST(stq) /* FIXME: need two registers for quad */ @@ -195,17 +230,18 @@ int do_test(char *test_name, void (*test_func)(char *, char *)) printf("\tDoing %s:\t", test_name); - fd = open("/dev/fb0", O_RDWR); + fd = open(cipath, O_RDWR); if (fd < 0) { printf("\n"); - perror("Can't open /dev/fb0 now?"); + perror("Can't open ci file now?"); return 1; } - ci0 = mmap(NULL, bufsize, PROT_WRITE, MAP_SHARED, - fd, 0x0); - ci1 = mmap(NULL, bufsize, PROT_WRITE, MAP_SHARED, - fd, bufsize); + ci0 = mmap(NULL, bufsize, PROT_WRITE | PROT_READ, MAP_SHARED, + fd, cioffset); + ci1 = mmap(NULL, bufsize, PROT_WRITE | PROT_READ, MAP_SHARED, + fd, cioffset + bufsize); + if ((ci0 == MAP_FAILED) || (ci1 == MAP_FAILED)) { printf("\n"); perror("mmap failed"); @@ -270,11 +306,11 @@ int do_test(char *test_name, void (*test_func)(char *, char *)) return rc; } -static bool can_open_fb0(void) +static bool can_open_cifile(void) { int fd; - fd = open("/dev/fb0", O_RDWR); + fd = open(cipath, O_RDWR); if (fd < 0) return false; @@ -286,7 +322,7 @@ int test_alignment_handler_vsx_206(void) { int rc = 0; - SKIP_IF(!can_open_fb0()); + SKIP_IF(!can_open_cifile()); SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_06)); printf("VSX: 2.06B\n"); @@ -304,7 +340,7 @@ int test_alignment_handler_vsx_207(void) { int rc = 0; - SKIP_IF(!can_open_fb0()); + SKIP_IF(!can_open_cifile()); SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07)); printf("VSX: 2.07B\n"); @@ -320,7 +356,7 @@ int test_alignment_handler_vsx_300(void) { int rc = 0; - SKIP_IF(!can_open_fb0()); + SKIP_IF(!can_open_cifile()); SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_3_00)); printf("VSX: 3.00B\n"); @@ -348,11 +384,30 @@ int test_alignment_handler_vsx_300(void) return rc; } +int test_alignment_handler_vsx_prefix(void) +{ + int rc = 0; + + SKIP_IF(!can_open_cifile()); + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_3_1)); + + printf("VSX: PREFIX\n"); + LOAD_VSX_8LS_PREFIX_TEST(PLXSD, 0); + LOAD_VSX_8LS_PREFIX_TEST(PLXSSP, 0); + LOAD_VSX_8LS_PREFIX_TEST(PLXV0, 0); + LOAD_VSX_8LS_PREFIX_TEST(PLXV1, 1); + STORE_VSX_8LS_PREFIX_TEST(PSTXSD, 0); + STORE_VSX_8LS_PREFIX_TEST(PSTXSSP, 0); + STORE_VSX_8LS_PREFIX_TEST(PSTXV0, 0); + STORE_VSX_8LS_PREFIX_TEST(PSTXV1, 1); + return rc; +} + int test_alignment_handler_integer(void) { int rc = 0; - SKIP_IF(!can_open_fb0()); + SKIP_IF(!can_open_cifile()); printf("Integer\n"); LOAD_DFORM_TEST(lbz); @@ -408,7 +463,7 @@ int test_alignment_handler_integer_206(void) { int rc = 0; - SKIP_IF(!can_open_fb0()); + SKIP_IF(!can_open_cifile()); SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_06)); printf("Integer: 2.06\n"); @@ -419,11 +474,32 @@ int test_alignment_handler_integer_206(void) return rc; } +int test_alignment_handler_integer_prefix(void) +{ + int rc = 0; + + SKIP_IF(!can_open_cifile()); + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_3_1)); + + printf("Integer: PREFIX\n"); + LOAD_MLS_PREFIX_TEST(PLBZ); + LOAD_MLS_PREFIX_TEST(PLHZ); + LOAD_MLS_PREFIX_TEST(PLHA); + LOAD_MLS_PREFIX_TEST(PLWZ); + LOAD_8LS_PREFIX_TEST(PLWA); + LOAD_8LS_PREFIX_TEST(PLD); + STORE_MLS_PREFIX_TEST(PSTB); + STORE_MLS_PREFIX_TEST(PSTH); + STORE_MLS_PREFIX_TEST(PSTW); + STORE_8LS_PREFIX_TEST(PSTD); + return rc; +} + int test_alignment_handler_vmx(void) { int rc = 0; - SKIP_IF(!can_open_fb0()); + SKIP_IF(!can_open_cifile()); SKIP_IF(!have_hwcap(PPC_FEATURE_HAS_ALTIVEC)); printf("VMX\n"); @@ -451,7 +527,7 @@ int test_alignment_handler_fp(void) { int rc = 0; - SKIP_IF(!can_open_fb0()); + SKIP_IF(!can_open_cifile()); printf("Floating point\n"); LOAD_FLOAT_DFORM_TEST(lfd); @@ -479,7 +555,7 @@ int test_alignment_handler_fp_205(void) { int rc = 0; - SKIP_IF(!can_open_fb0()); + SKIP_IF(!can_open_cifile()); SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_05)); printf("Floating point: 2.05\n"); @@ -497,7 +573,7 @@ int test_alignment_handler_fp_206(void) { int rc = 0; - SKIP_IF(!can_open_fb0()); + SKIP_IF(!can_open_cifile()); SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_06)); printf("Floating point: 2.06\n"); @@ -507,13 +583,32 @@ int test_alignment_handler_fp_206(void) return rc; } + +int test_alignment_handler_fp_prefix(void) +{ + int rc = 0; + + SKIP_IF(!can_open_cifile()); + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_3_1)); + + printf("Floating point: PREFIX\n"); + LOAD_FLOAT_DFORM_TEST(lfs); + LOAD_FLOAT_MLS_PREFIX_TEST(PLFS); + LOAD_FLOAT_MLS_PREFIX_TEST(PLFD); + STORE_FLOAT_MLS_PREFIX_TEST(PSTFS); + STORE_FLOAT_MLS_PREFIX_TEST(PSTFD); + return rc; +} + void usage(char *prog) { - printf("Usage: %s [options]\n", prog); + printf("Usage: %s [options] [path [offset]]\n", prog); printf(" -d Enable debug error output\n"); printf("\n"); - printf("This test requires a POWER8 or POWER9 CPU and a usable "); - printf("framebuffer at /dev/fb0.\n"); + printf("This test requires a POWER8, POWER9 or POWER10 CPU "); + printf("and either a usable framebuffer at /dev/fb0 or "); + printf("the path to usable cache inhibited memory and optional "); + printf("offset to be provided\n"); } int main(int argc, char *argv[]) @@ -533,6 +628,13 @@ int main(int argc, char *argv[]) exit(1); } } + argc -= optind; + argv += optind; + + if (argc > 0) + cipath = argv[0]; + if (argc > 1) + cioffset = strtol(argv[1], 0, 0x10); bufsize = getpagesize(); @@ -552,10 +654,14 @@ int main(int argc, char *argv[]) "test_alignment_handler_vsx_207"); rc |= test_harness(test_alignment_handler_vsx_300, "test_alignment_handler_vsx_300"); + rc |= test_harness(test_alignment_handler_vsx_prefix, + "test_alignment_handler_vsx_prefix"); rc |= test_harness(test_alignment_handler_integer, "test_alignment_handler_integer"); rc |= test_harness(test_alignment_handler_integer_206, "test_alignment_handler_integer_206"); + rc |= test_harness(test_alignment_handler_integer_prefix, + "test_alignment_handler_integer_prefix"); rc |= test_harness(test_alignment_handler_vmx, "test_alignment_handler_vmx"); rc |= test_harness(test_alignment_handler_fp, @@ -564,5 +670,7 @@ int main(int argc, char *argv[]) "test_alignment_handler_fp_205"); rc |= test_harness(test_alignment_handler_fp_206, "test_alignment_handler_fp_206"); + rc |= test_harness(test_alignment_handler_fp_prefix, + "test_alignment_handler_fp_prefix"); return rc; } diff --git a/tools/testing/selftests/powerpc/benchmarks/context_switch.c b/tools/testing/selftests/powerpc/benchmarks/context_switch.c index a2e8c9da7fa5..d50cc05df495 100644 --- a/tools/testing/selftests/powerpc/benchmarks/context_switch.c +++ b/tools/testing/selftests/powerpc/benchmarks/context_switch.c @@ -19,6 +19,7 @@ #include <limits.h> #include <sys/time.h> #include <sys/syscall.h> +#include <sys/sysinfo.h> #include <sys/types.h> #include <sys/shm.h> #include <linux/futex.h> @@ -104,8 +105,9 @@ static void start_thread_on(void *(*fn)(void *), void *arg, unsigned long cpu) static void start_process_on(void *(*fn)(void *), void *arg, unsigned long cpu) { - int pid; - cpu_set_t cpuset; + int pid, ncpus; + cpu_set_t *cpuset; + size_t size; pid = fork(); if (pid == -1) { @@ -116,14 +118,23 @@ static void start_process_on(void *(*fn)(void *), void *arg, unsigned long cpu) if (pid) return; - CPU_ZERO(&cpuset); - CPU_SET(cpu, &cpuset); + ncpus = get_nprocs(); + size = CPU_ALLOC_SIZE(ncpus); + cpuset = CPU_ALLOC(ncpus); + if (!cpuset) { + perror("malloc"); + exit(1); + } + CPU_ZERO_S(size, cpuset); + CPU_SET_S(cpu, size, cpuset); - if (sched_setaffinity(0, sizeof(cpuset), &cpuset)) { + if (sched_setaffinity(0, size, cpuset)) { perror("sched_setaffinity"); + CPU_FREE(cpuset); exit(1); } + CPU_FREE(cpuset); fn(arg); exit(0); diff --git a/tools/testing/selftests/powerpc/eeh/eeh-functions.sh b/tools/testing/selftests/powerpc/eeh/eeh-functions.sh index f52ed92b53e7..00dc32c0ed75 100755 --- a/tools/testing/selftests/powerpc/eeh/eeh-functions.sh +++ b/tools/testing/selftests/powerpc/eeh/eeh-functions.sh @@ -5,12 +5,17 @@ pe_ok() { local dev="$1" local path="/sys/bus/pci/devices/$dev/eeh_pe_state" - if ! [ -e "$path" ] ; then + # if a driver doesn't support the error handling callbacks then the + # device is recovered by removing and re-probing it. This causes the + # sysfs directory to disappear so read the PE state once and squash + # any potential error messages + local eeh_state="$(cat $path 2>/dev/null)" + if [ -z "$eeh_state" ]; then return 1; fi - local fw_state="$(cut -d' ' -f1 < $path)" - local sw_state="$(cut -d' ' -f2 < $path)" + local fw_state="$(echo $eeh_state | cut -d' ' -f1)" + local sw_state="$(echo $eeh_state | cut -d' ' -f2)" # If EEH_PE_ISOLATED or EEH_PE_RECOVERING are set then the PE is in an # error state or being recovered. Either way, not ok. diff --git a/tools/testing/selftests/powerpc/include/instructions.h b/tools/testing/selftests/powerpc/include/instructions.h index f36061eb6f0f..4efa6314bd96 100644 --- a/tools/testing/selftests/powerpc/include/instructions.h +++ b/tools/testing/selftests/powerpc/include/instructions.h @@ -66,4 +66,81 @@ static inline int paste_last(void *i) #define PPC_INST_PASTE __PASTE(0, 0, 0, 0) #define PPC_INST_PASTE_LAST __PASTE(0, 0, 1, 1) +/* This defines the prefixed load/store instructions */ +#ifdef __ASSEMBLY__ +# define stringify_in_c(...) __VA_ARGS__ +#else +# define __stringify_in_c(...) #__VA_ARGS__ +# define stringify_in_c(...) __stringify_in_c(__VA_ARGS__) " " +#endif + +#define __PPC_RA(a) (((a) & 0x1f) << 16) +#define __PPC_RS(s) (((s) & 0x1f) << 21) +#define __PPC_RT(t) __PPC_RS(t) +#define __PPC_PREFIX_R(r) (((r) & 0x1) << 20) + +#define PPC_PREFIX_MLS 0x06000000 +#define PPC_PREFIX_8LS 0x04000000 + +#define PPC_INST_LBZ 0x88000000 +#define PPC_INST_LHZ 0xa0000000 +#define PPC_INST_LHA 0xa8000000 +#define PPC_INST_LWZ 0x80000000 +#define PPC_INST_STB 0x98000000 +#define PPC_INST_STH 0xb0000000 +#define PPC_INST_STW 0x90000000 +#define PPC_INST_STD 0xf8000000 +#define PPC_INST_LFS 0xc0000000 +#define PPC_INST_LFD 0xc8000000 +#define PPC_INST_STFS 0xd0000000 +#define PPC_INST_STFD 0xd8000000 + +#define PREFIX_MLS(instr, t, a, r, d) stringify_in_c(.balign 64, , 4;) \ + stringify_in_c(.long PPC_PREFIX_MLS | \ + __PPC_PREFIX_R(r) | \ + (((d) >> 16) & 0x3ffff);) \ + stringify_in_c(.long (instr) | \ + __PPC_RT(t) | \ + __PPC_RA(a) | \ + ((d) & 0xffff);\n) + +#define PREFIX_8LS(instr, t, a, r, d) stringify_in_c(.balign 64, , 4;) \ + stringify_in_c(.long PPC_PREFIX_8LS | \ + __PPC_PREFIX_R(r) | \ + (((d) >> 16) & 0x3ffff);) \ + stringify_in_c(.long (instr) | \ + __PPC_RT(t) | \ + __PPC_RA(a) | \ + ((d) & 0xffff);\n) + +/* Prefixed Integer Load/Store instructions */ +#define PLBZ(t, a, r, d) PREFIX_MLS(PPC_INST_LBZ, t, a, r, d) +#define PLHZ(t, a, r, d) PREFIX_MLS(PPC_INST_LHZ, t, a, r, d) +#define PLHA(t, a, r, d) PREFIX_MLS(PPC_INST_LHA, t, a, r, d) +#define PLWZ(t, a, r, d) PREFIX_MLS(PPC_INST_LWZ, t, a, r, d) +#define PLWA(t, a, r, d) PREFIX_8LS(0xa4000000, t, a, r, d) +#define PLD(t, a, r, d) PREFIX_8LS(0xe4000000, t, a, r, d) +#define PLQ(t, a, r, d) PREFIX_8LS(0xe0000000, t, a, r, d) +#define PSTB(s, a, r, d) PREFIX_MLS(PPC_INST_STB, s, a, r, d) +#define PSTH(s, a, r, d) PREFIX_MLS(PPC_INST_STH, s, a, r, d) +#define PSTW(s, a, r, d) PREFIX_MLS(PPC_INST_STW, s, a, r, d) +#define PSTD(s, a, r, d) PREFIX_8LS(0xf4000000, s, a, r, d) +#define PSTQ(s, a, r, d) PREFIX_8LS(0xf0000000, s, a, r, d) + +/* Prefixed Floating-Point Load/Store Instructions */ +#define PLFS(frt, a, r, d) PREFIX_MLS(PPC_INST_LFS, frt, a, r, d) +#define PLFD(frt, a, r, d) PREFIX_MLS(PPC_INST_LFD, frt, a, r, d) +#define PSTFS(frs, a, r, d) PREFIX_MLS(PPC_INST_STFS, frs, a, r, d) +#define PSTFD(frs, a, r, d) PREFIX_MLS(PPC_INST_STFD, frs, a, r, d) + +/* Prefixed VSX Load/Store Instructions */ +#define PLXSD(vrt, a, r, d) PREFIX_8LS(0xa8000000, vrt, a, r, d) +#define PLXSSP(vrt, a, r, d) PREFIX_8LS(0xac000000, vrt, a, r, d) +#define PLXV0(s, a, r, d) PREFIX_8LS(0xc8000000, s, a, r, d) +#define PLXV1(s, a, r, d) PREFIX_8LS(0xcc000000, s, a, r, d) +#define PSTXSD(vrs, a, r, d) PREFIX_8LS(0xb8000000, vrs, a, r, d) +#define PSTXSSP(vrs, a, r, d) PREFIX_8LS(0xbc000000, vrs, a, r, d) +#define PSTXV0(s, a, r, d) PREFIX_8LS(0xd8000000, s, a, r, d) +#define PSTXV1(s, a, r, d) PREFIX_8LS(0xdc000000, s, a, r, d) + #endif /* _SELFTESTS_POWERPC_INSTRUCTIONS_H */ diff --git a/tools/testing/selftests/powerpc/include/pkeys.h b/tools/testing/selftests/powerpc/include/pkeys.h new file mode 100644 index 000000000000..3312cb1b058d --- /dev/null +++ b/tools/testing/selftests/powerpc/include/pkeys.h @@ -0,0 +1,136 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright 2020, Sandipan Das, IBM Corp. + */ + +#ifndef _SELFTESTS_POWERPC_PKEYS_H +#define _SELFTESTS_POWERPC_PKEYS_H + +#include <sys/mman.h> + +#include "reg.h" +#include "utils.h" + +/* + * Older versions of libc use the Intel-specific access rights. + * Hence, override the definitions as they might be incorrect. + */ +#undef PKEY_DISABLE_ACCESS +#define PKEY_DISABLE_ACCESS 0x3 + +#undef PKEY_DISABLE_WRITE +#define PKEY_DISABLE_WRITE 0x2 + +#undef PKEY_DISABLE_EXECUTE +#define PKEY_DISABLE_EXECUTE 0x4 + +/* Older versions of libc do not not define this */ +#ifndef SEGV_PKUERR +#define SEGV_PKUERR 4 +#endif + +#define SI_PKEY_OFFSET 0x20 + +#define __NR_pkey_mprotect 386 +#define __NR_pkey_alloc 384 +#define __NR_pkey_free 385 + +#define PKEY_BITS_PER_PKEY 2 +#define NR_PKEYS 32 +#define PKEY_BITS_MASK ((1UL << PKEY_BITS_PER_PKEY) - 1) + +inline unsigned long pkeyreg_get(void) +{ + return mfspr(SPRN_AMR); +} + +inline void pkeyreg_set(unsigned long amr) +{ + set_amr(amr); +} + +void pkey_set_rights(int pkey, unsigned long rights) +{ + unsigned long amr, shift; + + shift = (NR_PKEYS - pkey - 1) * PKEY_BITS_PER_PKEY; + amr = pkeyreg_get(); + amr &= ~(PKEY_BITS_MASK << shift); + amr |= (rights & PKEY_BITS_MASK) << shift; + pkeyreg_set(amr); +} + +int sys_pkey_mprotect(void *addr, size_t len, int prot, int pkey) +{ + return syscall(__NR_pkey_mprotect, addr, len, prot, pkey); +} + +int sys_pkey_alloc(unsigned long flags, unsigned long rights) +{ + return syscall(__NR_pkey_alloc, flags, rights); +} + +int sys_pkey_free(int pkey) +{ + return syscall(__NR_pkey_free, pkey); +} + +int pkeys_unsupported(void) +{ + bool hash_mmu = false; + int pkey; + + /* Protection keys are currently supported on Hash MMU only */ + FAIL_IF(using_hash_mmu(&hash_mmu)); + SKIP_IF(!hash_mmu); + + /* Check if the system call is supported */ + pkey = sys_pkey_alloc(0, 0); + SKIP_IF(pkey < 0); + sys_pkey_free(pkey); + + return 0; +} + +int siginfo_pkey(siginfo_t *si) +{ + /* + * In older versions of libc, siginfo_t does not have si_pkey as + * a member. + */ +#ifdef si_pkey + return si->si_pkey; +#else + return *((int *)(((char *) si) + SI_PKEY_OFFSET)); +#endif +} + +#define pkey_rights(r) ({ \ + static char buf[4] = "rwx"; \ + unsigned int amr_bits; \ + if ((r) & PKEY_DISABLE_EXECUTE) \ + buf[2] = '-'; \ + amr_bits = (r) & PKEY_BITS_MASK; \ + if (amr_bits & PKEY_DISABLE_WRITE) \ + buf[1] = '-'; \ + if (amr_bits & PKEY_DISABLE_ACCESS & ~PKEY_DISABLE_WRITE) \ + buf[0] = '-'; \ + buf; \ +}) + +unsigned long next_pkey_rights(unsigned long rights) +{ + if (rights == PKEY_DISABLE_ACCESS) + return PKEY_DISABLE_EXECUTE; + else if (rights == (PKEY_DISABLE_ACCESS | PKEY_DISABLE_EXECUTE)) + return 0; + + if ((rights & PKEY_BITS_MASK) == 0) + rights |= PKEY_DISABLE_WRITE; + else if ((rights & PKEY_BITS_MASK) == PKEY_DISABLE_WRITE) + rights |= PKEY_DISABLE_ACCESS; + + return rights; +} + +#endif /* _SELFTESTS_POWERPC_PKEYS_H */ diff --git a/tools/testing/selftests/powerpc/include/reg.h b/tools/testing/selftests/powerpc/include/reg.h index 022c5076b2c5..c0f2742a3a59 100644 --- a/tools/testing/selftests/powerpc/include/reg.h +++ b/tools/testing/selftests/powerpc/include/reg.h @@ -57,6 +57,12 @@ #define SPRN_PPR 896 /* Program Priority Register */ #define SPRN_AMR 13 /* Authority Mask Register - problem state */ +#define set_amr(v) asm volatile("isync;" \ + "mtspr " __stringify(SPRN_AMR) ",%0;" \ + "isync" : \ + : "r" ((unsigned long)(v)) \ + : "memory") + /* TEXASR register bits */ #define TEXASR_FC 0xFE00000000000000 #define TEXASR_FP 0x0100000000000000 diff --git a/tools/testing/selftests/powerpc/include/utils.h b/tools/testing/selftests/powerpc/include/utils.h index e089a0c30d9a..71d2924f5b8b 100644 --- a/tools/testing/selftests/powerpc/include/utils.h +++ b/tools/testing/selftests/powerpc/include/utils.h @@ -42,6 +42,16 @@ int perf_event_enable(int fd); int perf_event_disable(int fd); int perf_event_reset(int fd); +#if !defined(__GLIBC_PREREQ) || !__GLIBC_PREREQ(2, 30) +#include <unistd.h> +#include <sys/syscall.h> + +static inline pid_t gettid(void) +{ + return syscall(SYS_gettid); +} +#endif + static inline bool have_hwcap(unsigned long ftr) { return ((unsigned long)get_auxv_entry(AT_HWCAP) & ftr) == ftr; @@ -60,6 +70,7 @@ static inline bool have_hwcap2(unsigned long ftr2) #endif bool is_ppc64le(void); +int using_hash_mmu(bool *using_hash); /* Yes, this is evil */ #define FAIL_IF(x) \ @@ -71,6 +82,15 @@ do { \ } \ } while (0) +#define FAIL_IF_EXIT(x) \ +do { \ + if ((x)) { \ + fprintf(stderr, \ + "[FAIL] Test FAILED on line %d\n", __LINE__); \ + _exit(1); \ + } \ +} while (0) + /* The test harness uses this, yes it's gross */ #define MAGIC_SKIP_RETURN_VALUE 99 @@ -96,11 +116,20 @@ do { \ #define _str(s) #s #define str(s) _str(s) +#define sigsafe_err(msg) ({ \ + ssize_t nbytes __attribute__((unused)); \ + nbytes = write(STDERR_FILENO, msg, strlen(msg)); }) + /* POWER9 feature */ #ifndef PPC_FEATURE2_ARCH_3_00 #define PPC_FEATURE2_ARCH_3_00 0x00800000 #endif +/* POWER10 feature */ +#ifndef PPC_FEATURE2_ARCH_3_1 +#define PPC_FEATURE2_ARCH_3_1 0x00040000 +#endif + #if defined(__powerpc64__) #define UCONTEXT_NIA(UC) (UC)->uc_mcontext.gp_regs[PT_NIP] #define UCONTEXT_MSR(UC) (UC)->uc_mcontext.gp_regs[PT_MSR] diff --git a/tools/testing/selftests/powerpc/math/.gitignore b/tools/testing/selftests/powerpc/math/.gitignore index e31ca6f453ed..d0c23b2e4b60 100644 --- a/tools/testing/selftests/powerpc/math/.gitignore +++ b/tools/testing/selftests/powerpc/math/.gitignore @@ -6,3 +6,4 @@ vmx_preempt fpu_signal vmx_signal vsx_preempt +fpu_denormal diff --git a/tools/testing/selftests/powerpc/math/Makefile b/tools/testing/selftests/powerpc/math/Makefile index 11a10d7a2bbd..fcc91c205984 100644 --- a/tools/testing/selftests/powerpc/math/Makefile +++ b/tools/testing/selftests/powerpc/math/Makefile @@ -1,5 +1,5 @@ # SPDX-License-Identifier: GPL-2.0 -TEST_GEN_PROGS := fpu_syscall fpu_preempt fpu_signal vmx_syscall vmx_preempt vmx_signal vsx_preempt +TEST_GEN_PROGS := fpu_syscall fpu_preempt fpu_signal fpu_denormal vmx_syscall vmx_preempt vmx_signal vsx_preempt top_srcdir = ../../../../.. include ../../lib.mk @@ -11,9 +11,9 @@ $(OUTPUT)/fpu_syscall: fpu_asm.S $(OUTPUT)/fpu_preempt: fpu_asm.S $(OUTPUT)/fpu_signal: fpu_asm.S -$(OUTPUT)/vmx_syscall: vmx_asm.S -$(OUTPUT)/vmx_preempt: vmx_asm.S -$(OUTPUT)/vmx_signal: vmx_asm.S +$(OUTPUT)/vmx_syscall: vmx_asm.S ../utils.c +$(OUTPUT)/vmx_preempt: vmx_asm.S ../utils.c +$(OUTPUT)/vmx_signal: vmx_asm.S ../utils.c $(OUTPUT)/vsx_preempt: CFLAGS += -mvsx -$(OUTPUT)/vsx_preempt: vsx_asm.S +$(OUTPUT)/vsx_preempt: vsx_asm.S ../utils.c diff --git a/tools/testing/selftests/powerpc/math/fpu_denormal.c b/tools/testing/selftests/powerpc/math/fpu_denormal.c new file mode 100644 index 000000000000..5f96682abaa8 --- /dev/null +++ b/tools/testing/selftests/powerpc/math/fpu_denormal.c @@ -0,0 +1,38 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright IBM Corp. 2020 + * + * This test attempts to cause a FP denormal exception on POWER8 CPUs. Unfortunately + * if the denormal handler is not configured or working properly, this can cause a bad + * crash in kernel mode when the kernel tries to save FP registers when the process + * exits. + */ + +#include <stdio.h> +#include <string.h> + +#include "utils.h" + +static int test_denormal_fpu(void) +{ + unsigned int m32; + unsigned long m64; + volatile float f; + volatile double d; + + /* try to induce lfs <denormal> ; stfd */ + + m32 = 0x00715fcf; /* random denormal */ + memcpy((float *)&f, &m32, sizeof(f)); + d = f; + memcpy(&m64, (double *)&d, sizeof(d)); + + FAIL_IF((long)(m64 != 0x380c57f3c0000000)); /* renormalised value */ + + return 0; +} + +int main(int argc, char *argv[]) +{ + return test_harness(test_denormal_fpu, "fpu_denormal"); +} diff --git a/tools/testing/selftests/powerpc/math/vmx_preempt.c b/tools/testing/selftests/powerpc/math/vmx_preempt.c index 2e059f154e77..6761d6ce30ec 100644 --- a/tools/testing/selftests/powerpc/math/vmx_preempt.c +++ b/tools/testing/selftests/powerpc/math/vmx_preempt.c @@ -57,6 +57,9 @@ int test_preempt_vmx(void) int i, rc, threads; pthread_t *tids; + // vcmpequd used in vmx_asm.S is v2.07 + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07)); + threads = sysconf(_SC_NPROCESSORS_ONLN) * THREAD_FACTOR; tids = malloc(threads * sizeof(pthread_t)); FAIL_IF(!tids); diff --git a/tools/testing/selftests/powerpc/math/vmx_signal.c b/tools/testing/selftests/powerpc/math/vmx_signal.c index 785a48e0976f..b340a5c4e79d 100644 --- a/tools/testing/selftests/powerpc/math/vmx_signal.c +++ b/tools/testing/selftests/powerpc/math/vmx_signal.c @@ -96,6 +96,9 @@ int test_signal_vmx(void) void *rc_p; pthread_t *tids; + // vcmpequd used in vmx_asm.S is v2.07 + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07)); + threads = sysconf(_SC_NPROCESSORS_ONLN) * THREAD_FACTOR; tids = malloc(threads * sizeof(pthread_t)); FAIL_IF(!tids); diff --git a/tools/testing/selftests/powerpc/math/vmx_syscall.c b/tools/testing/selftests/powerpc/math/vmx_syscall.c index 9ee293cc868e..03c78dfe3444 100644 --- a/tools/testing/selftests/powerpc/math/vmx_syscall.c +++ b/tools/testing/selftests/powerpc/math/vmx_syscall.c @@ -49,9 +49,14 @@ int test_vmx_syscall(void) * Setup an environment with much context switching */ pid_t pid2; - pid_t pid = fork(); + pid_t pid; int ret; int child_ret; + + // vcmpequd used in vmx_asm.S is v2.07 + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07)); + + pid = fork(); FAIL_IF(pid == -1); pid2 = fork(); diff --git a/tools/testing/selftests/powerpc/math/vsx_preempt.c b/tools/testing/selftests/powerpc/math/vsx_preempt.c index 63de9c6e2cd3..d1601bb889d4 100644 --- a/tools/testing/selftests/powerpc/math/vsx_preempt.c +++ b/tools/testing/selftests/powerpc/math/vsx_preempt.c @@ -92,6 +92,8 @@ int test_preempt_vsx(void) int i, rc, threads; pthread_t *tids; + SKIP_IF(!have_hwcap(PPC_FEATURE_HAS_VSX)); + threads = sysconf(_SC_NPROCESSORS_ONLN) * THREAD_FACTOR; tids = malloc(threads * sizeof(pthread_t)); FAIL_IF(!tids); diff --git a/tools/testing/selftests/powerpc/mm/.gitignore b/tools/testing/selftests/powerpc/mm/.gitignore index 2ca523255b1b..91c775c23c66 100644 --- a/tools/testing/selftests/powerpc/mm/.gitignore +++ b/tools/testing/selftests/powerpc/mm/.gitignore @@ -2,9 +2,12 @@ hugetlb_vs_thp_test subpage_prot tempfile -prot_sao segv_errors wild_bctr large_vm_fork_separation bad_accesses tlbie_test +pkey_exec_prot +pkey_siginfo +stack_expansion_ldst +stack_expansion_signal diff --git a/tools/testing/selftests/powerpc/mm/Makefile b/tools/testing/selftests/powerpc/mm/Makefile index b9103c4bb414..250ce172e0da 100644 --- a/tools/testing/selftests/powerpc/mm/Makefile +++ b/tools/testing/selftests/powerpc/mm/Makefile @@ -2,23 +2,31 @@ noarg: $(MAKE) -C ../ -TEST_GEN_PROGS := hugetlb_vs_thp_test subpage_prot prot_sao segv_errors wild_bctr \ - large_vm_fork_separation bad_accesses +TEST_GEN_PROGS := hugetlb_vs_thp_test subpage_prot segv_errors wild_bctr \ + large_vm_fork_separation bad_accesses pkey_exec_prot \ + pkey_siginfo stack_expansion_signal stack_expansion_ldst + TEST_GEN_PROGS_EXTENDED := tlbie_test TEST_GEN_FILES := tempfile top_srcdir = ../../../../.. include ../../lib.mk -$(TEST_GEN_PROGS): ../harness.c - -$(OUTPUT)/prot_sao: ../utils.c +$(TEST_GEN_PROGS): ../harness.c ../utils.c $(OUTPUT)/wild_bctr: CFLAGS += -m64 $(OUTPUT)/large_vm_fork_separation: CFLAGS += -m64 $(OUTPUT)/bad_accesses: CFLAGS += -m64 +$(OUTPUT)/pkey_exec_prot: CFLAGS += -m64 +$(OUTPUT)/pkey_siginfo: CFLAGS += -m64 + +$(OUTPUT)/stack_expansion_signal: ../utils.c ../pmu/lib.c + +$(OUTPUT)/stack_expansion_ldst: CFLAGS += -fno-stack-protector +$(OUTPUT)/stack_expansion_ldst: ../utils.c $(OUTPUT)/tempfile: dd if=/dev/zero of=$@ bs=64k count=1 $(OUTPUT)/tlbie_test: LDLIBS += -lpthread +$(OUTPUT)/pkey_siginfo: LDLIBS += -lpthread diff --git a/tools/testing/selftests/powerpc/mm/bad_accesses.c b/tools/testing/selftests/powerpc/mm/bad_accesses.c index adc465f499ef..a864ed7e2008 100644 --- a/tools/testing/selftests/powerpc/mm/bad_accesses.c +++ b/tools/testing/selftests/powerpc/mm/bad_accesses.c @@ -64,34 +64,6 @@ int bad_access(char *p, bool write) return 0; } -static int using_hash_mmu(bool *using_hash) -{ - char line[128]; - FILE *f; - int rc; - - f = fopen("/proc/cpuinfo", "r"); - FAIL_IF(!f); - - rc = 0; - while (fgets(line, sizeof(line), f) != NULL) { - if (strcmp(line, "MMU : Hash\n") == 0) { - *using_hash = true; - goto out; - } - - if (strcmp(line, "MMU : Radix\n") == 0) { - *using_hash = false; - goto out; - } - } - - rc = -1; -out: - fclose(f); - return rc; -} - static int test(void) { unsigned long i, j, addr, region_shift, page_shift, page_size; diff --git a/tools/testing/selftests/powerpc/mm/pkey_exec_prot.c b/tools/testing/selftests/powerpc/mm/pkey_exec_prot.c new file mode 100644 index 000000000000..9e5c7f3f498a --- /dev/null +++ b/tools/testing/selftests/powerpc/mm/pkey_exec_prot.c @@ -0,0 +1,294 @@ +// SPDX-License-Identifier: GPL-2.0+ + +/* + * Copyright 2020, Sandipan Das, IBM Corp. + * + * Test if applying execute protection on pages using memory + * protection keys works as expected. + */ + +#define _GNU_SOURCE +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <signal.h> + +#include <unistd.h> + +#include "pkeys.h" + +#define PPC_INST_NOP 0x60000000 +#define PPC_INST_TRAP 0x7fe00008 +#define PPC_INST_BLR 0x4e800020 + +static volatile sig_atomic_t fault_pkey, fault_code, fault_type; +static volatile sig_atomic_t remaining_faults; +static volatile unsigned int *fault_addr; +static unsigned long pgsize, numinsns; +static unsigned int *insns; + +static void trap_handler(int signum, siginfo_t *sinfo, void *ctx) +{ + /* Check if this fault originated from the expected address */ + if (sinfo->si_addr != (void *) fault_addr) + sigsafe_err("got a fault for an unexpected address\n"); + + _exit(1); +} + +static void segv_handler(int signum, siginfo_t *sinfo, void *ctx) +{ + int signal_pkey; + + signal_pkey = siginfo_pkey(sinfo); + fault_code = sinfo->si_code; + + /* Check if this fault originated from the expected address */ + if (sinfo->si_addr != (void *) fault_addr) { + sigsafe_err("got a fault for an unexpected address\n"); + _exit(1); + } + + /* Check if too many faults have occurred for a single test case */ + if (!remaining_faults) { + sigsafe_err("got too many faults for the same address\n"); + _exit(1); + } + + + /* Restore permissions in order to continue */ + switch (fault_code) { + case SEGV_ACCERR: + if (mprotect(insns, pgsize, PROT_READ | PROT_WRITE)) { + sigsafe_err("failed to set access permissions\n"); + _exit(1); + } + break; + case SEGV_PKUERR: + if (signal_pkey != fault_pkey) { + sigsafe_err("got a fault for an unexpected pkey\n"); + _exit(1); + } + + switch (fault_type) { + case PKEY_DISABLE_ACCESS: + pkey_set_rights(fault_pkey, 0); + break; + case PKEY_DISABLE_EXECUTE: + /* + * Reassociate the exec-only pkey with the region + * to be able to continue. Unlike AMR, we cannot + * set IAMR directly from userspace to restore the + * permissions. + */ + if (mprotect(insns, pgsize, PROT_EXEC)) { + sigsafe_err("failed to set execute permissions\n"); + _exit(1); + } + break; + default: + sigsafe_err("got a fault with an unexpected type\n"); + _exit(1); + } + break; + default: + sigsafe_err("got a fault with an unexpected code\n"); + _exit(1); + } + + remaining_faults--; +} + +static int test(void) +{ + struct sigaction segv_act, trap_act; + unsigned long rights; + int pkey, ret, i; + + ret = pkeys_unsupported(); + if (ret) + return ret; + + /* Setup SIGSEGV handler */ + segv_act.sa_handler = 0; + segv_act.sa_sigaction = segv_handler; + FAIL_IF(sigprocmask(SIG_SETMASK, 0, &segv_act.sa_mask) != 0); + segv_act.sa_flags = SA_SIGINFO; + segv_act.sa_restorer = 0; + FAIL_IF(sigaction(SIGSEGV, &segv_act, NULL) != 0); + + /* Setup SIGTRAP handler */ + trap_act.sa_handler = 0; + trap_act.sa_sigaction = trap_handler; + FAIL_IF(sigprocmask(SIG_SETMASK, 0, &trap_act.sa_mask) != 0); + trap_act.sa_flags = SA_SIGINFO; + trap_act.sa_restorer = 0; + FAIL_IF(sigaction(SIGTRAP, &trap_act, NULL) != 0); + + /* Setup executable region */ + pgsize = getpagesize(); + numinsns = pgsize / sizeof(unsigned int); + insns = (unsigned int *) mmap(NULL, pgsize, PROT_READ | PROT_WRITE, + MAP_PRIVATE | MAP_ANONYMOUS, -1, 0); + FAIL_IF(insns == MAP_FAILED); + + /* Write the instruction words */ + for (i = 1; i < numinsns - 1; i++) + insns[i] = PPC_INST_NOP; + + /* + * Set the first instruction as an unconditional trap. If + * the last write to this address succeeds, this should + * get overwritten by a no-op. + */ + insns[0] = PPC_INST_TRAP; + + /* + * Later, to jump to the executable region, we use a branch + * and link instruction (bctrl) which sets the return address + * automatically in LR. Use that to return back. + */ + insns[numinsns - 1] = PPC_INST_BLR; + + /* Allocate a pkey that restricts execution */ + rights = PKEY_DISABLE_EXECUTE; + pkey = sys_pkey_alloc(0, rights); + FAIL_IF(pkey < 0); + + /* + * Pick the first instruction's address from the executable + * region. + */ + fault_addr = insns; + + /* The following two cases will avoid SEGV_PKUERR */ + fault_type = -1; + fault_pkey = -1; + + /* + * Read an instruction word from the address when AMR bits + * are not set i.e. the pkey permits both read and write + * access. + * + * This should not generate a fault as having PROT_EXEC + * implies PROT_READ on GNU systems. The pkey currently + * restricts execution only based on the IAMR bits. The + * AMR bits are cleared. + */ + remaining_faults = 0; + FAIL_IF(sys_pkey_mprotect(insns, pgsize, PROT_EXEC, pkey) != 0); + printf("read from %p, pkey permissions are %s\n", fault_addr, + pkey_rights(rights)); + i = *fault_addr; + FAIL_IF(remaining_faults != 0); + + /* + * Write an instruction word to the address when AMR bits + * are not set i.e. the pkey permits both read and write + * access. + * + * This should generate an access fault as having just + * PROT_EXEC also restricts writes. The pkey currently + * restricts execution only based on the IAMR bits. The + * AMR bits are cleared. + */ + remaining_faults = 1; + FAIL_IF(sys_pkey_mprotect(insns, pgsize, PROT_EXEC, pkey) != 0); + printf("write to %p, pkey permissions are %s\n", fault_addr, + pkey_rights(rights)); + *fault_addr = PPC_INST_TRAP; + FAIL_IF(remaining_faults != 0 || fault_code != SEGV_ACCERR); + + /* The following three cases will generate SEGV_PKUERR */ + rights |= PKEY_DISABLE_ACCESS; + fault_type = PKEY_DISABLE_ACCESS; + fault_pkey = pkey; + + /* + * Read an instruction word from the address when AMR bits + * are set i.e. the pkey permits neither read nor write + * access. + * + * This should generate a pkey fault based on AMR bits only + * as having PROT_EXEC implicitly allows reads. + */ + remaining_faults = 1; + FAIL_IF(sys_pkey_mprotect(insns, pgsize, PROT_EXEC, pkey) != 0); + pkey_set_rights(pkey, rights); + printf("read from %p, pkey permissions are %s\n", fault_addr, + pkey_rights(rights)); + i = *fault_addr; + FAIL_IF(remaining_faults != 0 || fault_code != SEGV_PKUERR); + + /* + * Write an instruction word to the address when AMR bits + * are set i.e. the pkey permits neither read nor write + * access. + * + * This should generate two faults. First, a pkey fault + * based on AMR bits and then an access fault since + * PROT_EXEC does not allow writes. + */ + remaining_faults = 2; + FAIL_IF(sys_pkey_mprotect(insns, pgsize, PROT_EXEC, pkey) != 0); + pkey_set_rights(pkey, rights); + printf("write to %p, pkey permissions are %s\n", fault_addr, + pkey_rights(rights)); + *fault_addr = PPC_INST_NOP; + FAIL_IF(remaining_faults != 0 || fault_code != SEGV_ACCERR); + + /* Free the current pkey */ + sys_pkey_free(pkey); + + rights = 0; + do { + /* + * Allocate pkeys with all valid combinations of read, + * write and execute restrictions. + */ + pkey = sys_pkey_alloc(0, rights); + FAIL_IF(pkey < 0); + + /* + * Jump to the executable region. AMR bits may or may not + * be set but they should not affect execution. + * + * This should generate pkey faults based on IAMR bits which + * may be set to restrict execution. + * + * The first iteration also checks if the overwrite of the + * first instruction word from a trap to a no-op succeeded. + */ + fault_pkey = pkey; + fault_type = -1; + remaining_faults = 0; + if (rights & PKEY_DISABLE_EXECUTE) { + fault_type = PKEY_DISABLE_EXECUTE; + remaining_faults = 1; + } + + FAIL_IF(sys_pkey_mprotect(insns, pgsize, PROT_EXEC, pkey) != 0); + printf("execute at %p, pkey permissions are %s\n", fault_addr, + pkey_rights(rights)); + asm volatile("mtctr %0; bctrl" : : "r"(insns)); + FAIL_IF(remaining_faults != 0); + if (rights & PKEY_DISABLE_EXECUTE) + FAIL_IF(fault_code != SEGV_PKUERR); + + /* Free the current pkey */ + sys_pkey_free(pkey); + + /* Find next valid combination of pkey rights */ + rights = next_pkey_rights(rights); + } while (rights); + + /* Cleanup */ + munmap((void *) insns, pgsize); + + return 0; +} + +int main(void) +{ + test_harness(test, "pkey_exec_prot"); +} diff --git a/tools/testing/selftests/powerpc/mm/pkey_siginfo.c b/tools/testing/selftests/powerpc/mm/pkey_siginfo.c new file mode 100644 index 000000000000..4f815d7c1214 --- /dev/null +++ b/tools/testing/selftests/powerpc/mm/pkey_siginfo.c @@ -0,0 +1,333 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright 2020, Sandipan Das, IBM Corp. + * + * Test if the signal information reports the correct memory protection + * key upon getting a key access violation fault for a page that was + * attempted to be protected by two different keys from two competing + * threads at the same time. + */ + +#define _GNU_SOURCE +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <signal.h> + +#include <unistd.h> +#include <pthread.h> +#include <sys/mman.h> + +#include "pkeys.h" + +#define PPC_INST_NOP 0x60000000 +#define PPC_INST_BLR 0x4e800020 +#define PROT_RWX (PROT_READ | PROT_WRITE | PROT_EXEC) + +#define NUM_ITERATIONS 1000000 + +static volatile sig_atomic_t perm_pkey, rest_pkey; +static volatile sig_atomic_t rights, fault_count; +static volatile unsigned int *volatile fault_addr; +static pthread_barrier_t iteration_barrier; + +static void segv_handler(int signum, siginfo_t *sinfo, void *ctx) +{ + void *pgstart; + size_t pgsize; + int pkey; + + pkey = siginfo_pkey(sinfo); + + /* Check if this fault originated from a pkey access violation */ + if (sinfo->si_code != SEGV_PKUERR) { + sigsafe_err("got a fault for an unexpected reason\n"); + _exit(1); + } + + /* Check if this fault originated from the expected address */ + if (sinfo->si_addr != (void *) fault_addr) { + sigsafe_err("got a fault for an unexpected address\n"); + _exit(1); + } + + /* Check if this fault originated from the restrictive pkey */ + if (pkey != rest_pkey) { + sigsafe_err("got a fault for an unexpected pkey\n"); + _exit(1); + } + + /* Check if too many faults have occurred for the same iteration */ + if (fault_count > 0) { + sigsafe_err("got too many faults for the same address\n"); + _exit(1); + } + + pgsize = getpagesize(); + pgstart = (void *) ((unsigned long) fault_addr & ~(pgsize - 1)); + + /* + * If the current fault occurred due to lack of execute rights, + * reassociate the page with the exec-only pkey since execute + * rights cannot be changed directly for the faulting pkey as + * IAMR is inaccessible from userspace. + * + * Otherwise, if the current fault occurred due to lack of + * read-write rights, change the AMR permission bits for the + * pkey. + * + * This will let the test continue. + */ + if (rights == PKEY_DISABLE_EXECUTE && + mprotect(pgstart, pgsize, PROT_EXEC)) + _exit(1); + else + pkey_set_rights(pkey, 0); + + fault_count++; +} + +struct region { + unsigned long rights; + unsigned int *base; + size_t size; +}; + +static void *protect(void *p) +{ + unsigned long rights; + unsigned int *base; + size_t size; + int tid, i; + + tid = gettid(); + base = ((struct region *) p)->base; + size = ((struct region *) p)->size; + FAIL_IF_EXIT(!base); + + /* No read, write and execute restrictions */ + rights = 0; + + printf("tid %d, pkey permissions are %s\n", tid, pkey_rights(rights)); + + /* Allocate the permissive pkey */ + perm_pkey = sys_pkey_alloc(0, rights); + FAIL_IF_EXIT(perm_pkey < 0); + + /* + * Repeatedly try to protect the common region with a permissive + * pkey + */ + for (i = 0; i < NUM_ITERATIONS; i++) { + /* + * Wait until the other thread has finished allocating the + * restrictive pkey or until the next iteration has begun + */ + pthread_barrier_wait(&iteration_barrier); + + /* Try to associate the permissive pkey with the region */ + FAIL_IF_EXIT(sys_pkey_mprotect(base, size, PROT_RWX, + perm_pkey)); + } + + /* Free the permissive pkey */ + sys_pkey_free(perm_pkey); + + return NULL; +} + +static void *protect_access(void *p) +{ + size_t size, numinsns; + unsigned int *base; + int tid, i; + + tid = gettid(); + base = ((struct region *) p)->base; + size = ((struct region *) p)->size; + rights = ((struct region *) p)->rights; + numinsns = size / sizeof(base[0]); + FAIL_IF_EXIT(!base); + + /* Allocate the restrictive pkey */ + rest_pkey = sys_pkey_alloc(0, rights); + FAIL_IF_EXIT(rest_pkey < 0); + + printf("tid %d, pkey permissions are %s\n", tid, pkey_rights(rights)); + printf("tid %d, %s randomly in range [%p, %p]\n", tid, + (rights == PKEY_DISABLE_EXECUTE) ? "execute" : + (rights == PKEY_DISABLE_WRITE) ? "write" : "read", + base, base + numinsns); + + /* + * Repeatedly try to protect the common region with a restrictive + * pkey and read, write or execute from it + */ + for (i = 0; i < NUM_ITERATIONS; i++) { + /* + * Wait until the other thread has finished allocating the + * permissive pkey or until the next iteration has begun + */ + pthread_barrier_wait(&iteration_barrier); + + /* Try to associate the restrictive pkey with the region */ + FAIL_IF_EXIT(sys_pkey_mprotect(base, size, PROT_RWX, + rest_pkey)); + + /* Choose a random instruction word address from the region */ + fault_addr = base + (rand() % numinsns); + fault_count = 0; + + switch (rights) { + /* Read protection test */ + case PKEY_DISABLE_ACCESS: + /* + * Read an instruction word from the region and + * verify if it has not been overwritten to + * something unexpected + */ + FAIL_IF_EXIT(*fault_addr != PPC_INST_NOP && + *fault_addr != PPC_INST_BLR); + break; + + /* Write protection test */ + case PKEY_DISABLE_WRITE: + /* + * Write an instruction word to the region and + * verify if the overwrite has succeeded + */ + *fault_addr = PPC_INST_BLR; + FAIL_IF_EXIT(*fault_addr != PPC_INST_BLR); + break; + + /* Execute protection test */ + case PKEY_DISABLE_EXECUTE: + /* Jump to the region and execute instructions */ + asm volatile( + "mtctr %0; bctrl" + : : "r"(fault_addr) : "ctr", "lr"); + break; + } + + /* + * Restore the restrictions originally imposed by the + * restrictive pkey as the signal handler would have + * cleared out the corresponding AMR bits + */ + pkey_set_rights(rest_pkey, rights); + } + + /* Free restrictive pkey */ + sys_pkey_free(rest_pkey); + + return NULL; +} + +static void reset_pkeys(unsigned long rights) +{ + int pkeys[NR_PKEYS], i; + + /* Exhaustively allocate all available pkeys */ + for (i = 0; i < NR_PKEYS; i++) + pkeys[i] = sys_pkey_alloc(0, rights); + + /* Free all allocated pkeys */ + for (i = 0; i < NR_PKEYS; i++) + sys_pkey_free(pkeys[i]); +} + +static int test(void) +{ + pthread_t prot_thread, pacc_thread; + struct sigaction act; + pthread_attr_t attr; + size_t numinsns; + struct region r; + int ret, i; + + srand(time(NULL)); + ret = pkeys_unsupported(); + if (ret) + return ret; + + /* Allocate the region */ + r.size = getpagesize(); + r.base = mmap(NULL, r.size, PROT_RWX, + MAP_PRIVATE | MAP_ANONYMOUS, -1, 0); + FAIL_IF(r.base == MAP_FAILED); + + /* + * Fill the region with no-ops with a branch at the end + * for returning to the caller + */ + numinsns = r.size / sizeof(r.base[0]); + for (i = 0; i < numinsns - 1; i++) + r.base[i] = PPC_INST_NOP; + r.base[i] = PPC_INST_BLR; + + /* Setup SIGSEGV handler */ + act.sa_handler = 0; + act.sa_sigaction = segv_handler; + FAIL_IF(sigprocmask(SIG_SETMASK, 0, &act.sa_mask) != 0); + act.sa_flags = SA_SIGINFO; + act.sa_restorer = 0; + FAIL_IF(sigaction(SIGSEGV, &act, NULL) != 0); + + /* + * For these tests, the parent process should clear all bits of + * AMR and IAMR, i.e. impose no restrictions, for all available + * pkeys. This will be the base for the initial AMR and IAMR + * values for all the test thread pairs. + * + * If the AMR and IAMR bits of all available pkeys are cleared + * before running the tests and a fault is generated when + * attempting to read, write or execute instructions from a + * pkey protected region, the pkey responsible for this must be + * the one from the protect-and-access thread since the other + * one is fully permissive. Despite that, if the pkey reported + * by siginfo is not the restrictive pkey, then there must be a + * kernel bug. + */ + reset_pkeys(0); + + /* Setup barrier for protect and protect-and-access threads */ + FAIL_IF(pthread_attr_init(&attr) != 0); + FAIL_IF(pthread_barrier_init(&iteration_barrier, NULL, 2) != 0); + + /* Setup and start protect and protect-and-read threads */ + puts("starting thread pair (protect, protect-and-read)"); + r.rights = PKEY_DISABLE_ACCESS; + FAIL_IF(pthread_create(&prot_thread, &attr, &protect, &r) != 0); + FAIL_IF(pthread_create(&pacc_thread, &attr, &protect_access, &r) != 0); + FAIL_IF(pthread_join(prot_thread, NULL) != 0); + FAIL_IF(pthread_join(pacc_thread, NULL) != 0); + + /* Setup and start protect and protect-and-write threads */ + puts("starting thread pair (protect, protect-and-write)"); + r.rights = PKEY_DISABLE_WRITE; + FAIL_IF(pthread_create(&prot_thread, &attr, &protect, &r) != 0); + FAIL_IF(pthread_create(&pacc_thread, &attr, &protect_access, &r) != 0); + FAIL_IF(pthread_join(prot_thread, NULL) != 0); + FAIL_IF(pthread_join(pacc_thread, NULL) != 0); + + /* Setup and start protect and protect-and-execute threads */ + puts("starting thread pair (protect, protect-and-execute)"); + r.rights = PKEY_DISABLE_EXECUTE; + FAIL_IF(pthread_create(&prot_thread, &attr, &protect, &r) != 0); + FAIL_IF(pthread_create(&pacc_thread, &attr, &protect_access, &r) != 0); + FAIL_IF(pthread_join(prot_thread, NULL) != 0); + FAIL_IF(pthread_join(pacc_thread, NULL) != 0); + + /* Cleanup */ + FAIL_IF(pthread_attr_destroy(&attr) != 0); + FAIL_IF(pthread_barrier_destroy(&iteration_barrier) != 0); + munmap(r.base, r.size); + + return 0; +} + +int main(void) +{ + test_harness(test, "pkey_siginfo"); +} diff --git a/tools/testing/selftests/powerpc/mm/prot_sao.c b/tools/testing/selftests/powerpc/mm/prot_sao.c deleted file mode 100644 index e2eed65b7735..000000000000 --- a/tools/testing/selftests/powerpc/mm/prot_sao.c +++ /dev/null @@ -1,42 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0-only -/* - * Copyright 2016, Michael Ellerman, IBM Corp. - */ - -#include <stdio.h> -#include <stdlib.h> -#include <string.h> -#include <sys/mman.h> - -#include <asm/cputable.h> - -#include "utils.h" - -#define SIZE (64 * 1024) - -int test_prot_sao(void) -{ - char *p; - - /* 2.06 or later should support SAO */ - SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_06)); - - /* - * Ensure we can ask for PROT_SAO. - * We can't really verify that it does the right thing, but at least we - * confirm the kernel will accept it. - */ - p = mmap(NULL, SIZE, PROT_READ | PROT_WRITE | PROT_SAO, - MAP_ANONYMOUS | MAP_PRIVATE, -1, 0); - FAIL_IF(p == MAP_FAILED); - - /* Write to the mapping, to at least cause a fault */ - memset(p, 0xaa, SIZE); - - return 0; -} - -int main(void) -{ - return test_harness(test_prot_sao, "prot-sao"); -} diff --git a/tools/testing/selftests/powerpc/mm/stack_expansion_ldst.c b/tools/testing/selftests/powerpc/mm/stack_expansion_ldst.c new file mode 100644 index 000000000000..ed9143990888 --- /dev/null +++ b/tools/testing/selftests/powerpc/mm/stack_expansion_ldst.c @@ -0,0 +1,202 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Test that loads/stores expand the stack segment, or trigger a SEGV, in + * various conditions. + * + * Based on test code by Tom Lane. + */ + +#undef NDEBUG +#include <assert.h> + +#include <err.h> +#include <errno.h> +#include <stdio.h> +#include <signal.h> +#include <stdlib.h> +#include <string.h> +#include <sys/resource.h> +#include <sys/time.h> +#include <sys/types.h> +#include <sys/wait.h> +#include <unistd.h> + +#define _KB (1024) +#define _MB (1024 * 1024) + +volatile char *stack_top_ptr; +volatile unsigned long stack_top_sp; +volatile char c; + +enum access_type { + LOAD, + STORE, +}; + +/* + * Consume stack until the stack pointer is below @target_sp, then do an access + * (load or store) at offset @delta from either the base of the stack or the + * current stack pointer. + */ +__attribute__ ((noinline)) +int consume_stack(unsigned long target_sp, unsigned long stack_high, int delta, enum access_type type) +{ + unsigned long target; + char stack_cur; + + if ((unsigned long)&stack_cur > target_sp) + return consume_stack(target_sp, stack_high, delta, type); + else { + // We don't really need this, but without it GCC might not + // generate a recursive call above. + stack_top_ptr = &stack_cur; + +#ifdef __powerpc__ + asm volatile ("mr %[sp], %%r1" : [sp] "=r" (stack_top_sp)); +#else + asm volatile ("mov %%rsp, %[sp]" : [sp] "=r" (stack_top_sp)); +#endif + target = stack_high - delta + 1; + volatile char *p = (char *)target; + + if (type == STORE) + *p = c; + else + c = *p; + + // Do something to prevent the stack frame being popped prior to + // our access above. + getpid(); + } + + return 0; +} + +static int search_proc_maps(char *needle, unsigned long *low, unsigned long *high) +{ + unsigned long start, end; + static char buf[4096]; + char name[128]; + FILE *f; + int rc; + + f = fopen("/proc/self/maps", "r"); + if (!f) { + perror("fopen"); + return -1; + } + + while (fgets(buf, sizeof(buf), f)) { + rc = sscanf(buf, "%lx-%lx %*c%*c%*c%*c %*x %*d:%*d %*d %127s\n", + &start, &end, name); + if (rc == 2) + continue; + + if (rc != 3) { + printf("sscanf errored\n"); + rc = -1; + break; + } + + if (strstr(name, needle)) { + *low = start; + *high = end - 1; + rc = 0; + break; + } + } + + fclose(f); + + return rc; +} + +int child(unsigned int stack_used, int delta, enum access_type type) +{ + unsigned long low, stack_high; + + assert(search_proc_maps("[stack]", &low, &stack_high) == 0); + + assert(consume_stack(stack_high - stack_used, stack_high, delta, type) == 0); + + printf("Access OK: %s delta %-7d used size 0x%06x stack high 0x%lx top_ptr %p top sp 0x%lx actual used 0x%lx\n", + type == LOAD ? "load" : "store", delta, stack_used, stack_high, + stack_top_ptr, stack_top_sp, stack_high - stack_top_sp + 1); + + return 0; +} + +static int test_one(unsigned int stack_used, int delta, enum access_type type) +{ + pid_t pid; + int rc; + + pid = fork(); + if (pid == 0) + exit(child(stack_used, delta, type)); + + assert(waitpid(pid, &rc, 0) != -1); + + if (WIFEXITED(rc) && WEXITSTATUS(rc) == 0) + return 0; + + // We don't expect a non-zero exit that's not a signal + assert(!WIFEXITED(rc)); + + printf("Faulted: %s delta %-7d used size 0x%06x signal %d\n", + type == LOAD ? "load" : "store", delta, stack_used, + WTERMSIG(rc)); + + return 1; +} + +// This is fairly arbitrary but is well below any of the targets below, +// so that the delta between the stack pointer and the target is large. +#define DEFAULT_SIZE (32 * _KB) + +static void test_one_type(enum access_type type, unsigned long page_size, unsigned long rlim_cur) +{ + unsigned long delta; + + // We should be able to access anywhere within the rlimit + for (delta = page_size; delta <= rlim_cur; delta += page_size) + assert(test_one(DEFAULT_SIZE, delta, type) == 0); + + assert(test_one(DEFAULT_SIZE, rlim_cur, type) == 0); + + // But if we go past the rlimit it should fail + assert(test_one(DEFAULT_SIZE, rlim_cur + 1, type) != 0); +} + +static int test(void) +{ + unsigned long page_size; + struct rlimit rlimit; + + page_size = getpagesize(); + getrlimit(RLIMIT_STACK, &rlimit); + printf("Stack rlimit is 0x%lx\n", rlimit.rlim_cur); + + printf("Testing loads ...\n"); + test_one_type(LOAD, page_size, rlimit.rlim_cur); + printf("Testing stores ...\n"); + test_one_type(STORE, page_size, rlimit.rlim_cur); + + printf("All OK\n"); + + return 0; +} + +#ifdef __powerpc__ +#include "utils.h" + +int main(void) +{ + return test_harness(test, "stack_expansion_ldst"); +} +#else +int main(void) +{ + return test(); +} +#endif diff --git a/tools/testing/selftests/powerpc/mm/stack_expansion_signal.c b/tools/testing/selftests/powerpc/mm/stack_expansion_signal.c new file mode 100644 index 000000000000..c8b32a29e274 --- /dev/null +++ b/tools/testing/selftests/powerpc/mm/stack_expansion_signal.c @@ -0,0 +1,118 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Test that signal delivery is able to expand the stack segment without + * triggering a SEGV. + * + * Based on test code by Tom Lane. + */ + +#include <err.h> +#include <stdio.h> +#include <stdlib.h> +#include <signal.h> +#include <sys/types.h> +#include <unistd.h> + +#include "../pmu/lib.h" +#include "utils.h" + +#define _KB (1024) +#define _MB (1024 * 1024) + +static char *stack_base_ptr; +static char *stack_top_ptr; + +static volatile sig_atomic_t sig_occurred = 0; + +static void sigusr1_handler(int signal_arg) +{ + sig_occurred = 1; +} + +static int consume_stack(unsigned int stack_size, union pipe write_pipe) +{ + char stack_cur; + + if ((stack_base_ptr - &stack_cur) < stack_size) + return consume_stack(stack_size, write_pipe); + else { + stack_top_ptr = &stack_cur; + + FAIL_IF(notify_parent(write_pipe)); + + while (!sig_occurred) + barrier(); + } + + return 0; +} + +static int child(unsigned int stack_size, union pipe write_pipe) +{ + struct sigaction act; + char stack_base; + + act.sa_handler = sigusr1_handler; + sigemptyset(&act.sa_mask); + act.sa_flags = 0; + if (sigaction(SIGUSR1, &act, NULL) < 0) + err(1, "sigaction"); + + stack_base_ptr = (char *) (((size_t) &stack_base + 65535) & ~65535UL); + + FAIL_IF(consume_stack(stack_size, write_pipe)); + + printf("size 0x%06x: OK, stack base %p top %p (%zx used)\n", + stack_size, stack_base_ptr, stack_top_ptr, + stack_base_ptr - stack_top_ptr); + + return 0; +} + +static int test_one_size(unsigned int stack_size) +{ + union pipe read_pipe, write_pipe; + pid_t pid; + + FAIL_IF(pipe(read_pipe.fds) == -1); + FAIL_IF(pipe(write_pipe.fds) == -1); + + pid = fork(); + if (pid == 0) { + close(read_pipe.read_fd); + close(write_pipe.write_fd); + exit(child(stack_size, read_pipe)); + } + + close(read_pipe.write_fd); + close(write_pipe.read_fd); + FAIL_IF(sync_with_child(read_pipe, write_pipe)); + + kill(pid, SIGUSR1); + + FAIL_IF(wait_for_child(pid)); + + close(read_pipe.read_fd); + close(write_pipe.write_fd); + + return 0; +} + +int test(void) +{ + unsigned int i, size; + + // Test with used stack from 1MB - 64K to 1MB + 64K + // Increment by 64 to get more coverage of odd sizes + for (i = 0; i < (128 * _KB); i += 64) { + size = i + (1 * _MB) - (64 * _KB); + FAIL_IF(test_one_size(size)); + } + + return 0; +} + +int main(void) +{ + return test_harness(test, "stack_expansion_signal"); +} diff --git a/tools/testing/selftests/powerpc/pmu/count_stcx_fail.c b/tools/testing/selftests/powerpc/pmu/count_stcx_fail.c index 7b4ac4537702..2980abca31e0 100644 --- a/tools/testing/selftests/powerpc/pmu/count_stcx_fail.c +++ b/tools/testing/selftests/powerpc/pmu/count_stcx_fail.c @@ -9,6 +9,7 @@ #include <stdbool.h> #include <string.h> #include <sys/prctl.h> +#include <asm/cputable.h> #include "event.h" #include "utils.h" @@ -104,6 +105,9 @@ static int test_body(void) struct event events[3]; u64 overhead; + // The STCX_FAIL event we use works on Power8 or later + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07)); + setup_event(&events[0], PERF_COUNT_HW_INSTRUCTIONS, PERF_TYPE_HARDWARE, "instructions"); setup_event(&events[1], PERF_COUNT_HW_CPU_CYCLES, PERF_TYPE_HARDWARE, "cycles"); setup_event(&events[2], PM_STCX_FAIL, PERF_TYPE_RAW, "stcx_fail"); diff --git a/tools/testing/selftests/powerpc/pmu/ebb/back_to_back_ebbs_test.c b/tools/testing/selftests/powerpc/pmu/ebb/back_to_back_ebbs_test.c index a2d7b0e3dca9..a26ac122c759 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/back_to_back_ebbs_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/back_to_back_ebbs_test.c @@ -91,8 +91,6 @@ int back_to_back_ebbs(void) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(1, sample_period); - dump_ebb_state(); event_close(&event); diff --git a/tools/testing/selftests/powerpc/pmu/ebb/cycles_test.c b/tools/testing/selftests/powerpc/pmu/ebb/cycles_test.c index bc893813483e..bb9f587fa76e 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/cycles_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/cycles_test.c @@ -42,8 +42,6 @@ int cycles(void) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(1, sample_period); - dump_ebb_state(); event_close(&event); diff --git a/tools/testing/selftests/powerpc/pmu/ebb/cycles_with_freeze_test.c b/tools/testing/selftests/powerpc/pmu/ebb/cycles_with_freeze_test.c index dcd351d20328..9ae795ce314e 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/cycles_with_freeze_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/cycles_with_freeze_test.c @@ -99,8 +99,6 @@ int cycles_with_freeze(void) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(1, sample_period); - dump_ebb_state(); printf("EBBs while frozen %d\n", ebbs_while_frozen); diff --git a/tools/testing/selftests/powerpc/pmu/ebb/cycles_with_mmcr2_test.c b/tools/testing/selftests/powerpc/pmu/ebb/cycles_with_mmcr2_test.c index 94c99c12c0f2..4b45a2e70f62 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/cycles_with_mmcr2_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/cycles_with_mmcr2_test.c @@ -71,8 +71,6 @@ int cycles_with_mmcr2(void) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(1, sample_period); - dump_ebb_state(); event_close(&event); diff --git a/tools/testing/selftests/powerpc/pmu/ebb/ebb.c b/tools/testing/selftests/powerpc/pmu/ebb/ebb.c index dfbc5c3ad52d..21537d6eb6b7 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/ebb.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/ebb.c @@ -396,8 +396,6 @@ int ebb_child(union pipe read_pipe, union pipe write_pipe) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(1, sample_period); - dump_ebb_state(); event_close(&event); diff --git a/tools/testing/selftests/powerpc/pmu/ebb/ebb_on_willing_child_test.c b/tools/testing/selftests/powerpc/pmu/ebb/ebb_on_willing_child_test.c index ca2f7d729155..b208bf6ad58d 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/ebb_on_willing_child_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/ebb_on_willing_child_test.c @@ -38,8 +38,6 @@ static int victim_child(union pipe read_pipe, union pipe write_pipe) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(1, sample_period); - dump_ebb_state(); FAIL_IF(ebb_state.stats.ebb_count == 0); diff --git a/tools/testing/selftests/powerpc/pmu/ebb/lost_exception_test.c b/tools/testing/selftests/powerpc/pmu/ebb/lost_exception_test.c index ac3e6e182614..ba2681a12cc7 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/lost_exception_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/lost_exception_test.c @@ -75,7 +75,6 @@ static int test_body(void) ebb_freeze_pmcs(); ebb_global_disable(); - count_pmc(4, sample_period); mtspr(SPRN_PMC4, 0xdead); dump_summary_ebb_state(); diff --git a/tools/testing/selftests/powerpc/pmu/ebb/multi_counter_test.c b/tools/testing/selftests/powerpc/pmu/ebb/multi_counter_test.c index b8242e9d97d2..791d37ba327b 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/multi_counter_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/multi_counter_test.c @@ -70,13 +70,6 @@ int multi_counter(void) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(1, sample_period); - count_pmc(2, sample_period); - count_pmc(3, sample_period); - count_pmc(4, sample_period); - count_pmc(5, sample_period); - count_pmc(6, sample_period); - dump_ebb_state(); for (i = 0; i < 6; i++) diff --git a/tools/testing/selftests/powerpc/pmu/ebb/multi_ebb_procs_test.c b/tools/testing/selftests/powerpc/pmu/ebb/multi_ebb_procs_test.c index a05c0e18ded6..9b0f70d59702 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/multi_ebb_procs_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/multi_ebb_procs_test.c @@ -61,8 +61,6 @@ static int cycles_child(void) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(1, sample_period); - dump_summary_ebb_state(); event_close(&event); diff --git a/tools/testing/selftests/powerpc/pmu/ebb/pmae_handling_test.c b/tools/testing/selftests/powerpc/pmu/ebb/pmae_handling_test.c index 153ebc92234f..2904c741e04e 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/pmae_handling_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/pmae_handling_test.c @@ -82,8 +82,6 @@ static int test_body(void) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(1, sample_period); - dump_ebb_state(); if (mmcr0_mismatch) diff --git a/tools/testing/selftests/powerpc/pmu/ebb/pmc56_overflow_test.c b/tools/testing/selftests/powerpc/pmu/ebb/pmc56_overflow_test.c index eadad75ed7e6..b29f8ba22d1e 100644 --- a/tools/testing/selftests/powerpc/pmu/ebb/pmc56_overflow_test.c +++ b/tools/testing/selftests/powerpc/pmu/ebb/pmc56_overflow_test.c @@ -76,8 +76,6 @@ int pmc56_overflow(void) ebb_global_disable(); ebb_freeze_pmcs(); - count_pmc(2, sample_period); - dump_ebb_state(); printf("PMC5/6 overflow %d\n", pmc56_overflowed); diff --git a/tools/testing/selftests/powerpc/pmu/lib.h b/tools/testing/selftests/powerpc/pmu/lib.h index fa12e7d0b4d3..bf1bec013bbb 100644 --- a/tools/testing/selftests/powerpc/pmu/lib.h +++ b/tools/testing/selftests/powerpc/pmu/lib.h @@ -6,6 +6,7 @@ #ifndef __SELFTESTS_POWERPC_PMU_LIB_H #define __SELFTESTS_POWERPC_PMU_LIB_H +#include <stdbool.h> #include <stdio.h> #include <stdint.h> #include <string.h> diff --git a/tools/testing/selftests/powerpc/pmu/per_event_excludes.c b/tools/testing/selftests/powerpc/pmu/per_event_excludes.c index 2756fe2efdc5..2d37942bf72b 100644 --- a/tools/testing/selftests/powerpc/pmu/per_event_excludes.c +++ b/tools/testing/selftests/powerpc/pmu/per_event_excludes.c @@ -12,6 +12,8 @@ #include <string.h> #include <sys/prctl.h> +#include <asm/cputable.h> + #include "event.h" #include "lib.h" #include "utils.h" @@ -23,12 +25,9 @@ static int per_event_excludes(void) { struct event *e, events[4]; - char *platform; int i; - platform = (char *)get_auxv_entry(AT_BASE_PLATFORM); - FAIL_IF(!platform); - SKIP_IF(strcmp(platform, "power8") != 0); + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07)); /* * We need to create the events disabled, otherwise the running/enabled diff --git a/tools/testing/selftests/powerpc/ptrace/core-pkey.c b/tools/testing/selftests/powerpc/ptrace/core-pkey.c index d5c64fee032d..bbc05ffc5860 100644 --- a/tools/testing/selftests/powerpc/ptrace/core-pkey.c +++ b/tools/testing/selftests/powerpc/ptrace/core-pkey.c @@ -150,7 +150,7 @@ static int child(struct shared_info *info) printf("%-30s AMR: %016lx pkey1: %d pkey2: %d pkey3: %d\n", user_write, info->amr, pkey1, pkey2, pkey3); - mtspr(SPRN_AMR, info->amr); + set_amr(info->amr); /* * We won't use pkey3. This tests whether the kernel restores the UAMOR diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-pkey.c b/tools/testing/selftests/powerpc/ptrace/ptrace-pkey.c index bdbbbe8431e0..bc454f899124 100644 --- a/tools/testing/selftests/powerpc/ptrace/ptrace-pkey.c +++ b/tools/testing/selftests/powerpc/ptrace/ptrace-pkey.c @@ -44,7 +44,7 @@ struct shared_info { unsigned long amr2; /* AMR value that ptrace should refuse to write to the child. */ - unsigned long amr3; + unsigned long invalid_amr; /* IAMR value the parent expects to read from the child. */ unsigned long expected_iamr; @@ -57,8 +57,8 @@ struct shared_info { * (even though they're valid ones) because userspace doesn't have * access to those registers. */ - unsigned long new_iamr; - unsigned long new_uamor; + unsigned long invalid_iamr; + unsigned long invalid_uamor; }; static int sys_pkey_alloc(unsigned long flags, unsigned long init_access_rights) @@ -66,11 +66,6 @@ static int sys_pkey_alloc(unsigned long flags, unsigned long init_access_rights) return syscall(__NR_pkey_alloc, flags, init_access_rights); } -static int sys_pkey_free(int pkey) -{ - return syscall(__NR_pkey_free, pkey); -} - static int child(struct shared_info *info) { unsigned long reg; @@ -100,33 +95,37 @@ static int child(struct shared_info *info) info->amr1 |= 3ul << pkeyshift(pkey1); info->amr2 |= 3ul << pkeyshift(pkey2); - info->amr3 |= info->amr2 | 3ul << pkeyshift(pkey3); + /* + * invalid amr value where we try to force write + * things which are deined by a uamor setting. + */ + info->invalid_amr = info->amr2 | (~0x0UL & ~info->expected_uamor); + /* + * if PKEY_DISABLE_EXECUTE succeeded we should update the expected_iamr + */ if (disable_execute) info->expected_iamr |= 1ul << pkeyshift(pkey1); else info->expected_iamr &= ~(1ul << pkeyshift(pkey1)); - info->expected_iamr &= ~(1ul << pkeyshift(pkey2) | 1ul << pkeyshift(pkey3)); - - info->expected_uamor |= 3ul << pkeyshift(pkey1) | - 3ul << pkeyshift(pkey2); - info->new_iamr |= 1ul << pkeyshift(pkey1) | 1ul << pkeyshift(pkey2); - info->new_uamor |= 3ul << pkeyshift(pkey1); + /* + * We allocated pkey2 and pkey 3 above. Clear the IAMR bits. + */ + info->expected_iamr &= ~(1ul << pkeyshift(pkey2)); + info->expected_iamr &= ~(1ul << pkeyshift(pkey3)); /* - * We won't use pkey3. We just want a plausible but invalid key to test - * whether ptrace will let us write to AMR bits we are not supposed to. - * - * This also tests whether the kernel restores the UAMOR permissions - * after a key is freed. + * Create an IAMR value different from expected value. + * Kernel will reject an IAMR and UAMOR change. */ - sys_pkey_free(pkey3); + info->invalid_iamr = info->expected_iamr | (1ul << pkeyshift(pkey1) | 1ul << pkeyshift(pkey2)); + info->invalid_uamor = info->expected_uamor & ~(0x3ul << pkeyshift(pkey1)); printf("%-30s AMR: %016lx pkey1: %d pkey2: %d pkey3: %d\n", user_write, info->amr1, pkey1, pkey2, pkey3); - mtspr(SPRN_AMR, info->amr1); + set_amr(info->amr1); /* Wait for parent to read our AMR value and write a new one. */ ret = prod_parent(&info->child_sync); @@ -196,9 +195,9 @@ static int parent(struct shared_info *info, pid_t pid) PARENT_SKIP_IF_UNSUPPORTED(ret, &info->child_sync); PARENT_FAIL_IF(ret, &info->child_sync); - info->amr1 = info->amr2 = info->amr3 = regs[0]; - info->expected_iamr = info->new_iamr = regs[1]; - info->expected_uamor = info->new_uamor = regs[2]; + info->amr1 = info->amr2 = regs[0]; + info->expected_iamr = regs[1]; + info->expected_uamor = regs[2]; /* Wake up child so that it can set itself up. */ ret = prod_child(&info->child_sync); @@ -234,10 +233,10 @@ static int parent(struct shared_info *info, pid_t pid) return ret; /* Write invalid AMR value in child. */ - ret = ptrace_write_regs(pid, NT_PPC_PKEY, &info->amr3, 1); + ret = ptrace_write_regs(pid, NT_PPC_PKEY, &info->invalid_amr, 1); PARENT_FAIL_IF(ret, &info->child_sync); - printf("%-30s AMR: %016lx\n", ptrace_write_running, info->amr3); + printf("%-30s AMR: %016lx\n", ptrace_write_running, info->invalid_amr); /* Wake up child so that it can verify it didn't change. */ ret = prod_child(&info->child_sync); @@ -249,7 +248,7 @@ static int parent(struct shared_info *info, pid_t pid) /* Try to write to IAMR. */ regs[0] = info->amr1; - regs[1] = info->new_iamr; + regs[1] = info->invalid_iamr; ret = ptrace_write_regs(pid, NT_PPC_PKEY, regs, 2); PARENT_FAIL_IF(!ret, &info->child_sync); @@ -257,7 +256,7 @@ static int parent(struct shared_info *info, pid_t pid) ptrace_write_running, regs[0], regs[1]); /* Try to write to IAMR and UAMOR. */ - regs[2] = info->new_uamor; + regs[2] = info->invalid_uamor; ret = ptrace_write_regs(pid, NT_PPC_PKEY, regs, 3); PARENT_FAIL_IF(!ret, &info->child_sync); diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-tar.c b/tools/testing/selftests/powerpc/ptrace/ptrace-tar.c index 58cb1a860cc9..4436ca9d3caf 100644 --- a/tools/testing/selftests/powerpc/ptrace/ptrace-tar.c +++ b/tools/testing/selftests/powerpc/ptrace/ptrace-tar.c @@ -78,6 +78,9 @@ int ptrace_tar(void) pid_t pid; int ret, status; + // TAR was added in v2.07 + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07)); + shm_id = shmget(IPC_PRIVATE, sizeof(int) * 3, 0777|IPC_CREAT); pid = fork(); if (pid < 0) { diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-vsx.c b/tools/testing/selftests/powerpc/ptrace/ptrace-vsx.c index c4fe0e893306..cb9875f764ca 100644 --- a/tools/testing/selftests/powerpc/ptrace/ptrace-vsx.c +++ b/tools/testing/selftests/powerpc/ptrace/ptrace-vsx.c @@ -61,6 +61,8 @@ int ptrace_vsx(void) pid_t pid; int ret, status, i; + SKIP_IF(!have_hwcap(PPC_FEATURE_HAS_VSX)); + shm_id = shmget(IPC_PRIVATE, sizeof(int) * 2, 0777|IPC_CREAT); for (i = 0; i < VEC_MAX; i++) diff --git a/tools/testing/selftests/powerpc/security/spectre_v2.c b/tools/testing/selftests/powerpc/security/spectre_v2.c index 8c6b982af2a8..c8d82b784102 100644 --- a/tools/testing/selftests/powerpc/security/spectre_v2.c +++ b/tools/testing/selftests/powerpc/security/spectre_v2.c @@ -183,6 +183,16 @@ int spectre_v2_test(void) if (miss_percent > 15) { printf("Branch misses > 15%% unexpected in this configuration!\n"); printf("Possible mis-match between reported & actual mitigation\n"); + /* + * Such a mismatch may be caused by a guest system + * reporting as vulnerable when the host is mitigated. + * Return skip code to avoid detecting this as an error. + * We are not vulnerable and reporting otherwise, so + * missing such a mismatch is safe. + */ + if (state == VULNERABLE) + return 4; + return 1; } break; diff --git a/tools/testing/selftests/powerpc/stringloops/Makefile b/tools/testing/selftests/powerpc/stringloops/Makefile index 7fc0623d85c3..9c39f55a58ff 100644 --- a/tools/testing/selftests/powerpc/stringloops/Makefile +++ b/tools/testing/selftests/powerpc/stringloops/Makefile @@ -8,7 +8,7 @@ build_32bit = $(shell if ($(CC) $(CFLAGS) -m32 -o /dev/null memcmp.c >/dev/null TEST_GEN_PROGS := memcmp_64 strlen -$(OUTPUT)/memcmp_64: memcmp.c +$(OUTPUT)/memcmp_64: memcmp.c ../utils.c $(OUTPUT)/memcmp_64: CFLAGS += -m64 -maltivec ifeq ($(build_32bit),1) diff --git a/tools/testing/selftests/powerpc/stringloops/memcmp.c b/tools/testing/selftests/powerpc/stringloops/memcmp.c index b1fa7546957f..979df3d98368 100644 --- a/tools/testing/selftests/powerpc/stringloops/memcmp.c +++ b/tools/testing/selftests/powerpc/stringloops/memcmp.c @@ -2,7 +2,9 @@ #include <malloc.h> #include <stdlib.h> #include <string.h> +#include <sys/mman.h> #include <time.h> +#include <asm/cputable.h> #include "utils.h" #define SIZE 256 @@ -13,6 +15,9 @@ #define LARGE_MAX_OFFSET 32 #define LARGE_SIZE_START 4096 +/* This is big enough to fit LARGE_SIZE and works on 4K & 64K kernels */ +#define MAP_SIZE (64 * 1024) + #define MAX_OFFSET_DIFF_S1_S2 48 int vmx_count; @@ -68,25 +73,25 @@ static void test_one(char *s1, char *s2, unsigned long max_offset, static int testcase(bool islarge) { - char *s1; - char *s2; - unsigned long i; - - unsigned long comp_size = (islarge ? LARGE_SIZE : SIZE); - unsigned long alloc_size = comp_size + MAX_OFFSET_DIFF_S1_S2; - int iterations = islarge ? LARGE_ITERATIONS : ITERATIONS; - - s1 = memalign(128, alloc_size); - if (!s1) { - perror("memalign"); - exit(1); - } + unsigned long i, comp_size, alloc_size; + char *p, *s1, *s2; + int iterations; - s2 = memalign(128, alloc_size); - if (!s2) { - perror("memalign"); - exit(1); - } + comp_size = (islarge ? LARGE_SIZE : SIZE); + alloc_size = comp_size + MAX_OFFSET_DIFF_S1_S2; + iterations = islarge ? LARGE_ITERATIONS : ITERATIONS; + + p = mmap(NULL, 4 * MAP_SIZE, PROT_READ | PROT_WRITE, + MAP_ANONYMOUS | MAP_PRIVATE, -1, 0); + FAIL_IF(p == MAP_FAILED); + + /* Put s1/s2 at the end of a page */ + s1 = p + MAP_SIZE - alloc_size; + s2 = p + 3 * MAP_SIZE - alloc_size; + + /* And unmap the subsequent page to force a fault if we overread */ + munmap(p + MAP_SIZE, MAP_SIZE); + munmap(p + 3 * MAP_SIZE, MAP_SIZE); srandom(time(0)); @@ -147,6 +152,11 @@ static int testcase(bool islarge) static int testcases(void) { +#ifdef __powerpc64__ + // vcmpequd used in memcmp_64.S is v2.07 + SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07)); +#endif + testcase(0); testcase(1); return 0; diff --git a/tools/testing/selftests/powerpc/utils.c b/tools/testing/selftests/powerpc/utils.c index 5ee0e98c4896..18b6a773d5c7 100644 --- a/tools/testing/selftests/powerpc/utils.c +++ b/tools/testing/selftests/powerpc/utils.c @@ -16,6 +16,7 @@ #include <string.h> #include <sys/ioctl.h> #include <sys/stat.h> +#include <sys/sysinfo.h> #include <sys/types.h> #include <sys/utsname.h> #include <unistd.h> @@ -88,28 +89,40 @@ void *get_auxv_entry(int type) int pick_online_cpu(void) { - cpu_set_t mask; - int cpu; + int ncpus, cpu = -1; + cpu_set_t *mask; + size_t size; + + ncpus = get_nprocs_conf(); + size = CPU_ALLOC_SIZE(ncpus); + mask = CPU_ALLOC(ncpus); + if (!mask) { + perror("malloc"); + return -1; + } - CPU_ZERO(&mask); + CPU_ZERO_S(size, mask); - if (sched_getaffinity(0, sizeof(mask), &mask)) { + if (sched_getaffinity(0, size, mask)) { perror("sched_getaffinity"); - return -1; + goto done; } /* We prefer a primary thread, but skip 0 */ - for (cpu = 8; cpu < CPU_SETSIZE; cpu += 8) - if (CPU_ISSET(cpu, &mask)) - return cpu; + for (cpu = 8; cpu < ncpus; cpu += 8) + if (CPU_ISSET_S(cpu, size, mask)) + goto done; /* Search for anything, but in reverse */ - for (cpu = CPU_SETSIZE - 1; cpu >= 0; cpu--) - if (CPU_ISSET(cpu, &mask)) - return cpu; + for (cpu = ncpus - 1; cpu >= 0; cpu--) + if (CPU_ISSET_S(cpu, size, mask)) + goto done; printf("No cpus in affinity mask?!\n"); - return -1; + +done: + CPU_FREE(mask); + return cpu; } bool is_ppc64le(void) @@ -293,3 +306,31 @@ void set_dscr(unsigned long val) asm volatile("mtspr %1,%0" : : "r" (val), "i" (SPRN_DSCR)); } + +int using_hash_mmu(bool *using_hash) +{ + char line[128]; + FILE *f; + int rc; + + f = fopen("/proc/cpuinfo", "r"); + FAIL_IF(!f); + + rc = 0; + while (fgets(line, sizeof(line), f) != NULL) { + if (strcmp(line, "MMU : Hash\n") == 0) { + *using_hash = true; + goto out; + } + + if (strcmp(line, "MMU : Radix\n") == 0) { + *using_hash = false; + goto out; + } + } + + rc = -1; +out: + fclose(f); + return rc; +} diff --git a/tools/testing/selftests/splice/.gitignore b/tools/testing/selftests/splice/.gitignore index d5a2da428752..be8266f5d04c 100644 --- a/tools/testing/selftests/splice/.gitignore +++ b/tools/testing/selftests/splice/.gitignore @@ -1,2 +1,3 @@ # SPDX-License-Identifier: GPL-2.0-only default_file_splice_read +splice_read diff --git a/tools/testing/selftests/splice/Makefile b/tools/testing/selftests/splice/Makefile index e519b159b60d..541cd826d5a5 100644 --- a/tools/testing/selftests/splice/Makefile +++ b/tools/testing/selftests/splice/Makefile @@ -1,5 +1,5 @@ # SPDX-License-Identifier: GPL-2.0 -TEST_PROGS := default_file_splice_read.sh -TEST_GEN_PROGS_EXTENDED := default_file_splice_read +TEST_PROGS := default_file_splice_read.sh short_splice_read.sh +TEST_GEN_PROGS_EXTENDED := default_file_splice_read splice_read include ../lib.mk diff --git a/tools/testing/selftests/splice/config b/tools/testing/selftests/splice/config new file mode 100644 index 000000000000..058c928368b8 --- /dev/null +++ b/tools/testing/selftests/splice/config @@ -0,0 +1 @@ +CONFIG_TEST_LKM=m diff --git a/tools/testing/selftests/splice/settings b/tools/testing/selftests/splice/settings new file mode 100644 index 000000000000..89cedfc0d12b --- /dev/null +++ b/tools/testing/selftests/splice/settings @@ -0,0 +1 @@ +timeout=5 diff --git a/tools/testing/selftests/splice/short_splice_read.sh b/tools/testing/selftests/splice/short_splice_read.sh new file mode 100755 index 000000000000..7810d3589d9a --- /dev/null +++ b/tools/testing/selftests/splice/short_splice_read.sh @@ -0,0 +1,56 @@ +#!/bin/sh +# SPDX-License-Identifier: GPL-2.0 +set -e + +ret=0 + +do_splice() +{ + filename="$1" + bytes="$2" + expected="$3" + + out=$(./splice_read "$filename" "$bytes" | cat) + if [ "$out" = "$expected" ] ; then + echo "ok: $filename $bytes" + else + echo "FAIL: $filename $bytes" + ret=1 + fi +} + +test_splice() +{ + filename="$1" + + full=$(cat "$filename") + two=$(echo "$full" | grep -m1 . | cut -c-2) + + # Make sure full splice has the same contents as a standard read. + do_splice "$filename" 4096 "$full" + + # Make sure a partial splice see the first two characters. + do_splice "$filename" 2 "$two" +} + +# proc_single_open(), seq_read() +test_splice /proc/$$/limits +# special open, seq_read() +test_splice /proc/$$/comm + +# proc_handler, proc_dointvec_minmax +test_splice /proc/sys/fs/nr_open +# proc_handler, proc_dostring +test_splice /proc/sys/kernel/modprobe +# proc_handler, special read +test_splice /proc/sys/kernel/version + +if ! [ -d /sys/module/test_module/sections ] ; then + modprobe test_module +fi +# kernfs, attr +test_splice /sys/module/test_module/coresize +# kernfs, binattr +test_splice /sys/module/test_module/sections/.init.text + +exit $ret diff --git a/tools/testing/selftests/splice/splice_read.c b/tools/testing/selftests/splice/splice_read.c new file mode 100644 index 000000000000..46dae6a25cfb --- /dev/null +++ b/tools/testing/selftests/splice/splice_read.c @@ -0,0 +1,57 @@ +// SPDX-License-Identifier: GPL-2.0 +#define _GNU_SOURCE +#include <errno.h> +#include <fcntl.h> +#include <limits.h> +#include <stdio.h> +#include <stdlib.h> +#include <unistd.h> +#include <sys/types.h> +#include <sys/stat.h> + +int main(int argc, char *argv[]) +{ + int fd; + size_t size; + ssize_t spliced; + + if (argc < 2) { + fprintf(stderr, "Usage: %s INPUT [BYTES]\n", argv[0]); + return EXIT_FAILURE; + } + + fd = open(argv[1], O_RDONLY); + if (fd < 0) { + perror(argv[1]); + return EXIT_FAILURE; + } + + if (argc == 3) + size = atol(argv[2]); + else { + struct stat statbuf; + + if (fstat(fd, &statbuf) < 0) { + perror(argv[1]); + return EXIT_FAILURE; + } + + if (statbuf.st_size > INT_MAX) { + fprintf(stderr, "%s: Too big\n", argv[1]); + return EXIT_FAILURE; + } + + size = statbuf.st_size; + } + + /* splice(2) file to stdout. */ + spliced = splice(fd, NULL, STDOUT_FILENO, NULL, + size, SPLICE_F_MOVE); + if (spliced < 0) { + perror("splice"); + return EXIT_FAILURE; + } + + close(fd); + return EXIT_SUCCESS; +} diff --git a/tools/testing/selftests/vm/hmm-tests.c b/tools/testing/selftests/vm/hmm-tests.c index 91d38a29956b..93fc5cadce61 100644 --- a/tools/testing/selftests/vm/hmm-tests.c +++ b/tools/testing/selftests/vm/hmm-tests.c @@ -942,6 +942,41 @@ TEST_F(hmm, migrate_fault) } /* + * Migrate anonymous shared memory to device private memory. + */ +TEST_F(hmm, migrate_shared) +{ + struct hmm_buffer *buffer; + unsigned long npages; + unsigned long size; + int ret; + + npages = ALIGN(HMM_BUFFER_SIZE, self->page_size) >> self->page_shift; + ASSERT_NE(npages, 0); + size = npages << self->page_shift; + + buffer = malloc(sizeof(*buffer)); + ASSERT_NE(buffer, NULL); + + buffer->fd = -1; + buffer->size = size; + buffer->mirror = malloc(size); + ASSERT_NE(buffer->mirror, NULL); + + buffer->ptr = mmap(NULL, size, + PROT_READ | PROT_WRITE, + MAP_SHARED | MAP_ANONYMOUS, + buffer->fd, 0); + ASSERT_NE(buffer->ptr, MAP_FAILED); + + /* Migrate memory to device. */ + ret = hmm_dmirror_cmd(self->fd, HMM_DMIRROR_MIGRATE, buffer, npages); + ASSERT_EQ(ret, -ENOENT); + + hmm_buffer_free(buffer); +} + +/* * Try to migrate various memory types to device private memory. */ TEST_F(hmm2, migrate_mixed) |