February 2025 - Linux-kselftest-mirror

[PATCH v2] ww_mutex: convert self-test to KUnit

by Tamir Duberstein

Convert this unit test to a KUnit test. This allows the test to benefit from the KUnit tooling. Note that care is taken to avoid test-ending assertions in worker threads, which is unsafe in KUnit (and wasn't done before this change either). Signed-off-by: Tamir Duberstein <tamird(a)gmail.com> --- I tested this using: $ tools/testing/kunit/kunit.py run --arch arm64 --make_options LLVM=1 ww_mutex ; [12:48:16] ================== ww_mutex (5 subtests) =================== ; [12:48:16] ======================= test_mutex ======================== ; [12:48:16] [PASSED] flags=0 ; [12:48:16] [PASSED] flags=1 ; [12:48:16] [PASSED] flags=2 ; [12:48:16] [PASSED] flags=3 ; [12:48:16] [PASSED] flags=4 ; [12:48:17] [PASSED] flags=5 ; [12:48:17] [PASSED] flags=6 ; [12:48:17] [PASSED] flags=7 ; [12:48:17] =================== [PASSED] test_mutex ==================== ; [12:48:17] ========================= test_aa ========================= ; [12:48:17] [PASSED] lock ; [12:48:17] [PASSED] trylock ; [12:48:17] ===================== [PASSED] test_aa ===================== ; [12:48:17] ======================== test_abba ======================== ; [12:48:17] [PASSED] trylock=0,resolve=0 ; [12:48:17] [PASSED] trylock=1,resolve=1 ; [12:48:17] [PASSED] trylock=0,resolve=0 ; [12:48:17] [PASSED] trylock=1,resolve=1 ; [12:48:17] ==================== [PASSED] test_abba ==================== ; [12:48:17] ======================= test_cycle ======================== ; [12:48:17] [PASSED] nthreads=2 ; [12:48:17] =================== [PASSED] test_cycle ==================== ; [12:48:21] ========================= stress ========================== ; [12:48:21] [PASSED] nlocks=16,nthreads_per_cpu=2,flags=1 ; [12:48:23] [PASSED] nlocks=16,nthreads_per_cpu=2,flags=2 ; [12:48:23] [PASSED] nlocks=2046,nthreads_per_cpu=3,flags=7 ; [12:48:23] ===================== [PASSED] stress ====================== ; [12:48:23] ==================== [PASSED] ww_mutex ===================== ; [12:48:23] ============================================================ ; [12:48:23] Testing complete. Ran 18 tests: passed: 18 --- Changes in v2: - Avoid KUNIT_ASSERT_* in non-main thread. (David Gow) - Include rationale and details in the commit message. (Boqun Feng) - Introduce test_abba_param to avoid odd-looking bit shifting. (Dan Carpenter) - Rebase on linux-next. - Move the test to tests/ per KUnit style guide. - Link to v1: https://lore.kernel.org/r/20250210-ww_mutex-kunit-convert-v1-1-972f0201f71e… --- kernel/locking/Makefile | 3 +- kernel/locking/tests/Makefile | 3 + .../{test-ww_mutex.c => tests/ww_mutex_kunit.c} | 319 +++++++++++---------- lib/Kconfig.debug | 12 +- tools/testing/selftests/locking/ww_mutex.sh | 19 -- 5 files changed, 176 insertions(+), 180 deletions(-) diff --git a/kernel/locking/Makefile b/kernel/locking/Makefile index 0db4093d17b8..b5fa68d18823 100644 --- a/kernel/locking/Makefile +++ b/kernel/locking/Makefile @@ -30,5 +30,6 @@ obj-$(CONFIG_DEBUG_SPINLOCK) += spinlock.o obj-$(CONFIG_DEBUG_SPINLOCK) += spinlock_debug.o obj-$(CONFIG_QUEUED_RWLOCKS) += qrwlock.o obj-$(CONFIG_LOCK_TORTURE_TEST) += locktorture.o -obj-$(CONFIG_WW_MUTEX_SELFTEST) += test-ww_mutex.o obj-$(CONFIG_LOCK_EVENT_COUNTS) += lock_events.o + +obj-y += tests/ diff --git a/kernel/locking/tests/Makefile b/kernel/locking/tests/Makefile new file mode 100644 index 000000000000..8ddde822c793 --- /dev/null +++ b/kernel/locking/tests/Makefile @@ -0,0 +1,3 @@ +# SPDX-License-Identifier: GPL-2.0-only + +obj-$(CONFIG_WW_MUTEX_KUNIT_TEST) += ww_mutex_kunit.o diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/tests/ww_mutex_kunit.c similarity index 66% rename from kernel/locking/test-ww_mutex.c rename to kernel/locking/tests/ww_mutex_kunit.c index bcb1b9fea588..a8d412231735 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/tests/ww_mutex_kunit.c @@ -3,10 +3,10 @@ * Module-based API test facility for ww_mutexes */ -#include <linux/kernel.h> - +#include <kunit/test.h> #include <linux/completion.h> #include <linux/delay.h> +#include <linux/kernel.h> #include <linux/kthread.h> #include <linux/module.h> #include <linux/prandom.h> @@ -54,12 +54,39 @@ static void test_mutex_work(struct work_struct *work) ww_mutex_unlock(&mtx->mutex); } -static int __test_mutex(unsigned int flags) +static const unsigned int *gen_range( + unsigned int *storage, + const unsigned int min, + const unsigned int max, + const int *prev) +{ + if (prev != NULL) { + if (*prev >= max) + return NULL; + *storage = *prev + 1; + } else { + *storage = min; + } + return storage; +} + +static const void *test_mutex_gen_params(const void *prev, char *desc) +{ + static unsigned int storage; + const unsigned int *next = gen_range(&storage, 0, __TEST_MTX_LAST - 1, prev); + + if (next != NULL) + snprintf(desc, KUNIT_PARAM_DESC_SIZE, "flags=%x", *next); + return next; +} + +static void test_mutex(struct kunit *test) { #define TIMEOUT (HZ / 16) + const unsigned int *param = test->param_value; + const unsigned int flags = *param; struct test_mutex mtx; struct ww_acquire_ctx ctx; - int ret; ww_mutex_init(&mtx.mutex, &ww_class); if (flags & TEST_MTX_CTX) @@ -79,53 +106,42 @@ static int __test_mutex(unsigned int flags) if (flags & TEST_MTX_SPIN) { unsigned long timeout = jiffies + TIMEOUT; - ret = 0; do { if (completion_done(&mtx.done)) { - ret = -EINVAL; + KUNIT_FAIL(test, "mutual exclusion failure"); break; } cond_resched(); } while (time_before(jiffies, timeout)); } else { - ret = wait_for_completion_timeout(&mtx.done, TIMEOUT); + KUNIT_EXPECT_EQ(test, wait_for_completion_timeout(&mtx.done, TIMEOUT), 0); } ww_mutex_unlock(&mtx.mutex); if (flags & TEST_MTX_CTX) ww_acquire_fini(&ctx); - if (ret) { - pr_err("%s(flags=%x): mutual exclusion failure\n", - __func__, flags); - ret = -EINVAL; - } - flush_work(&mtx.work); destroy_work_on_stack(&mtx.work); - return ret; #undef TIMEOUT } -static int test_mutex(void) +static const void *test_aa_gen_params(const void *prev, char *desc) { - int ret; - int i; - - for (i = 0; i < __TEST_MTX_LAST; i++) { - ret = __test_mutex(i); - if (ret) - return ret; - } + static unsigned int storage; + const unsigned int *next = gen_range(&storage, 0, 1, prev); - return 0; + if (next != NULL) + snprintf(desc, KUNIT_PARAM_DESC_SIZE, *next ? "trylock" : "lock"); + return next; } -static int test_aa(bool trylock) +static void test_aa(struct kunit *test) { + const unsigned int *param = test->param_value; + const bool trylock = *param; struct ww_mutex mutex; struct ww_acquire_ctx ctx; int ret; - const char *from = trylock ? "trylock" : "lock"; ww_mutex_init(&mutex, &ww_class); ww_acquire_init(&ctx, &ww_class); @@ -133,46 +149,42 @@ static int test_aa(bool trylock) if (!trylock) { ret = ww_mutex_lock(&mutex, &ctx); if (ret) { - pr_err("%s: initial lock failed!\n", __func__); + KUNIT_FAIL(test, "initial lock failed, ret=%d", ret); goto out; } } else { ret = !ww_mutex_trylock(&mutex, &ctx); if (ret) { - pr_err("%s: initial trylock failed!\n", __func__); + KUNIT_FAIL(test, "initial trylock failed, ret=%d", ret); goto out; } } - if (ww_mutex_trylock(&mutex, NULL)) { - pr_err("%s: trylocked itself without context from %s!\n", __func__, from); + ret = ww_mutex_trylock(&mutex, NULL); + if (ret) { + KUNIT_FAIL(test, "trylocked itself without context, ret=%d", ret); ww_mutex_unlock(&mutex); - ret = -EINVAL; goto out; } - if (ww_mutex_trylock(&mutex, &ctx)) { - pr_err("%s: trylocked itself with context from %s!\n", __func__, from); + ret = ww_mutex_trylock(&mutex, &ctx); + if (ret) { + KUNIT_FAIL(test, "trylocked itself with context, ret=%d", ret); ww_mutex_unlock(&mutex); - ret = -EINVAL; goto out; } ret = ww_mutex_lock(&mutex, &ctx); if (ret != -EALREADY) { - pr_err("%s: missed deadlock for recursing, ret=%d from %s\n", - __func__, ret, from); + KUNIT_FAIL(test, "missed deadlock for recursing, ret=%d", ret); if (!ret) ww_mutex_unlock(&mutex); - ret = -EINVAL; goto out; } ww_mutex_unlock(&mutex); - ret = 0; out: ww_acquire_fini(&ctx); - return ret; } struct test_abba { @@ -217,11 +229,36 @@ static void test_abba_work(struct work_struct *work) abba->result = err; } -static int test_abba(bool trylock, bool resolve) +union test_abba_param { + unsigned int value; + struct { + unsigned int trylock : 1; + unsigned int resolve : 1; + }; +}; + +static const void *test_abba_gen_params(const void *prev, char *desc) { + static unsigned int storage; + const unsigned int *next = gen_range(&storage, 0b00, 0b11, prev); + + if (next != NULL) { + const union test_abba_param param = { .value = *next }; + + snprintf(desc, KUNIT_PARAM_DESC_SIZE, "trylock=%d,resolve=%d", + param.trylock, param.resolve); + } + return next; +} + +static void test_abba(struct kunit *test) +{ + const union test_abba_param *param = test->param_value; + const bool trylock = param->trylock; + const bool resolve = param->resolve; struct test_abba abba; struct ww_acquire_ctx ctx; - int err, ret; + int err; ww_mutex_init(&abba.a_mutex, &ww_class); ww_mutex_init(&abba.b_mutex, &ww_class); @@ -259,21 +296,17 @@ static int test_abba(bool trylock, bool resolve) flush_work(&abba.work); destroy_work_on_stack(&abba.work); - ret = 0; if (resolve) { if (err || abba.result) { - pr_err("%s: failed to resolve ABBA deadlock, A err=%d, B err=%d\n", - __func__, err, abba.result); - ret = -EINVAL; + KUNIT_FAIL(test, "failed to resolve ABBA deadlock, A err=%d, B err=%d", + err, abba.result); } } else { if (err != -EDEADLK && abba.result != -EDEADLK) { - pr_err("%s: missed ABBA deadlock, A err=%d, B err=%d\n", - __func__, err, abba.result); - ret = -EINVAL; + KUNIT_FAIL(test, "missed ABBA deadlock, A err=%d, B err=%d", + err, abba.result); } } - return ret; } struct test_cycle { @@ -314,15 +347,25 @@ static void test_cycle_work(struct work_struct *work) cycle->result = err ?: erra; } -static int __test_cycle(unsigned int nthreads) +static const void *test_cycle_gen_params(const void *prev, char *desc) { + static unsigned int storage; + const unsigned int *next = gen_range(&storage, 2, num_online_cpus(), prev); + + if (next != NULL) + snprintf(desc, KUNIT_PARAM_DESC_SIZE, "nthreads=%d", *next); + return next; +} + +static void test_cycle(struct kunit *test) +{ + const unsigned int *param = test->param_value; + const unsigned int nthreads = *param; struct test_cycle *cycles; unsigned int n, last = nthreads - 1; - int ret; - cycles = kmalloc_array(nthreads, sizeof(*cycles), GFP_KERNEL); - if (!cycles) - return -ENOMEM; + cycles = kunit_kmalloc_array(test, nthreads, sizeof(*cycles), GFP_KERNEL); + KUNIT_ASSERT_NOT_NULL(test, cycles); for (n = 0; n < nthreads; n++) { struct test_cycle *cycle = &cycles[n]; @@ -348,41 +391,24 @@ static int __test_cycle(unsigned int nthreads) flush_workqueue(wq); - ret = 0; for (n = 0; n < nthreads; n++) { struct test_cycle *cycle = &cycles[n]; if (!cycle->result) continue; - pr_err("cyclic deadlock not resolved, ret[%d/%d] = %d\n", - n, nthreads, cycle->result); - ret = -EINVAL; + KUNIT_FAIL(test, "cyclic deadlock not resolved, ret[%d/%d] = %d", + n, nthreads, cycle->result); break; } for (n = 0; n < nthreads; n++) ww_mutex_destroy(&cycles[n].a_mutex); - kfree(cycles); - return ret; -} - -static int test_cycle(unsigned int ncpus) -{ - unsigned int n; - int ret; - - for (n = 2; n <= ncpus + 1; n++) { - ret = __test_cycle(n); - if (ret) - return ret; - } - - return 0; } struct stress { struct work_struct work; + struct kunit *test; struct ww_mutex *locks; unsigned long timeout; int nlocks; @@ -401,12 +427,12 @@ static inline u32 prandom_u32_below(u32 ceil) return ret; } -static int *get_random_order(int count) +static int *get_random_order(struct kunit *test, int count) { int *order; int n, r; - order = kmalloc_array(count, sizeof(*order), GFP_KERNEL); + order = kunit_kmalloc_array(test, count, sizeof(*order), GFP_KERNEL); if (!order) return order; @@ -435,7 +461,8 @@ static void stress_inorder_work(struct work_struct *work) struct ww_acquire_ctx ctx; int *order; - order = get_random_order(nlocks); + order = get_random_order(stress->test, nlocks); + KUNIT_EXPECT_NOT_NULL(stress->test, order); if (!order) return; @@ -472,13 +499,10 @@ static void stress_inorder_work(struct work_struct *work) ww_acquire_fini(&ctx); if (err) { - pr_err_once("stress (%s) failed with %d\n", - __func__, err); + KUNIT_FAIL(stress->test, "lock[%d] failed, err=%d", n, err); break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(order); } struct reorder_lock { @@ -495,19 +519,19 @@ static void stress_reorder_work(struct work_struct *work) int *order; int n, err; - order = get_random_order(stress->nlocks); + order = get_random_order(stress->test, stress->nlocks); + KUNIT_EXPECT_NOT_NULL(stress->test, order); if (!order) return; for (n = 0; n < stress->nlocks; n++) { - ll = kmalloc(sizeof(*ll), GFP_KERNEL); + ll = kunit_kmalloc(stress->test, sizeof(*ll), GFP_KERNEL); + KUNIT_EXPECT_NOT_NULL(stress->test, ll); if (!ll) - goto out; - + return; ll->lock = &stress->locks[order[n]]; list_add(&ll->link, &locks); } - kfree(order); order = NULL; do { @@ -523,8 +547,7 @@ static void stress_reorder_work(struct work_struct *work) ww_mutex_unlock(ln->lock); if (err != -EDEADLK) { - pr_err_once("stress (%s) failed with %d\n", - __func__, err); + KUNIT_FAIL(stress->test, "lock failed, err=%d", err); break; } @@ -538,11 +561,6 @@ static void stress_reorder_work(struct work_struct *work) ww_acquire_fini(&ctx); } while (!time_after(jiffies, stress->timeout)); - -out: - list_for_each_entry_safe(ll, ln, &locks, link) - kfree(ll); - kfree(order); } static void stress_one_work(struct work_struct *work) @@ -558,8 +576,7 @@ static void stress_one_work(struct work_struct *work) dummy_load(stress); ww_mutex_unlock(lock); } else { - pr_err_once("stress (%s) failed with %d\n", - __func__, err); + KUNIT_FAIL(stress->test, "lock failed, err=%d", err); break; } } while (!time_after(jiffies, stress->timeout)); @@ -570,22 +587,41 @@ static void stress_one_work(struct work_struct *work) #define STRESS_ONE BIT(2) #define STRESS_ALL (STRESS_INORDER | STRESS_REORDER | STRESS_ONE) -static int stress(int nlocks, int nthreads, unsigned int flags) +struct stress_case { + int nlocks; + int nthreads_per_cpu; + unsigned int flags; +}; + +static const struct stress_case stress_cases[] = { + { 16, 2, STRESS_INORDER }, + { 16, 2, STRESS_REORDER }, + { 2046, hweight32(STRESS_ALL), STRESS_ALL }, +}; + +static void stress_case_to_desc(const struct stress_case *param, char *desc) { + snprintf(desc, KUNIT_PARAM_DESC_SIZE, "nlocks=%d,nthreads_per_cpu=%d,flags=%x", + param->nlocks, param->nthreads_per_cpu, param->flags); +} + +KUNIT_ARRAY_PARAM(stress_cases, stress_cases, stress_case_to_desc); + +static void stress(struct kunit *test) +{ + const struct stress_case *param = test->param_value; + const int nlocks = param->nlocks; + int nthreads = param->nthreads_per_cpu * num_online_cpus(); + const unsigned int flags = param->flags; struct ww_mutex *locks; struct stress *stress_array; int n, count; - locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); - if (!locks) - return -ENOMEM; + locks = kunit_kmalloc_array(test, nlocks, sizeof(*locks), GFP_KERNEL); + KUNIT_ASSERT_NOT_NULL(test, locks); - stress_array = kmalloc_array(nthreads, sizeof(*stress_array), - GFP_KERNEL); - if (!stress_array) { - kfree(locks); - return -ENOMEM; - } + stress_array = kunit_kmalloc_array(test, nthreads, sizeof(*stress_array), GFP_KERNEL); + KUNIT_ASSERT_NOT_NULL(test, stress_array); for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); @@ -617,6 +653,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); + stress->test = test; stress->locks = locks; stress->nlocks = nlocks; stress->timeout = jiffies + 2*HZ; @@ -629,70 +666,42 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); - kfree(stress_array); - kfree(locks); - - return 0; } -static int __init test_ww_mutex_init(void) +static int ww_mutex_suite_init(struct kunit_suite *suite) { - int ncpus = num_online_cpus(); - int ret, i; - - printk(KERN_INFO "Beginning ww mutex selftests\n"); - - prandom_seed_state(&rng, get_random_u64()); - wq = alloc_workqueue("test-ww_mutex", WQ_UNBOUND, 0); if (!wq) return -ENOMEM; - ret = test_mutex(); - if (ret) - return ret; - - ret = test_aa(false); - if (ret) - return ret; - - ret = test_aa(true); - if (ret) - return ret; - - for (i = 0; i < 4; i++) { - ret = test_abba(i & 1, i & 2); - if (ret) - return ret; - } - - ret = test_cycle(ncpus); - if (ret) - return ret; - - ret = stress(16, 2*ncpus, STRESS_INORDER); - if (ret) - return ret; - - ret = stress(16, 2*ncpus, STRESS_REORDER); - if (ret) - return ret; - - ret = stress(2046, hweight32(STRESS_ALL)*ncpus, STRESS_ALL); - if (ret) - return ret; + prandom_seed_state(&rng, get_random_u64()); - printk(KERN_INFO "All ww mutex selftests passed\n"); return 0; } -static void __exit test_ww_mutex_exit(void) +static void ww_mutex_suite_exit(struct kunit_suite *suite) { - destroy_workqueue(wq); + if (wq) + destroy_workqueue(wq); } -module_init(test_ww_mutex_init); -module_exit(test_ww_mutex_exit); +static struct kunit_case ww_mutex_cases[] = { + KUNIT_CASE_PARAM(test_mutex, test_mutex_gen_params), + KUNIT_CASE_PARAM(test_aa, test_aa_gen_params), + KUNIT_CASE_PARAM(test_abba, test_abba_gen_params), + KUNIT_CASE_PARAM(test_cycle, test_cycle_gen_params), + KUNIT_CASE_PARAM_ATTR(stress, stress_cases_gen_params, {.speed = KUNIT_SPEED_SLOW}), + {}, +}; + +static struct kunit_suite ww_mutex_suite = { + .name = "ww_mutex", + .suite_init = ww_mutex_suite_init, + .suite_exit = ww_mutex_suite_exit, + .test_cases = ww_mutex_cases, +}; + +kunit_test_suite(ww_mutex_suite); MODULE_LICENSE("GPL"); MODULE_AUTHOR("Intel Corporation"); diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 85b95d645b10..2c9da8647eaf 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -1595,16 +1595,18 @@ config LOCK_TORTURE_TEST Say M if you want these torture tests to build as a module. Say N if you are unsure. -config WW_MUTEX_SELFTEST - tristate "Wait/wound mutex selftests" +config WW_MUTEX_KUNIT_TEST + tristate "KUnit test for wait/wound mutex" if !KUNIT_ALL_TESTS + depends on KUNIT + default KUNIT_ALL_TESTS help - This option provides a kernel module that runs tests on the - on the struct ww_mutex locking API. + This option provides a KUnit test that exercises the struct + ww_mutex locking API. It is recommended to enable DEBUG_WW_MUTEX_SLOWPATH in conjunction with this test harness. - Say M if you want these self tests to build as a module. + Say M if you want these tests to build as a module. Say N if you are unsure. config SCF_TORTURE_TEST diff --git a/tools/testing/selftests/locking/ww_mutex.sh b/tools/testing/selftests/locking/ww_mutex.sh deleted file mode 100755 index 91e4ac7566af..000000000000 --- a/tools/testing/selftests/locking/ww_mutex.sh +++ /dev/null @@ -1,19 +0,0 @@ -#!/bin/sh -# SPDX-License-Identifier: GPL-2.0 - -# Kselftest framework requirement - SKIP code is 4. -ksft_skip=4 - -# Runs API tests for struct ww_mutex (Wait/Wound mutexes) -if ! /sbin/modprobe -q -n test-ww_mutex; then - echo "ww_mutex: module test-ww_mutex is not found [SKIP]" - exit $ksft_skip -fi - -if /sbin/modprobe -q test-ww_mutex; then - /sbin/modprobe -q -r test-ww_mutex - echo "locking/ww_mutex: ok" -else - echo "locking/ww_mutex: [FAIL]" - exit 1 -fi --- base-commit: 7b7a883c7f4de1ee5040bd1c32aabaafde54d209 change-id: 20250208-ww_mutex-kunit-convert-71842a7a6be2 Best regards, -- Tamir Duberstein <tamird(a)gmail.com>

1 week, 6 days

2
5
0 0

[PATCH bpf-next v2] arm64, bpf: Add 12-argument support for bpf trampoline

by Puranjay Mohan

The arm64 bpf JIT currently supports attaching the trampoline to functions with <= 8 arguments. This is because up to 8 arguments can be passed in registers r0-r7. If there are more than 8 arguments then the 9th and later arguments are passed on the stack, with SP pointing to the first stacked argument. See aapcs64[1] for more details. If the 8th argument is a structure of size > 8B, then it is passed fully on stack and r7 is not used for passing any argument. If there is a 9th argument, it will be passed on the stack, even though r7 is available. Add the support of storing and restoring arguments passed on the stack to the arm64 bpf trampoline. This will allow attaching the trampoline to functions that take up to 12 arguments. [1] https://github.com/ARM-software/abi-aa/blob/main/aapcs64/aapcs64.rst#parame… Signed-off-by: Puranjay Mohan <puranjay(a)kernel.org> --- Changes in V1 -> V2: V1: https://lore.kernel.org/all/20240704173227.130491-1-puranjay@kernel.org/ - Fixed the argument handling for composite types (structs) --- arch/arm64/net/bpf_jit_comp.c | 139 ++++++++++++++----- tools/testing/selftests/bpf/DENYLIST.aarch64 | 3 - 2 files changed, 107 insertions(+), 35 deletions(-) diff --git a/arch/arm64/net/bpf_jit_comp.c b/arch/arm64/net/bpf_jit_comp.c index 751331f5ba90..063bf5e11fc6 100644 --- a/arch/arm64/net/bpf_jit_comp.c +++ b/arch/arm64/net/bpf_jit_comp.c @@ -30,6 +30,8 @@ #define TMP_REG_3 (MAX_BPF_JIT_REG + 3) #define FP_BOTTOM (MAX_BPF_JIT_REG + 4) #define ARENA_VM_START (MAX_BPF_JIT_REG + 5) +/* Up to eight function arguments are passed in registers r0-r7 */ +#define ARM64_MAX_REG_ARGS 8 #define check_imm(bits, imm) do { \ if ((((imm) > 0) && ((imm) >> (bits))) || \ @@ -2001,26 +2003,51 @@ static void invoke_bpf_mod_ret(struct jit_ctx *ctx, struct bpf_tramp_links *tl, } } -static void save_args(struct jit_ctx *ctx, int args_off, int nregs) +static void save_args(struct jit_ctx *ctx, int args_off, int orig_sp_off, + int nargs, int nreg_args) { + const u8 tmp = bpf2a64[TMP_REG_1]; + int arg_pos; int i; - for (i = 0; i < nregs; i++) { - emit(A64_STR64I(i, A64_SP, args_off), ctx); + for (i = 0; i < nargs; i++) { + if (i < nreg_args) { + emit(A64_STR64I(i, A64_SP, args_off), ctx); + } else { + arg_pos = orig_sp_off + (i - nreg_args) * 8; + emit(A64_LDR64I(tmp, A64_SP, arg_pos), ctx); + emit(A64_STR64I(tmp, A64_SP, args_off), ctx); + } args_off += 8; } } -static void restore_args(struct jit_ctx *ctx, int args_off, int nregs) +static void restore_args(struct jit_ctx *ctx, int args_off, int nreg_args) { int i; - for (i = 0; i < nregs; i++) { + for (i = 0; i < nreg_args; i++) { emit(A64_LDR64I(i, A64_SP, args_off), ctx); args_off += 8; } } +static void restore_stack_args(struct jit_ctx *ctx, int args_off, int stk_arg_off, + int nargs, int nreg_args) +{ + const u8 tmp = bpf2a64[TMP_REG_1]; + int arg_pos; + int i; + + for (i = nreg_args; i < nargs; i++) { + arg_pos = args_off + i * 8; + emit(A64_LDR64I(tmp, A64_SP, arg_pos), ctx); + emit(A64_STR64I(tmp, A64_SP, stk_arg_off), ctx); + + stk_arg_off += 8; + } +} + /* Based on the x86's implementation of arch_prepare_bpf_trampoline(). * * bpf prog and function entry before bpf trampoline hooked: @@ -2034,15 +2061,17 @@ static void restore_args(struct jit_ctx *ctx, int args_off, int nregs) */ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, struct bpf_tramp_links *tlinks, void *func_addr, - int nregs, u32 flags) + int nargs, int nreg_args, u32 flags) { int i; int stack_size; + int stk_arg_off; + int orig_sp_off; int retaddr_off; int regs_off; int retval_off; int args_off; - int nregs_off; + int nargs_off; int ip_off; int run_ctx_off; struct bpf_tramp_links *fentry = &tlinks[BPF_TRAMP_FENTRY]; @@ -2052,6 +2081,7 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, __le32 **branches = NULL; /* trampoline stack layout: + * SP + orig_sp_off [ first stack arg ] if nargs > 8 * [ parent ip ] * [ FP ] * SP + retaddr_off [ self ip ] @@ -2069,14 +2099,24 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, * [ ... ] * SP + args_off [ arg reg 1 ] * - * SP + nregs_off [ arg regs count ] + * SP + nargs_off [ arg count ] * * SP + ip_off [ traced function ] BPF_TRAMP_F_IP_ARG flag * * SP + run_ctx_off [ bpf_tramp_run_ctx ] + * + * [ stack_argN ] + * [ ... ] + * SP + stk_arg_off [ stack_arg1 ] BPF_TRAMP_F_CALL_ORIG */ stack_size = 0; + stk_arg_off = stack_size; + if ((flags & BPF_TRAMP_F_CALL_ORIG) && (nargs - nreg_args > 0)) { + /* room for saving arguments passed on stack */ + stack_size += (nargs - nreg_args) * 8; + } + run_ctx_off = stack_size; /* room for bpf_tramp_run_ctx */ stack_size += round_up(sizeof(struct bpf_tramp_run_ctx), 8); @@ -2086,13 +2126,13 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, if (flags & BPF_TRAMP_F_IP_ARG) stack_size += 8; - nregs_off = stack_size; + nargs_off = stack_size; /* room for args count */ stack_size += 8; args_off = stack_size; /* room for args */ - stack_size += nregs * 8; + stack_size += nargs * 8; /* room for return value */ retval_off = stack_size; @@ -2110,6 +2150,11 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, /* return address locates above FP */ retaddr_off = stack_size + 8; + /* original SP position + * stack_size + parent function frame + patched function frame + */ + orig_sp_off = stack_size + 32; + /* bpf trampoline may be invoked by 3 instruction types: * 1. bl, attached to bpf prog or kernel function via short jump * 2. br, attached to bpf prog or kernel function via long jump @@ -2135,12 +2180,12 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, emit(A64_STR64I(A64_R(10), A64_SP, ip_off), ctx); } - /* save arg regs count*/ - emit(A64_MOVZ(1, A64_R(10), nregs, 0), ctx); - emit(A64_STR64I(A64_R(10), A64_SP, nregs_off), ctx); + /* save argument count */ + emit(A64_MOVZ(1, A64_R(10), nargs, 0), ctx); + emit(A64_STR64I(A64_R(10), A64_SP, nargs_off), ctx); - /* save arg regs */ - save_args(ctx, args_off, nregs); + /* save arguments passed in regs and on the stack */ + save_args(ctx, args_off, orig_sp_off, nargs, nreg_args); /* save callee saved registers */ emit(A64_STR64I(A64_R(19), A64_SP, regs_off), ctx); @@ -2167,7 +2212,10 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, } if (flags & BPF_TRAMP_F_CALL_ORIG) { - restore_args(ctx, args_off, nregs); + /* restore arguments that were passed in registers */ + restore_args(ctx, args_off, nreg_args); + /* restore arguments that were passed on the stack */ + restore_stack_args(ctx, args_off, stk_arg_off, nargs, nreg_args); /* call original func */ emit(A64_LDR64I(A64_R(10), A64_SP, retaddr_off), ctx); emit(A64_ADR(A64_LR, AARCH64_INSN_SIZE * 2), ctx); @@ -2196,7 +2244,7 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, } if (flags & BPF_TRAMP_F_RESTORE_REGS) - restore_args(ctx, args_off, nregs); + restore_args(ctx, args_off, nreg_args); /* restore callee saved register x19 and x20 */ emit(A64_LDR64I(A64_R(19), A64_SP, regs_off), ctx); @@ -2228,19 +2276,42 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, return ctx->idx; } -static int btf_func_model_nregs(const struct btf_func_model *m) +static int btf_func_model_nargs(const struct btf_func_model *m) { - int nregs = m->nr_args; + int nargs = m->nr_args; int i; - /* extra registers needed for struct argument */ + /* extra registers or stack slots needed for struct argument */ for (i = 0; i < MAX_BPF_FUNC_ARGS; i++) { /* The arg_size is at most 16 bytes, enforced by the verifier. */ if (m->arg_flags[i] & BTF_FMODEL_STRUCT_ARG) - nregs += (m->arg_size[i] + 7) / 8 - 1; + nargs += (m->arg_size[i] + 7) / 8 - 1; } - return nregs; + return nargs; +} + +/* get the count of the regs that are used to pass arguments */ +static int btf_func_model_nreg_args(const struct btf_func_model *m) +{ + int nargs = m->nr_args; + int nreg_args = 0; + int i; + + for (i = 0; i < nargs; i++) { + /* The arg_size is at most 16 bytes, enforced by the verifier. */ + if (m->arg_flags[i] & BTF_FMODEL_STRUCT_ARG) { + /* struct members are all in the registers or all + * on the stack. + */ + if (nreg_args + ((m->arg_size[i] + 7) / 8 - 1) > 7) + break; + nreg_args += (m->arg_size[i] + 7) / 8 - 1; + } + nreg_args++; + } + + return (nreg_args > ARM64_MAX_REG_ARGS ? ARM64_MAX_REG_ARGS : nreg_args); } int arch_bpf_trampoline_size(const struct btf_func_model *m, u32 flags, @@ -2251,14 +2322,16 @@ int arch_bpf_trampoline_size(const struct btf_func_model *m, u32 flags, .idx = 0, }; struct bpf_tramp_image im; - int nregs, ret; + int nargs, nreg_args, ret; - nregs = btf_func_model_nregs(m); - /* the first 8 registers are used for arguments */ - if (nregs > 8) + nargs = btf_func_model_nargs(m); + if (nargs > MAX_BPF_FUNC_ARGS) return -ENOTSUPP; - ret = prepare_trampoline(&ctx, &im, tlinks, func_addr, nregs, flags); + nreg_args = btf_func_model_nreg_args(m); + + ret = prepare_trampoline(&ctx, &im, tlinks, func_addr, nargs, nreg_args, + flags); if (ret < 0) return ret; @@ -2285,7 +2358,7 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *ro_image, u32 flags, struct bpf_tramp_links *tlinks, void *func_addr) { - int ret, nregs; + int ret, nargs, nreg_args; void *image, *tmp; u32 size = ro_image_end - ro_image; @@ -2302,13 +2375,15 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *ro_image, .idx = 0, }; - nregs = btf_func_model_nregs(m); - /* the first 8 registers are used for arguments */ - if (nregs > 8) + nargs = btf_func_model_nargs(m); + if (nargs > MAX_BPF_FUNC_ARGS) return -ENOTSUPP; + nreg_args = btf_func_model_nreg_args(m); + jit_fill_hole(image, (unsigned int)(ro_image_end - ro_image)); - ret = prepare_trampoline(&ctx, im, tlinks, func_addr, nregs, flags); + ret = prepare_trampoline(&ctx, im, tlinks, func_addr, nargs, nreg_args, + flags); if (ret > 0 && validate_code(&ctx) < 0) { ret = -EINVAL; diff --git a/tools/testing/selftests/bpf/DENYLIST.aarch64 b/tools/testing/selftests/bpf/DENYLIST.aarch64 index 3c7c3e79aa93..e865451e90d2 100644 --- a/tools/testing/selftests/bpf/DENYLIST.aarch64 +++ b/tools/testing/selftests/bpf/DENYLIST.aarch64 @@ -4,9 +4,6 @@ fexit_sleep # The test never returns. The r kprobe_multi_bench_attach # needs CONFIG_FPROBE kprobe_multi_test # needs CONFIG_FPROBE module_attach # prog 'kprobe_multi': failed to auto-attach: -95 -fentry_test/fentry_many_args # fentry_many_args:FAIL:fentry_many_args_attach unexpected error: -524 -fexit_test/fexit_many_args # fexit_many_args:FAIL:fexit_many_args_attach unexpected error: -524 -tracing_struct/struct_many_args # struct_many_args:FAIL:tracing_struct_many_args__attach unexpected error: -524 fill_link_info/kprobe_multi_link_info # bpf_program__attach_kprobe_multi_opts unexpected error: -95 fill_link_info/kretprobe_multi_link_info # bpf_program__attach_kprobe_multi_opts unexpected error: -95 fill_link_info/kprobe_multi_invalid_ubuff # bpf_program__attach_kprobe_multi_opts unexpected error: -95 -- 2.40.1

1 week, 6 days

3
4
0 0

[PATCH] selftests/landlock: add binaries to gitignore

by Bharadwaj Raju

Building the test creates binaries 'wait-pipe' and 'sandbox-and-launch' which need to be gitignore'd. Signed-off-by: Bharadwaj Raju <bharadwaj.raju777(a)gmail.com> --- tools/testing/selftests/landlock/.gitignore | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tools/testing/selftests/landlock/.gitignore b/tools/testing/selftests/landlock/.gitignore index 470203a7cd73..0566c50dfcad 100644 --- a/tools/testing/selftests/landlock/.gitignore +++ b/tools/testing/selftests/landlock/.gitignore @@ -1,2 +1,4 @@ /*_test /true +/wait-pipe +/sandbox-and-launch -- 2.43.0

1 week, 6 days

2
1
0 0

[PATCH v6 00/14] iommufd: Add vIOMMU infrastructure (Part-3: vEVENTQ)

by Nicolin Chen

As the vIOMMU infrastructure series part-3, this introduces a new vEVENTQ object. The existing FAULT object provides a nice notification pathway to the user space with a queue already, so let vEVENTQ reuse that. Mimicing the HWPT structure, add a common EVENTQ structure to support its derivatives: IOMMUFD_OBJ_FAULT (existing) and IOMMUFD_OBJ_VEVENTQ (new). An IOMMUFD_CMD_VEVENTQ_ALLOC is introduced to allocate vEVENTQ object for vIOMMUs. One vIOMMU can have multiple vEVENTQs in different types but can not support multiple vEVENTQs in the same type. The forwarding part is fairly simple but might need to replace a physical device ID with a virtual device ID in a driver-level event data structure. So, this also adds some helpers for drivers to use. As usual, this series comes with the selftest coverage for this new ioctl and with a real world use case in the ARM SMMUv3 driver. This is on Github: https://github.com/nicolinc/iommufd/commits/iommufd_veventq-v6 Testing with RMR patches for MSI: https://github.com/nicolinc/iommufd/commits/iommufd_veventq-v6-with-rmr Paring QEMU branch for testing: https://github.com/nicolinc/qemu/commits/wip/for_iommufd_veventq-v6 Changelog v6 * Drop supports_veventq viommu op * Split bug/cosmetics fixes out of the series * Drop the blocking mutex around copy_to_user() * Add veventq_depth in uAPI to limit vEVENTQ size * Revise the documentation for a clear description * Fix sparse warnings in arm_vmaster_report_event() * Rework iommufd_viommu_get_vdev_id() to return -ENOENT v.s. 0 * Allow Abort/Bypass STEs to allocate vEVENTQ and set STE.MEV for DoS mitigations v5 https://lore.kernel.org/all/cover.1736237481.git.nicolinc@nvidia.com/ * Add Reviewed-by from Baolu * Reorder the OBJ list as well * Fix alphabetical order after renaming in v4 * Add supports_veventq viommu op for vEVENTQ type validation v4 https://lore.kernel.org/all/cover.1735933254.git.nicolinc@nvidia.com/ * Rename "vIRQ" to "vEVENTQ" * Use flexible array in struct iommufd_vevent * Add the new ioctl command to union ucmd_buffer * Fix the alphabetical order in union ucmd_buffer too * Rename _TYPE_NONE to _TYPE_DEFAULT aligning with vIOMMU naming v3 https://lore.kernel.org/all/cover.1734477608.git.nicolinc@nvidia.com/ * Rebase on Will's for-joerg/arm-smmu/updates for arm_smmu_event series * Add "Reviewed-by" lines from Kevin * Fix typos in comments, kdocs, and jump tags * Add a patch to sort struct iommufd_ioctl_op * Update iommufd's userpsace-api documentation * Update uAPI kdoc to quote SMMUv3 offical spec * Drop the unused workqueue in struct iommufd_virq * Drop might_sleep() in iommufd_viommu_report_irq() helper * Add missing "break" in iommufd_viommu_get_vdev_id() helper * Shrink the scope of the vmaster's read lock in SMMUv3 driver * Pass in two arguments to iommufd_eventq_virq_handler() helper * Move "!ops || !ops->read" validation into iommufd_eventq_init() * Move "fault->ictx = ictx" closer to iommufd_ctx_get(fault->ictx) * Update commit message for arm_smmu_attach_prepare/commit_vmaster() * Keep "iommufd_fault" as-is and rename "iommufd_eventq_virq" to just "iommufd_virq" v2 https://lore.kernel.org/all/cover.1733263737.git.nicolinc@nvidia.com/ * Rebase on v6.13-rc1 * Add IOPF and vIRQ in iommufd.rst (userspace-api) * Add a proper locking in iommufd_event_virq_destroy * Add iommufd_event_virq_abort with a lockdep_assert_held * Rename "EVENT_*" to "EVENTQ_*" to describe the objects better * Reorganize flows in iommufd_eventq_virq_alloc for abort() to work * Adde struct arm_smmu_vmaster to store vSID upon attaching to a nested domain, calling a newly added iommufd_viommu_get_vdev_id helper * Adde an arm_vmaster_report_event helper in arm-smmu-v3-iommufd file to simplify the routine in arm_smmu_handle_evt() of the main driver v1 https://lore.kernel.org/all/cover.1724777091.git.nicolinc@nvidia.com/ Thanks! Nicolin Nicolin Chen (14): iommufd/fault: Move two fault functions out of the header iommufd/fault: Add an iommufd_fault_init() helper iommufd: Abstract an iommufd_eventq from iommufd_fault iommufd: Rename fault.c to eventq.c iommufd: Add IOMMUFD_OBJ_VEVENTQ and IOMMUFD_CMD_VEVENTQ_ALLOC iommufd/viommu: Add iommufd_viommu_get_vdev_id helper iommufd/viommu: Add iommufd_viommu_report_event helper iommufd/selftest: Require vdev_id when attaching to a nested domain iommufd/selftest: Add IOMMU_TEST_OP_TRIGGER_VEVENT for vEVENTQ coverage iommufd/selftest: Add IOMMU_VEVENTQ_ALLOC test coverage Documentation: userspace-api: iommufd: Update FAULT and VEVENTQ iommu/arm-smmu-v3: Introduce struct arm_smmu_vmaster iommu/arm-smmu-v3: Report events that belong to devices attached to vIOMMU iommu/arm-smmu-v3: Set MEV bit in nested STE for DoS mitigations drivers/iommu/iommufd/Makefile | 2 +- drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h | 31 ++ drivers/iommu/iommufd/iommufd_private.h | 141 +++++-- drivers/iommu/iommufd/iommufd_test.h | 10 + include/linux/iommufd.h | 23 ++ include/uapi/linux/iommufd.h | 100 +++++ tools/testing/selftests/iommu/iommufd_utils.h | 115 ++++++ .../arm/arm-smmu-v3/arm-smmu-v3-iommufd.c | 62 +++ drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 94 +++-- drivers/iommu/iommufd/driver.c | 69 ++++ drivers/iommu/iommufd/{fault.c => eventq.c} | 364 +++++++++++++++--- drivers/iommu/iommufd/hw_pagetable.c | 6 +- drivers/iommu/iommufd/main.c | 7 + drivers/iommu/iommufd/selftest.c | 54 +++ drivers/iommu/iommufd/viommu.c | 2 + tools/testing/selftests/iommu/iommufd.c | 36 ++ .../selftests/iommu/iommufd_fail_nth.c | 7 + Documentation/userspace-api/iommufd.rst | 17 + 18 files changed, 1018 insertions(+), 122 deletions(-) rename drivers/iommu/iommufd/{fault.c => eventq.c} (50%) base-commit: e94dc6ddda8dd3770879a132d577accd2cce25f9 prerequisite-patch-id: bc39b89c8e2b8298a337943610e1cfd84d9b7d7d prerequisite-patch-id: 5cd371c3fddec696510e3e9c4f449dc60bd7c2ae prerequisite-patch-id: adbc6b7916b03f56eff01a9f1b33a7832fe0884e prerequisite-patch-id: c62d01dcfe8faeb928847fb4e51f82eebafe6ae3 prerequisite-patch-id: 0000000000000000000000000000000000000000 -- 2.43.0

1 week, 6 days

5
59
0 0

[PATCH] selftests: sched: add sched as a default selftest target

by Sinadin Shan

The sched tests are missing a target entry and hence out-of-tree build support. For instance: make -C tools/testing/selftests install INSTALL_LOCATION=/foo/bar is expected to build the sched tests and place them at /foo/bar. But this is not observed since a TARGET entry is not present for sched. This was suggested by Shuah in this conversation Link: https://lore.kernel.org/linux-kselftest/60dd0240-8e45-4958-acf2-7eeee917785… Add support for sched selftests by adding sched as a default TARGET Signed-off-by: Sinadin Shan <sinadin.shan(a)oracle.com> --- tools/testing/selftests/Makefile | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/testing/selftests/Makefile b/tools/testing/selftests/Makefile index 8daac70c2f9d2..e2d0d389ad912 100644 --- a/tools/testing/selftests/Makefile +++ b/tools/testing/selftests/Makefile @@ -91,6 +91,7 @@ TARGETS += rlimits TARGETS += rseq TARGETS += rtc TARGETS += rust +TARGETS += sched TARGETS += sched_ext TARGETS += seccomp TARGETS += sgx -- 2.43.5

1 week, 6 days

4
8
0 0

[PATCH] selftests/sysctl: fix wording of help messages

by Bharadwaj Raju

Fix paramter -> parameter, and recomended -> recommended in sysctl.sh's help messages. Also correct grammar: "number amount of times is recommended" etc -> "the recommended number of times". Signed-off-by: Bharadwaj Raju <bharadwaj.raju777(a)gmail.com> --- tools/testing/selftests/sysctl/sysctl.sh | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tools/testing/selftests/sysctl/sysctl.sh b/tools/testing/selftests/sysctl/sysctl.sh index 84472b436c07..a8e5736e89a8 100755 --- a/tools/testing/selftests/sysctl/sysctl.sh +++ b/tools/testing/selftests/sysctl/sysctl.sh @@ -764,7 +764,7 @@ sysctl_test_0007() fi if [ ! -f /proc/cmdline ]; then - echo -e "SKIPPING\nThere is no /proc/cmdline to check for paramter" + echo -e "SKIPPING\nThere is no /proc/cmdline to check for parameter" return $ksft_skip fi @@ -884,7 +884,7 @@ usage() echo "Valid tests: 0001-$MAX_TEST" echo "" echo " all Runs all tests (default)" - echo " -t Run test ID the number amount of times is recommended" + echo " -t Run test ID the recommended number of times" echo " -w Watch test ID run until it runs into an error" echo " -c Run test ID once" echo " -s Run test ID x test-count number of times" @@ -898,7 +898,7 @@ usage() echo Example uses: echo echo "$TEST_NAME.sh -- executes all tests" - echo "$TEST_NAME.sh -t 0002 -- Executes test ID 0002 number of times is recomended" + echo "$TEST_NAME.sh -t 0002 -- Executes test ID 0002 the recommended number of times" echo "$TEST_NAME.sh -w 0002 -- Watch test ID 0002 run until an error occurs" echo "$TEST_NAME.sh -s 0002 -- Run test ID 0002 once" echo "$TEST_NAME.sh -c 0002 3 -- Run test ID 0002 three times" -- 2.48.1

1 week, 6 days

2
1
0 0

[PATCH] selftests: fix spelling/grammar errors in sysctl/sysctl.sh

by Chandra Pratap

Fix the grammatical/spelling errors in sysctl/sysctl.sh. This fixes all errors pointed out by codespell in the file. Signed-off-by: Chandra Pratap <chandrapratap3519(a)gmail.com> --- tools/testing/selftests/sysctl/sysctl.sh | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tools/testing/selftests/sysctl/sysctl.sh b/tools/testing/selftests/sysctl/sysctl.sh index 84472b436c07..f6e129a82ffd 100755 --- a/tools/testing/selftests/sysctl/sysctl.sh +++ b/tools/testing/selftests/sysctl/sysctl.sh @@ -21,7 +21,7 @@ TEST_FILE=$(mktemp) # ENABLED: 1 if enabled, 0 otherwise # TARGET: test target file required on the test_sysctl module # SKIP_NO_TARGET: 1 skip if TARGET not there -# 0 run eventhough TARGET not there +# 0 run even though TARGET not there # # Once these are enabled please leave them as-is. Write your own test, # we have tons of space. @@ -764,7 +764,7 @@ sysctl_test_0007() fi if [ ! -f /proc/cmdline ]; then - echo -e "SKIPPING\nThere is no /proc/cmdline to check for paramter" + echo -e "SKIPPING\nThere is no /proc/cmdline to check for parameter" return $ksft_skip fi @@ -898,7 +898,7 @@ usage() echo Example uses: echo echo "$TEST_NAME.sh -- executes all tests" - echo "$TEST_NAME.sh -t 0002 -- Executes test ID 0002 number of times is recomended" + echo "$TEST_NAME.sh -t 0002 -- Executes test ID 0002 number of times is recommended" echo "$TEST_NAME.sh -w 0002 -- Watch test ID 0002 run until an error occurs" echo "$TEST_NAME.sh -s 0002 -- Run test ID 0002 once" echo "$TEST_NAME.sh -c 0002 3 -- Run test ID 0002 three times" -- 2.48.0-rc1

1 week, 6 days

2
1
0 0

[PATCH bpf-next v12 0/5] xsk: TX metadata Launch Time support

by Song Yoong Siang

This series expands the XDP TX metadata framework to allow user applications to pass per packet 64-bit launch time directly to the kernel driver, requesting launch time hardware offload support. The XDP TX metadata framework will not perform any clock conversion or packet reordering. Please note that the role of Tx metadata is just to pass the launch time, not to enable the offload feature. Users will need to enable the launch time hardware offload feature of the device by using the respective command, such as the tc-etf command. Although some devices use the tc-etf command to enable their launch time hardware offload feature, xsk packets will not go through the etf qdisc. Therefore, in my opinion, the launch time should always be based on the PTP Hardware Clock (PHC). Thus, i did not include a clock ID to indicate the clock source. To simplify the test steps, I modified the xdp_hw_metadata bpf self-test tool in such a way that it will set the launch time based on the offset provided by the user and the value of the Receive Hardware Timestamp, which is against the PHC. This will eliminate the need to discipline System Clock with the PHC and then use clock_gettime() to get the time. Please note that AF_XDP lacks a feedback mechanism to inform the application if the requested launch time is invalid. So, users are expected to familiar with the horizon of the launch time of the device they use and not request a launch time that is beyond the horizon. Otherwise, the driver might interpret the launch time incorrectly and react wrongly. For stmmac and igc, where modulo computation is used, a launch time larger than the horizon will cause the device to transmit the packet earlier that the requested launch time. Although there is no feedback mechanism for the launch time request for now, user still can check whether the requested launch time is working or not, by requesting the Transmit Completion Hardware Timestamp. v12: - Fix the comment in include/uapi/linux/if_xdp.h to allign with what is generated by ./tools/net/ynl/ynl-regen.sh to avoid dirty tree error in the netdev/ynl checks. v11: https://lore.kernel.org/netdev/20250216074302.956937-1-yoong.siang.song@int… - regenerate netdev_xsk_flags based on latest netdev.yaml (Jakub) v10: https://lore.kernel.org/netdev/20250207021943.814768-1-yoong.siang.song@int… - use net_err_ratelimited(), instead of net_ratelimit() (Maciej) - accumulate the amount of used descs in local variable and update the igc_metadata_request::used_desc once (Maciej) - Ensure reverse christmas tree rule (Maciej) V9: https://lore.kernel.org/netdev/20250206060408.808325-1-yoong.siang.song@int… - Remove the igc_desc_unused() checking (Maciej) - Ensure that skb allocation and DMA mapping work before proceeding to fill in igc_tx_buffer info, context desc, and data desc (Maciej) - Rate limit the error messages (Maciej) - Update the comment to indicate that the 2 descriptors needed by the empty frame are already taken into consideration (Maciej) - Handle the case where the insertion of an empty frame fails and explain the reason behind (Maciej) - put self SOB tag as last tag (Maciej) V8: https://lore.kernel.org/netdev/20250205024116.798862-1-yoong.siang.song@int… - check the number of used descriptor in xsk_tx_metadata_request() by using used_desc of struct igc_metadata_request, and then decreases the budget with it (Maciej) - submit another bug fix patch to set the buffer type for empty frame (Maciej): https://lore.kernel.org/netdev/20250205023603.798819-1-yoong.siang.song@int… V7: https://lore.kernel.org/netdev/20250204004907.789330-1-yoong.siang.song@int… - split the refactoring code of igc empty packet insertion into a separate commit (Faizal) - add explanation on why the value "4" is used as igc transmit budget (Faizal) - perform a stress test by sending 1000 packets with 10ms interval and launch time set to 500us in the future (Faizal & Yong Liang) V6: https://lore.kernel.org/netdev/20250116155350.555374-1-yoong.siang.song@int… - fix selftest build errors by using asprintf() and realloc(), instead of managing the buffer sizes manually (Daniel, Stanislav) V5: https://lore.kernel.org/netdev/20250114152718.120588-1-yoong.siang.song@int… - change netdev feature name from tx-launch-time to tx-launch-time-fifo to explicitly state the FIFO behaviour (Stanislav) - improve the looping of xdp_hw_metadata app to wait for packet tx completion to be more readable by using clock_gettime() (Stanislav) - add launch time setup steps into xdp_hw_metadata app (Stanislav) V4: https://lore.kernel.org/netdev/20250106135506.9687-1-yoong.siang.song@intel… - added XDP launch time support to the igc driver (Jesper & Florian) - added per-driver launch time limitation on xsk-tx-metadata.rst (Jesper) - added explanation on FIFO behavior on xsk-tx-metadata.rst (Jakub) - added step to enable launch time in the commit message (Jesper & Willem) - explicitly documented the type of launch_time and which clock source it is against (Willem) V3: https://lore.kernel.org/netdev/20231203165129.1740512-1-yoong.siang.song@in… - renamed to use launch time (Jesper & Willem) - changed the default launch time in xdp_hw_metadata apps from 1s to 0.1s because some NICs do not support such a large future time. V2: https://lore.kernel.org/netdev/20231201062421.1074768-1-yoong.siang.song@in… - renamed to use Earliest TxTime First (Willem) - renamed to use txtime (Willem) V1: https://lore.kernel.org/netdev/20231130162028.852006-1-yoong.siang.song@int… Song Yoong Siang (5): xsk: Add launch time hardware offload support to XDP Tx metadata selftests/bpf: Add launch time request to xdp_hw_metadata net: stmmac: Add launch time support to XDP ZC igc: Refactor empty frame insertion for launch time support igc: Add launch time support to XDP ZC Documentation/netlink/specs/netdev.yaml | 4 + Documentation/networking/xsk-tx-metadata.rst | 62 +++++++ drivers/net/ethernet/intel/igc/igc.h | 1 + drivers/net/ethernet/intel/igc/igc_main.c | 143 +++++++++++---- drivers/net/ethernet/stmicro/stmmac/stmmac.h | 2 + .../net/ethernet/stmicro/stmmac/stmmac_main.c | 13 ++ include/net/xdp_sock.h | 10 ++ include/net/xdp_sock_drv.h | 1 + include/uapi/linux/if_xdp.h | 10 ++ include/uapi/linux/netdev.h | 3 + net/core/netdev-genl.c | 2 + net/xdp/xsk.c | 3 + tools/include/uapi/linux/if_xdp.h | 10 ++ tools/include/uapi/linux/netdev.h | 3 + tools/testing/selftests/bpf/xdp_hw_metadata.c | 168 +++++++++++++++++- 15 files changed, 396 insertions(+), 39 deletions(-) -- 2.34.1

1 week, 6 days

3
7
0 0

[PATCH net-next v6 0/7] Some pktgen fixes/improvments (part I)

by Peter Seiderer

While taking a look at '[PATCH net] pktgen: Avoid out-of-range in get_imix_entries' ([1]) and '[PATCH net v2] pktgen: Avoid out-of-bounds access in get_imix_entries' ([2], [3]) and doing some tests and code review I detected that the /proc/net/pktgen/... parsing logic does not honour the user given buffer bounds (resulting in out-of-bounds access). This can be observed e.g. by the following simple test (sometimes the old/'longer' previous value is re-read from the buffer): $ echo add_device lo@0 > /proc/net/pktgen/kpktgend_0 $ echo "min_pkt_size 12345" > /proc/net/pktgen/lo\@0 && grep min_pkt_size /proc/net/pktgen/lo\@0 Params: count 1000 min_pkt_size: 12345 max_pkt_size: 0 Result: OK: min_pkt_size=12345 $ echo -n "min_pkt_size 123" > /proc/net/pktgen/lo\@0 && grep min_pkt_size /proc/net/pktgen/lo\@0 Params: count 1000 min_pkt_size: 12345 max_pkt_size: 0 Result: OK: min_pkt_size=12345 $ echo "min_pkt_size 123" > /proc/net/pktgen/lo\@0 && grep min_pkt_size /proc/net/pktgen/lo\@0 Params: count 1000 min_pkt_size: 123 max_pkt_size: 0 Result: OK: min_pkt_size=123 So fix the out-of-bounds access (and some minor findings) and add a simple proc_net_pktgen selftest... Patch set splited into part I (this one) - net: pktgen: replace ENOTSUPP with EOPNOTSUPP - net: pktgen: enable 'param=value' parsing - net: pktgen: fix hex32_arg parsing for short reads - net: pktgen: fix 'rate 0' error handling (return -EINVAL) - net: pktgen: fix 'ratep 0' error handling (return -EINVAL) - net: pktgen: fix ctrl interface command parsing - net: pktgen: fix access outside of user given buffer in pktgen_thread_write() And part II (will follow): - net: pktgen: fix mix of int/long - net: pktgen: remove extra tmp variable (re-use len instead) - net: pktgen: remove some superfluous variable initializing - net: pktgen: fix mpls maximum labels list parsing - net: pktgen: fix access outside of user given buffer in pktgen_if_write() - net: pktgen: fix mpls reset parsing - net: pktgen: remove all superfluous index assignements - selftest: net: add proc_net_pktgen Regards, Peter Changes v5 -> v6: - add rev-by Simon Horman - drop patch 'net: pktgen: use defines for the various dec/hex number parsing digits lengths' Changes v4 -> v5: - split up patchset into part i/ii (suggested by Simon Horman) Changes v3 -> v4: - add rev-by Simon Horman - new patch 'net: pktgen: use defines for the various dec/hex number parsing digits lengths' (suggested by Simon Horman) - replace C99 comment (suggested by Paolo Abeni) - drop available characters check in strn_len() (suggested by Paolo Abeni) - factored out patch 'net: pktgen: align some variable declarations to the most common pattern' (suggested by Paolo Abeni) - factored out patch 'net: pktgen: remove extra tmp variable (re-use len instead)' (suggested by Paolo Abeni) - factored out patch 'net: pktgen: remove some superfluous variable initializing' (suggested by Paolo Abeni) - factored out patch 'net: pktgen: fix mpls maximum labels list parsing' (suggested by Paolo Abeni) - factored out 'net: pktgen: hex32_arg/num_arg error out in case no characters are available' (suggested by Paolo Abeni) - factored out 'net: pktgen: num_arg error out in case no valid character is parsed' (suggested by Paolo Abeni) Changes v2 -> v3: - new patch: 'net: pktgen: fix ctrl interface command parsing' - new patch: 'net: pktgen: fix mpls reset parsing' - tools/testing/selftests/net/proc_net_pktgen.c: - fix typo in change description ('v1 -> v1' and tyop) - rename some vars to better match usage add_loopback_0 -> thr_cmd_add_loopback_0 rm_loopback_0 -> thr_cmd_rm_loopback_0 wrong_ctrl_cmd -> wrong_thr_cmd legacy_ctrl_cmd -> legacy_thr_cmd ctrl_fd -> thr_fd - add ctrl interface tests Changes v1 -> v2: - new patch: 'net: pktgen: fix hex32_arg parsing for short reads' - new patch: 'net: pktgen: fix 'rate 0' error handling (return -EINVAL)' - new patch: 'net: pktgen: fix 'ratep 0' error handling (return -EINVAL)' - net/core/pktgen.c: additional fix get_imix_entries() and get_labels() - tools/testing/selftests/net/proc_net_pktgen.c: - fix tyop not vs. nod (suggested by Jakub Kicinski) - fix misaligned line (suggested by Jakub Kicinski) - enable fomerly commented out CONFIG_XFRM dependent test (command spi), as CONFIG_XFRM is enabled via tools/testing/selftests/net/config CONFIG_XFRM_INTERFACE/CONFIG_XFRM_USER (suggestex by Jakub Kicinski) - add CONFIG_NET_PKTGEN=m to tools/testing/selftests/net/config (suggested by Jakub Kicinski) - add modprobe pktgen to FIXTURE_SETUP() (suggested by Jakub Kicinski) - fix some checkpatch warnings (Missing a blank line after declarations) - shrink line length by re-naming some variables (command -> cmd, device -> dev) - add 'rate 0' testcase - add 'ratep 0' testcase [1] https://lore.kernel.org/netdev/20241006221221.3744995-1-artem.chernyshev@re… [2] https://lore.kernel.org/netdev/20250109083039.14004-1-pchelkin@ispras.ru/ [3] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… Peter Seiderer (7): net: pktgen: replace ENOTSUPP with EOPNOTSUPP net: pktgen: enable 'param=value' parsing net: pktgen: fix hex32_arg parsing for short reads net: pktgen: fix 'rate 0' error handling (return -EINVAL) net: pktgen: fix 'ratep 0' error handling (return -EINVAL) net: pktgen: fix ctrl interface command parsing net: pktgen: fix access outside of user given buffer in pktgen_thread_write() net/core/pktgen.c | 39 ++++++++++++++++++++++----------------- 1 file changed, 22 insertions(+), 17 deletions(-) -- 2.48.1

1 week, 6 days

2
8
0 0

[PATCH net-next v2] tun: Pad virtio headers

by Akihiko Odaki

tun simply advances iov_iter when it needs to pad virtio header, which leaves the garbage in the buffer as is. This will become especially problematic when tun starts to allow enabling the hash reporting feature; even if the feature is enabled, the packet may lack a hash value and may contain a hole in the virtio header because the packet arrived before the feature gets enabled or does not contain the header fields to be hashed. If the hole is not filled with zero, it is impossible to tell if the packet lacks a hash value. In theory, a user of tun can fill the buffer with zero before calling read() to avoid such a problem, but leaving the garbage in the buffer is awkward anyway so replace advancing the iterator with writing zeros. A user might have initialized the buffer to some non-zero value, expecting tun to skip writing it. As this was never a documented feature, this seems unlikely. The overhead of filling the hole in the header is negligible when the header size is specified according to the specification as doing so will not make another cache line dirty under a reasonable assumption. Below is a proof of this statement: The first 10 bytes of the header is always written and tun also writes the packet itself immediately after the packet unless the packet is empty. This makes a hole between these writes whose size is: sz - 10 where sz is the specified header size. Therefore, we will never make another cache line dirty when: sz < L1_CACHE_BYTES + 10 where L1_CACHE_BYTES is the cache line size. Assuming L1_CACHE_BYTES >= 16, this inequation holds when: sz < 26. sz <= 20 according to the current specification so we even have a margin of 5 bytes in case that the header size grows in a future version of the specification. Signed-off-by: Akihiko Odaki <akihiko.odaki(a)daynix.com> --- Changes in v2: - Dropped the code to set num_buffers to 1. - Incorporated a grammatical improvement suggested by Michael S. Tsirkin. - Added an explanation of this patch's risk suggested by Michael S. Tsirkin. - Noted that it will not make another cache line dirty. - Added an error check. - Link to v1: https://lore.kernel.org/r/20250213-buffers-v1-1-ec4a0821957a@daynix.com --- drivers/net/tun_vnet.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/tun_vnet.h b/drivers/net/tun_vnet.h index fd7411c4447ffb180e032fe3e22f6709c30da8e9..58b9ac7a5fc4084c789fe94fe36b5f8631bf1fa4 100644 --- a/drivers/net/tun_vnet.h +++ b/drivers/net/tun_vnet.h @@ -143,7 +143,8 @@ static inline int tun_vnet_hdr_put(int sz, struct iov_iter *iter, if (unlikely(copy_to_iter(hdr, sizeof(*hdr), iter) != sizeof(*hdr))) return -EFAULT; - iov_iter_advance(iter, sz - sizeof(*hdr)); + if (iov_iter_zero(sz - sizeof(*hdr), iter) != sz - sizeof(*hdr)) + return -EFAULT; return 0; } --- base-commit: f54eab84fc17ef79b701e29364b7d08ca3a1d2f6 change-id: 20250116-buffers-96e14bf023fc prerequisite-change-id: 20241230-tun-66e10a49b0c7:v6 prerequisite-patch-id: 871dc5f146fb6b0e3ec8612971a8e8190472c0fb prerequisite-patch-id: 2797ed249d32590321f088373d4055ff3f430a0e prerequisite-patch-id: ea3370c72d4904e2f0536ec76ba5d26784c0cede prerequisite-patch-id: 837e4cf5d6b451424f9b1639455e83a260c4440d prerequisite-patch-id: ea701076f57819e844f5a35efe5cbc5712d3080d prerequisite-patch-id: 701646fb43ad04cc64dd2bf13c150ccbe6f828ce prerequisite-patch-id: 53176dae0c003f5b6c114d43f936cf7140d31bb5 Best regards, -- Akihiko Odaki <akihiko.odaki(a)daynix.com>

1 week, 6 days

5
4
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-kselftest-mirror February 2025