mirrors
/
linux-stable
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git
1
0
Fork 0
Code Issues Releases Wiki Activity
linux-stable/tools/testing/selftests/bpf/test_loader.c

737 lines
19 KiB
C
Raw Normal View History

selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
// SPDX-License-Identifier: GPL-2.0-only
/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
#include <linux/capability.h>
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
#include <stdlib.h>
#include <test_progs.h>
#include <bpf/btf.h>
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
#include "autoconf_helper.h"
#include "unpriv_helpers.h"
#include "cap_helpers.h"
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
#define str_has_pfx(str, pfx) \
(strncmp(str, pfx, __builtin_constant_p(pfx) ? sizeof(pfx) - 1 : strlen(pfx)) == 0)
selftests/bpf: Double the size of test_loader log Testing long jumps requires having >32k instructions. That many instructions require the verifier log buffer of 2 megabytes. The regular test_progs run doesn't need an increased buffer, since gotol test with 40k instructions doesn't request a log, but test_progs -v will set the verifier log level. Hence to avoid breaking gotol test with -v increase the buffer size. Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com> Acked-by: Yonghong Song <yonghong.song@linux.dev> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/r/20240102193531.3169422-3-iii@linux.ibm.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2024-01-02 19:30:36 +00:00
#define TEST_LOADER_LOG_BUF_SZ 2097152
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
#define TEST_TAG_EXPECT_FAILURE "comment:test_expect_failure"
#define TEST_TAG_EXPECT_SUCCESS "comment:test_expect_success"
#define TEST_TAG_EXPECT_MSG_PFX "comment:test_expect_msg="
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
#define TEST_TAG_EXPECT_FAILURE_UNPRIV "comment:test_expect_failure_unpriv"
#define TEST_TAG_EXPECT_SUCCESS_UNPRIV "comment:test_expect_success_unpriv"
#define TEST_TAG_EXPECT_MSG_PFX_UNPRIV "comment:test_expect_msg_unpriv="
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
#define TEST_TAG_LOG_LEVEL_PFX "comment:test_log_level="
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
#define TEST_TAG_PROG_FLAGS_PFX "comment:test_prog_flags="
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
#define TEST_TAG_DESCRIPTION_PFX "comment:test_description="
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
#define TEST_TAG_RETVAL_PFX "comment:test_retval="
#define TEST_TAG_RETVAL_PFX_UNPRIV "comment:test_retval_unpriv="
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
#define TEST_TAG_AUXILIARY "comment:test_auxiliary"
#define TEST_TAG_AUXILIARY_UNPRIV "comment:test_auxiliary_unpriv"
bpf: selftests: test_loader: Support __btf_path() annotation This commit adds support for per-prog btf_custom_path. This is necessary for testing CO-RE relocations on non-vmlinux types using test_loader infrastructure. Acked-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Daniel Xu <dxu@dxuuu.xyz> Link: https://lore.kernel.org/r/660ea7f2fdbdd5103bc1af87c9fc931f05327926.1702325874.git.dxu@dxuuu.xyz Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-12-11 20:20:08 +00:00
#define TEST_BTF_PATH "comment:test_btf_path="
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
/* Warning: duplicated in bpf_misc.h */
#define POINTER_VALUE 0xcafe4all
#define TEST_DATA_LEN 64
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
#ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
#define EFFICIENT_UNALIGNED_ACCESS 1
#else
#define EFFICIENT_UNALIGNED_ACCESS 0
#endif
static int sysctl_unpriv_disabled = -1;
enum mode {
PRIV = 1,
UNPRIV = 2
};
struct test_subspec {
char *name;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
bool expect_failure;
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
const char **expect_msgs;
size_t expect_msg_cnt;
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
int retval;
bool execute;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
};
struct test_spec {
const char *prog_name;
struct test_subspec priv;
struct test_subspec unpriv;
bpf: selftests: test_loader: Support __btf_path() annotation This commit adds support for per-prog btf_custom_path. This is necessary for testing CO-RE relocations on non-vmlinux types using test_loader infrastructure. Acked-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Daniel Xu <dxu@dxuuu.xyz> Link: https://lore.kernel.org/r/660ea7f2fdbdd5103bc1af87c9fc931f05327926.1702325874.git.dxu@dxuuu.xyz Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-12-11 20:20:08 +00:00
const char *btf_custom_path;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
int log_level;
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
int prog_flags;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
int mode_mask;
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
bool auxiliary;
bool valid;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
};
static int tester_init(struct test_loader *tester)
{
if (!tester->log_buf) {
tester->log_buf_sz = TEST_LOADER_LOG_BUF_SZ;
selftests/bpf: Fix compiler warnings reported in -O2 mode Fix a bunch of potentially unitialized variable usage warnings that are reported by GCC in -O2 mode. Also silence overzealous stringop-truncation class of warnings. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Jiri Olsa <jolsa@kernel.org> Link: https://lore.kernel.org/bpf/20231006175744.3136675-1-andrii@kernel.org
2023-10-06 17:57:42 +00:00
tester->log_buf = calloc(tester->log_buf_sz, 1);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
if (!ASSERT_OK_PTR(tester->log_buf, "tester_log_buf"))
return -ENOMEM;
}
return 0;
}
void test_loader_fini(struct test_loader *tester)
{
if (!tester)
return;
free(tester->log_buf);
}
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
static void free_test_spec(struct test_spec *spec)
{
free(spec->priv.name);
free(spec->unpriv.name);
free(spec->priv.expect_msgs);
free(spec->unpriv.expect_msgs);
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
spec->priv.name = NULL;
spec->unpriv.name = NULL;
spec->priv.expect_msgs = NULL;
spec->unpriv.expect_msgs = NULL;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
}
static int push_msg(const char *msg, struct test_subspec *subspec)
{
void *tmp;
tmp = realloc(subspec->expect_msgs, (1 + subspec->expect_msg_cnt) * sizeof(void *));
if (!tmp) {
ASSERT_FAIL("failed to realloc memory for messages\n");
return -ENOMEM;
}
subspec->expect_msgs = tmp;
subspec->expect_msgs[subspec->expect_msg_cnt++] = msg;
return 0;
}
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
static int parse_int(const char *str, int *val, const char *name)
{
char *end;
long tmp;
errno = 0;
if (str_has_pfx(str, "0x"))
tmp = strtol(str + 2, &end, 16);
else
tmp = strtol(str, &end, 10);
if (errno || end[0] != '\0') {
PRINT_FAIL("failed to parse %s from '%s'\n", name, str);
return -EINVAL;
}
*val = tmp;
return 0;
}
static int parse_retval(const char *str, int *val, const char *name)
{
struct {
char *name;
int val;
} named_values[] = {
{ "INT_MIN" , INT_MIN },
{ "POINTER_VALUE", POINTER_VALUE },
{ "TEST_DATA_LEN", TEST_DATA_LEN },
};
int i;
for (i = 0; i < ARRAY_SIZE(named_values); ++i) {
if (strcmp(str, named_values[i].name) != 0)
continue;
*val = named_values[i].val;
return 0;
}
return parse_int(str, val, name);
}
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
static void update_flags(int *flags, int flag, bool clear)
{
if (clear)
*flags &= ~flag;
else
*flags |= flag;
}
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
/* Uses btf_decl_tag attributes to describe the expected test
* behavior, see bpf_misc.h for detailed description of each attribute
* and attribute combinations.
*/
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
static int parse_test_spec(struct test_loader *tester,
struct bpf_object *obj,
struct bpf_program *prog,
struct test_spec *spec)
{
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
const char *description = NULL;
bool has_unpriv_result = false;
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
bool has_unpriv_retval = false;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
int func_id, i, err = 0;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
struct btf *btf;
memset(spec, 0, sizeof(*spec));
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
spec->prog_name = bpf_program__name(prog);
selftests/bpf: detect testing prog flags support Various tests specify extra testing prog_flags when loading BPF programs, like BPF_F_TEST_RND_HI32, and more recently also BPF_F_TEST_REG_INVARIANTS. While BPF_F_TEST_RND_HI32 is old enough to not cause much problem on older kernels, BPF_F_TEST_REG_INVARIANTS is very fresh and unconditionally specifying it causes selftests to fail on even slightly outdated kernels. This breaks libbpf CI test against 4.9 and 5.15 kernels, it can break some local development (done outside of VM), etc. To prevent this, and guard against similar problems in the future, do runtime detection of supported "testing flags", and only provide those that host kernel recognizes. Acked-by: Song Liu <song@kernel.org> Acked-by: Jiri Olsa <jolsa@kernel.org> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20240109231738.575844-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2024-01-09 23:17:38 +00:00
spec->prog_flags = testing_prog_flags();
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
btf = bpf_object__btf(obj);
if (!btf) {
ASSERT_FAIL("BPF object has no BTF");
return -EINVAL;
}
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
func_id = btf__find_by_name_kind(btf, spec->prog_name, BTF_KIND_FUNC);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
if (func_id < 0) {
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
ASSERT_FAIL("failed to find FUNC BTF type for '%s'", spec->prog_name);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
return -EINVAL;
}
for (i = 1; i < btf__type_cnt(btf); i++) {
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
const char *s, *val, *msg;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
const struct btf_type *t;
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
bool clear;
int flags;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
t = btf__type_by_id(btf, i);
if (!btf_is_decl_tag(t))
continue;
if (t->type != func_id || btf_decl_tag(t)->component_idx != -1)
continue;
s = btf__str_by_offset(btf, t->name_off);
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
if (str_has_pfx(s, TEST_TAG_DESCRIPTION_PFX)) {
description = s + sizeof(TEST_TAG_DESCRIPTION_PFX) - 1;
} else if (strcmp(s, TEST_TAG_EXPECT_FAILURE) == 0) {
spec->priv.expect_failure = true;
spec->mode_mask |= PRIV;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
} else if (strcmp(s, TEST_TAG_EXPECT_SUCCESS) == 0) {
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
spec->priv.expect_failure = false;
spec->mode_mask |= PRIV;
} else if (strcmp(s, TEST_TAG_EXPECT_FAILURE_UNPRIV) == 0) {
spec->unpriv.expect_failure = true;
spec->mode_mask |= UNPRIV;
has_unpriv_result = true;
} else if (strcmp(s, TEST_TAG_EXPECT_SUCCESS_UNPRIV) == 0) {
spec->unpriv.expect_failure = false;
spec->mode_mask |= UNPRIV;
has_unpriv_result = true;
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
} else if (strcmp(s, TEST_TAG_AUXILIARY) == 0) {
spec->auxiliary = true;
spec->mode_mask |= PRIV;
} else if (strcmp(s, TEST_TAG_AUXILIARY_UNPRIV) == 0) {
spec->auxiliary = true;
spec->mode_mask |= UNPRIV;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
} else if (str_has_pfx(s, TEST_TAG_EXPECT_MSG_PFX)) {
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
msg = s + sizeof(TEST_TAG_EXPECT_MSG_PFX) - 1;
err = push_msg(msg, &spec->priv);
if (err)
goto cleanup;
spec->mode_mask |= PRIV;
} else if (str_has_pfx(s, TEST_TAG_EXPECT_MSG_PFX_UNPRIV)) {
msg = s + sizeof(TEST_TAG_EXPECT_MSG_PFX_UNPRIV) - 1;
err = push_msg(msg, &spec->unpriv);
if (err)
goto cleanup;
spec->mode_mask |= UNPRIV;
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
} else if (str_has_pfx(s, TEST_TAG_RETVAL_PFX)) {
val = s + sizeof(TEST_TAG_RETVAL_PFX) - 1;
err = parse_retval(val, &spec->priv.retval, "__retval");
if (err)
goto cleanup;
spec->priv.execute = true;
spec->mode_mask |= PRIV;
} else if (str_has_pfx(s, TEST_TAG_RETVAL_PFX_UNPRIV)) {
val = s + sizeof(TEST_TAG_RETVAL_PFX_UNPRIV) - 1;
err = parse_retval(val, &spec->unpriv.retval, "__retval_unpriv");
if (err)
goto cleanup;
spec->mode_mask |= UNPRIV;
spec->unpriv.execute = true;
has_unpriv_retval = true;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
} else if (str_has_pfx(s, TEST_TAG_LOG_LEVEL_PFX)) {
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
val = s + sizeof(TEST_TAG_LOG_LEVEL_PFX) - 1;
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
err = parse_int(val, &spec->log_level, "test log level");
if (err)
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
goto cleanup;
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
} else if (str_has_pfx(s, TEST_TAG_PROG_FLAGS_PFX)) {
val = s + sizeof(TEST_TAG_PROG_FLAGS_PFX) - 1;
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
clear = val[0] == '!';
if (clear)
val++;
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
if (strcmp(val, "BPF_F_STRICT_ALIGNMENT") == 0) {
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
update_flags(&spec->prog_flags, BPF_F_STRICT_ALIGNMENT, clear);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
} else if (strcmp(val, "BPF_F_ANY_ALIGNMENT") == 0) {
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
update_flags(&spec->prog_flags, BPF_F_ANY_ALIGNMENT, clear);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
} else if (strcmp(val, "BPF_F_TEST_RND_HI32") == 0) {
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
update_flags(&spec->prog_flags, BPF_F_TEST_RND_HI32, clear);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
} else if (strcmp(val, "BPF_F_TEST_STATE_FREQ") == 0) {
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
update_flags(&spec->prog_flags, BPF_F_TEST_STATE_FREQ, clear);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
} else if (strcmp(val, "BPF_F_SLEEPABLE") == 0) {
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
update_flags(&spec->prog_flags, BPF_F_SLEEPABLE, clear);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
} else if (strcmp(val, "BPF_F_XDP_HAS_FRAGS") == 0) {
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
update_flags(&spec->prog_flags, BPF_F_XDP_HAS_FRAGS, clear);
bpf: rename BPF_F_TEST_SANITY_STRICT to BPF_F_TEST_REG_INVARIANTS Rename verifier internal flag BPF_F_TEST_SANITY_STRICT to more neutral BPF_F_TEST_REG_INVARIANTS. This is a follow up to [0]. A few selftests and veristat need to be adjusted in the same patch as well. [0] https://patchwork.kernel.org/project/netdevbpf/patch/20231112010609.848406-5-andrii@kernel.org/ Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231117171404.225508-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-17 17:14:04 +00:00
} else if (strcmp(val, "BPF_F_TEST_REG_INVARIANTS") == 0) {
update_flags(&spec->prog_flags, BPF_F_TEST_REG_INVARIANTS, clear);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
} else /* assume numeric value */ {
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
err = parse_int(val, &flags, "test prog flags");
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
if (err)
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
goto cleanup;
selftests/bpf: set BPF_F_TEST_SANITY_SCRIPT by default Make sure to set BPF_F_TEST_SANITY_STRICT program flag by default across most verifier tests (and a bunch of others that set custom prog flags). There are currently two tests that do fail validation, if enforced strictly: verifier_bounds/crossing_64_bit_signed_boundary_2 and verifier_bounds/crossing_32_bit_signed_boundary_2. To accommodate them, we teach test_loader a flag negation: __flag(!<flagname>) will *clear* specified flag, allowing easy opt-out. We apply __flag(!BPF_F_TEST_SANITY_STRICT) to these to tests. Also sprinkle BPF_F_TEST_SANITY_STRICT everywhere where we already set test-only BPF_F_TEST_RND_HI32 flag, for completeness. Acked-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20231112010609.848406-12-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:07 +00:00
update_flags(&spec->prog_flags, flags, clear);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
}
bpf: selftests: test_loader: Support __btf_path() annotation This commit adds support for per-prog btf_custom_path. This is necessary for testing CO-RE relocations on non-vmlinux types using test_loader infrastructure. Acked-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Daniel Xu <dxu@dxuuu.xyz> Link: https://lore.kernel.org/r/660ea7f2fdbdd5103bc1af87c9fc931f05327926.1702325874.git.dxu@dxuuu.xyz Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-12-11 20:20:08 +00:00
} else if (str_has_pfx(s, TEST_BTF_PATH)) {
spec->btf_custom_path = s + sizeof(TEST_BTF_PATH) - 1;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
}
}
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
if (spec->mode_mask == 0)
spec->mode_mask = PRIV;
if (!description)
description = spec->prog_name;
if (spec->mode_mask & PRIV) {
spec->priv.name = strdup(description);
if (!spec->priv.name) {
PRINT_FAIL("failed to allocate memory for priv.name\n");
err = -ENOMEM;
goto cleanup;
}
}
if (spec->mode_mask & UNPRIV) {
int descr_len = strlen(description);
const char *suffix = " @unpriv";
char *name;
name = malloc(descr_len + strlen(suffix) + 1);
if (!name) {
PRINT_FAIL("failed to allocate memory for unpriv.name\n");
err = -ENOMEM;
goto cleanup;
}
strcpy(name, description);
strcpy(&name[descr_len], suffix);
spec->unpriv.name = name;
}
if (spec->mode_mask & (PRIV | UNPRIV)) {
if (!has_unpriv_result)
spec->unpriv.expect_failure = spec->priv.expect_failure;
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
if (!has_unpriv_retval) {
spec->unpriv.retval = spec->priv.retval;
spec->unpriv.execute = spec->priv.execute;
}
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
if (!spec->unpriv.expect_msgs) {
size_t sz = spec->priv.expect_msg_cnt * sizeof(void *);
spec->unpriv.expect_msgs = malloc(sz);
if (!spec->unpriv.expect_msgs) {
PRINT_FAIL("failed to allocate memory for unpriv.expect_msgs\n");
err = -ENOMEM;
goto cleanup;
}
memcpy(spec->unpriv.expect_msgs, spec->priv.expect_msgs, sz);
spec->unpriv.expect_msg_cnt = spec->priv.expect_msg_cnt;
}
}
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
spec->valid = true;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
return 0;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
cleanup:
free_test_spec(spec);
return err;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
}
static void prepare_case(struct test_loader *tester,
struct test_spec *spec,
struct bpf_object *obj,
struct bpf_program *prog)
{
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
int min_log_level = 0, prog_flags;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
if (env.verbosity > VERBOSE_NONE)
min_log_level = 1;
if (env.verbosity > VERBOSE_VERY)
min_log_level = 2;
bpf_program__set_log_buf(prog, tester->log_buf, tester->log_buf_sz);
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
/* Make sure we set at least minimal log level, unless test requires
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
* even higher level already. Make sure to preserve independent log
* level 4 (verifier stats), though.
*/
if ((spec->log_level & 3) < min_log_level)
bpf_program__set_log_level(prog, (spec->log_level & 4) | min_log_level);
else
bpf_program__set_log_level(prog, spec->log_level);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
prog_flags = bpf_program__flags(prog);
bpf_program__set_flags(prog, prog_flags | spec->prog_flags);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
tester->log_buf[0] = '\0';
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
tester->next_match_pos = 0;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
}
static void emit_verifier_log(const char *log_buf, bool force)
{
if (!force && env.verbosity == VERBOSE_NONE)
return;
fprintf(stdout, "VERIFIER LOG:\n=============\n%s=============\n", log_buf);
}
static void validate_case(struct test_loader *tester,
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
struct test_subspec *subspec,
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
struct bpf_object *obj,
struct bpf_program *prog,
int load_err)
{
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
int i, j;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
for (i = 0; i < subspec->expect_msg_cnt; i++) {
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
char *match;
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
const char *expect_msg;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
expect_msg = subspec->expect_msgs[i];
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
match = strstr(tester->log_buf + tester->next_match_pos, expect_msg);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
if (!ASSERT_OK_PTR(match, "expect_msg")) {
/* if we are in verbose mode, we've already emitted log */
if (env.verbosity == VERBOSE_NONE)
emit_verifier_log(tester->log_buf, true /*force*/);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
for (j = 0; j < i; j++)
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
fprintf(stderr,
"MATCHED MSG: '%s'\n", subspec->expect_msgs[j]);
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
fprintf(stderr, "EXPECTED MSG: '%s'\n", expect_msg);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
return;
}
selftests/bpf: Support custom per-test flags and multiple expected messages Extend __flag attribute by allowing to specify one of the following: * BPF_F_STRICT_ALIGNMENT * BPF_F_ANY_ALIGNMENT * BPF_F_TEST_RND_HI32 * BPF_F_TEST_STATE_FREQ * BPF_F_SLEEPABLE * BPF_F_XDP_HAS_FRAGS * Some numeric value Extend __msg attribute by allowing to specify multiple exepcted messages. All messages are expected to be present in the verifier log in the order of application. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20230301175417.3146070-2-eddyz87@gmail.com [ Eduard: added commit message, formatting, comments ]
2023-03-01 17:54:17 +00:00
tester->next_match_pos = match - tester->log_buf + strlen(expect_msg);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
}
}
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
struct cap_state {
__u64 old_caps;
bool initialized;
};
static int drop_capabilities(struct cap_state *caps)
{
const __u64 caps_to_drop = (1ULL << CAP_SYS_ADMIN | 1ULL << CAP_NET_ADMIN |
1ULL << CAP_PERFMON | 1ULL << CAP_BPF);
int err;
err = cap_disable_effective(caps_to_drop, &caps->old_caps);
if (err) {
PRINT_FAIL("failed to drop capabilities: %i, %s\n", err, strerror(err));
return err;
}
caps->initialized = true;
return 0;
}
static int restore_capabilities(struct cap_state *caps)
{
int err;
if (!caps->initialized)
return 0;
err = cap_enable_effective(caps->old_caps, NULL);
if (err)
PRINT_FAIL("failed to restore capabilities: %i, %s\n", err, strerror(err));
caps->initialized = false;
return err;
}
static bool can_execute_unpriv(struct test_loader *tester, struct test_spec *spec)
{
if (sysctl_unpriv_disabled < 0)
sysctl_unpriv_disabled = get_unpriv_disabled() ? 1 : 0;
if (sysctl_unpriv_disabled)
return false;
if ((spec->prog_flags & BPF_F_ANY_ALIGNMENT) && !EFFICIENT_UNALIGNED_ACCESS)
return false;
return true;
}
static bool is_unpriv_capable_map(struct bpf_map *map)
{
enum bpf_map_type type;
__u32 flags;
type = bpf_map__type(map);
switch (type) {
case BPF_MAP_TYPE_HASH:
case BPF_MAP_TYPE_PERCPU_HASH:
case BPF_MAP_TYPE_HASH_OF_MAPS:
flags = bpf_map__map_flags(map);
return !(flags & BPF_F_ZERO_SEED);
case BPF_MAP_TYPE_PERCPU_CGROUP_STORAGE:
case BPF_MAP_TYPE_ARRAY:
case BPF_MAP_TYPE_RINGBUF:
case BPF_MAP_TYPE_PROG_ARRAY:
case BPF_MAP_TYPE_CGROUP_ARRAY:
case BPF_MAP_TYPE_PERCPU_ARRAY:
case BPF_MAP_TYPE_USER_RINGBUF:
case BPF_MAP_TYPE_ARRAY_OF_MAPS:
case BPF_MAP_TYPE_CGROUP_STORAGE:
case BPF_MAP_TYPE_PERF_EVENT_ARRAY:
return true;
default:
return false;
}
}
selftests/bpf: Add unit tests for bpf_arena_alloc/free_pages Add unit tests for bpf_arena_alloc/free_pages() functionality and bpf_arena_common.h with a set of common helpers and macros that is used in this test and the following patches. Also modify test_loader that didn't support running bpf_prog_type_syscall programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20240308010812.89848-13-alexei.starovoitov@gmail.com
2024-03-08 01:08:10 +00:00
static int do_prog_test_run(int fd_prog, int *retval, bool empty_opts)
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
{
__u8 tmp_out[TEST_DATA_LEN << 2] = {};
__u8 tmp_in[TEST_DATA_LEN] = {};
int err, saved_errno;
LIBBPF_OPTS(bpf_test_run_opts, topts,
.data_in = tmp_in,
.data_size_in = sizeof(tmp_in),
.data_out = tmp_out,
.data_size_out = sizeof(tmp_out),
.repeat = 1,
);
selftests/bpf: Add unit tests for bpf_arena_alloc/free_pages Add unit tests for bpf_arena_alloc/free_pages() functionality and bpf_arena_common.h with a set of common helpers and macros that is used in this test and the following patches. Also modify test_loader that didn't support running bpf_prog_type_syscall programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20240308010812.89848-13-alexei.starovoitov@gmail.com
2024-03-08 01:08:10 +00:00
if (empty_opts) {
memset(&topts, 0, sizeof(struct bpf_test_run_opts));
topts.sz = sizeof(struct bpf_test_run_opts);
}
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
err = bpf_prog_test_run_opts(fd_prog, &topts);
saved_errno = errno;
if (err) {
PRINT_FAIL("FAIL: Unexpected bpf_prog_test_run error: %d (%s) ",
saved_errno, strerror(saved_errno));
return err;
}
ASSERT_OK(0, "bpf_prog_test_run");
*retval = topts.retval;
return 0;
}
static bool should_do_test_run(struct test_spec *spec, struct test_subspec *subspec)
{
if (!subspec->execute)
return false;
if (subspec->expect_failure)
return false;
if ((spec->prog_flags & BPF_F_ANY_ALIGNMENT) && !EFFICIENT_UNALIGNED_ACCESS) {
if (env.verbosity != VERBOSE_NONE)
printf("alignment prevents execution\n");
return false;
}
return true;
}
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
/* this function is forced noinline and has short generic name to look better
* in test_progs output (in case of a failure)
*/
static noinline
void run_subtest(struct test_loader *tester,
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
struct bpf_object_open_opts *open_opts,
const void *obj_bytes,
size_t obj_byte_cnt,
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
struct test_spec *specs,
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
struct test_spec *spec,
bool unpriv)
{
struct test_subspec *subspec = unpriv ? &spec->unpriv : &spec->priv;
selftests/bpf: Fix compiler warnings reported in -O2 mode Fix a bunch of potentially unitialized variable usage warnings that are reported by GCC in -O2 mode. Also silence overzealous stringop-truncation class of warnings. Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Jiri Olsa <jolsa@kernel.org> Link: https://lore.kernel.org/bpf/20231006175744.3136675-1-andrii@kernel.org
2023-10-06 17:57:42 +00:00
struct bpf_program *tprog = NULL, *tprog_iter;
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
struct test_spec *spec_iter;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
struct cap_state caps = {};
struct bpf_object *tobj;
struct bpf_map *map;
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
int retval, err, i;
bool should_load;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
if (!test__start_subtest(subspec->name))
return;
if (unpriv) {
if (!can_execute_unpriv(tester, spec)) {
test__skip();
test__end_subtest();
return;
}
if (drop_capabilities(&caps)) {
test__end_subtest();
return;
}
}
bpf: selftests: test_loader: Support __btf_path() annotation This commit adds support for per-prog btf_custom_path. This is necessary for testing CO-RE relocations on non-vmlinux types using test_loader infrastructure. Acked-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Daniel Xu <dxu@dxuuu.xyz> Link: https://lore.kernel.org/r/660ea7f2fdbdd5103bc1af87c9fc931f05327926.1702325874.git.dxu@dxuuu.xyz Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-12-11 20:20:08 +00:00
/* Implicitly reset to NULL if next test case doesn't specify */
open_opts->btf_custom_path = spec->btf_custom_path;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
tobj = bpf_object__open_mem(obj_bytes, obj_byte_cnt, open_opts);
if (!ASSERT_OK_PTR(tobj, "obj_open_mem")) /* shouldn't happen */
goto subtest_cleanup;
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
i = 0;
bpf_object__for_each_program(tprog_iter, tobj) {
spec_iter = &specs[i++];
should_load = false;
if (spec_iter->valid) {
if (strcmp(bpf_program__name(tprog_iter), spec->prog_name) == 0) {
tprog = tprog_iter;
should_load = true;
}
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
if (spec_iter->auxiliary &&
spec_iter->mode_mask & (unpriv ? UNPRIV : PRIV))
should_load = true;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
}
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
bpf_program__set_autoload(tprog_iter, should_load);
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
}
prepare_case(tester, spec, tobj, tprog);
/* By default bpf_object__load() automatically creates all
* maps declared in the skeleton. Some map types are only
* allowed in priv mode. Disable autoload for such maps in
* unpriv mode.
*/
bpf_object__for_each_map(map, tobj)
bpf_map__set_autocreate(map, !unpriv || is_unpriv_capable_map(map));
err = bpf_object__load(tobj);
if (subspec->expect_failure) {
if (!ASSERT_ERR(err, "unexpected_load_success")) {
emit_verifier_log(tester->log_buf, false /*force*/);
goto tobj_cleanup;
}
} else {
if (!ASSERT_OK(err, "unexpected_load_failure")) {
emit_verifier_log(tester->log_buf, true /*force*/);
goto tobj_cleanup;
}
}
emit_verifier_log(tester->log_buf, false /*force*/);
validate_case(tester, subspec, tobj, tprog, err);
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
if (should_do_test_run(spec, subspec)) {
/* For some reason test_verifier executes programs
* with all capabilities restored. Do the same here.
*/
selftests/bpf: fix __retval() being always ignored Florian Westphal found a bug in and suggested a fix for test_loader.c processing of __retval tag. Because of this bug the function test_loader.c:do_prog_test_run() never executed and all __retval test tags were ignored. If this bug is fixed a number of test cases from progs/verifier_array_access.c fail with retval not matching the expected value. This test was recently converted to use test_loader.c and inline assembly in [1]. When doing the conversion I missed the important detail of test_verifier.c operation: when it creates fixup_map_array_ro, fixup_map_array_wo and fixup_map_array_small it populates these maps with a dummy record. Disabling the __retval checks for the affected verifier_array_access in this commit to avoid false-postivies in any potential bisects. The issue is addressed in the next patch. I verified that the __retval tags are now respected by changing expected return values for all tests annotated with __retval, and checking that these tests started to fail. [1] https://lore.kernel.org/bpf/20230325025524.144043-1-eddyz87@gmail.com/ Fixes: 19a8e06f5f91 ("selftests/bpf: Tests execution support for test_loader.c") Reported-by: Florian Westphal <fw@strlen.de> Link: https://lore.kernel.org/bpf/f4c4aee644425842ee6aa8edf1da68f0a8260e7c.camel@gmail.com/T/ Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230420232317.2181776-3-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-20 23:23:15 +00:00
if (restore_capabilities(&caps))
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
goto tobj_cleanup;
selftests/bpf: add pre bpf_prog_test_run_opts() callback for test_loader When a test case is annotated with __retval tag the test_loader engine would use libbpf's bpf_prog_test_run_opts() to do a test run of the program and compare retvals. This commit allows to perform arbitrary actions on bpf object right before test loader invokes bpf_prog_test_run_opts(). This could be used to setup some state for program execution, e.g. fill some maps. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230420232317.2181776-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-20 23:23:16 +00:00
if (tester->pre_execution_cb) {
err = tester->pre_execution_cb(tobj);
if (err) {
PRINT_FAIL("pre_execution_cb failed: %d\n", err);
goto tobj_cleanup;
}
}
selftests/bpf: Add unit tests for bpf_arena_alloc/free_pages Add unit tests for bpf_arena_alloc/free_pages() functionality and bpf_arena_common.h with a set of common helpers and macros that is used in this test and the following patches. Also modify test_loader that didn't support running bpf_prog_type_syscall programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20240308010812.89848-13-alexei.starovoitov@gmail.com
2024-03-08 01:08:10 +00:00
do_prog_test_run(bpf_program__fd(tprog), &retval,
bpf_program__type(tprog) == BPF_PROG_TYPE_SYSCALL ? true : false);
selftests/bpf: Tests execution support for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute BPF_PROG_TEST_RUN bpf command for selected programs. This is similar to functionality provided by test_verifier. Adds the following new attributes controlling test_loader behavior: __retval(...) __retval_unpriv(...) * If any of these attributes is present, the annotated program would be executed using libbpf's bpf_prog_test_run_opts() function. * If __retval is present, the test run would be done for program loaded in privileged mode. * If __retval_unpriv is present, the test run would be done for program loaded in unprivileged mode. * To mimic test_verifier behavior, the actual run is initiated in privileged mode. * The value returned by a test run is compared against retval parameter. The retval attribute takes one of the following parameters: - a decimal number - a hexadecimal number (must start from '0x') - any of a three special literals (provided for compatibility with test_verifier): - INT_MIN - POINTER_VALUE - TEST_DATA_LEN An example of the attribute usage: SEC("socket") __description("return 42") __success __success_unpriv __retval(42) __naked void the_42_test(void) { asm volatile (" \ r0 = 42; \ exit; \ " ::: __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-5-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:45 +00:00
if (retval != subspec->retval && subspec->retval != POINTER_VALUE) {
PRINT_FAIL("Unexpected retval: %d != %d\n", retval, subspec->retval);
goto tobj_cleanup;
}
}
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
tobj_cleanup:
bpf_object__close(tobj);
subtest_cleanup:
test__end_subtest();
restore_capabilities(&caps);
}
static void process_subtest(struct test_loader *tester,
const char *skel_name,
skel_elf_bytes_fn elf_bytes_factory)
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
{
LIBBPF_OPTS(bpf_object_open_opts, open_opts, .object_name = skel_name);
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
struct test_spec *specs = NULL;
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
struct bpf_object *obj = NULL;
struct bpf_program *prog;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
const void *obj_bytes;
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
int err, i, nr_progs;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
size_t obj_byte_cnt;
if (tester_init(tester) < 0)
return; /* failed to initialize tester */
obj_bytes = elf_bytes_factory(&obj_byte_cnt);
obj = bpf_object__open_mem(obj_bytes, obj_byte_cnt, &open_opts);
if (!ASSERT_OK_PTR(obj, "obj_open_mem"))
return;
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
nr_progs = 0;
bpf_object__for_each_program(prog, obj)
++nr_progs;
specs = calloc(nr_progs, sizeof(struct test_spec));
selftests/bpf: fix test_loader check message Seeing: process_subtest:PASS:Can't alloc specs array 0 nsec ... in verbose successful test log is very confusing. Use smaller identifier-like test tag to denote that we are asserting specs array allocation success. Now it's much less distracting: process_subtest:PASS:specs_alloc 0 nsec Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Acked-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20240105000909.2818934-2-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2024-01-05 00:09:02 +00:00
if (!ASSERT_OK_PTR(specs, "specs_alloc"))
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
return;
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
i = 0;
bpf_object__for_each_program(prog, obj) {
/* ignore tests for which we can't derive test specification */
err = parse_test_spec(tester, obj, prog, &specs[i++]);
if (err)
selftests/bpf: Report program name on parse_test_spec error Change test_loader.c:run_subtest() behavior to show BPF program name when test spec for that program can't be parsed. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:42 +00:00
PRINT_FAIL("Can't parse test spec for program '%s'\n",
bpf_program__name(prog));
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
}
i = 0;
bpf_object__for_each_program(prog, obj) {
struct test_spec *spec = &specs[i++];
if (!spec->valid || spec->auxiliary)
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
continue;
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
if (spec->mode_mask & PRIV)
run_subtest(tester, &open_opts, obj_bytes, obj_byte_cnt,
specs, spec, false);
if (spec->mode_mask & UNPRIV)
run_subtest(tester, &open_opts, obj_bytes, obj_byte_cnt,
specs, spec, true);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
}
selftests/bpf: Add notion of auxiliary programs for test_loader In order to express test cases that use bpf_tail_call() intrinsic it is necessary to have several programs to be loaded at a time. This commit adds __auxiliary annotation to the set of annotations supported by test_loader.c. Programs marked as auxiliary are always loaded but are not treated as a separate test. For example: void dummy_prog1(void); struct { __uint(type, BPF_MAP_TYPE_PROG_ARRAY); __uint(max_entries, 4); __uint(key_size, sizeof(int)); __array(values, void (void)); } prog_map SEC(".maps") = { .values = { [0] = (void *) &dummy_prog1, }, }; SEC("tc") __auxiliary __naked void dummy_prog1(void) { asm volatile ("r0 = 42; exit;"); } SEC("tc") __description("reference tracking: check reference or tail call") __success __retval(0) __naked void check_reference_or_tail_call(void) { asm volatile ( "r2 = %[prog_map] ll;" "r3 = 0;" "call %[bpf_tail_call];" "r0 = 0;" "exit;" :: __imm(bpf_tail_call), : __clobber_all); } Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-21 17:42:11 +00:00
for (i = 0; i < nr_progs; ++i)
free_test_spec(&specs[i]);
free(specs);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
bpf_object__close(obj);
}
void test_loader__run_subtests(struct test_loader *tester,
const char *skel_name,
skel_elf_bytes_fn elf_bytes_factory)
{
/* see comment in run_subtest() for why we do this function nesting */
selftests/bpf: Unprivileged tests for test_loader.c Extends test_loader.c:test_loader__run_subtests() by allowing to execute tests in unprivileged mode, similar to test_verifier.c. Adds the following new attributes controlling test_loader behavior: __msg_unpriv __success_unpriv __failure_unpriv * If any of these attributes is present the test would be loaded in unprivileged mode. * If only "privileged" attributes are present the test would be loaded only in privileged mode. * If both "privileged" and "unprivileged" attributes are present the test would be loaded in both modes. * If test has to be executed in both modes, __msg(text) is specified and __msg_unpriv is not specified the behavior is the same as if __msg_unpriv(text) is specified. * For test filtering purposes the name of the program loaded in unprivileged mode is derived from the usual program name by adding `@unpriv' suffix. Also adds attribute '__description'. This attribute specifies text to be used instead of a program name for display and filtering purposes. Signed-off-by: Eduard Zingerman <eddyz87@gmail.com> Link: https://lore.kernel.org/r/20230325025524.144043-4-eddyz87@gmail.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 02:54:44 +00:00
process_subtest(tester, skel_name, elf_bytes_factory);
selftests/bpf: add generic BPF program tester-loader It's become a common pattern to have a collection of small BPF programs in one BPF object file, each representing one test case. On user-space side of such tests we maintain a table of program names and expected failure or success, along with optional expected verifier log message. This works, but each set of tests reimplement this mundane code over and over again, which is a waste of time for anyone trying to add a new set of tests. Furthermore, it's quite error prone as it's way too easy to miss some entries in these manually maintained test tables (as evidences by dynptr_fail tests, in which ringbuf_release_uninit_dynptr subtest was accidentally missed; this is fixed in next patch). So this patch implements generic test_loader, which accepts skeleton name and handles the rest of details: opens and loads BPF object file, making sure each program is tested in isolation. Optionally each test case can specify expected BPF verifier log message. In case of failure, tester makes sure to report verifier log, but it also reports verifier log in verbose mode unconditionally. Now, the interesting deviation from existing custom implementations is the use of btf_decl_tag attribute to specify expected-to-fail vs expected-to-succeed markers and, optionally, expected log message directly next to BPF program source code, eliminating the need to manually create and update table of tests. We define few macros wrapping btf_decl_tag with a convention that all values of btf_decl_tag start with "comment:" prefix, and then utilizing a very simple "just_some_text_tag" or "some_key_name=<value>" pattern to define things like expected success/failure, expected verifier message, extra verifier log level (if necessary). This approach is demonstrated by next patch in which two existing sets of failure tests are converted. Tester supports both expected-to-fail and expected-to-succeed programs, though this patch set didn't convert any existing expected-to-succeed programs yet, as existing tests couple BPF program loading with their further execution through attach or test_prog_run. One way to allow testing scenarios like this would be ability to specify custom callback, executed for each successfully loaded BPF program. This is left for follow up patches, after some more analysis of existing test cases. This test_loader is, hopefully, a start of a test_verifier-like runner, but integrated into test_progs infrastructure. It will allow much better "user experience" of defining low-level verification tests that can take advantage of all the libbpf-provided nicety features on BPF side: global variables, declarative maps, etc. All while having a choice of defining it in C or as BPF assembly (through __attribute__((naked)) functions and using embedded asm), depending on what makes most sense in each particular case. This will be explored in follow up patches as well. Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/r/20221207201648.2990661-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-12-07 20:16:47 +00:00
}