From 1967939462641d8b36bcb3fcf06d48e66cd67a4f Mon Sep 17 00:00:00 2001 From: Masahiro Yamada Date: Sun, 29 Nov 2020 04:33:35 +0900 Subject: Compiler Attributes: remove CONFIG_ENABLE_MUST_CHECK Revert commit cebc04ba9aeb ("add CONFIG_ENABLE_MUST_CHECK"). A lot of warn_unused_result warnings existed in 2006, but until now they have been fixed thanks to people doing allmodconfig tests. Our goal is to always enable __must_check where appropriate, so this CONFIG option is no longer needed. I see a lot of defconfig (arch/*/configs/*_defconfig) files having: # CONFIG_ENABLE_MUST_CHECK is not set I did not touch them for now since it would be a big churn. If arch maintainers want to clean them up, please go ahead. While I was here, I also moved __must_check to compiler_attributes.h from compiler_types.h Signed-off-by: Masahiro Yamada Acked-by: Jason A. Donenfeld Acked-by: Nathan Chancellor Reviewed-by: Nick Desaulniers [Moved addition in compiler_attributes.h to keep it sorted] Signed-off-by: Miguel Ojeda --- tools/testing/selftests/wireguard/qemu/debug.config | 1 - 1 file changed, 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/wireguard/qemu/debug.config b/tools/testing/selftests/wireguard/qemu/debug.config index b50c2085c1ac..fe07d97df9fa 100644 --- a/tools/testing/selftests/wireguard/qemu/debug.config +++ b/tools/testing/selftests/wireguard/qemu/debug.config @@ -1,5 +1,4 @@ CONFIG_LOCALVERSION="-debug" -CONFIG_ENABLE_MUST_CHECK=y CONFIG_FRAME_POINTER=y CONFIG_STACK_VALIDATION=y CONFIG_DEBUG_KERNEL=y -- cgit v1.2.3-59-g8ed1b From 1a3449c19407a28f7019a887cdf0d6ba2444751a Mon Sep 17 00:00:00 2001 From: Kamal Mostafa Date: Tue, 15 Dec 2020 10:20:10 -0800 Subject: selftests/bpf: Clarify build error if no vmlinux If Makefile cannot find any of the vmlinux's in its VMLINUX_BTF_PATHS list, it tries to run btftool incorrectly, with VMLINUX_BTF unset: bpftool btf dump file $(VMLINUX_BTF) format c Such that the keyword 'format' is misinterpreted as the path to vmlinux. The resulting build error message is fairly cryptic: GEN vmlinux.h Error: failed to load BTF from format: No such file or directory This patch makes the failure reason clearer by yielding this instead: Makefile:...: *** Cannot find a vmlinux for VMLINUX_BTF at any of "{paths}". Stop. Fixes: acbd06206bbb ("selftests/bpf: Add vmlinux.h selftest exercising tracing of syscalls") Signed-off-by: Kamal Mostafa Signed-off-by: Andrii Nakryiko Link: https://lore.kernel.org/bpf/20201215182011.15755-1-kamal@canonical.com --- tools/testing/selftests/bpf/Makefile | 3 +++ 1 file changed, 3 insertions(+) (limited to 'tools') diff --git a/tools/testing/selftests/bpf/Makefile b/tools/testing/selftests/bpf/Makefile index 8c33e999319a..c51df6b91bef 100644 --- a/tools/testing/selftests/bpf/Makefile +++ b/tools/testing/selftests/bpf/Makefile @@ -121,6 +121,9 @@ VMLINUX_BTF_PATHS ?= $(if $(O),$(O)/vmlinux) \ /sys/kernel/btf/vmlinux \ /boot/vmlinux-$(shell uname -r) VMLINUX_BTF ?= $(abspath $(firstword $(wildcard $(VMLINUX_BTF_PATHS)))) +ifeq ($(VMLINUX_BTF),) +$(error Cannot find a vmlinux for VMLINUX_BTF at any of "$(VMLINUX_BTF_PATHS)") +endif # Define simple and short `make test_progs`, `make test_sysctl`, etc targets # to build individual tests. -- cgit v1.2.3-59-g8ed1b From e79bb299ccad6983876686a4d8c87c92ebbe5657 Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Mon, 14 Dec 2020 22:35:39 +0000 Subject: selftests/bpf: Fix spelling mistake "tranmission" -> "transmission" There are two spelling mistakes in output messages. Fix these. Signed-off-by: Colin Ian King Signed-off-by: Daniel Borkmann Link: https://lore.kernel.org/bpf/20201214223539.83168-1-colin.king@canonical.com --- tools/testing/selftests/bpf/xdpxceiver.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/bpf/xdpxceiver.c b/tools/testing/selftests/bpf/xdpxceiver.c index 014dedaa4dd2..1e722ee76b1f 100644 --- a/tools/testing/selftests/bpf/xdpxceiver.c +++ b/tools/testing/selftests/bpf/xdpxceiver.c @@ -715,7 +715,7 @@ static void worker_pkt_dump(void) int payload = *((uint32_t *)(pkt_buf[iter]->payload + PKT_HDR_SIZE)); if (payload == EOT) { - ksft_print_msg("End-of-tranmission frame received\n"); + ksft_print_msg("End-of-transmission frame received\n"); fprintf(stdout, "---------------------------------------\n"); break; } @@ -747,7 +747,7 @@ static void worker_pkt_validate(void) } if (payloadseqnum == EOT) { - ksft_print_msg("End-of-tranmission frame received: PASS\n"); + ksft_print_msg("End-of-transmission frame received: PASS\n"); sigvar = 1; break; } -- cgit v1.2.3-59-g8ed1b From 11b844b0b7c7c3dc8e8f4d0bbaad5e798351862c Mon Sep 17 00:00:00 2001 From: Andrii Nakryiko Date: Wed, 23 Dec 2020 12:06:52 -0800 Subject: selftests/bpf: Work-around EBUSY errors from hashmap update/delete 20b6cc34ea74 ("bpf: Avoid hashtab deadlock with map_locked") introduced a possibility of getting EBUSY error on lock contention, which seems to happen very deterministically in test_maps when running 1024 threads on low-CPU machine. In libbpf CI case, it's a 2 CPU VM and it's hitting this 100% of the time. Work around by retrying on EBUSY (and EAGAIN, while we are at it) after a small sleep. sched_yield() is too agressive and fails even after 20 retries, so I went with usleep(1) for backoff. Also log actual error returned to make it easier to see what's going on. Fixes: 20b6cc34ea74 ("bpf: Avoid hashtab deadlock with map_locked") Signed-off-by: Andrii Nakryiko Signed-off-by: Daniel Borkmann Acked-by: Song Liu Link: https://lore.kernel.org/bpf/20201223200652.3417075-1-andrii@kernel.org --- tools/testing/selftests/bpf/test_maps.c | 48 ++++++++++++++++++++++++++++----- 1 file changed, 42 insertions(+), 6 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/bpf/test_maps.c b/tools/testing/selftests/bpf/test_maps.c index 0ad3e6305ff0..51adc42b2b40 100644 --- a/tools/testing/selftests/bpf/test_maps.c +++ b/tools/testing/selftests/bpf/test_maps.c @@ -1312,22 +1312,58 @@ static void test_map_stress(void) #define DO_UPDATE 1 #define DO_DELETE 0 +#define MAP_RETRIES 20 + +static int map_update_retriable(int map_fd, const void *key, const void *value, + int flags, int attempts) +{ + while (bpf_map_update_elem(map_fd, key, value, flags)) { + if (!attempts || (errno != EAGAIN && errno != EBUSY)) + return -errno; + + usleep(1); + attempts--; + } + + return 0; +} + +static int map_delete_retriable(int map_fd, const void *key, int attempts) +{ + while (bpf_map_delete_elem(map_fd, key)) { + if (!attempts || (errno != EAGAIN && errno != EBUSY)) + return -errno; + + usleep(1); + attempts--; + } + + return 0; +} + static void test_update_delete(unsigned int fn, void *data) { int do_update = ((int *)data)[1]; int fd = ((int *)data)[0]; - int i, key, value; + int i, key, value, err; for (i = fn; i < MAP_SIZE; i += TASKS) { key = value = i; if (do_update) { - assert(bpf_map_update_elem(fd, &key, &value, - BPF_NOEXIST) == 0); - assert(bpf_map_update_elem(fd, &key, &value, - BPF_EXIST) == 0); + err = map_update_retriable(fd, &key, &value, BPF_NOEXIST, MAP_RETRIES); + if (err) + printf("error %d %d\n", err, errno); + assert(err == 0); + err = map_update_retriable(fd, &key, &value, BPF_EXIST, MAP_RETRIES); + if (err) + printf("error %d %d\n", err, errno); + assert(err == 0); } else { - assert(bpf_map_delete_elem(fd, &key) == 0); + err = map_delete_retriable(fd, &key, MAP_RETRIES); + if (err) + printf("error %d %d\n", err, errno); + assert(err == 0); } } } -- cgit v1.2.3-59-g8ed1b From da4282c17d695b9311608aa63b3c633e649aadea Mon Sep 17 00:00:00 2001 From: Jiang Wang Date: Thu, 24 Dec 2020 01:12:42 +0000 Subject: selftests/bpf: Fix a compile error for BPF_F_BPRM_SECUREEXEC When CONFIG_BPF_LSM is not configured, running bpf selftesting will show BPF_F_BPRM_SECUREEXEC undefined error for bprm_opts.c. The problem is that bprm_opts.c includes vmliunx.h. The vmlinux.h is generated by "bpftool btf dump file ./vmlinux format c". On the other hand, BPF_F_BPRM_SECUREEXEC is defined in include/uapi/linux/bpf.h and used only in bpf_lsm.c. When CONFIG_BPF_LSM is not set, bpf_lsm will not be compiled, so vmlinux.h will not include definition of BPF_F_BPRM_SECUREEXEC. Ideally, we want to compile bpf selftest regardless of the configuration setting, so change the include file from vmlinux.h to bpf.h. Signed-off-by: Jiang Wang Signed-off-by: Daniel Borkmann Acked-by: Song Liu Acked-by: John Fastabend Link: https://lore.kernel.org/bpf/20201224011242.585967-1-jiang.wang@bytedance.com --- tools/testing/selftests/bpf/progs/bprm_opts.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/bpf/progs/bprm_opts.c b/tools/testing/selftests/bpf/progs/bprm_opts.c index 5bfef2887e70..418d9c6d4952 100644 --- a/tools/testing/selftests/bpf/progs/bprm_opts.c +++ b/tools/testing/selftests/bpf/progs/bprm_opts.c @@ -4,7 +4,7 @@ * Copyright 2020 Google LLC. */ -#include "vmlinux.h" +#include #include #include #include -- cgit v1.2.3-59-g8ed1b From 7cf22a1c88c05ea3807f95b1edfebb729016ae52 Mon Sep 17 00:00:00 2001 From: Harish Date: Tue, 29 Dec 2020 15:14:22 -0800 Subject: selftests/vm: fix building protection keys test Commit d8cbe8bfa7d ("tools/testing/selftests/vm: fix build error") tried to include a ARCH check for powerpc, however ARCH is not defined in the Makefile before including lib.mk. This makes test building to skip on both x86 and powerpc. Fix the arch check by replacing it using machine type as it is already defined and used in the test. Link: https://lkml.kernel.org/r/20201215100402.257376-1-harish@linux.ibm.com Fixes: d8cbe8bfa7d ("tools/testing/selftests/vm: fix build error") Signed-off-by: Harish Reviewed-by: Sandipan Das Cc: Shuah Khan Cc: Sandipan Das Cc: John Hubbard Cc: Dave Hansen Cc: Kirill A. Shutemov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- tools/testing/selftests/vm/Makefile | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/vm/Makefile b/tools/testing/selftests/vm/Makefile index 9a25307f6115..d42115e4284d 100644 --- a/tools/testing/selftests/vm/Makefile +++ b/tools/testing/selftests/vm/Makefile @@ -4,7 +4,7 @@ include local_config.mk uname_M := $(shell uname -m 2>/dev/null || echo not) -MACHINE ?= $(shell echo $(uname_M) | sed -e 's/aarch64.*/arm64/') +MACHINE ?= $(shell echo $(uname_M) | sed -e 's/aarch64.*/arm64/' -e 's/ppc64.*/ppc64/') # Without this, failed build products remain, with up-to-date timestamps, # thus tricking Make (and you!) into believing that All Is Well, in subsequent @@ -43,7 +43,7 @@ TEST_GEN_FILES += thuge-gen TEST_GEN_FILES += transhuge-stress TEST_GEN_FILES += userfaultfd -ifeq ($(ARCH),x86_64) +ifeq ($(MACHINE),x86_64) CAN_BUILD_I386 := $(shell ./../x86/check_cc.sh $(CC) ../x86/trivial_32bit_program.c -m32) CAN_BUILD_X86_64 := $(shell ./../x86/check_cc.sh $(CC) ../x86/trivial_64bit_program.c) CAN_BUILD_WITH_NOPIE := $(shell ./../x86/check_cc.sh $(CC) ../x86/trivial_program.c -no-pie) @@ -65,13 +65,13 @@ TEST_GEN_FILES += $(BINARIES_64) endif else -ifneq (,$(findstring $(ARCH),powerpc)) +ifneq (,$(findstring $(MACHINE),ppc64)) TEST_GEN_FILES += protection_keys endif endif -ifneq (,$(filter $(MACHINE),arm64 ia64 mips64 parisc64 ppc64 ppc64le riscv64 s390x sh64 sparc64 x86_64)) +ifneq (,$(filter $(MACHINE),arm64 ia64 mips64 parisc64 ppc64 riscv64 s390x sh64 sparc64 x86_64)) TEST_GEN_FILES += va_128TBswitch TEST_GEN_FILES += virtual_address_range TEST_GEN_FILES += write_to_hugetlbfs @@ -84,7 +84,7 @@ TEST_FILES := test_vmalloc.sh KSFT_KHDR_INSTALL := 1 include ../lib.mk -ifeq ($(ARCH),x86_64) +ifeq ($(MACHINE),x86_64) BINARIES_32 := $(patsubst %,$(OUTPUT)/%,$(BINARIES_32)) BINARIES_64 := $(patsubst %,$(OUTPUT)/%,$(BINARIES_64)) -- cgit v1.2.3-59-g8ed1b From 65a4e5299739abe0888cda0938d21f8ea3b5c606 Mon Sep 17 00:00:00 2001 From: David Gow Date: Mon, 21 Dec 2020 23:39:00 -0800 Subject: kunit: tool: Force the use of the 'tty' console for UML kunit_tool relies on the UML console outputting printk() output to the tty in order to get results. Since the default console driver could change, pass 'console=tty' to the kernel. This is triggered by a change[1] to use ttynull as a fallback console driver which -- by chance or by design -- seems to have changed the default console output on UML, breaking kunit_tool. While this may be fixed, we should be less fragile to such changes in the default. [1]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=757055ae8dedf5333af17b3b5b4b70ba9bc9da4e Signed-off-by: David Gow Fixes: 757055ae8ded ("init/console: Use ttynull as a fallback when there is no console") Reported-by: Andy Shevchenko Tested-by: Andy Shevchenko Acked-by: Brendan Higgins Signed-off-by: Shuah Khan --- tools/testing/kunit/kunit_kernel.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/kunit/kunit_kernel.py b/tools/testing/kunit/kunit_kernel.py index 57c1724b7e5d..698358c9c0d6 100644 --- a/tools/testing/kunit/kunit_kernel.py +++ b/tools/testing/kunit/kunit_kernel.py @@ -198,7 +198,7 @@ class LinuxSourceTree(object): return self.validate_config(build_dir) def run_kernel(self, args=[], build_dir='', timeout=None): - args.extend(['mem=1G']) + args.extend(['mem=1G', 'console=tty']) self._ops.linux_bin(args, timeout, build_dir) outfile = get_outfile_path(build_dir) subprocess.call(['stty', 'sane']) -- cgit v1.2.3-59-g8ed1b From 3b4cf848dad5dad4bf239ba664c809c8cf29f1ed Mon Sep 17 00:00:00 2001 From: Tobias Klauser Date: Thu, 17 Dec 2020 17:31:40 +0100 Subject: selftests/vDSO: add additional binaries to .gitignore Add the test binaries introduced by commit 693f5ca08ca0 ("kselftest: Extend vDSO selftest"), commit 03f55c7952c9 ("kselftest: Extend vDSO selftest to clock_getres") and commit c7e5789b24d3 ("kselftest: Move test_vdso to the vDSO test suite") to .gitignore. Signed-off-by: Tobias Klauser Signed-off-by: Shuah Khan --- tools/testing/selftests/vDSO/.gitignore | 3 +++ 1 file changed, 3 insertions(+) (limited to 'tools') diff --git a/tools/testing/selftests/vDSO/.gitignore b/tools/testing/selftests/vDSO/.gitignore index 5eb64d41e541..a8dc51af5a9c 100644 --- a/tools/testing/selftests/vDSO/.gitignore +++ b/tools/testing/selftests/vDSO/.gitignore @@ -1,5 +1,8 @@ # SPDX-License-Identifier: GPL-2.0-only vdso_test +vdso_test_abi +vdso_test_clock_getres +vdso_test_correctness vdso_test_gettimeofday vdso_test_getcpu vdso_standalone_test_x86 -- cgit v1.2.3-59-g8ed1b From df00d02989024d193a6efd1a85513a5658c6a10f Mon Sep 17 00:00:00 2001 From: Tobias Klauser Date: Thu, 17 Dec 2020 17:32:35 +0100 Subject: selftests/vDSO: fix -Wformat warning in vdso_test_correctness MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix the following -Wformat warnings in vdso_test_correctness.c: vdso_test_correctness.c: In function ‘test_one_clock_gettime64’: vdso_test_correctness.c:352:21: warning: format ‘%ld’ expects argument of type ‘long int’, but argument 3 has type ‘long long int’ [-Wformat=] 352 | printf("\t%llu.%09ld %llu.%09ld %llu.%09ld\n", | ~~~~^ | | | long int | %09lld 353 | (unsigned long long)start.tv_sec, start.tv_nsec, | ~~~~~~~~~~~~~ | | | long long int vdso_test_correctness.c:352:32: warning: format ‘%ld’ expects argument of type ‘long int’, but argument 5 has type ‘long long int’ [-Wformat=] 352 | printf("\t%llu.%09ld %llu.%09ld %llu.%09ld\n", | ~~~~^ | | | long int | %09lld 353 | (unsigned long long)start.tv_sec, start.tv_nsec, 354 | (unsigned long long)vdso.tv_sec, vdso.tv_nsec, | ~~~~~~~~~~~~ | | | long long int vdso_test_correctness.c:352:43: warning: format ‘%ld’ expects argument of type ‘long int’, but argument 7 has type ‘long long int’ [-Wformat=] The tv_sec member of __kernel_timespec is long long, both in uapi/linux/time_types.h and locally in vdso_test_correctness.c. Signed-off-by: Tobias Klauser Signed-off-by: Shuah Khan --- tools/testing/selftests/vDSO/vdso_test_correctness.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/vDSO/vdso_test_correctness.c b/tools/testing/selftests/vDSO/vdso_test_correctness.c index 5029ef9b228c..c4aea794725a 100644 --- a/tools/testing/selftests/vDSO/vdso_test_correctness.c +++ b/tools/testing/selftests/vDSO/vdso_test_correctness.c @@ -349,7 +349,7 @@ static void test_one_clock_gettime64(int clock, const char *name) return; } - printf("\t%llu.%09ld %llu.%09ld %llu.%09ld\n", + printf("\t%llu.%09lld %llu.%09lld %llu.%09lld\n", (unsigned long long)start.tv_sec, start.tv_nsec, (unsigned long long)vdso.tv_sec, vdso.tv_nsec, (unsigned long long)end.tv_sec, end.tv_nsec); -- cgit v1.2.3-59-g8ed1b From 7a6eb7c34a78498742b5f82543b7a68c1c443329 Mon Sep 17 00:00:00 2001 From: Mark Brown Date: Thu, 10 Dec 2020 18:52:33 +0000 Subject: selftests: Skip BPF seftests by default The BPF selftests have build time dependencies on cutting edge versions of tools in the BPF ecosystem including LLVM which are more involved to satisfy than more typical requirements like installing a package from your distribution. This causes issues for users looking at kselftest in as a whole who find that a default build of kselftest fails and that resolving this is time consuming and adds administrative overhead. The fast pace of BPF development and the need for a full BPF stack to do substantial development or validation work on the code mean that people working directly on it don't see a reasonable way to keep supporting older environments without causing problems with the usability of the BPF tests in BPF development so these requirements are unlikely to be relaxed in the immediate future. There is already support for skipping targets so in order to reduce the barrier to entry for people interested in kselftest as a whole let's use that to skip the BPF tests by default when people work with the top level kselftest build system. Users can still build the BPF selftests as part of the wider kselftest build by specifying SKIP_TARGETS, including setting an empty SKIP_TARGETS to build everything. They can also continue to build the BPF selftests individually in cases where they are specifically focused on BPF. This isn't ideal since it means people will need to take special steps to build the BPF tests but the dependencies mean that realistically this is already the case to some extent and it makes it easier for people to pick up and work with the other selftests which is hopefully a net win. Signed-off-by: Mark Brown Signed-off-by: Shuah Khan --- tools/testing/selftests/Makefile | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/Makefile b/tools/testing/selftests/Makefile index afbab4aeef3c..8a917cb4426a 100644 --- a/tools/testing/selftests/Makefile +++ b/tools/testing/selftests/Makefile @@ -77,8 +77,10 @@ TARGETS += zram TARGETS_HOTPLUG = cpu-hotplug TARGETS_HOTPLUG += memory-hotplug -# User can optionally provide a TARGETS skiplist. -SKIP_TARGETS ?= +# User can optionally provide a TARGETS skiplist. By default we skip +# BPF since it has cutting edge build time dependencies which require +# more effort to install. +SKIP_TARGETS ?= bpf ifneq ($(SKIP_TARGETS),) TMP := $(filter-out $(SKIP_TARGETS), $(TARGETS)) override TARGETS := $(TMP) -- cgit v1.2.3-59-g8ed1b From 2ff2c7e274392871bfdee00ff2adbb8ebae5d240 Mon Sep 17 00:00:00 2001 From: Ido Schimmel Date: Wed, 30 Dec 2020 13:42:51 +0200 Subject: selftests: mlxsw: Set headroom size of correct port The test was setting the headroom size of the wrong port. This was not visible because of a firmware bug that canceled this bug. Set the headroom size of the correct port, so that the test will pass with both old and new firmware versions. Fixes: bfa804784e32 ("selftests: mlxsw: Add a PFC test") Signed-off-by: Ido Schimmel Reviewed-by: Petr Machata Link: https://lore.kernel.org/r/20201230114251.394009-1-idosch@idosch.org Signed-off-by: Jakub Kicinski --- tools/testing/selftests/drivers/net/mlxsw/qos_pfc.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/drivers/net/mlxsw/qos_pfc.sh b/tools/testing/selftests/drivers/net/mlxsw/qos_pfc.sh index 4d900bc1f76c..5c7700212f75 100755 --- a/tools/testing/selftests/drivers/net/mlxsw/qos_pfc.sh +++ b/tools/testing/selftests/drivers/net/mlxsw/qos_pfc.sh @@ -230,7 +230,7 @@ switch_create() __mlnx_qos -i $swp4 --pfc=0,1,0,0,0,0,0,0 >/dev/null # PG0 will get autoconfigured to Xoff, give PG1 arbitrarily 100K, which # is (-2*MTU) about 80K of delay provision. - __mlnx_qos -i $swp3 --buffer_size=0,$_100KB,0,0,0,0,0,0 >/dev/null + __mlnx_qos -i $swp4 --buffer_size=0,$_100KB,0,0,0,0,0,0 >/dev/null # bridges # ------- -- cgit v1.2.3-59-g8ed1b From 3503ee6c0bec5f173d606359e6384a5ef85492fb Mon Sep 17 00:00:00 2001 From: Po-Hsu Lin Date: Tue, 5 Jan 2021 18:17:40 +0800 Subject: selftests: fix the return value for UDP GRO test The udpgro.sh will always return 0 (unless the bpf selftest was not build first) even if there are some failed sub test-cases. Therefore the kselftest framework will report this case is OK. Check and return the exit status of each test to make it easier to spot real failures. Signed-off-by: Po-Hsu Lin Signed-off-by: David S. Miller --- tools/testing/selftests/net/udpgro.sh | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+) (limited to 'tools') diff --git a/tools/testing/selftests/net/udpgro.sh b/tools/testing/selftests/net/udpgro.sh index ac2a30be9b32..f8a19f548ae9 100755 --- a/tools/testing/selftests/net/udpgro.sh +++ b/tools/testing/selftests/net/udpgro.sh @@ -5,6 +5,14 @@ readonly PEER_NS="ns-peer-$(mktemp -u XXXXXX)" +# set global exit status, but never reset nonzero one. +check_err() +{ + if [ $ret -eq 0 ]; then + ret=$1 + fi +} + cleanup() { local -r jobs="$(jobs -p)" local -r ns="$(ip netns list|grep $PEER_NS)" @@ -44,7 +52,9 @@ run_one() { # Hack: let bg programs complete the startup sleep 0.1 ./udpgso_bench_tx ${tx_args} + ret=$? wait $(jobs -p) + return $ret } run_test() { @@ -87,8 +97,10 @@ run_one_nat() { sleep 0.1 ./udpgso_bench_tx ${tx_args} + ret=$? kill -INT $pid wait $(jobs -p) + return $ret } run_one_2sock() { @@ -110,7 +122,9 @@ run_one_2sock() { sleep 0.1 # first UDP GSO socket should be closed at this point ./udpgso_bench_tx ${tx_args} + ret=$? wait $(jobs -p) + return $ret } run_nat_test() { @@ -131,36 +145,54 @@ run_all() { local -r core_args="-l 4" local -r ipv4_args="${core_args} -4 -D 192.168.1.1" local -r ipv6_args="${core_args} -6 -D 2001:db8::1" + ret=0 echo "ipv4" run_test "no GRO" "${ipv4_args} -M 10 -s 1400" "-4 -n 10 -l 1400" + check_err $? # explicitly check we are not receiving UDP_SEGMENT cmsg (-S -1) # when GRO does not take place run_test "no GRO chk cmsg" "${ipv4_args} -M 10 -s 1400" "-4 -n 10 -l 1400 -S -1" + check_err $? # the GSO packets are aggregated because: # * veth schedule napi after each xmit # * segmentation happens in BH context, veth napi poll is delayed after # the transmission of the last segment run_test "GRO" "${ipv4_args} -M 1 -s 14720 -S 0 " "-4 -n 1 -l 14720" + check_err $? run_test "GRO chk cmsg" "${ipv4_args} -M 1 -s 14720 -S 0 " "-4 -n 1 -l 14720 -S 1472" + check_err $? run_test "GRO with custom segment size" "${ipv4_args} -M 1 -s 14720 -S 500 " "-4 -n 1 -l 14720" + check_err $? run_test "GRO with custom segment size cmsg" "${ipv4_args} -M 1 -s 14720 -S 500 " "-4 -n 1 -l 14720 -S 500" + check_err $? run_nat_test "bad GRO lookup" "${ipv4_args} -M 1 -s 14720 -S 0" "-n 10 -l 1472" + check_err $? run_2sock_test "multiple GRO socks" "${ipv4_args} -M 1 -s 14720 -S 0 " "-4 -n 1 -l 14720 -S 1472" + check_err $? echo "ipv6" run_test "no GRO" "${ipv6_args} -M 10 -s 1400" "-n 10 -l 1400" + check_err $? run_test "no GRO chk cmsg" "${ipv6_args} -M 10 -s 1400" "-n 10 -l 1400 -S -1" + check_err $? run_test "GRO" "${ipv6_args} -M 1 -s 14520 -S 0" "-n 1 -l 14520" + check_err $? run_test "GRO chk cmsg" "${ipv6_args} -M 1 -s 14520 -S 0" "-n 1 -l 14520 -S 1452" + check_err $? run_test "GRO with custom segment size" "${ipv6_args} -M 1 -s 14520 -S 500" "-n 1 -l 14520" + check_err $? run_test "GRO with custom segment size cmsg" "${ipv6_args} -M 1 -s 14520 -S 500" "-n 1 -l 14520 -S 500" + check_err $? run_nat_test "bad GRO lookup" "${ipv6_args} -M 1 -s 14520 -S 0" "-n 10 -l 1452" + check_err $? run_2sock_test "multiple GRO socks" "${ipv6_args} -M 1 -s 14520 -S 0 " "-n 1 -l 14520 -S 1452" + check_err $? + return $ret } if [ ! -f ../bpf/xdp_dummy.o ]; then @@ -180,3 +212,5 @@ elif [[ $1 == "__subprocess_2sock" ]]; then shift run_one_2sock $@ fi + +exit $? -- cgit v1.2.3-59-g8ed1b From 67208692802ce3cacfa00fe586dc0cb1bef0a51c Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Wed, 6 Jan 2021 00:42:19 +0100 Subject: tools/resolve_btfids: Warn when having multiple IDs for single type The kernel image can contain multiple types (structs/unions) with the same name. This causes distinct type hierarchies in BTF data and makes resolve_btfids fail with error like: BTFIDS vmlinux FAILED unresolved symbol udp6_sock as reported by Qais Yousef [1]. This change adds warning when multiple types of the same name are detected: BTFIDS vmlinux WARN: multiple IDs found for 'file': 526, 113351 - using 526 WARN: multiple IDs found for 'sk_buff': 2744, 113958 - using 2744 We keep the lower ID for the given type instance and let the build continue. Also changing the 'nr' variable name to 'nr_types' to avoid confusion. [1] https://lore.kernel.org/lkml/20201229151352.6hzmjvu3qh6p2qgg@e107158-lin/ Signed-off-by: Jiri Olsa Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Link: https://lore.kernel.org/bpf/20210105234219.970039-1-jolsa@kernel.org --- tools/bpf/resolve_btfids/main.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) (limited to 'tools') diff --git a/tools/bpf/resolve_btfids/main.c b/tools/bpf/resolve_btfids/main.c index e3ea569ee125..7409d7860aa6 100644 --- a/tools/bpf/resolve_btfids/main.c +++ b/tools/bpf/resolve_btfids/main.c @@ -139,6 +139,8 @@ int eprintf(int level, int var, const char *fmt, ...) #define pr_debug2(fmt, ...) pr_debugN(2, pr_fmt(fmt), ##__VA_ARGS__) #define pr_err(fmt, ...) \ eprintf(0, verbose, pr_fmt(fmt), ##__VA_ARGS__) +#define pr_info(fmt, ...) \ + eprintf(0, verbose, pr_fmt(fmt), ##__VA_ARGS__) static bool is_btf_id(const char *name) { @@ -472,7 +474,7 @@ static int symbols_resolve(struct object *obj) int nr_funcs = obj->nr_funcs; int err, type_id; struct btf *btf; - __u32 nr; + __u32 nr_types; btf = btf__parse(obj->btf ?: obj->path, NULL); err = libbpf_get_error(btf); @@ -483,12 +485,12 @@ static int symbols_resolve(struct object *obj) } err = -1; - nr = btf__get_nr_types(btf); + nr_types = btf__get_nr_types(btf); /* * Iterate all the BTF types and search for collected symbol IDs. */ - for (type_id = 1; type_id <= nr; type_id++) { + for (type_id = 1; type_id <= nr_types; type_id++) { const struct btf_type *type; struct rb_root *root; struct btf_id *id; @@ -526,8 +528,13 @@ static int symbols_resolve(struct object *obj) id = btf_id__find(root, str); if (id) { - id->id = type_id; - (*nr)--; + if (id->id) { + pr_info("WARN: multiple IDs found for '%s': %d, %d - using %d\n", + str, id->id, type_id, id->id); + } else { + id->id = type_id; + (*nr)--; + } } } -- cgit v1.2.3-59-g8ed1b From 6f02b540d7597f357bc6ee711346761045d4e108 Mon Sep 17 00:00:00 2001 From: Alan Maguire Date: Wed, 6 Jan 2021 15:59:06 +0000 Subject: bpftool: Fix compilation failure for net.o with older glibc MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit For older glibc ~2.17, #include'ing both linux/if.h and net/if.h fails due to complaints about redefinition of interface flags: CC net.o In file included from net.c:13:0: /usr/include/linux/if.h:71:2: error: redeclaration of enumerator ‘IFF_UP’ IFF_UP = 1<<0, /* sysfs */ ^ /usr/include/net/if.h:44:5: note: previous definition of ‘IFF_UP’ was here IFF_UP = 0x1, /* Interface is up. */ The issue was fixed in kernel headers in [1], but since compilation of net.c picks up system headers the problem can recur. Dropping #include resolves the issue and it is not needed for compilation anyhow. [1] https://lore.kernel.org/netdev/1461512707-23058-1-git-send-email-mikko.rapeli__34748.27880641$1462831734$gmane$org@iki.fi/ Fixes: f6f3bac08ff9 ("tools/bpf: bpftool: add net support") Signed-off-by: Alan Maguire Signed-off-by: Andrii Nakryiko Acked-by: Martin KaFai Lau Link: https://lore.kernel.org/bpf/1609948746-15369-1-git-send-email-alan.maguire@oracle.com --- tools/bpf/bpftool/net.c | 1 - 1 file changed, 1 deletion(-) (limited to 'tools') diff --git a/tools/bpf/bpftool/net.c b/tools/bpf/bpftool/net.c index 3fae61ef6339..ff3aa0cf3997 100644 --- a/tools/bpf/bpftool/net.c +++ b/tools/bpf/bpftool/net.c @@ -11,7 +11,6 @@ #include #include #include -#include #include #include #include -- cgit v1.2.3-59-g8ed1b From 5316a7c0130acf09bfc8bb0092407006010fcccc Mon Sep 17 00:00:00 2001 From: Sean Tranchetti Date: Tue, 5 Jan 2021 16:22:26 -0800 Subject: tools: selftests: add test for changing routes with PTMU exceptions Adds new 2 new tests to the PTMU script: pmtu_ipv4/6_route_change. These tests explicitly test for a recently discovered problem in the IPv6 routing framework where PMTU exceptions were not properly released when replacing a route via "ip route change ...". After creating PMTU exceptions, the route from the device A to R1 will be replaced with a new route, then device A will be deleted. If the PMTU exceptions were properly cleaned up by the kernel, this device deletion will succeed. Otherwise, the unregistration of the device will stall, and messages such as the following will be logged in dmesg: unregister_netdevice: waiting for veth_A-R1 to become free. Usage count = 4 Signed-off-by: Sean Tranchetti Reviewed-by: David Ahern Link: https://lore.kernel.org/r/1609892546-11389-2-git-send-email-stranche@quicinc.com Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/pmtu.sh | 71 +++++++++++++++++++++++++++++++++++-- 1 file changed, 69 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/net/pmtu.sh b/tools/testing/selftests/net/pmtu.sh index 464e31eabc73..64cd2e23c568 100755 --- a/tools/testing/selftests/net/pmtu.sh +++ b/tools/testing/selftests/net/pmtu.sh @@ -162,7 +162,15 @@ # - list_flush_ipv6_exception # Using the same topology as in pmtu_ipv6, create exceptions, and check # they are shown when listing exception caches, gone after flushing them - +# +# - pmtu_ipv4_route_change +# Use the same topology as in pmtu_ipv4, but issue a route replacement +# command and delete the corresponding device afterward. This tests for +# proper cleanup of the PMTU exceptions by the route replacement path. +# Device unregistration should complete successfully +# +# - pmtu_ipv6_route_change +# Same as above but with IPv6 # Kselftest framework requirement - SKIP code is 4. ksft_skip=4 @@ -224,7 +232,9 @@ tests=" cleanup_ipv4_exception ipv4: cleanup of cached exceptions 1 cleanup_ipv6_exception ipv6: cleanup of cached exceptions 1 list_flush_ipv4_exception ipv4: list and flush cached exceptions 1 - list_flush_ipv6_exception ipv6: list and flush cached exceptions 1" + list_flush_ipv6_exception ipv6: list and flush cached exceptions 1 + pmtu_ipv4_route_change ipv4: PMTU exception w/route replace 1 + pmtu_ipv6_route_change ipv6: PMTU exception w/route replace 1" NS_A="ns-A" NS_B="ns-B" @@ -1782,6 +1792,63 @@ test_list_flush_ipv6_exception() { return ${fail} } +test_pmtu_ipvX_route_change() { + family=${1} + + setup namespaces routing || return 2 + trace "${ns_a}" veth_A-R1 "${ns_r1}" veth_R1-A \ + "${ns_r1}" veth_R1-B "${ns_b}" veth_B-R1 \ + "${ns_a}" veth_A-R2 "${ns_r2}" veth_R2-A \ + "${ns_r2}" veth_R2-B "${ns_b}" veth_B-R2 + + if [ ${family} -eq 4 ]; then + ping=ping + dst1="${prefix4}.${b_r1}.1" + dst2="${prefix4}.${b_r2}.1" + gw="${prefix4}.${a_r1}.2" + else + ping=${ping6} + dst1="${prefix6}:${b_r1}::1" + dst2="${prefix6}:${b_r2}::1" + gw="${prefix6}:${a_r1}::2" + fi + + # Set up initial MTU values + mtu "${ns_a}" veth_A-R1 2000 + mtu "${ns_r1}" veth_R1-A 2000 + mtu "${ns_r1}" veth_R1-B 1400 + mtu "${ns_b}" veth_B-R1 1400 + + mtu "${ns_a}" veth_A-R2 2000 + mtu "${ns_r2}" veth_R2-A 2000 + mtu "${ns_r2}" veth_R2-B 1500 + mtu "${ns_b}" veth_B-R2 1500 + + # Create route exceptions + run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst1} + run_cmd ${ns_a} ${ping} -q -M want -i 0.1 -w 1 -s 1800 ${dst2} + + # Check that exceptions have been created with the correct PMTU + pmtu_1="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst1})" + check_pmtu_value "1400" "${pmtu_1}" "exceeding MTU" || return 1 + pmtu_2="$(route_get_dst_pmtu_from_exception "${ns_a}" ${dst2})" + check_pmtu_value "1500" "${pmtu_2}" "exceeding MTU" || return 1 + + # Replace the route from A to R1 + run_cmd ${ns_a} ip route change default via ${gw} + + # Delete the device in A + run_cmd ${ns_a} ip link del "veth_A-R1" +} + +test_pmtu_ipv4_route_change() { + test_pmtu_ipvX_route_change 4 +} + +test_pmtu_ipv6_route_change() { + test_pmtu_ipvX_route_change 6 +} + usage() { echo echo "$0 [OPTIONS] [TEST]..." -- cgit v1.2.3-59-g8ed1b From 9e7a67dee27902fedab880b9af909bd4acd0fba9 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 6 Jan 2021 00:15:21 +0100 Subject: selftests: netfilter: add selftest for ipip pmtu discovery with enabled connection tracking Convert Christians bug description into a reproducer. Cc: Shuah Khan Reported-by: Christian Perle Signed-off-by: Florian Westphal Acked-by: Pablo Neira Ayuso Signed-off-by: Jakub Kicinski --- tools/testing/selftests/netfilter/Makefile | 3 +- .../selftests/netfilter/ipip-conntrack-mtu.sh | 206 +++++++++++++++++++++ 2 files changed, 208 insertions(+), 1 deletion(-) create mode 100755 tools/testing/selftests/netfilter/ipip-conntrack-mtu.sh (limited to 'tools') diff --git a/tools/testing/selftests/netfilter/Makefile b/tools/testing/selftests/netfilter/Makefile index a374e10ef506..3006a8e5b41a 100644 --- a/tools/testing/selftests/netfilter/Makefile +++ b/tools/testing/selftests/netfilter/Makefile @@ -4,7 +4,8 @@ TEST_PROGS := nft_trans_stress.sh nft_nat.sh bridge_brouter.sh \ conntrack_icmp_related.sh nft_flowtable.sh ipvs.sh \ nft_concat_range.sh nft_conntrack_helper.sh \ - nft_queue.sh nft_meta.sh + nft_queue.sh nft_meta.sh \ + ipip-conntrack-mtu.sh LDLIBS = -lmnl TEST_GEN_FILES = nf-queue diff --git a/tools/testing/selftests/netfilter/ipip-conntrack-mtu.sh b/tools/testing/selftests/netfilter/ipip-conntrack-mtu.sh new file mode 100755 index 000000000000..4a6f5c3b3215 --- /dev/null +++ b/tools/testing/selftests/netfilter/ipip-conntrack-mtu.sh @@ -0,0 +1,206 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0 + +# Kselftest framework requirement - SKIP code is 4. +ksft_skip=4 + +# Conntrack needs to reassemble fragments in order to have complete +# packets for rule matching. Reassembly can lead to packet loss. + +# Consider the following setup: +# +--------+ +---------+ +--------+ +# |Router A|-------|Wanrouter|-------|Router B| +# | |.IPIP..| |..IPIP.| | +# +--------+ +---------+ +--------+ +# / mtu 1400 \ +# / \ +#+--------+ +--------+ +#|Client A| |Client B| +#| | | | +#+--------+ +--------+ + +# Router A and Router B use IPIP tunnel interfaces to tunnel traffic +# between Client A and Client B over WAN. Wanrouter has MTU 1400 set +# on its interfaces. + +rnd=$(mktemp -u XXXXXXXX) +rx=$(mktemp) + +r_a="ns-ra-$rnd" +r_b="ns-rb-$rnd" +r_w="ns-rw-$rnd" +c_a="ns-ca-$rnd" +c_b="ns-cb-$rnd" + +checktool (){ + if ! $1 > /dev/null 2>&1; then + echo "SKIP: Could not $2" + exit $ksft_skip + fi +} + +checktool "iptables --version" "run test without iptables" +checktool "ip -Version" "run test without ip tool" +checktool "which nc" "run test without nc (netcat)" +checktool "ip netns add ${r_a}" "create net namespace" + +for n in ${r_b} ${r_w} ${c_a} ${c_b};do + ip netns add ${n} +done + +cleanup() { + for n in ${r_a} ${r_b} ${r_w} ${c_a} ${c_b};do + ip netns del ${n} + done + rm -f ${rx} +} + +trap cleanup EXIT + +test_path() { + msg="$1" + + ip netns exec ${c_b} nc -n -w 3 -q 3 -u -l -p 5000 > ${rx} < /dev/null & + + sleep 1 + for i in 1 2 3; do + head -c1400 /dev/zero | tr "\000" "a" | ip netns exec ${c_a} nc -n -w 1 -u 192.168.20.2 5000 + done + + wait + + bytes=$(wc -c < ${rx}) + + if [ $bytes -eq 1400 ];then + echo "OK: PMTU $msg connection tracking" + else + echo "FAIL: PMTU $msg connection tracking: got $bytes, expected 1400" + exit 1 + fi +} + +# Detailed setup for Router A +# --------------------------- +# Interfaces: +# eth0: 10.2.2.1/24 +# eth1: 192.168.10.1/24 +# ipip0: No IP address, local 10.2.2.1 remote 10.4.4.1 +# Routes: +# 192.168.20.0/24 dev ipip0 (192.168.20.0/24 is subnet of Client B) +# 10.4.4.1 via 10.2.2.254 (Router B via Wanrouter) +# No iptables rules at all. + +ip link add veth0 netns ${r_a} type veth peer name veth0 netns ${r_w} +ip link add veth1 netns ${r_a} type veth peer name veth0 netns ${c_a} + +l_addr="10.2.2.1" +r_addr="10.4.4.1" +ip netns exec ${r_a} ip link add ipip0 type ipip local ${l_addr} remote ${r_addr} mode ipip || exit $ksft_skip + +for dev in lo veth0 veth1 ipip0; do + ip -net ${r_a} link set $dev up +done + +ip -net ${r_a} addr add 10.2.2.1/24 dev veth0 +ip -net ${r_a} addr add 192.168.10.1/24 dev veth1 + +ip -net ${r_a} route add 192.168.20.0/24 dev ipip0 +ip -net ${r_a} route add 10.4.4.0/24 via 10.2.2.254 + +ip netns exec ${r_a} sysctl -q net.ipv4.conf.all.forwarding=1 > /dev/null + +# Detailed setup for Router B +# --------------------------- +# Interfaces: +# eth0: 10.4.4.1/24 +# eth1: 192.168.20.1/24 +# ipip0: No IP address, local 10.4.4.1 remote 10.2.2.1 +# Routes: +# 192.168.10.0/24 dev ipip0 (192.168.10.0/24 is subnet of Client A) +# 10.2.2.1 via 10.4.4.254 (Router A via Wanrouter) +# No iptables rules at all. + +ip link add veth0 netns ${r_b} type veth peer name veth1 netns ${r_w} +ip link add veth1 netns ${r_b} type veth peer name veth0 netns ${c_b} + +l_addr="10.4.4.1" +r_addr="10.2.2.1" + +ip netns exec ${r_b} ip link add ipip0 type ipip local ${l_addr} remote ${r_addr} mode ipip || exit $ksft_skip + +for dev in lo veth0 veth1 ipip0; do + ip -net ${r_b} link set $dev up +done + +ip -net ${r_b} addr add 10.4.4.1/24 dev veth0 +ip -net ${r_b} addr add 192.168.20.1/24 dev veth1 + +ip -net ${r_b} route add 192.168.10.0/24 dev ipip0 +ip -net ${r_b} route add 10.2.2.0/24 via 10.4.4.254 +ip netns exec ${r_b} sysctl -q net.ipv4.conf.all.forwarding=1 > /dev/null + +# Client A +ip -net ${c_a} addr add 192.168.10.2/24 dev veth0 +ip -net ${c_a} link set dev lo up +ip -net ${c_a} link set dev veth0 up +ip -net ${c_a} route add default via 192.168.10.1 + +# Client A +ip -net ${c_b} addr add 192.168.20.2/24 dev veth0 +ip -net ${c_b} link set dev veth0 up +ip -net ${c_b} link set dev lo up +ip -net ${c_b} route add default via 192.168.20.1 + +# Wan +ip -net ${r_w} addr add 10.2.2.254/24 dev veth0 +ip -net ${r_w} addr add 10.4.4.254/24 dev veth1 + +ip -net ${r_w} link set dev lo up +ip -net ${r_w} link set dev veth0 up mtu 1400 +ip -net ${r_w} link set dev veth1 up mtu 1400 + +ip -net ${r_a} link set dev veth0 mtu 1400 +ip -net ${r_b} link set dev veth0 mtu 1400 + +ip netns exec ${r_w} sysctl -q net.ipv4.conf.all.forwarding=1 > /dev/null + +# Path MTU discovery +# ------------------ +# Running tracepath from Client A to Client B shows PMTU discovery is working +# as expected: +# +# clienta:~# tracepath 192.168.20.2 +# 1?: [LOCALHOST] pmtu 1500 +# 1: 192.168.10.1 0.867ms +# 1: 192.168.10.1 0.302ms +# 2: 192.168.10.1 0.312ms pmtu 1480 +# 2: no reply +# 3: 192.168.10.1 0.510ms pmtu 1380 +# 3: 192.168.20.2 2.320ms reached +# Resume: pmtu 1380 hops 3 back 3 + +# ip netns exec ${c_a} traceroute --mtu 192.168.20.2 + +# Router A has learned PMTU (1400) to Router B from Wanrouter. +# Client A has learned PMTU (1400 - IPIP overhead = 1380) to Client B +# from Router A. + +#Send large UDP packet +#--------------------- +#Now we send a 1400 bytes UDP packet from Client A to Client B: + +# clienta:~# head -c1400 /dev/zero | tr "\000" "a" | nc -u 192.168.20.2 5000 +test_path "without" + +# The IPv4 stack on Client A already knows the PMTU to Client B, so the +# UDP packet is sent as two fragments (1380 + 20). Router A forwards the +# fragments between eth1 and ipip0. The fragments fit into the tunnel and +# reach their destination. + +#When sending the large UDP packet again, Router A now reassembles the +#fragments before routing the packet over ipip0. The resulting IPIP +#packet is too big (1400) for the tunnel PMTU (1380) to Router B, it is +#dropped on Router A before sending. + +ip netns exec ${r_a} iptables -A FORWARD -m conntrack --ctstate NEW +test_path "with" -- cgit v1.2.3-59-g8ed1b From e42ac777d661e878c3b9bac56df11e226cab3010 Mon Sep 17 00:00:00 2001 From: Andrew Jones Date: Fri, 18 Dec 2020 15:17:32 +0100 Subject: KVM: selftests: Factor out guest mode code demand_paging_test, dirty_log_test, and dirty_log_perf_test have redundant guest mode code. Factor it out. Also, while adding a new include, remove the ones we don't need. Reviewed-by: Ben Gardon Reviewed-by: Peter Xu Signed-off-by: Andrew Jones Message-Id: <20201218141734.54359-2-drjones@redhat.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/Makefile | 2 +- tools/testing/selftests/kvm/demand_paging_test.c | 107 +++++------------- tools/testing/selftests/kvm/dirty_log_perf_test.c | 121 ++++++--------------- tools/testing/selftests/kvm/dirty_log_test.c | 125 +++++++--------------- tools/testing/selftests/kvm/include/guest_modes.h | 21 ++++ tools/testing/selftests/kvm/lib/guest_modes.c | 70 ++++++++++++ 6 files changed, 189 insertions(+), 257 deletions(-) create mode 100644 tools/testing/selftests/kvm/include/guest_modes.h create mode 100644 tools/testing/selftests/kvm/lib/guest_modes.c (limited to 'tools') diff --git a/tools/testing/selftests/kvm/Makefile b/tools/testing/selftests/kvm/Makefile index c7ca4faba272..a7286a08c3ae 100644 --- a/tools/testing/selftests/kvm/Makefile +++ b/tools/testing/selftests/kvm/Makefile @@ -33,7 +33,7 @@ ifeq ($(ARCH),s390) UNAME_M := s390x endif -LIBKVM = lib/assert.c lib/elf.c lib/io.c lib/kvm_util.c lib/sparsebit.c lib/test_util.c +LIBKVM = lib/assert.c lib/elf.c lib/io.c lib/kvm_util.c lib/sparsebit.c lib/test_util.c lib/guest_modes.c LIBKVM_x86_64 = lib/x86_64/processor.c lib/x86_64/vmx.c lib/x86_64/svm.c lib/x86_64/ucall.c lib/x86_64/handlers.S LIBKVM_aarch64 = lib/aarch64/processor.c lib/aarch64/ucall.c LIBKVM_s390x = lib/s390x/processor.c lib/s390x/ucall.c lib/s390x/diag318_test_handler.c diff --git a/tools/testing/selftests/kvm/demand_paging_test.c b/tools/testing/selftests/kvm/demand_paging_test.c index 3d96a7bfaff3..946161a9ce2d 100644 --- a/tools/testing/selftests/kvm/demand_paging_test.c +++ b/tools/testing/selftests/kvm/demand_paging_test.c @@ -7,23 +7,20 @@ * Copyright (C) 2019, Google, Inc. */ -#define _GNU_SOURCE /* for program_invocation_name */ +#define _GNU_SOURCE /* for program_invocation_name and pipe2 */ #include #include -#include -#include -#include #include #include #include -#include -#include #include +#include -#include "perf_test_util.h" -#include "processor.h" +#include "kvm_util.h" #include "test_util.h" +#include "perf_test_util.h" +#include "guest_modes.h" #ifdef __NR_userfaultfd @@ -248,9 +245,14 @@ static int setup_demand_paging(struct kvm_vm *vm, return 0; } -static void run_test(enum vm_guest_mode mode, bool use_uffd, - useconds_t uffd_delay) +struct test_params { + bool use_uffd; + useconds_t uffd_delay; +}; + +static void run_test(enum vm_guest_mode mode, void *arg) { + struct test_params *p = arg; pthread_t *vcpu_threads; pthread_t *uffd_handler_threads = NULL; struct uffd_handler_args *uffd_args = NULL; @@ -275,7 +277,7 @@ static void run_test(enum vm_guest_mode mode, bool use_uffd, add_vcpus(vm, nr_vcpus, guest_percpu_mem_size); - if (use_uffd) { + if (p->use_uffd) { uffd_handler_threads = malloc(nr_vcpus * sizeof(*uffd_handler_threads)); TEST_ASSERT(uffd_handler_threads, "Memory allocation failed"); @@ -308,7 +310,7 @@ static void run_test(enum vm_guest_mode mode, bool use_uffd, r = setup_demand_paging(vm, &uffd_handler_threads[vcpu_id], pipefds[vcpu_id * 2], - uffd_delay, &uffd_args[vcpu_id], + p->uffd_delay, &uffd_args[vcpu_id], vcpu_hva, guest_percpu_mem_size); if (r < 0) exit(-r); @@ -339,7 +341,7 @@ static void run_test(enum vm_guest_mode mode, bool use_uffd, pr_info("All vCPU threads joined\n"); - if (use_uffd) { + if (p->use_uffd) { char c; /* Tell the user fault fd handler threads to quit */ @@ -362,38 +364,19 @@ static void run_test(enum vm_guest_mode mode, bool use_uffd, free(guest_data_prototype); free(vcpu_threads); - if (use_uffd) { + if (p->use_uffd) { free(uffd_handler_threads); free(uffd_args); free(pipefds); } } -struct guest_mode { - bool supported; - bool enabled; -}; -static struct guest_mode guest_modes[NUM_VM_MODES]; - -#define guest_mode_init(mode, supported, enabled) ({ \ - guest_modes[mode] = (struct guest_mode){ supported, enabled }; \ -}) - static void help(char *name) { - int i; - puts(""); printf("usage: %s [-h] [-m mode] [-u] [-d uffd_delay_usec]\n" " [-b memory] [-v vcpus]\n", name); - printf(" -m: specify the guest mode ID to test\n" - " (default: test all supported modes)\n" - " This option may be used multiple times.\n" - " Guest mode IDs:\n"); - for (i = 0; i < NUM_VM_MODES; ++i) { - printf(" %d: %s%s\n", i, vm_guest_mode_string(i), - guest_modes[i].supported ? " (supported)" : ""); - } + guest_modes_help(); printf(" -u: use User Fault FD to handle vCPU page\n" " faults.\n"); printf(" -d: add a delay in usec to the User Fault\n" @@ -410,53 +393,22 @@ static void help(char *name) int main(int argc, char *argv[]) { int max_vcpus = kvm_check_cap(KVM_CAP_MAX_VCPUS); - bool mode_selected = false; - unsigned int mode; - int opt, i; - bool use_uffd = false; - useconds_t uffd_delay = 0; - -#ifdef __x86_64__ - guest_mode_init(VM_MODE_PXXV48_4K, true, true); -#endif -#ifdef __aarch64__ - guest_mode_init(VM_MODE_P40V48_4K, true, true); - guest_mode_init(VM_MODE_P40V48_64K, true, true); - { - unsigned int limit = kvm_check_cap(KVM_CAP_ARM_VM_IPA_SIZE); - - if (limit >= 52) - guest_mode_init(VM_MODE_P52V48_64K, true, true); - if (limit >= 48) { - guest_mode_init(VM_MODE_P48V48_4K, true, true); - guest_mode_init(VM_MODE_P48V48_64K, true, true); - } - } -#endif -#ifdef __s390x__ - guest_mode_init(VM_MODE_P40V48_4K, true, true); -#endif + struct test_params p = {}; + int opt; + + guest_modes_append_default(); while ((opt = getopt(argc, argv, "hm:ud:b:v:")) != -1) { switch (opt) { case 'm': - if (!mode_selected) { - for (i = 0; i < NUM_VM_MODES; ++i) - guest_modes[i].enabled = false; - mode_selected = true; - } - mode = strtoul(optarg, NULL, 10); - TEST_ASSERT(mode < NUM_VM_MODES, - "Guest mode ID %d too big", mode); - guest_modes[mode].enabled = true; + guest_modes_cmdline(optarg); break; case 'u': - use_uffd = true; + p.use_uffd = true; break; case 'd': - uffd_delay = strtoul(optarg, NULL, 0); - TEST_ASSERT(uffd_delay >= 0, - "A negative UFFD delay is not supported."); + p.uffd_delay = strtoul(optarg, NULL, 0); + TEST_ASSERT(p.uffd_delay >= 0, "A negative UFFD delay is not supported."); break; case 'b': guest_percpu_mem_size = parse_size(optarg); @@ -473,14 +425,7 @@ int main(int argc, char *argv[]) } } - for (i = 0; i < NUM_VM_MODES; ++i) { - if (!guest_modes[i].enabled) - continue; - TEST_ASSERT(guest_modes[i].supported, - "Guest mode ID %d (%s) not supported.", - i, vm_guest_mode_string(i)); - run_test(i, use_uffd, uffd_delay); - } + for_each_guest_mode(run_test, &p); return 0; } diff --git a/tools/testing/selftests/kvm/dirty_log_perf_test.c b/tools/testing/selftests/kvm/dirty_log_perf_test.c index 9c6a7be31e03..506741eb5d7f 100644 --- a/tools/testing/selftests/kvm/dirty_log_perf_test.c +++ b/tools/testing/selftests/kvm/dirty_log_perf_test.c @@ -12,16 +12,14 @@ #include #include -#include #include #include #include -#include #include "kvm_util.h" -#include "perf_test_util.h" -#include "processor.h" #include "test_util.h" +#include "perf_test_util.h" +#include "guest_modes.h" /* How many host loops to run by default (one KVM_GET_DIRTY_LOG for each loop)*/ #define TEST_HOST_LOOP_N 2UL @@ -89,9 +87,15 @@ static void *vcpu_worker(void *data) return NULL; } -static void run_test(enum vm_guest_mode mode, unsigned long iterations, - uint64_t phys_offset, int wr_fract) +struct test_params { + unsigned long iterations; + uint64_t phys_offset; + int wr_fract; +}; + +static void run_test(enum vm_guest_mode mode, void *arg) { + struct test_params *p = arg; pthread_t *vcpu_threads; struct kvm_vm *vm; unsigned long *bmap; @@ -108,7 +112,7 @@ static void run_test(enum vm_guest_mode mode, unsigned long iterations, vm = create_vm(mode, nr_vcpus, guest_percpu_mem_size); - perf_test_args.wr_fract = wr_fract; + perf_test_args.wr_fract = p->wr_fract; guest_num_pages = (nr_vcpus * guest_percpu_mem_size) >> vm_get_page_shift(vm); guest_num_pages = vm_adjust_num_guest_pages(mode, guest_num_pages); @@ -156,7 +160,7 @@ static void run_test(enum vm_guest_mode mode, unsigned long iterations, pr_info("Enabling dirty logging time: %ld.%.9lds\n\n", ts_diff.tv_sec, ts_diff.tv_nsec); - while (iteration < iterations) { + while (iteration < p->iterations) { /* * Incrementing the iteration number will start the vCPUs * dirtying memory again. @@ -210,15 +214,15 @@ static void run_test(enum vm_guest_mode mode, unsigned long iterations, pr_info("Disabling dirty logging time: %ld.%.9lds\n", ts_diff.tv_sec, ts_diff.tv_nsec); - avg = timespec_div(get_dirty_log_total, iterations); + avg = timespec_div(get_dirty_log_total, p->iterations); pr_info("Get dirty log over %lu iterations took %ld.%.9lds. (Avg %ld.%.9lds/iteration)\n", - iterations, get_dirty_log_total.tv_sec, + p->iterations, get_dirty_log_total.tv_sec, get_dirty_log_total.tv_nsec, avg.tv_sec, avg.tv_nsec); if (dirty_log_manual_caps) { - avg = timespec_div(clear_dirty_log_total, iterations); + avg = timespec_div(clear_dirty_log_total, p->iterations); pr_info("Clear dirty log over %lu iterations took %ld.%.9lds. (Avg %ld.%.9lds/iteration)\n", - iterations, clear_dirty_log_total.tv_sec, + p->iterations, clear_dirty_log_total.tv_sec, clear_dirty_log_total.tv_nsec, avg.tv_sec, avg.tv_nsec); } @@ -228,20 +232,8 @@ static void run_test(enum vm_guest_mode mode, unsigned long iterations, kvm_vm_free(vm); } -struct guest_mode { - bool supported; - bool enabled; -}; -static struct guest_mode guest_modes[NUM_VM_MODES]; - -#define guest_mode_init(mode, supported, enabled) ({ \ - guest_modes[mode] = (struct guest_mode){ supported, enabled }; \ -}) - static void help(char *name) { - int i; - puts(""); printf("usage: %s [-h] [-i iterations] [-p offset] " "[-m mode] [-b vcpu bytes] [-v vcpus]\n", name); @@ -250,14 +242,7 @@ static void help(char *name) TEST_HOST_LOOP_N); printf(" -p: specify guest physical test memory offset\n" " Warning: a low offset can conflict with the loaded test code.\n"); - printf(" -m: specify the guest mode ID to test " - "(default: test all supported modes)\n" - " This option may be used multiple times.\n" - " Guest mode IDs:\n"); - for (i = 0; i < NUM_VM_MODES; ++i) { - printf(" %d: %s%s\n", i, vm_guest_mode_string(i), - guest_modes[i].supported ? " (supported)" : ""); - } + guest_modes_help(); printf(" -b: specify the size of the memory region which should be\n" " dirtied by each vCPU. e.g. 10M or 3G.\n" " (default: 1G)\n"); @@ -272,74 +257,43 @@ static void help(char *name) int main(int argc, char *argv[]) { - unsigned long iterations = TEST_HOST_LOOP_N; - bool mode_selected = false; - uint64_t phys_offset = 0; - unsigned int mode; - int opt, i; - int wr_fract = 1; + int max_vcpus = kvm_check_cap(KVM_CAP_MAX_VCPUS); + struct test_params p = { + .iterations = TEST_HOST_LOOP_N, + .wr_fract = 1, + }; + int opt; dirty_log_manual_caps = kvm_check_cap(KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2); dirty_log_manual_caps &= (KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE | KVM_DIRTY_LOG_INITIALLY_SET); -#ifdef __x86_64__ - guest_mode_init(VM_MODE_PXXV48_4K, true, true); -#endif -#ifdef __aarch64__ - guest_mode_init(VM_MODE_P40V48_4K, true, true); - guest_mode_init(VM_MODE_P40V48_64K, true, true); - - { - unsigned int limit = kvm_check_cap(KVM_CAP_ARM_VM_IPA_SIZE); - - if (limit >= 52) - guest_mode_init(VM_MODE_P52V48_64K, true, true); - if (limit >= 48) { - guest_mode_init(VM_MODE_P48V48_4K, true, true); - guest_mode_init(VM_MODE_P48V48_64K, true, true); - } - } -#endif -#ifdef __s390x__ - guest_mode_init(VM_MODE_P40V48_4K, true, true); -#endif + guest_modes_append_default(); while ((opt = getopt(argc, argv, "hi:p:m:b:f:v:")) != -1) { switch (opt) { case 'i': - iterations = strtol(optarg, NULL, 10); + p.iterations = strtol(optarg, NULL, 10); break; case 'p': - phys_offset = strtoull(optarg, NULL, 0); + p.phys_offset = strtoull(optarg, NULL, 0); break; case 'm': - if (!mode_selected) { - for (i = 0; i < NUM_VM_MODES; ++i) - guest_modes[i].enabled = false; - mode_selected = true; - } - mode = strtoul(optarg, NULL, 10); - TEST_ASSERT(mode < NUM_VM_MODES, - "Guest mode ID %d too big", mode); - guest_modes[mode].enabled = true; + guest_modes_cmdline(optarg); break; case 'b': guest_percpu_mem_size = parse_size(optarg); break; case 'f': - wr_fract = atoi(optarg); - TEST_ASSERT(wr_fract >= 1, + p.wr_fract = atoi(optarg); + TEST_ASSERT(p.wr_fract >= 1, "Write fraction cannot be less than one"); break; case 'v': nr_vcpus = atoi(optarg); - TEST_ASSERT(nr_vcpus > 0, - "Must have a positive number of vCPUs"); - TEST_ASSERT(nr_vcpus <= MAX_VCPUS, - "This test does not currently support\n" - "more than %d vCPUs.", MAX_VCPUS); + TEST_ASSERT(nr_vcpus > 0 && nr_vcpus <= max_vcpus, + "Invalid number of vcpus, must be between 1 and %d", max_vcpus); break; case 'h': default: @@ -348,18 +302,11 @@ int main(int argc, char *argv[]) } } - TEST_ASSERT(iterations >= 2, "The test should have at least two iterations"); + TEST_ASSERT(p.iterations >= 2, "The test should have at least two iterations"); - pr_info("Test iterations: %"PRIu64"\n", iterations); + pr_info("Test iterations: %"PRIu64"\n", p.iterations); - for (i = 0; i < NUM_VM_MODES; ++i) { - if (!guest_modes[i].enabled) - continue; - TEST_ASSERT(guest_modes[i].supported, - "Guest mode ID %d (%s) not supported.", - i, vm_guest_mode_string(i)); - run_test(i, iterations, phys_offset, wr_fract); - } + for_each_guest_mode(run_test, &p); return 0; } diff --git a/tools/testing/selftests/kvm/dirty_log_test.c b/tools/testing/selftests/kvm/dirty_log_test.c index 471baecb7772..bb2752d78fe3 100644 --- a/tools/testing/selftests/kvm/dirty_log_test.c +++ b/tools/testing/selftests/kvm/dirty_log_test.c @@ -9,8 +9,6 @@ #include #include -#include -#include #include #include #include @@ -20,8 +18,9 @@ #include #include -#include "test_util.h" #include "kvm_util.h" +#include "test_util.h" +#include "guest_modes.h" #include "processor.h" #define VCPU_ID 1 @@ -673,9 +672,15 @@ static struct kvm_vm *create_vm(enum vm_guest_mode mode, uint32_t vcpuid, #define DIRTY_MEM_BITS 30 /* 1G */ #define PAGE_SHIFT_4K 12 -static void run_test(enum vm_guest_mode mode, unsigned long iterations, - unsigned long interval, uint64_t phys_offset) +struct test_params { + unsigned long iterations; + unsigned long interval; + uint64_t phys_offset; +}; + +static void run_test(enum vm_guest_mode mode, void *arg) { + struct test_params *p = arg; struct kvm_vm *vm; unsigned long *bmap; @@ -709,12 +714,12 @@ static void run_test(enum vm_guest_mode mode, unsigned long iterations, host_page_size = getpagesize(); host_num_pages = vm_num_host_pages(mode, guest_num_pages); - if (!phys_offset) { + if (!p->phys_offset) { guest_test_phys_mem = (vm_get_max_gfn(vm) - guest_num_pages) * guest_page_size; guest_test_phys_mem &= ~(host_page_size - 1); } else { - guest_test_phys_mem = phys_offset; + guest_test_phys_mem = p->phys_offset; } #ifdef __s390x__ @@ -758,9 +763,9 @@ static void run_test(enum vm_guest_mode mode, unsigned long iterations, pthread_create(&vcpu_thread, NULL, vcpu_worker, vm); - while (iteration < iterations) { + while (iteration < p->iterations) { /* Give the vcpu thread some time to dirty some pages */ - usleep(interval * 1000); + usleep(p->interval * 1000); log_mode_collect_dirty_pages(vm, TEST_MEM_SLOT_INDEX, bmap, host_num_pages); vm_dirty_log_verify(mode, bmap); @@ -783,20 +788,8 @@ static void run_test(enum vm_guest_mode mode, unsigned long iterations, kvm_vm_free(vm); } -struct guest_mode { - bool supported; - bool enabled; -}; -static struct guest_mode guest_modes[NUM_VM_MODES]; - -#define guest_mode_init(mode, supported, enabled) ({ \ - guest_modes[mode] = (struct guest_mode){ supported, enabled }; \ -}) - static void help(char *name) { - int i; - puts(""); printf("usage: %s [-h] [-i iterations] [-I interval] " "[-p offset] [-m mode]\n", name); @@ -813,51 +806,23 @@ static void help(char *name) printf(" -M: specify the host logging mode " "(default: run all log modes). Supported modes: \n\t"); log_modes_dump(); - printf(" -m: specify the guest mode ID to test " - "(default: test all supported modes)\n" - " This option may be used multiple times.\n" - " Guest mode IDs:\n"); - for (i = 0; i < NUM_VM_MODES; ++i) { - printf(" %d: %s%s\n", i, vm_guest_mode_string(i), - guest_modes[i].supported ? " (supported)" : ""); - } + guest_modes_help(); puts(""); exit(0); } int main(int argc, char *argv[]) { - unsigned long iterations = TEST_HOST_LOOP_N; - unsigned long interval = TEST_HOST_LOOP_INTERVAL; - bool mode_selected = false; - uint64_t phys_offset = 0; - unsigned int mode; - int opt, i, j; + struct test_params p = { + .iterations = TEST_HOST_LOOP_N, + .interval = TEST_HOST_LOOP_INTERVAL, + }; + int opt, i; sem_init(&dirty_ring_vcpu_stop, 0, 0); sem_init(&dirty_ring_vcpu_cont, 0, 0); -#ifdef __x86_64__ - guest_mode_init(VM_MODE_PXXV48_4K, true, true); -#endif -#ifdef __aarch64__ - guest_mode_init(VM_MODE_P40V48_4K, true, true); - guest_mode_init(VM_MODE_P40V48_64K, true, true); - - { - unsigned int limit = kvm_check_cap(KVM_CAP_ARM_VM_IPA_SIZE); - - if (limit >= 52) - guest_mode_init(VM_MODE_P52V48_64K, true, true); - if (limit >= 48) { - guest_mode_init(VM_MODE_P48V48_4K, true, true); - guest_mode_init(VM_MODE_P48V48_64K, true, true); - } - } -#endif -#ifdef __s390x__ - guest_mode_init(VM_MODE_P40V48_4K, true, true); -#endif + guest_modes_append_default(); while ((opt = getopt(argc, argv, "c:hi:I:p:m:M:")) != -1) { switch (opt) { @@ -865,24 +830,16 @@ int main(int argc, char *argv[]) test_dirty_ring_count = strtol(optarg, NULL, 10); break; case 'i': - iterations = strtol(optarg, NULL, 10); + p.iterations = strtol(optarg, NULL, 10); break; case 'I': - interval = strtol(optarg, NULL, 10); + p.interval = strtol(optarg, NULL, 10); break; case 'p': - phys_offset = strtoull(optarg, NULL, 0); + p.phys_offset = strtoull(optarg, NULL, 0); break; case 'm': - if (!mode_selected) { - for (i = 0; i < NUM_VM_MODES; ++i) - guest_modes[i].enabled = false; - mode_selected = true; - } - mode = strtoul(optarg, NULL, 10); - TEST_ASSERT(mode < NUM_VM_MODES, - "Guest mode ID %d too big", mode); - guest_modes[mode].enabled = true; + guest_modes_cmdline(optarg); break; case 'M': if (!strcmp(optarg, "all")) { @@ -911,32 +868,24 @@ int main(int argc, char *argv[]) } } - TEST_ASSERT(iterations > 2, "Iterations must be greater than two"); - TEST_ASSERT(interval > 0, "Interval must be greater than zero"); + TEST_ASSERT(p.iterations > 2, "Iterations must be greater than two"); + TEST_ASSERT(p.interval > 0, "Interval must be greater than zero"); pr_info("Test iterations: %"PRIu64", interval: %"PRIu64" (ms)\n", - iterations, interval); + p.iterations, p.interval); srandom(time(0)); - for (i = 0; i < NUM_VM_MODES; ++i) { - if (!guest_modes[i].enabled) - continue; - TEST_ASSERT(guest_modes[i].supported, - "Guest mode ID %d (%s) not supported.", - i, vm_guest_mode_string(i)); - if (host_log_mode_option == LOG_MODE_ALL) { - /* Run each log mode */ - for (j = 0; j < LOG_MODE_NUM; j++) { - pr_info("Testing Log Mode '%s'\n", - log_modes[j].name); - host_log_mode = j; - run_test(i, iterations, interval, phys_offset); - } - } else { - host_log_mode = host_log_mode_option; - run_test(i, iterations, interval, phys_offset); + if (host_log_mode_option == LOG_MODE_ALL) { + /* Run each log mode */ + for (i = 0; i < LOG_MODE_NUM; i++) { + pr_info("Testing Log Mode '%s'\n", log_modes[i].name); + host_log_mode = i; + for_each_guest_mode(run_test, &p); } + } else { + host_log_mode = host_log_mode_option; + for_each_guest_mode(run_test, &p); } return 0; diff --git a/tools/testing/selftests/kvm/include/guest_modes.h b/tools/testing/selftests/kvm/include/guest_modes.h new file mode 100644 index 000000000000..b691df33e64e --- /dev/null +++ b/tools/testing/selftests/kvm/include/guest_modes.h @@ -0,0 +1,21 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2020, Red Hat, Inc. + */ +#include "kvm_util.h" + +struct guest_mode { + bool supported; + bool enabled; +}; + +extern struct guest_mode guest_modes[NUM_VM_MODES]; + +#define guest_mode_append(mode, supported, enabled) ({ \ + guest_modes[mode] = (struct guest_mode){ supported, enabled }; \ +}) + +void guest_modes_append_default(void); +void for_each_guest_mode(void (*func)(enum vm_guest_mode, void *), void *arg); +void guest_modes_help(void); +void guest_modes_cmdline(const char *arg); diff --git a/tools/testing/selftests/kvm/lib/guest_modes.c b/tools/testing/selftests/kvm/lib/guest_modes.c new file mode 100644 index 000000000000..25bff307c71f --- /dev/null +++ b/tools/testing/selftests/kvm/lib/guest_modes.c @@ -0,0 +1,70 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2020, Red Hat, Inc. + */ +#include "guest_modes.h" + +struct guest_mode guest_modes[NUM_VM_MODES]; + +void guest_modes_append_default(void) +{ + guest_mode_append(VM_MODE_DEFAULT, true, true); + +#ifdef __aarch64__ + guest_mode_append(VM_MODE_P40V48_64K, true, true); + { + unsigned int limit = kvm_check_cap(KVM_CAP_ARM_VM_IPA_SIZE); + if (limit >= 52) + guest_mode_append(VM_MODE_P52V48_64K, true, true); + if (limit >= 48) { + guest_mode_append(VM_MODE_P48V48_4K, true, true); + guest_mode_append(VM_MODE_P48V48_64K, true, true); + } + } +#endif +} + +void for_each_guest_mode(void (*func)(enum vm_guest_mode, void *), void *arg) +{ + int i; + + for (i = 0; i < NUM_VM_MODES; ++i) { + if (!guest_modes[i].enabled) + continue; + TEST_ASSERT(guest_modes[i].supported, + "Guest mode ID %d (%s) not supported.", + i, vm_guest_mode_string(i)); + func(i, arg); + } +} + +void guest_modes_help(void) +{ + int i; + + printf(" -m: specify the guest mode ID to test\n" + " (default: test all supported modes)\n" + " This option may be used multiple times.\n" + " Guest mode IDs:\n"); + for (i = 0; i < NUM_VM_MODES; ++i) { + printf(" %d: %s%s\n", i, vm_guest_mode_string(i), + guest_modes[i].supported ? " (supported)" : ""); + } +} + +void guest_modes_cmdline(const char *arg) +{ + static bool mode_selected; + unsigned int mode; + int i; + + if (!mode_selected) { + for (i = 0; i < NUM_VM_MODES; ++i) + guest_modes[i].enabled = false; + mode_selected = true; + } + + mode = strtoul(optarg, NULL, 10); + TEST_ASSERT(mode < NUM_VM_MODES, "Guest mode ID %d too big", mode); + guest_modes[mode].enabled = true; +} -- cgit v1.2.3-59-g8ed1b From 1133e17ea7c9929ff7b90e81d8926f9e870748e9 Mon Sep 17 00:00:00 2001 From: Andrew Jones Date: Fri, 18 Dec 2020 15:17:33 +0100 Subject: KVM: selftests: Use vm_create_with_vcpus in create_vm Reviewed-by: Ben Gardon Signed-off-by: Andrew Jones Message-Id: <20201218141734.54359-3-drjones@redhat.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/demand_paging_test.c | 2 +- tools/testing/selftests/kvm/dirty_log_perf_test.c | 2 - tools/testing/selftests/kvm/include/kvm_util.h | 8 ++++ .../testing/selftests/kvm/include/perf_test_util.h | 47 +++++----------------- tools/testing/selftests/kvm/lib/kvm_util.c | 9 +---- 5 files changed, 21 insertions(+), 47 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/kvm/demand_paging_test.c b/tools/testing/selftests/kvm/demand_paging_test.c index 946161a9ce2d..b0c41de32e9b 100644 --- a/tools/testing/selftests/kvm/demand_paging_test.c +++ b/tools/testing/selftests/kvm/demand_paging_test.c @@ -7,7 +7,7 @@ * Copyright (C) 2019, Google, Inc. */ -#define _GNU_SOURCE /* for program_invocation_name and pipe2 */ +#define _GNU_SOURCE /* for pipe2 */ #include #include diff --git a/tools/testing/selftests/kvm/dirty_log_perf_test.c b/tools/testing/selftests/kvm/dirty_log_perf_test.c index 506741eb5d7f..36bea75a8d6f 100644 --- a/tools/testing/selftests/kvm/dirty_log_perf_test.c +++ b/tools/testing/selftests/kvm/dirty_log_perf_test.c @@ -8,8 +8,6 @@ * Copyright (C) 2020, Google, Inc. */ -#define _GNU_SOURCE /* for program_invocation_name */ - #include #include #include diff --git a/tools/testing/selftests/kvm/include/kvm_util.h b/tools/testing/selftests/kvm/include/kvm_util.h index dfa9d369e8fc..149766ecd68b 100644 --- a/tools/testing/selftests/kvm/include/kvm_util.h +++ b/tools/testing/selftests/kvm/include/kvm_util.h @@ -70,6 +70,14 @@ enum vm_guest_mode { #define vm_guest_mode_string(m) vm_guest_mode_string[m] extern const char * const vm_guest_mode_string[]; +struct vm_guest_mode_params { + unsigned int pa_bits; + unsigned int va_bits; + unsigned int page_size; + unsigned int page_shift; +}; +extern const struct vm_guest_mode_params vm_guest_mode_params[]; + enum vm_mem_backing_src_type { VM_MEM_SRC_ANONYMOUS, VM_MEM_SRC_ANONYMOUS_THP, diff --git a/tools/testing/selftests/kvm/include/perf_test_util.h b/tools/testing/selftests/kvm/include/perf_test_util.h index 239421e4f6b8..cd4c258f458d 100644 --- a/tools/testing/selftests/kvm/include/perf_test_util.h +++ b/tools/testing/selftests/kvm/include/perf_test_util.h @@ -13,9 +13,6 @@ #define MAX_VCPUS 512 -#define PAGE_SHIFT_4K 12 -#define PTES_PER_4K_PT 512 - #define TEST_MEM_SLOT_INDEX 1 /* Default guest test virtual memory offset */ @@ -94,41 +91,26 @@ static struct kvm_vm *create_vm(enum vm_guest_mode mode, int vcpus, uint64_t vcpu_memory_bytes) { struct kvm_vm *vm; - uint64_t pages = DEFAULT_GUEST_PHY_PAGES; uint64_t guest_num_pages; - /* Account for a few pages per-vCPU for stacks */ - pages += DEFAULT_STACK_PGS * vcpus; - - /* - * Reserve twice the ammount of memory needed to map the test region and - * the page table / stacks region, at 4k, for page tables. Do the - * calculation with 4K page size: the smallest of all archs. (e.g., 64K - * page size guest will need even less memory for page tables). - */ - pages += (2 * pages) / PTES_PER_4K_PT; - pages += ((2 * vcpus * vcpu_memory_bytes) >> PAGE_SHIFT_4K) / - PTES_PER_4K_PT; - pages = vm_adjust_num_guest_pages(mode, pages); - pr_info("Testing guest mode: %s\n", vm_guest_mode_string(mode)); - vm = vm_create(mode, pages, O_RDWR); - kvm_vm_elf_load(vm, program_invocation_name, 0, 0); -#ifdef __x86_64__ - vm_create_irqchip(vm); -#endif - - perf_test_args.vm = vm; - perf_test_args.guest_page_size = vm_get_page_size(vm); perf_test_args.host_page_size = getpagesize(); + perf_test_args.guest_page_size = vm_guest_mode_params[mode].page_size; + guest_num_pages = vm_adjust_num_guest_pages(mode, + (vcpus * vcpu_memory_bytes) / perf_test_args.guest_page_size); + + TEST_ASSERT(vcpu_memory_bytes % perf_test_args.host_page_size == 0, + "Guest memory size is not host page size aligned."); TEST_ASSERT(vcpu_memory_bytes % perf_test_args.guest_page_size == 0, "Guest memory size is not guest page size aligned."); - guest_num_pages = (vcpus * vcpu_memory_bytes) / - perf_test_args.guest_page_size; - guest_num_pages = vm_adjust_num_guest_pages(mode, guest_num_pages); + vm = vm_create_with_vcpus(mode, vcpus, + (vcpus * vcpu_memory_bytes) / perf_test_args.guest_page_size, + 0, guest_code, NULL); + + perf_test_args.vm = vm; /* * If there should be more memory in the guest test region than there @@ -140,18 +122,13 @@ static struct kvm_vm *create_vm(enum vm_guest_mode mode, int vcpus, guest_num_pages, vm_get_max_gfn(vm), vcpus, vcpu_memory_bytes); - TEST_ASSERT(vcpu_memory_bytes % perf_test_args.host_page_size == 0, - "Guest memory size is not host page size aligned."); - guest_test_phys_mem = (vm_get_max_gfn(vm) - guest_num_pages) * perf_test_args.guest_page_size; guest_test_phys_mem &= ~(perf_test_args.host_page_size - 1); - #ifdef __s390x__ /* Align to 1M (segment size) */ guest_test_phys_mem &= ~((1 << 20) - 1); #endif - pr_info("guest physical test memory offset: 0x%lx\n", guest_test_phys_mem); /* Add an extra memory slot for testing */ @@ -177,8 +154,6 @@ static void add_vcpus(struct kvm_vm *vm, int vcpus, uint64_t vcpu_memory_bytes) for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) { vcpu_args = &perf_test_args.vcpu_args[vcpu_id]; - vm_vcpu_add_default(vm, vcpu_id, guest_code); - vcpu_args->vcpu_id = vcpu_id; vcpu_args->gva = guest_test_virt_mem + (vcpu_id * vcpu_memory_bytes); diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c index 88ef7067f1e6..fa5a90e6c6f0 100644 --- a/tools/testing/selftests/kvm/lib/kvm_util.c +++ b/tools/testing/selftests/kvm/lib/kvm_util.c @@ -153,14 +153,7 @@ const char * const vm_guest_mode_string[] = { _Static_assert(sizeof(vm_guest_mode_string)/sizeof(char *) == NUM_VM_MODES, "Missing new mode strings?"); -struct vm_guest_mode_params { - unsigned int pa_bits; - unsigned int va_bits; - unsigned int page_size; - unsigned int page_shift; -}; - -static const struct vm_guest_mode_params vm_guest_mode_params[] = { +const struct vm_guest_mode_params vm_guest_mode_params[] = { { 52, 48, 0x1000, 12 }, { 52, 48, 0x10000, 16 }, { 48, 48, 0x1000, 12 }, -- cgit v1.2.3-59-g8ed1b From b268b6f0bd36322358accb15c45683a9e1220231 Mon Sep 17 00:00:00 2001 From: Andrew Jones Date: Fri, 18 Dec 2020 15:17:34 +0100 Subject: KVM: selftests: Implement perf_test_util more conventionally It's not conventional C to put non-inline functions in header files. Create a source file for the functions instead. Also reduce the amount of globals and rename the functions to something less generic. Reviewed-by: Ben Gardon Reviewed-by: Peter Xu Signed-off-by: Andrew Jones Message-Id: <20201218141734.54359-4-drjones@redhat.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/Makefile | 2 +- tools/testing/selftests/kvm/demand_paging_test.c | 11 +- tools/testing/selftests/kvm/dirty_log_perf_test.c | 22 ++-- tools/testing/selftests/kvm/include/kvm_util.h | 1 + .../testing/selftests/kvm/include/perf_test_util.h | 142 ++------------------- tools/testing/selftests/kvm/lib/perf_test_util.c | 134 +++++++++++++++++++ 6 files changed, 166 insertions(+), 146 deletions(-) create mode 100644 tools/testing/selftests/kvm/lib/perf_test_util.c (limited to 'tools') diff --git a/tools/testing/selftests/kvm/Makefile b/tools/testing/selftests/kvm/Makefile index a7286a08c3ae..fe41c6a0fa67 100644 --- a/tools/testing/selftests/kvm/Makefile +++ b/tools/testing/selftests/kvm/Makefile @@ -33,7 +33,7 @@ ifeq ($(ARCH),s390) UNAME_M := s390x endif -LIBKVM = lib/assert.c lib/elf.c lib/io.c lib/kvm_util.c lib/sparsebit.c lib/test_util.c lib/guest_modes.c +LIBKVM = lib/assert.c lib/elf.c lib/io.c lib/kvm_util.c lib/sparsebit.c lib/test_util.c lib/guest_modes.c lib/perf_test_util.c LIBKVM_x86_64 = lib/x86_64/processor.c lib/x86_64/vmx.c lib/x86_64/svm.c lib/x86_64/ucall.c lib/x86_64/handlers.S LIBKVM_aarch64 = lib/aarch64/processor.c lib/aarch64/ucall.c LIBKVM_s390x = lib/s390x/processor.c lib/s390x/ucall.c lib/s390x/diag318_test_handler.c diff --git a/tools/testing/selftests/kvm/demand_paging_test.c b/tools/testing/selftests/kvm/demand_paging_test.c index b0c41de32e9b..cdad1eca72f7 100644 --- a/tools/testing/selftests/kvm/demand_paging_test.c +++ b/tools/testing/selftests/kvm/demand_paging_test.c @@ -36,12 +36,14 @@ #define PER_VCPU_DEBUG(...) _no_printf(__VA_ARGS__) #endif +static int nr_vcpus = 1; +static uint64_t guest_percpu_mem_size = DEFAULT_PER_VCPU_MEM_SIZE; static char *guest_data_prototype; static void *vcpu_worker(void *data) { int ret; - struct vcpu_args *vcpu_args = (struct vcpu_args *)data; + struct perf_test_vcpu_args *vcpu_args = (struct perf_test_vcpu_args *)data; int vcpu_id = vcpu_args->vcpu_id; struct kvm_vm *vm = perf_test_args.vm; struct kvm_run *run; @@ -263,7 +265,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) int vcpu_id; int r; - vm = create_vm(mode, nr_vcpus, guest_percpu_mem_size); + vm = perf_test_create_vm(mode, nr_vcpus, guest_percpu_mem_size); perf_test_args.wr_fract = 1; @@ -275,7 +277,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - add_vcpus(vm, nr_vcpus, guest_percpu_mem_size); + perf_test_setup_vcpus(vm, nr_vcpus, guest_percpu_mem_size); if (p->use_uffd) { uffd_handler_threads = @@ -359,8 +361,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) perf_test_args.vcpu_args[0].pages * nr_vcpus / ((double)ts_diff.tv_sec + (double)ts_diff.tv_nsec / 100000000.0)); - ucall_uninit(vm); - kvm_vm_free(vm); + perf_test_destroy_vm(vm); free(guest_data_prototype); free(vcpu_threads); diff --git a/tools/testing/selftests/kvm/dirty_log_perf_test.c b/tools/testing/selftests/kvm/dirty_log_perf_test.c index 36bea75a8d6f..2283a0ec74a9 100644 --- a/tools/testing/selftests/kvm/dirty_log_perf_test.c +++ b/tools/testing/selftests/kvm/dirty_log_perf_test.c @@ -22,11 +22,14 @@ /* How many host loops to run by default (one KVM_GET_DIRTY_LOG for each loop)*/ #define TEST_HOST_LOOP_N 2UL +static int nr_vcpus = 1; +static uint64_t guest_percpu_mem_size = DEFAULT_PER_VCPU_MEM_SIZE; + /* Host variables */ static u64 dirty_log_manual_caps; static bool host_quit; static uint64_t iteration; -static uint64_t vcpu_last_completed_iteration[MAX_VCPUS]; +static uint64_t vcpu_last_completed_iteration[KVM_MAX_VCPUS]; static void *vcpu_worker(void *data) { @@ -38,7 +41,7 @@ static void *vcpu_worker(void *data) struct timespec ts_diff; struct timespec total = (struct timespec){0}; struct timespec avg; - struct vcpu_args *vcpu_args = (struct vcpu_args *)data; + struct perf_test_vcpu_args *vcpu_args = (struct perf_test_vcpu_args *)data; int vcpu_id = vcpu_args->vcpu_id; vcpu_args_set(vm, vcpu_id, 1, vcpu_id); @@ -108,7 +111,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) struct kvm_enable_cap cap = {}; struct timespec clear_dirty_log_total = (struct timespec){0}; - vm = create_vm(mode, nr_vcpus, guest_percpu_mem_size); + vm = perf_test_create_vm(mode, nr_vcpus, guest_percpu_mem_size); perf_test_args.wr_fract = p->wr_fract; @@ -126,7 +129,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads)); TEST_ASSERT(vcpu_threads, "Memory allocation failed"); - add_vcpus(vm, nr_vcpus, guest_percpu_mem_size); + perf_test_setup_vcpus(vm, nr_vcpus, guest_percpu_mem_size); sync_global_to_guest(vm, perf_test_args); @@ -152,7 +155,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) /* Enable dirty logging */ clock_gettime(CLOCK_MONOTONIC, &start); - vm_mem_region_set_flags(vm, TEST_MEM_SLOT_INDEX, + vm_mem_region_set_flags(vm, PERF_TEST_MEM_SLOT_INDEX, KVM_MEM_LOG_DIRTY_PAGES); ts_diff = timespec_diff_now(start); pr_info("Enabling dirty logging time: %ld.%.9lds\n\n", @@ -179,7 +182,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) iteration, ts_diff.tv_sec, ts_diff.tv_nsec); clock_gettime(CLOCK_MONOTONIC, &start); - kvm_vm_get_dirty_log(vm, TEST_MEM_SLOT_INDEX, bmap); + kvm_vm_get_dirty_log(vm, PERF_TEST_MEM_SLOT_INDEX, bmap); ts_diff = timespec_diff_now(start); get_dirty_log_total = timespec_add(get_dirty_log_total, @@ -189,7 +192,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) if (dirty_log_manual_caps) { clock_gettime(CLOCK_MONOTONIC, &start); - kvm_vm_clear_dirty_log(vm, TEST_MEM_SLOT_INDEX, bmap, 0, + kvm_vm_clear_dirty_log(vm, PERF_TEST_MEM_SLOT_INDEX, bmap, 0, host_num_pages); ts_diff = timespec_diff_now(start); @@ -207,7 +210,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) /* Disable dirty logging */ clock_gettime(CLOCK_MONOTONIC, &start); - vm_mem_region_set_flags(vm, TEST_MEM_SLOT_INDEX, 0); + vm_mem_region_set_flags(vm, PERF_TEST_MEM_SLOT_INDEX, 0); ts_diff = timespec_diff_now(start); pr_info("Disabling dirty logging time: %ld.%.9lds\n", ts_diff.tv_sec, ts_diff.tv_nsec); @@ -226,8 +229,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) free(bmap); free(vcpu_threads); - ucall_uninit(vm); - kvm_vm_free(vm); + perf_test_destroy_vm(vm); } static void help(char *name) diff --git a/tools/testing/selftests/kvm/include/kvm_util.h b/tools/testing/selftests/kvm/include/kvm_util.h index 149766ecd68b..5cbb861525ed 100644 --- a/tools/testing/selftests/kvm/include/kvm_util.h +++ b/tools/testing/selftests/kvm/include/kvm_util.h @@ -16,6 +16,7 @@ #include "sparsebit.h" +#define KVM_MAX_VCPUS 512 /* * Callers of kvm_util only have an incomplete/opaque description of the diff --git a/tools/testing/selftests/kvm/include/perf_test_util.h b/tools/testing/selftests/kvm/include/perf_test_util.h index cd4c258f458d..b1188823c31b 100644 --- a/tools/testing/selftests/kvm/include/perf_test_util.h +++ b/tools/testing/selftests/kvm/include/perf_test_util.h @@ -9,35 +9,15 @@ #define SELFTEST_KVM_PERF_TEST_UTIL_H #include "kvm_util.h" -#include "processor.h" - -#define MAX_VCPUS 512 - -#define TEST_MEM_SLOT_INDEX 1 /* Default guest test virtual memory offset */ #define DEFAULT_GUEST_TEST_MEM 0xc0000000 #define DEFAULT_PER_VCPU_MEM_SIZE (1 << 30) /* 1G */ -/* - * Guest physical memory offset of the testing memory slot. - * This will be set to the topmost valid physical address minus - * the test memory size. - */ -static uint64_t guest_test_phys_mem; - -/* - * Guest virtual memory offset of the testing memory slot. - * Must not conflict with identity mapped test code. - */ -static uint64_t guest_test_virt_mem = DEFAULT_GUEST_TEST_MEM; -static uint64_t guest_percpu_mem_size = DEFAULT_PER_VCPU_MEM_SIZE; - -/* Number of VCPUs for the test */ -static int nr_vcpus = 1; +#define PERF_TEST_MEM_SLOT_INDEX 1 -struct vcpu_args { +struct perf_test_vcpu_args { uint64_t gva; uint64_t pages; @@ -51,119 +31,21 @@ struct perf_test_args { uint64_t guest_page_size; int wr_fract; - struct vcpu_args vcpu_args[MAX_VCPUS]; + struct perf_test_vcpu_args vcpu_args[KVM_MAX_VCPUS]; }; -static struct perf_test_args perf_test_args; +extern struct perf_test_args perf_test_args; /* - * Continuously write to the first 8 bytes of each page in the - * specified region. + * Guest physical memory offset of the testing memory slot. + * This will be set to the topmost valid physical address minus + * the test memory size. */ -static void guest_code(uint32_t vcpu_id) -{ - struct vcpu_args *vcpu_args = &perf_test_args.vcpu_args[vcpu_id]; - uint64_t gva; - uint64_t pages; - int i; - - /* Make sure vCPU args data structure is not corrupt. */ - GUEST_ASSERT(vcpu_args->vcpu_id == vcpu_id); - - gva = vcpu_args->gva; - pages = vcpu_args->pages; - - while (true) { - for (i = 0; i < pages; i++) { - uint64_t addr = gva + (i * perf_test_args.guest_page_size); - - if (i % perf_test_args.wr_fract == 0) - *(uint64_t *)addr = 0x0123456789ABCDEF; - else - READ_ONCE(*(uint64_t *)addr); - } - - GUEST_SYNC(1); - } -} - -static struct kvm_vm *create_vm(enum vm_guest_mode mode, int vcpus, - uint64_t vcpu_memory_bytes) -{ - struct kvm_vm *vm; - uint64_t guest_num_pages; - - pr_info("Testing guest mode: %s\n", vm_guest_mode_string(mode)); - - perf_test_args.host_page_size = getpagesize(); - perf_test_args.guest_page_size = vm_guest_mode_params[mode].page_size; - - guest_num_pages = vm_adjust_num_guest_pages(mode, - (vcpus * vcpu_memory_bytes) / perf_test_args.guest_page_size); - - TEST_ASSERT(vcpu_memory_bytes % perf_test_args.host_page_size == 0, - "Guest memory size is not host page size aligned."); - TEST_ASSERT(vcpu_memory_bytes % perf_test_args.guest_page_size == 0, - "Guest memory size is not guest page size aligned."); - - vm = vm_create_with_vcpus(mode, vcpus, - (vcpus * vcpu_memory_bytes) / perf_test_args.guest_page_size, - 0, guest_code, NULL); - - perf_test_args.vm = vm; - - /* - * If there should be more memory in the guest test region than there - * can be pages in the guest, it will definitely cause problems. - */ - TEST_ASSERT(guest_num_pages < vm_get_max_gfn(vm), - "Requested more guest memory than address space allows.\n" - " guest pages: %lx max gfn: %x vcpus: %d wss: %lx]\n", - guest_num_pages, vm_get_max_gfn(vm), vcpus, - vcpu_memory_bytes); - - guest_test_phys_mem = (vm_get_max_gfn(vm) - guest_num_pages) * - perf_test_args.guest_page_size; - guest_test_phys_mem &= ~(perf_test_args.host_page_size - 1); -#ifdef __s390x__ - /* Align to 1M (segment size) */ - guest_test_phys_mem &= ~((1 << 20) - 1); -#endif - pr_info("guest physical test memory offset: 0x%lx\n", guest_test_phys_mem); - - /* Add an extra memory slot for testing */ - vm_userspace_mem_region_add(vm, VM_MEM_SRC_ANONYMOUS, - guest_test_phys_mem, - TEST_MEM_SLOT_INDEX, - guest_num_pages, 0); - - /* Do mapping for the demand paging memory slot */ - virt_map(vm, guest_test_virt_mem, guest_test_phys_mem, guest_num_pages, 0); - - ucall_init(vm, NULL); - - return vm; -} - -static void add_vcpus(struct kvm_vm *vm, int vcpus, uint64_t vcpu_memory_bytes) -{ - vm_paddr_t vcpu_gpa; - struct vcpu_args *vcpu_args; - int vcpu_id; - - for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) { - vcpu_args = &perf_test_args.vcpu_args[vcpu_id]; - - vcpu_args->vcpu_id = vcpu_id; - vcpu_args->gva = guest_test_virt_mem + - (vcpu_id * vcpu_memory_bytes); - vcpu_args->pages = vcpu_memory_bytes / - perf_test_args.guest_page_size; +extern uint64_t guest_test_phys_mem; - vcpu_gpa = guest_test_phys_mem + (vcpu_id * vcpu_memory_bytes); - pr_debug("Added VCPU %d with test mem gpa [%lx, %lx)\n", - vcpu_id, vcpu_gpa, vcpu_gpa + vcpu_memory_bytes); - } -} +struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, + uint64_t vcpu_memory_bytes); +void perf_test_destroy_vm(struct kvm_vm *vm); +void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, uint64_t vcpu_memory_bytes); #endif /* SELFTEST_KVM_PERF_TEST_UTIL_H */ diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c new file mode 100644 index 000000000000..9be1944c2d1c --- /dev/null +++ b/tools/testing/selftests/kvm/lib/perf_test_util.c @@ -0,0 +1,134 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2020, Google LLC. + */ + +#include "kvm_util.h" +#include "perf_test_util.h" +#include "processor.h" + +struct perf_test_args perf_test_args; + +uint64_t guest_test_phys_mem; + +/* + * Guest virtual memory offset of the testing memory slot. + * Must not conflict with identity mapped test code. + */ +static uint64_t guest_test_virt_mem = DEFAULT_GUEST_TEST_MEM; + +/* + * Continuously write to the first 8 bytes of each page in the + * specified region. + */ +static void guest_code(uint32_t vcpu_id) +{ + struct perf_test_vcpu_args *vcpu_args = &perf_test_args.vcpu_args[vcpu_id]; + uint64_t gva; + uint64_t pages; + int i; + + /* Make sure vCPU args data structure is not corrupt. */ + GUEST_ASSERT(vcpu_args->vcpu_id == vcpu_id); + + gva = vcpu_args->gva; + pages = vcpu_args->pages; + + while (true) { + for (i = 0; i < pages; i++) { + uint64_t addr = gva + (i * perf_test_args.guest_page_size); + + if (i % perf_test_args.wr_fract == 0) + *(uint64_t *)addr = 0x0123456789ABCDEF; + else + READ_ONCE(*(uint64_t *)addr); + } + + GUEST_SYNC(1); + } +} + +struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus, + uint64_t vcpu_memory_bytes) +{ + struct kvm_vm *vm; + uint64_t guest_num_pages; + + pr_info("Testing guest mode: %s\n", vm_guest_mode_string(mode)); + + perf_test_args.host_page_size = getpagesize(); + perf_test_args.guest_page_size = vm_guest_mode_params[mode].page_size; + + guest_num_pages = vm_adjust_num_guest_pages(mode, + (vcpus * vcpu_memory_bytes) / perf_test_args.guest_page_size); + + TEST_ASSERT(vcpu_memory_bytes % perf_test_args.host_page_size == 0, + "Guest memory size is not host page size aligned."); + TEST_ASSERT(vcpu_memory_bytes % perf_test_args.guest_page_size == 0, + "Guest memory size is not guest page size aligned."); + + vm = vm_create_with_vcpus(mode, vcpus, + (vcpus * vcpu_memory_bytes) / perf_test_args.guest_page_size, + 0, guest_code, NULL); + + perf_test_args.vm = vm; + + /* + * If there should be more memory in the guest test region than there + * can be pages in the guest, it will definitely cause problems. + */ + TEST_ASSERT(guest_num_pages < vm_get_max_gfn(vm), + "Requested more guest memory than address space allows.\n" + " guest pages: %lx max gfn: %x vcpus: %d wss: %lx]\n", + guest_num_pages, vm_get_max_gfn(vm), vcpus, + vcpu_memory_bytes); + + guest_test_phys_mem = (vm_get_max_gfn(vm) - guest_num_pages) * + perf_test_args.guest_page_size; + guest_test_phys_mem &= ~(perf_test_args.host_page_size - 1); +#ifdef __s390x__ + /* Align to 1M (segment size) */ + guest_test_phys_mem &= ~((1 << 20) - 1); +#endif + pr_info("guest physical test memory offset: 0x%lx\n", guest_test_phys_mem); + + /* Add an extra memory slot for testing */ + vm_userspace_mem_region_add(vm, VM_MEM_SRC_ANONYMOUS, + guest_test_phys_mem, + PERF_TEST_MEM_SLOT_INDEX, + guest_num_pages, 0); + + /* Do mapping for the demand paging memory slot */ + virt_map(vm, guest_test_virt_mem, guest_test_phys_mem, guest_num_pages, 0); + + ucall_init(vm, NULL); + + return vm; +} + +void perf_test_destroy_vm(struct kvm_vm *vm) +{ + ucall_uninit(vm); + kvm_vm_free(vm); +} + +void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus, uint64_t vcpu_memory_bytes) +{ + vm_paddr_t vcpu_gpa; + struct perf_test_vcpu_args *vcpu_args; + int vcpu_id; + + for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) { + vcpu_args = &perf_test_args.vcpu_args[vcpu_id]; + + vcpu_args->vcpu_id = vcpu_id; + vcpu_args->gva = guest_test_virt_mem + + (vcpu_id * vcpu_memory_bytes); + vcpu_args->pages = vcpu_memory_bytes / + perf_test_args.guest_page_size; + + vcpu_gpa = guest_test_phys_mem + (vcpu_id * vcpu_memory_bytes); + pr_debug("Added VCPU %d with test mem gpa [%lx, %lx)\n", + vcpu_id, vcpu_gpa, vcpu_gpa + vcpu_memory_bytes); + } +} -- cgit v1.2.3-59-g8ed1b From a5c9ca76a1c61fb5e4c35de8eb25aa925b03c9e4 Mon Sep 17 00:00:00 2001 From: Ido Schimmel Date: Thu, 7 Jan 2021 16:48:24 +0200 Subject: selftests: fib_nexthops: Fix wrong mausezahn invocation For IPv6 traffic, mausezahn needs to be invoked with '-6'. Otherwise an error is returned: # ip netns exec me mausezahn veth1 -B 2001:db8:101::2 -A 2001:db8:91::1 -c 0 -t tcp "dp=1-1023, flags=syn" Failed to set source IPv4 address. Please check if source is set to a valid IPv4 address. Invalid command line parameters! Fixes: 7c741868ceab ("selftests: Add torture tests to nexthop tests") Signed-off-by: Ido Schimmel Reviewed-by: Petr Machata Reviewed-by: David Ahern Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/fib_nexthops.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/net/fib_nexthops.sh b/tools/testing/selftests/net/fib_nexthops.sh index eb693a3b7b4a..4c7d33618437 100755 --- a/tools/testing/selftests/net/fib_nexthops.sh +++ b/tools/testing/selftests/net/fib_nexthops.sh @@ -869,7 +869,7 @@ ipv6_torture() pid3=$! ip netns exec me ping -f 2001:db8:101::2 >/dev/null 2>&1 & pid4=$! - ip netns exec me mausezahn veth1 -B 2001:db8:101::2 -A 2001:db8:91::1 -c 0 -t tcp "dp=1-1023, flags=syn" >/dev/null 2>&1 & + ip netns exec me mausezahn -6 veth1 -B 2001:db8:101::2 -A 2001:db8:91::1 -c 0 -t tcp "dp=1-1023, flags=syn" >/dev/null 2>&1 & pid5=$! sleep 300 -- cgit v1.2.3-59-g8ed1b From 3502bd9b5762154ff11665f3f18f6d7dcc6f781c Mon Sep 17 00:00:00 2001 From: Vadim Fedorenko Date: Sat, 9 Jan 2021 00:37:45 +0300 Subject: selftests/tls: fix selftests after adding ChaCha20-Poly1305 TLS selftests where broken because of wrong variable types used. Fix it by changing u16 -> uint16_t Fixes: 4f336e88a870 ("selftests/tls: add CHACHA20-POLY1305 to tls selftests") Reported-by: kernel test robot Signed-off-by: Vadim Fedorenko Link: https://lore.kernel.org/r/1610141865-7142-1-git-send-email-vfedorenko@novek.ru Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/tls.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/net/tls.c b/tools/testing/selftests/net/tls.c index cb0d1890a860..e0088c2d38a5 100644 --- a/tools/testing/selftests/net/tls.c +++ b/tools/testing/selftests/net/tls.c @@ -103,8 +103,8 @@ FIXTURE(tls) FIXTURE_VARIANT(tls) { - u16 tls_version; - u16 cipher_type; + uint16_t tls_version; + uint16_t cipher_type; }; FIXTURE_VARIANT_ADD(tls, 12_gcm) -- cgit v1.2.3-59-g8ed1b From fab336b42441e0b2eb1d81becedb45fbdf99606e Mon Sep 17 00:00:00 2001 From: Chen Yi Date: Tue, 5 Jan 2021 23:31:20 +0800 Subject: selftests: netfilter: Pass family parameter "-f" to conntrack tool Fix nft_conntrack_helper.sh false fail report: 1) Conntrack tool need "-f ipv6" parameter to show out ipv6 traffic items. 2) Sleep 1 second after background nc send packet, to make sure check is after this statement executed. False report: FAIL: ns1-lkjUemYw did not show attached helper ip set via ruleset PASS: ns1-lkjUemYw connection on port 2121 has ftp helper attached ... After fix: PASS: ns1-2hUniwU2 connection on port 2121 has ftp helper attached PASS: ns2-2hUniwU2 connection on port 2121 has ftp helper attached ... Fixes: 619ae8e0697a6 ("selftests: netfilter: add test case for conntrack helper assignment") Signed-off-by: Chen Yi Signed-off-by: Pablo Neira Ayuso --- tools/testing/selftests/netfilter/nft_conntrack_helper.sh | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/netfilter/nft_conntrack_helper.sh b/tools/testing/selftests/netfilter/nft_conntrack_helper.sh index edf0a48da6bf..bf6b9626c7dd 100755 --- a/tools/testing/selftests/netfilter/nft_conntrack_helper.sh +++ b/tools/testing/selftests/netfilter/nft_conntrack_helper.sh @@ -94,7 +94,13 @@ check_for_helper() local message=$2 local port=$3 - ip netns exec ${netns} conntrack -L -p tcp --dport $port 2> /dev/null |grep -q 'helper=ftp' + if echo $message |grep -q 'ipv6';then + local family="ipv6" + else + local family="ipv4" + fi + + ip netns exec ${netns} conntrack -L -f $family -p tcp --dport $port 2> /dev/null |grep -q 'helper=ftp' if [ $? -ne 0 ] ; then echo "FAIL: ${netns} did not show attached helper $message" 1>&2 ret=1 @@ -111,8 +117,8 @@ test_helper() sleep 3 | ip netns exec ${ns2} nc -w 2 -l -p $port > /dev/null & - sleep 1 sleep 1 | ip netns exec ${ns1} nc -w 2 10.0.1.2 $port > /dev/null & + sleep 1 check_for_helper "$ns1" "ip $msg" $port check_for_helper "$ns2" "ip $msg" $port @@ -128,8 +134,8 @@ test_helper() sleep 3 | ip netns exec ${ns2} nc -w 2 -6 -l -p $port > /dev/null & - sleep 1 sleep 1 | ip netns exec ${ns1} nc -w 2 -6 dead:1::2 $port > /dev/null & + sleep 1 check_for_helper "$ns1" "ipv6 $msg" $port check_for_helper "$ns2" "ipv6 $msg" $port -- cgit v1.2.3-59-g8ed1b From 55ed4560774d81d7343223b8fd2784c530a9c6c1 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Wed, 9 Dec 2020 14:27:44 +0900 Subject: tools/bootconfig: Add tracing_on support to helper scripts Add ftrace.instance.INSTANCE.tracing_on support to ftrace2bconf.sh and bconf2ftrace.sh. commit 8490db06f914 ("tracing/boot: Add per-instance tracing_on option support") added the per-instance tracing_on option, but forgot to update the helper scripts. Link: https://lkml.kernel.org/r/160749166410.3497930.14204335886811029800.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 8490db06f914 ("tracing/boot: Add per-instance tracing_on option support") Signed-off-by: Masami Hiramatsu Signed-off-by: Steven Rostedt (VMware) --- tools/bootconfig/scripts/bconf2ftrace.sh | 1 + tools/bootconfig/scripts/ftrace2bconf.sh | 4 ++++ 2 files changed, 5 insertions(+) (limited to 'tools') diff --git a/tools/bootconfig/scripts/bconf2ftrace.sh b/tools/bootconfig/scripts/bconf2ftrace.sh index 595e164dc352..feb30c2c7881 100755 --- a/tools/bootconfig/scripts/bconf2ftrace.sh +++ b/tools/bootconfig/scripts/bconf2ftrace.sh @@ -152,6 +152,7 @@ setup_instance() { # [instance] set_array_of ${instance}.options ${instancedir}/trace_options set_value_of ${instance}.trace_clock ${instancedir}/trace_clock set_value_of ${instance}.cpumask ${instancedir}/tracing_cpumask + set_value_of ${instance}.tracing_on ${instancedir}/tracing_on set_value_of ${instance}.tracer ${instancedir}/current_tracer set_array_of ${instance}.ftrace.filters \ ${instancedir}/set_ftrace_filter diff --git a/tools/bootconfig/scripts/ftrace2bconf.sh b/tools/bootconfig/scripts/ftrace2bconf.sh index 6c0d4b61e0c2..a0c3bcc6da4f 100755 --- a/tools/bootconfig/scripts/ftrace2bconf.sh +++ b/tools/bootconfig/scripts/ftrace2bconf.sh @@ -221,6 +221,10 @@ instance_options() { # [instance-name] if [ `echo $val | sed -e s/f//g`x != x ]; then emit_kv $PREFIX.cpumask = $val fi + val=`cat $INSTANCE/tracing_on` + if [ `echo $val | sed -e s/f//g`x != x ]; then + emit_kv $PREFIX.tracing_on = $val + fi val= for i in `cat $INSTANCE/set_event`; do -- cgit v1.2.3-59-g8ed1b From 3a57a643a851dbb1c4a1819394ca009e3bfa4813 Mon Sep 17 00:00:00 2001 From: Mark Brown Date: Fri, 8 Jan 2021 18:31:44 +0000 Subject: arm64: selftests: Fix spelling of 'Mismatch' The SVE and FPSIMD stress tests have a spelling mistake in the output, fix it. Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20210108183144.673-1-broonie@kernel.org Signed-off-by: Catalin Marinas --- tools/testing/selftests/arm64/fp/fpsimd-test.S | 2 +- tools/testing/selftests/arm64/fp/sve-test.S | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/arm64/fp/fpsimd-test.S b/tools/testing/selftests/arm64/fp/fpsimd-test.S index 1c5556bdd11d..0dbd594c2747 100644 --- a/tools/testing/selftests/arm64/fp/fpsimd-test.S +++ b/tools/testing/selftests/arm64/fp/fpsimd-test.S @@ -457,7 +457,7 @@ function barf mov x11, x1 // actual data mov x12, x2 // data size - puts "Mistatch: PID=" + puts "Mismatch: PID=" mov x0, x20 bl putdec puts ", iteration=" diff --git a/tools/testing/selftests/arm64/fp/sve-test.S b/tools/testing/selftests/arm64/fp/sve-test.S index f95074c9b48b..9210691aa998 100644 --- a/tools/testing/selftests/arm64/fp/sve-test.S +++ b/tools/testing/selftests/arm64/fp/sve-test.S @@ -625,7 +625,7 @@ function barf mov x11, x1 // actual data mov x12, x2 // data size - puts "Mistatch: PID=" + puts "Mismatch: PID=" mov x0, x20 bl putdec puts ", iteration=" -- cgit v1.2.3-59-g8ed1b From 301f0203e04293c13372c032198665bd75adf81b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 29 Dec 2020 15:41:19 -0300 Subject: perf bpf examples: Fix bpf.h header include directive in 5sec.c example It was looking at bpf/bpf.h, which caused this problem: # perf trace -e tools/perf/examples/bpf/5sec.c /home/acme/git/perf/tools/perf/examples/bpf/5sec.c:42:10: fatal error: 'bpf/bpf.h' file not found #include ^~~~~~~~~~~ 1 error generated. ERROR: unable to compile tools/perf/examples/bpf/5sec.c Hint: Check error message shown above. Hint: You can also pre-compile it into .o using: clang -target bpf -O2 -c tools/perf/examples/bpf/5sec.c with proper -I and -D options. event syntax error: 'tools/perf/examples/bpf/5sec.c' \___ Failed to load tools/perf/examples/bpf/5sec.c from source: Error when compiling BPF scriptlet # Change that to plain bpf.h, to make it work again: # perf trace -e tools/perf/examples/bpf/5sec.c sleep 5s 0.000 perf_bpf_probe:hrtimer_nanosleep(__probe_ip: -1776891872, rqtp: 5000000000) # perf trace -e tools/perf/examples/bpf/5sec.c/max-stack=16/ sleep 5s 0.000 perf_bpf_probe:hrtimer_nanosleep(__probe_ip: -1776891872, rqtp: 5000000000) hrtimer_nanosleep ([kernel.kallsyms]) common_nsleep ([kernel.kallsyms]) __x64_sys_clock_nanosleep ([kernel.kallsyms]) do_syscall_64 ([kernel.kallsyms]) entry_SYSCALL_64_after_hwframe ([kernel.kallsyms]) __clock_nanosleep_2 (/usr/lib64/libc-2.32.so) # perf trace -e tools/perf/examples/bpf/5sec.c sleep 4s # Cc: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Cc: Namhyung Kim Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/examples/bpf/5sec.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/examples/bpf/5sec.c b/tools/perf/examples/bpf/5sec.c index 65c4ff6892d9..e6b6181c6dc6 100644 --- a/tools/perf/examples/bpf/5sec.c +++ b/tools/perf/examples/bpf/5sec.c @@ -39,7 +39,7 @@ Copyright (C) 2018 Red Hat, Inc., Arnaldo Carvalho de Melo */ -#include +#include #define NSEC_PER_SEC 1000000000L -- cgit v1.2.3-59-g8ed1b From 38c53947a7dcb6d295769830c9085b0409921ec9 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2021 12:26:08 -0300 Subject: tools headers UAPI: Sync kvm.h headers with the kernel sources To pick the changes in: 647daca25d24fb6e ("KVM: SVM: Add support for booting APs in an SEV-ES guest") That don't cause any tooling change, just silences this perf build warning: Warning: Kernel ABI header at 'tools/include/uapi/linux/kvm.h' differs from latest version at 'include/uapi/linux/kvm.h' diff -u tools/include/uapi/linux/kvm.h include/uapi/linux/kvm.h Cc: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paolo Bonzini Cc: Tom Lendacky Signed-off-by: Arnaldo Carvalho de Melo --- tools/include/uapi/linux/kvm.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'tools') diff --git a/tools/include/uapi/linux/kvm.h b/tools/include/uapi/linux/kvm.h index 886802b8ffba..374c67875cdb 100644 --- a/tools/include/uapi/linux/kvm.h +++ b/tools/include/uapi/linux/kvm.h @@ -251,6 +251,7 @@ struct kvm_hyperv_exit { #define KVM_EXIT_X86_RDMSR 29 #define KVM_EXIT_X86_WRMSR 30 #define KVM_EXIT_DIRTY_RING_FULL 31 +#define KVM_EXIT_AP_RESET_HOLD 32 /* For KVM_EXIT_INTERNAL_ERROR */ /* Emulate instruction failed. */ @@ -573,6 +574,7 @@ struct kvm_vapic_addr { #define KVM_MP_STATE_CHECK_STOP 6 #define KVM_MP_STATE_OPERATING 7 #define KVM_MP_STATE_LOAD 8 +#define KVM_MP_STATE_AP_RESET_HOLD 9 struct kvm_mp_state { __u32 mp_state; -- cgit v1.2.3-59-g8ed1b From addbdff24293ef772a1b8e5d127b570e70f08cdc Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2021 12:35:27 -0300 Subject: tools headers: Syncronize linux/build_bug.h with the kernel sources To pick up the changes in: 3a176b94609a18f5 ("Revert "kbuild: avoid static_assert for genksyms"") And silence this perf build warning: Warning: Kernel ABI header at 'tools/include/linux/build_bug.h' differs from latest version at 'include/linux/build_bug.h' diff -u tools/include/linux/build_bug.h include/linux/build_bug.h Cc: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Cc: Masahiro Yamada Cc: Namhyung Kim Signed-off-by: Arnaldo Carvalho de Melo --- tools/include/linux/build_bug.h | 5 ----- 1 file changed, 5 deletions(-) (limited to 'tools') diff --git a/tools/include/linux/build_bug.h b/tools/include/linux/build_bug.h index ce365d212768..cc7070c7439b 100644 --- a/tools/include/linux/build_bug.h +++ b/tools/include/linux/build_bug.h @@ -79,9 +79,4 @@ #define __static_assert(expr, msg, ...) _Static_assert(expr, msg) #endif // static_assert -#ifdef __GENKSYMS__ -/* genksyms gets confused by _Static_assert */ -#define _Static_assert(expr, ...) -#endif - #endif /* _LINUX_BUILD_BUG_H */ -- cgit v1.2.3-59-g8ed1b From a042a82ddbb3434f523c0671f5301d1fe796b4eb Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 14 Jan 2021 14:06:09 +0900 Subject: perf test: Fix shadow stat test for non-bash shells It was using some bash-specific features and failed to parse when running with a different shell like below: root@kbl-ppc:~/kbl-ws/perf-dev/lck-9077/acme.tmp/tools/perf# ./perf test 83 -vv 83: perf stat metrics (shadow stat) test : --- start --- test child forked, pid 3922 ./tests/shell/stat+shadow_stat.sh: 19: ./tests/shell/stat+shadow_stat.sh: [[: not found ./tests/shell/stat+shadow_stat.sh: 24: ./tests/shell/stat+shadow_stat.sh: [[: not found ./tests/shell/stat+shadow_stat.sh: 30: ./tests/shell/stat+shadow_stat.sh: [[: not found (standard_in) 2: syntax error ./tests/shell/stat+shadow_stat.sh: 36: ./tests/shell/stat+shadow_stat.sh: [[: not found ./tests/shell/stat+shadow_stat.sh: 19: ./tests/shell/stat+shadow_stat.sh: [[: not found ./tests/shell/stat+shadow_stat.sh: 24: ./tests/shell/stat+shadow_stat.sh: [[: not found ./tests/shell/stat+shadow_stat.sh: 30: ./tests/shell/stat+shadow_stat.sh: [[: not found (standard_in) 2: syntax error ./tests/shell/stat+shadow_stat.sh: 36: ./tests/shell/stat+shadow_stat.sh: [[: not found ./tests/shell/stat+shadow_stat.sh: 45: ./tests/shell/stat+shadow_stat.sh: declare: not found test child finished with -1 ---- end ---- perf stat metrics (shadow stat) test: FAILED! Reported-by: Jin Yao Signed-off-by: Namhyung Kim Acked-by: Jiri Olsa Cc: Alexander Shishkin Cc: Andi Kleen Cc: David Laight Cc: Ian Rogers Cc: Mark Rutland Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lore.kernel.org/lkml/20210114050609.1258820-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/tests/shell/stat+shadow_stat.sh | 30 ++++++++++++++---------------- 1 file changed, 14 insertions(+), 16 deletions(-) (limited to 'tools') diff --git a/tools/perf/tests/shell/stat+shadow_stat.sh b/tools/perf/tests/shell/stat+shadow_stat.sh index 249dfe48cf6a..ebebd3596cf9 100755 --- a/tools/perf/tests/shell/stat+shadow_stat.sh +++ b/tools/perf/tests/shell/stat+shadow_stat.sh @@ -9,31 +9,29 @@ perf stat -a true > /dev/null 2>&1 || exit 2 test_global_aggr() { - local cyc - perf stat -a --no-big-num -e cycles,instructions sleep 1 2>&1 | \ grep -e cycles -e instructions | \ while read num evt hash ipc rest do # skip not counted events - if [[ $num == "&1 | \ grep ^CPU | \ while read cpu num evt hash ipc rest do # skip not counted events - if [[ $num == " Date: Thu, 14 Jan 2021 13:23:04 -0800 Subject: libperf tests: Avoid uninitialized variable warning The variable 'bf' is read (for a write call) without being initialized triggering a memory sanitizer warning. Use 'bf' in the read and switch the write to reading from a string. Signed-off-by: Ian Rogers Acked-by: Jiri Olsa Acked-by: Namhyung Kim Cc: Alexander Shishkin Cc: Mark Rutland Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lore.kernel.org/lkml/20210114212304.4018119-1-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/lib/perf/tests/test-evlist.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/lib/perf/tests/test-evlist.c b/tools/lib/perf/tests/test-evlist.c index 6d8ebe0c2504..1b225fe34a72 100644 --- a/tools/lib/perf/tests/test-evlist.c +++ b/tools/lib/perf/tests/test-evlist.c @@ -208,7 +208,6 @@ static int test_mmap_thread(void) char path[PATH_MAX]; int id, err, pid, go_pipe[2]; union perf_event *event; - char bf; int count = 0; snprintf(path, PATH_MAX, "%s/kernel/debug/tracing/events/syscalls/sys_enter_prctl/id", @@ -229,6 +228,7 @@ static int test_mmap_thread(void) pid = fork(); if (!pid) { int i; + char bf; read(go_pipe[0], &bf, 1); @@ -266,7 +266,7 @@ static int test_mmap_thread(void) perf_evlist__enable(evlist); /* kick the child and wait for it to finish */ - write(go_pipe[1], &bf, 1); + write(go_pipe[1], "A", 1); waitpid(pid, NULL, 0); /* -- cgit v1.2.3-59-g8ed1b From bba2ea17ef553aea0df80cb64399fe2f70f225dd Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 14 Jan 2021 10:02:49 -0800 Subject: libperf tests: If a test fails return non-zero If a test fails return -1 rather than 0. This is consistent with the return value in test-cpumap.c Signed-off-by: Ian Rogers Acked-by: Jiri Olsa Cc: Alexander Shishkin Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lore.kernel.org/lkml/20210114180250.3853825-1-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/lib/perf/tests/test-cpumap.c | 2 +- tools/lib/perf/tests/test-evlist.c | 2 +- tools/lib/perf/tests/test-evsel.c | 2 +- tools/lib/perf/tests/test-threadmap.c | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/lib/perf/tests/test-cpumap.c b/tools/lib/perf/tests/test-cpumap.c index c8d45091e7c2..c70e9e03af3e 100644 --- a/tools/lib/perf/tests/test-cpumap.c +++ b/tools/lib/perf/tests/test-cpumap.c @@ -27,5 +27,5 @@ int main(int argc, char **argv) perf_cpu_map__put(cpus); __T_END; - return 0; + return tests_failed == 0 ? 0 : -1; } diff --git a/tools/lib/perf/tests/test-evlist.c b/tools/lib/perf/tests/test-evlist.c index 1b225fe34a72..220a95be47c3 100644 --- a/tools/lib/perf/tests/test-evlist.c +++ b/tools/lib/perf/tests/test-evlist.c @@ -409,5 +409,5 @@ int main(int argc, char **argv) test_mmap_cpus(); __T_END; - return 0; + return tests_failed == 0 ? 0 : -1; } diff --git a/tools/lib/perf/tests/test-evsel.c b/tools/lib/perf/tests/test-evsel.c index 135722ac965b..0ad82d7a2a51 100644 --- a/tools/lib/perf/tests/test-evsel.c +++ b/tools/lib/perf/tests/test-evsel.c @@ -131,5 +131,5 @@ int main(int argc, char **argv) test_stat_thread_enable(); __T_END; - return 0; + return tests_failed == 0 ? 0 : -1; } diff --git a/tools/lib/perf/tests/test-threadmap.c b/tools/lib/perf/tests/test-threadmap.c index 7dc4d6fbedde..384471441b48 100644 --- a/tools/lib/perf/tests/test-threadmap.c +++ b/tools/lib/perf/tests/test-threadmap.c @@ -27,5 +27,5 @@ int main(int argc, char **argv) perf_thread_map__put(threads); __T_END; - return 0; + return tests_failed == 0 ? 0 : -1; } -- cgit v1.2.3-59-g8ed1b From 66dd86b2a2bee129c70f7ff054d3a6a2e5f8eb20 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Thu, 14 Jan 2021 10:02:50 -0800 Subject: libperf tests: Fail when failing to get a tracepoint id Permissions are necessary to get a tracepoint id. Fail the test when the read fails. Signed-off-by: Ian Rogers Acked-by: Jiri Olsa Cc: Alexander Shishkin Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lore.kernel.org/lkml/20210114180250.3853825-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/lib/perf/tests/test-evlist.c | 1 + 1 file changed, 1 insertion(+) (limited to 'tools') diff --git a/tools/lib/perf/tests/test-evlist.c b/tools/lib/perf/tests/test-evlist.c index 220a95be47c3..e2ac0b7f432e 100644 --- a/tools/lib/perf/tests/test-evlist.c +++ b/tools/lib/perf/tests/test-evlist.c @@ -214,6 +214,7 @@ static int test_mmap_thread(void) sysfs__mountpoint()); if (filename__read_int(path, &id)) { + tests_failed++; fprintf(stderr, "error: failed to get tracepoint id: %s\n", path); return -1; } -- cgit v1.2.3-59-g8ed1b From 3ff1e7180abc7f6db413933c110df69157216715 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 15 Jan 2021 16:11:38 +0900 Subject: perf stat: Introduce struct runtime_stat_data To pass more info to the saved_value in the runtime_stat, add a new struct runtime_stat_data. Currently it only has 'ctx' field but later patch will add more. Note that we intentionally pass 0 as ctx to clock-related events for compatibility. It was already there in a few places. So move the code into the saved_value_lookup() explicitly and add a comment. Suggested-by: Andi Kleen Signed-off-by: Namhyung Kim Acked-by: Jiri Olsa Cc: Alexander Shishkin Cc: Andi Kleen Cc: Ian Rogers Cc: Jin Yao Cc: Mark Rutland Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lore.kernel.org/lkml/20210115071139.257042-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/stat-shadow.c | 346 +++++++++++++++++++++--------------------- 1 file changed, 173 insertions(+), 173 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/stat-shadow.c b/tools/perf/util/stat-shadow.c index 901265127e36..a1565b6e38f2 100644 --- a/tools/perf/util/stat-shadow.c +++ b/tools/perf/util/stat-shadow.c @@ -114,6 +114,10 @@ static struct saved_value *saved_value_lookup(struct evsel *evsel, rblist = &st->value_list; + /* don't use context info for clock events */ + if (type == STAT_NSECS) + dm.ctx = 0; + nd = rblist__find(rblist, &dm); if (nd) return container_of(nd, struct saved_value, rb_node); @@ -191,12 +195,17 @@ void perf_stat__reset_shadow_per_stat(struct runtime_stat *st) reset_stat(st); } +struct runtime_stat_data { + int ctx; +}; + static void update_runtime_stat(struct runtime_stat *st, enum stat_type type, - int ctx, int cpu, u64 count) + int cpu, u64 count, + struct runtime_stat_data *rsd) { - struct saved_value *v = saved_value_lookup(NULL, cpu, true, - type, ctx, st); + struct saved_value *v = saved_value_lookup(NULL, cpu, true, type, + rsd->ctx, st); if (v) update_stats(&v->stats, count); @@ -210,73 +219,75 @@ static void update_runtime_stat(struct runtime_stat *st, void perf_stat__update_shadow_stats(struct evsel *counter, u64 count, int cpu, struct runtime_stat *st) { - int ctx = evsel_context(counter); u64 count_ns = count; struct saved_value *v; + struct runtime_stat_data rsd = { + .ctx = evsel_context(counter), + }; count *= counter->scale; if (evsel__is_clock(counter)) - update_runtime_stat(st, STAT_NSECS, 0, cpu, count_ns); + update_runtime_stat(st, STAT_NSECS, cpu, count_ns, &rsd); else if (evsel__match(counter, HARDWARE, HW_CPU_CYCLES)) - update_runtime_stat(st, STAT_CYCLES, ctx, cpu, count); + update_runtime_stat(st, STAT_CYCLES, cpu, count, &rsd); else if (perf_stat_evsel__is(counter, CYCLES_IN_TX)) - update_runtime_stat(st, STAT_CYCLES_IN_TX, ctx, cpu, count); + update_runtime_stat(st, STAT_CYCLES_IN_TX, cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TRANSACTION_START)) - update_runtime_stat(st, STAT_TRANSACTION, ctx, cpu, count); + update_runtime_stat(st, STAT_TRANSACTION, cpu, count, &rsd); else if (perf_stat_evsel__is(counter, ELISION_START)) - update_runtime_stat(st, STAT_ELISION, ctx, cpu, count); + update_runtime_stat(st, STAT_ELISION, cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TOPDOWN_TOTAL_SLOTS)) update_runtime_stat(st, STAT_TOPDOWN_TOTAL_SLOTS, - ctx, cpu, count); + cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TOPDOWN_SLOTS_ISSUED)) update_runtime_stat(st, STAT_TOPDOWN_SLOTS_ISSUED, - ctx, cpu, count); + cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TOPDOWN_SLOTS_RETIRED)) update_runtime_stat(st, STAT_TOPDOWN_SLOTS_RETIRED, - ctx, cpu, count); + cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TOPDOWN_FETCH_BUBBLES)) update_runtime_stat(st, STAT_TOPDOWN_FETCH_BUBBLES, - ctx, cpu, count); + cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TOPDOWN_RECOVERY_BUBBLES)) update_runtime_stat(st, STAT_TOPDOWN_RECOVERY_BUBBLES, - ctx, cpu, count); + cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TOPDOWN_RETIRING)) update_runtime_stat(st, STAT_TOPDOWN_RETIRING, - ctx, cpu, count); + cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TOPDOWN_BAD_SPEC)) update_runtime_stat(st, STAT_TOPDOWN_BAD_SPEC, - ctx, cpu, count); + cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TOPDOWN_FE_BOUND)) update_runtime_stat(st, STAT_TOPDOWN_FE_BOUND, - ctx, cpu, count); + cpu, count, &rsd); else if (perf_stat_evsel__is(counter, TOPDOWN_BE_BOUND)) update_runtime_stat(st, STAT_TOPDOWN_BE_BOUND, - ctx, cpu, count); + cpu, count, &rsd); else if (evsel__match(counter, HARDWARE, HW_STALLED_CYCLES_FRONTEND)) update_runtime_stat(st, STAT_STALLED_CYCLES_FRONT, - ctx, cpu, count); + cpu, count, &rsd); else if (evsel__match(counter, HARDWARE, HW_STALLED_CYCLES_BACKEND)) update_runtime_stat(st, STAT_STALLED_CYCLES_BACK, - ctx, cpu, count); + cpu, count, &rsd); else if (evsel__match(counter, HARDWARE, HW_BRANCH_INSTRUCTIONS)) - update_runtime_stat(st, STAT_BRANCHES, ctx, cpu, count); + update_runtime_stat(st, STAT_BRANCHES, cpu, count, &rsd); else if (evsel__match(counter, HARDWARE, HW_CACHE_REFERENCES)) - update_runtime_stat(st, STAT_CACHEREFS, ctx, cpu, count); + update_runtime_stat(st, STAT_CACHEREFS, cpu, count, &rsd); else if (evsel__match(counter, HW_CACHE, HW_CACHE_L1D)) - update_runtime_stat(st, STAT_L1_DCACHE, ctx, cpu, count); + update_runtime_stat(st, STAT_L1_DCACHE, cpu, count, &rsd); else if (evsel__match(counter, HW_CACHE, HW_CACHE_L1I)) - update_runtime_stat(st, STAT_L1_ICACHE, ctx, cpu, count); + update_runtime_stat(st, STAT_L1_ICACHE, cpu, count, &rsd); else if (evsel__match(counter, HW_CACHE, HW_CACHE_LL)) - update_runtime_stat(st, STAT_LL_CACHE, ctx, cpu, count); + update_runtime_stat(st, STAT_LL_CACHE, cpu, count, &rsd); else if (evsel__match(counter, HW_CACHE, HW_CACHE_DTLB)) - update_runtime_stat(st, STAT_DTLB_CACHE, ctx, cpu, count); + update_runtime_stat(st, STAT_DTLB_CACHE, cpu, count, &rsd); else if (evsel__match(counter, HW_CACHE, HW_CACHE_ITLB)) - update_runtime_stat(st, STAT_ITLB_CACHE, ctx, cpu, count); + update_runtime_stat(st, STAT_ITLB_CACHE, cpu, count, &rsd); else if (perf_stat_evsel__is(counter, SMI_NUM)) - update_runtime_stat(st, STAT_SMI_NUM, ctx, cpu, count); + update_runtime_stat(st, STAT_SMI_NUM, cpu, count, &rsd); else if (perf_stat_evsel__is(counter, APERF)) - update_runtime_stat(st, STAT_APERF, ctx, cpu, count); + update_runtime_stat(st, STAT_APERF, cpu, count, &rsd); if (counter->collect_stat) { v = saved_value_lookup(counter, cpu, true, STAT_NONE, 0, st); @@ -422,11 +433,12 @@ void perf_stat__collect_metric_expr(struct evlist *evsel_list) } static double runtime_stat_avg(struct runtime_stat *st, - enum stat_type type, int ctx, int cpu) + enum stat_type type, int cpu, + struct runtime_stat_data *rsd) { struct saved_value *v; - v = saved_value_lookup(NULL, cpu, false, type, ctx, st); + v = saved_value_lookup(NULL, cpu, false, type, rsd->ctx, st); if (!v) return 0.0; @@ -434,11 +446,12 @@ static double runtime_stat_avg(struct runtime_stat *st, } static double runtime_stat_n(struct runtime_stat *st, - enum stat_type type, int ctx, int cpu) + enum stat_type type, int cpu, + struct runtime_stat_data *rsd) { struct saved_value *v; - v = saved_value_lookup(NULL, cpu, false, type, ctx, st); + v = saved_value_lookup(NULL, cpu, false, type, rsd->ctx, st); if (!v) return 0.0; @@ -446,16 +459,15 @@ static double runtime_stat_n(struct runtime_stat *st, } static void print_stalled_cycles_frontend(struct perf_stat_config *config, - int cpu, - struct evsel *evsel, double avg, + int cpu, double avg, struct perf_stat_output_ctx *out, - struct runtime_stat *st) + struct runtime_stat *st, + struct runtime_stat_data *rsd) { double total, ratio = 0.0; const char *color; - int ctx = evsel_context(evsel); - total = runtime_stat_avg(st, STAT_CYCLES, ctx, cpu); + total = runtime_stat_avg(st, STAT_CYCLES, cpu, rsd); if (total) ratio = avg / total * 100.0; @@ -470,16 +482,15 @@ static void print_stalled_cycles_frontend(struct perf_stat_config *config, } static void print_stalled_cycles_backend(struct perf_stat_config *config, - int cpu, - struct evsel *evsel, double avg, + int cpu, double avg, struct perf_stat_output_ctx *out, - struct runtime_stat *st) + struct runtime_stat *st, + struct runtime_stat_data *rsd) { double total, ratio = 0.0; const char *color; - int ctx = evsel_context(evsel); - total = runtime_stat_avg(st, STAT_CYCLES, ctx, cpu); + total = runtime_stat_avg(st, STAT_CYCLES, cpu, rsd); if (total) ratio = avg / total * 100.0; @@ -490,17 +501,15 @@ static void print_stalled_cycles_backend(struct perf_stat_config *config, } static void print_branch_misses(struct perf_stat_config *config, - int cpu, - struct evsel *evsel, - double avg, + int cpu, double avg, struct perf_stat_output_ctx *out, - struct runtime_stat *st) + struct runtime_stat *st, + struct runtime_stat_data *rsd) { double total, ratio = 0.0; const char *color; - int ctx = evsel_context(evsel); - total = runtime_stat_avg(st, STAT_BRANCHES, ctx, cpu); + total = runtime_stat_avg(st, STAT_BRANCHES, cpu, rsd); if (total) ratio = avg / total * 100.0; @@ -511,18 +520,15 @@ static void print_branch_misses(struct perf_stat_config *config, } static void print_l1_dcache_misses(struct perf_stat_config *config, - int cpu, - struct evsel *evsel, - double avg, + int cpu, double avg, struct perf_stat_output_ctx *out, - struct runtime_stat *st) - + struct runtime_stat *st, + struct runtime_stat_data *rsd) { double total, ratio = 0.0; const char *color; - int ctx = evsel_context(evsel); - total = runtime_stat_avg(st, STAT_L1_DCACHE, ctx, cpu); + total = runtime_stat_avg(st, STAT_L1_DCACHE, cpu, rsd); if (total) ratio = avg / total * 100.0; @@ -533,18 +539,15 @@ static void print_l1_dcache_misses(struct perf_stat_config *config, } static void print_l1_icache_misses(struct perf_stat_config *config, - int cpu, - struct evsel *evsel, - double avg, + int cpu, double avg, struct perf_stat_output_ctx *out, - struct runtime_stat *st) - + struct runtime_stat *st, + struct runtime_stat_data *rsd) { double total, ratio = 0.0; const char *color; - int ctx = evsel_context(evsel); - total = runtime_stat_avg(st, STAT_L1_ICACHE, ctx, cpu); + total = runtime_stat_avg(st, STAT_L1_ICACHE, cpu, rsd); if (total) ratio = avg / total * 100.0; @@ -554,17 +557,15 @@ static void print_l1_icache_misses(struct perf_stat_config *config, } static void print_dtlb_cache_misses(struct perf_stat_config *config, - int cpu, - struct evsel *evsel, - double avg, + int cpu, double avg, struct perf_stat_output_ctx *out, - struct runtime_stat *st) + struct runtime_stat *st, + struct runtime_stat_data *rsd) { double total, ratio = 0.0; const char *color; - int ctx = evsel_context(evsel); - total = runtime_stat_avg(st, STAT_DTLB_CACHE, ctx, cpu); + total = runtime_stat_avg(st, STAT_DTLB_CACHE, cpu, rsd); if (total) ratio = avg / total * 100.0; @@ -574,17 +575,15 @@ static void print_dtlb_cache_misses(struct perf_stat_config *config, } static void print_itlb_cache_misses(struct perf_stat_config *config, - int cpu, - struct evsel *evsel, - double avg, + int cpu, double avg, struct perf_stat_output_ctx *out, - struct runtime_stat *st) + struct runtime_stat *st, + struct runtime_stat_data *rsd) { double total, ratio = 0.0; const char *color; - int ctx = evsel_context(evsel); - total = runtime_stat_avg(st, STAT_ITLB_CACHE, ctx, cpu); + total = runtime_stat_avg(st, STAT_ITLB_CACHE, cpu, rsd); if (total) ratio = avg / total * 100.0; @@ -594,17 +593,15 @@ static void print_itlb_cache_misses(struct perf_stat_config *config, } static void print_ll_cache_misses(struct perf_stat_config *config, - int cpu, - struct evsel *evsel, - double avg, + int cpu, double avg, struct perf_stat_output_ctx *out, - struct runtime_stat *st) + struct runtime_stat *st, + struct runtime_stat_data *rsd) { double total, ratio = 0.0; const char *color; - int ctx = evsel_context(evsel); - total = runtime_stat_avg(st, STAT_LL_CACHE, ctx, cpu); + total = runtime_stat_avg(st, STAT_LL_CACHE, cpu, rsd); if (total) ratio = avg / total * 100.0; @@ -662,56 +659,61 @@ static double sanitize_val(double x) return x; } -static double td_total_slots(int ctx, int cpu, struct runtime_stat *st) +static double td_total_slots(int cpu, struct runtime_stat *st, + struct runtime_stat_data *rsd) { - return runtime_stat_avg(st, STAT_TOPDOWN_TOTAL_SLOTS, ctx, cpu); + return runtime_stat_avg(st, STAT_TOPDOWN_TOTAL_SLOTS, cpu, rsd); } -static double td_bad_spec(int ctx, int cpu, struct runtime_stat *st) +static double td_bad_spec(int cpu, struct runtime_stat *st, + struct runtime_stat_data *rsd) { double bad_spec = 0; double total_slots; double total; - total = runtime_stat_avg(st, STAT_TOPDOWN_SLOTS_ISSUED, ctx, cpu) - - runtime_stat_avg(st, STAT_TOPDOWN_SLOTS_RETIRED, ctx, cpu) + - runtime_stat_avg(st, STAT_TOPDOWN_RECOVERY_BUBBLES, ctx, cpu); + total = runtime_stat_avg(st, STAT_TOPDOWN_SLOTS_ISSUED, cpu, rsd) - + runtime_stat_avg(st, STAT_TOPDOWN_SLOTS_RETIRED, cpu, rsd) + + runtime_stat_avg(st, STAT_TOPDOWN_RECOVERY_BUBBLES, cpu, rsd); - total_slots = td_total_slots(ctx, cpu, st); + total_slots = td_total_slots(cpu, st, rsd); if (total_slots) bad_spec = total / total_slots; return sanitize_val(bad_spec); } -static double td_retiring(int ctx, int cpu, struct runtime_stat *st) +static double td_retiring(int cpu, struct runtime_stat *st, + struct runtime_stat_data *rsd) { double retiring = 0; - double total_slots = td_total_slots(ctx, cpu, st); + double total_slots = td_total_slots(cpu, st, rsd); double ret_slots = runtime_stat_avg(st, STAT_TOPDOWN_SLOTS_RETIRED, - ctx, cpu); + cpu, rsd); if (total_slots) retiring = ret_slots / total_slots; return retiring; } -static double td_fe_bound(int ctx, int cpu, struct runtime_stat *st) +static double td_fe_bound(int cpu, struct runtime_stat *st, + struct runtime_stat_data *rsd) { double fe_bound = 0; - double total_slots = td_total_slots(ctx, cpu, st); + double total_slots = td_total_slots(cpu, st, rsd); double fetch_bub = runtime_stat_avg(st, STAT_TOPDOWN_FETCH_BUBBLES, - ctx, cpu); + cpu, rsd); if (total_slots) fe_bound = fetch_bub / total_slots; return fe_bound; } -static double td_be_bound(int ctx, int cpu, struct runtime_stat *st) +static double td_be_bound(int cpu, struct runtime_stat *st, + struct runtime_stat_data *rsd) { - double sum = (td_fe_bound(ctx, cpu, st) + - td_bad_spec(ctx, cpu, st) + - td_retiring(ctx, cpu, st)); + double sum = (td_fe_bound(cpu, st, rsd) + + td_bad_spec(cpu, st, rsd) + + td_retiring(cpu, st, rsd)); if (sum == 0) return 0; return sanitize_val(1.0 - sum); @@ -722,15 +724,15 @@ static double td_be_bound(int ctx, int cpu, struct runtime_stat *st) * the ratios we need to recreate the sum. */ -static double td_metric_ratio(int ctx, int cpu, - enum stat_type type, - struct runtime_stat *stat) +static double td_metric_ratio(int cpu, enum stat_type type, + struct runtime_stat *stat, + struct runtime_stat_data *rsd) { - double sum = runtime_stat_avg(stat, STAT_TOPDOWN_RETIRING, ctx, cpu) + - runtime_stat_avg(stat, STAT_TOPDOWN_FE_BOUND, ctx, cpu) + - runtime_stat_avg(stat, STAT_TOPDOWN_BE_BOUND, ctx, cpu) + - runtime_stat_avg(stat, STAT_TOPDOWN_BAD_SPEC, ctx, cpu); - double d = runtime_stat_avg(stat, type, ctx, cpu); + double sum = runtime_stat_avg(stat, STAT_TOPDOWN_RETIRING, cpu, rsd) + + runtime_stat_avg(stat, STAT_TOPDOWN_FE_BOUND, cpu, rsd) + + runtime_stat_avg(stat, STAT_TOPDOWN_BE_BOUND, cpu, rsd) + + runtime_stat_avg(stat, STAT_TOPDOWN_BAD_SPEC, cpu, rsd); + double d = runtime_stat_avg(stat, type, cpu, rsd); if (sum) return d / sum; @@ -742,34 +744,33 @@ static double td_metric_ratio(int ctx, int cpu, * We allow two missing. */ -static bool full_td(int ctx, int cpu, - struct runtime_stat *stat) +static bool full_td(int cpu, struct runtime_stat *stat, + struct runtime_stat_data *rsd) { int c = 0; - if (runtime_stat_avg(stat, STAT_TOPDOWN_RETIRING, ctx, cpu) > 0) + if (runtime_stat_avg(stat, STAT_TOPDOWN_RETIRING, cpu, rsd) > 0) c++; - if (runtime_stat_avg(stat, STAT_TOPDOWN_BE_BOUND, ctx, cpu) > 0) + if (runtime_stat_avg(stat, STAT_TOPDOWN_BE_BOUND, cpu, rsd) > 0) c++; - if (runtime_stat_avg(stat, STAT_TOPDOWN_FE_BOUND, ctx, cpu) > 0) + if (runtime_stat_avg(stat, STAT_TOPDOWN_FE_BOUND, cpu, rsd) > 0) c++; - if (runtime_stat_avg(stat, STAT_TOPDOWN_BAD_SPEC, ctx, cpu) > 0) + if (runtime_stat_avg(stat, STAT_TOPDOWN_BAD_SPEC, cpu, rsd) > 0) c++; return c >= 2; } -static void print_smi_cost(struct perf_stat_config *config, - int cpu, struct evsel *evsel, +static void print_smi_cost(struct perf_stat_config *config, int cpu, struct perf_stat_output_ctx *out, - struct runtime_stat *st) + struct runtime_stat *st, + struct runtime_stat_data *rsd) { double smi_num, aperf, cycles, cost = 0.0; - int ctx = evsel_context(evsel); const char *color = NULL; - smi_num = runtime_stat_avg(st, STAT_SMI_NUM, ctx, cpu); - aperf = runtime_stat_avg(st, STAT_APERF, ctx, cpu); - cycles = runtime_stat_avg(st, STAT_CYCLES, ctx, cpu); + smi_num = runtime_stat_avg(st, STAT_SMI_NUM, cpu, rsd); + aperf = runtime_stat_avg(st, STAT_APERF, cpu, rsd); + cycles = runtime_stat_avg(st, STAT_CYCLES, cpu, rsd); if ((cycles == 0) || (aperf == 0)) return; @@ -930,12 +931,14 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, print_metric_t print_metric = out->print_metric; double total, ratio = 0.0, total2; const char *color = NULL; - int ctx = evsel_context(evsel); + struct runtime_stat_data rsd = { + .ctx = evsel_context(evsel), + }; struct metric_event *me; int num = 1; if (evsel__match(evsel, HARDWARE, HW_INSTRUCTIONS)) { - total = runtime_stat_avg(st, STAT_CYCLES, ctx, cpu); + total = runtime_stat_avg(st, STAT_CYCLES, cpu, &rsd); if (total) { ratio = avg / total; @@ -945,12 +948,11 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, print_metric(config, ctxp, NULL, NULL, "insn per cycle", 0); } - total = runtime_stat_avg(st, STAT_STALLED_CYCLES_FRONT, - ctx, cpu); + total = runtime_stat_avg(st, STAT_STALLED_CYCLES_FRONT, cpu, &rsd); total = max(total, runtime_stat_avg(st, STAT_STALLED_CYCLES_BACK, - ctx, cpu)); + cpu, &rsd)); if (total && avg) { out->new_line(config, ctxp); @@ -960,8 +962,8 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, ratio); } } else if (evsel__match(evsel, HARDWARE, HW_BRANCH_MISSES)) { - if (runtime_stat_n(st, STAT_BRANCHES, ctx, cpu) != 0) - print_branch_misses(config, cpu, evsel, avg, out, st); + if (runtime_stat_n(st, STAT_BRANCHES, cpu, &rsd) != 0) + print_branch_misses(config, cpu, avg, out, st, &rsd); else print_metric(config, ctxp, NULL, NULL, "of all branches", 0); } else if ( @@ -970,8 +972,8 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, ((PERF_COUNT_HW_CACHE_OP_READ) << 8) | ((PERF_COUNT_HW_CACHE_RESULT_MISS) << 16))) { - if (runtime_stat_n(st, STAT_L1_DCACHE, ctx, cpu) != 0) - print_l1_dcache_misses(config, cpu, evsel, avg, out, st); + if (runtime_stat_n(st, STAT_L1_DCACHE, cpu, &rsd) != 0) + print_l1_dcache_misses(config, cpu, avg, out, st, &rsd); else print_metric(config, ctxp, NULL, NULL, "of all L1-dcache accesses", 0); } else if ( @@ -980,8 +982,8 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, ((PERF_COUNT_HW_CACHE_OP_READ) << 8) | ((PERF_COUNT_HW_CACHE_RESULT_MISS) << 16))) { - if (runtime_stat_n(st, STAT_L1_ICACHE, ctx, cpu) != 0) - print_l1_icache_misses(config, cpu, evsel, avg, out, st); + if (runtime_stat_n(st, STAT_L1_ICACHE, cpu, &rsd) != 0) + print_l1_icache_misses(config, cpu, avg, out, st, &rsd); else print_metric(config, ctxp, NULL, NULL, "of all L1-icache accesses", 0); } else if ( @@ -990,8 +992,8 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, ((PERF_COUNT_HW_CACHE_OP_READ) << 8) | ((PERF_COUNT_HW_CACHE_RESULT_MISS) << 16))) { - if (runtime_stat_n(st, STAT_DTLB_CACHE, ctx, cpu) != 0) - print_dtlb_cache_misses(config, cpu, evsel, avg, out, st); + if (runtime_stat_n(st, STAT_DTLB_CACHE, cpu, &rsd) != 0) + print_dtlb_cache_misses(config, cpu, avg, out, st, &rsd); else print_metric(config, ctxp, NULL, NULL, "of all dTLB cache accesses", 0); } else if ( @@ -1000,8 +1002,8 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, ((PERF_COUNT_HW_CACHE_OP_READ) << 8) | ((PERF_COUNT_HW_CACHE_RESULT_MISS) << 16))) { - if (runtime_stat_n(st, STAT_ITLB_CACHE, ctx, cpu) != 0) - print_itlb_cache_misses(config, cpu, evsel, avg, out, st); + if (runtime_stat_n(st, STAT_ITLB_CACHE, cpu, &rsd) != 0) + print_itlb_cache_misses(config, cpu, avg, out, st, &rsd); else print_metric(config, ctxp, NULL, NULL, "of all iTLB cache accesses", 0); } else if ( @@ -1010,27 +1012,27 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, ((PERF_COUNT_HW_CACHE_OP_READ) << 8) | ((PERF_COUNT_HW_CACHE_RESULT_MISS) << 16))) { - if (runtime_stat_n(st, STAT_LL_CACHE, ctx, cpu) != 0) - print_ll_cache_misses(config, cpu, evsel, avg, out, st); + if (runtime_stat_n(st, STAT_LL_CACHE, cpu, &rsd) != 0) + print_ll_cache_misses(config, cpu, avg, out, st, &rsd); else print_metric(config, ctxp, NULL, NULL, "of all LL-cache accesses", 0); } else if (evsel__match(evsel, HARDWARE, HW_CACHE_MISSES)) { - total = runtime_stat_avg(st, STAT_CACHEREFS, ctx, cpu); + total = runtime_stat_avg(st, STAT_CACHEREFS, cpu, &rsd); if (total) ratio = avg * 100 / total; - if (runtime_stat_n(st, STAT_CACHEREFS, ctx, cpu) != 0) + if (runtime_stat_n(st, STAT_CACHEREFS, cpu, &rsd) != 0) print_metric(config, ctxp, NULL, "%8.3f %%", "of all cache refs", ratio); else print_metric(config, ctxp, NULL, NULL, "of all cache refs", 0); } else if (evsel__match(evsel, HARDWARE, HW_STALLED_CYCLES_FRONTEND)) { - print_stalled_cycles_frontend(config, cpu, evsel, avg, out, st); + print_stalled_cycles_frontend(config, cpu, avg, out, st, &rsd); } else if (evsel__match(evsel, HARDWARE, HW_STALLED_CYCLES_BACKEND)) { - print_stalled_cycles_backend(config, cpu, evsel, avg, out, st); + print_stalled_cycles_backend(config, cpu, avg, out, st, &rsd); } else if (evsel__match(evsel, HARDWARE, HW_CPU_CYCLES)) { - total = runtime_stat_avg(st, STAT_NSECS, 0, cpu); + total = runtime_stat_avg(st, STAT_NSECS, cpu, &rsd); if (total) { ratio = avg / total; @@ -1039,7 +1041,7 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, print_metric(config, ctxp, NULL, NULL, "Ghz", 0); } } else if (perf_stat_evsel__is(evsel, CYCLES_IN_TX)) { - total = runtime_stat_avg(st, STAT_CYCLES, ctx, cpu); + total = runtime_stat_avg(st, STAT_CYCLES, cpu, &rsd); if (total) print_metric(config, ctxp, NULL, @@ -1049,8 +1051,8 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, print_metric(config, ctxp, NULL, NULL, "transactional cycles", 0); } else if (perf_stat_evsel__is(evsel, CYCLES_IN_TX_CP)) { - total = runtime_stat_avg(st, STAT_CYCLES, ctx, cpu); - total2 = runtime_stat_avg(st, STAT_CYCLES_IN_TX, ctx, cpu); + total = runtime_stat_avg(st, STAT_CYCLES, cpu, &rsd); + total2 = runtime_stat_avg(st, STAT_CYCLES_IN_TX, cpu, &rsd); if (total2 < avg) total2 = avg; @@ -1060,21 +1062,19 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, else print_metric(config, ctxp, NULL, NULL, "aborted cycles", 0); } else if (perf_stat_evsel__is(evsel, TRANSACTION_START)) { - total = runtime_stat_avg(st, STAT_CYCLES_IN_TX, - ctx, cpu); + total = runtime_stat_avg(st, STAT_CYCLES_IN_TX, cpu, &rsd); if (avg) ratio = total / avg; - if (runtime_stat_n(st, STAT_CYCLES_IN_TX, ctx, cpu) != 0) + if (runtime_stat_n(st, STAT_CYCLES_IN_TX, cpu, &rsd) != 0) print_metric(config, ctxp, NULL, "%8.0f", "cycles / transaction", ratio); else print_metric(config, ctxp, NULL, NULL, "cycles / transaction", 0); } else if (perf_stat_evsel__is(evsel, ELISION_START)) { - total = runtime_stat_avg(st, STAT_CYCLES_IN_TX, - ctx, cpu); + total = runtime_stat_avg(st, STAT_CYCLES_IN_TX, cpu, &rsd); if (avg) ratio = total / avg; @@ -1087,28 +1087,28 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, else print_metric(config, ctxp, NULL, NULL, "CPUs utilized", 0); } else if (perf_stat_evsel__is(evsel, TOPDOWN_FETCH_BUBBLES)) { - double fe_bound = td_fe_bound(ctx, cpu, st); + double fe_bound = td_fe_bound(cpu, st, &rsd); if (fe_bound > 0.2) color = PERF_COLOR_RED; print_metric(config, ctxp, color, "%8.1f%%", "frontend bound", fe_bound * 100.); } else if (perf_stat_evsel__is(evsel, TOPDOWN_SLOTS_RETIRED)) { - double retiring = td_retiring(ctx, cpu, st); + double retiring = td_retiring(cpu, st, &rsd); if (retiring > 0.7) color = PERF_COLOR_GREEN; print_metric(config, ctxp, color, "%8.1f%%", "retiring", retiring * 100.); } else if (perf_stat_evsel__is(evsel, TOPDOWN_RECOVERY_BUBBLES)) { - double bad_spec = td_bad_spec(ctx, cpu, st); + double bad_spec = td_bad_spec(cpu, st, &rsd); if (bad_spec > 0.1) color = PERF_COLOR_RED; print_metric(config, ctxp, color, "%8.1f%%", "bad speculation", bad_spec * 100.); } else if (perf_stat_evsel__is(evsel, TOPDOWN_SLOTS_ISSUED)) { - double be_bound = td_be_bound(ctx, cpu, st); + double be_bound = td_be_bound(cpu, st, &rsd); const char *name = "backend bound"; static int have_recovery_bubbles = -1; @@ -1121,43 +1121,43 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, if (be_bound > 0.2) color = PERF_COLOR_RED; - if (td_total_slots(ctx, cpu, st) > 0) + if (td_total_slots(cpu, st, &rsd) > 0) print_metric(config, ctxp, color, "%8.1f%%", name, be_bound * 100.); else print_metric(config, ctxp, NULL, NULL, name, 0); } else if (perf_stat_evsel__is(evsel, TOPDOWN_RETIRING) && - full_td(ctx, cpu, st)) { - double retiring = td_metric_ratio(ctx, cpu, - STAT_TOPDOWN_RETIRING, st); - + full_td(cpu, st, &rsd)) { + double retiring = td_metric_ratio(cpu, + STAT_TOPDOWN_RETIRING, st, + &rsd); if (retiring > 0.7) color = PERF_COLOR_GREEN; print_metric(config, ctxp, color, "%8.1f%%", "retiring", retiring * 100.); } else if (perf_stat_evsel__is(evsel, TOPDOWN_FE_BOUND) && - full_td(ctx, cpu, st)) { - double fe_bound = td_metric_ratio(ctx, cpu, - STAT_TOPDOWN_FE_BOUND, st); - + full_td(cpu, st, &rsd)) { + double fe_bound = td_metric_ratio(cpu, + STAT_TOPDOWN_FE_BOUND, st, + &rsd); if (fe_bound > 0.2) color = PERF_COLOR_RED; print_metric(config, ctxp, color, "%8.1f%%", "frontend bound", fe_bound * 100.); } else if (perf_stat_evsel__is(evsel, TOPDOWN_BE_BOUND) && - full_td(ctx, cpu, st)) { - double be_bound = td_metric_ratio(ctx, cpu, - STAT_TOPDOWN_BE_BOUND, st); - + full_td(cpu, st, &rsd)) { + double be_bound = td_metric_ratio(cpu, + STAT_TOPDOWN_BE_BOUND, st, + &rsd); if (be_bound > 0.2) color = PERF_COLOR_RED; print_metric(config, ctxp, color, "%8.1f%%", "backend bound", be_bound * 100.); } else if (perf_stat_evsel__is(evsel, TOPDOWN_BAD_SPEC) && - full_td(ctx, cpu, st)) { - double bad_spec = td_metric_ratio(ctx, cpu, - STAT_TOPDOWN_BAD_SPEC, st); - + full_td(cpu, st, &rsd)) { + double bad_spec = td_metric_ratio(cpu, + STAT_TOPDOWN_BAD_SPEC, st, + &rsd); if (bad_spec > 0.1) color = PERF_COLOR_RED; print_metric(config, ctxp, color, "%8.1f%%", "bad speculation", @@ -1165,11 +1165,11 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, } else if (evsel->metric_expr) { generic_metric(config, evsel->metric_expr, evsel->metric_events, NULL, evsel->name, evsel->metric_name, NULL, 1, cpu, out, st); - } else if (runtime_stat_n(st, STAT_NSECS, 0, cpu) != 0) { + } else if (runtime_stat_n(st, STAT_NSECS, cpu, &rsd) != 0) { char unit = 'M'; char unit_buf[10]; - total = runtime_stat_avg(st, STAT_NSECS, 0, cpu); + total = runtime_stat_avg(st, STAT_NSECS, cpu, &rsd); if (total) ratio = 1000.0 * avg / total; @@ -1180,7 +1180,7 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, snprintf(unit_buf, sizeof(unit_buf), "%c/sec", unit); print_metric(config, ctxp, NULL, "%8.3f", unit_buf, ratio); } else if (perf_stat_evsel__is(evsel, SMI_NUM)) { - print_smi_cost(config, cpu, evsel, out, st); + print_smi_cost(config, cpu, out, st, &rsd); } else { num = 0; } -- cgit v1.2.3-59-g8ed1b From a1bf23052bdfe30ec3c693cf32feb2d79114ac16 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 15 Jan 2021 16:11:39 +0900 Subject: perf stat: Take cgroups into account for shadow stats As of now it doesn't consider cgroups when collecting shadow stats and metrics so counter values from different cgroups will be saved in a same slot. This resulted in incorrect numbers when those cgroups have different workloads. For example, let's look at the scenario below: cgroups A and C runs same workload which burns a cpu while cgroup B runs a light workload. $ perf stat -a -e cycles,instructions --for-each-cgroup A,B,C sleep 1 Performance counter stats for 'system wide': 3,958,116,522 cycles A 6,722,650,929 instructions A # 2.53 insn per cycle 1,132,741 cycles B 571,743 instructions B # 0.00 insn per cycle 4,007,799,935 cycles C 6,793,181,523 instructions C # 2.56 insn per cycle 1.001050869 seconds time elapsed When I run 'perf stat' with single workload, it usually shows IPC around 1.7. We can verify it (6,722,650,929.0 / 3,958,116,522 = 1.698) for cgroup A. But in this case, since cgroups are ignored, cycles are averaged so it used the lower value for IPC calculation and resulted in around 2.5. avg cycle: (3958116522 + 1132741 + 4007799935) / 3 = 2655683066 IPC (A) : 6722650929 / 2655683066 = 2.531 IPC (B) : 571743 / 2655683066 = 0.0002 IPC (C) : 6793181523 / 2655683066 = 2.557 We can simply compare cgroup pointers in the evsel and it'll be NULL when cgroups are not specified. With this patch, I can see correct numbers like below: $ perf stat -a -e cycles,instructions --for-each-cgroup A,B,C sleep 1 Performance counter stats for 'system wide': 4,171,051,687 cycles A 7,219,793,922 instructions A # 1.73 insn per cycle 1,051,189 cycles B 583,102 instructions B # 0.55 insn per cycle 4,171,124,710 cycles C 7,192,944,580 instructions C # 1.72 insn per cycle 1.007909814 seconds time elapsed Signed-off-by: Namhyung Kim Acked-by: Jiri Olsa Cc: Alexander Shishkin Cc: Andi Kleen Cc: Ian Rogers Cc: Jin Yao Cc: Mark Rutland Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lore.kernel.org/lkml/20210115071139.257042-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/stat-shadow.c | 26 +++++++++++++++++++------- 1 file changed, 19 insertions(+), 7 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/stat-shadow.c b/tools/perf/util/stat-shadow.c index a1565b6e38f2..12eafd12a693 100644 --- a/tools/perf/util/stat-shadow.c +++ b/tools/perf/util/stat-shadow.c @@ -8,6 +8,7 @@ #include "evlist.h" #include "expr.h" #include "metricgroup.h" +#include "cgroup.h" #include /* @@ -28,6 +29,7 @@ struct saved_value { enum stat_type type; int ctx; int cpu; + struct cgroup *cgrp; struct runtime_stat *stat; struct stats stats; u64 metric_total; @@ -57,6 +59,9 @@ static int saved_value_cmp(struct rb_node *rb_node, const void *entry) if (a->ctx != b->ctx) return a->ctx - b->ctx; + if (a->cgrp != b->cgrp) + return (char *)a->cgrp < (char *)b->cgrp ? -1 : +1; + if (a->evsel == NULL && b->evsel == NULL) { if (a->stat == b->stat) return 0; @@ -100,7 +105,8 @@ static struct saved_value *saved_value_lookup(struct evsel *evsel, bool create, enum stat_type type, int ctx, - struct runtime_stat *st) + struct runtime_stat *st, + struct cgroup *cgrp) { struct rblist *rblist; struct rb_node *nd; @@ -110,6 +116,7 @@ static struct saved_value *saved_value_lookup(struct evsel *evsel, .type = type, .ctx = ctx, .stat = st, + .cgrp = cgrp, }; rblist = &st->value_list; @@ -197,6 +204,7 @@ void perf_stat__reset_shadow_per_stat(struct runtime_stat *st) struct runtime_stat_data { int ctx; + struct cgroup *cgrp; }; static void update_runtime_stat(struct runtime_stat *st, @@ -205,7 +213,7 @@ static void update_runtime_stat(struct runtime_stat *st, struct runtime_stat_data *rsd) { struct saved_value *v = saved_value_lookup(NULL, cpu, true, type, - rsd->ctx, st); + rsd->ctx, st, rsd->cgrp); if (v) update_stats(&v->stats, count); @@ -223,6 +231,7 @@ void perf_stat__update_shadow_stats(struct evsel *counter, u64 count, struct saved_value *v; struct runtime_stat_data rsd = { .ctx = evsel_context(counter), + .cgrp = counter->cgrp, }; count *= counter->scale; @@ -290,13 +299,14 @@ void perf_stat__update_shadow_stats(struct evsel *counter, u64 count, update_runtime_stat(st, STAT_APERF, cpu, count, &rsd); if (counter->collect_stat) { - v = saved_value_lookup(counter, cpu, true, STAT_NONE, 0, st); + v = saved_value_lookup(counter, cpu, true, STAT_NONE, 0, st, + rsd.cgrp); update_stats(&v->stats, count); if (counter->metric_leader) v->metric_total += count; } else if (counter->metric_leader) { v = saved_value_lookup(counter->metric_leader, - cpu, true, STAT_NONE, 0, st); + cpu, true, STAT_NONE, 0, st, rsd.cgrp); v->metric_total += count; v->metric_other++; } @@ -438,7 +448,7 @@ static double runtime_stat_avg(struct runtime_stat *st, { struct saved_value *v; - v = saved_value_lookup(NULL, cpu, false, type, rsd->ctx, st); + v = saved_value_lookup(NULL, cpu, false, type, rsd->ctx, st, rsd->cgrp); if (!v) return 0.0; @@ -451,7 +461,7 @@ static double runtime_stat_n(struct runtime_stat *st, { struct saved_value *v; - v = saved_value_lookup(NULL, cpu, false, type, rsd->ctx, st); + v = saved_value_lookup(NULL, cpu, false, type, rsd->ctx, st, rsd->cgrp); if (!v) return 0.0; @@ -805,7 +815,8 @@ static int prepare_metric(struct evsel **metric_events, scale = 1e-9; } else { v = saved_value_lookup(metric_events[i], cpu, false, - STAT_NONE, 0, st); + STAT_NONE, 0, st, + metric_events[i]->cgrp); if (!v) break; stats = &v->stats; @@ -933,6 +944,7 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config, const char *color = NULL; struct runtime_stat_data rsd = { .ctx = evsel_context(evsel), + .cgrp = evsel->cgrp, }; struct metric_event *me; int num = 1; -- cgit v1.2.3-59-g8ed1b From 5501e9229a80d95a1ea68609f44c447a75d23ed5 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Thu, 7 Jan 2021 19:41:59 +0200 Subject: perf intel-pt: Fix 'CPU too large' error In some cases, the number of cpus (nr_cpus_online) is confused with the maximum cpu number (nr_cpus_avail), which results in the error in the example below: Example on system with 8 cpus: Before: # echo 0 > /sys/devices/system/cpu/cpu2/online # ./perf record --kcore -e intel_pt// taskset --cpu-list 7 uname Linux [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.147 MB perf.data ] # ./perf script --itrace=e Requested CPU 7 too large. Consider raising MAX_NR_CPUS 0x25908 [0x8]: failed to process type: 68 [Invalid argument] After: # ./perf script --itrace=e # Fixes: 8c7274691f0d ("perf machine: Replace MAX_NR_CPUS with perf_env::nr_cpus_online") Fixes: 7df4e36a4785 ("perf session: Replace MAX_NR_CPUS with perf_env::nr_cpus_online") Signed-off-by: Adrian Hunter Tested-by: Kan Liang Cc: Jiri Olsa Cc: stable@vger.kernel.org Link: http://lore.kernel.org/lkml/20210107174159.24897-1-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 4 ++-- tools/perf/util/session.c | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index f841f3503cae..1e9d3f982b47 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -2980,7 +2980,7 @@ int machines__for_each_thread(struct machines *machines, pid_t machine__get_current_tid(struct machine *machine, int cpu) { - int nr_cpus = min(machine->env->nr_cpus_online, MAX_NR_CPUS); + int nr_cpus = min(machine->env->nr_cpus_avail, MAX_NR_CPUS); if (cpu < 0 || cpu >= nr_cpus || !machine->current_tid) return -1; @@ -2992,7 +2992,7 @@ int machine__set_current_tid(struct machine *machine, int cpu, pid_t pid, pid_t tid) { struct thread *thread; - int nr_cpus = min(machine->env->nr_cpus_online, MAX_NR_CPUS); + int nr_cpus = min(machine->env->nr_cpus_avail, MAX_NR_CPUS); if (cpu < 0) return -EINVAL; diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 50ff9795a4f1..25adbcce0281 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -2404,7 +2404,7 @@ int perf_session__cpu_bitmap(struct perf_session *session, { int i, err = -1; struct perf_cpu_map *map; - int nr_cpus = min(session->header.env.nr_cpus_online, MAX_NR_CPUS); + int nr_cpus = min(session->header.env.nr_cpus_avail, MAX_NR_CPUS); for (i = 0; i < PERF_TYPE_MAX; ++i) { struct evsel *evsel; -- cgit v1.2.3-59-g8ed1b From 648b054a4647cd62e13ba79f398b8b97a7c82b19 Mon Sep 17 00:00:00 2001 From: Al Grant Date: Tue, 24 Nov 2020 19:58:17 +0000 Subject: perf inject: Correct event attribute sizes When 'perf inject' reads a perf.data file from an older version of perf, it writes event attributes into the output with the original size field, but lays them out as if they had the size currently used. Readers see a corrupt file. Update the size field to match the layout. Signed-off-by: Al Grant Acked-by: Jiri Olsa Cc: Alexander Shishkin Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lore.kernel.org/lkml/20201124195818.30603-1-al.grant@arm.com Signed-off-by: Denis Nikitin Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/header.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 062383e225a3..c4ed3dc2c8f4 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -3323,6 +3323,14 @@ int perf_session__write_header(struct perf_session *session, attr_offset = lseek(ff.fd, 0, SEEK_CUR); evlist__for_each_entry(evlist, evsel) { + if (evsel->core.attr.size < sizeof(evsel->core.attr)) { + /* + * We are likely in "perf inject" and have read + * from an older file. Update attr size so that + * reader gets the right offset to the ids. + */ + evsel->core.attr.size = sizeof(evsel->core.attr); + } f_attr = (struct perf_file_attr){ .attr = evsel->core.attr, .ids = { -- cgit v1.2.3-59-g8ed1b