Implement debugfs entries for serializing profiles and resetting counters/bitmaps. Also adds Source-based Code Coverage flags and kconfig options. This work reuses a portion of code from a previous effort by Sami Tolvanen et al. [1], specifically its debugfs interface and the underlying profile processing, but discards all its PGO-specific parts, notably value profiling. To our end (code coverage, like MC/DC, required for high assurance), we do instrumentation at the compiler frontend, instead of IR; we care about counters and bitmaps, but not value profiling. [1] https://lore.kernel.org/lkml/20210407211704.367039-1-morbo@xxxxxxxxxx/ Signed-off-by: Wentao Zhang <wentaoz5@xxxxxxxxxxxx> Signed-off-by: Chuck Wolber <chuck.wolber@xxxxxxxxxx> --- Makefile | 3 + arch/Kconfig | 1 + arch/x86/Kconfig | 1 + arch/x86/kernel/vmlinux.lds.S | 2 + include/asm-generic/vmlinux.lds.h | 38 +++++ kernel/Makefile | 1 + kernel/llvm-cov/Kconfig | 29 ++++ kernel/llvm-cov/Makefile | 5 + kernel/llvm-cov/fs.c | 253 ++++++++++++++++++++++++++++++ kernel/llvm-cov/llvm-cov.h | 156 ++++++++++++++++++ scripts/Makefile.lib | 10 ++ scripts/mod/modpost.c | 2 + 12 files changed, 501 insertions(+) create mode 100644 kernel/llvm-cov/Kconfig create mode 100644 kernel/llvm-cov/Makefile create mode 100644 kernel/llvm-cov/fs.c create mode 100644 kernel/llvm-cov/llvm-cov.h diff --git a/Makefile b/Makefile index 68ebd6d6b444..1750a2b7dfe8 100644 --- a/Makefile +++ b/Makefile @@ -737,6 +737,9 @@ endif # KBUILD_EXTMOD # Defaults to vmlinux, but the arch makefile usually adds further targets all: vmlinux +CFLAGS_LLVM_COV := -fprofile-instr-generate -fcoverage-mapping +export CFLAGS_LLVM_COV + CFLAGS_GCOV := -fprofile-arcs -ftest-coverage ifdef CONFIG_CC_IS_GCC CFLAGS_GCOV += -fno-tree-loop-im diff --git a/arch/Kconfig b/arch/Kconfig index 975dd22a2dbd..0727265f677f 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -1601,6 +1601,7 @@ config ARCH_HAS_KERNEL_FPU_SUPPORT the kernel, as described in Documentation/core-api/floating-point.rst. source "kernel/gcov/Kconfig" +source "kernel/llvm-cov/Kconfig" source "scripts/gcc-plugins/Kconfig" diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 007bab9f2a0e..87a793b82bab 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -85,6 +85,7 @@ config X86 select ARCH_HAS_FORTIFY_SOURCE select ARCH_HAS_GCOV_PROFILE_ALL select ARCH_HAS_KCOV if X86_64 + select ARCH_HAS_LLVM_COV if X86_64 select ARCH_HAS_KERNEL_FPU_SUPPORT select ARCH_HAS_MEM_ENCRYPT select ARCH_HAS_MEMBARRIER_SYNC_CORE diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S index 6e73403e874f..90433772240a 100644 --- a/arch/x86/kernel/vmlinux.lds.S +++ b/arch/x86/kernel/vmlinux.lds.S @@ -191,6 +191,8 @@ SECTIONS BUG_TABLE + LLVM_COV_DATA + ORC_UNWIND_TABLE . = ALIGN(PAGE_SIZE); diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index 1ae44793132a..770ff8469dcd 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -334,6 +334,44 @@ #define THERMAL_TABLE(name) #endif +#ifdef CONFIG_LLVM_COV_KERNEL +#define LLVM_COV_DATA \ + __llvm_prf_data : AT(ADDR(__llvm_prf_data) - LOAD_OFFSET) { \ + __llvm_prf_start = .; \ + __llvm_prf_data_start = .; \ + *(__llvm_prf_data) \ + __llvm_prf_data_end = .; \ + } \ + __llvm_prf_cnts : AT(ADDR(__llvm_prf_cnts) - LOAD_OFFSET) { \ + __llvm_prf_cnts_start = .; \ + *(__llvm_prf_cnts) \ + __llvm_prf_cnts_end = .; \ + } \ + __llvm_prf_names : AT(ADDR(__llvm_prf_names) - LOAD_OFFSET) { \ + __llvm_prf_names_start = .; \ + *(__llvm_prf_names) \ + __llvm_prf_names_end = .; \ + } \ + __llvm_prf_bits : AT(ADDR(__llvm_prf_bits) - LOAD_OFFSET) { \ + __llvm_prf_bits_start = .; \ + *(__llvm_prf_bits) \ + __llvm_prf_bits_end = .; \ + } \ + __llvm_covfun : AT(ADDR(__llvm_covfun) - LOAD_OFFSET) { \ + __llvm_covfun_start = .; \ + *(__llvm_covfun) \ + __llvm_covfun_end = .; \ + } \ + __llvm_covmap : AT(ADDR(__llvm_covmap) - LOAD_OFFSET) { \ + __llvm_covmap_start = .; \ + *(__llvm_covmap) \ + __llvm_covmap_end = .; \ + __llvm_prf_end = .; \ + } +#else +#define LLVM_COV_DATA +#endif + #define KERNEL_DTB() \ STRUCT_ALIGN(); \ __dtb_start = .; \ diff --git a/kernel/Makefile b/kernel/Makefile index 3c13240dfc9f..773e6a9ee00a 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -117,6 +117,7 @@ obj-$(CONFIG_HAVE_STATIC_CALL) += static_call.o obj-$(CONFIG_HAVE_STATIC_CALL_INLINE) += static_call_inline.o obj-$(CONFIG_CFI_CLANG) += cfi.o obj-$(CONFIG_NUMA) += numa.o +obj-$(CONFIG_LLVM_COV_KERNEL) += llvm-cov/ obj-$(CONFIG_PERF_EVENTS) += events/ diff --git a/kernel/llvm-cov/Kconfig b/kernel/llvm-cov/Kconfig new file mode 100644 index 000000000000..505eba5bd23c --- /dev/null +++ b/kernel/llvm-cov/Kconfig @@ -0,0 +1,29 @@ +# SPDX-License-Identifier: GPL-2.0-only +menu "Clang's source-based kernel coverage measurement (EXPERIMENTAL)" + +config ARCH_HAS_LLVM_COV + bool + +config LLVM_COV_KERNEL + bool "Enable Clang's source-based kernel coverage measurement" + depends on DEBUG_FS + depends on ARCH_HAS_LLVM_COV + depends on CC_IS_CLANG && CLANG_VERSION >= 180000 + default n + help + This option enables Clang's Source-based Code Coverage. + + If unsure, say N. + + On a kernel compiled with this option, run your test suites, and + download the raw profile from /sys/kernel/debug/llvm-cov/profraw. + This file can then be converted into the indexed format with + llvm-profdata and used to generate coverage reports with llvm-cov. + + Note that a kernel compiled with coverage flags will be significantly + larger and run slower. + + Note that the debugfs filesystem has to be mounted to access the raw + profile. + +endmenu diff --git a/kernel/llvm-cov/Makefile b/kernel/llvm-cov/Makefile new file mode 100644 index 000000000000..a0f45e05f8fb --- /dev/null +++ b/kernel/llvm-cov/Makefile @@ -0,0 +1,5 @@ +# SPDX-License-Identifier: GPL-2.0 +GCOV_PROFILE := n +LLVM_COV_PROFILE := n + +obj-y += fs.o \ No newline at end of file diff --git a/kernel/llvm-cov/fs.c b/kernel/llvm-cov/fs.c new file mode 100644 index 000000000000..c56f660a174e --- /dev/null +++ b/kernel/llvm-cov/fs.c @@ -0,0 +1,253 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2019 Sami Tolvanen <samitolvanen@xxxxxxxxxx>, Google, Inc. + * Copyright (C) 2024 Jinghao Jia <jinghao7@xxxxxxxxxxxx>, UIUC + * Copyright (C) 2024 Wentao Zhang <wentaoz5@xxxxxxxxxxxx>, UIUC + * + * This software is licensed under the terms of the GNU General Public + * License version 2, as published by the Free Software Foundation, and + * may be copied, distributed, and modified under those terms. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + */ + +#define pr_fmt(fmt) "llvm-cov: " fmt + +#include <linux/kernel.h> +#include <linux/debugfs.h> +#include <linux/fs.h> +#include <linux/module.h> +#include <linux/slab.h> +#include <linux/vmalloc.h> +#include <linux/spinlock.h> +#include "llvm-cov.h" + +/* + * This lock guards both counter/bitmap reset and serialization of the + * raw profile data. Keeping both of these activities separate via locking + * ensures that we don't try to serialize data that's being reset. + */ +DEFINE_SPINLOCK(llvm_cov_lock); + +static struct dentry *directory; + +struct llvm_cov_private_data { + char *buffer; + unsigned long size; +}; + +/* + * Raw profile data format: + * https://llvm.org/docs/InstrProfileFormat.html#raw-profile-format. We will + * only populate information that's relevant to basic Source-based Code Coverage + * before serialization. Other features like binary IDs, continuous mode, + * single-byte mode, value profiling, type profiling etc are not implemented. + */ + +static void llvm_cov_fill_raw_profile_header(void **buffer) +{ + struct __llvm_profile_header *header = *(struct __llvm_profile_header **)buffer; + + header->magic = INSTR_PROF_RAW_MAGIC_64; + header->version = INSTR_PROF_RAW_VERSION; + header->binary_ids_size = 0; + header->num_data = __llvm_prf_data_count(); + header->padding_bytes_before_counters = 0; + header->num_counters = __llvm_prf_cnts_count(); + header->padding_bytes_after_counters = + __llvm_prf_get_padding(__llvm_prf_cnts_size()); + header->num_bitmap_bytes = __llvm_prf_bits_size(); + header->padding_bytes_after_bitmap_bytes = + __llvm_prf_get_padding(__llvm_prf_bits_size()); + header->names_size = __llvm_prf_names_size(); + header->counters_delta = (u64)__llvm_prf_cnts_start - + (u64)__llvm_prf_data_start; + header->bitmap_delta = (u64)__llvm_prf_bits_start - + (u64)__llvm_prf_data_start; + header->names_delta = (u64)__llvm_prf_names_start; +#if defined(CONFIG_CC_IS_CLANG) && CONFIG_CLANG_VERSION >= 190000 + header->num_v_tables = 0; + header->v_names_size = 0; +#endif + header->value_kind_last = IPVK_LAST; + + *buffer += sizeof(*header); +} + +/* + * Copy the source into the buffer, incrementing the pointer into buffer in the + * process. + */ +static void llvm_cov_copy_section_to_buffer(void **buffer, void *src, + unsigned long size) +{ + memcpy(*buffer, src, size); + *buffer += size; +} + +static unsigned long llvm_cov_get_raw_profile_size(void) +{ + return sizeof(struct __llvm_profile_header) + + __llvm_prf_data_size() + + __llvm_prf_cnts_size() + + __llvm_prf_get_padding(__llvm_prf_cnts_size()) + + __llvm_prf_bits_size() + + __llvm_prf_get_padding(__llvm_prf_bits_size()) + + __llvm_prf_names_size() + + __llvm_prf_get_padding(__llvm_prf_names_size()); +} + +/* + * Serialize in-memory data into a format LLVM tools can understand + * (https://llvm.org/docs/InstrProfileFormat.html#raw-profile-format) + */ +static int llvm_cov_serialize_raw_profile(struct llvm_cov_private_data *p) +{ + int err = 0; + void *buffer; + + p->size = llvm_cov_get_raw_profile_size(); + p->buffer = vzalloc(p->size); + + if (!p->buffer) { + err = -ENOMEM; + goto out; + } + + buffer = p->buffer; + + llvm_cov_fill_raw_profile_header(&buffer); + llvm_cov_copy_section_to_buffer(&buffer, __llvm_prf_data_start, + __llvm_prf_data_size()); + llvm_cov_copy_section_to_buffer(&buffer, __llvm_prf_cnts_start, + __llvm_prf_cnts_size()); + buffer += __llvm_prf_get_padding(__llvm_prf_cnts_size()); + llvm_cov_copy_section_to_buffer(&buffer, __llvm_prf_bits_start, + __llvm_prf_bits_size()); + buffer += __llvm_prf_get_padding(__llvm_prf_bits_size()); + llvm_cov_copy_section_to_buffer(&buffer, __llvm_prf_names_start, + __llvm_prf_names_size()); + buffer += __llvm_prf_get_padding(__llvm_prf_names_size()); + +out: + return err; +} + +/* open() implementation for llvm-cov data file. */ +static int llvm_cov_open(struct inode *inode, struct file *file) +{ + struct llvm_cov_private_data *data; + unsigned long flags; + int err; + + data = kzalloc(sizeof(*data), GFP_KERNEL); + if (!data) { + err = -ENOMEM; + goto out; + } + + flags = llvm_cov_claim_lock(); + + err = llvm_cov_serialize_raw_profile(data); + if (unlikely(err)) { + kfree(data); + goto out_unlock; + } + + file->private_data = data; + +out_unlock: + llvm_cov_release_lock(flags); +out: + return err; +} + +/* read() implementation for llvm-cov data file. */ +static ssize_t llvm_cov_read(struct file *file, char __user *buf, size_t count, + loff_t *ppos) +{ + struct llvm_cov_private_data *data = file->private_data; + + if (!data) + return -EBADF; + + return simple_read_from_buffer(buf, count, ppos, data->buffer, + data->size); +} + +/* release() implementation for llvm-cov data file. */ +static int llvm_cov_release(struct inode *inode, struct file *file) +{ + struct llvm_cov_private_data *data = file->private_data; + + if (data) { + vfree(data->buffer); + kfree(data); + } + + return 0; +} + +static const struct file_operations llvm_cov_data_fops = { + .owner = THIS_MODULE, + .open = llvm_cov_open, + .read = llvm_cov_read, + .llseek = default_llseek, + .release = llvm_cov_release +}; + +/* write() implementation for llvm-cov reset file */ +static ssize_t reset_write(struct file *file, const char __user *addr, + size_t len, loff_t *pos) +{ + unsigned long flags; + + flags = llvm_cov_claim_lock(); + memset(__llvm_prf_cnts_start, 0, __llvm_prf_cnts_size()); + memset(__llvm_prf_bits_start, 0, __llvm_prf_bits_size()); + llvm_cov_release_lock(flags); + + return len; +} + +static const struct file_operations llvm_cov_reset_fops = { + .owner = THIS_MODULE, + .write = reset_write, + .llseek = noop_llseek, +}; + +/* Create debugfs entries. */ +static int __init llvm_cov_init(void) +{ + directory = debugfs_create_dir("llvm-cov", NULL); + if (!directory) + goto err_remove; + + if (!debugfs_create_file("profraw", 0400, directory, NULL, + &llvm_cov_data_fops)) + goto err_remove; + + if (!debugfs_create_file("reset", 0200, directory, NULL, + &llvm_cov_reset_fops)) + goto err_remove; + + return 0; + +err_remove: + debugfs_remove_recursive(directory); + pr_err("initialization failed\n"); + return -EIO; +} + +/* Remove debugfs entries. */ +static void __exit llvm_cov_exit(void) +{ + debugfs_remove_recursive(directory); +} + +module_init(llvm_cov_init); +module_exit(llvm_cov_exit); diff --git a/kernel/llvm-cov/llvm-cov.h b/kernel/llvm-cov/llvm-cov.h new file mode 100644 index 000000000000..d9551a685756 --- /dev/null +++ b/kernel/llvm-cov/llvm-cov.h @@ -0,0 +1,156 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2019 Sami Tolvanen <samitolvanen@xxxxxxxxxx>, Google, Inc. + * Copyright (C) 2024 Jinghao Jia <jinghao7@xxxxxxxxxxxx>, UIUC + * Copyright (C) 2024 Wentao Zhang <wentaoz5@xxxxxxxxxxxx>, UIUC + * + * This software is licensed under the terms of the GNU General Public + * License version 2, as published by the Free Software Foundation, and + * may be copied, distributed, and modified under those terms. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + */ + +#ifndef _LLVM_COV_H +#define _LLVM_COV_H + +extern spinlock_t llvm_cov_lock; + +static __always_inline unsigned long llvm_cov_claim_lock(void) +{ + unsigned long flags; + + spin_lock_irqsave(&llvm_cov_lock, flags); + + return flags; +} + +static __always_inline void llvm_cov_release_lock(unsigned long flags) +{ + spin_unlock_irqrestore(&llvm_cov_lock, flags); +} + +/* + * Note: These internal LLVM definitions must match the compiler version. + * See llvm/include/llvm/ProfileData/InstrProfData.inc in LLVM's source code. + */ + +#define INSTR_PROF_RAW_MAGIC_64 \ + ((u64)255 << 56 | \ + (u64)'l' << 48 | \ + (u64)'p' << 40 | \ + (u64)'r' << 32 | \ + (u64)'o' << 24 | \ + (u64)'f' << 16 | \ + (u64)'r' << 8 | \ + (u64)129) + +#if defined(CONFIG_CC_IS_CLANG) && CONFIG_CLANG_VERSION >= 190000 +#define INSTR_PROF_RAW_VERSION 10 +#define INSTR_PROF_DATA_ALIGNMENT 8 +#define IPVK_LAST 2 +#elif defined(CONFIG_CC_IS_CLANG) && CONFIG_CLANG_VERSION >= 180000 +#define INSTR_PROF_RAW_VERSION 9 +#define INSTR_PROF_DATA_ALIGNMENT 8 +#define IPVK_LAST 1 +#endif + +/** + * struct __llvm_profile_header - represents the raw profile header data + * structure. Description of each member can be found here: + * https://llvm.org/docs/InstrProfileFormat.html#header. + */ +struct __llvm_profile_header { + u64 magic; + u64 version; + u64 binary_ids_size; + u64 num_data; + u64 padding_bytes_before_counters; + u64 num_counters; + u64 padding_bytes_after_counters; + u64 num_bitmap_bytes; + u64 padding_bytes_after_bitmap_bytes; + u64 names_size; + u64 counters_delta; + u64 bitmap_delta; + u64 names_delta; +#if defined(CONFIG_CC_IS_CLANG) && CONFIG_CLANG_VERSION >= 190000 + u64 num_v_tables; + u64 v_names_size; +#endif + u64 value_kind_last; +}; + +/** + * struct __llvm_profile_data - represents the per-function control structure. + * Description of each member can be found here: + * https://llvm.org/docs/InstrProfileFormat.html#profile-metadata. To measure + * Source-based Code Coverage, the internals of this struct don't matter at run + * time. The only purpose of the definition below is to run sizeof() against it + * so that we can calculate the "num_data" field in header. + */ +struct __llvm_profile_data { + const u64 name_ref; + const u64 func_hash; + const void *counter_ptr; + const void *bitmap_ptr; + const void *function_pointer; + void *values; + const u32 num_counters; + const u16 num_value_sites[IPVK_LAST + 1]; + const u32 num_bitmap_bytes; +} __aligned(INSTR_PROF_DATA_ALIGNMENT); + +/* Payload sections */ + +extern struct __llvm_profile_data __llvm_prf_data_start[]; +extern struct __llvm_profile_data __llvm_prf_data_end[]; + +extern u64 __llvm_prf_cnts_start[]; +extern u64 __llvm_prf_cnts_end[]; + +extern char __llvm_prf_names_start[]; +extern char __llvm_prf_names_end[]; + +extern char __llvm_prf_bits_start[]; +extern char __llvm_prf_bits_end[]; + +#define __DEFINE_SECTION_SIZE(s) \ + static inline unsigned long __llvm_prf_ ## s ## _size(void) \ + { \ + unsigned long start = \ + (unsigned long)__llvm_prf_ ## s ## _start; \ + unsigned long end = \ + (unsigned long)__llvm_prf_ ## s ## _end; \ + return end - start; \ + } +#define __DEFINE_SECTION_COUNT(s) \ + static inline unsigned long __llvm_prf_ ## s ## _count(void) \ + { \ + return __llvm_prf_ ## s ## _size() / \ + sizeof(__llvm_prf_ ## s ## _start[0]); \ + } + +__DEFINE_SECTION_SIZE(data) +__DEFINE_SECTION_SIZE(cnts) +__DEFINE_SECTION_SIZE(names) +__DEFINE_SECTION_SIZE(bits) + +__DEFINE_SECTION_COUNT(data) +__DEFINE_SECTION_COUNT(cnts) +__DEFINE_SECTION_COUNT(names) +__DEFINE_SECTION_COUNT(bits) + +#undef __DEFINE_SECTION_SIZE +#undef __DEFINE_SECTION_COUNT + +static inline unsigned long __llvm_prf_get_padding(unsigned long size) +{ + return 7 & (sizeof(u64) - size % sizeof(u64)); +} + +#endif /* _LLVM_COV_H */ diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib index fe3668dc4954..b9ceaee34b28 100644 --- a/scripts/Makefile.lib +++ b/scripts/Makefile.lib @@ -158,6 +158,16 @@ _c_flags += $(if $(patsubst n%,, \ $(CFLAGS_GCOV)) endif +# +# Enable Clang's Source-based Code Coverage flags for a file or directory +# depending on variables LLVM_COV_PROFILE_obj.o and LLVM_COV_PROFILE. +# +ifeq ($(CONFIG_LLVM_COV_KERNEL),y) +_c_flags += $(if $(patsubst n%,, \ + $(LLVM_COV_PROFILE_$(basetarget).o)$(LLVM_COV_PROFILE)y), \ + $(CFLAGS_LLVM_COV)) +endif + # # Enable address sanitizer flags for kernel except some files or directories # we don't want to check (depends on variables KASAN_SANITIZE_obj.o, KASAN_SANITIZE) diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c index d16d0ace2775..836c2289b8d8 100644 --- a/scripts/mod/modpost.c +++ b/scripts/mod/modpost.c @@ -743,6 +743,8 @@ static const char *const section_white_list[] = ".gnu.lto*", ".discard.*", ".llvm.call-graph-profile", /* call graph */ + "__llvm_covfun", + "__llvm_covmap", NULL }; -- 2.45.2
