kexec: Use BPF lskel to enable kexec to load PE format boot image

arch/arm64/Kconfig

@@ -1608,6 +1608,7 @@ config ARCH_SELECTS_KEXEC_FILE
 	def_bool y
 	depends on KEXEC_FILE
 	select HAVE_IMA_KEXEC if IMA
+	select KEXEC_PE_IMAGE
 
 config ARCH_SUPPORTS_KEXEC_SIG
 	def_bool y

arch/arm64/include/asm/kexec.h

@@ -120,6 +120,7 @@ struct kimage_arch {
 
 #ifdef CONFIG_KEXEC_FILE
 extern const struct kexec_file_ops kexec_image_ops;
+extern const struct kexec_file_ops kexec_pe_image_ops;
 
 int arch_kimage_file_post_load_cleanup(struct kimage *image);
 #define arch_kimage_file_post_load_cleanup arch_kimage_file_post_load_cleanup

arch/arm64/kernel/machine_kexec_file.c

@@ -24,6 +24,9 @@
 
 const struct kexec_file_ops * const kexec_file_loaders[] = {
 	&kexec_image_ops,
+#ifdef CONFIG_KEXEC_PE_IMAGE
+	&kexec_pe_image_ops,
+#endif
 	NULL
 };
 

include/linux/bpf.h

@@ -3709,4 +3709,46 @@ int bpf_prog_get_file_line(struct bpf_prog *prog, unsigned long ip, const char *
 			   const char **linep, int *nump);
 struct bpf_prog *bpf_prog_find_from_stack(void);
 
+enum alloc_type {
+	TYPE_KALLOC,
+	TYPE_VMALLOC,
+	TYPE_VMAP,
+};
+
+struct mem_range_result {
+	struct kref ref;
+	char *buf;
+	uint32_t buf_sz;
+	uint32_t data_sz;
+	/* kmalloc-ed, vmalloc-ed, or vmap-ed */
+	enum alloc_type alloc_type;
+	/* Valid if vmap-ed */
+	struct page **pages;
+	unsigned int pg_cnt;
+	int status;
+	struct mem_cgroup *memcg;
+};
+
+struct mem_range_result *mem_range_result_alloc(void);
+void mem_range_result_get(struct mem_range_result *r);
+void mem_range_result_put(struct mem_range_result *r);
+
+__bpf_kfunc int bpf_mem_range_result_put(struct mem_range_result *result);
+__bpf_kfunc int bpf_copy_to_kernel(const char *name, char *buf, int size);
+
+typedef int (*resource_handler)(const char *name, struct mem_range_result *r);
+
+struct carrier_listener {
+	struct hlist_node node;
+	char *name;
+	resource_handler handler;
+	/*
+	 * bpf_copy_to_kernel() knows the size in advance, so vmap-ed is not
+	 * supported.
+	 */
+	enum alloc_type alloc_type;
+};
+
+int register_carrier_listener(struct carrier_listener *listener);
+int unregister_carrier_listener(char *str);
 #endif /* _LINUX_BPF_H */

include/linux/decompress/mm.h

@@ -92,7 +92,14 @@ MALLOC_VISIBLE void free(void *where)
 #define large_malloc(a) vmalloc(a)
 #define large_free(a) vfree(a)
 
+#ifdef CONFIG_KEEP_DECOMPRESSOR
+#define INIT
+#define INITCONST
+#else
 #define INIT __init
+#define INITCONST __initconst
+#endif
+
 #define STATIC
 
 #include <linux/init.h>

include/linux/kexec.h

@@ -23,6 +23,10 @@
 #include <uapi/linux/kexec.h>
 #include <linux/verification.h>
 
+#if defined(CONFIG_ARCH_SUPPORTS_KEXEC_PURGATORY) || defined(CONFIG_KEXEC_PE_IMAGE)
+#include <linux/module.h>
+#endif
+
 extern note_buf_t __percpu *crash_notes;
 
 #ifdef CONFIG_CRASH_DUMP
@@ -158,6 +162,7 @@ extern const struct kexec_file_ops * const kexec_file_loaders[];
 
 int kexec_image_probe_default(struct kimage *image, void *buf,
 			      unsigned long buf_len);
+void *kexec_image_load_default(struct kimage *image);
 int kexec_image_post_load_cleanup_default(struct kimage *image);
 
 /*
@@ -443,6 +448,7 @@ static inline int machine_kexec_post_load(struct kimage *image) { return 0; }
 
 extern struct kimage *kexec_image;
 extern struct kimage *kexec_crash_image;
+extern const struct kexec_file_ops pe_image_ops;
 
 bool kexec_load_permitted(int kexec_image_type);
 
@@ -548,6 +554,10 @@ void set_kexec_sig_enforced(void);
 static inline void set_kexec_sig_enforced(void) {}
 #endif
 
+#if defined(CONFIG_ARCH_SUPPORTS_KEXEC_PURGATORY) || defined(CONFIG_KEXEC_PE_IMAGE)
+const Elf_Sym *elf_find_symbol(const Elf_Ehdr *ehdr, const char *name);
+#endif
+
 #endif /* !defined(__ASSEBMLY__) */
 
 #endif /* LINUX_KEXEC_H */

kernel/Kconfig.kexec

@@ -46,6 +46,15 @@ config KEXEC_FILE
 	  for kernel and initramfs as opposed to list of segments as
 	  accepted by kexec system call.
 
+config KEXEC_PE_IMAGE
+	bool "Enable parsing UEFI PE file through kexec file based system call"
+	select KEEP_DECOMPRESSOR
+	depends on KEXEC_FILE
+	depends on DEBUG_INFO_BTF && BPF_SYSCALL
+	help
+	  This option makes the kexec_file_load() syscall cooperates with bpf-prog
+	  to parse PE format file
+
 config KEXEC_SIG
 	bool "Verify kernel signature during kexec_file_load() syscall"
 	depends on ARCH_SUPPORTS_KEXEC_SIG

kernel/Makefile

@@ -80,6 +80,7 @@ obj-$(CONFIG_CRASH_DUMP) += crash_core.o
 obj-$(CONFIG_CRASH_DM_CRYPT) += crash_dump_dm_crypt.o
 obj-$(CONFIG_KEXEC) += kexec.o
 obj-$(CONFIG_KEXEC_FILE) += kexec_file.o
+obj-$(CONFIG_KEXEC_PE_IMAGE) += kexec_pe_image.o
 obj-$(CONFIG_KEXEC_ELF) += kexec_elf.o
 obj-$(CONFIG_KEXEC_HANDOVER) += kexec_handover.o
 obj-$(CONFIG_BACKTRACE_SELF_TEST) += backtracetest.o
@@ -141,6 +142,7 @@ obj-$(CONFIG_SYSCTL_KUNIT_TEST) += sysctl-test.o
 
 CFLAGS_kstack_erase.o += $(DISABLE_KSTACK_ERASE)
 CFLAGS_kstack_erase.o += $(call cc-option,-mgeneral-regs-only)
+CFLAGS_kexec_pe_image.o += -I$(srctree)/tools/lib
 obj-$(CONFIG_KSTACK_ERASE) += kstack_erase.o
 KASAN_SANITIZE_kstack_erase.o := n
 KCSAN_SANITIZE_kstack_erase.o := n

kernel/bpf/Makefile

@@ -56,6 +56,9 @@ obj-$(CONFIG_BPF_SYSCALL) += kmem_cache_iter.o
 ifeq ($(CONFIG_DMA_SHARED_BUFFER),y)
 obj-$(CONFIG_BPF_SYSCALL) += dmabuf_iter.o
 endif
+ifeq ($(CONFIG_KEXEC_PE_IMAGE),y)
+obj-$(CONFIG_BPF_SYSCALL) += helpers_carrier.o
+endif
 
 CFLAGS_REMOVE_percpu_freelist.o = $(CC_FLAGS_FTRACE)
 CFLAGS_REMOVE_bpf_lru_list.o = $(CC_FLAGS_FTRACE)