alloc_tag: support for page allocation tag compression

Implement support for storing page allocation tag references directly in
the page flags instead of page extensions.  sysctl.vm.mem_profiling boot
parameter it extended to provide a way for a user to request this mode. 
Enabling compression eliminates memory overhead caused by page_ext and
results in better performance for page allocations.  However this mode
will not work if the number of available page flag bits is insufficient to
address all kernel allocations.  Such condition can happen during boot or
when loading a module.  If this condition is detected, memory allocation
profiling gets disabled with an appropriate warning.  By default
compression mode is disabled.

Link: https://lkml.kernel.org/r/20241023170759.999909-7-surenb@google.com
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Reviewed-by: Pasha Tatashin <pasha.tatashin@soleen.com>
Cc: Ard Biesheuvel <ardb@kernel.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Borislav Petkov (AMD) <bp@alien8.de>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Daniel Gomez <da.gomez@samsung.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: David Rientjes <rientjes@google.com>
Cc: Dennis Zhou <dennis@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Kalesh Singh <kaleshsingh@google.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Liam R. Howlett <Liam.Howlett@Oracle.com>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport (Microsoft) <rppt@kernel.org>
Cc: Minchan Kim <minchan@google.com>
Cc: Paul E. McKenney <paulmck@kernel.org>
Cc: Petr Pavlu <petr.pavlu@suse.com>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Sami Tolvanen <samitolvanen@google.com>
Cc: Sourav Panda <souravpanda@google.com>
Cc: Steven Rostedt (Google) <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Huth <thuth@redhat.com>
Cc: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Xiongwei Song <xiongwei.song@windriver.com>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

Documentation/mm/allocation-profiling.rst

@@ -18,12 +18,17 @@ kconfig options:
   missing annotation
 
 Boot parameter:
-  sysctl.vm.mem_profiling=0|1|never
+  sysctl.vm.mem_profiling={0|1|never}[,compressed]
 
   When set to "never", memory allocation profiling overhead is minimized and it
   cannot be enabled at runtime (sysctl becomes read-only).
   When CONFIG_MEM_ALLOC_PROFILING_ENABLED_BY_DEFAULT=y, default value is "1".
   When CONFIG_MEM_ALLOC_PROFILING_ENABLED_BY_DEFAULT=n, default value is "never".
+  "compressed" optional parameter will try to store page tag references in a
+  compact format, avoiding page extensions. This results in improved performance
+  and memory consumption, however it might fail depending on system configuration.
+  If compression fails, a warning is issued and memory allocation profiling gets
+  disabled.
 
 sysctl:
   /proc/sys/vm/mem_profiling

include/linux/alloc_tag.h

@@ -30,8 +30,16 @@ struct alloc_tag {
 	struct alloc_tag_counters __percpu	*counters;
 } __aligned(8);
 
+struct alloc_tag_kernel_section {
+	struct alloc_tag *first_tag;
+	unsigned long count;
+};
+
 struct alloc_tag_module_section {
-	unsigned long start_addr;
+	union {
+		unsigned long start_addr;
+		struct alloc_tag *first_tag;
+	};
 	unsigned long end_addr;
 	/* used size */
 	unsigned long size;

include/linux/codetag.h

@@ -13,6 +13,9 @@ struct codetag_module;
 struct seq_buf;
 struct module;
 
+#define CODETAG_SECTION_START_PREFIX	"__start_"
+#define CODETAG_SECTION_STOP_PREFIX	"__stop_"
+
 /*
  * An instance of this structure is created in a special ELF section at every
  * code location being tagged.  At runtime, the special section is treated as

include/linux/page-flags-layout.h

@@ -111,5 +111,12 @@
 			    ZONES_WIDTH - LRU_GEN_WIDTH - SECTIONS_WIDTH - \
 			    NODES_WIDTH - KASAN_TAG_WIDTH - LAST_CPUPID_WIDTH)
 
+#define NR_NON_PAGEFLAG_BITS	(SECTIONS_WIDTH + NODES_WIDTH + ZONES_WIDTH + \
+				LAST_CPUPID_SHIFT + KASAN_TAG_WIDTH + \
+				LRU_GEN_WIDTH + LRU_REFS_WIDTH)
+
+#define NR_UNUSED_PAGEFLAG_BITS	(BITS_PER_LONG - \
+				(NR_NON_PAGEFLAG_BITS + NR_PAGEFLAGS))
+
 #endif
 #endif /* _LINUX_PAGE_FLAGS_LAYOUT */

include/linux/pgalloc_tag.h

@@ -11,29 +11,118 @@
 
 #include <linux/page_ext.h>
 
+extern struct page_ext_operations page_alloc_tagging_ops;
+extern unsigned long alloc_tag_ref_mask;
+extern int alloc_tag_ref_offs;
+extern struct alloc_tag_kernel_section kernel_tags;
+
+DECLARE_STATIC_KEY_FALSE(mem_profiling_compressed);
+
+typedef u16	pgalloc_tag_idx;
+
 union pgtag_ref_handle {
 	union codetag_ref *ref;	/* reference in page extension */
+	struct page *page;	/* reference in page flags */
 };
 
-extern struct page_ext_operations page_alloc_tagging_ops;
+/* Reserved indexes */
+#define CODETAG_ID_NULL		0
+#define CODETAG_ID_EMPTY	1
+#define CODETAG_ID_FIRST	2
+
+#ifdef CONFIG_MODULES
+
+extern struct alloc_tag_module_section module_tags;
+
+static inline struct alloc_tag *module_idx_to_tag(pgalloc_tag_idx idx)
+{
+	return &module_tags.first_tag[idx - kernel_tags.count];
+}
+
+static inline pgalloc_tag_idx module_tag_to_idx(struct alloc_tag *tag)
+{
+	return CODETAG_ID_FIRST + kernel_tags.count + (tag - module_tags.first_tag);
+}
+
+#else /* CONFIG_MODULES */
+
+static inline struct alloc_tag *module_idx_to_tag(pgalloc_tag_idx idx)
+{
+	pr_warn("invalid page tag reference %lu\n", (unsigned long)idx);
+	return NULL;
+}
+
+static inline pgalloc_tag_idx module_tag_to_idx(struct alloc_tag *tag)
+{
+	pr_warn("invalid page tag 0x%lx\n", (unsigned long)tag);
+	return CODETAG_ID_NULL;
+}
+
+#endif /* CONFIG_MODULES */
+
+static inline void idx_to_ref(pgalloc_tag_idx idx, union codetag_ref *ref)
+{
+	switch (idx) {
+	case (CODETAG_ID_NULL):
+		ref->ct = NULL;
+		break;
+	case (CODETAG_ID_EMPTY):
+		set_codetag_empty(ref);
+		break;
+	default:
+		idx -= CODETAG_ID_FIRST;
+		ref->ct = idx < kernel_tags.count ?
+			&kernel_tags.first_tag[idx].ct :
+			&module_idx_to_tag(idx)->ct;
+		break;
+	}
+}
+
+static inline pgalloc_tag_idx ref_to_idx(union codetag_ref *ref)
+{
+	struct alloc_tag *tag;
+
+	if (!ref->ct)
+		return CODETAG_ID_NULL;
+
+	if (is_codetag_empty(ref))
+		return CODETAG_ID_EMPTY;
+
+	tag = ct_to_alloc_tag(ref->ct);
+	if (tag >= kernel_tags.first_tag && tag < kernel_tags.first_tag + kernel_tags.count)
+		return CODETAG_ID_FIRST + (tag - kernel_tags.first_tag);
+
+	return module_tag_to_idx(tag);
+}
+
+
 
 /* Should be called only if mem_alloc_profiling_enabled() */
 static inline bool get_page_tag_ref(struct page *page, union codetag_ref *ref,
 				    union pgtag_ref_handle *handle)
 {
-	struct page_ext *page_ext;
-	union codetag_ref *tmp;
-
 	if (!page)
 		return false;
 
-	page_ext = page_ext_get(page);
-	if (!page_ext)
-		return false;
+	if (static_key_enabled(&mem_profiling_compressed)) {
+		pgalloc_tag_idx idx;
+
+		idx = (page->flags >> alloc_tag_ref_offs) & alloc_tag_ref_mask;
+		idx_to_ref(idx, ref);
+		handle->page = page;
+	} else {
+		struct page_ext *page_ext;
+		union codetag_ref *tmp;
+
+		page_ext = page_ext_get(page);
+		if (!page_ext)
+			return false;
+
+		tmp = (union codetag_ref *)page_ext_data(page_ext, &page_alloc_tagging_ops);
+		ref->ct = tmp->ct;
+		handle->ref = tmp;
+	}
 
-	tmp = (union codetag_ref *)page_ext_data(page_ext, &page_alloc_tagging_ops);
-	ref->ct = tmp->ct;
-	handle->ref = tmp;
 	return true;
 }
 
@@ -42,16 +131,35 @@ static inline void put_page_tag_ref(union pgtag_ref_handle handle)
 	if (WARN_ON(!handle.ref))
 		return;
 
-	page_ext_put((void *)handle.ref - page_alloc_tagging_ops.offset);
+	if (!static_key_enabled(&mem_profiling_compressed))
+		page_ext_put((void *)handle.ref - page_alloc_tagging_ops.offset);
 }
 
-static inline void update_page_tag_ref(union pgtag_ref_handle handle,
-				       union codetag_ref *ref)
+static inline void update_page_tag_ref(union pgtag_ref_handle handle, union codetag_ref *ref)
 {
-	if (WARN_ON(!handle.ref || !ref))
-		return;
+	if (static_key_enabled(&mem_profiling_compressed)) {
+		struct page *page = handle.page;
+		unsigned long old_flags;
+		unsigned long flags;
+		unsigned long idx;
 
-	handle.ref->ct = ref->ct;
+		if (WARN_ON(!page || !ref))
+			return;
+
+		idx = (unsigned long)ref_to_idx(ref);
+		idx = (idx & alloc_tag_ref_mask) << alloc_tag_ref_offs;
+		do {
+			old_flags = READ_ONCE(page->flags);
+			flags = old_flags;
+			flags &= ~(alloc_tag_ref_mask << alloc_tag_ref_offs);
+			flags |= idx;
+		} while (unlikely(!try_cmpxchg(&page->flags, &old_flags, flags)));
+	} else {
+		if (WARN_ON(!handle.ref || !ref))
+			return;
+
+		handle.ref->ct = ref->ct;
+	}
 }
 
 static inline void clear_page_tag_ref(struct page *page)
@@ -122,6 +230,8 @@ static inline void pgalloc_tag_sub_pages(struct alloc_tag *tag, unsigned int nr)
 		this_cpu_sub(tag->counters->bytes, PAGE_SIZE * nr);
 }
 
+void __init alloc_tag_sec_init(void);
+
 #else /* CONFIG_MEM_ALLOC_PROFILING */
 
 static inline void clear_page_tag_ref(struct page *page) {}
@@ -130,6 +240,7 @@ static inline void pgalloc_tag_add(struct page *page, struct task_struct *task,
 static inline void pgalloc_tag_sub(struct page *page, unsigned int nr) {}
 static inline struct alloc_tag *pgalloc_tag_get(struct page *page) { return NULL; }
 static inline void pgalloc_tag_sub_pages(struct alloc_tag *tag, unsigned int nr) {}
+static inline void alloc_tag_sec_init(void) {}
 
 #endif /* CONFIG_MEM_ALLOC_PROFILING */
 

lib/alloc_tag.c

@@ -3,6 +3,7 @@
 #include <linux/execmem.h>
 #include <linux/fs.h>
 #include <linux/gfp.h>
+#include <linux/kallsyms.h>
 #include <linux/module.h>
 #include <linux/page_ext.h>
 #include <linux/proc_fs.h>
@@ -12,6 +13,8 @@
 
 #define ALLOCINFO_FILE_NAME		"allocinfo"
 #define MODULE_ALLOC_TAG_VMAP_SIZE	(100000UL * sizeof(struct alloc_tag))
+#define SECTION_START(NAME)		(CODETAG_SECTION_START_PREFIX NAME)
+#define SECTION_STOP(NAME)		(CODETAG_SECTION_STOP_PREFIX NAME)
 
 #ifdef CONFIG_MEM_ALLOC_PROFILING_ENABLED_BY_DEFAULT
 static bool mem_profiling_support = true;
@@ -26,6 +29,11 @@ EXPORT_SYMBOL(_shared_alloc_tag);
 
 DEFINE_STATIC_KEY_MAYBE(CONFIG_MEM_ALLOC_PROFILING_ENABLED_BY_DEFAULT,
 			mem_alloc_profiling_key);
+DEFINE_STATIC_KEY_FALSE(mem_profiling_compressed);
+
+struct alloc_tag_kernel_section kernel_tags = { NULL, 0 };
+unsigned long alloc_tag_ref_mask;
+int alloc_tag_ref_offs;
 
 struct allocinfo_private {
 	struct codetag_iterator iter;
@@ -155,7 +163,7 @@ size_t alloc_tag_top_users(struct codetag_bytes *tags, size_t count, bool can_sl
 	return nr;
 }
 
-static void shutdown_mem_profiling(void)
+static void shutdown_mem_profiling(bool remove_file)
 {
 	if (mem_alloc_profiling_enabled())
 		static_branch_disable(&mem_alloc_profiling_key);
@@ -163,6 +171,8 @@ static void shutdown_mem_profiling(void)
 	if (!mem_profiling_support)
 		return;
 
+	if (remove_file)
+		remove_proc_entry(ALLOCINFO_FILE_NAME, NULL);
 	mem_profiling_support = false;
 }
 
@@ -173,10 +183,40 @@ static void __init procfs_init(void)
 
 	if (!proc_create_seq(ALLOCINFO_FILE_NAME, 0400, NULL, &allocinfo_seq_op)) {
 		pr_err("Failed to create %s file\n", ALLOCINFO_FILE_NAME);
-		shutdown_mem_profiling();
+		shutdown_mem_profiling(false);
 	}
 }
 
+void __init alloc_tag_sec_init(void)
+{
+	struct alloc_tag *last_codetag;
+
+	if (!mem_profiling_support)
+		return;
+
+	if (!static_key_enabled(&mem_profiling_compressed))
+		return;
+
+	kernel_tags.first_tag = (struct alloc_tag *)kallsyms_lookup_name(
+					SECTION_START(ALLOC_TAG_SECTION_NAME));
+	last_codetag = (struct alloc_tag *)kallsyms_lookup_name(
+					SECTION_STOP(ALLOC_TAG_SECTION_NAME));
+	kernel_tags.count = last_codetag - kernel_tags.first_tag;
+
+	/* Check if kernel tags fit into page flags */
+	if (kernel_tags.count > (1UL << NR_UNUSED_PAGEFLAG_BITS)) {
+		shutdown_mem_profiling(false); /* allocinfo file does not exist yet */
+		pr_err("%lu allocation tags cannot be references using %d available page flag bits. Memory allocation profiling is disabled!\n",
+			kernel_tags.count, NR_UNUSED_PAGEFLAG_BITS);
+		return;
+	}
+
+	alloc_tag_ref_offs = (LRU_REFS_PGOFF - NR_UNUSED_PAGEFLAG_BITS);
+	alloc_tag_ref_mask = ((1UL << NR_UNUSED_PAGEFLAG_BITS) - 1);
+	pr_debug("Memory allocation profiling compression is using %d page flag bits!\n",
+		 NR_UNUSED_PAGEFLAG_BITS);
+}
+
 #ifdef CONFIG_MODULES
 
 static struct maple_tree mod_area_mt = MTREE_INIT(mod_area_mt, MT_FLAGS_ALLOC_RANGE);
@@ -186,10 +226,59 @@ static struct module unloaded_mod;
 /* A dummy object used to indicate a module prepended area */
 static struct module prepend_mod;
 
-static struct alloc_tag_module_section module_tags;
+struct alloc_tag_module_section module_tags;
+
+static inline unsigned long alloc_tag_align(unsigned long val)
+{
+	if (!static_key_enabled(&mem_profiling_compressed)) {
+		/* No alignment requirements when we are not indexing the tags */
+		return val;
+	}
+
+	if (val % sizeof(struct alloc_tag) == 0)
+		return val;
+	return ((val / sizeof(struct alloc_tag)) + 1) * sizeof(struct alloc_tag);
+}
+
+static bool ensure_alignment(unsigned long align, unsigned int *prepend)
+{
+	if (!static_key_enabled(&mem_profiling_compressed)) {
+		/* No alignment requirements when we are not indexing the tags */
+		return true;
+	}
+
+	/*
+	 * If alloc_tag size is not a multiple of required alignment, tag
+	 * indexing does not work.
+	 */
+	if (!IS_ALIGNED(sizeof(struct alloc_tag), align))
+		return false;
+
+	/* Ensure prepend consumes multiple of alloc_tag-sized blocks */
+	if (*prepend)
+		*prepend = alloc_tag_align(*prepend);
+
+	return true;
+}
+
+static inline bool tags_addressable(void)
+{
+	unsigned long tag_idx_count;
+
+	if (!static_key_enabled(&mem_profiling_compressed))
+		return true; /* with page_ext tags are always addressable */
+
+	tag_idx_count = CODETAG_ID_FIRST + kernel_tags.count +
+			module_tags.size / sizeof(struct alloc_tag);
+
+	return tag_idx_count < (1UL << NR_UNUSED_PAGEFLAG_BITS);
+}
 
 static bool needs_section_mem(struct module *mod, unsigned long size)
 {
+	if (!mem_profiling_support)
+		return false;
+
 	return size >= sizeof(struct alloc_tag);
 }
 
@@ -300,6 +389,13 @@ static void *reserve_module_tags(struct module *mod, unsigned long size,
 	if (!align)
 		align = 1;
 
+	if (!ensure_alignment(align, &prepend)) {
+		shutdown_mem_profiling(true);
+		pr_err("%s: alignment %lu is incompatible with allocation tag indexing. Memory allocation profiling is disabled!\n",
+			mod->name, align);
+		return ERR_PTR(-EINVAL);
+	}
+
 	mas_lock(&mas);
 	if (!find_aligned_area(&mas, section_size, size, prepend, align)) {
 		ret = ERR_PTR(-ENOMEM);
@@ -343,9 +439,15 @@ unlock:
 		int grow_res;
 
 		module_tags.size = offset + size;
+		if (mem_alloc_profiling_enabled() && !tags_addressable()) {
+			shutdown_mem_profiling(true);
+			pr_warn("With module %s there are too many tags to fit in %d page flag bits. Memory allocation profiling is disabled!\n",
+				mod->name, NR_UNUSED_PAGEFLAG_BITS);
+		}
+
 		grow_res = vm_module_tags_populate();
 		if (grow_res) {
-			shutdown_mem_profiling();
+			shutdown_mem_profiling(true);
 			pr_err("Failed to allocate memory for allocation tags in the module %s. Memory allocation profiling is disabled!\n",
 			       mod->name);
 			return ERR_PTR(grow_res);
@@ -429,6 +531,8 @@ static int __init alloc_mod_tags_mem(void)
 
 	module_tags.start_addr = (unsigned long)vm_module_tags->addr;
 	module_tags.end_addr = module_tags.start_addr + MODULE_ALLOC_TAG_VMAP_SIZE;
+	/* Ensure the base is alloc_tag aligned when required for indexing */
+	module_tags.start_addr = alloc_tag_align(module_tags.start_addr);
 
 	return 0;
 }
@@ -451,8 +555,10 @@ static inline void free_mod_tags_mem(void) {}
 
 #endif /* CONFIG_MODULES */
 
+/* See: Documentation/mm/allocation-profiling.rst */
 static int __init setup_early_mem_profiling(char *str)
 {
+	bool compressed = false;
 	bool enable;
 
 	if (!str || !str[0])
@@ -461,22 +567,37 @@ static int __init setup_early_mem_profiling(char *str)
 	if (!strncmp(str, "never", 5)) {
 		enable = false;
 		mem_profiling_support = false;
+		pr_info("Memory allocation profiling is disabled!\n");
 	} else {
-		int res;
+		char *token = strsep(&str, ",");
 
-		res = kstrtobool(str, &enable);
-		if (res)
-			return res;
+		if (kstrtobool(token, &enable))
+			return -EINVAL;
 
+		if (str) {
+
+			if (strcmp(str, "compressed"))
+				return -EINVAL;
+
+			compressed = true;
+		}
 		mem_profiling_support = true;
+		pr_info("Memory allocation profiling is enabled %s compression and is turned %s!\n",
+			compressed ? "with" : "without", enable ? "on" : "off");
 	}
 
-	if (enable != static_key_enabled(&mem_alloc_profiling_key)) {
+	if (enable != mem_alloc_profiling_enabled()) {
 		if (enable)
 			static_branch_enable(&mem_alloc_profiling_key);
 		else
 			static_branch_disable(&mem_alloc_profiling_key);
 	}
+	if (compressed != static_key_enabled(&mem_profiling_compressed)) {
+		if (compressed)
+			static_branch_enable(&mem_profiling_compressed);
+		else
+			static_branch_disable(&mem_profiling_compressed);
+	}
 
 	return 0;
 }
@@ -484,6 +605,9 @@ early_param("sysctl.vm.mem_profiling", setup_early_mem_profiling);
 
 static __init bool need_page_alloc_tagging(void)
 {
+	if (static_key_enabled(&mem_profiling_compressed))
+		return false;
+
 	return mem_profiling_support;
 }
 

lib/codetag.c

@@ -149,8 +149,8 @@ static struct codetag_range get_section_range(struct module *mod,
 					      const char *section)
 {
 	return (struct codetag_range) {
-		get_symbol(mod, "__start_", section),
-		get_symbol(mod, "__stop_", section),
+		get_symbol(mod, CODETAG_SECTION_START_PREFIX, section),
+		get_symbol(mod, CODETAG_SECTION_STOP_PREFIX, section),
 	};
 }
 

mm/mm_init.c

@@ -83,8 +83,7 @@ void __init mminit_verify_pageflags_layout(void)
 	unsigned long or_mask, add_mask;
 
 	shift = BITS_PER_LONG;
-	width = shift - SECTIONS_WIDTH - NODES_WIDTH - ZONES_WIDTH
-		- LAST_CPUPID_SHIFT - KASAN_TAG_WIDTH - LRU_GEN_WIDTH - LRU_REFS_WIDTH;
+	width = shift - NR_NON_PAGEFLAG_BITS;
 	mminit_dprintk(MMINIT_TRACE, "pageflags_layout_widths",
 		"Section %d Node %d Zone %d Lastcpupid %d Kasantag %d Gen %d Tier %d Flags %d\n",
 		SECTIONS_WIDTH,
@@ -2639,7 +2638,7 @@ void __init mm_core_init(void)
 	BUILD_BUG_ON(MAX_ZONELISTS > 2);
 	build_all_zonelists(NULL);
 	page_alloc_init_cpuhp();
-
+	alloc_tag_sec_init();
 	/*
 	 * page_ext requires contiguous pages,
 	 * bigger than MAX_PAGE_ORDER unless SPARSEMEM.