x86/fpu: Fix FPU context sizing boot regression, introduce dynamic task_struct

index 0637826292de..c49c5173158e 100644
@@ -189,6 +189,7 @@ union fpregs_state {
 	struct fxregs_state		fxsave;
 	struct swregs_state		soft;
 	struct xregs_state		xsave;
+	u8 __padding[PAGE_SIZE];
 };
 
 /*

@@ -198,40 +199,6 @@ union fpregs_state {
  */
 struct fpu {
 	/*
-	 * @state:
-	 *
-	 * In-memory copy of all FPU registers that we save/restore
-	 * over context switches. If the task is using the FPU then
-	 * the registers in the FPU are more recent than this state
-	 * copy. If the task context-switches away then they get
-	 * saved here and represent the FPU state.
-	 *
-	 * After context switches there may be a (short) time period
-	 * during which the in-FPU hardware registers are unchanged
-	 * and still perfectly match this state, if the tasks
-	 * scheduled afterwards are not using the FPU.
-	 *
-	 * This is the 'lazy restore' window of optimization, which
-	 * we track though 'fpu_fpregs_owner_ctx' and 'fpu->last_cpu'.
-	 *
-	 * We detect whether a subsequent task uses the FPU via setting
-	 * CR0::TS to 1, which causes any FPU use to raise a #NM fault.
-	 *
-	 * During this window, if the task gets scheduled again, we
-	 * might be able to skip having to do a restore from this
-	 * memory buffer to the hardware registers - at the cost of
-	 * incurring the overhead of #NM fault traps.
-	 *
-	 * Note that on modern CPUs that support the XSAVEOPT (or other
-	 * optimized XSAVE instructions), we don't use #NM traps anymore,
-	 * as the hardware can track whether FPU registers need saving
-	 * or not. On such CPUs we activate the non-lazy ('eagerfpu')
-	 * logic, which unconditionally saves/restores all FPU state
-	 * across context switches. (if FPU state exists.)
-	 */
-	union fpregs_state		state;
-
-	/*
 	 * @last_cpu:
 	 *
 	 * Records the last CPU on which this context was loaded into

@@ -288,6 +255,43 @@ struct fpu {
 	 * deal with bursty apps that only use the FPU for a short time:
 	 */
 	unsigned char			counter;
+	/*
+	 * @state:
+	 *
+	 * In-memory copy of all FPU registers that we save/restore
+	 * over context switches. If the task is using the FPU then
+	 * the registers in the FPU are more recent than this state
+	 * copy. If the task context-switches away then they get
+	 * saved here and represent the FPU state.
+	 *
+	 * After context switches there may be a (short) time period
+	 * during which the in-FPU hardware registers are unchanged
+	 * and still perfectly match this state, if the tasks
+	 * scheduled afterwards are not using the FPU.
+	 *
+	 * This is the 'lazy restore' window of optimization, which
+	 * we track though 'fpu_fpregs_owner_ctx' and 'fpu->last_cpu'.
+	 *
+	 * We detect whether a subsequent task uses the FPU via setting
+	 * CR0::TS to 1, which causes any FPU use to raise a #NM fault.
+	 *
+	 * During this window, if the task gets scheduled again, we
+	 * might be able to skip having to do a restore from this
+	 * memory buffer to the hardware registers - at the cost of
+	 * incurring the overhead of #NM fault traps.
+	 *
+	 * Note that on modern CPUs that support the XSAVEOPT (or other
+	 * optimized XSAVE instructions), we don't use #NM traps anymore,
+	 * as the hardware can track whether FPU registers need saving
+	 * or not. On such CPUs we activate the non-lazy ('eagerfpu')
+	 * logic, which unconditionally saves/restores all FPU state
+	 * across context switches. (if FPU state exists.)
+	 */
+	union fpregs_state		state;
+	/*
+	 * WARNING: 'state' is dynamically-sized.  Do not put
+	 * anything after it here.
+	 */
 };
 
 #endif /* _ASM_X86_FPU_H */

index 43e6519df0d5..944f1785ed0d 100644
@@ -390,9 +390,6 @@ struct thread_struct {
 #endif
 	unsigned long		gs;
 
-	/* Floating point and extended processor state */
-	struct fpu		fpu;
-
 	/* Save middle states of ptrace breakpoints */
 	struct perf_event	*ptrace_bps[HBP_NUM];
 	/* Debug status used for traps, single steps, etc... */

@@ -418,6 +415,13 @@ struct thread_struct {
 	unsigned long		iopl;
 	/* Max allowed port in the bitmap, in bytes: */
 	unsigned		io_bitmap_max;
+
+	/* Floating point and extended processor state */
+	struct fpu		fpu;
+	/*
+	 * WARNING: 'fpu' is dynamically-sized.  It *MUST* be at
+	 * the end.
+	 */
 };
 
 /*

index 32826791e675..deacbfa6b33e 100644
@@ -136,6 +136,45 @@ static void __init fpu__init_system_generic(void)
 unsigned int xstate_size;
 EXPORT_SYMBOL_GPL(xstate_size);
 
+#define CHECK_MEMBER_AT_END_OF(TYPE, MEMBER)	\
+	BUILD_BUG_ON((sizeof(TYPE) -			\
+			offsetof(TYPE, MEMBER) -	\
+			sizeof(((TYPE *)0)->MEMBER)) > 	\
+			0)				\
+
+/*
+ * We append the 'struct fpu' to the task_struct.
+ */
+int __weak arch_task_struct_size(void)
+{
+	int task_size = sizeof(struct task_struct);
+
+	/*
+	 * Subtract off the static size of the register state.
+	 * It potentially has a bunch of padding.
+	 */
+	task_size -= sizeof(((struct task_struct *)0)->thread.fpu.state);
+
+	/*
+	 * Add back the dynamically-calculated register state
+	 * size.
+	 */
+	task_size += xstate_size;
+
+	/*
+	 * We dynamically size 'struct fpu', so we require that
+	 * it be at the end of 'thread_struct' and that
+	 * 'thread_struct' be at the end of 'task_struct'.  If
+	 * you hit a compile error here, check the structure to
+	 * see if something got added to the end.
+	 */
+	CHECK_MEMBER_AT_END_OF(struct fpu, state);
+	CHECK_MEMBER_AT_END_OF(struct thread_struct, fpu);
+	CHECK_MEMBER_AT_END_OF(struct task_struct, thread);
+
+	return task_size;
+}
+
 /*
  * Set up the xstate_size based on the legacy FPU context size.
  *

index 9cad694ed7c4..975420eac105 100644
@@ -81,7 +81,7 @@ EXPORT_SYMBOL_GPL(idle_notifier_unregister);
  */
 int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src)
 {
-	*dst = *src;
+	memcpy(dst, src, arch_task_struct_size());
 
 	return fpu__copy(&dst->thread.fpu, &src->thread.fpu);
 }

index 91a4e6426321..a0fe99485687 100644
@@ -92,7 +92,7 @@ static size_t get_kcore_size(int *nphdr, size_t *elf_buflen)
 			     roundup(sizeof(CORE_STR), 4)) +
 			roundup(sizeof(struct elf_prstatus), 4) +
 			roundup(sizeof(struct elf_prpsinfo), 4) +
-			roundup(sizeof(struct task_struct), 4);
+			roundup(arch_task_struct_size(), 4);
 	*elf_buflen = PAGE_ALIGN(*elf_buflen);
 	return size + *elf_buflen;
 }

@@ -415,7 +415,7 @@ static void elf_kcore_store_hdr(char *bufp, int nphdr, int dataoff)
 	/* set up the task structure */
 	notes[2].name	= CORE_STR;
 	notes[2].type	= NT_TASKSTRUCT;
-	notes[2].datasz	= sizeof(struct task_struct);
+	notes[2].datasz	= arch_task_struct_size();
 	notes[2].data	= current;
 
 	nhdr->p_filesz	+= notesize(&notes[2]);

index ae21f1591615..e43a41d892b6 100644
@@ -1522,8 +1522,6 @@ struct task_struct {
 /* hung task detection */
 	unsigned long last_switch_count;
 #endif
-/* CPU-specific state of this task */
-	struct thread_struct thread;
 /* filesystem information */
 	struct fs_struct *fs;
 /* open file information */

@@ -1778,8 +1776,18 @@ struct task_struct {
 	unsigned long	task_state_change;
 #endif
 	int pagefault_disabled;
+/* CPU-specific state of this task */
+	struct thread_struct thread;
+/*
+ * WARNING: on x86, 'thread_struct' contains a variable-sized
+ * structure.  It *MUST* be at the end of 'task_struct'.
+ *
+ * Do not put anything below here!
+ */
 };
 
+extern int arch_task_struct_size(void);
+
 /* Future-safe accessor for struct task_struct's cpus_allowed. */
 #define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
 

index 1bfefc6f96a4..431b67a6098c 100644
@@ -287,15 +287,21 @@ static void set_max_threads(unsigned int max_threads_suggested)
 	max_threads = clamp_t(u64, threads, MIN_THREADS, MAX_THREADS);
 }
 
+int __weak arch_task_struct_size(void)
+{
+	return sizeof(struct task_struct);
+}
+
 void __init fork_init(void)
 {
+	int task_struct_size = arch_task_struct_size();
 #ifndef CONFIG_ARCH_TASK_STRUCT_ALLOCATOR
 #ifndef ARCH_MIN_TASKALIGN
 #define ARCH_MIN_TASKALIGN	L1_CACHE_BYTES
 #endif
 	/* create a slab on which task_structs can be allocated */
 	task_struct_cachep =
-		kmem_cache_create("task_struct", sizeof(struct task_struct),
+		kmem_cache_create("task_struct", task_struct_size,
 			ARCH_MIN_TASKALIGN, SLAB_PANIC | SLAB_NOTRACK, NULL);
 #endif
 

index deacbfa6b33e..0b39173dd971 100644
@@ -4,6 +4,8 @@
 #include <asm/fpu/internal.h>
 #include <asm/tlbflush.h>
 
+#include <linux/sched.h>
+
 /*
  * Initialize the TS bit in CR0 according to the style of context-switches
  * we are using:

@@ -136,16 +138,14 @@ static void __init fpu__init_system_generic(void)
 unsigned int xstate_size;
 EXPORT_SYMBOL_GPL(xstate_size);
 
-#define CHECK_MEMBER_AT_END_OF(TYPE, MEMBER)	\
-	BUILD_BUG_ON((sizeof(TYPE) -			\
-			offsetof(TYPE, MEMBER) -	\
-			sizeof(((TYPE *)0)->MEMBER)) > 	\
-			0)				\
+/* Enforce that 'MEMBER' is the last field of 'TYPE': */
+#define CHECK_MEMBER_AT_END_OF(TYPE, MEMBER) \
+	BUILD_BUG_ON(sizeof(TYPE) != offsetofend(TYPE, MEMBER))
 
 /*
- * We append the 'struct fpu' to the task_struct.
+ * We append the 'struct fpu' to the task_struct:
  */
-int __weak arch_task_struct_size(void)
+static void __init fpu__init_task_struct_size(void)
 {
 	int task_size = sizeof(struct task_struct);
 

@@ -172,7 +172,7 @@ int __weak arch_task_struct_size(void)
 	CHECK_MEMBER_AT_END_OF(struct thread_struct, fpu);
 	CHECK_MEMBER_AT_END_OF(struct task_struct, thread);
 
-	return task_size;
+	arch_task_struct_size = task_size;
 }
 
 /*

@@ -326,6 +326,7 @@ void __init fpu__init_system(struct cpuinfo_x86 *c)
 	fpu__init_system_generic();
 	fpu__init_system_xstate_size_legacy();
 	fpu__init_system_xstate();
+	fpu__init_task_struct_size();
 
 	fpu__init_system_ctx_switch();
 }

index 975420eac105..397688beed4b 100644
@@ -81,7 +81,7 @@ EXPORT_SYMBOL_GPL(idle_notifier_unregister);
  */
 int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src)
 {
-	memcpy(dst, src, arch_task_struct_size());
+	memcpy(dst, src, arch_task_struct_size);
 
 	return fpu__copy(&dst->thread.fpu, &src->thread.fpu);
 }

index a0fe99485687..92e6726f6e37 100644
@@ -92,7 +92,7 @@ static size_t get_kcore_size(int *nphdr, size_t *elf_buflen)
 			     roundup(sizeof(CORE_STR), 4)) +
 			roundup(sizeof(struct elf_prstatus), 4) +
 			roundup(sizeof(struct elf_prpsinfo), 4) +
-			roundup(arch_task_struct_size(), 4);
+			roundup(arch_task_struct_size, 4);
 	*elf_buflen = PAGE_ALIGN(*elf_buflen);
 	return size + *elf_buflen;
 }

@@ -415,7 +415,7 @@ static void elf_kcore_store_hdr(char *bufp, int nphdr, int dataoff)
 	/* set up the task structure */
 	notes[2].name	= CORE_STR;
 	notes[2].type	= NT_TASKSTRUCT;
-	notes[2].datasz	= arch_task_struct_size();
+	notes[2].datasz	= arch_task_struct_size;
 	notes[2].data	= current;
 
 	nhdr->p_filesz	+= notesize(&notes[2]);

index e43a41d892b6..04b5ada460b4 100644
@@ -1786,7 +1786,11 @@ struct task_struct {
  */
 };
 
-extern int arch_task_struct_size(void);
+#ifdef CONFIG_ARCH_WANTS_DYNAMIC_TASK_STRUCT
+extern int arch_task_struct_size __read_mostly;
+#else
+# define arch_task_struct_size (sizeof(struct task_struct))
+#endif
 
 /* Future-safe accessor for struct task_struct's cpus_allowed. */
 #define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)

index 431b67a6098c..dbd9b8d7b7cc 100644
@@ -287,21 +287,20 @@ static void set_max_threads(unsigned int max_threads_suggested)
 	max_threads = clamp_t(u64, threads, MIN_THREADS, MAX_THREADS);
 }
 
-int __weak arch_task_struct_size(void)
-{
-	return sizeof(struct task_struct);
-}
+#ifdef CONFIG_ARCH_WANTS_DYNAMIC_TASK_STRUCT
+/* Initialized by the architecture: */
+int arch_task_struct_size __read_mostly;
+#endif
 
 void __init fork_init(void)
 {
-	int task_struct_size = arch_task_struct_size();
 #ifndef CONFIG_ARCH_TASK_STRUCT_ALLOCATOR
 #ifndef ARCH_MIN_TASKALIGN
 #define ARCH_MIN_TASKALIGN	L1_CACHE_BYTES
 #endif
 	/* create a slab on which task_structs can be allocated */
 	task_struct_cachep =
-		kmem_cache_create("task_struct", task_struct_size,
+		kmem_cache_create("task_struct", arch_task_struct_size,
 			ARCH_MIN_TASKALIGN, SLAB_PANIC | SLAB_NOTRACK, NULL);
 #endif