Merge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar: "Microcode fixes, a Xen fix and a KASLR boot loading fix with certain memory layouts" * 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: x86, microcode, AMD: Fix ucode patch stashing on 32-bit x86/core, x86/xen/smp: Use 'die_complete' completion when taking CPU down x86, microcode: Fix accessing dis_ucode_ldr on 32-bit x86, kaslr: Prevent .bss from overlaping initrd x86, microcode, AMD: Fix early ucode loading on 32-bit

Merge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Ingo Molnar: "Microcode fixes, a Xen fix and a KASLR boot loading fix with certain memory layouts" * 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: x86, microcode, AMD: Fix ucode patch stashing on 32-bit x86/core, x86/xen/smp: Use 'die_complete' completion when taking CPU down x86, microcode: Fix accessing dis_ucode_ldr on 32-bit x86, kaslr: Prevent .bss from overlaping initrd x86, microcode, AMD: Fix early ucode loading on 32-bit
Linus Torvalds
2 parents 3b91270a0a 0cafa3e714
Showing 11 changed files Side-by-side Diff
arch/x86/boot/compressed/Makefile
arch/x86/boot/compressed/head_32.S
arch/x86/boot/compressed/head_64.S
arch/x86/boot/compressed/misc.c
arch/x86/boot/compressed/mkpiggy.c
arch/x86/include/asm/smp.h
arch/x86/kernel/cpu/microcode/amd_early.c
arch/x86/kernel/cpu/microcode/core_early.c
arch/x86/kernel/smpboot.c
arch/x86/tools/calc_run_size.pl
arch/x86/xen/smp.c
@@ -76,8 +76,10 @@
 suffix-$(CONFIG_KERNEL_LZO) 	:= lzo
 suffix-$(CONFIG_KERNEL_LZ4) 	:= lz4
  
+RUN_SIZE = $(shell objdump -h vmlinux | \
+	     perl $(srctree)/arch/x86/tools/calc_run_size.pl)
 quiet_cmd_mkpiggy = MKPIGGY $@
-      cmd_mkpiggy = $(obj)/mkpiggy $< > $@ || ( rm -f $@ ; false )
+      cmd_mkpiggy = $(obj)/mkpiggy $< $(RUN_SIZE) > $@ || ( rm -f $@ ; false )
  
 targets += piggy.S
 $(obj)/piggy.S: $(obj)/vmlinux.bin.$(suffix-y) $(obj)/mkpiggy FORCE
@@ -207,7 +207,8 @@
  * Do the decompression, and jump to the new kernel..
  */
 				/* push arguments for decompress_kernel: */
-	pushl	$z_output_len	/* decompressed length */
+	pushl	$z_run_size	/* size of kernel with .bss and .brk */
+	pushl	$z_output_len	/* decompressed length, end of relocs */
 	leal	z_extract_offset_negative(%ebx), %ebp
 	pushl	%ebp		/* output address */
 	pushl	$z_input_len	/* input_len */
@@ -217,7 +218,7 @@
 	pushl	%eax		/* heap area */
 	pushl	%esi		/* real mode pointer */
 	call	decompress_kernel /* returns kernel location in %eax */
-	addl	$24, %esp
+	addl	$28, %esp
  
 /*
  * Jump to the decompressed kernel.
@@ -402,13 +402,16 @@
  * Do the decompression, and jump to the new kernel..
  */
 	pushq	%rsi			/* Save the real mode argument */
+	movq	$z_run_size, %r9	/* size of kernel with .bss and .brk */
+	pushq	%r9
 	movq	%rsi, %rdi		/* real mode address */
 	leaq	boot_heap(%rip), %rsi	/* malloc area for uncompression */
 	leaq	input_data(%rip), %rdx  /* input_data */
 	movl	$z_input_len, %ecx	/* input_len */
 	movq	%rbp, %r8		/* output target address */
-	movq	$z_output_len, %r9	/* decompressed length */
+	movq	$z_output_len, %r9	/* decompressed length, end of relocs */
 	call	decompress_kernel	/* returns kernel location in %rax */
+	popq	%r9
 	popq	%rsi
  
 /*
@@ -358,7 +358,8 @@
 				  unsigned char *input_data,
 				  unsigned long input_len,
 				  unsigned char *output,
-				  unsigned long output_len)
+				  unsigned long output_len,
+				  unsigned long run_size)
 {
 	real_mode = rmode;
  
@@ -381,8 +382,14 @@
 	free_mem_ptr     = heap;	/* Heap */
 	free_mem_end_ptr = heap + BOOT_HEAP_SIZE;
  
-	output = choose_kernel_location(input_data, input_len,
-					output, output_len);
+	/*
+	 * The memory hole needed for the kernel is the larger of either
+	 * the entire decompressed kernel plus relocation table, or the
+	 * entire decompressed kernel plus .bss and .brk sections.
+	 */
+	output = choose_kernel_location(input_data, input_len, output,
+					output_len > run_size ? output_len
+							      : run_size);
  
 	/* Validate memory location choices. */
 	if ((unsigned long)output & (MIN_KERNEL_ALIGN - 1))
@@ -36,11 +36,13 @@
 	uint32_t olen;
 	long ilen;
 	unsigned long offs;
+	unsigned long run_size;
 	FILE *f = NULL;
 	int retval = 1;
  
-	if (argc < 2) {
-		fprintf(stderr, "Usage: %s compressed_file\n", argv[0]);
+	if (argc < 3) {
+		fprintf(stderr, "Usage: %s compressed_file run_size\n",
+				argv[0]);
 		goto bail;
 	}
  
@@ -74,6 +76,7 @@
 	offs += olen >> 12;	/* Add 8 bytes for each 32K block */
 	offs += 64*1024 + 128;	/* Add 64K + 128 bytes slack */
 	offs = (offs+4095) & ~4095; /* Round to a 4K boundary */
+	run_size = atoi(argv[2]);
  
 	printf(".section \".rodata..compressed\",\"a\",@progbits\n");
 	printf(".globl z_input_len\n");
@@ -85,6 +88,8 @@
 	/* z_extract_offset_negative allows simplification of head_32.S */
 	printf(".globl z_extract_offset_negative\n");
 	printf("z_extract_offset_negative = -0x%lx\n", offs);
+	printf(".globl z_run_size\n");
+	printf("z_run_size = %lu\n", run_size);
  
 	printf(".globl input_data, input_data_end\n");
 	printf("input_data:\n");
@@ -150,6 +150,7 @@
 }
  
 void cpu_disable_common(void);
+void cpu_die_common(unsigned int cpu);
 void native_smp_prepare_boot_cpu(void);
 void native_smp_prepare_cpus(unsigned int max_cpus);
 void native_smp_cpus_done(unsigned int max_cpus);
@@ -108,12 +108,13 @@
  * load_microcode_amd() to save equivalent cpu table and microcode patches in
  * kernel heap memory.
  */
-static void apply_ucode_in_initrd(void *ucode, size_t size)
+static void apply_ucode_in_initrd(void *ucode, size_t size, bool save_patch)
 {
 	struct equiv_cpu_entry *eq;
 	size_t *cont_sz;
 	u32 *header;
 	u8  *data, **cont;
+	u8 (*patch)[PATCH_MAX_SIZE];
 	u16 eq_id = 0;
 	int offset, left;
 	u32 rev, eax, ebx, ecx, edx;
  
@@ -123,10 +124,12 @@
 	new_rev = (u32 *)__pa_nodebug(&ucode_new_rev);
 	cont_sz = (size_t *)__pa_nodebug(&container_size);
 	cont	= (u8 **)__pa_nodebug(&container);
+	patch	= (u8 (*)[PATCH_MAX_SIZE])__pa_nodebug(&amd_ucode_patch);
 #else
 	new_rev = &ucode_new_rev;
 	cont_sz = &container_size;
 	cont	= &container;
+	patch	= &amd_ucode_patch;
 #endif
  
 	data   = ucode;
@@ -213,9 +216,9 @@
 				rev = mc->hdr.patch_id;
 				*new_rev = rev;
  
-				/* save ucode patch */
-				memcpy(amd_ucode_patch, mc,
-				       min_t(u32, header[1], PATCH_MAX_SIZE));
+				if (save_patch)
+					memcpy(patch, mc,
+					       min_t(u32, header[1], PATCH_MAX_SIZE));
 			}
 		}
  
@@ -246,7 +249,7 @@
 	*data = cp.data;
 	*size = cp.size;
  
-	apply_ucode_in_initrd(cp.data, cp.size);
+	apply_ucode_in_initrd(cp.data, cp.size, true);
 }
  
 #ifdef CONFIG_X86_32
@@ -263,7 +266,7 @@
 	size_t *usize;
 	void **ucode;
  
-	mc = (struct microcode_amd *)__pa(amd_ucode_patch);
+	mc = (struct microcode_amd *)__pa_nodebug(amd_ucode_patch);
 	if (mc->hdr.patch_id && mc->hdr.processor_rev_id) {
 		__apply_microcode_amd(mc);
 		return;
@@ -275,7 +278,7 @@
 	if (!*ucode || !*usize)
 		return;
  
-	apply_ucode_in_initrd(*ucode, *usize);
+	apply_ucode_in_initrd(*ucode, *usize, false);
 }
  
 static void __init collect_cpu_sig_on_bsp(void *arg)
@@ -339,7 +342,7 @@
 		 * AP has a different equivalence ID than BSP, looks like
 		 * mixed-steppings silicon so go through the ucode blob anew.
 		 */
-		apply_ucode_in_initrd(ucode_cpio.data, ucode_cpio.size);
+		apply_ucode_in_initrd(ucode_cpio.data, ucode_cpio.size, false);
 	}
 }
 #endif
  
@@ -347,7 +350,9 @@
 int __init save_microcode_in_initrd_amd(void)
 {
 	unsigned long cont;
+	int retval = 0;
 	enum ucode_state ret;
+	u8 *cont_va;
 	u32 eax;
  
 	if (!container)
  
@@ -355,13 +360,15 @@
  
 #ifdef CONFIG_X86_32
 	get_bsp_sig();
-	cont = (unsigned long)container;
+	cont	= (unsigned long)container;
+	cont_va = __va(container);
 #else
 	/*
 	 * We need the physical address of the container for both bitness since
 	 * boot_params.hdr.ramdisk_image is a physical address.
 	 */
-	cont = __pa(container);
+	cont    = __pa(container);
+	cont_va = container;
 #endif
  
 	/*
@@ -372,6 +379,8 @@
 	if (relocated_ramdisk)
 		container = (u8 *)(__va(relocated_ramdisk) +
 			     (cont - boot_params.hdr.ramdisk_image));
+	else
+		container = cont_va;
  
 	if (ucode_new_rev)
 		pr_info("microcode: updated early to new patch_level=0x%08x\n",
@@ -382,7 +391,7 @@
  
 	ret = load_microcode_amd(eax, container, container_size);
 	if (ret != UCODE_OK)
-		return -EINVAL;
+		retval = -EINVAL;
  
 	/*
 	 * This will be freed any msec now, stash patches for the current
@@ -391,6 +400,6 @@
 	container = NULL;
 	container_size = 0;
  
-	return 0;
+	return retval;
 }
@@ -124,7 +124,7 @@
 static bool check_loader_disabled_ap(void)
 {
 #ifdef CONFIG_X86_32
-	return __pa_nodebug(dis_ucode_ldr);
+	return *((bool *)__pa_nodebug(&dis_ucode_ldr));
 #else
 	return dis_ucode_ldr;
 #endif
@@ -1303,10 +1303,14 @@
 	numa_remove_cpu(cpu);
 }
  
+static DEFINE_PER_CPU(struct completion, die_complete);
+
 void cpu_disable_common(void)
 {
 	int cpu = smp_processor_id();
  
+	init_completion(&per_cpu(die_complete, smp_processor_id()));
+
 	remove_siblinginfo(cpu);
  
 	/* It's now safe to remove this processor from the online map */
@@ -1316,8 +1320,6 @@
 	fixup_irqs();
 }
  
-static DEFINE_PER_CPU(struct completion, die_complete);
-
 int native_cpu_disable(void)
 {
 	int ret;
  
  
@@ -1327,16 +1329,21 @@
 		return ret;
  
 	clear_local_APIC();
-	init_completion(&per_cpu(die_complete, smp_processor_id()));
 	cpu_disable_common();
  
 	return 0;
 }
  
+void cpu_die_common(unsigned int cpu)
+{
+	wait_for_completion_timeout(&per_cpu(die_complete, cpu), HZ);
+}
+
 void native_cpu_die(unsigned int cpu)
 {
 	/* We don't do anything here: idle task is faking death itself. */
-	wait_for_completion_timeout(&per_cpu(die_complete, cpu), HZ);
+
+	cpu_die_common(cpu);
  
 	/* They ack this in play_dead() by setting CPU_DEAD */
 	if (per_cpu(cpu_state, cpu) == CPU_DEAD) {
+#!/usr/bin/perl
+#
+# Calculate the amount of space needed to run the kernel, including room for
+# the .bss and .brk sections.
+#
+# Usage:
+# objdump -h a.out | perl calc_run_size.pl
+use strict;
+
+my $mem_size = 0;
+my $file_offset = 0;
+
+my $sections=" *[0-9]+ \.(?:bss|brk) +";
+while (<>) {
+	if (/^$sections([0-9a-f]+) +(?:[0-9a-f]+ +){2}([0-9a-f]+)/) {
+		my $size = hex($1);
+		my $offset = hex($2);
+		$mem_size += $size;
+		if ($file_offset == 0) {
+			$file_offset = $offset;
+		} elsif ($file_offset != $offset) {
+			die ".bss and .brk lack common file offset\n";
+		}
+	}
+}
+
+if ($file_offset == 0) {
+	die "Never found .bss or .brk file offset\n";
+}
+printf("%d\n", $mem_size + $file_offset);
@@ -510,6 +510,9 @@
 		current->state = TASK_UNINTERRUPTIBLE;
 		schedule_timeout(HZ/10);
 	}
+
+	cpu_die_common(cpu);
+
 	xen_smp_intr_free(cpu);
 	xen_uninit_lock_cpu(cpu);
 	xen_teardown_timer(cpu);
...	...	@@ -76,8 +76,10 @@
76	76	suffix-$(CONFIG_KERNEL_LZO) := lzo
77	77	suffix-$(CONFIG_KERNEL_LZ4) := lz4
78	78
	79	+RUN_SIZE = $(shell objdump -h vmlinux \| \
	80	+ perl $(srctree)/arch/x86/tools/calc_run_size.pl)
79	81	quiet_cmd_mkpiggy = MKPIGGY $@
80		- cmd_mkpiggy = $(obj)/mkpiggy $< > $@ \|\| ( rm -f $@ ; false )
	82	+ cmd_mkpiggy = $(obj)/mkpiggy $< $(RUN_SIZE) > $@ \|\| ( rm -f $@ ; false )
81	83
82	84	targets += piggy.S
83	85	$(obj)/piggy.S: $(obj)/vmlinux.bin.$(suffix-y) $(obj)/mkpiggy FORCE
...	...	@@ -207,7 +207,8 @@
207	207	* Do the decompression, and jump to the new kernel..
208	208	*/
209	209	/* push arguments for decompress_kernel: */
210		- pushl $z_output_len /* decompressed length */
	210	+ pushl $z_run_size /* size of kernel with .bss and .brk */
	211	+ pushl $z_output_len /* decompressed length, end of relocs */
211	212	leal z_extract_offset_negative(%ebx), %ebp
212	213	pushl %ebp /* output address */
213	214	pushl $z_input_len /* input_len */
...	...	@@ -217,7 +218,7 @@
217	218	pushl %eax /* heap area */
218	219	pushl %esi /* real mode pointer */
219	220	call decompress_kernel /* returns kernel location in %eax */
220		- addl $24, %esp
	221	+ addl $28, %esp
221	222
222	223	/*
223	224	* Jump to the decompressed kernel.
...	...	@@ -402,13 +402,16 @@
402	402	* Do the decompression, and jump to the new kernel..
403	403	*/
404	404	pushq %rsi /* Save the real mode argument */
	405	+ movq $z_run_size, %r9 /* size of kernel with .bss and .brk */
	406	+ pushq %r9
405	407	movq %rsi, %rdi /* real mode address */
406	408	leaq boot_heap(%rip), %rsi /* malloc area for uncompression */
407	409	leaq input_data(%rip), %rdx /* input_data */
408	410	movl $z_input_len, %ecx /* input_len */
409	411	movq %rbp, %r8 /* output target address */
410		- movq $z_output_len, %r9 /* decompressed length */
	412	+ movq $z_output_len, %r9 /* decompressed length, end of relocs */
411	413	call decompress_kernel /* returns kernel location in %rax */
	414	+ popq %r9
412	415	popq %rsi
413	416
414	417	/*
...	...	@@ -358,7 +358,8 @@
358	358	unsigned char *input_data,
359	359	unsigned long input_len,
360	360	unsigned char *output,
361		- unsigned long output_len)
	361	+ unsigned long output_len,
	362	+ unsigned long run_size)
362	363	{
363	364	real_mode = rmode;
364	365
...	...	@@ -381,8 +382,14 @@
381	382	free_mem_ptr = heap; /* Heap */
382	383	free_mem_end_ptr = heap + BOOT_HEAP_SIZE;
383	384
384		- output = choose_kernel_location(input_data, input_len,
385		- output, output_len);
	385	+ /*
	386	+ * The memory hole needed for the kernel is the larger of either
	387	+ * the entire decompressed kernel plus relocation table, or the
	388	+ * entire decompressed kernel plus .bss and .brk sections.
	389	+ */
	390	+ output = choose_kernel_location(input_data, input_len, output,
	391	+ output_len > run_size ? output_len
	392	+ : run_size);
386	393
387	394	/* Validate memory location choices. */
388	395	if ((unsigned long)output & (MIN_KERNEL_ALIGN - 1))
...	...	@@ -36,11 +36,13 @@
36	36	uint32_t olen;
37	37	long ilen;
38	38	unsigned long offs;
	39	+ unsigned long run_size;
39	40	FILE *f = NULL;
40	41	int retval = 1;
41	42
42		- if (argc < 2) {
43		- fprintf(stderr, "Usage: %s compressed_file\n", argv[0]);
	43	+ if (argc < 3) {
	44	+ fprintf(stderr, "Usage: %s compressed_file run_size\n",
	45	+ argv[0]);
44	46	goto bail;
45	47	}
46	48
...	...	@@ -74,6 +76,7 @@
74	76	offs += olen >> 12; /* Add 8 bytes for each 32K block */
75	77	offs += 641024 + 128; / Add 64K + 128 bytes slack */
76	78	offs = (offs+4095) & ~4095; /* Round to a 4K boundary */
	79	+ run_size = atoi(argv[2]);
77	80
78	81	printf(".section \".rodata..compressed\",\"a\",@progbits\n");
79	82	printf(".globl z_input_len\n");
...	...	@@ -85,6 +88,8 @@
85	88	/* z_extract_offset_negative allows simplification of head_32.S */
86	89	printf(".globl z_extract_offset_negative\n");
87	90	printf("z_extract_offset_negative = -0x%lx\n", offs);
	91	+ printf(".globl z_run_size\n");
	92	+ printf("z_run_size = %lu\n", run_size);
88	93
89	94	printf(".globl input_data, input_data_end\n");
90	95	printf("input_data:\n");
...	...	@@ -150,6 +150,7 @@
150	150	}
151	151
152	152	void cpu_disable_common(void);
	153	+void cpu_die_common(unsigned int cpu);
153	154	void native_smp_prepare_boot_cpu(void);
154	155	void native_smp_prepare_cpus(unsigned int max_cpus);
155	156	void native_smp_cpus_done(unsigned int max_cpus);
...	...	@@ -108,12 +108,13 @@
108	108	* load_microcode_amd() to save equivalent cpu table and microcode patches in
109	109	* kernel heap memory.
110	110	*/
111		-static void apply_ucode_in_initrd(void *ucode, size_t size)
	111	+static void apply_ucode_in_initrd(void *ucode, size_t size, bool save_patch)
112	112	{
113	113	struct equiv_cpu_entry *eq;
114	114	size_t *cont_sz;
115	115	u32 *header;
116	116	u8 data, *cont;
	117	+ u8 (*patch)[PATCH_MAX_SIZE];
117	118	u16 eq_id = 0;
118	119	int offset, left;
119	120	u32 rev, eax, ebx, ecx, edx;
120	121
...	...	@@ -123,10 +124,12 @@
123	124	new_rev = (u32 *)__pa_nodebug(&ucode_new_rev);
124	125	cont_sz = (size_t *)__pa_nodebug(&container_size);
125	126	cont = (u8 **)__pa_nodebug(&container);
	127	+ patch = (u8 (*)[PATCH_MAX_SIZE])__pa_nodebug(&amd_ucode_patch);
126	128	#else
127	129	new_rev = &ucode_new_rev;
128	130	cont_sz = &container_size;
129	131	cont = &container;
	132	+ patch = &amd_ucode_patch;
130	133	#endif
131	134
132	135	data = ucode;
...	...	@@ -213,9 +216,9 @@
213	216	rev = mc->hdr.patch_id;
214	217	*new_rev = rev;
215	218
216		- /* save ucode patch */
217		- memcpy(amd_ucode_patch, mc,
218		- min_t(u32, header[1], PATCH_MAX_SIZE));
	219	+ if (save_patch)
	220	+ memcpy(patch, mc,
	221	+ min_t(u32, header[1], PATCH_MAX_SIZE));
219	222	}
220	223	}
221	224
...	...	@@ -246,7 +249,7 @@
246	249	*data = cp.data;
247	250	*size = cp.size;
248	251
249		- apply_ucode_in_initrd(cp.data, cp.size);
	252	+ apply_ucode_in_initrd(cp.data, cp.size, true);
250	253	}
251	254
252	255	#ifdef CONFIG_X86_32
...	...	@@ -263,7 +266,7 @@
263	266	size_t *usize;
264	267	void **ucode;
265	268
266		- mc = (struct microcode_amd *)__pa(amd_ucode_patch);
	269	+ mc = (struct microcode_amd *)__pa_nodebug(amd_ucode_patch);
267	270	if (mc->hdr.patch_id && mc->hdr.processor_rev_id) {
268	271	__apply_microcode_amd(mc);
269	272	return;
...	...	@@ -275,7 +278,7 @@
275	278	if (!ucode \|\| !usize)
276	279	return;
277	280
278		- apply_ucode_in_initrd(ucode, usize);
	281	+ apply_ucode_in_initrd(ucode, usize, false);
279	282	}
280	283
281	284	static void __init collect_cpu_sig_on_bsp(void *arg)
...	...	@@ -339,7 +342,7 @@
339	342	* AP has a different equivalence ID than BSP, looks like
340	343	* mixed-steppings silicon so go through the ucode blob anew.
341	344	*/
342		- apply_ucode_in_initrd(ucode_cpio.data, ucode_cpio.size);
	345	+ apply_ucode_in_initrd(ucode_cpio.data, ucode_cpio.size, false);
343	346	}
344	347	}
345	348	#endif
346	349
...	...	@@ -347,7 +350,9 @@
347	350	int __init save_microcode_in_initrd_amd(void)
348	351	{
349	352	unsigned long cont;
	353	+ int retval = 0;
350	354	enum ucode_state ret;
	355	+ u8 *cont_va;
351	356	u32 eax;
352	357
353	358	if (!container)
354	359
...	...	@@ -355,13 +360,15 @@
355	360
356	361	#ifdef CONFIG_X86_32
357	362	get_bsp_sig();
358		- cont = (unsigned long)container;
	363	+ cont = (unsigned long)container;
	364	+ cont_va = __va(container);
359	365	#else
360	366	/*
361	367	* We need the physical address of the container for both bitness since
362	368	* boot_params.hdr.ramdisk_image is a physical address.
363	369	*/
364		- cont = __pa(container);
	370	+ cont = __pa(container);
	371	+ cont_va = container;
365	372	#endif
366	373
367	374	/*
...	...	@@ -372,6 +379,8 @@
372	379	if (relocated_ramdisk)
373	380	container = (u8 *)(__va(relocated_ramdisk) +
374	381	(cont - boot_params.hdr.ramdisk_image));
	382	+ else
	383	+ container = cont_va;
375	384
376	385	if (ucode_new_rev)
377	386	pr_info("microcode: updated early to new patch_level=0x%08x\n",
...	...	@@ -382,7 +391,7 @@
382	391
383	392	ret = load_microcode_amd(eax, container, container_size);
384	393	if (ret != UCODE_OK)
385		- return -EINVAL;
	394	+ retval = -EINVAL;
386	395
387	396	/*
388	397	* This will be freed any msec now, stash patches for the current
...	...	@@ -391,6 +400,6 @@
391	400	container = NULL;
392	401	container_size = 0;
393	402
394		- return 0;
	403	+ return retval;
395	404	}
...	...	@@ -124,7 +124,7 @@
124	124	static bool check_loader_disabled_ap(void)
125	125	{
126	126	#ifdef CONFIG_X86_32
127		- return __pa_nodebug(dis_ucode_ldr);
	127	+ return ((bool )__pa_nodebug(&dis_ucode_ldr));
128	128	#else
129	129	return dis_ucode_ldr;
130	130	#endif
...	...	@@ -1303,10 +1303,14 @@
1303	1303	numa_remove_cpu(cpu);
1304	1304	}
1305	1305
	1306	+static DEFINE_PER_CPU(struct completion, die_complete);
	1307	+
1306	1308	void cpu_disable_common(void)
1307	1309	{
1308	1310	int cpu = smp_processor_id();
1309	1311
	1312	+ init_completion(&per_cpu(die_complete, smp_processor_id()));
	1313	+
1310	1314	remove_siblinginfo(cpu);
1311	1315
1312	1316	/* It's now safe to remove this processor from the online map */
...	...	@@ -1316,8 +1320,6 @@
1316	1320	fixup_irqs();
1317	1321	}
1318	1322
1319		-static DEFINE_PER_CPU(struct completion, die_complete);
1320		-
1321	1323	int native_cpu_disable(void)
1322	1324	{
1323	1325	int ret;
1324	1326
1325	1327
...	...	@@ -1327,16 +1329,21 @@
1327	1329	return ret;
1328	1330
1329	1331	clear_local_APIC();
1330		- init_completion(&per_cpu(die_complete, smp_processor_id()));
1331	1332	cpu_disable_common();
1332	1333
1333	1334	return 0;
1334	1335	}
1335	1336
	1337	+void cpu_die_common(unsigned int cpu)
	1338	+{
	1339	+ wait_for_completion_timeout(&per_cpu(die_complete, cpu), HZ);
	1340	+}
	1341	+
1336	1342	void native_cpu_die(unsigned int cpu)
1337	1343	{
1338	1344	/* We don't do anything here: idle task is faking death itself. */
1339		- wait_for_completion_timeout(&per_cpu(die_complete, cpu), HZ);
	1345	+
	1346	+ cpu_die_common(cpu);
1340	1347
1341	1348	/* They ack this in play_dead() by setting CPU_DEAD */
1342	1349	if (per_cpu(cpu_state, cpu) == CPU_DEAD) {
	1	+#!/usr/bin/perl
	2	+#
	3	+# Calculate the amount of space needed to run the kernel, including room for
	4	+# the .bss and .brk sections.
	5	+#
	6	+# Usage:
	7	+# objdump -h a.out \| perl calc_run_size.pl
	8	+use strict;
	9	+
	10	+my $mem_size = 0;
	11	+my $file_offset = 0;
	12	+
	13	+my $sections=" *[0-9]+ \.(?:bss\|brk) +";
	14	+while (<>) {
	15	+ if (/^$sections([0-9a-f]+) +(?:[0-9a-f]+ +){2}([0-9a-f]+)/) {
	16	+ my $size = hex($1);
	17	+ my $offset = hex($2);
	18	+ $mem_size += $size;
	19	+ if ($file_offset == 0) {
	20	+ $file_offset = $offset;
	21	+ } elsif ($file_offset != $offset) {
	22	+ die ".bss and .brk lack common file offset\n";
	23	+ }
	24	+ }
	25	+}
	26	+
	27	+if ($file_offset == 0) {
	28	+ die "Never found .bss or .brk file offset\n";
	29	+}
	30	+printf("%d\n", $mem_size + $file_offset);