Skip to contentSkip to conversation

x86 fixes

GIT PULL Ingo Molnar 1 participant 1 message Jun 26, 2009

Subsystems (8)
lore
Share & tools
lore
#0 lore raw

Linus,

Please pull the latest x86-fixes-for-linus git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.git x86-fixes-for-linus

out-of-topic modifications in x86-fixes-for-linus:
--------------------------------------------------

include/linux/kernel.h             # 5211a24: x86: Add sysctl to allow panic on
kernel/sysctl.c                    # 5211a24: x86: Add sysctl to allow panic on
mm/percpu.c                        # fa8a709: x86: implement percpu_alloc kerne

# 85ae87c: percpu: fix too lazy vunmap cache

This is somewhat larger than usual due to the somewhat late percpu/PAT fixes.

Thanks,

Ingo

------------------>
Andreas Herrmann (1):
x86: Set cpu_llc_id on AMD CPUs

Cliff Wickman (1):
x86: Fix uv bau sending buffer initialization

H. Peter Anvin (1):
x86, setup: correct include file in <asm/boot.h>

Hidetoshi Seto (2):
x86, mce: Fix mce resume on 32bit
x86, mce: percpu mcheck_timer should be pinned

Kurt Garloff (1):
x86: Add sysctl to allow panic on IOCK NMI error

Pallipadi, Venkatesh (1):
x86, delay: tsc based udelay should have rdtsc_barrier

Pekka J Enberg (1):
x86: Move init_gbpages() to setup_arch()

Robert P. J. Day (1):

      x86, setup: Fix typo "CONFIG_x86_64" in <asm/boot.h>

Tejun Heo (8):
percpu: fix too lazy vunmap cache flushing
x86: fix duplicate free in setup_pcpu_remap() failure path
x86: rename remap percpu first chunk allocator to lpage
x86: prepare setup_pcpu_lpage() for pageattr fix
x86: reorganize cpa_process_alias()
x86: fix pageattr handling for lpage percpu allocator and re-enable it
x86: implement percpu_alloc kernel parameter
x86: ensure percpu lpage doesn't consume too much vmalloc space

 Documentation/kernel-parameters.txt |    6 +
 arch/x86/include/asm/boot.h         |    6 +-
 arch/x86/include/asm/percpu.h       |   10 ++
 arch/x86/kernel/cpu/amd.c           |    4 +-
 arch/x86/kernel/cpu/mcheck/mce.c    |    4 +-
 arch/x86/kernel/dumpstack.c         |    1 +
 arch/x86/kernel/setup.c             |   16 +++
 arch/x86/kernel/setup_percpu.c      |  219 ++++++++++++++++++++++++++---------
 arch/x86/kernel/tlb_uv.c            |    9 +-
 arch/x86/kernel/traps.c             |    3 +
 arch/x86/lib/delay.c                |    3 +
 arch/x86/mm/init.c                  |   17 ---
 arch/x86/mm/pageattr.c              |   65 +++++++----
 arch/x86/power/cpu.c                |    2 +-
 include/linux/kernel.h              |    1 +
 kernel/sysctl.c                     |    8 ++
 mm/percpu.c                         |   24 +++--
 17 files changed, 275 insertions(+), 123 deletions(-)
Documentation/kernel-parameters.txt +6 -0
index 08def8d..ecad946 100644
@@ -1882,6 +1882,12 @@ and is between 256 and 4096 characters. It is defined in the file
 			Format: { 0 | 1 }
 			See arch/parisc/kernel/pdc_chassis.c
 
+	percpu_alloc=	[X86] Select which percpu first chunk allocator to use.
+			Allowed values are one of "lpage", "embed" and "4k".
+			See comments in arch/x86/kernel/setup_percpu.c for
+			details on each allocator.  This parameter is primarily
+			for debugging and performance comparison.
+
 	pf.		[PARIDE]
 			See Documentation/blockdev/paride.txt.
 
arch/x86/include/asm/boot.h +3 -3
index 418e632..7a10659 100644
@@ -8,7 +8,7 @@
 
 #ifdef __KERNEL__
 
-#include <asm/page_types.h>
+#include <asm/pgtable_types.h>
 
 /* Physical address where kernel should be loaded. */
 #define LOAD_PHYSICAL_ADDR ((CONFIG_PHYSICAL_START \
@@ -16,10 +16,10 @@
 				& ~(CONFIG_PHYSICAL_ALIGN - 1))
 
 /* Minimum kernel alignment, as a power of two */
-#ifdef CONFIG_x86_64
+#ifdef CONFIG_X86_64
 #define MIN_KERNEL_ALIGN_LG2	PMD_SHIFT
 #else
-#define MIN_KERNEL_ALIGN_LG2	(PAGE_SHIFT+1)
+#define MIN_KERNEL_ALIGN_LG2	(PAGE_SHIFT + THREAD_ORDER)
 #endif
 #define MIN_KERNEL_ALIGN	(_AC(1, UL) << MIN_KERNEL_ALIGN_LG2)
 
arch/x86/include/asm/percpu.h +10 -0
index 02ecb30..103f1dd 100644
@@ -42,6 +42,7 @@
 
 #else /* ...!ASSEMBLY */
 
+#include <linux/kernel.h>
 #include <linux/stringify.h>
 
 #ifdef CONFIG_SMP
@@ -155,6 +156,15 @@ do {							\
 /* We can use this directly for local CPU (faster). */
 DECLARE_PER_CPU(unsigned long, this_cpu_off);
 
+#ifdef CONFIG_NEED_MULTIPLE_NODES
+void *pcpu_lpage_remapped(void *kaddr);
+#else
+static inline void *pcpu_lpage_remapped(void *kaddr)
+{
+	return NULL;
+}
+#endif
+
 #endif /* !__ASSEMBLY__ */
 
 #ifdef CONFIG_SMP
arch/x86/kernel/cpu/amd.c +3 -1
index e5b27d8..28e5f59 100644
@@ -258,13 +258,15 @@ static void __cpuinit amd_detect_cmp(struct cpuinfo_x86 *c)
 {
 #ifdef CONFIG_X86_HT
 	unsigned bits;
+	int cpu = smp_processor_id();
 
 	bits = c->x86_coreid_bits;
-
 	/* Low order bits define the core id (index of core in socket) */
 	c->cpu_core_id = c->initial_apicid & ((1 << bits)-1);
 	/* Convert the initial APIC ID into the socket ID */
 	c->phys_proc_id = c->initial_apicid >> bits;
+	/* use socket ID also for last level cache */
+	per_cpu(cpu_llc_id, cpu) = c->phys_proc_id;
 #endif
 }
 
arch/x86/kernel/cpu/mcheck/mce.c +2 -2
index 284d1de..af425b8 100644
@@ -1117,7 +1117,7 @@ static void mcheck_timer(unsigned long data)
 		*n = min(*n*2, (int)round_jiffies_relative(check_interval*HZ));
 
 	t->expires = jiffies + *n;
-	add_timer(t);
+	add_timer_on(t, smp_processor_id());
 }
 
 static void mce_do_trigger(struct work_struct *work)
@@ -1321,7 +1321,7 @@ static void mce_init_timer(void)
 		return;
 	setup_timer(t, mcheck_timer, smp_processor_id());
 	t->expires = round_jiffies(jiffies + *n);
-	add_timer(t);
+	add_timer_on(t, smp_processor_id());
 }
 
 /*
arch/x86/kernel/dumpstack.c +1 -0
index 95ea5fa..c840571 100644
@@ -22,6 +22,7 @@
 #include "dumpstack.h"
 
 int panic_on_unrecovered_nmi;
+int panic_on_io_nmi;
 unsigned int code_bytes = 64;
 int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE;
 static int die_counter;
arch/x86/kernel/setup.c +16 -0
index be5ae80..de2cab1 100644
@@ -289,6 +289,20 @@ void * __init extend_brk(size_t size, size_t align)
 	return ret;
 }
 
+#ifdef CONFIG_X86_64
+static void __init init_gbpages(void)
+{
+	if (direct_gbpages && cpu_has_gbpages)
+		printk(KERN_INFO "Using GB pages for direct mapping\n");
+	else
+		direct_gbpages = 0;
+}
+#else
+static inline void init_gbpages(void)
+{
+}
+#endif
+
 static void __init reserve_brk(void)
 {
 	if (_brk_end > _brk_start)
@@ -871,6 +885,8 @@ void __init setup_arch(char **cmdline_p)
 
 	reserve_brk();
 
+	init_gbpages();
+
 	/* max_pfn_mapped is updated here */
 	max_low_pfn_mapped = init_memory_mapping(0, max_low_pfn<<PAGE_SHIFT);
 	max_pfn_mapped = max_low_pfn_mapped;
arch/x86/kernel/setup_percpu.c +162 -57

322 lines not on this page · raw message

arch/x86/kernel/tlb_uv.c +2 -7
index 124d40c..8ccabb8 100644
@@ -711,7 +711,6 @@ uv_activation_descriptor_init(int node, int pnode)
 	unsigned long pa;
 	unsigned long m;
 	unsigned long n;
-	unsigned long mmr_image;
 	struct bau_desc *adp;
 	struct bau_desc *ad2;
 
@@ -727,12 +726,8 @@ uv_activation_descriptor_init(int node, int pnode)
 	n = pa >> uv_nshift;
 	m = pa & uv_mmask;
 
-	mmr_image = uv_read_global_mmr64(pnode, UVH_LB_BAU_SB_DESCRIPTOR_BASE);
-	if (mmr_image) {
-		uv_write_global_mmr64(pnode, (unsigned long)
-				      UVH_LB_BAU_SB_DESCRIPTOR_BASE,
-				      (n << UV_DESC_BASE_PNODE_SHIFT | m));
-	}
+	uv_write_global_mmr64(pnode, UVH_LB_BAU_SB_DESCRIPTOR_BASE,
+			      (n << UV_DESC_BASE_PNODE_SHIFT | m));
 
 	/*
 	 * initializing all 8 (UV_ITEMS_PER_DESCRIPTOR) descriptors for each
arch/x86/kernel/traps.c +3 -0
index a0f48f5..5204332 100644
@@ -346,6 +346,9 @@ io_check_error(unsigned char reason, struct pt_regs *regs)
 	printk(KERN_EMERG "NMI: IOCK error (debug interrupt?)\n");
 	show_registers(regs);
 
+	if (panic_on_io_nmi)
+		panic("NMI IOCK error: Not continuing");
+
 	/* Re-enable the IOCK line, wait for a few seconds */
 	reason = (reason & 0xf) | 8;
 	outb(reason, 0x61);
arch/x86/lib/delay.c +3 -0
index f456860..ff485d3 100644
@@ -55,8 +55,10 @@ static void delay_tsc(unsigned long loops)
 
 	preempt_disable();
 	cpu = smp_processor_id();
+	rdtsc_barrier();
 	rdtscl(bclock);
 	for (;;) {
+		rdtsc_barrier();
 		rdtscl(now);
 		if ((now - bclock) >= loops)
 			break;
@@ -78,6 +80,7 @@ static void delay_tsc(unsigned long loops)
 		if (unlikely(cpu != smp_processor_id())) {
 			loops -= (now - bclock);
 			cpu = smp_processor_id();
+			rdtsc_barrier();
 			rdtscl(bclock);
 		}
 	}
arch/x86/mm/init.c +0 -17
index f53b57e..47ce9a2 100644
@@ -177,20 +177,6 @@ static int __meminit save_mr(struct map_range *mr, int nr_range,
 	return nr_range;
 }
 
-#ifdef CONFIG_X86_64
-static void __init init_gbpages(void)
-{
-	if (direct_gbpages && cpu_has_gbpages)
-		printk(KERN_INFO "Using GB pages for direct mapping\n");
-	else
-		direct_gbpages = 0;
-}
-#else
-static inline void init_gbpages(void)
-{
-}
-#endif
-
 /*
  * Setup the direct mapping of the physical memory at PAGE_OFFSET.
  * This runs before bootmem is initialized and gets pages directly from
@@ -210,9 +196,6 @@ unsigned long __init_refok init_memory_mapping(unsigned long start,
 
 	printk(KERN_INFO "init_memory_mapping: %016lx-%016lx\n", start, end);
 
-	if (!after_bootmem)
-		init_gbpages();
-
 #if defined(CONFIG_DEBUG_PAGEALLOC) || defined(CONFIG_KMEMCHECK)
 	/*
 	 * For CONFIG_DEBUG_PAGEALLOC, identity mapping will use small pages.
arch/x86/mm/pageattr.c +40 -25

96 lines not on this page · raw message

arch/x86/power/cpu.c +1 -1
index d277ef1..b3d20b9 100644
@@ -244,7 +244,7 @@ static void __restore_processor_state(struct saved_context *ctxt)
 	do_fpu_end();
 	mtrr_ap_init();
 
-#ifdef CONFIG_X86_32
+#ifdef CONFIG_X86_OLD_MCE
 	mcheck_init(&boot_cpu_data);
 #endif
 }
include/linux/kernel.h +1 -0
index fac104e..d6320a3 100644
@@ -303,6 +303,7 @@ extern int oops_in_progress;		/* If set, an oops, panic(), BUG() or die() is in
 extern int panic_timeout;
 extern int panic_on_oops;
 extern int panic_on_unrecovered_nmi;
+extern int panic_on_io_nmi;
 extern const char *print_tainted(void);
 extern void add_taint(unsigned flag);
 extern int test_taint(unsigned flag);
kernel/sysctl.c +8 -0
index 62e4ff9..fba42ed 100644
@@ -744,6 +744,14 @@ static struct ctl_table kern_table[] = {
 		.proc_handler	= &proc_dointvec,
 	},
 	{
+		.ctl_name	= CTL_UNNUMBERED,
+		.procname	= "panic_on_io_nmi",
+		.data		= &panic_on_io_nmi,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
+	},
+	{
 		.ctl_name	= KERN_BOOTLOADER_TYPE,
 		.procname	= "bootloader_type",
 		.data		= &bootloader_type,
mm/percpu.c +14 -10
index c0b2c1a..b70f2ac 100644
@@ -549,14 +549,14 @@ static void pcpu_free_area(struct pcpu_chunk *chunk, int freeme)
  * @chunk: chunk of interest
  * @page_start: page index of the first page to unmap
  * @page_end: page index of the last page to unmap + 1
- * @flush: whether to flush cache and tlb or not
+ * @flush_tlb: whether to flush tlb or not
  *
  * For each cpu, unmap pages [@page_start,@page_end) out of @chunk.
  * If @flush is true, vcache is flushed before unmapping and tlb
  * after.
  */
 static void pcpu_unmap(struct pcpu_chunk *chunk, int page_start, int page_end,
-		       bool flush)
+		       bool flush_tlb)
 {
 	unsigned int last = num_possible_cpus() - 1;
 	unsigned int cpu;
@@ -569,9 +569,8 @@ static void pcpu_unmap(struct pcpu_chunk *chunk, int page_start, int page_end,
 	 * the whole region at once rather than doing it for each cpu.
 	 * This could be an overkill but is more scalable.
 	 */
-	if (flush)
-		flush_cache_vunmap(pcpu_chunk_addr(chunk, 0, page_start),
-				   pcpu_chunk_addr(chunk, last, page_end));
+	flush_cache_vunmap(pcpu_chunk_addr(chunk, 0, page_start),
+			   pcpu_chunk_addr(chunk, last, page_end));
 
 	for_each_possible_cpu(cpu)
 		unmap_kernel_range_noflush(
@@ -579,7 +578,7 @@ static void pcpu_unmap(struct pcpu_chunk *chunk, int page_start, int page_end,
 				(page_end - page_start) << PAGE_SHIFT);
 
 	/* ditto as flush_cache_vunmap() */
-	if (flush)
+	if (flush_tlb)
 		flush_tlb_kernel_range(pcpu_chunk_addr(chunk, 0, page_start),
 				       pcpu_chunk_addr(chunk, last, page_end));
 }
@@ -1234,6 +1233,7 @@ static struct page * __init pcpue_get_page(unsigned int cpu, int pageno)
 ssize_t __init pcpu_embed_first_chunk(size_t static_size, size_t reserved_size,
 				      ssize_t dyn_size, ssize_t unit_size)
 {
+	size_t chunk_size;
 	unsigned int cpu;
 
 	/* determine parameters and allocate */
@@ -1248,11 +1248,15 @@ ssize_t __init pcpu_embed_first_chunk(size_t static_size, size_t reserved_size,
 	} else
 		pcpue_unit_size = max_t(size_t, pcpue_size, PCPU_MIN_UNIT_SIZE);
 
-	pcpue_ptr = __alloc_bootmem_nopanic(
-					num_possible_cpus() * pcpue_unit_size,
-					PAGE_SIZE, __pa(MAX_DMA_ADDRESS));
-	if (!pcpue_ptr)
+	chunk_size = pcpue_unit_size * num_possible_cpus();
+
+	pcpue_ptr = __alloc_bootmem_nopanic(chunk_size, PAGE_SIZE,
+					    __pa(MAX_DMA_ADDRESS));
+	if (!pcpue_ptr) {
+		pr_warning("PERCPU: failed to allocate %zu bytes for "
+			   "embedding\n", chunk_size);
 		return -ENOMEM;
+	}
 
 	/* return the leftover and copy */
 	for_each_possible_cpu(cpu) {