Re: [RFC][PATCH 1/3] Embed vmcoreinfo into kernel file

["Ken'ichi Ohmichi" <oomichi@mxs.nes.nec.co.jp>]

Hi Vivek,

2007/08/17 17:58:03 +0530, Vivek Goyal <vgoyal@in.ibm.com> wrote:
>On Fri, Aug 17, 2007 at 03:43:31PM +0900, Ken'ichi Ohmichi wrote:
>[..]
>> >- There is another important field which I would like to see in vmcoreinfo
>> >  and that is time of crash (lets say CRASH_TIME). This will indicate the
>> >  timestamp when did system actually crash. One can read the time in
>> >  crash_kexec(), fill in the field and then save vmcore info note.
>> >
>> >  For this, either you need to scan the vmcoreinfo note again and fill in
>> >  the time stamp. Or you need to do vmcoreinfo note saving after crash
>> >  instead of boot time.
>> 
>> Is it necessary of the field for timestamp ?
>> The crash utility can display the time of crash already like the following.
>> 
>> # crash vmlinux vmcore
>> [snip]
>>       KERNEL: vmlinux
>>     DUMPFILE: vmcore
>>         CPUS: 2
>>         DATE: Tue Jul 10 20:41:50 2007  <- Here
>>       UPTIME: 00:04:58
>> LOAD AVERAGE: 0.12, 0.22, 0.11
>>        TASKS: 88
>>     NODENAME: peak
>>      RELEASE: 2.6.16.46-0.12-default
>>      VERSION: #1 SMP Thu May 17 14:00:09 UTC 2007
>>      MACHINE: ia64  (1000 Mhz)
>>       MEMORY: 5.5 GB
>>        PANIC: "SysRq : Trigger a crashdump"
>>          PID: 3455
>>      COMMAND: "bash"
>>         TASK: e0000040f2f40000  [THREAD_INFO: e0000040f2f410d0]
>>          CPU: 0
>>        STATE: TASK_RUNNING (SYSRQ)
>> 
>> 
>
>But this requires a debug compiled vmlinux to be available on the system
>and that's a rarity. For example, we are exporting UTSNAME also which 
>can be retrieved by opening vmcore with crash.
>
>By exporting this, now one can write a small utility in second kernel,
>which can just look at the /proc/vmcore in second kernel, determine that
>machine had crashed and send a notification over network to other machine
>with crash time stamp.

OK, the attached patch includes the field CRASHTIME.


>> +#define SYMBOL(name) \
>> +	vmcoreinfo_append_str("SYMBOL(%s)=%lx\n", #name, (unsigned long)&name)
>> +#define SIZE(name) \
>> +	vmcoreinfo_append_str("SIZE(%s)=%d\n", #name, sizeof(struct name))
>> +#define OFFSET(name, field) \
>> +	vmcoreinfo_append_str("OFFSET(%s.%s)=%d\n", #name, #field, &(((struct name *)0)->field))
>> +#define LENGTH(name, value) \
>> +	vmcoreinfo_append_str("LENGTH(%s)=%d\n", #name, value)
>> +#define CONFIG(name) \
>> +	vmcoreinfo_append_str("CONFIG_%s=%c\n", #name, CONFIG_##name)
>> +
>
>Have you tested CONFIG(). For me it did not work? Are you getting output
>like CONFIG_KEXEC=y with above macro?

You're right, I fixed it.

And Daisuke Nishimura updated the patch for translating the virtual
address of vmcoreinfo_note into its physical address on ia64.


Thanks
Ken'ichi Ohmichi

---
Signed-off-by: Dan Aloni <da-x@monatomic.org>
Signed-off-by: Ken'ichi Ohmichi <oomichi@mxs.nes.nec.co.jp>
Signed-off-by: Bernhard Walle <bwalle@suse.de>
Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>

---
diff -rpuN a/arch/i386/kernel/machine_kexec.c b/arch/i386/kernel/machine_kexec.c
--- a/arch/i386/kernel/machine_kexec.c	2007-08-23 00:03:30.000000000 +0900
+++ b/arch/i386/kernel/machine_kexec.c	2007-08-23 00:47:38.000000000 +0900
@@ -10,6 +10,7 @@
 #include <linux/kexec.h>
 #include <linux/delay.h>
 #include <linux/init.h>
+#include <linux/numa.h>
 #include <asm/pgtable.h>
 #include <asm/pgalloc.h>
 #include <asm/tlbflush.h>
@@ -169,3 +170,15 @@ static int __init parse_crashkernel(char
 	return 0;
 }
 early_param("crashkernel", parse_crashkernel);
+
+void arch_crash_save_vmcoreinfo(void)
+{
+#ifdef CONFIG_ARCH_DISCONTIGMEM_ENABLE
+	SYMBOL(node_data);
+	LENGTH(node_data, MAX_NUMNODES);
+#endif
+#ifdef CONFIG_X86_PAE
+	CONFIG(X86_PAE);
+#endif
+}
+
diff -rpuN a/arch/ia64/kernel/machine_kexec.c b/arch/ia64/kernel/machine_kexec.c
--- a/arch/ia64/kernel/machine_kexec.c	2007-08-23 00:03:30.000000000 +0900
+++ b/arch/ia64/kernel/machine_kexec.c	2007-08-23 00:47:38.000000000 +0900
@@ -15,6 +15,8 @@
 #include <linux/cpu.h>
 #include <linux/irq.h>
 #include <linux/efi.h>
+#include <linux/numa.h>
+#include <linux/mmzone.h>
 #include <asm/mmu_context.h>
 #include <asm/setup.h>
 #include <asm/delay.h>
@@ -125,3 +127,31 @@ void machine_kexec(struct kimage *image)
 	unw_init_running(ia64_machine_kexec, image);
 	for(;;);
 }
+
+void arch_crash_save_vmcoreinfo(void)
+{
+#ifdef CONFIG_ARCH_DISCONTIGMEM_ENABLE
+	SYMBOL(pgdat_list);
+	LENGTH(pgdat_list, MAX_NUMNODES);
+
+	SYMBOL(node_memblk);
+	LENGTH(node_memblk, NR_NODE_MEMBLKS);
+	SIZE(node_memblk_s);
+	OFFSET(node_memblk_s, start_paddr);
+	OFFSET(node_memblk_s, size);
+#endif
+#ifdef CONFIG_PGTABLE_3
+	CONFIG(PGTABLE_3);
+#elif  CONFIG_PGTABLE_4
+	CONFIG(PGTABLE_4);
+#endif
+}
+
+unsigned long paddr_vmcoreinfo_note(void)
+{
+	unsigned long vaddr, paddr;
+	vaddr = (unsigned long)(char *)&vmcoreinfo_note;
+	asm volatile ("tpa %0 = %1" : "=r"(paddr) : "r"(vaddr) : "memory");
+	return paddr;
+}
+
diff -rpuN a/arch/ia64/mm/discontig.c b/arch/ia64/mm/discontig.c
--- a/arch/ia64/mm/discontig.c	2007-08-23 00:03:30.000000000 +0900
+++ b/arch/ia64/mm/discontig.c	2007-08-23 00:47:38.000000000 +0900
@@ -47,7 +47,7 @@ struct early_node_data {
 static struct early_node_data mem_data[MAX_NUMNODES] __initdata;
 static nodemask_t memory_less_mask __initdata;
 
-static pg_data_t *pgdat_list[MAX_NUMNODES];
+pg_data_t *pgdat_list[MAX_NUMNODES];
 
 /*
  * To prevent cache aliasing effects, align per-node structures so that they
diff -rpuN a/arch/x86_64/kernel/machine_kexec.c b/arch/x86_64/kernel/machine_kexec.c
--- a/arch/x86_64/kernel/machine_kexec.c	2007-08-23 00:03:30.000000000 +0900
+++ b/arch/x86_64/kernel/machine_kexec.c	2007-08-23 00:47:38.000000000 +0900
@@ -10,6 +10,7 @@
 #include <linux/kexec.h>
 #include <linux/string.h>
 #include <linux/reboot.h>
+#include <linux/numa.h>
 #include <asm/pgtable.h>
 #include <asm/tlbflush.h>
 #include <asm/mmu_context.h>
@@ -257,3 +258,11 @@ static int __init setup_crashkernel(char
 }
 early_param("crashkernel", setup_crashkernel);
 
+void arch_crash_save_vmcoreinfo(void)
+{
+#ifdef CONFIG_ARCH_DISCONTIGMEM_ENABLE
+	SYMBOL(node_data);
+	LENGTH(node_data, MAX_NUMNODES);
+#endif
+}
+
diff -rpuN a/include/asm-ia64/numa.h b/include/asm-ia64/numa.h
--- a/include/asm-ia64/numa.h	2007-08-23 00:03:33.000000000 +0900
+++ b/include/asm-ia64/numa.h	2007-08-23 00:47:38.000000000 +0900
@@ -24,6 +24,7 @@
 
 extern u16 cpu_to_node_map[NR_CPUS] __cacheline_aligned;
 extern cpumask_t node_to_cpu_mask[MAX_NUMNODES] __cacheline_aligned;
+extern pg_data_t *pgdat_list[MAX_NUMNODES];
 
 /* Stuff below this line could be architecture independent */
 
diff -rpuN a/include/linux/kexec.h b/include/linux/kexec.h
--- a/include/linux/kexec.h	2007-08-23 00:03:33.000000000 +0900
+++ b/include/linux/kexec.h	2007-08-23 00:47:38.000000000 +0900
@@ -121,6 +121,23 @@ extern struct page *kimage_alloc_control
 extern void crash_kexec(struct pt_regs *);
 int kexec_should_crash(struct task_struct *);
 void crash_save_cpu(struct pt_regs *regs, int cpu);
+void crash_save_vmcoreinfo(void);
+void arch_crash_save_vmcoreinfo(void);
+void vmcoreinfo_append_str(const char *fmt, ...);
+unsigned long paddr_vmcoreinfo_note(void);
+
+#define SYMBOL(name) \
+	vmcoreinfo_append_str("SYMBOL(%s)=%lx\n", #name, (unsigned long)&name)
+#define SIZE(name) \
+	vmcoreinfo_append_str("SIZE(%s)=%d\n", #name, sizeof(struct name))
+#define OFFSET(name, field) \
+	vmcoreinfo_append_str("OFFSET(%s.%s)=%d\n", #name, #field, \
+			      &(((struct name *)0)->field))
+#define LENGTH(name, value) \
+	vmcoreinfo_append_str("LENGTH(%s)=%d\n", #name, value)
+#define CONFIG(name) \
+	vmcoreinfo_append_str("CONFIG_%s=y\n", #name)
+
 extern struct kimage *kexec_image;
 extern struct kimage *kexec_crash_image;
 
@@ -148,11 +165,20 @@ extern struct kimage *kexec_crash_image;
 
 #define KEXEC_FLAGS    (KEXEC_ON_CRASH)  /* List of defined/legal kexec flags */
 
+#define VMCOREINFO_BYTES           (4096)
+#define VMCOREINFO_NOTE_NAME       "VMCOREINFO"
+#define VMCOREINFO_NOTE_NAME_BYTES ALIGN(sizeof(VMCOREINFO_NOTE_NAME), 4)
+#define VMCOREINFO_NOTE_SIZE       (KEXEC_NOTE_HEAD_BYTES*2 + VMCOREINFO_BYTES \
+				    + VMCOREINFO_NOTE_NAME_BYTES)
+
 /* Location of a reserved region to hold the crash kernel.
  */
 extern struct resource crashk_res;
 typedef u32 note_buf_t[KEXEC_NOTE_BYTES/4];
 extern note_buf_t *crash_notes;
+extern u32 vmcoreinfo_note[VMCOREINFO_NOTE_SIZE/4];
+extern unsigned int vmcoreinfo_size;
+extern unsigned int vmcoreinfo_max_size;
 
 
 #else /* !CONFIG_KEXEC */
diff -rpuN a/kernel/kexec.c b/kernel/kexec.c
--- a/kernel/kexec.c	2007-08-23 00:03:30.000000000 +0900
+++ b/kernel/kexec.c	2007-08-23 00:48:40.000000000 +0900
@@ -22,16 +22,26 @@
 #include <linux/hardirq.h>
 #include <linux/elf.h>
 #include <linux/elfcore.h>
+#include <linux/utsrelease.h>
+#include <linux/utsname.h>
+#include <linux/numa.h>
 
 #include <asm/page.h>
 #include <asm/uaccess.h>
 #include <asm/io.h>
 #include <asm/system.h>
 #include <asm/semaphore.h>
+#include <asm/sections.h>
 
 /* Per cpu memory for storing cpu states in case of system crash. */
 note_buf_t* crash_notes;
 
+/* vmcoreinfo stuff */
+unsigned char vmcoreinfo_data[VMCOREINFO_BYTES];
+u32 vmcoreinfo_note[VMCOREINFO_NOTE_SIZE/4];
+unsigned int vmcoreinfo_size = 0;
+unsigned int vmcoreinfo_max_size = sizeof(vmcoreinfo_data);
+
 /* Location of the reserved area for the crash kernel */
 struct resource crashk_res = {
 	.name  = "Crash kernel",
@@ -1061,6 +1071,7 @@ void crash_kexec(struct pt_regs *regs)
 		if (kexec_crash_image) {
 			struct pt_regs fixed_regs;
 			crash_setup_regs(&fixed_regs, regs);
+			crash_save_vmcoreinfo();
 			machine_crash_shutdown(&fixed_regs);
 			machine_kexec(kexec_crash_image);
 		}
@@ -1135,3 +1146,102 @@ static int __init crash_notes_memory_ini
 	return 0;
 }
 module_init(crash_notes_memory_init)
+
+void crash_save_vmcoreinfo(void)
+{
+	u32 *buf;
+
+	if (!vmcoreinfo_size)
+		return;
+
+	vmcoreinfo_append_str("CRASHTIME=%d", xtime.tv_sec);
+
+	buf = (u32 *)vmcoreinfo_note;
+
+	buf = append_elf_note(buf, VMCOREINFO_NOTE_NAME, 0, vmcoreinfo_data,
+			      vmcoreinfo_size);
+
+	final_note(buf);
+}
+
+void vmcoreinfo_append_str(const char *fmt, ...)
+{
+	va_list args;
+	char buf[0x50];
+	int r;
+
+	va_start(args, fmt);
+	r = vsnprintf(buf, sizeof(buf), fmt, args);
+	va_end(args);
+
+	if (r + vmcoreinfo_size > vmcoreinfo_max_size)
+		r = vmcoreinfo_max_size - vmcoreinfo_size;
+
+	memcpy(&vmcoreinfo_data[vmcoreinfo_size], buf, r);
+
+	vmcoreinfo_size += r;
+}
+
+/*
+ * provide an empty default implementation here -- architecture
+ * code may override this
+ */
+void __attribute__ ((weak)) arch_crash_save_vmcoreinfo(void)
+{}
+
+unsigned long __attribute__ ((weak)) paddr_vmcoreinfo_note(void)
+{
+	return __pa((unsigned long)(char *)&vmcoreinfo_note);
+}
+
+static int __init crash_save_vmcoreinfo_init(void)
+{
+	vmcoreinfo_append_str("OSRELEASE=%s\n", UTS_RELEASE);
+	vmcoreinfo_append_str("PAGESIZE=%d\n", PAGE_SIZE);
+
+	SYMBOL(init_uts_ns);
+	SYMBOL(node_online_map);
+	SYMBOL(swapper_pg_dir);
+	SYMBOL(_stext);
+
+#ifndef CONFIG_NEED_MULTIPLE_NODES
+	SYMBOL(mem_map);
+	SYMBOL(contig_page_data);
+#endif
+#ifdef CONFIG_SPARSEMEM
+	SYMBOL(mem_section);
+	LENGTH(mem_section, NR_SECTION_ROOTS);
+	SIZE(mem_section);
+	OFFSET(mem_section, section_mem_map);
+#endif
+	SIZE(page);
+	SIZE(pglist_data);
+	SIZE(zone);
+	SIZE(free_area);
+	SIZE(list_head);
+	OFFSET(page, flags);
+	OFFSET(page, _count);
+	OFFSET(page, mapping);
+	OFFSET(page, lru);
+	OFFSET(pglist_data, node_zones);
+	OFFSET(pglist_data, nr_zones);
+#ifdef CONFIG_FLAT_NODE_MEM_MAP
+	OFFSET(pglist_data, node_mem_map);
+#endif
+	OFFSET(pglist_data, node_start_pfn);
+	OFFSET(pglist_data, node_spanned_pages);
+	OFFSET(pglist_data, node_id);
+	OFFSET(zone, free_area);
+	OFFSET(zone, vm_stat);
+	OFFSET(zone, spanned_pages);
+	OFFSET(free_area, free_list);
+	OFFSET(list_head, next);
+	OFFSET(list_head, prev);
+	LENGTH(zone.free_area, MAX_ORDER);
+
+	arch_crash_save_vmcoreinfo();
+
+	return 0;
+}
+
+module_init(crash_save_vmcoreinfo_init)
diff -rpuN a/kernel/ksysfs.c b/kernel/ksysfs.c
--- a/kernel/ksysfs.c	2007-08-23 00:03:30.000000000 +0900
+++ b/kernel/ksysfs.c	2007-08-23 00:47:53.000000000 +0900
@@ -60,6 +60,15 @@ static ssize_t kexec_crash_loaded_show(s
 	return sprintf(page, "%d\n", !!kexec_crash_image);
 }
 KERNEL_ATTR_RO(kexec_crash_loaded);
+
+static ssize_t vmcoreinfo_show(struct kset *kset, char *page)
+{
+	return sprintf(page, "%lx %x\n",
+		       paddr_vmcoreinfo_note(),
+		       vmcoreinfo_max_size);
+}
+KERNEL_ATTR_RO(vmcoreinfo);
+
 #endif /* CONFIG_KEXEC */
 
 decl_subsys(kernel, NULL, NULL);
@@ -73,6 +82,7 @@ static struct attribute * kernel_attrs[]
 #ifdef CONFIG_KEXEC
 	&kexec_loaded_attr.attr,
 	&kexec_crash_loaded_attr.attr,
+	&vmcoreinfo_attr.attr,
 #endif
 	NULL
 };
_

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec