IA64 Kexec/Kdump patch for 2.6.18-rc6
Subsystems (3)
PATCH 1/1 IA64 Kexec/Kdump patch for 2.6.18-rc6
Hi,
Here is a new version of IA64 Kexec/Kdump patch. Update since last patch.
1. Ignore offset in crashkernel=size@offset kernel parameter. kernel
will find crashkernel region according to size at boot time. However crashkernel parameter format is not changed to keep compatibility with other archs
2. send EOI to iosapic
3. Patch from HP to clean interrupt at shutdown time.
4. Enhanced OS_INIT handle patch base on Takao Indoh and comments
from Keith Owens.
arch/ia64/Kconfig +23 -0
@@ -427,6 +427,29 @@ config SGI_SN source "drivers/sn/Kconfig" +config KEXEC + bool "kexec system call (EXPERIMENTAL)" + depends on EXPERIMENTAL + help + kexec is a system call that implements the ability to shutdown your + current kernel, and to start another kernel. It is like a reboot + but it is indepedent of the system firmware. And like a reboot + you can start any kernel with it, not just Linux. + + The name comes from the similiarity to the exec system call. + + It is an ongoing process to be certain the hardware in a machine + is properly shutdown, so do not be surprised if this code does not + initially work for you. It may help to enable device hotplugging + support. As of this writing the exact hardware interface is + strongly in flux, so no good recommendation can be made. + +config CRASH_DUMP + bool "kernel crash dumps (EXPERIMENTAL)" + depends on EXPERIMENTAL && IA64_MCA_RECOVERY + help + Generate crash dump after being started by kexec. + source "drivers/firmware/Kconfig" source "fs/Kconfig.binfmt"
arch/ia64/kernel/crash.c +224 -0
224 lines not on this page · raw message
arch/ia64/kernel/efi.c +62 -3
107 lines not on this page · raw message
arch/ia64/kernel/entry.S +1 -1
@@ -1575,7 +1575,7 @@ sys_call_table: data8 sys_mq_timedreceive // 1265 data8 sys_mq_notify data8 sys_mq_getsetattr - data8 sys_ni_syscall // reserved for kexec_load + data8 sys_kexec_load data8 sys_ni_syscall // reserved for vserver data8 sys_waitid // 1270 data8 sys_add_key
arch/ia64/kernel/iosapic.c +21 -0
@@ -288,6 +288,27 @@ nop (unsigned int irq)/* do nothing... */ } + +#ifdef CONFIG_KEXEC +void +kexec_disable_iosapic (void ) +{ +struct iosapic_intr_info *info; +struct iosapic_rte_info *rte; +u8 vec; +for (info = iosapic_intr_info; info < + iosapic_intr_info +IA64_NUM_VECTORS ; ++info) { +list_for_each_entry (rte, &info->rtes, + rte_list) { +iosapic_write (rte->addr, +IOSAPIC_RTE_LOW (rte->rte_index), +IOSAPIC_MASK ); +iosapic_eoi (rte->addr, vec); + } + } +} +#endif +static void mask_irq (unsigned int irq) {
arch/ia64/kernel/machine_kexec.c +133 -0
133 lines not on this page · raw message
arch/ia64/kernel/Makefile +1 -0
@@ -28,6 +28,7 @@ obj-$(CONFIG_IA64_CYCLONE) += cyclone.o obj-$(CONFIG_CPU_FREQ) += cpufreq/ obj-$(CONFIG_IA64_MCA_RECOVERY) += mca_recovery.o obj-$(CONFIG_KPROBES) += kprobes.o jprobes.o +obj-$(CONFIG_KEXEC) += machine_kexec.o relocate_kernel.o crash.o obj-$(CONFIG_IA64_UNCACHED_ALLOCATOR) += uncached.o obj-$(CONFIG_AUDIT) += audit.o mca_recovery-y += mca_drv.o mca_drv_asm.o
arch/ia64/kernel/relocate_kernel.S +490 -0
490 lines not on this page · raw message
arch/ia64/kernel/setup.c +37 -0
@@ -43,6 +43,8 @@#include <linux/initrd.h> #include <linux/pm.h> #include <linux/cpufreq.h> +#include <linux/kexec.h> +#include <linux/crash_dump.h> #include <asm/ia32.h> #include <asm/machvec.h>
@@ -253,6 +255,42 @@ reserve_memory (void)efi_memmap_init (&rsvd_region[n].start, &rsvd_region[n].end); n++; +#ifdef CONFIG_KEXEC +/* crashkernel=size@offset specifies the size to reserve for a crash +* kernel.(offset is ingored for keep compatibility with other archs) +* By reserving this memory we guarantee that linux +* never set's it up as a DMA target. +* Useful for holding code to do something appropriate +* after a kernel panic. +*/ + { +char *from =strstr (saved_command_line,"crashkernel=" ); +unsigned long base, size; +if (from) { + size =memparse (from +12 , &from); +if (size) { +sort_regions (rsvd_region, n); + base =kdump_find_rsvd_region (size, + rsvd_region, n); +if (base != ~0UL ) { + rsvd_region[n].start + (unsigned long )__va (base); + rsvd_region[n].end + (unsigned long )__va (base + size); + n++; + crashk_res.start = base; + crashk_res.end = base + size -1 ; + } + } + } + efi_memmap_res.start = ia64_boot_param->efi_memmap; + efi_memmap_res.end = efi_memmap_res.start + + ia64_boot_param->efi_memmap_size; + boot_param_res.start =__pa (ia64_boot_param); + boot_param_res.end = boot_param_res.start + +sizeof (*ia64_boot_param); + } +#endif /* end of memory marker */ rsvd_region[n].start = ~0UL ; rsvd_region[n].end = ~0UL ;
@@ -264,6 +302,7 @@ reserve_memory (void)sort_regions (rsvd_region, num_rsvd_regions); } +/** * find_initrd - get initrd parameters from the boot parameter structure *
arch/ia64/kernel/smp.c +55 -1
86 lines not on this page · raw message
include/asm-ia64/kexec.h +53 -0
53 lines not on this page · raw message
include/asm-ia64/meminit.h +2 -1
@@ -15,11 +15,12 @@* - initrd (optional) * - command line string * - kernel code & data +* - crash dumping code reserved region * - Kernel memory map built from EFI memory map * * More could be added if necessary */ -#define IA64_MAX_RSVD_REGIONS 6 +#define IA64_MAX_RSVD_REGIONS 7struct rsvd_region {unsigned long start;/* virtual address of beginning of element */
include/asm-ia64/smp.h +3 -0
@@ -128,6 +128,9 @@ extern void smp_send_reschedule (int cpuextern void lock_ipi_calllock (void );extern void unlock_ipi_calllock (void );extern void identify_siblings (struct cpuinfo_ia64 *); +#ifdef CONFIG_KEXEC +extern void kexec_stop_this_cpu (void *); +#endif #else
include/linux/kexec.h +5 -0
@@ -108,6 +108,10 @@ int kexec_should_crash(struct task_strucextern struct kimage *kexec_image;extern struct kimage *kexec_crash_image; +#ifndef kexec_flush_icache_page +#define kexec_flush_icache_page (page) +#endif +#define KEXEC_ON_CRASH 0x00000001#define KEXEC_ARCH_MASK 0xffff0000
@@ -131,6 +135,7 @@ extern struct resource crashk_res;typedef u32 note_buf_t [MAX_NOTE_BYTES /4 ];extern note_buf_t *crash_notes; +#else /* !CONFIG_KEXEC */ struct pt_regs ;struct task_struct ;
include/linux/sysctl.h +1 -0
@@ -150,6 +150,7 @@ enum KERN_IA64_UNALIGNEDr,/* int: ia64 unaligned userland trap enable */ KERN_COMPAT_LOGs,/* int: print compat layer messages */ KERN_MAX_LOCK_DEPTHt, + KERN_KDUMP_ON_INITu, /* int: ia64 kdump with INIT */ };
kernel/kexec.c +1 -0
@@ -851,6 +851,7 @@ static int kimage_load_crash_segment(strmemset (ptr + uchunk,0 , mchunk - uchunk); } result =copy_from_user (ptr, buf, uchunk); +kexec_flush_icache_page (page);kunmap (page);if (result) { result = (result <0 ) ? result : -EIO ;
Hi,
here is an incremental version of this patch.
It is incremental against the previous version that you posted according to my records.
The entire series, for 2.6.18-rc6 or Tony Luck's ia64 test branch can be found at http://www.vergenet.net/~horms/patches/ia64-kexec/kernel/
Tony, Nan Hai, do you have any thoughts on merging this series into ia64 test? Its really far to big to be managed as a single unit. I would be much more comfortable if
we could track small incremental changes. Even the diff between the current and previous version (as below) is quite large.
Alternatively, if an ia64 kexec tree, branched either of ia64-test, or Linus' tree is better then I am happy to set that up.
Signature · 5 lines
Horms H: http://www.vergenet.net/~horms/ W: http://www.valinux.co.jp/en/ [IA64] Kexec/Kdump patch for 2.6.18-rc6
Quoted 17 lines
From: Zou Nan hai <[email protected]>
Subject: [Patch] IA64 Kexec/Kdump patch for 2.6.18-rc6
Archived-At: http://permalink.gmane.org/gmane.linux.ports.ia64/14988Hi,
Here is a new version of IA64 Kexec/Kdump patch. Update since last patch.1. Ignore offset in crashkernel=size@offset kernel parameter. kernel
will find crashkernel region according to size at boot time. However crashkernel parameter format is not changed to keep compatibility with other archs
2. send EOI to iosapic
3. Patch from HP to clean interrupt at shutdown time.
4. EnhancedOS_INIThandle patch base on Takao Indoh\011and comments
from Keith Owens.
Quoted 1 line
Signed-off-by Zou Nan hai <[email protected]>
This is an incremental version of the above patch
Index: linux-2.6/arch/ia64/Kconfig
=================================--- linux-2.6.orig/arch/ia64/Kconfig 2006-09-20 10:44:38.000000000 +0900
+++ linux-2.6/arch/ia64/Kconfig 2006-09-20 10:45:16.000000000 +0900
@@ -444,7 +446,7 @@
config CRASH_DUMP
bool "kernel crash dumps (EXPERIMENTAL)"
- depends on EXPERIMENTAL
+ depends on EXPERIMENTAL && IA64_MCA_RECOVERY
help
Generate crash dump after being started by kexec.
Index: linux-2.6/arch/ia64/kernel/crash.c
=================================--- linux-2.6.orig/arch/ia64/kernel/crash.c 2006-09-20 10:44:38.000000000 +0900
+++ linux-2.6/arch/ia64/kernel/crash.c 2006-09-20 10:45:16.000000000 +0900
@@ -8,44 +8,39 @@
* Copyright (C) 2005 Intel Corp Zou Nan hai <[email protected]>
*
*/
-#include <linux/init.h>
-#include <linux/types.h>
-#include <linux/kernel.h>
#include <linux/smp.h>
-#include <linux/irq.h>
-#include <linux/pci.h>
-#include <linux/reboot.h>
-#include <linux/kexec.h>
-#include <linux/irq.h>
#include <linux/delay.h>
-#include <linux/elf.h>
+#include <linux/crash_dump.h>
+#include <linux/bootmem.h>
+#include <linux/kexec.h>
#include <linux/elfcore.h>
-#include <linux/device.h>
-#include <asm/uaccess.h>
+#include <linux/sysctl.h>
+#include <linux/init.h>
-size_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
-{
- void *vaddr;
+#include <asm/kdebug.h>
+#include <asm/mca.h>
+#include <asm/uaccess.h>
- if (!csize)
- return 0;
- vaddr = __va(pfn<<PAGE_SHIFT);
- if (userbuf) {
- if (copy_to_user(buf, (vaddr + offset), csize)) {
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
- return csize;
-}
-
-static void device_shootdown(void)
-{
- kdump_disable_iosapic();
-#ifdef CONFIG_IA64_HP_ZX1
- ioc_iova_disable();
-#endif
+int kdump_status[NR_CPUS];
+atomic_t kdump_cpu_freezed;
+int kdump_on_init = 1;
+
+ssize_t
+copy_oldmem_page(unsigned long pfn, char *buf,
+ size_t csize, unsigned long offset, int userbuf)
+{
+ void *vaddr;
+
+ if (!csize)
+ return 0;
+ vaddr = __va(pfn<<PAGE_SHIFT);
+ if (userbuf) {
+ if (copy_to_user(buf, (vaddr + offset), csize)) {
+ return -EFAULT;
+ }
+ } else
+ memcpy(buf, (vaddr + offset), csize);
+ return csize;
}
static inline Elf64_Word
@@ -88,20 +83,33 @@
prstatus->pr_pid = current->pid;
ia64_dump_cpu_regs(dst);
- cfm = dst[43]; - sol = (cfm >> 7) & 0x7f; - sof = cfm & 0x7f; - dst[46] = (unsigned long)ia64_rse_skip_regs((unsigned long *)dst[46],
- sof - sol);
+ cfm = dst[43];
+ sol = (cfm >> 7) & 0x7f;
+ sof = cfm & 0x7f;
+ dst[46] = (unsigned long)ia64_rse_skip_regs((unsigned long *)dst[46],
+ sof - sol);
- buf = (u64 *) per_cpu_ptr(crash_notes, cpu);
+ buf = (u64 *) per_cpu_ptr(crash_notes, cpu);
if (!buf)
return;
buf = append_elf_note(buf, "CORE", NT_PRSTATUS, prstatus,
- sizeof(*prstatus));
+ sizeof(*prstatus));
final_note(buf);
}
+static int
+kdump_wait_cpu_freeze(void)
+{
+ int cpu_num = num_online_cpus() - 1;
+ int timeout = 1000;
+ while(timeout-- > 0) {
+ if (atomic_read(&kdump_cpu_freezed) = cpu_num)
+ return 0;
+ udelay(1000);
+ }
+ return 1;
+}
+
void
machine_crash_shutdown(struct pt_regs *pt)
{
@@ -113,11 +121,104 @@
* In practice this means shooting down the other cpus in
* an SMP system.
*/
- if (in_interrupt())
- ia64_eoi();
- device_shootdown();
+ kexec_disable_iosapic();
#ifdef CONFIG_SMP
kdump_smp_send_stop();
+ if (kdump_wait_cpu_freeze() && kdump_on_init) {
+ //not all cpu response to IPI, send INIT to freeze them
+ kdump_smp_send_init();
+ }
#endif
- udelay(1000000);
}
+
+static void
+machine_kdump_on_init(void)
+{
+ local_irq_disable();
+ kexec_disable_iosapic();
+ machine_kexec(ia64_kimage);
+}
+
+void
+kdump_cpu_freeze(struct unw_frame_info *info, void *arg)
+{
+ local_irq_disable();
+ crash_save_this_cpu();
+ current->thread.ksp = (__u64)info->sw - 16;
+ atomic_inc(&kdump_cpu_freezed);
+ kdump_status[smp_processor_id()] = 1;
+ mb();
+ for (;;)
+ cpu_relax();
+}
+
+static int
+kdump_init_notifier(struct notifier_block *self, unsigned long val, void *data) +{
+ struct ia64_mca_notify_die *nd;
+ struct die_args *args = data;
+
+ if (!kdump_on_init)
+ return NOTIFY_DONE;
+ nd = (struct ia64_mca_notify_die *)args->err;
+ /* Reason code 1 means machine check rendezous*/
+ if (nd->sos->rv_rc = 1)
+ return NOTIFY_DONE;
+
+ switch (val) {
+ case DIE_INIT_MONARCH_ENTER:
+ machine_kdump_on_init();
+ break;
+ case DIE_INIT_SLAVE_ENTER:
+ unw_init_running(kdump_cpu_freeze, NULL);
+ break;
+ }
+ return NOTIFY_DONE;
+}
+
+#ifdef CONFIG_SYSCTL
+static ctl_table kdump_on_init_table[] = {
+ {
+ .ctl_name = KERN_KDUMP_ON_INIT,
+ .procname = "kdump_on_init",
+ .data = &kdump_on_init,
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = &proc_dointvec,
+ },
+ { .ctl_name = 0 }
+};
+
+static ctl_table sys_table[] = {
+ {
+ .ctl_name = CTL_KERN, + .procname = "kernel", + .mode = 0555, + .child = kdump_on_init_table,
+ },
+ { .ctl_name = 0 }
+};
+#endif
+
+static int
+machine_crash_setup(void)
+{
+ char *from = strstr(saved_command_line, "elfcorehdr=");
+ static struct notifier_block kdump_init_notifier_nb = {
+ .notifier_call = kdump_init_notifier,
+ };
+ int ret;
+ if (from)
+ elfcorehdr_addr = memparse(from+11, &from);
+ saved_max_pfn = (unsigned long)-1;
+ if((ret = register_die_notifier(&kdump_init_notifier_nb)) != 0)
+ return ret;
+#ifdef CONFIG_SYSCTL
+ if ((ret = register_sysctl_table(sys_table, 0)) != 0)
+ return ret;
+#endif
+ return 0;
+}
+
+__initcall(machine_crash_setup);
+
Index: linux-2.6/arch/ia64/kernel/efi.c
=================================--- linux-2.6.orig/arch/ia64/kernel/efi.c 2006-09-20 10:44:38.000000000 +0900
+++ linux-2.6/arch/ia64/kernel/efi.c 2006-09-20 10:45:16.000000000 +0900
@@ -1126,9 +1126,55 @@
#ifdef CONFIG_KEXEC
insert_resource(res, &efi_memmap_res);
insert_resource(res, &boot_param_res);- if (crashk_res.end > crashk_res.start)
+ if (crashk_res.end > crashk_res.start)
insert_resource(res, &crashk_res);
#endif
}
}
}
+
+#ifdef CONFIG_KEXEC
+/* find a block of memory aligned to 64M exclude reserved regions
+ rsvd_regions are sorted
+ */
+unsigned long
+kdump_find_rsvd_region (unsigned long size,
+ struct rsvd_region *r, int n)
+{
+ int i; + u64 start, end; + u64 alignment = 1UL << _PAGE_SIZE_64M; + void *efi_map_start, *efi_map_end, *p; + efi_memory_desc_t *md; + u64 efi_desc_size; + + efi_map_start = __va(ia64_boot_param->efi_memmap); + efi_map_end = efi_map_start + ia64_boot_param->efi_memmap_size; + efi_desc_size = ia64_boot_param->efi_memdesc_size;
+
+ for (p = efi_map_start; p < efi_map_end; p += efi_desc_size) {
+ md = p;
+ if (!efi_wb(md))
+ continue;
+ start = ALIGN(md->phys_addr, alignment);
+ end = efi_md_end(md);
+ for (i = 0; i < n; i++) {+ if (__pa(r[i].start) >= start && __pa(r[i].end) < end) {
+ if (__pa(r[i].start) > start + size)
+ return start;
+ start = ALIGN(__pa(r[i].end), alignment);
+ if (i < n-1 && __pa(r[i+1].start) < start + size)
+ continue;
+ else
+ break;
+ }
+ }
+ if (end > start + size)
+ return start;
+ }
+
+ printk(KERN_WARNING "Cannot reserve 0x%lx byte of memory for crashdump\n",
+ size);
+ return ~0UL;
+}
+#endif
Index: linux-2.6/arch/ia64/kernel/iosapic.c
=================================--- linux-2.6.orig/arch/ia64/kernel/iosapic.c 2006-09-20 10:44:38.000000000 +0900
+++ linux-2.6/arch/ia64/kernel/iosapic.c 2006-09-20 10:45:16.000000000 +0900
@@ -288,20 +288,22 @@
/* do nothing... */
}
-#ifdef CONFIG_CRASH_DUMP
+
+#ifdef CONFIG_KEXEC
void
-kdump_disable_iosapic(void)
+kexec_disable_iosapic(void)
{
- u32 low32;
struct iosapic_intr_info *info;
struct iosapic_rte_info *rte;
+ u8 vec;
for (info = iosapic_intr_info; info <
iosapic_intr_info + IA64_NUM_VECTORS; ++info) {
- low32 = info->low32 |= IOSAPIC_MASK;
list_for_each_entry(rte, &info->rtes,
rte_list) {
iosapic_write(rte->addr,
- IOSAPIC_RTE_LOW(rte->rte_index), low32);
+ IOSAPIC_RTE_LOW(rte->rte_index),
+ IOSAPIC_MASK);
+ iosapic_eoi(rte->addr, vec);
}
}
}
Index: linux-2.6/arch/ia64/kernel/machine_kexec.c
=================================--- linux-2.6.orig/arch/ia64/kernel/machine_kexec.c 2006-09-20 10:44:38.000000000 +0900
+++ linux-2.6/arch/ia64/kernel/machine_kexec.c 2006-09-20 10:45:16.000000000 +0900
@@ -10,24 +10,20 @@
* Version 2. See the file COPYING for more details.
*/
-#include <linux/kernel.h>
-#include <linux/config.h>
#include <linux/mm.h>
#include <linux/kexec.h>
-#include <linux/pci.h>
#include <linux/cpu.h>
+#include <linux/irq.h>
#include <asm/mmu_context.h>
#include <asm/setup.h>
-#include <asm/mca.h>
-#include <asm/page.h>
-#include <asm/bitops.h>
-#include <asm/tlbflush.h>
#include <asm/delay.h>
#include <asm/meminit.h>
typedef void (*relocate_new_kernel_t)(unsigned long, unsigned long,
struct ia64_boot_param *, unsigned long);
-static struct kimage *ia64_kimage;
+
+struct kimage *ia64_kimage;
+
struct resource efi_memmap_res = {
.name = "EFI Memory Map",
.start = 0,
@@ -83,28 +79,7 @@
#elif defined(CONFIG_SMP)
smp_call_function(kexec_stop_this_cpu, (void *)ia64_kimage->start, 0, 0);
#endif
-#ifdef CONFIG_PCI
- {
- struct pci_dev *dev = NULL;
- irq_desc_t *idesc;
- /* Disable all PCI devices */
- while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) {
- if (!(dev->is_enabled))
- continue;
- idesc = irq_desc + dev->irq;
- if (!idesc||!idesc->chip)
- continue;
- disable_irq_nosync(dev->irq);
- idesc->chip->end(dev->irq);
- idesc->action = NULL;
- pci_disable_device(dev);
- }
- }
-#endif
-
-#ifdef CONFIG_IA64_HP_ZX1
- ioc_iova_disable();
-#endif
+ kexec_disable_iosapic();
}
/*
@@ -118,14 +93,33 @@
relocate_new_kernel_t rnk;
void *pal_addr = efi_get_pal_addr();
unsigned long code_addr = (unsigned long)page_address(image->control_code_page);
+ unsigned long vector;
+
if (image->type = KEXEC_TYPE_CRASH) {
crash_save_this_cpu();
current->thread.ksp = (__u64)info->sw - 16;
}
/* Interrupts aren't acceptable while we reboot */
- ia64_set_itv(1<<16);
local_irq_disable();
+
+ /* Mask CMC and Performance Monitor interrupts */
+ ia64_setreg(_IA64_REG_CR_PMV, 1 << 16);
+ ia64_setreg(_IA64_REG_CR_CMCV, 1 << 16);
+
+ /* Mask ITV and Local Redirect Registers */
+ ia64_set_itv(1 << 16);
+ ia64_set_lrr0(1 << 16);
+ ia64_set_lrr1(1 << 16);
+
+ /* unmask TPR and clear any pending interrupts */
+ ia64_setreg(_IA64_REG_CR_TPR, 0);
+ ia64_srlz_d();
+ vector = ia64_get_ivr();
+ while (vector != IA64_SPURIOUS_INT_VECTOR) {
+ ia64_eoi();
+ vector = ia64_get_ivr();
+ }
rnk = (relocate_new_kernel_t)&code_addr; (*rnk)(image->head, image->start, ia64_boot_param,
GRANULEROUNDDOWN((unsigned long) pal_addr));
Index: linux-2.6/arch/ia64/kernel/setup.c
=================================--- linux-2.6.orig/arch/ia64/kernel/setup.c 2006-09-20 10:44:38.000000000 +0900
+++ linux-2.6/arch/ia64/kernel/setup.c 2006-09-20 10:45:16.000000000 +0900
@@ -252,27 +252,35 @@
}
#endif
+ efi_memmap_init(&rsvd_region[n].start, &rsvd_region[n].end);
+ n++;
+
#ifdef CONFIG_KEXEC
- /* crashkernel=size@addr specifies the location to reserve for
- * a crash kernel. By reserving this memory we guarantee
- * that linux never set's it up as a DMA target.
+ /* crashkernel=size@offset specifies the size to reserve for a crash
+ * kernel.(offset is ingored for keep compatibility with other archs)
+ * By reserving this memory we guarantee that linux
+ * never set's it up as a DMA target.
* Useful for holding code to do something appropriate
* after a kernel panic.
*/
{
char *from = strstr(saved_command_line, "crashkernel=");
+ unsigned long base, size;
if (from) {
- unsigned long size, base;
size = memparse(from + 12, &from);
- if (*from = '@') {
- base = memparse(from + 1, &from);
- rsvd_region[n].start - (unsigned long)__va(base);
- rsvd_region[n].end - (unsigned long)__va(base + size);
- crashk_res.start = base;
- crashk_res.end = base + size - 1;
- n++;
+ if (size) {
+ sort_regions(rsvd_region, n);
+ base = kdump_find_rsvd_region(size,
+ rsvd_region, n);
+ if (base != ~0UL) {
+ rsvd_region[n].start + (unsigned long)__va(base);
+ rsvd_region[n].end + (unsigned long)__va(base + size);
+ n++;
+ crashk_res.start = base;
+ crashk_res.end = base + size - 1;
+ }
}
}
efi_memmap_res.start = ia64_boot_param->efi_memmap;
@@ -283,10 +291,6 @@
sizeof(*ia64_boot_param);
}
#endif
-
- efi_memmap_init(&rsvd_region[n].start, &rsvd_region[n].end);
- n++;
-
/* end of memory marker */ rsvd_region[n].start = ~0UL;
rsvd_region[n].end = ~0UL;
@@ -298,6 +302,7 @@
sort_regions(rsvd_region, num_rsvd_regions);
}
+
/**
* find_initrd - get initrd parameters from the boot parameter structure *
@@ -518,16 +523,6 @@
if (!nomca)
ia64_mca_init();
-#ifdef CONFIG_CRASH_DUMP
- {
- char *from = strstr(saved_command_line, "elfcorehdr=");
-
- if (from)
- elfcorehdr_addr = memparse(from+11, &from);
- saved_max_pfn = (unsigned long) -1;
- }
-#endif
-
platform_setup(cmdline_p);
paging_init();
}
Index: linux-2.6/arch/ia64/kernel/smp.c
=================================--- linux-2.6.orig/arch/ia64/kernel/smp.c 2006-09-20 10:44:38.000000000 +0900
+++ linux-2.6/arch/ia64/kernel/smp.c 2006-09-20 10:45:16.000000000 +0900
@@ -126,19 +126,6 @@
cpu_halt();
}
-#ifdef CONFIG_CRASH_DUMP
-static void
-kdump_cpu_freeze(struct unw_frame_info *info, void *arg) -{
- local_irq_disable();
- crash_save_this_cpu();
- current->thread.ksp = (__u64)info->sw - 16;
- for (;;)
- ia64_hint(ia64_hint_pause);
-}
-#endif
-
-
void
cpu_die(void)
{
@@ -266,6 +253,19 @@
{
send_IPI_allbutself(IPI_KDUMP_CPU_STOP);
}
+
+void
+kdump_smp_send_init()
+{
+ unsigned int cpu, self_cpu;
+ self_cpu = smp_processor_id();
+ for_each_online_cpu(cpu) {
+ if (cpu != self_cpu) {
+ if(kdump_status[cpu] = 0)
+ platform_send_ipi(cpu, 0, IA64_IPI_DM_INIT, 0);
+ }
+ }
+}
#endif
/*
* Called with preeemption disabled.
Index: linux-2.6/include/asm-ia64/kexec.h
=================================--- linux-2.6.orig/include/asm-ia64/kexec.h 2006-09-20 10:44:38.000000000 +0900
+++ linux-2.6/include/asm-ia64/kexec.h 2006-09-20 10:45:16.000000000 +0900
@@ -25,6 +25,7 @@
flush_icache_range(page_addr, page_addr + PAGE_SIZE); \
} while(0)+extern struct kimage *ia64_kimage;
DECLARE_PER_CPU(u64, ia64_mca_pal_base);
const extern unsigned int relocate_new_kernel_size; volatile extern long kexec_rendez;
@@ -39,7 +40,14 @@
extern struct resource efi_memmap_res;
extern struct resource boot_param_res;
extern void kdump_smp_send_stop(void);
-extern void kdump_disable_iosapic(void);
+extern void kdump_smp_send_init(void);
+extern void kexec_disable_iosapic(void);
extern void crash_save_this_cpu(void);
+struct rsvd_region;
+extern unsigned long kdump_find_rsvd_region(unsigned long size,
+ struct rsvd_region *rsvd_regions, int n);
+extern void kdump_cpu_freeze(struct unw_frame_info *info, void *arg); +extern int kdump_status[];
+extern atomic_t kdump_cpu_freezed;
#endif /* _ASM_IA64_KEXEC_H */
Index: linux-2.6/include/linux/sysctl.h
=================================--- linux-2.6.orig/include/linux/sysctl.h 2006-09-20 10:44:39.000000000 +0900
+++ linux-2.6/include/linux/sysctl.h 2006-09-20 10:45:16.000000000 +0900
@@ -150,6 +150,7 @@
KERN_IA64_UNALIGNEDr, /* int: ia64 unaligned userland trap enable */
KERN_COMPAT_LOGs, /* int: print compat layer messages */
KERN_MAX_LOCK_DEPTHt,
+ KERN_KDUMP_ON_INITu, /* int: ia64 kdump with INIT */
};
Quoted 5 lines
Tony, Nan Hai, do you have any thoughts on merging this series into ia64 test? Its really far to big to be managed as a single unit. I would be much more comfortable if
we could track small incremental changes. Even the diff between the current and previous version (as below) is quite large.
I'm open to suggestions. My current inclination is to abandon the existing kexec/kdump patches that are currently in my test tree[1] and replace them with the latest set of patches. So a re-diff of all that is considered good against 2.6.18 would be ideal for that.
Breaking the patch into some smaller pieces would also be good, but I don't think this needs to be taken to extremes (just define a few basic areas and split the patch ... I don't thing it is worth spending days/weeks to split this into a sequence of 37 individually crafted patches, with the kernel being fully functional at every one of the 36 intermediate points).
-Tony
[1] My test tree is not a guaranteed monotonic increasing sequence of changesets. I've already blown it away once before to clean out all the crufty "Auto-update from upstream" commits that my workflow generates. It's due for another clean-up anyway.
Quoted 4 lines · Zou Nan hai
Hi,
Here is a new version of IA64 Kexec/Kdump patch. Update since last patch.
Quoted 8 lines
1. Ignore offset in crashkernel=size@offset kernel parameter. kernel
will find crashkernel region according to size at boot time. However crashkernel parameter format is not changed to keep compatibility with other archs
2. send EOI to iosapic
3. Patch from HP to clean interrupt at shutdown time.
4. EnhancedOS_INIThandle patch base on Takao Indoh and comments
from Keith Owens.
This patch fails our buncho read_oops_irq test because of this change (as displayed in Horms' incremental patch):
@@ -113,11 +121,104 @@
* In practice this means shooting down the other cpus in
* an SMP system.
*/
- if (in_interrupt())
- ia64_eoi();
- device_shootdown();
+ kexec_disable_iosapic();
#ifdef CONFIG_SMP
Our read_oops_irq test attempts to simulate an oops from an interrupt handler by sending an IPI to a processor and having it generate an oops
from within the handler. With the new patch we see this:
...
<0>Kernel panic - not syncing: Aiee, killing interrupt handler!
Linux version 2.6.18-rc6-15sep (bobm@hpde-erix2) (gcc version 3.3.5
(Debian 1:3.3.5-12)) #9 SMP Wed Sep 20 20:09:10 MDT 2006 Ignoring memory below 128MB
Ignoring memory above 384MB
EFI v1.10 by HP: SALsystab=0x3ee7a000 ACPI 2.0=0x3fe34000 SMBIOS=0x3ee7c000 HCDP=0x3fe32000
booting generic kernel on platform dig
PCDP: v3 at 0x3fe32000
Early serial console at MMIO 0xf4050000 (options '9600n8') SAL 3.1: HP version 1.11
SAL Platform features: None
SAL: AP wakeup using external interrupt vector 0xff
BUG: warning at arch/ia64/kernel/sal.c:251/check_sal_cache_flush()
(and then the console hangs)
Upon reboot, the crash_kexec'd system BUGs and hangs in check_sal_cache_flush because ia64_get_ivr returns IA64_SPURIOUS_INT_VECTOR instead of the expected IA64_TIMER_VECTOR. I believe it does this because the processor still has an in-service flag for the IPI interrupt because the handler dies before doing an ia64_eoi().
The old kdump patch checked in_interrupt(), a software construct that keeps track of interrupt nesting, I think, and executed an ia64_eoi() if nonzero. But that got removed in this new patch, leading to our test failures.
The old code wasn't really correct because it only issued one ia64_eoi(). Because of the capability of nesting interrupts, it should be possible to have either 16 (priority classes?) or 256 - 16 (prioritized interrupt vectors?) levels of nested interrupt at the time of the crash. Which is it? Do we believe this comment in arch/ia64/kernel/irq_ia64.c?
/*
* Always set TPR to limit maximum interrupt nesting depth to
* 16 (without this, it would be ~240, which could easily lead
* to kernel stack overflows).
*/
We might be able to trust the software in_interrupt mechanism and count it down to issue ia64_eoi's, but it seems that it's just as easy on our way down to issue a bunch of ia64_eoi's equal to the maximum possible nesting level. I can't see any indication in the docs that it's bad to do ia64_eoi if an interrupt is not currently in-service.
This has to occur before the pending interrupt clearing loop done in ia64_machine_kexec, because any in_service interrupts could cause this loop to terminate early with the IA64_SPURIOUS_INT_VECTOR also, rendering it ineffective:
/* unmask TPR and clear any pending interrupts */
ia64_setreg(_IA64_REG_CR_TPR, 0);
ia64_srlz_d();
vector = ia64_get_ivr();
while (vector != IA64_SPURIOUS_INT_VECTOR) {
ia64_eoi();
vector = ia64_get_ivr();
}My prosposed fix appears below:
Bob Montgomery
Working at HP
arch/ia64/kernel/machine_kexec.c 2006-09-20 20:36:21.000000000 -0600 renamed from arch/ia64/kernel/machine_kexec.c.orig 2006-09-19 10:17:48.000000000 -0600 +5 -0
@@ -94,6 +94,7 @@ static void ia64_machine_kexec(struct un void *pal_addr = efi_get_pal_addr(); unsigned long code_addr = (unsigned long)page_address(image->control_code_page); unsigned long vector; + int ii; if (image->type = KEXEC_TYPE_CRASH) { crash_save_this_cpu();
@@ -112,6 +113,10 @@ static void ia64_machine_kexec(struct un ia64_set_lrr0(1 << 16); ia64_set_lrr1(1 << 16); + /* terminate possibly nested in-service interrupts */ + for (ii = 0; ii < 16; ii++) + ia64_eoi(); + /* unmask TPR and clear any pending interrupts */ ia64_setreg(_IA64_REG_CR_TPR, 0); ia64_srlz_d();
Quoted 25 lines
-----Original Message-----
From: Luck, Tony
Sent: 2006Äê9ÔÂ21ÈÕ 0:43
To: Horms; Zou, Nanhai
Cc: Linux-IA64; fastboot
Subject: RE: [Patch] IA64 Kexec/Kdump patch for 2.6.18-rc6Quoted 5 lines
Tony, Nan Hai, do you have any thoughts on merging this series into ia64 test? Its really far to big to be managed as a single unit. I would be much more comfortable if
we could track small incremental changes. Even the diff between the current and previous version (as below) is quite large.I'm open to suggestions. My current inclination is to abandon the existing kexec/kdump patches that are currently in my test tree[1] and replace them with the latest set of patches. So a re-diff of all that is considered good against 2.6.18 would be ideal for that.
Breaking the patch into some smaller pieces would also be good, but I don't think this needs to be taken to extremes (just define a few basic areas and split the patch ... I don't thing it is worth spending days/weeks to split this into a sequence of 37 individually crafted patches, with the kernel being fully functional at every one of the 36 intermediate points).
Quoted 6 lines
-Tony
[1] My test tree is not a guaranteed monotonic increasing sequence of changesets. I've already blown it away once before to clean out all the crufty "Auto-update from upstream" commits that my workflow generates. It's due for another clean-up anyway.
I put those in a big patch to help people testing the IA64 kdump.
People can grab a base kernel and the patch to test.
Thanks
Zou Nan hai
Quoted 10 lines · Tony
Quoted 5 lines
Tony, Nan Hai, do you have any thoughts on merging this series into ia64 test? Its really far to big to be managed as a single unit. I would be much more comfortable if
we could track small incremental changes. Even the diff between the current and previous version (as below) is quite large.I'm open to suggestions. My current inclination is to abandon the existing kexec/kdump patches that are currently in my test tree[1] and replace them with the latest set of patches. So a re-diff of all that is considered good against 2.6.18 would be ideal for that.
I believe that the patch that Nanhai posted a few days ago is currently the best candidate we have. That could just go straight in. Alternatively, I can easily make a single patch that moves your test tree to that point without having to revert the kexec patches that you have already merged.
Quoted 6 lines
Breaking the patch into some smaller pieces would also be good, but I don't think this needs to be taken to extremes (just define a few basic areas and split the patch ... I don't thing it is worth spending days/weeks to split this into a sequence of 37 individually crafted patches, with the kernel being fully functional at every one of the 36 intermediate points).
Understood. My main concern is being able to track changes moving forward.
Signature · 3 lines
Horms H: http://www.vergenet.net/~horms/ W: http://www.valinux.co.jp/en/
PATCH 1/1 ia64 Kexec/Kdump patch for 2.6.18
In reply to Zou Nan haiFixes since last patch:
1. Check val in kdump_init_notifier.
2. write EOI to clean in-service flag to make kdump work when crash happen inside irq handler.
3. Make CONFIG_KEXEC and CONFIG_CRASH_DUMP depends on !CONFIG_HP_SIM.
arch/ia64/Kconfig +23 -0
@@ -431,6 +431,29 @@ config SGI_SN source "drivers/sn/Kconfig" +config KEXEC + bool "kexec system call (EXPERIMENTAL)" + depends on EXPERIMENTAL && !IA64_HP_SIM + help + kexec is a system call that implements the ability to shutdown your + current kernel, and to start another kernel. It is like a reboot + but it is indepedent of the system firmware. And like a reboot + you can start any kernel with it, not just Linux. + + The name comes from the similiarity to the exec system call. + + It is an ongoing process to be certain the hardware in a machine + is properly shutdown, so do not be surprised if this code does not + initially work for you. It may help to enable device hotplugging + support. As of this writing the exact hardware interface is + strongly in flux, so no good recommendation can be made. + +config CRASH_DUMP + bool "kernel crash dumps (EXPERIMENTAL)" + depends on EXPERIMENTAL && IA64_MCA_RECOVERY && !IA64_HP_SIM + help + Generate crash dump after being started by kexec. + source "drivers/firmware/Kconfig" source "fs/Kconfig.binfmt"
arch/ia64/kernel/crash.c +227 -0
227 lines not on this page · raw message
arch/ia64/kernel/efi.c +62 -3
107 lines not on this page · raw message
arch/ia64/kernel/entry.S +1 -1
@@ -1575,7 +1575,7 @@ sys_call_table: data8 sys_mq_timedreceive // 1265 data8 sys_mq_notify data8 sys_mq_getsetattr - data8 sys_ni_syscall // reserved for kexec_load + data8 sys_kexec_load data8 sys_ni_syscall // reserved for vserver data8 sys_waitid // 1270 data8 sys_add_key
arch/ia64/kernel/iosapic.c +21 -0
@@ -288,6 +288,27 @@ nop (unsigned int irq)/* do nothing... */ } + +#ifdef CONFIG_KEXEC +void +kexec_disable_iosapic (void ) +{ +struct iosapic_intr_info *info; +struct iosapic_rte_info *rte; +u8 vec; +for (info = iosapic_intr_info; info < + iosapic_intr_info +IA64_NUM_VECTORS ; ++info) { +list_for_each_entry (rte, &info->rtes, + rte_list) { +iosapic_write (rte->addr, +IOSAPIC_RTE_LOW (rte->rte_index), +IOSAPIC_MASK ); +iosapic_eoi (rte->addr, vec); + } + } +} +#endif +static void mask_irq (unsigned int irq) {
arch/ia64/kernel/machine_kexec.c +138 -0
138 lines not on this page · raw message
arch/ia64/kernel/Makefile +1 -0
@@ -28,6 +28,7 @@ obj-$(CONFIG_IA64_CYCLONE) += cyclone.o obj-$(CONFIG_CPU_FREQ) += cpufreq/ obj-$(CONFIG_IA64_MCA_RECOVERY) += mca_recovery.o obj-$(CONFIG_KPROBES) += kprobes.o jprobes.o +obj-$(CONFIG_KEXEC) += machine_kexec.o relocate_kernel.o crash.o obj-$(CONFIG_IA64_UNCACHED_ALLOCATOR) += uncached.o obj-$(CONFIG_AUDIT) += audit.o mca_recovery-y += mca_drv.o mca_drv_asm.o
arch/ia64/kernel/relocate_kernel.S +490 -0
490 lines not on this page · raw message
arch/ia64/kernel/setup.c +36 -0
@@ -43,6 +43,8 @@#include <linux/initrd.h> #include <linux/pm.h> #include <linux/cpufreq.h> +#include <linux/kexec.h> +#include <linux/crash_dump.h> #include <asm/ia32.h> #include <asm/machvec.h>
@@ -253,6 +255,41 @@ reserve_memory (void)efi_memmap_init (&rsvd_region[n].start, &rsvd_region[n].end); n++; +#ifdef CONFIG_KEXEC +/* crashkernel=size@offset specifies the size to reserve for a crash +* kernel.(offset is ingored for keep compatibility with other archs) +* By reserving this memory we guarantee that linux never set's it +* up as a DMA target.Useful for holding code to do something +* appropriate after a kernel panic. +*/ + { +char *from =strstr (saved_command_line,"crashkernel=" ); +unsigned long base, size; +if (from) { + size =memparse (from +12 , &from); +if (size) { +sort_regions (rsvd_region, n); + base =kdump_find_rsvd_region (size, + rsvd_region, n); +if (base != ~0UL ) { + rsvd_region[n].start + (unsigned long )__va (base); + rsvd_region[n].end + (unsigned long )__va (base + size); + n++; + crashk_res.start = base; + crashk_res.end = base + size -1 ; + } + } + } + efi_memmap_res.start = ia64_boot_param->efi_memmap; + efi_memmap_res.end = efi_memmap_res.start + + ia64_boot_param->efi_memmap_size; + boot_param_res.start =__pa (ia64_boot_param); + boot_param_res.end = boot_param_res.start + +sizeof (*ia64_boot_param); + } +#endif /* end of memory marker */ rsvd_region[n].start = ~0UL ; rsvd_region[n].end = ~0UL ;
@@ -264,6 +301,7 @@ reserve_memory (void)sort_regions (rsvd_region, num_rsvd_regions); } +/** * find_initrd - get initrd parameters from the boot parameter structure *
arch/ia64/kernel/smp.c +55 -1
86 lines not on this page · raw message
include/asm-ia64/kexec.h +53 -0
53 lines not on this page · raw message
include/asm-ia64/meminit.h +2 -1
@@ -15,11 +15,12 @@* - initrd (optional) * - command line string * - kernel code & data +* - crash dumping code reserved region * - Kernel memory map built from EFI memory map * * More could be added if necessary */ -#define IA64_MAX_RSVD_REGIONS 6 +#define IA64_MAX_RSVD_REGIONS 7struct rsvd_region {unsigned long start;/* virtual address of beginning of element */
include/asm-ia64/smp.h +3 -0
@@ -128,6 +128,9 @@ extern void smp_send_reschedule (int cpuextern void lock_ipi_calllock (void );extern void unlock_ipi_calllock (void );extern void identify_siblings (struct cpuinfo_ia64 *); +#ifdef CONFIG_KEXEC +extern void kexec_stop_this_cpu (void *); +#endif #else
include/linux/kexec.h +5 -0
@@ -108,6 +108,10 @@ int kexec_should_crash(struct task_strucextern struct kimage *kexec_image;extern struct kimage *kexec_crash_image; +#ifndef kexec_flush_icache_page +#define kexec_flush_icache_page (page) +#endif +#define KEXEC_ON_CRASH 0x00000001#define KEXEC_ARCH_MASK 0xffff0000
@@ -131,6 +135,7 @@ extern struct resource crashk_res;typedef u32 note_buf_t [MAX_NOTE_BYTES /4 ];extern note_buf_t *crash_notes; +#else /* !CONFIG_KEXEC */ struct pt_regs ;struct task_struct ;
include/linux/sysctl.h +1 -0
@@ -150,6 +150,7 @@ enum KERN_IA64_UNALIGNEDr,/* int: ia64 unaligned userland trap enable */ KERN_COMPAT_LOGs,/* int: print compat layer messages */ KERN_MAX_LOCK_DEPTHt, + KERN_KDUMP_ON_INITu, /* int: ia64 kdump with INIT */ };
kernel/kexec.c +1 -0
@@ -851,6 +851,7 @@ static int kimage_load_crash_segment(strmemset (ptr + uchunk,0 , mchunk - uchunk); } result =copy_from_user (ptr, buf, uchunk); +kexec_flush_icache_page (page);kunmap (page);if (result) { result = (result <0 ) ? result : -EIO ;