freebsd-skq/sys/kern/kern_shutdown.c

894 lines
22 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1986, 1988, 1991, 1993
* The Regents of the University of California. All rights reserved.
* (c) UNIX System Laboratories, Inc.
* All or some portions of this file are derived from material licensed
* to the University of California by American Telephone and Telegraph
* Co. or Unix System Laboratories, Inc. and are reproduced herein with
* the permission of UNIX System Laboratories, Inc.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)kern_shutdown.c 8.3 (Berkeley) 1/21/94
*/
2003-06-11 00:56:59 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
Add textdump(4) facility, which provides an alternative form of kernel dump using mechanically generated/extracted debugging output rather than a simple memory dump. Current sources of debugging output are: - DDB output capture buffer, if there is captured output to save - Kernel message buffer - Kernel configuration, if included in kernel - Kernel version string - Panic message Textdumps are stored in swap/dump partitions as with regular dumps, but are laid out as ustar files in order to allow multiple parts to be stored as a stream of sequentially written blocks. Blocks are written out in reverse order, as the size of a textdump isn't known a priori. As with regular dumps, they will be extracted using savecore(8). One new DDB(4) command is added, "textdump", which accepts "set", "unset", and "status" arguments. By default, normal kernel dumps are generated unless "textdump set" is run in order to schedule a textdump. It can be canceled using "textdump unset" to restore generation of a normal kernel dump. Several sysctls exist to configure aspects of textdumps; debug.ddb.textdump.pending can be set to check whether a textdump is pending, or set/unset in order to control whether the next kernel dump will be a textdump from userspace. While textdumps don't have to be generated as a result of a DDB script run automatically as part of a kernel panic, this is a particular useful way to use them, as instead of generating a complete memory dump, a simple transcript of an automated DDB session can be captured using the DDB output capture and textdump facilities. This can be used to generate quite brief kernel bug reports rich in debugging information but not dependent on kernel symbol tables or precisely synchronized source code. Most textdumps I generate are less than 100k including the full message buffer. Using textdumps with an interactive debugging session is also useful, with capture being enabled/disabled in order to record some but not all of the DDB session. MFC after: 3 months
2007-12-26 11:32:33 +00:00
#include "opt_ddb.h"
#include "opt_kdb.h"
#include "opt_panic.h"
#include "opt_sched.h"
#include "opt_watchdog.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/bio.h>
#include <sys/buf.h>
#include <sys/conf.h>
#include <sys/cons.h>
#include <sys/eventhandler.h>
#include <sys/jail.h>
#include <sys/kdb.h>
#include <sys/kernel.h>
#include <sys/kerneldump.h>
#include <sys/kthread.h>
#include <sys/ktr.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#include <sys/priv.h>
#include <sys/proc.h>
#include <sys/reboot.h>
#include <sys/resourcevar.h>
#include <sys/rwlock.h>
#include <sys/sched.h>
#include <sys/smp.h>
#include <sys/sysctl.h>
#include <sys/sysproto.h>
#include <sys/vnode.h>
#include <sys/watchdog.h>
Add textdump(4) facility, which provides an alternative form of kernel dump using mechanically generated/extracted debugging output rather than a simple memory dump. Current sources of debugging output are: - DDB output capture buffer, if there is captured output to save - Kernel message buffer - Kernel configuration, if included in kernel - Kernel version string - Panic message Textdumps are stored in swap/dump partitions as with regular dumps, but are laid out as ustar files in order to allow multiple parts to be stored as a stream of sequentially written blocks. Blocks are written out in reverse order, as the size of a textdump isn't known a priori. As with regular dumps, they will be extracted using savecore(8). One new DDB(4) command is added, "textdump", which accepts "set", "unset", and "status" arguments. By default, normal kernel dumps are generated unless "textdump set" is run in order to schedule a textdump. It can be canceled using "textdump unset" to restore generation of a normal kernel dump. Several sysctls exist to configure aspects of textdumps; debug.ddb.textdump.pending can be set to check whether a textdump is pending, or set/unset in order to control whether the next kernel dump will be a textdump from userspace. While textdumps don't have to be generated as a result of a DDB script run automatically as part of a kernel panic, this is a particular useful way to use them, as instead of generating a complete memory dump, a simple transcript of an automated DDB session can be captured using the DDB output capture and textdump facilities. This can be used to generate quite brief kernel bug reports rich in debugging information but not dependent on kernel symbol tables or precisely synchronized source code. Most textdumps I generate are less than 100k including the full message buffer. Using textdumps with an interactive debugging session is also useful, with capture being enabled/disabled in order to record some but not all of the DDB session. MFC after: 3 months
2007-12-26 11:32:33 +00:00
#include <ddb/ddb.h>
#include <machine/cpu.h>
#include <machine/pcb.h>
#include <machine/smp.h>
#include <security/mac/mac_framework.h>
#include <vm/vm.h>
#include <vm/vm_object.h>
#include <vm/vm_page.h>
#include <vm/vm_pager.h>
#include <vm/swap_pager.h>
#include <sys/signalvar.h>
#ifndef PANIC_REBOOT_WAIT_TIME
#define PANIC_REBOOT_WAIT_TIME 15 /* default to 15 seconds */
#endif
/*
* Note that stdarg.h and the ANSI style va_start macro is used for both
* ANSI and traditional C compilers.
*/
#include <machine/stdarg.h>
#ifdef KDB
#ifdef KDB_UNATTENDED
int debugger_on_panic = 0;
#else
int debugger_on_panic = 1;
#endif
SYSCTL_INT(_debug, OID_AUTO, debugger_on_panic,
2011-12-14 02:31:32 +00:00
CTLFLAG_RW | CTLFLAG_SECURE | CTLFLAG_TUN,
&debugger_on_panic, 0, "Run debugger on kernel panic");
TUNABLE_INT("debug.debugger_on_panic", &debugger_on_panic);
#ifdef KDB_TRACE
static int trace_on_panic = 1;
#else
static int trace_on_panic = 0;
#endif
SYSCTL_INT(_debug, OID_AUTO, trace_on_panic,
2011-12-14 02:31:32 +00:00
CTLFLAG_RW | CTLFLAG_SECURE | CTLFLAG_TUN,
&trace_on_panic, 0, "Print stack trace on kernel panic");
TUNABLE_INT("debug.trace_on_panic", &trace_on_panic);
#endif /* KDB */
static int sync_on_panic = 0;
SYSCTL_INT(_kern, OID_AUTO, sync_on_panic, CTLFLAG_RW | CTLFLAG_TUN,
&sync_on_panic, 0, "Do a sync before rebooting from a panic");
TUNABLE_INT("kern.sync_on_panic", &sync_on_panic);
static SYSCTL_NODE(_kern, OID_AUTO, shutdown, CTLFLAG_RW, 0,
"Shutdown environment");
#ifndef DIAGNOSTIC
static int show_busybufs;
#else
static int show_busybufs = 1;
#endif
SYSCTL_INT(_kern_shutdown, OID_AUTO, show_busybufs, CTLFLAG_RW,
&show_busybufs, 0, "");
/*
* Variable panicstr contains argument to first call to panic; used as flag
* to indicate that the kernel has already called panic.
*/
const char *panicstr;
Here follows the new kernel dumping infrastructure. Caveats: The new savecore program is not complete in the sense that it emulates enough of the old savecores features to do the job, but implements none of the options yet. I would appreciate if a userland hacker could help me out getting savecore to do what we want it to do from a users point of view, compression, email-notification, space reservation etc etc. (send me email if you are interested). Currently, savecore will scan all devices marked as "swap" or "dump" in /etc/fstab _or_ any devices specified on the command-line. All architectures but i386 lack an implementation of dumpsys(), but looking at the i386 version it should be trivial for anybody familiar with the platform(s) to provide this function. Documentation is quite sparse at this time, more to come. Details: ATA and SCSI drivers should work as the dump formatting code has been removed. The IDA, TWE and AAC have not yet been converted. Dumpon now opens the device and uses ioctl(DIOCGKERNELDUMP) to set the device as dumpdev. To implement the "off" argument, /dev/null is used as the device. Savecore will fail if handed any options since they are not (yet) implemented. All devices marked "dump" or "swap" in /etc/fstab will be scanned and dumps found will be saved to diskfiles named from the MD5 hash of the header record. The header record is dumped in readable format in the .info file. The kernel is not saved. Only complete dumps will be saved. All maintainer rights for this code are disclaimed: feel free to improve and extend. Sponsored by: DARPA, NAI Labs
2002-03-31 22:37:00 +00:00
int dumping; /* system is dumping */
int rebooting; /* system is rebooting */
Here follows the new kernel dumping infrastructure. Caveats: The new savecore program is not complete in the sense that it emulates enough of the old savecores features to do the job, but implements none of the options yet. I would appreciate if a userland hacker could help me out getting savecore to do what we want it to do from a users point of view, compression, email-notification, space reservation etc etc. (send me email if you are interested). Currently, savecore will scan all devices marked as "swap" or "dump" in /etc/fstab _or_ any devices specified on the command-line. All architectures but i386 lack an implementation of dumpsys(), but looking at the i386 version it should be trivial for anybody familiar with the platform(s) to provide this function. Documentation is quite sparse at this time, more to come. Details: ATA and SCSI drivers should work as the dump formatting code has been removed. The IDA, TWE and AAC have not yet been converted. Dumpon now opens the device and uses ioctl(DIOCGKERNELDUMP) to set the device as dumpdev. To implement the "off" argument, /dev/null is used as the device. Savecore will fail if handed any options since they are not (yet) implemented. All devices marked "dump" or "swap" in /etc/fstab will be scanned and dumps found will be saved to diskfiles named from the MD5 hash of the header record. The header record is dumped in readable format in the .info file. The kernel is not saved. Only complete dumps will be saved. All maintainer rights for this code are disclaimed: feel free to improve and extend. Sponsored by: DARPA, NAI Labs
2002-03-31 22:37:00 +00:00
static struct dumperinfo dumper; /* our selected dumper */
/* Context information for dump-debuggers. */
static struct pcb dumppcb; /* Registers. */
lwpid_t dumptid; /* Thread ID. */
static void poweroff_wait(void *, int);
static void shutdown_halt(void *junk, int howto);
static void shutdown_panic(void *junk, int howto);
static void shutdown_reset(void *junk, int howto);
static void vpanic(const char *fmt, va_list ap) __dead2;
/* register various local shutdown events */
2003-02-14 12:44:48 +00:00
static void
shutdown_conf(void *unused)
{
2003-02-14 12:44:48 +00:00
EVENTHANDLER_REGISTER(shutdown_final, poweroff_wait, NULL,
SHUTDOWN_PRI_FIRST);
2003-02-14 12:44:48 +00:00
EVENTHANDLER_REGISTER(shutdown_final, shutdown_halt, NULL,
SHUTDOWN_PRI_LAST + 100);
EVENTHANDLER_REGISTER(shutdown_final, shutdown_panic, NULL,
SHUTDOWN_PRI_LAST + 100);
EVENTHANDLER_REGISTER(shutdown_final, shutdown_reset, NULL,
SHUTDOWN_PRI_LAST + 200);
}
1999-07-17 20:47:52 +00:00
SYSINIT(shutdown_conf, SI_SUB_INTRINSIC, SI_ORDER_ANY, shutdown_conf, NULL);
/*
* The system call that results in a reboot.
*/
/* ARGSUSED */
int
sys_reboot(struct thread *td, struct reboot_args *uap)
{
int error;
error = 0;
#ifdef MAC
error = mac_system_check_reboot(td->td_ucred, uap->opt);
#endif
if (error == 0)
error = priv_check(td, PRIV_REBOOT);
if (error == 0) {
mtx_lock(&Giant);
kern_reboot(uap->opt);
mtx_unlock(&Giant);
}
return (error);
}
/*
* Called by events that want to shut down.. e.g <CTL><ALT><DEL> on a PC
*/
static int shutdown_howto = 0;
void
shutdown_nice(int howto)
{
2003-02-14 12:44:48 +00:00
shutdown_howto = howto;
2003-02-14 12:44:48 +00:00
/* Send a signal to init(8) and have it shutdown the world */
if (initproc != NULL) {
PROC_LOCK(initproc);
kern_psignal(initproc, SIGINT);
PROC_UNLOCK(initproc);
} else {
/* No init(8) running, so simply reboot */
kern_reboot(RB_NOSYNC);
}
return;
}
static int waittime = -1;
static void
print_uptime(void)
{
int f;
struct timespec ts;
getnanouptime(&ts);
printf("Uptime: ");
f = 0;
if (ts.tv_sec >= 86400) {
printf("%ldd", (long)ts.tv_sec / 86400);
ts.tv_sec %= 86400;
f = 1;
}
if (f || ts.tv_sec >= 3600) {
printf("%ldh", (long)ts.tv_sec / 3600);
ts.tv_sec %= 3600;
f = 1;
}
if (f || ts.tv_sec >= 60) {
printf("%ldm", (long)ts.tv_sec / 60);
ts.tv_sec %= 60;
f = 1;
}
printf("%lds\n", (long)ts.tv_sec);
}
int
doadump(boolean_t textdump)
{
boolean_t coredump;
2003-02-14 12:44:48 +00:00
if (dumping)
return (EBUSY);
if (dumper.dumper == NULL)
return (ENXIO);
savectx(&dumppcb);
dumptid = curthread->td_tid;
dumping++;
coredump = TRUE;
Add textdump(4) facility, which provides an alternative form of kernel dump using mechanically generated/extracted debugging output rather than a simple memory dump. Current sources of debugging output are: - DDB output capture buffer, if there is captured output to save - Kernel message buffer - Kernel configuration, if included in kernel - Kernel version string - Panic message Textdumps are stored in swap/dump partitions as with regular dumps, but are laid out as ustar files in order to allow multiple parts to be stored as a stream of sequentially written blocks. Blocks are written out in reverse order, as the size of a textdump isn't known a priori. As with regular dumps, they will be extracted using savecore(8). One new DDB(4) command is added, "textdump", which accepts "set", "unset", and "status" arguments. By default, normal kernel dumps are generated unless "textdump set" is run in order to schedule a textdump. It can be canceled using "textdump unset" to restore generation of a normal kernel dump. Several sysctls exist to configure aspects of textdumps; debug.ddb.textdump.pending can be set to check whether a textdump is pending, or set/unset in order to control whether the next kernel dump will be a textdump from userspace. While textdumps don't have to be generated as a result of a DDB script run automatically as part of a kernel panic, this is a particular useful way to use them, as instead of generating a complete memory dump, a simple transcript of an automated DDB session can be captured using the DDB output capture and textdump facilities. This can be used to generate quite brief kernel bug reports rich in debugging information but not dependent on kernel symbol tables or precisely synchronized source code. Most textdumps I generate are less than 100k including the full message buffer. Using textdumps with an interactive debugging session is also useful, with capture being enabled/disabled in order to record some but not all of the DDB session. MFC after: 3 months
2007-12-26 11:32:33 +00:00
#ifdef DDB
if (textdump && textdump_pending) {
coredump = FALSE;
Add textdump(4) facility, which provides an alternative form of kernel dump using mechanically generated/extracted debugging output rather than a simple memory dump. Current sources of debugging output are: - DDB output capture buffer, if there is captured output to save - Kernel message buffer - Kernel configuration, if included in kernel - Kernel version string - Panic message Textdumps are stored in swap/dump partitions as with regular dumps, but are laid out as ustar files in order to allow multiple parts to be stored as a stream of sequentially written blocks. Blocks are written out in reverse order, as the size of a textdump isn't known a priori. As with regular dumps, they will be extracted using savecore(8). One new DDB(4) command is added, "textdump", which accepts "set", "unset", and "status" arguments. By default, normal kernel dumps are generated unless "textdump set" is run in order to schedule a textdump. It can be canceled using "textdump unset" to restore generation of a normal kernel dump. Several sysctls exist to configure aspects of textdumps; debug.ddb.textdump.pending can be set to check whether a textdump is pending, or set/unset in order to control whether the next kernel dump will be a textdump from userspace. While textdumps don't have to be generated as a result of a DDB script run automatically as part of a kernel panic, this is a particular useful way to use them, as instead of generating a complete memory dump, a simple transcript of an automated DDB session can be captured using the DDB output capture and textdump facilities. This can be used to generate quite brief kernel bug reports rich in debugging information but not dependent on kernel symbol tables or precisely synchronized source code. Most textdumps I generate are less than 100k including the full message buffer. Using textdumps with an interactive debugging session is also useful, with capture being enabled/disabled in order to record some but not all of the DDB session. MFC after: 3 months
2007-12-26 11:32:33 +00:00
textdump_dumpsys(&dumper);
}
Add textdump(4) facility, which provides an alternative form of kernel dump using mechanically generated/extracted debugging output rather than a simple memory dump. Current sources of debugging output are: - DDB output capture buffer, if there is captured output to save - Kernel message buffer - Kernel configuration, if included in kernel - Kernel version string - Panic message Textdumps are stored in swap/dump partitions as with regular dumps, but are laid out as ustar files in order to allow multiple parts to be stored as a stream of sequentially written blocks. Blocks are written out in reverse order, as the size of a textdump isn't known a priori. As with regular dumps, they will be extracted using savecore(8). One new DDB(4) command is added, "textdump", which accepts "set", "unset", and "status" arguments. By default, normal kernel dumps are generated unless "textdump set" is run in order to schedule a textdump. It can be canceled using "textdump unset" to restore generation of a normal kernel dump. Several sysctls exist to configure aspects of textdumps; debug.ddb.textdump.pending can be set to check whether a textdump is pending, or set/unset in order to control whether the next kernel dump will be a textdump from userspace. While textdumps don't have to be generated as a result of a DDB script run automatically as part of a kernel panic, this is a particular useful way to use them, as instead of generating a complete memory dump, a simple transcript of an automated DDB session can be captured using the DDB output capture and textdump facilities. This can be used to generate quite brief kernel bug reports rich in debugging information but not dependent on kernel symbol tables or precisely synchronized source code. Most textdumps I generate are less than 100k including the full message buffer. Using textdumps with an interactive debugging session is also useful, with capture being enabled/disabled in order to record some but not all of the DDB session. MFC after: 3 months
2007-12-26 11:32:33 +00:00
#endif
if (coredump)
Add textdump(4) facility, which provides an alternative form of kernel dump using mechanically generated/extracted debugging output rather than a simple memory dump. Current sources of debugging output are: - DDB output capture buffer, if there is captured output to save - Kernel message buffer - Kernel configuration, if included in kernel - Kernel version string - Panic message Textdumps are stored in swap/dump partitions as with regular dumps, but are laid out as ustar files in order to allow multiple parts to be stored as a stream of sequentially written blocks. Blocks are written out in reverse order, as the size of a textdump isn't known a priori. As with regular dumps, they will be extracted using savecore(8). One new DDB(4) command is added, "textdump", which accepts "set", "unset", and "status" arguments. By default, normal kernel dumps are generated unless "textdump set" is run in order to schedule a textdump. It can be canceled using "textdump unset" to restore generation of a normal kernel dump. Several sysctls exist to configure aspects of textdumps; debug.ddb.textdump.pending can be set to check whether a textdump is pending, or set/unset in order to control whether the next kernel dump will be a textdump from userspace. While textdumps don't have to be generated as a result of a DDB script run automatically as part of a kernel panic, this is a particular useful way to use them, as instead of generating a complete memory dump, a simple transcript of an automated DDB session can be captured using the DDB output capture and textdump facilities. This can be used to generate quite brief kernel bug reports rich in debugging information but not dependent on kernel symbol tables or precisely synchronized source code. Most textdumps I generate are less than 100k including the full message buffer. Using textdumps with an interactive debugging session is also useful, with capture being enabled/disabled in order to record some but not all of the DDB session. MFC after: 3 months
2007-12-26 11:32:33 +00:00
dumpsys(&dumper);
dumping--;
return (0);
}
static int
isbufbusy(struct buf *bp)
{
if (((bp->b_flags & (B_INVAL | B_PERSISTENT)) == 0 &&
BUF_ISLOCKED(bp)) ||
((bp->b_flags & (B_DELWRI | B_INVAL)) == B_DELWRI))
return (1);
return (0);
}
/*
* Shutdown the system cleanly to prepare for reboot, halt, or power off.
*/
void
kern_reboot(int howto)
{
static int first_buf_printf = 1;
#if defined(SMP)
/*
* Bind us to CPU 0 so that all shutdown code runs there. Some
* systems don't shutdown properly (i.e., ACPI power off) if we
* run on another processor.
*/
panic: add a switch and infrastructure for stopping other CPUs in SMP case Historical behavior of letting other CPUs merily go on is a default for time being. The new behavior can be switched on via kern.stop_scheduler_on_panic tunable and sysctl. Stopping of the CPUs has (at least) the following benefits: - more of the system state at panic time is preserved intact - threads and interrupts do not interfere with dumping of the system state Only one thread runs uninterrupted after panic if stop_scheduler_on_panic is set. That thread might call code that is also used in normal context and that code might use locks to prevent concurrent execution of certain parts. Those locks might be held by the stopped threads and would never be released. To work around this issue, it was decided that instead of explicit checks for panic context, we would rather put those checks inside the locking primitives. This change has substantial portions written and re-written by attilio and kib at various times. Other changes are heavily based on the ideas and patches submitted by jhb and mdf. bde has provided many insights into the details and history of the current code. The new behavior may cause problems for systems that use a USB keyboard for interfacing with system console. This is because of some unusual locking patterns in the ukbd code which have to be used because on one hand ukbd is below syscons, but on the other hand it has to interface with other usb code that uses regular mutexes/Giant for its concurrency protection. Dumping to USB-connected disks may also be affected. PR: amd64/139614 (at least) In cooperation with: attilio, jhb, kib, mdf Discussed with: arch@, bde Tested by: Eugene Grosbein <eugen@grosbein.net>, gnn, Steven Hartland <killing@multiplay.co.uk>, glebius, Andrew Boyer <aboyer@averesystems.com> (various versions of the patch) MFC after: 3 months (or never)
2011-12-11 21:02:01 +00:00
if (!SCHEDULER_STOPPED()) {
thread_lock(curthread);
sched_bind(curthread, 0);
thread_unlock(curthread);
KASSERT(PCPU_GET(cpuid) == 0, ("boot: not running on cpu 0"));
}
#endif
/* We're in the process of rebooting. */
rebooting = 1;
/* collect extra flags that shutdown_nice might have set */
howto |= shutdown_howto;
/* We are out of the debugger now. */
kdb_active = 0;
/*
* Do any callouts that should be done BEFORE syncing the filesystems.
*/
EVENTHANDLER_INVOKE(shutdown_pre_sync, howto);
/*
* Now sync filesystems
*/
if (!cold && (howto & RB_NOSYNC) == 0 && waittime < 0) {
register struct buf *bp;
int iter, nbusy, pbusy;
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
#ifndef PREEMPTION
int subiter;
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
#endif
waittime = 0;
wdog_kern_pat(WD_LASTVAL);
sys_sync(curthread, NULL);
/*
* With soft updates, some buffers that are
* written will be remarked as dirty until other
* buffers are written.
*/
for (iter = pbusy = 0; iter < 20; iter++) {
nbusy = 0;
for (bp = &buf[nbuf]; --bp >= buf; )
if (isbufbusy(bp))
nbusy++;
if (nbusy == 0) {
if (first_buf_printf)
printf("All buffers synced.");
break;
}
if (first_buf_printf) {
printf("Syncing disks, buffers remaining... ");
first_buf_printf = 0;
}
printf("%d ", nbusy);
if (nbusy < pbusy)
iter = 0;
pbusy = nbusy;
wdog_kern_pat(WD_LASTVAL);
sys_sync(curthread, NULL);
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
#ifdef PREEMPTION
/*
* Drop Giant and spin for a while to allow
* interrupt threads to run.
*/
DROP_GIANT();
DELAY(50000 * iter);
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
PICKUP_GIANT();
#else
/*
* Drop Giant and context switch several times to
* allow interrupt threads to run.
*/
DROP_GIANT();
for (subiter = 0; subiter < 50 * iter; subiter++) {
thread_lock(curthread);
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
mi_switch(SW_VOL, NULL);
thread_unlock(curthread);
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
DELAY(1000);
}
PICKUP_GIANT();
#endif
}
printf("\n");
/*
* Count only busy local buffers to prevent forcing
* a fsck if we're just a client of a wedged NFS server
*/
nbusy = 0;
for (bp = &buf[nbuf]; --bp >= buf; ) {
if (isbufbusy(bp)) {
2004-11-04 07:59:57 +00:00
#if 0
/* XXX: This is bogus. We should probably have a BO_REMOTE flag instead */
if (bp->b_dev == NULL) {
TAILQ_REMOVE(&mountlist,
bp->b_vp->v_mount, mnt_list);
continue;
}
2004-11-04 07:59:57 +00:00
#endif
nbusy++;
if (show_busybufs > 0) {
printf(
"%d: buf:%p, vnode:%p, flags:%0x, blkno:%jd, lblkno:%jd, buflock:",
nbusy, bp, bp->b_vp, bp->b_flags,
(intmax_t)bp->b_blkno,
(intmax_t)bp->b_lblkno);
BUF_LOCKPRINTINFO(bp);
if (show_busybufs > 1)
vn_printf(bp->b_vp,
"vnode content: ");
}
}
}
if (nbusy) {
/*
* Failed to sync all blocks. Indicate this and don't
* unmount filesystems (thus forcing an fsck on reboot).
*/
printf("Giving up on %d buffers\n", nbusy);
DELAY(5000000); /* 5 seconds */
} else {
if (!first_buf_printf)
printf("Final sync complete\n");
/*
* Unmount filesystems
*/
if (panicstr == 0)
vfs_unmountall();
}
swapoff_all();
DELAY(100000); /* wait for console output to finish */
}
print_uptime();
cngrab();
/*
* Ok, now do things that assume all filesystem activity has
* been completed.
*/
EVENTHANDLER_INVOKE(shutdown_post_sync, howto);
if ((howto & (RB_HALT|RB_DUMP)) == RB_DUMP && !cold && !dumping)
doadump(TRUE);
/* Now that we're going to really halt the system... */
EVENTHANDLER_INVOKE(shutdown_final, howto);
for(;;) ; /* safety against shutdown_reset not working */
/* NOTREACHED */
}
/*
* If the shutdown was a clean halt, behave accordingly.
*/
static void
shutdown_halt(void *junk, int howto)
{
2003-02-14 12:44:48 +00:00
if (howto & RB_HALT) {
printf("\n");
printf("The operating system has halted.\n");
printf("Please press any key to reboot.\n\n");
switch (cngetc()) {
case -1: /* No console, just die */
cpu_halt();
/* NOTREACHED */
default:
howto &= ~RB_HALT;
break;
}
}
}
/*
* Check to see if the system paniced, pause and then reboot
* according to the specified delay.
*/
static void
shutdown_panic(void *junk, int howto)
{
int loop;
if (howto & RB_DUMP) {
if (PANIC_REBOOT_WAIT_TIME != 0) {
if (PANIC_REBOOT_WAIT_TIME != -1) {
printf("Automatic reboot in %d seconds - "
"press a key on the console to abort\n",
PANIC_REBOOT_WAIT_TIME);
for (loop = PANIC_REBOOT_WAIT_TIME * 10;
loop > 0; --loop) {
DELAY(1000 * 100); /* 1/10th second */
/* Did user type a key? */
if (cncheckc() != -1)
break;
}
if (!loop)
return;
}
} else { /* zero time specified - reboot NOW */
return;
}
printf("--> Press a key on the console to reboot,\n");
printf("--> or switch off the system now.\n");
cngetc();
}
}
/*
* Everything done, now reset
*/
static void
shutdown_reset(void *junk, int howto)
{
2003-02-14 12:44:48 +00:00
printf("Rebooting...\n");
DELAY(1000000); /* wait 1 sec for printf's to complete and be read */
2009-08-13 17:09:45 +00:00
/*
* Acquiring smp_ipi_mtx here has a double effect:
* - it disables interrupts avoiding CPU0 preemption
* by fast handlers (thus deadlocking against other CPUs)
* - it avoids deadlocks against smp_rendezvous() or, more
* generally, threads busy-waiting, with this spinlock held,
* and waiting for responses by threads on other CPUs
* (ie. smp_tlb_shootdown()).
*
* For the !SMP case it just needs to handle the former problem.
2009-08-13 17:09:45 +00:00
*/
#ifdef SMP
mtx_lock_spin(&smp_ipi_mtx);
#else
spinlock_enter();
#endif
2009-08-13 17:09:45 +00:00
/* cpu_boot(howto); */ /* doesn't do anything at the moment */
cpu_reset();
/* NOTREACHED */ /* assuming reset worked */
}
#if defined(WITNESS) || defined(INVARIANTS)
static int kassert_warn_only = 0;
2012-12-10 23:11:26 +00:00
#ifdef KDB
static int kassert_do_kdb = 0;
#endif
#ifdef KTR
static int kassert_do_ktr = 0;
#endif
static int kassert_do_log = 1;
static int kassert_log_pps_limit = 4;
static int kassert_log_mute_at = 0;
static int kassert_log_panic_at = 0;
static int kassert_warnings = 0;
SYSCTL_NODE(_debug, OID_AUTO, kassert, CTLFLAG_RW, NULL, "kassert options");
SYSCTL_INT(_debug_kassert, OID_AUTO, warn_only, CTLFLAG_RW | CTLFLAG_TUN,
&kassert_warn_only, 0,
"KASSERT triggers a panic (1) or just a warning (0)");
TUNABLE_INT("debug.kassert.warn_only", &kassert_warn_only);
2012-12-10 23:11:26 +00:00
#ifdef KDB
SYSCTL_INT(_debug_kassert, OID_AUTO, do_kdb, CTLFLAG_RW | CTLFLAG_TUN,
&kassert_do_kdb, 0, "KASSERT will enter the debugger");
TUNABLE_INT("debug.kassert.do_kdb", &kassert_do_kdb);
#endif
#ifdef KTR
SYSCTL_UINT(_debug_kassert, OID_AUTO, do_ktr, CTLFLAG_RW | CTLFLAG_TUN,
&kassert_do_ktr, 0,
"KASSERT does a KTR, set this to the KTRMASK you want");
TUNABLE_INT("debug.kassert.do_ktr", &kassert_do_ktr);
#endif
SYSCTL_INT(_debug_kassert, OID_AUTO, do_log, CTLFLAG_RW | CTLFLAG_TUN,
&kassert_do_log, 0, "KASSERT triggers a panic (1) or just a warning (0)");
TUNABLE_INT("debug.kassert.do_log", &kassert_do_log);
SYSCTL_INT(_debug_kassert, OID_AUTO, warnings, CTLFLAG_RW | CTLFLAG_TUN,
&kassert_warnings, 0, "number of KASSERTs that have been triggered");
TUNABLE_INT("debug.kassert.warnings", &kassert_warnings);
SYSCTL_INT(_debug_kassert, OID_AUTO, log_panic_at, CTLFLAG_RW | CTLFLAG_TUN,
&kassert_log_panic_at, 0, "max number of KASSERTS before we will panic");
TUNABLE_INT("debug.kassert.log_panic_at", &kassert_log_panic_at);
SYSCTL_INT(_debug_kassert, OID_AUTO, log_pps_limit, CTLFLAG_RW | CTLFLAG_TUN,
&kassert_log_pps_limit, 0, "limit number of log messages per second");
TUNABLE_INT("debug.kassert.log_pps_limit", &kassert_log_pps_limit);
SYSCTL_INT(_debug_kassert, OID_AUTO, log_mute_at, CTLFLAG_RW | CTLFLAG_TUN,
&kassert_log_mute_at, 0, "max number of KASSERTS to log");
TUNABLE_INT("debug.kassert.log_mute_at", &kassert_log_mute_at);
static int kassert_sysctl_kassert(SYSCTL_HANDLER_ARGS);
SYSCTL_PROC(_debug_kassert, OID_AUTO, kassert,
CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_SECURE, NULL, 0,
kassert_sysctl_kassert, "I", "set to trigger a test kassert");
static int
kassert_sysctl_kassert(SYSCTL_HANDLER_ARGS)
{
int error, i;
error = sysctl_wire_old_buffer(req, sizeof(int));
if (error == 0) {
i = 0;
error = sysctl_handle_int(oidp, &i, 0, req);
}
if (error != 0 || req->newptr == NULL)
return (error);
KASSERT(0, ("kassert_sysctl_kassert triggered kassert %d", i));
return (0);
}
/*
* Called by KASSERT, this decides if we will panic
* or if we will log via printf and/or ktr.
*/
void
kassert_panic(const char *fmt, ...)
{
static char buf[256];
va_list ap;
va_start(ap, fmt);
(void)vsnprintf(buf, sizeof(buf), fmt, ap);
va_end(ap);
/*
* panic if we're not just warning, or if we've exceeded
* kassert_log_panic_at warnings.
*/
if (!kassert_warn_only ||
(kassert_log_panic_at > 0 &&
kassert_warnings >= kassert_log_panic_at)) {
va_start(ap, fmt);
vpanic(fmt, ap);
/* NORETURN */
}
#ifdef KTR
if (kassert_do_ktr)
CTR0(ktr_mask, buf);
#endif /* KTR */
/*
* log if we've not yet met the mute limit.
*/
if (kassert_do_log &&
(kassert_log_mute_at == 0 ||
kassert_warnings < kassert_log_mute_at)) {
static struct timeval lasterr;
static int curerr;
if (ppsratecheck(&lasterr, &curerr, kassert_log_pps_limit)) {
printf("KASSERT failed: %s\n", buf);
kdb_backtrace();
}
}
2012-12-10 23:11:26 +00:00
#ifdef KDB
if (kassert_do_kdb) {
kdb_enter(KDB_WHY_KASSERT, buf);
}
#endif
atomic_add_int(&kassert_warnings, 1);
}
#endif
/*
* Panic is called on unresolvable fatal errors. It prints "panic: mesg",
* and then reboots. If we are called twice, then we avoid trying to sync
* the disks as this often leads to recursive panics.
*/
void
panic(const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
vpanic(fmt, ap);
}
static void
vpanic(const char *fmt, va_list ap)
{
#ifdef SMP
panic: add a switch and infrastructure for stopping other CPUs in SMP case Historical behavior of letting other CPUs merily go on is a default for time being. The new behavior can be switched on via kern.stop_scheduler_on_panic tunable and sysctl. Stopping of the CPUs has (at least) the following benefits: - more of the system state at panic time is preserved intact - threads and interrupts do not interfere with dumping of the system state Only one thread runs uninterrupted after panic if stop_scheduler_on_panic is set. That thread might call code that is also used in normal context and that code might use locks to prevent concurrent execution of certain parts. Those locks might be held by the stopped threads and would never be released. To work around this issue, it was decided that instead of explicit checks for panic context, we would rather put those checks inside the locking primitives. This change has substantial portions written and re-written by attilio and kib at various times. Other changes are heavily based on the ideas and patches submitted by jhb and mdf. bde has provided many insights into the details and history of the current code. The new behavior may cause problems for systems that use a USB keyboard for interfacing with system console. This is because of some unusual locking patterns in the ukbd code which have to be used because on one hand ukbd is below syscons, but on the other hand it has to interface with other usb code that uses regular mutexes/Giant for its concurrency protection. Dumping to USB-connected disks may also be affected. PR: amd64/139614 (at least) In cooperation with: attilio, jhb, kib, mdf Discussed with: arch@, bde Tested by: Eugene Grosbein <eugen@grosbein.net>, gnn, Steven Hartland <killing@multiplay.co.uk>, glebius, Andrew Boyer <aboyer@averesystems.com> (various versions of the patch) MFC after: 3 months (or never)
2011-12-11 21:02:01 +00:00
cpuset_t other_cpus;
#endif
struct thread *td = curthread;
int bootopt, newpanic;
static char buf[256];
spinlock_enter();
panic: add a switch and infrastructure for stopping other CPUs in SMP case Historical behavior of letting other CPUs merily go on is a default for time being. The new behavior can be switched on via kern.stop_scheduler_on_panic tunable and sysctl. Stopping of the CPUs has (at least) the following benefits: - more of the system state at panic time is preserved intact - threads and interrupts do not interfere with dumping of the system state Only one thread runs uninterrupted after panic if stop_scheduler_on_panic is set. That thread might call code that is also used in normal context and that code might use locks to prevent concurrent execution of certain parts. Those locks might be held by the stopped threads and would never be released. To work around this issue, it was decided that instead of explicit checks for panic context, we would rather put those checks inside the locking primitives. This change has substantial portions written and re-written by attilio and kib at various times. Other changes are heavily based on the ideas and patches submitted by jhb and mdf. bde has provided many insights into the details and history of the current code. The new behavior may cause problems for systems that use a USB keyboard for interfacing with system console. This is because of some unusual locking patterns in the ukbd code which have to be used because on one hand ukbd is below syscons, but on the other hand it has to interface with other usb code that uses regular mutexes/Giant for its concurrency protection. Dumping to USB-connected disks may also be affected. PR: amd64/139614 (at least) In cooperation with: attilio, jhb, kib, mdf Discussed with: arch@, bde Tested by: Eugene Grosbein <eugen@grosbein.net>, gnn, Steven Hartland <killing@multiplay.co.uk>, glebius, Andrew Boyer <aboyer@averesystems.com> (various versions of the patch) MFC after: 3 months (or never)
2011-12-11 21:02:01 +00:00
#ifdef SMP
/*
* stop_cpus_hard(other_cpus) should prevent multiple CPUs from
* concurrently entering panic. Only the winner will proceed
* further.
*/
if (panicstr == NULL && !kdb_active) {
other_cpus = all_cpus;
CPU_CLR(PCPU_GET(cpuid), &other_cpus);
stop_cpus_hard(other_cpus);
panic: add a switch and infrastructure for stopping other CPUs in SMP case Historical behavior of letting other CPUs merily go on is a default for time being. The new behavior can be switched on via kern.stop_scheduler_on_panic tunable and sysctl. Stopping of the CPUs has (at least) the following benefits: - more of the system state at panic time is preserved intact - threads and interrupts do not interfere with dumping of the system state Only one thread runs uninterrupted after panic if stop_scheduler_on_panic is set. That thread might call code that is also used in normal context and that code might use locks to prevent concurrent execution of certain parts. Those locks might be held by the stopped threads and would never be released. To work around this issue, it was decided that instead of explicit checks for panic context, we would rather put those checks inside the locking primitives. This change has substantial portions written and re-written by attilio and kib at various times. Other changes are heavily based on the ideas and patches submitted by jhb and mdf. bde has provided many insights into the details and history of the current code. The new behavior may cause problems for systems that use a USB keyboard for interfacing with system console. This is because of some unusual locking patterns in the ukbd code which have to be used because on one hand ukbd is below syscons, but on the other hand it has to interface with other usb code that uses regular mutexes/Giant for its concurrency protection. Dumping to USB-connected disks may also be affected. PR: amd64/139614 (at least) In cooperation with: attilio, jhb, kib, mdf Discussed with: arch@, bde Tested by: Eugene Grosbein <eugen@grosbein.net>, gnn, Steven Hartland <killing@multiplay.co.uk>, glebius, Andrew Boyer <aboyer@averesystems.com> (various versions of the patch) MFC after: 3 months (or never)
2011-12-11 21:02:01 +00:00
}
/*
* We set stop_scheduler here and not in the block above,
* because we want to ensure that if panic has been called and
* stop_scheduler_on_panic is true, then stop_scheduler will
* always be set. Even if panic has been entered from kdb.
*/
td->td_stopsched = 1;
#endif
bootopt = RB_AUTOBOOT;
newpanic = 0;
if (panicstr)
bootopt |= RB_NOSYNC;
else {
bootopt |= RB_DUMP;
panicstr = fmt;
newpanic = 1;
}
if (newpanic) {
(void)vsnprintf(buf, sizeof(buf), fmt, ap);
panicstr = buf;
cngrab();
printf("panic: %s\n", buf);
} else {
printf("panic: ");
vprintf(fmt, ap);
printf("\n");
}
#ifdef SMP
printf("cpuid = %d\n", PCPU_GET(cpuid));
#endif
#ifdef KDB
if (newpanic && trace_on_panic)
kdb_backtrace();
if (debugger_on_panic)
kdb_enter(KDB_WHY_PANIC, "panic");
#endif
/*thread_lock(td); */
td->td_flags |= TDF_INPANIC;
/* thread_unlock(td); */
if (!sync_on_panic)
bootopt |= RB_NOSYNC;
kern_reboot(bootopt);
}
/*
* Support for poweroff delay.
*
* Please note that setting this delay too short might power off your machine
* before the write cache on your hard disk has been flushed, leading to
* soft-updates inconsistencies.
*/
#ifndef POWEROFF_DELAY
# define POWEROFF_DELAY 5000
#endif
static int poweroff_delay = POWEROFF_DELAY;
SYSCTL_INT(_kern_shutdown, OID_AUTO, poweroff_delay, CTLFLAG_RW,
&poweroff_delay, 0, "Delay before poweroff to write disk caches (msec)");
2003-02-14 12:44:48 +00:00
static void
poweroff_wait(void *junk, int howto)
{
2003-02-14 12:44:48 +00:00
if (!(howto & RB_POWEROFF) || poweroff_delay <= 0)
return;
DELAY(poweroff_delay * 1000);
}
/*
* Some system processes (e.g. syncer) need to be stopped at appropriate
* points in their main loops prior to a system shutdown, so that they
* won't interfere with the shutdown process (e.g. by holding a disk buf
* to cause sync to fail). For each of these system processes, register
* shutdown_kproc() as a handler for one of shutdown events.
*/
static int kproc_shutdown_wait = 60;
SYSCTL_INT(_kern_shutdown, OID_AUTO, kproc_shutdown_wait, CTLFLAG_RW,
&kproc_shutdown_wait, 0, "Max wait time (sec) to stop for each process");
void
kproc_shutdown(void *arg, int howto)
{
struct proc *p;
int error;
if (panicstr)
return;
p = (struct proc *)arg;
2004-07-30 01:30:05 +00:00
printf("Waiting (max %d seconds) for system process `%s' to stop...",
kproc_shutdown_wait, p->p_comm);
error = kproc_suspend(p, kproc_shutdown_wait * hz);
if (error == EWOULDBLOCK)
2004-07-30 01:30:05 +00:00
printf("timed out\n");
else
2004-07-30 01:30:05 +00:00
printf("done\n");
}
Here follows the new kernel dumping infrastructure. Caveats: The new savecore program is not complete in the sense that it emulates enough of the old savecores features to do the job, but implements none of the options yet. I would appreciate if a userland hacker could help me out getting savecore to do what we want it to do from a users point of view, compression, email-notification, space reservation etc etc. (send me email if you are interested). Currently, savecore will scan all devices marked as "swap" or "dump" in /etc/fstab _or_ any devices specified on the command-line. All architectures but i386 lack an implementation of dumpsys(), but looking at the i386 version it should be trivial for anybody familiar with the platform(s) to provide this function. Documentation is quite sparse at this time, more to come. Details: ATA and SCSI drivers should work as the dump formatting code has been removed. The IDA, TWE and AAC have not yet been converted. Dumpon now opens the device and uses ioctl(DIOCGKERNELDUMP) to set the device as dumpdev. To implement the "off" argument, /dev/null is used as the device. Savecore will fail if handed any options since they are not (yet) implemented. All devices marked "dump" or "swap" in /etc/fstab will be scanned and dumps found will be saved to diskfiles named from the MD5 hash of the header record. The header record is dumped in readable format in the .info file. The kernel is not saved. Only complete dumps will be saved. All maintainer rights for this code are disclaimed: feel free to improve and extend. Sponsored by: DARPA, NAI Labs
2002-03-31 22:37:00 +00:00
void
kthread_shutdown(void *arg, int howto)
{
struct thread *td;
int error;
if (panicstr)
return;
td = (struct thread *)arg;
printf("Waiting (max %d seconds) for system thread `%s' to stop...",
kproc_shutdown_wait, td->td_name);
error = kthread_suspend(td, kproc_shutdown_wait * hz);
if (error == EWOULDBLOCK)
printf("timed out\n");
else
printf("done\n");
}
static char dumpdevname[sizeof(((struct cdev*)NULL)->si_name)];
SYSCTL_STRING(_kern_shutdown, OID_AUTO, dumpdevname, CTLFLAG_RD,
dumpdevname, 0, "Device for kernel dumps");
Here follows the new kernel dumping infrastructure. Caveats: The new savecore program is not complete in the sense that it emulates enough of the old savecores features to do the job, but implements none of the options yet. I would appreciate if a userland hacker could help me out getting savecore to do what we want it to do from a users point of view, compression, email-notification, space reservation etc etc. (send me email if you are interested). Currently, savecore will scan all devices marked as "swap" or "dump" in /etc/fstab _or_ any devices specified on the command-line. All architectures but i386 lack an implementation of dumpsys(), but looking at the i386 version it should be trivial for anybody familiar with the platform(s) to provide this function. Documentation is quite sparse at this time, more to come. Details: ATA and SCSI drivers should work as the dump formatting code has been removed. The IDA, TWE and AAC have not yet been converted. Dumpon now opens the device and uses ioctl(DIOCGKERNELDUMP) to set the device as dumpdev. To implement the "off" argument, /dev/null is used as the device. Savecore will fail if handed any options since they are not (yet) implemented. All devices marked "dump" or "swap" in /etc/fstab will be scanned and dumps found will be saved to diskfiles named from the MD5 hash of the header record. The header record is dumped in readable format in the .info file. The kernel is not saved. Only complete dumps will be saved. All maintainer rights for this code are disclaimed: feel free to improve and extend. Sponsored by: DARPA, NAI Labs
2002-03-31 22:37:00 +00:00
/* Registration of dumpers */
int
set_dumper(struct dumperinfo *di, const char *devname)
Here follows the new kernel dumping infrastructure. Caveats: The new savecore program is not complete in the sense that it emulates enough of the old savecores features to do the job, but implements none of the options yet. I would appreciate if a userland hacker could help me out getting savecore to do what we want it to do from a users point of view, compression, email-notification, space reservation etc etc. (send me email if you are interested). Currently, savecore will scan all devices marked as "swap" or "dump" in /etc/fstab _or_ any devices specified on the command-line. All architectures but i386 lack an implementation of dumpsys(), but looking at the i386 version it should be trivial for anybody familiar with the platform(s) to provide this function. Documentation is quite sparse at this time, more to come. Details: ATA and SCSI drivers should work as the dump formatting code has been removed. The IDA, TWE and AAC have not yet been converted. Dumpon now opens the device and uses ioctl(DIOCGKERNELDUMP) to set the device as dumpdev. To implement the "off" argument, /dev/null is used as the device. Savecore will fail if handed any options since they are not (yet) implemented. All devices marked "dump" or "swap" in /etc/fstab will be scanned and dumps found will be saved to diskfiles named from the MD5 hash of the header record. The header record is dumped in readable format in the .info file. The kernel is not saved. Only complete dumps will be saved. All maintainer rights for this code are disclaimed: feel free to improve and extend. Sponsored by: DARPA, NAI Labs
2002-03-31 22:37:00 +00:00
{
size_t wantcopy;
2003-02-14 12:44:48 +00:00
Here follows the new kernel dumping infrastructure. Caveats: The new savecore program is not complete in the sense that it emulates enough of the old savecores features to do the job, but implements none of the options yet. I would appreciate if a userland hacker could help me out getting savecore to do what we want it to do from a users point of view, compression, email-notification, space reservation etc etc. (send me email if you are interested). Currently, savecore will scan all devices marked as "swap" or "dump" in /etc/fstab _or_ any devices specified on the command-line. All architectures but i386 lack an implementation of dumpsys(), but looking at the i386 version it should be trivial for anybody familiar with the platform(s) to provide this function. Documentation is quite sparse at this time, more to come. Details: ATA and SCSI drivers should work as the dump formatting code has been removed. The IDA, TWE and AAC have not yet been converted. Dumpon now opens the device and uses ioctl(DIOCGKERNELDUMP) to set the device as dumpdev. To implement the "off" argument, /dev/null is used as the device. Savecore will fail if handed any options since they are not (yet) implemented. All devices marked "dump" or "swap" in /etc/fstab will be scanned and dumps found will be saved to diskfiles named from the MD5 hash of the header record. The header record is dumped in readable format in the .info file. The kernel is not saved. Only complete dumps will be saved. All maintainer rights for this code are disclaimed: feel free to improve and extend. Sponsored by: DARPA, NAI Labs
2002-03-31 22:37:00 +00:00
if (di == NULL) {
bzero(&dumper, sizeof dumper);
dumpdevname[0] = '\0';
Here follows the new kernel dumping infrastructure. Caveats: The new savecore program is not complete in the sense that it emulates enough of the old savecores features to do the job, but implements none of the options yet. I would appreciate if a userland hacker could help me out getting savecore to do what we want it to do from a users point of view, compression, email-notification, space reservation etc etc. (send me email if you are interested). Currently, savecore will scan all devices marked as "swap" or "dump" in /etc/fstab _or_ any devices specified on the command-line. All architectures but i386 lack an implementation of dumpsys(), but looking at the i386 version it should be trivial for anybody familiar with the platform(s) to provide this function. Documentation is quite sparse at this time, more to come. Details: ATA and SCSI drivers should work as the dump formatting code has been removed. The IDA, TWE and AAC have not yet been converted. Dumpon now opens the device and uses ioctl(DIOCGKERNELDUMP) to set the device as dumpdev. To implement the "off" argument, /dev/null is used as the device. Savecore will fail if handed any options since they are not (yet) implemented. All devices marked "dump" or "swap" in /etc/fstab will be scanned and dumps found will be saved to diskfiles named from the MD5 hash of the header record. The header record is dumped in readable format in the .info file. The kernel is not saved. Only complete dumps will be saved. All maintainer rights for this code are disclaimed: feel free to improve and extend. Sponsored by: DARPA, NAI Labs
2002-03-31 22:37:00 +00:00
return (0);
}
if (dumper.dumper != NULL)
return (EBUSY);
dumper = *di;
wantcopy = strlcpy(dumpdevname, devname, sizeof(dumpdevname));
if (wantcopy >= sizeof(dumpdevname)) {
printf("set_dumper: device name truncated from '%s' -> '%s'\n",
devname, dumpdevname);
}
Here follows the new kernel dumping infrastructure. Caveats: The new savecore program is not complete in the sense that it emulates enough of the old savecores features to do the job, but implements none of the options yet. I would appreciate if a userland hacker could help me out getting savecore to do what we want it to do from a users point of view, compression, email-notification, space reservation etc etc. (send me email if you are interested). Currently, savecore will scan all devices marked as "swap" or "dump" in /etc/fstab _or_ any devices specified on the command-line. All architectures but i386 lack an implementation of dumpsys(), but looking at the i386 version it should be trivial for anybody familiar with the platform(s) to provide this function. Documentation is quite sparse at this time, more to come. Details: ATA and SCSI drivers should work as the dump formatting code has been removed. The IDA, TWE and AAC have not yet been converted. Dumpon now opens the device and uses ioctl(DIOCGKERNELDUMP) to set the device as dumpdev. To implement the "off" argument, /dev/null is used as the device. Savecore will fail if handed any options since they are not (yet) implemented. All devices marked "dump" or "swap" in /etc/fstab will be scanned and dumps found will be saved to diskfiles named from the MD5 hash of the header record. The header record is dumped in readable format in the .info file. The kernel is not saved. Only complete dumps will be saved. All maintainer rights for this code are disclaimed: feel free to improve and extend. Sponsored by: DARPA, NAI Labs
2002-03-31 22:37:00 +00:00
return (0);
}
/* Call dumper with bounds checking. */
int
dump_write(struct dumperinfo *di, void *virtual, vm_offset_t physical,
off_t offset, size_t length)
{
if (length != 0 && (offset < di->mediaoffset ||
offset - di->mediaoffset + length > di->mediasize)) {
printf("Attempt to write outside dump device boundaries.\n"
"offset(%jd), mediaoffset(%jd), length(%ju), mediasize(%jd).\n",
(intmax_t)offset, (intmax_t)di->mediaoffset,
(uintmax_t)length, (intmax_t)di->mediasize);
return (ENOSPC);
}
return (di->dumper(di->priv, virtual, physical, offset, length));
}
void
mkdumpheader(struct kerneldumpheader *kdh, char *magic, uint32_t archver,
uint64_t dumplen, uint32_t blksz)
{
bzero(kdh, sizeof(*kdh));
strncpy(kdh->magic, magic, sizeof(kdh->magic));
strncpy(kdh->architecture, MACHINE_ARCH, sizeof(kdh->architecture));
kdh->version = htod32(KERNELDUMPVERSION);
kdh->architectureversion = htod32(archver);
kdh->dumplength = htod64(dumplen);
kdh->dumptime = htod64(time_second);
kdh->blocksize = htod32(blksz);
strncpy(kdh->hostname, prison0.pr_hostname, sizeof(kdh->hostname));
strncpy(kdh->versionstring, version, sizeof(kdh->versionstring));
if (panicstr != NULL)
strncpy(kdh->panicstring, panicstr, sizeof(kdh->panicstring));
kdh->parity = kerneldump_parity(kdh);
}