+ print-utsname-on-oops-on-all-architectures.patch added to -mm tree

+ print-utsname-on-oops-on-all-architectures.patch added to -mm tree

[akpm]

The patch titled
     Print utsname on Oops on all architectures
has been added to the -mm tree.  Its filename is
     print-utsname-on-oops-on-all-architectures.patch

*** Remember to use Documentation/SubmitChecklist when testing your code ***

See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this

------------------------------------------------------
Subject: Print utsname on Oops on all architectures
From: Joshua Wise <jwise@google.com>

Background:
 This patch is a follow-on to "Info dump on Oops or panic()" [1].

 On some architectures, the kernel printed some information on the running
 kernel, but not on all architectures. The information printed was generally
 the version and build number, but it was not located in a consistant place,
 and some architectures did not print it at all.

Description:
 This patch uses the already-existing die_chain to print utsname information
 on Oops. This patch also removes the architecture-specific utsname
 printers. To avoid crashing the system further (and hence not printing the
 Oops) in the case where the system is so hopelessly smashed that utsname
 might be destroyed, we vsprintf the utsname data into a static buffer
 first, and then just print that on crash.

Testing:
 I wrote a module that does a *(int*)0 = 0; and observed that I got my
 utsname data printed.

Potential impact:
 This adds another line to the Oops output, causing the first few lines to
 potentially scroll off the screen. This also adds a few more pointer
 dereferences in the Oops path, because it adds to the die_chain notifier
 chain, reducing the likelihood that the Oops will be printed if there is
 very bad memory corruption.

Signed-off-by: Joshua Wise <jwise@google.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Andi Kleen <ak@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 arch/alpha/kernel/process.c   |    1 -
 arch/arm/kernel/process.c     |    7 ++-----
 arch/i386/kernel/process.c    |    7 ++-----
 arch/powerpc/kernel/process.c |    5 ++---
 arch/x86_64/kernel/process.c  |    8 ++------
 kernel/sys.c                  |   31 +++++++++++++++++++++++++++++++
 6 files changed, 39 insertions(+), 20 deletions(-)

diff -puN arch/alpha/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/alpha/kernel/process.c
--- a/arch/alpha/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/alpha/kernel/process.c
@@ -20,7 +20,6 @@
 #include <linux/slab.h>
 #include <linux/user.h>
 #include <linux/a.out.h>
-#include <linux/utsname.h>
 #include <linux/time.h>
 #include <linux/major.h>
 #include <linux/stat.h>
diff -puN arch/arm/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/arm/kernel/process.c
--- a/arch/arm/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/arm/kernel/process.c
@@ -28,7 +28,6 @@
 #include <linux/elfcore.h>
 #include <linux/pm.h>
 #include <linux/tick.h>
-#include <linux/utsname.h>
 
 #include <asm/leds.h>
 #include <asm/processor.h>
@@ -207,10 +206,8 @@ void __show_regs(struct pt_regs *regs)
 	unsigned long flags;
 	char buf[64];
 
-	printk("CPU: %d    %s  (%s %.*s)\n",
-		smp_processor_id(), print_tainted(), init_utsname()->release,
-		(int)strcspn(init_utsname()->version, " "),
-		init_utsname()->version);
+	printk("CPU: %d    %s\n",
+		smp_processor_id(), print_tainted());
 	print_symbol("PC is at %s\n", instruction_pointer(regs));
 	print_symbol("LR is at %s\n", regs->ARM_lr);
 	printk("pc : [<%08lx>]    lr : [<%08lx>]    psr: %08lx\n"
diff -puN arch/i386/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/i386/kernel/process.c
--- a/arch/i386/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/i386/kernel/process.c
@@ -27,7 +27,6 @@
 #include <linux/user.h>
 #include <linux/a.out.h>
 #include <linux/interrupt.h>
-#include <linux/utsname.h>
 #include <linux/delay.h>
 #include <linux/reboot.h>
 #include <linux/init.h>
@@ -311,10 +310,8 @@ void show_regs(struct pt_regs * regs)
 
 	if (user_mode_vm(regs))
 		printk(" ESP: %04x:%08lx",0xffff & regs->xss,regs->esp);
-	printk(" EFLAGS: %08lx    %s  (%s %.*s)\n",
-	       regs->eflags, print_tainted(), init_utsname()->release,
-	       (int)strcspn(init_utsname()->version, " "),
-	       init_utsname()->version);
+	printk(" EFLAGS: %08lx    %s\n",
+	       regs->eflags, print_tainted());
 	printk("EAX: %08lx EBX: %08lx ECX: %08lx EDX: %08lx\n",
 		regs->eax,regs->ebx,regs->ecx,regs->edx);
 	printk("ESI: %08lx EDI: %08lx EBP: %08lx",
diff -puN arch/powerpc/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/powerpc/kernel/process.c
--- a/arch/powerpc/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/powerpc/kernel/process.c
@@ -32,7 +32,6 @@
 #include <linux/kallsyms.h>
 #include <linux/mqueue.h>
 #include <linux/hardirq.h>
-#include <linux/utsname.h>
 
 #include <asm/pgtable.h>
 #include <asm/uaccess.h>
@@ -416,8 +415,8 @@ void show_regs(struct pt_regs * regs)
 
 	printk("NIP: "REG" LR: "REG" CTR: "REG"\n",
 	       regs->nip, regs->link, regs->ctr);
-	printk("REGS: %p TRAP: %04lx   %s  (%s)\n",
-	       regs, regs->trap, print_tainted(), init_utsname()->release);
+	printk("REGS: %p TRAP: %04lx   %s\n",
+	       regs, regs->trap, print_tainted());
 	printk("MSR: "REG" ", regs->msr);
 	printbits(regs->msr, msr_bits);
 	printk("  CR: %08lx  XER: %08lx\n", regs->ccr, regs->xer);
diff -puN arch/x86_64/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/x86_64/kernel/process.c
--- a/arch/x86_64/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/x86_64/kernel/process.c
@@ -32,7 +32,6 @@
 #include <linux/interrupt.h>
 #include <linux/delay.h>
 #include <linux/ptrace.h>
-#include <linux/utsname.h>
 #include <linux/random.h>
 #include <linux/notifier.h>
 #include <linux/kprobes.h>
@@ -316,11 +315,8 @@ void __show_regs(struct pt_regs * regs)
 
 	printk("\n");
 	print_modules();
-	printk("Pid: %d, comm: %.20s %s %s %.*s\n",
-		current->pid, current->comm, print_tainted(),
-		init_utsname()->release,
-		(int)strcspn(init_utsname()->version, " "),
-		init_utsname()->version);
+	printk("Pid: %d, comm: %.20s %s\n",
+		current->pid, current->comm, print_tainted());
 	printk("RIP: %04lx:[<%016lx>] ", regs->cs & 0xffff, regs->rip);
 	printk_address(regs->rip); 
 	printk("RSP: %04lx:%016lx  EFLAGS: %08lx\n", regs->ss, regs->rsp,
diff -puN kernel/sys.c~print-utsname-on-oops-on-all-architectures kernel/sys.c
--- a/kernel/sys.c~print-utsname-on-oops-on-all-architectures
+++ a/kernel/sys.c
@@ -27,6 +27,7 @@
 #include <linux/dcookies.h>
 #include <linux/suspend.h>
 #include <linux/tty.h>
+#include <linux/kdebug.h>
 #include <linux/signal.h>
 #include <linux/cn_proc.h>
 #include <linux/getcpu.h>
@@ -107,6 +108,36 @@ void (*pm_power_off_prepare)(void);
 EXPORT_SYMBOL(pm_power_off_prepare);
 
 /*
+ * Dump out UTS info on oops / panic.  We pregenerate utsname_info so that
+ * we don't access utsname() if the machine is in a potentially bad state.
+ */
+
+static char *utsname_info;
+
+static int dump_utsname(struct notifier_block *self, unsigned long diecode,
+		void *p)
+{
+	printk(KERN_EMERG "%s\n", utsname_info);
+	return 0;
+}
+
+static struct notifier_block utsname_notifier = {
+	.notifier_call = dump_utsname
+};
+
+static int __init register_utsname_dump(void)
+{
+	utsname_info = kasprintf(GFP_KERNEL, "%s %s %s %s",
+		utsname()->sysname,
+		utsname()->release,
+		utsname()->version,
+		utsname()->machine);
+	register_die_notifier(&utsname_notifier);
+	return 0;
+}
+module_init(register_utsname_dump);
+
+/*
  *	Notifier list for kernel code which wants to be called
  *	at shutdown. This is used to stop any idling DMA operations
  *	and the like. 
_

Patches currently in -mm which might be from jwise@google.com are

print-utsname-on-oops-on-all-architectures.patch
print-utsname-on-oops-on-all-architectures-fix.patch

+ print-utsname-on-oops-on-all-architectures.patch added to -mm tree

[akpm]

The patch titled
     Print utsname on Oops on all architectures
has been added to the -mm tree.  Its filename is
     print-utsname-on-oops-on-all-architectures.patch

*** Remember to use Documentation/SubmitChecklist when testing your code ***

See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this

------------------------------------------------------
Subject: Print utsname on Oops on all architectures
From: Joshua Wise <jwise@google.com>

Background:
 This patch is a follow-on to "Info dump on Oops or panic()" [1].

 On some architectures, the kernel printed some information on the running
 kernel, but not on all architectures. The information printed was generally
 the version and build number, but it was not located in a consistant place,
 and some architectures did not print it at all.

Description:
 This patch uses the already-existing die_chain to print utsname information
 on Oops. This patch also removes the architecture-specific utsname
 printers. To avoid crashing the system further (and hence not printing the
 Oops) in the case where the system is so hopelessly smashed that utsname
 might be destroyed, we vsprintf the utsname data into a static buffer
 first, and then just print that on crash.

Testing:
 I wrote a module that does a *(int*)0 = 0; and observed that I got my
 utsname data printed.

Potential impact:
 This adds another line to the Oops output, causing the first few lines to
 potentially scroll off the screen. This also adds a few more pointer
 dereferences in the Oops path, because it adds to the die_chain notifier
 chain, reducing the likelihood that the Oops will be printed if there is
 very bad memory corruption.

Signed-off-by: Joshua Wise <jwise@google.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Andi Kleen <ak@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 arch/alpha/kernel/process.c   |    1 -
 arch/arm/kernel/process.c     |    7 ++-----
 arch/i386/kernel/process.c    |    7 ++-----
 arch/powerpc/kernel/process.c |    5 ++---
 arch/x86_64/kernel/process.c  |    8 ++------
 kernel/sys.c                  |   31 +++++++++++++++++++++++++++++++
 6 files changed, 39 insertions(+), 20 deletions(-)

diff -puN arch/alpha/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/alpha/kernel/process.c
--- a/arch/alpha/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/alpha/kernel/process.c
@@ -20,7 +20,6 @@
 #include <linux/slab.h>
 #include <linux/user.h>
 #include <linux/a.out.h>
-#include <linux/utsname.h>
 #include <linux/time.h>
 #include <linux/major.h>
 #include <linux/stat.h>
diff -puN arch/arm/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/arm/kernel/process.c
--- a/arch/arm/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/arm/kernel/process.c
@@ -28,7 +28,6 @@
 #include <linux/elfcore.h>
 #include <linux/pm.h>
 #include <linux/tick.h>
-#include <linux/utsname.h>
 
 #include <asm/leds.h>
 #include <asm/processor.h>
@@ -207,10 +206,8 @@ void __show_regs(struct pt_regs *regs)
 	unsigned long flags;
 	char buf[64];
 
-	printk("CPU: %d    %s  (%s %.*s)\n",
-		smp_processor_id(), print_tainted(), init_utsname()->release,
-		(int)strcspn(init_utsname()->version, " "),
-		init_utsname()->version);
+	printk("CPU: %d    %s\n",
+		smp_processor_id(), print_tainted());
 	print_symbol("PC is at %s\n", instruction_pointer(regs));
 	print_symbol("LR is at %s\n", regs->ARM_lr);
 	printk("pc : [<%08lx>]    lr : [<%08lx>]    psr: %08lx\n"
diff -puN arch/i386/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/i386/kernel/process.c
--- a/arch/i386/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/i386/kernel/process.c
@@ -27,7 +27,6 @@
 #include <linux/user.h>
 #include <linux/a.out.h>
 #include <linux/interrupt.h>
-#include <linux/utsname.h>
 #include <linux/delay.h>
 #include <linux/reboot.h>
 #include <linux/init.h>
@@ -311,10 +310,8 @@ void show_regs(struct pt_regs * regs)
 
 	if (user_mode_vm(regs))
 		printk(" ESP: %04x:%08lx",0xffff & regs->xss,regs->esp);
-	printk(" EFLAGS: %08lx    %s  (%s %.*s)\n",
-	       regs->eflags, print_tainted(), init_utsname()->release,
-	       (int)strcspn(init_utsname()->version, " "),
-	       init_utsname()->version);
+	printk(" EFLAGS: %08lx    %s\n",
+	       regs->eflags, print_tainted());
 	printk("EAX: %08lx EBX: %08lx ECX: %08lx EDX: %08lx\n",
 		regs->eax,regs->ebx,regs->ecx,regs->edx);
 	printk("ESI: %08lx EDI: %08lx EBP: %08lx",
diff -puN arch/powerpc/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/powerpc/kernel/process.c
--- a/arch/powerpc/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/powerpc/kernel/process.c
@@ -32,7 +32,6 @@
 #include <linux/kallsyms.h>
 #include <linux/mqueue.h>
 #include <linux/hardirq.h>
-#include <linux/utsname.h>
 
 #include <asm/pgtable.h>
 #include <asm/uaccess.h>
@@ -416,8 +415,8 @@ void show_regs(struct pt_regs * regs)
 
 	printk("NIP: "REG" LR: "REG" CTR: "REG"\n",
 	       regs->nip, regs->link, regs->ctr);
-	printk("REGS: %p TRAP: %04lx   %s  (%s)\n",
-	       regs, regs->trap, print_tainted(), init_utsname()->release);
+	printk("REGS: %p TRAP: %04lx   %s\n",
+	       regs, regs->trap, print_tainted());
 	printk("MSR: "REG" ", regs->msr);
 	printbits(regs->msr, msr_bits);
 	printk("  CR: %08lx  XER: %08lx\n", regs->ccr, regs->xer);
diff -puN arch/x86_64/kernel/process.c~print-utsname-on-oops-on-all-architectures arch/x86_64/kernel/process.c
--- a/arch/x86_64/kernel/process.c~print-utsname-on-oops-on-all-architectures
+++ a/arch/x86_64/kernel/process.c
@@ -32,7 +32,6 @@
 #include <linux/interrupt.h>
 #include <linux/delay.h>
 #include <linux/ptrace.h>
-#include <linux/utsname.h>
 #include <linux/random.h>
 #include <linux/notifier.h>
 #include <linux/kprobes.h>
@@ -316,11 +315,8 @@ void __show_regs(struct pt_regs * regs)
 
 	printk("\n");
 	print_modules();
-	printk("Pid: %d, comm: %.20s %s %s %.*s\n",
-		current->pid, current->comm, print_tainted(),
-		init_utsname()->release,
-		(int)strcspn(init_utsname()->version, " "),
-		init_utsname()->version);
+	printk("Pid: %d, comm: %.20s %s\n",
+		current->pid, current->comm, print_tainted());
 	printk("RIP: %04lx:[<%016lx>] ", regs->cs & 0xffff, regs->rip);
 	printk_address(regs->rip); 
 	printk("RSP: %04lx:%016lx  EFLAGS: %08lx\n", regs->ss, regs->rsp,
diff -puN kernel/sys.c~print-utsname-on-oops-on-all-architectures kernel/sys.c
--- a/kernel/sys.c~print-utsname-on-oops-on-all-architectures
+++ a/kernel/sys.c
@@ -27,6 +27,7 @@
 #include <linux/dcookies.h>
 #include <linux/suspend.h>
 #include <linux/tty.h>
+#include <linux/kdebug.h>
 #include <linux/signal.h>
 #include <linux/cn_proc.h>
 #include <linux/getcpu.h>
@@ -107,6 +108,36 @@ void (*pm_power_off_prepare)(void);
 EXPORT_SYMBOL(pm_power_off_prepare);
 
 /*
+ * Dump out UTS info on oops / panic.  We pregenerate utsname_info so that
+ * we don't access utsname() if the machine is in a potentially bad state.
+ */
+
+static char *utsname_info;
+
+static int dump_utsname(struct notifier_block *self, unsigned long diecode,
+		void *p)
+{
+	printk(KERN_EMERG "%s\n", utsname_info);
+	return 0;
+}
+
+static struct notifier_block utsname_notifier = {
+	.notifier_call = dump_utsname
+};
+
+static int __init register_utsname_dump(void)
+{
+	utsname_info = kasprintf(GFP_KERNEL, "%s %s %s %s",
+		utsname()->sysname,
+		utsname()->release,
+		utsname()->version,
+		utsname()->machine);
+	register_die_notifier(&utsname_notifier);
+	return 0;
+}
+module_init(register_utsname_dump);
+
+/*
  *	Notifier list for kernel code which wants to be called
  *	at shutdown. This is used to stop any idling DMA operations
  *	and the like. 
_

Patches currently in -mm which might be from jwise@google.com are

print-utsname-on-oops-on-all-architectures.patch

Re: + print-utsname-on-oops-on-all-architectures.patch added to -mm tree

[Russell King]

On Fri, Jul 06, 2007 at 10:54:33AM -0700, akpm@linux-foundation.org wrote:
> Potential impact:
>  This adds another line to the Oops output, causing the first few lines to
>  potentially scroll off the screen. This also adds a few more pointer
>  dereferences in the Oops path, because it adds to the die_chain notifier
>  chain, reducing the likelihood that the Oops will be printed if there is
>  very bad memory corruption.

Plus we don't get the utsname information on oops dumps during the kernel
initialisation.  Not good - I'd rather keep things as is rather than loose
that facility which we've just gained on ARM.

Instead, can we have this as a call-able function which returns the
pointer, or just make the pointer global.  That way we can also eliminate
the need for another additional line in the oops output.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:

Re: + print-utsname-on-oops-on-all-architectures.patch added to -mm tree

[Joshua Wise]

On Fri, 6 Jul 2007, Russell King wrote:
> On Fri, Jul 06, 2007 at 10:54:33AM -0700, akpm@linux-foundation.org wrote:
> > Potential impact:
> >  This adds another line to the Oops output, causing the first few lines to
> >  potentially scroll off the screen. This also adds a few more pointer
> >  dereferences in the Oops path, because it adds to the die_chain notifier
> >  chain, reducing the likelihood that the Oops will be printed if there is
> >  very bad memory corruption.
> 
> Plus we don't get the utsname information on oops dumps during the kernel
> initialisation.  Not good - I'd rather keep things as is rather than loose
> that facility which we've just gained on ARM.

Ok -- would it be more agreeable to make that a core_initcall?

> Instead, can we have this as a call-able function which returns the
> pointer, or just make the pointer global.  That way we can also eliminate
> the need for another additional line in the oops output.

Hm. I don't really like the precedent being set for sucking arbitrary
globals into the Oops message. It's not particularly extensible, and the
additional line at the top doesn't seem like it will kill anyone...

In particular, I am also considering a patch to print machine check
statistics (if we took any machine check exceptions) at Oops-time, and I
would implement that in a similar manner to this.

joshua

Re: + print-utsname-on-oops-on-all-architectures.patch added to -mm tree

[Russell King]

On Fri, Jul 06, 2007 at 02:03:45PM -0700, Joshua Wise wrote:
> On Fri, 6 Jul 2007, Russell King wrote:
> > On Fri, Jul 06, 2007 at 10:54:33AM -0700, akpm@linux-foundation.org wrote:
> > > Potential impact:
> > >  This adds another line to the Oops output, causing the first few lines to
> > >  potentially scroll off the screen. This also adds a few more pointer
> > >  dereferences in the Oops path, because it adds to the die_chain notifier
> > >  chain, reducing the likelihood that the Oops will be printed if there is
> > >  very bad memory corruption.
> > 
> > Plus we don't get the utsname information on oops dumps during the kernel
> > initialisation.  Not good - I'd rather keep things as is rather than loose
> > that facility which we've just gained on ARM.
> 
> Ok -- would it be more agreeable to make that a core_initcall?

That just makes the reporting slightly earlier and doesn't solve the
real problem.

> > Instead, can we have this as a call-able function which returns the
> > pointer, or just make the pointer global.  That way we can also eliminate
> > the need for another additional line in the oops output.
> 
> Hm. I don't really like the precedent being set for sucking arbitrary
> globals into the Oops message.

It's much safer than walking notifier lists and such like.

> In particular, I am also considering a patch to print machine check
> statistics (if we took any machine check exceptions) at Oops-time, and I
> would implement that in a similar manner to this.

I'm also going to raise an additional objection - whoever introduced
notify_die() hasn't bothered fixing up the ARM implementation, so your
patch actually _removes_ functionality recently introduced.

In addition, I notice that the introduction of notify_die() didn't take
account of what other architectures do (eg passing siginfo structures),
so to convert ARM to using it would either require us to drop some 
functionality or re-engineer notify_die().

Therefore, I'm going to ask you to remove the changes from your patch
which touch the ARM architecture since the overall effect is a net
removal of functionality I've only recently merged.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of: