[PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[jmerkey]

Centralized external definitions into include files and corrected 
the file to conform to Linux coding practices.  fixed word wrap
problems with patches.

Corrected kprobes section for int3 trap where CONFIG_MDB is always 
true.

Signed-off-by: Jeffrey Vernon Merkey (jmerkey@wolfmountaingroup.com)

--- a/arch/x86/kernel/traps_32.c	2008-08-04 15:53:24.000000000 -0600
+++ b/arch/x86/kernel/traps_32.c	2008-08-04 16:00:25.000000000 -0600
@@ -46,6 +46,10 @@
 #include <linux/edac.h>
 #endif
 
+#ifdef CONFIG_MDB
+#include <linux/mdb.h>
+#endif // CONFIG_MDB
+
 #include <asm/arch_hooks.h>
 #include <asm/stacktrace.h>
 #include <asm/processor.h>
@@ -452,8 +456,15 @@
  */
 void die(const char *str, struct pt_regs *regs, long err)
 {
-	unsigned long flags = oops_begin();
+#ifdef	CONFIG_MDB
+	unsigned long flags;
 
+        mdb_oops = (unsigned char *)str;
+        mdb(SOFTWARE_EXCEPTION, err, regs); 
+	flags = oops_begin();
+#else
+	unsigned long flags = oops_begin();
+#endif // CONFIG_MDB	
 	if (die_nest_count < 3) {
 		report_bug(regs->ip, regs);
 
@@ -573,9 +584,9 @@
 }
 
 DO_VM86_ERROR_INFO(0, SIGFPE, "divide error", divide_error, FPE_INTDIV, regs->ip)
-#ifndef CONFIG_KPROBES
+#if !defined(CONFIG_KPROBES) && !defined(CONFIG_MDB)
 DO_VM86_ERROR(3, SIGTRAP, "int3", int3)
-#endif
+#endif // CONFIG_MDB
 DO_VM86_ERROR(4, SIGSEGV, "overflow", overflow)
 DO_VM86_ERROR(5, SIGSEGV, "bounds", bounds)
 DO_ERROR_INFO(6, SIGILL, "invalid opcode", invalid_op, ILL_ILLOPN, regs->ip, 0)
@@ -718,6 +729,9 @@
 {
 	if (notify_die(DIE_NMIUNKNOWN, "nmi", regs, reason, 2, SIGINT) == NOTIFY_STOP)
 		return;
+#ifdef	CONFIG_MDB
+        mdb(NMI_EXCEPTION, reason, regs); // nmi is code 2
+#endif // CONFIG_MDB
 #ifdef CONFIG_MCA
 	/*
 	 * Might actually be able to figure out what the guilty party
@@ -756,6 +770,9 @@
 	printk(" on CPU%d, ip %08lx, registers:\n",
 		smp_processor_id(), regs->ip);
 	show_registers(regs);
+#ifdef	CONFIG_MDB
+        mdb(NMI_EXCEPTION, 0, regs); // nmi is code 2
+#endif // CONFIG_MDB
 	if (do_panic)
 		panic("Non maskable interrupt");
 	console_silent();
@@ -785,6 +802,10 @@
 	if (!cpu)
 		reason = get_nmi_reason();
 
+#if defined(CONFIG_SMP) && defined(CONFIG_MDB)
+        if (mdb(NMI_EXCEPTION, 0, regs))
+           return;
+#endif // CONFIG_MDB
 	if (!(reason & 0xc0)) {
 		if (notify_die(DIE_NMI_IPI, "nmi_ipi", regs, reason, 2, SIGINT)
 								== NOTIFY_STOP)
@@ -850,6 +871,10 @@
 #ifdef CONFIG_KPROBES
 void __kprobes do_int3(struct pt_regs *regs, long error_code)
 {
+#ifdef	CONFIG_MDB
+        if (mdb(BREAKPOINT_EXCEPTION, error_code, regs))
+           return;
+#endif // CONFIG_MDB
 	trace_hardirqs_fixup();
 
 	if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP)
@@ -865,6 +890,16 @@
 }
 #endif
 
+#if !defined(CONFIG_KPROBES)
+fastcall void do_int3(struct pt_regs * regs, long error_code)
+{
+#if defined(CONFIG_MDB)
+        if (mdb(BREAKPOINT_EXCEPTION, error_code, regs))
+           return;
+#endif // CONFIG_MDB
+	do_trap(3, SIGTRAP, "int3", 1, regs, error_code, NULL);
+}
+#endif	
 /*
  * Our handling of the processor debug registers is non-trivial.
  * We do not clear them on entry and exit from the kernel. Therefore
@@ -895,6 +930,10 @@
 	trace_hardirqs_fixup();
 
 	get_debugreg(condition, 6);
+#ifdef	CONFIG_MDB
+        if (mdb(DEBUGGER_EXCEPTION, error_code, regs))
+           return;
+#endif // CONFIG_MDB	
 
 	/*
 	 * The processor cleared BTF, so don't mark that we need it set.

Re: [PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[jmerkey]

> Hi Jeffery,
>
> Could you explain what is MDB and how it works?


MDB is a kernel debugger that uses the same general architecture
as the kernel debugger in Netware.  Unlike the variants of GNU
debuggers, MDB is geared towards folks from the intel/microsoft worlds
who are more accustomed to the old school Intel assembler model and
not the GNU variant that looks like an angry typewriter cursing at
the screen (%#*&!, etc.).

MDB has some features not yet in this first patch, but I will be adding to
subsequent patches like the ability to create time based traces of
execution of code just like an ice does so you can set trigger breakpoints
(kdb does not, nor kgdb), then go back ad look what happened in windows of
time, as well as a much easier interface, smaller size, more flexible and
pluggable command architecture -- a best of all -- it works and does not
crash every other release or on every third vendors PC architecture like
kdb does all the time.

It has been in use by me since 1997, when I wrote it after I left Novell.
It is mature, hardened, and very stable -- and follows an architecture
prove in the industry for over 20 years on Netware.


> And please check your patch style by using scripts/checkpatch.pl.
> Your cording style seem to be different from linux coding style...
>
> Thank you,


I was unaware of checkpatch.pl, but will run it later today and submit a
git6 patch for any areas not in compliance with linux coding standards.

Jeff

>
> jmerkey@wolfmountaingroup.com wrote:
>> Centralized external definitions into include files and corrected
>> the file to conform to Linux coding practices.  fixed word wrap
>> problems with patches.
>>
>> Corrected kprobes section for int3 trap where CONFIG_MDB is always
>> true.
>>
>> Signed-off-by: Jeffrey Vernon Merkey (jmerkey@wolfmountaingroup.com)
>>
>> --- a/arch/x86/kernel/traps_32.c	2008-08-04 15:53:24.000000000 -0600
>> +++ b/arch/x86/kernel/traps_32.c	2008-08-04 16:00:25.000000000 -0600
>> @@ -46,6 +46,10 @@
>>  #include <linux/edac.h>
>>  #endif
>>
>> +#ifdef CONFIG_MDB
>> +#include <linux/mdb.h>
>> +#endif // CONFIG_MDB
>> +
>>  #include <asm/arch_hooks.h>
>>  #include <asm/stacktrace.h>
>>  #include <asm/processor.h>
>> @@ -452,8 +456,15 @@
>>   */
>>  void die(const char *str, struct pt_regs *regs, long err)
>>  {
>> -	unsigned long flags = oops_begin();
>> +#ifdef	CONFIG_MDB
>> +	unsigned long flags;
>>
>> +        mdb_oops = (unsigned char *)str;
>> +        mdb(SOFTWARE_EXCEPTION, err, regs);
>> +	flags = oops_begin();
>> +#else
>> +	unsigned long flags = oops_begin();
>> +#endif // CONFIG_MDB
>>  	if (die_nest_count < 3) {
>>  		report_bug(regs->ip, regs);
>>
>> @@ -573,9 +584,9 @@
>>  }
>>
>>  DO_VM86_ERROR_INFO(0, SIGFPE, "divide error", divide_error, FPE_INTDIV,
>> regs->ip)
>> -#ifndef CONFIG_KPROBES
>> +#if !defined(CONFIG_KPROBES) && !defined(CONFIG_MDB)
>>  DO_VM86_ERROR(3, SIGTRAP, "int3", int3)
>> -#endif
>> +#endif // CONFIG_MDB
>>  DO_VM86_ERROR(4, SIGSEGV, "overflow", overflow)
>>  DO_VM86_ERROR(5, SIGSEGV, "bounds", bounds)
>>  DO_ERROR_INFO(6, SIGILL, "invalid opcode", invalid_op, ILL_ILLOPN,
>> regs->ip, 0)
>> @@ -718,6 +729,9 @@
>>  {
>>  	if (notify_die(DIE_NMIUNKNOWN, "nmi", regs, reason, 2, SIGINT) ==
>> NOTIFY_STOP)
>>  		return;
>> +#ifdef	CONFIG_MDB
>> +        mdb(NMI_EXCEPTION, reason, regs); // nmi is code 2
>> +#endif // CONFIG_MDB
>>  #ifdef CONFIG_MCA
>>  	/*
>>  	 * Might actually be able to figure out what the guilty party
>> @@ -756,6 +770,9 @@
>>  	printk(" on CPU%d, ip %08lx, registers:\n",
>>  		smp_processor_id(), regs->ip);
>>  	show_registers(regs);
>> +#ifdef	CONFIG_MDB
>> +        mdb(NMI_EXCEPTION, 0, regs); // nmi is code 2
>> +#endif // CONFIG_MDB
>>  	if (do_panic)
>>  		panic("Non maskable interrupt");
>>  	console_silent();
>> @@ -785,6 +802,10 @@
>>  	if (!cpu)
>>  		reason = get_nmi_reason();
>>
>> +#if defined(CONFIG_SMP) && defined(CONFIG_MDB)
>> +        if (mdb(NMI_EXCEPTION, 0, regs))
>> +           return;
>> +#endif // CONFIG_MDB
>>  	if (!(reason & 0xc0)) {
>>  		if (notify_die(DIE_NMI_IPI, "nmi_ipi", regs, reason, 2, SIGINT)
>>  								== NOTIFY_STOP)
>> @@ -850,6 +871,10 @@
>>  #ifdef CONFIG_KPROBES
>>  void __kprobes do_int3(struct pt_regs *regs, long error_code)
>>  {
>> +#ifdef	CONFIG_MDB
>> +        if (mdb(BREAKPOINT_EXCEPTION, error_code, regs))
>> +           return;
>> +#endif // CONFIG_MDB
>>  	trace_hardirqs_fixup();
>>
>>  	if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP)
>> @@ -865,6 +890,16 @@
>>  }
>>  #endif
>>
>> +#if !defined(CONFIG_KPROBES)
>> +fastcall void do_int3(struct pt_regs * regs, long error_code)
>> +{
>> +#if defined(CONFIG_MDB)
>> +        if (mdb(BREAKPOINT_EXCEPTION, error_code, regs))
>> +           return;
>> +#endif // CONFIG_MDB
>> +	do_trap(3, SIGTRAP, "int3", 1, regs, error_code, NULL);
>> +}
>> +#endif
>>  /*
>>   * Our handling of the processor debug registers is non-trivial.
>>   * We do not clear them on entry and exit from the kernel. Therefore
>> @@ -895,6 +930,10 @@
>>  	trace_hardirqs_fixup();
>>
>>  	get_debugreg(condition, 6);
>> +#ifdef	CONFIG_MDB
>> +        if (mdb(DEBUGGER_EXCEPTION, error_code, regs))
>> +           return;
>> +#endif // CONFIG_MDB
>>
>>  	/*
>>  	 * The processor cleared BTF, so don't mark that we need it set.
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel"
>> in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/
>
> --
> Masami Hiramatsu
>
> Software Engineer
> Hitachi Computer Products (America) Inc.
> Software Solutions Division
>
> e-mail: mhiramat@redhat.com
>
>

Re: [PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[Masami Hiramatsu]

Hi Jeffery,

Could you explain what is MDB and how it works?
And please check your patch style by using scripts/checkpatch.pl.
Your cording style seem to be different from linux coding style...

Thank you,

jmerkey@wolfmountaingroup.com wrote:
> Centralized external definitions into include files and corrected 
> the file to conform to Linux coding practices.  fixed word wrap
> problems with patches.
> 
> Corrected kprobes section for int3 trap where CONFIG_MDB is always 
> true.
> 
> Signed-off-by: Jeffrey Vernon Merkey (jmerkey@wolfmountaingroup.com)
> 
> --- a/arch/x86/kernel/traps_32.c	2008-08-04 15:53:24.000000000 -0600
> +++ b/arch/x86/kernel/traps_32.c	2008-08-04 16:00:25.000000000 -0600
> @@ -46,6 +46,10 @@
>  #include <linux/edac.h>
>  #endif
>  
> +#ifdef CONFIG_MDB
> +#include <linux/mdb.h>
> +#endif // CONFIG_MDB
> +
>  #include <asm/arch_hooks.h>
>  #include <asm/stacktrace.h>
>  #include <asm/processor.h>
> @@ -452,8 +456,15 @@
>   */
>  void die(const char *str, struct pt_regs *regs, long err)
>  {
> -	unsigned long flags = oops_begin();
> +#ifdef	CONFIG_MDB
> +	unsigned long flags;
>  
> +        mdb_oops = (unsigned char *)str;
> +        mdb(SOFTWARE_EXCEPTION, err, regs); 
> +	flags = oops_begin();
> +#else
> +	unsigned long flags = oops_begin();
> +#endif // CONFIG_MDB	
>  	if (die_nest_count < 3) {
>  		report_bug(regs->ip, regs);
>  
> @@ -573,9 +584,9 @@
>  }
>  
>  DO_VM86_ERROR_INFO(0, SIGFPE, "divide error", divide_error, FPE_INTDIV, regs->ip)
> -#ifndef CONFIG_KPROBES
> +#if !defined(CONFIG_KPROBES) && !defined(CONFIG_MDB)
>  DO_VM86_ERROR(3, SIGTRAP, "int3", int3)
> -#endif
> +#endif // CONFIG_MDB
>  DO_VM86_ERROR(4, SIGSEGV, "overflow", overflow)
>  DO_VM86_ERROR(5, SIGSEGV, "bounds", bounds)
>  DO_ERROR_INFO(6, SIGILL, "invalid opcode", invalid_op, ILL_ILLOPN, regs->ip, 0)
> @@ -718,6 +729,9 @@
>  {
>  	if (notify_die(DIE_NMIUNKNOWN, "nmi", regs, reason, 2, SIGINT) == NOTIFY_STOP)
>  		return;
> +#ifdef	CONFIG_MDB
> +        mdb(NMI_EXCEPTION, reason, regs); // nmi is code 2
> +#endif // CONFIG_MDB
>  #ifdef CONFIG_MCA
>  	/*
>  	 * Might actually be able to figure out what the guilty party
> @@ -756,6 +770,9 @@
>  	printk(" on CPU%d, ip %08lx, registers:\n",
>  		smp_processor_id(), regs->ip);
>  	show_registers(regs);
> +#ifdef	CONFIG_MDB
> +        mdb(NMI_EXCEPTION, 0, regs); // nmi is code 2
> +#endif // CONFIG_MDB
>  	if (do_panic)
>  		panic("Non maskable interrupt");
>  	console_silent();
> @@ -785,6 +802,10 @@
>  	if (!cpu)
>  		reason = get_nmi_reason();
>  
> +#if defined(CONFIG_SMP) && defined(CONFIG_MDB)
> +        if (mdb(NMI_EXCEPTION, 0, regs))
> +           return;
> +#endif // CONFIG_MDB
>  	if (!(reason & 0xc0)) {
>  		if (notify_die(DIE_NMI_IPI, "nmi_ipi", regs, reason, 2, SIGINT)
>  								== NOTIFY_STOP)
> @@ -850,6 +871,10 @@
>  #ifdef CONFIG_KPROBES
>  void __kprobes do_int3(struct pt_regs *regs, long error_code)
>  {
> +#ifdef	CONFIG_MDB
> +        if (mdb(BREAKPOINT_EXCEPTION, error_code, regs))
> +           return;
> +#endif // CONFIG_MDB
>  	trace_hardirqs_fixup();
>  
>  	if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP)
> @@ -865,6 +890,16 @@
>  }
>  #endif
>  
> +#if !defined(CONFIG_KPROBES)
> +fastcall void do_int3(struct pt_regs * regs, long error_code)
> +{
> +#if defined(CONFIG_MDB)
> +        if (mdb(BREAKPOINT_EXCEPTION, error_code, regs))
> +           return;
> +#endif // CONFIG_MDB
> +	do_trap(3, SIGTRAP, "int3", 1, regs, error_code, NULL);
> +}
> +#endif	
>  /*
>   * Our handling of the processor debug registers is non-trivial.
>   * We do not clear them on entry and exit from the kernel. Therefore
> @@ -895,6 +930,10 @@
>  	trace_hardirqs_fixup();
>  
>  	get_debugreg(condition, 6);
> +#ifdef	CONFIG_MDB
> +        if (mdb(DEBUGGER_EXCEPTION, error_code, regs))
> +           return;
> +#endif // CONFIG_MDB	
>  
>  	/*
>  	 * The processor cleared BTF, so don't mark that we need it set.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

-- 
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America) Inc.
Software Solutions Division

e-mail: mhiramat@redhat.com

Re: [PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[Stefan Richter]

jmerkey@wolfmountaingroup.com wrote on 2008-08-05:
> I was unaware of checkpatch.pl, but will run it later today and submit a
> git6 patch for any areas not in compliance with linux coding standards.

I guess you got complaints from scripts/checkpatch.pl for at least 3/4 
of all lines by now.  Some or many of the complaints will probably look 
silly.  Have a look at Documentation/CodingStyle then because it gives 
context for many of the complaints.  It also gives some rules of thumb 
for things which checkpatch.pl cannot check, e.g. legibility of global 
names.

 From what I saw, there are _many_ whitespace "deviations" in your code 
(which perhaps scripts/Lindent can help to adjust), upper case vs. lower 
case in names, and occasional functions which are deeper nested than 
probably necessary, as already mentioned.
-- 
Stefan Richter
-=====-==--- =--- --==-
http://arcgraph.de/sr/

Re: [PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[Ananth N Mavinakayanahalli]

On Tue, Aug 05, 2008 at 12:26:35AM -0600, jmerkey@wolfmountaingroup.com wrote:
...

> @@ -850,6 +871,10 @@
>  #ifdef CONFIG_KPROBES
>  void __kprobes do_int3(struct pt_regs *regs, long error_code)
>  {
> +#ifdef	CONFIG_MDB
> +        if (mdb(BREAKPOINT_EXCEPTION, error_code, regs))
> +           return;
> +#endif // CONFIG_MDB
>  	trace_hardirqs_fixup();
> 
>  	if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP)

Kprobes needs to be the first consumer of these exceptions (via
notify_die); if the exception was due to a kprobe, it does its thing
without requiring any user intervention. Here, for example, you can get
into mdb for a kprobe breakpoint hit.

Please move the mdb hooks to after kprobes has been notified. Better
still, integrate mdb to use the notify_die infrastructure and use a
lower priority than what kprobes does for it.

Ananth

Re: [PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[Abhishek Sagar]

On Tue, Aug 5, 2008 at 11:56 AM,  <jmerkey@wolfmountaingroup.com> wrote:
> @@ -850,6 +871,10 @@
>  #ifdef CONFIG_KPROBES
>  void __kprobes do_int3(struct pt_regs *regs, long error_code)
>  {
> +#ifdef CONFIG_MDB
> +        if (mdb(BREAKPOINT_EXCEPTION, error_code, regs))
> +           return;
> +#endif // CONFIG_MDB
>        trace_hardirqs_fixup();
>
>        if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP)
> @@ -865,6 +890,16 @@
>  }
>  #endif

If the mdb() call modifies the IF flag in regs, then it should
probably be called after trace_hardirqs_fixup. Best to keep it behind
notify_die() ensuring that kprobes handles all breakpoints first.
Otherwise at least the mdb function (and others in its call chain)
should be marked as __kprobes.

-- Abhishek Sagar

Re: [PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[jmerkey]

> On Tue, Aug 05, 2008 at 12:26:35AM -0600, jmerkey@wolfmountaingroup.com
> wrote:
> ...
>
>> @@ -850,6 +871,10 @@
>>  #ifdef CONFIG_KPROBES
>>  void __kprobes do_int3(struct pt_regs *regs, long error_code)
>>  {
>> +#ifdef	CONFIG_MDB
>> +        if (mdb(BREAKPOINT_EXCEPTION, error_code, regs))
>> +           return;
>> +#endif // CONFIG_MDB
>>  	trace_hardirqs_fixup();
>>
>>  	if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP)
>
> Kprobes needs to be the first consumer of these exceptions (via
> notify_die); if the exception was due to a kprobe, it does its thing
> without requiring any user intervention. Here, for example, you can get
> into mdb for a kprobe breakpoint hit.
>
> Please move the mdb hooks to after kprobes has been notified. Better
> still, integrate mdb to use the notify_die infrastructure and use a
> lower priority than what kprobes does for it.
>
> Ananth
>

I have removed these hooks already and converted to debugger to use
notify_die -- finished it last night as per Andi's suggestions.  This code
section has already been removed from reboot.c and traps_XX.c.

I'll post the new patch with the module based debugger later today.

Jeff

Re: [PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[Stefan Richter]

jmerkey@wolfmountaingroup.com wrote:
> Haven't gotten any complaints on this code so far actually from
> checkpatch.pl since it does not seem to exist in the 2.6.27 tree. (????)

It should exist in the scripts/ subdirectory.  Check whether execute
permission is set for the file.

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=scripts/checkpatch.pl
-- 
Stefan Richter
-=====-==--- =--- --==-
http://arcgraph.de/sr/

Re: [PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[Masami Hiramatsu]

jmerkey@wolfmountaingroup.com wrote:
>> Hi Jeffery,
>>
>> Could you explain what is MDB and how it works?
> 
> 
> MDB is a kernel debugger that uses the same general architecture
> as the kernel debugger in Netware.  Unlike the variants of GNU
> debuggers, MDB is geared towards folks from the intel/microsoft worlds
> who are more accustomed to the old school Intel assembler model and
> not the GNU variant that looks like an angry typewriter cursing at
> the screen (%#*&!, etc.).

Would you have any document about MDB, or project-site url?
I think it is hard for anyone to understand what you like to
do only from these big patchset.

> MDB has some features not yet in this first patch, but I will be adding to
> subsequent patches like the ability to create time based traces of
> execution of code just like an ice does so you can set trigger breakpoints
> (kdb does not, nor kgdb), then go back ad look what happened in windows of
> time, as well as a much easier interface, smaller size, more flexible and
> pluggable command architecture -- a best of all -- it works and does not
> crash every other release or on every third vendors PC architecture like
> kdb does all the time.

Would you have talked with kdb/kgdb people?
Since linux kernel already have kgdb, I think you might better talk
about merging your functionality (and stability) to kgdb, or at least
talk about what part both debugger can share.

> It has been in use by me since 1997, when I wrote it after I left Novell.
> It is mature, hardened, and very stable -- and follows an architecture
> prove in the industry for over 20 years on Netware.

Out of curiously, is there anyone except for you using MDB?

Thank you,

>> And please check your patch style by using scripts/checkpatch.pl.
>> Your cording style seem to be different from linux coding style...
>>
>> Thank you,
> 


-- 
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America) Inc.
Software Solutions Division

e-mail: mhiramat@redhat.com

Re: [PATCH 2.6.27-rc1-git5 2/26] mdb: correct kprobes int3 trap

[jmerkey]

> jmerkey@wolfmountaingroup.com wrote:
>>> Hi Jeffery,
>>>
>>> Could you explain what is MDB and how it works?
>>
>>
>> MDB is a kernel debugger that uses the same general architecture
>> as the kernel debugger in Netware.  Unlike the variants of GNU
>> debuggers, MDB is geared towards folks from the intel/microsoft worlds
>> who are more accustomed to the old school Intel assembler model and
>> not the GNU variant that looks like an angry typewriter cursing at
>> the screen (%#*&!, etc.).
>
> Would you have any document about MDB, or project-site url?
> I think it is hard for anyone to understand what you like to
> do only from these big patchset.

If you install the patch and enter the debugger, you will find quite
excellent documentation **inside** the debugger.  start with "help help"
at the debugger console.

>
>> MDB has some features not yet in this first patch, but I will be adding
>> to
>> subsequent patches like the ability to create time based traces of
>> execution of code just like an ice does so you can set trigger
>> breakpoints
>> (kdb does not, nor kgdb), then go back ad look what happened in windows
>> of
>> time, as well as a much easier interface, smaller size, more flexible
>> and
>> pluggable command architecture -- a best of all -- it works and does not
>> crash every other release or on every third vendors PC architecture like
>> kdb does all the time.
>
> Would you have talked with kdb/kgdb people?
> Since linux kernel already have kgdb, I think you might better talk
> about merging your functionality (and stability) to kgdb, or at least
> talk about what part both debugger can share.
>
>> It has been in use by me since 1997, when I wrote it after I left
>> Novell.
>> It is mature, hardened, and very stable -- and follows an architecture
>> prove in the industry for over 20 years on Netware.
>
> Out of curiously, is there anyone except for you using MDB?
>
> Thank you,


Network General (now NetScout)
Several Others I do not have authority to name because of NDA agreements.

Jeff




>
>>> And please check your patch style by using scripts/checkpatch.pl.
>>> Your cording style seem to be different from linux coding style...
>>>
>>> Thank you,
>>
>
>
> --
> Masami Hiramatsu
>
> Software Engineer
> Hitachi Computer Products (America) Inc.
> Software Solutions Division
>
> e-mail: mhiramat@redhat.com
>
>