The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an entry
break point early in the boot process, but it no longer works. With 7/8 you
have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
Hello Gary,
Can you supply specific instructions to repro the failure? I am
unable to cause any Ctrl+K settings to fail. We always welcome bug
reports to xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit
the power switch) without the kernel notifying the debugger that it’s
rebooted, then you’ll have to resync manually for the initial bp to
work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry
break point early in the boot process, but it no longer works. With 7/8
you
have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
When starting a debug session I use Ctl+Alt+K to set “Will request initial
breakpoint at next boot.” In versions past, this would always result in a
breakpoint when the target loads. This has worked this way, to my knowledge,
for at least 6 years and through a multitude of vesions. It does not,
however, work in 6.5.0003.7/8. In 7/8 it is at best hit and miss, and mostly
miss.
Instructions:
This works in 6.4 and earlier versions of WinDbg.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
Hello Gary,
Can you supply specific instructions to repro the failure? I am
unable to cause any Ctrl+K settings to fail. We always welcome bug
reports to xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit
the power switch) without the kernel notifying the debugger that it’s
rebooted, then you’ll have to resync manually for the initial bp to
work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry
break point early in the boot process, but it no longer works. With 7/8
you
have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
We are unable to repro this problem. You say it is hit or miss, do you
mean it only fails sometimes or it always fails but only on certain
systems? What kind of system is it failing on? What OS is running on
the target?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 12:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
When starting a debug session I use Ctl+Alt+K to set “Will request
initial breakpoint at next boot.” In versions past, this would always
result in a breakpoint when the target loads. This has worked this way,
to my knowledge, for at least 6 years and through a multitude of
vesions. It does not, however, work in 6.5.0003.7/8. In 7/8 it is at
best hit and miss, and mostly miss.
Instructions:
This works in 6.4 and earlier versions of WinDbg.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
Hello Gary,
Can you supply specific instructions to repro the failure? I am unable
to cause any Ctrl+K settings to fail. We always welcome bug reports to
xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit the power
switch) without the kernel notifying the debugger that it’s rebooted,
then you’ll have to resync manually for the initial bp to work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry break point early in the boot process, but it no longer works.
With 7/8 you have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com
To unsubscribe send a blank email to xxxxx@lists.osr.com
It’s a 3 month old XP SP2 on a 1GB RAM, Dual Xeon system with HT enabled
processors. The host is an Inspiron 5150 with 1/2 GB RAM, HT enabled PIV
using Firewire. Note that serial is completely useless on version 7/8. I get
the same response on Vista Beta 1 on this system.
By being able to not repro, do you mean it ALWAYS throws a BP when you
reboot the target? I mean it only works extremely rarely, and I use
“.reboot” several times an hour.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We are unable to repro this problem. You say it is hit or miss, do you
mean it only fails sometimes or it always fails but only on certain
systems? What kind of system is it failing on? What OS is running on
the target?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 12:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
When starting a debug session I use Ctl+Alt+K to set “Will request
initial breakpoint at next boot.” In versions past, this would always
result in a breakpoint when the target loads. This has worked this way,
to my knowledge, for at least 6 years and through a multitude of
vesions. It does not, however, work in 6.5.0003.7/8. In 7/8 it is at
best hit and miss, and mostly miss.
Instructions:
1. Start WinDbg on host.
2. At “Waiting to connect …” press Ctl+Alt+K 3. Boot or power on
target.
4. Shortly after the target System Uptime is displayed a BP should be
encountered.
This works in 6.4 and earlier versions of WinDbg.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
Hello Gary,
Can you supply specific instructions to repro the failure? I am unable
to cause any Ctrl+K settings to fail. We always welcome bug reports to
xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit the power
switch) without the kernel notifying the debugger that it’s rebooted,
then you’ll have to resync manually for the initial bp to work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry break point early in the boot process, but it no longer works.
With 7/8 you have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com
To unsubscribe send a blank email to xxxxx@lists.osr.com
By no repro I mean we have never observed unexpected behavior in any of
our testing.
I believe we’re both using serial connections, though, so our test is
not comparable. Jason, what kind of kd connection were you using?
What problem are you having with the serial connection? If it’s
unreliable connections, have you disabled legacy USB support in the
target’s BIOS settings?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Saturday, November 05, 2005 6:39 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
It’s a 3 month old XP SP2 on a 1GB RAM, Dual Xeon system with HT enabled
processors. The host is an Inspiron 5150 with 1/2 GB RAM, HT enabled PIV
using Firewire. Note that serial is completely useless on version 7/8. I
get the same response on Vista Beta 1 on this system.
By being able to not repro, do you mean it ALWAYS throws a BP when you
reboot the target? I mean it only works extremely rarely, and I use
“.reboot” several times an hour.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We are unable to repro this problem. You say it is hit or miss, do you
mean it only fails sometimes or it always fails but only on certain
systems? What kind of system is it failing on? What OS is running on
the target?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 12:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
When starting a debug session I use Ctl+Alt+K to set “Will request
initial breakpoint at next boot.” In versions past, this would always
result in a breakpoint when the target loads. This has worked this way,
to my knowledge, for at least 6 years and through a multitude of
vesions. It does not, however, work in 6.5.0003.7/8. In 7/8 it is at
best hit and miss, and mostly miss.
Instructions:
1. Start WinDbg on host.
2. At “Waiting to connect …” press Ctl+Alt+K 3. Boot or power on
target.
4. Shortly after the target System Uptime is displayed a BP should be
encountered.
This works in 6.4 and earlier versions of WinDbg.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
Hello Gary,
Can you supply specific instructions to repro the failure? I am unable
to cause any Ctrl+K settings to fail. We always welcome bug reports to
xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit the power
switch) without the kernel notifying the debugger that it’s rebooted,
then you’ll have to resync manually for the initial bp to work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry break point early in the boot process, but it no longer works.
With 7/8 you have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
I was using a serial connection when I last tried to repro.
-----Original Message-----
From: Drew Bliss
Sent: Sunday, November 06, 2005 4:05 PM
To: Kernel Debugging Interest List; Jason Cunningham
Subject: RE: [windbg] 6.5.0003.8 redux
By no repro I mean we have never observed unexpected behavior in any of
our testing.
I believe we’re both using serial connections, though, so our test is
not comparable. Jason, what kind of kd connection were you using?
What problem are you having with the serial connection? If it’s
unreliable connections, have you disabled legacy USB support in the
target’s BIOS settings?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Saturday, November 05, 2005 6:39 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
It’s a 3 month old XP SP2 on a 1GB RAM, Dual Xeon system with HT enabled
processors. The host is an Inspiron 5150 with 1/2 GB RAM, HT enabled PIV
using Firewire. Note that serial is completely useless on version 7/8. I
get the same response on Vista Beta 1 on this system.
By being able to not repro, do you mean it ALWAYS throws a BP when you
reboot the target? I mean it only works extremely rarely, and I use
“.reboot” several times an hour.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We are unable to repro this problem. You say it is hit or miss, do you
mean it only fails sometimes or it always fails but only on certain
systems? What kind of system is it failing on? What OS is running on
the target?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 12:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
When starting a debug session I use Ctl+Alt+K to set “Will request
initial breakpoint at next boot.” In versions past, this would always
result in a breakpoint when the target loads. This has worked this way,
to my knowledge, for at least 6 years and through a multitude of
vesions. It does not, however, work in 6.5.0003.7/8. In 7/8 it is at
best hit and miss, and mostly miss.
Instructions:
1. Start WinDbg on host.
2. At “Waiting to connect …” press Ctl+Alt+K 3. Boot or power on
target.
4. Shortly after the target System Uptime is displayed a BP should be
encountered.
This works in 6.4 and earlier versions of WinDbg.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
Hello Gary,
Can you supply specific instructions to repro the failure? I am unable
to cause any Ctrl+K settings to fail. We always welcome bug reports to
xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit the power
switch) without the kernel notifying the debugger that it’s rebooted,
then you’ll have to resync manually for the initial bp to work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry break point early in the boot process, but it no longer works.
With 7/8 you have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
Once I found a firewire card I could use, I have used it, and not returned
to serial for debug at all. I found such things as 20 minute, or longer boot
times, as well as minutes spent waiting for symbols to load. Those were
consistent on 7/8, and were not present in 6.4.
The same thing applies to Ctl+Alt+K in WinDbg. It ALWAYS works in 6.4, and
almost never works in 6.5.0003.7/8. I have had the same results for 7/8 on
two different dual processor systems.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
I was using a serial connection when I last tried to repro.
-----Original Message-----
From: Drew Bliss
Sent: Sunday, November 06, 2005 4:05 PM
To: Kernel Debugging Interest List; Jason Cunningham
Subject: RE: [windbg] 6.5.0003.8 redux
By no repro I mean we have never observed unexpected behavior in any of
our testing.
I believe we’re both using serial connections, though, so our test is
not comparable. Jason, what kind of kd connection were you using?
What problem are you having with the serial connection? If it’s
unreliable connections, have you disabled legacy USB support in the
target’s BIOS settings?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Saturday, November 05, 2005 6:39 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
It’s a 3 month old XP SP2 on a 1GB RAM, Dual Xeon system with HT enabled
processors. The host is an Inspiron 5150 with 1/2 GB RAM, HT enabled PIV
using Firewire. Note that serial is completely useless on version 7/8. I
get the same response on Vista Beta 1 on this system.
By being able to not repro, do you mean it ALWAYS throws a BP when you
reboot the target? I mean it only works extremely rarely, and I use
“.reboot” several times an hour.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We are unable to repro this problem. You say it is hit or miss, do you
mean it only fails sometimes or it always fails but only on certain
systems? What kind of system is it failing on? What OS is running on
the target?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 12:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
When starting a debug session I use Ctl+Alt+K to set “Will request
initial breakpoint at next boot.” In versions past, this would always
result in a breakpoint when the target loads. This has worked this way,
to my knowledge, for at least 6 years and through a multitude of
vesions. It does not, however, work in 6.5.0003.7/8. In 7/8 it is at
best hit and miss, and mostly miss.
Instructions:
1. Start WinDbg on host.
2. At “Waiting to connect …” press Ctl+Alt+K 3. Boot or power on
target.
4. Shortly after the target System Uptime is displayed a BP should be
encountered.
This works in 6.4 and earlier versions of WinDbg.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
Hello Gary,
Can you supply specific instructions to repro the failure? I am unable
to cause any Ctrl+K settings to fail. We always welcome bug reports to
xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit the power
switch) without the kernel notifying the debugger that it’s rebooted,
then you’ll have to resync manually for the initial bp to work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry break point early in the boot process, but it no longer works.
With 7/8 you have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
We’ll see if we can repro with a 1394 connection.
If you do find yourself using a serial connection at some point and have
slowness it’s usually due to serial line errors caused by Legacy USB
support.
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Monday, November 07, 2005 7:01 AM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
Once I found a firewire card I could use, I have used it, and not
returned to serial for debug at all. I found such things as 20 minute,
or longer boot times, as well as minutes spent waiting for symbols to
load. Those were consistent on 7/8, and were not present in 6.4.
The same thing applies to Ctl+Alt+K in WinDbg. It ALWAYS works in 6.4,
and almost never works in 6.5.0003.7/8. I have had the same results for
7/8 on two different dual processor systems.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
I was using a serial connection when I last tried to repro.
-----Original Message-----
From: Drew Bliss
Sent: Sunday, November 06, 2005 4:05 PM
To: Kernel Debugging Interest List; Jason Cunningham
Subject: RE: [windbg] 6.5.0003.8 redux
By no repro I mean we have never observed unexpected behavior in any of
our testing.
I believe we’re both using serial connections, though, so our test is
not comparable. Jason, what kind of kd connection were you using?
What problem are you having with the serial connection? If it’s
unreliable connections, have you disabled legacy USB support in the
target’s BIOS settings?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Saturday, November 05, 2005 6:39 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
It’s a 3 month old XP SP2 on a 1GB RAM, Dual Xeon system with HT enabled
processors. The host is an Inspiron 5150 with 1/2 GB RAM, HT enabled PIV
using Firewire. Note that serial is completely useless on version 7/8. I
get the same response on Vista Beta 1 on this system.
By being able to not repro, do you mean it ALWAYS throws a BP when you
reboot the target? I mean it only works extremely rarely, and I use
“.reboot” several times an hour.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We are unable to repro this problem. You say it is hit or miss, do you
mean it only fails sometimes or it always fails but only on certain
systems? What kind of system is it failing on? What OS is running on
the target?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 12:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
When starting a debug session I use Ctl+Alt+K to set “Will request
initial breakpoint at next boot.” In versions past, this would always
result in a breakpoint when the target loads. This has worked this way,
to my knowledge, for at least 6 years and through a multitude of
vesions. It does not, however, work in 6.5.0003.7/8. In 7/8 it is at
best hit and miss, and mostly miss.
Instructions:
1. Start WinDbg on host.
2. At “Waiting to connect …” press Ctl+Alt+K 3. Boot or power on
target.
4. Shortly after the target System Uptime is displayed a BP should be
encountered.
This works in 6.4 and earlier versions of WinDbg.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
Hello Gary,
Can you supply specific instructions to repro the failure? I am unable
to cause any Ctrl+K settings to fail. We always welcome bug reports to
xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit the power
switch) without the kernel notifying the debugger that it’s rebooted,
then you’ll have to resync manually for the initial bp to work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry break point early in the boot process, but it no longer works.
With 7/8 you have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
What constitutes “Legacy USB” support? Is the tardiness of 7/8 serial due to
the use of “Legacy USB” on the host or the target? On the target I do not
use any of the available USB connections, at all.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We’ll see if we can repro with a 1394 connection.
If you do find yourself using a serial connection at some point and have
slowness it’s usually due to serial line errors caused by Legacy USB
support.
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Monday, November 07, 2005 7:01 AM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
Once I found a firewire card I could use, I have used it, and not
returned to serial for debug at all. I found such things as 20 minute,
or longer boot times, as well as minutes spent waiting for symbols to
load. Those were consistent on 7/8, and were not present in 6.4.
The same thing applies to Ctl+Alt+K in WinDbg. It ALWAYS works in 6.4,
and almost never works in 6.5.0003.7/8. I have had the same results for
7/8 on two different dual processor systems.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
I was using a serial connection when I last tried to repro.
-----Original Message-----
From: Drew Bliss
Sent: Sunday, November 06, 2005 4:05 PM
To: Kernel Debugging Interest List; Jason Cunningham
Subject: RE: [windbg] 6.5.0003.8 redux
By no repro I mean we have never observed unexpected behavior in any of
our testing.
I believe we’re both using serial connections, though, so our test is
not comparable. Jason, what kind of kd connection were you using?
What problem are you having with the serial connection? If it’s
unreliable connections, have you disabled legacy USB support in the
target’s BIOS settings?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Saturday, November 05, 2005 6:39 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
It’s a 3 month old XP SP2 on a 1GB RAM, Dual Xeon system with HT enabled
processors. The host is an Inspiron 5150 with 1/2 GB RAM, HT enabled PIV
using Firewire. Note that serial is completely useless on version 7/8. I
get the same response on Vista Beta 1 on this system.
By being able to not repro, do you mean it ALWAYS throws a BP when you
reboot the target? I mean it only works extremely rarely, and I use
“.reboot” several times an hour.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We are unable to repro this problem. You say it is hit or miss, do you
mean it only fails sometimes or it always fails but only on certain
systems? What kind of system is it failing on? What OS is running on
the target?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 12:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
When starting a debug session I use Ctl+Alt+K to set “Will request
initial breakpoint at next boot.” In versions past, this would always
result in a breakpoint when the target loads. This has worked this way,
to my knowledge, for at least 6 years and through a multitude of
vesions. It does not, however, work in 6.5.0003.7/8. In 7/8 it is at
best hit and miss, and mostly miss.
Instructions:
1. Start WinDbg on host.
2. At “Waiting to connect …” press Ctl+Alt+K 3. Boot or power on
target.
4. Shortly after the target System Uptime is displayed a BP should be
encountered.
This works in 6.4 and earlier versions of WinDbg.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
Hello Gary,
Can you supply specific instructions to repro the failure? I am unable
to cause any Ctrl+K settings to fail. We always welcome bug reports to
xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit the power
switch) without the kernel notifying the debugger that it’s rebooted,
then you’ll have to resync manually for the initial bp to work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry break point early in the boot process, but it no longer works.
With 7/8 you have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
Legacy USB support is a BIOS setting; some BIOSes don’t support it.
You’d need to disable it, if present, on the target. It doesn’t matter
if you are using USB or not; the legacy support can cause problems with
the serial conn. This is a problem that spans debugger versions (since
it’s not a debugger issue) & is not specific to any particular release.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Tuesday, November 08, 2005 7:10 AM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
What constitutes “Legacy USB” support? Is the tardiness of 7/8 serial
due to
the use of “Legacy USB” on the host or the target? On the target I do
not
use any of the available USB connections, at all.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We’ll see if we can repro with a 1394 connection.
If you do find yourself using a serial connection at some point and have
slowness it’s usually due to serial line errors caused by Legacy USB
support.
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Monday, November 07, 2005 7:01 AM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
Once I found a firewire card I could use, I have used it, and not
returned to serial for debug at all. I found such things as 20 minute,
or longer boot times, as well as minutes spent waiting for symbols to
load. Those were consistent on 7/8, and were not present in 6.4.
The same thing applies to Ctl+Alt+K in WinDbg. It ALWAYS works in 6.4,
and almost never works in 6.5.0003.7/8. I have had the same results for
7/8 on two different dual processor systems.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
I was using a serial connection when I last tried to repro.
-----Original Message-----
From: Drew Bliss
Sent: Sunday, November 06, 2005 4:05 PM
To: Kernel Debugging Interest List; Jason Cunningham
Subject: RE: [windbg] 6.5.0003.8 redux
By no repro I mean we have never observed unexpected behavior in any of
our testing.
I believe we’re both using serial connections, though, so our test is
not comparable. Jason, what kind of kd connection were you using?
What problem are you having with the serial connection? If it’s
unreliable connections, have you disabled legacy USB support in the
target’s BIOS settings?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Saturday, November 05, 2005 6:39 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
It’s a 3 month old XP SP2 on a 1GB RAM, Dual Xeon system with HT enabled
processors. The host is an Inspiron 5150 with 1/2 GB RAM, HT enabled PIV
using Firewire. Note that serial is completely useless on version 7/8. I
get the same response on Vista Beta 1 on this system.
By being able to not repro, do you mean it ALWAYS throws a BP when you
reboot the target? I mean it only works extremely rarely, and I use
“.reboot” several times an hour.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We are unable to repro this problem. You say it is hit or miss, do you
mean it only fails sometimes or it always fails but only on certain
systems? What kind of system is it failing on? What OS is running on
the target?
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 12:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
When starting a debug session I use Ctl+Alt+K to set “Will request
initial breakpoint at next boot.” In versions past, this would always
result in a breakpoint when the target loads. This has worked this way,
to my knowledge, for at least 6 years and through a multitude of
vesions. It does not, however, work in 6.5.0003.7/8. In 7/8 it is at
best hit and miss, and mostly miss.
Instructions:
1. Start WinDbg on host.
2. At “Waiting to connect …” press Ctl+Alt+K 3. Boot or power on
target.
4. Shortly after the target System Uptime is displayed a BP should be
encountered.
This works in 6.4 and earlier versions of WinDbg.
–
The personal opinion of
Gary G. Little
“Jason Cunningham” wrote in message
news:xxxxx@windbg…
Hello Gary,
Can you supply specific instructions to repro the failure? I am unable
to cause any Ctrl+K settings to fail. We always welcome bug reports to
xxxxx@microsoft.com.
If the target reboots out from under kd (for instance, you hit the power
switch) without the kernel notifying the debugger that it’s rebooted,
then you’ll have to resync manually for the initial bp to work.
Thanks,
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Friday, November 04, 2005 7:49 AM
To: Kernel Debugging Interest List
Subject: [windbg] 6.5.0003.8 redux
The WinDbg Ctl+Alt+K sequence is broken. I have used it to force an
entry break point early in the boot process, but it no longer works.
With 7/8 you have to manually force a breakpoint.
I guess it is just the general malaise of the 7/8 version of the
debugger.
–
Gary G. Little
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
“Legacy USB Support” is usually an option you can turn on or off in your
system BIOS. If you go into the setup the next time you reboot, and turn
it OFF, then you turn off this source of interference.
You will lose the ability to use USB keyboards for programs that use the
legacy keyboard real-mode interface – which would be pretty much
nothing these days 
Cheers,
/ h+
Gary G. Little wrote:
What constitutes “Legacy USB” support? Is the tardiness of 7/8 serial due to
the use of “Legacy USB” on the host or the target? On the target I do not
use any of the available USB connections, at all.
That is all well and good, but the original question concerns the inability
of Ctl+Alt+K to inject a breakpoint at the next system start. I use Firewire
and really not interested in serial debugging.
–
The personal opinion of
Gary G. Little
“Jon Watte” wrote in message
news:xxxxx@windbg…
>
> “Legacy USB Support” is usually an option you can turn on or off in your
> system BIOS. If you go into the setup the next time you reboot, and turn
> it OFF, then you turn off this source of interference.
>
> You will lose the ability to use USB keyboards for programs that use the
> legacy keyboard real-mode interface – which would be pretty much nothing
> these days
>
> Cheers,
>
> / h+
>
>
> Gary G. Little wrote:
>> What constitutes “Legacy USB” support? Is the tardiness of 7/8 serial due
>> to the use of “Legacy USB” on the host or the target? On the target I do
>> not use any of the available USB connections, at all.
>>
>
>
We have identified a case where the Ctrl+Alt+K setting will be ignored.
We were able to reproduce no break in this case. If Ctrl+Alt+K is done
before .reboot the break occurs as expected, so the workaround would be
to make sure to set the initial break before .reboot.
We’ll fix this in the next debugger drop.
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Tuesday, November 08, 2005 1:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
That is all well and good, but the original question concerns the
inability of Ctl+Alt+K to inject a breakpoint at the next system start.
I use Firewire and really not interested in serial debugging.
–
The personal opinion of
Gary G. Little
“Jon Watte” wrote in message
news:xxxxx@windbg…
>
> “Legacy USB Support” is usually an option you can turn on or off in
> your system BIOS. If you go into the setup the next time you reboot,
> and turn it OFF, then you turn off this source of interference.
>
> You will lose the ability to use USB keyboards for programs that use
> the legacy keyboard real-mode interface – which would be pretty much
> nothing these days
>
> Cheers,
>
> / h+
>
>
> Gary G. Little wrote:
>> What constitutes “Legacy USB” support? Is the tardiness of 7/8 serial
>> due to the use of “Legacy USB” on the host or the target? On the
>> target I do not use any of the available USB connections, at all.
>>
>
>
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
There seems to be another caveat as well. If you press C+A+K twice to
request a breakpoint on first symbol load, the BP will occur.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We have identified a case where the Ctrl+Alt+K setting will be ignored.
1. 1394 only.
2. An initial break is not set at the time .reboot is used.
3. Ctrl+Alt+K is done after .reboot.
We were able to reproduce no break in this case. If Ctrl+Alt+K is done
before .reboot the break occurs as expected, so the workaround would be
to make sure to set the initial break before .reboot.
We’ll fix this in the next debugger drop.
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Tuesday, November 08, 2005 1:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
That is all well and good, but the original question concerns the
inability of Ctl+Alt+K to inject a breakpoint at the next system start.
I use Firewire and really not interested in serial debugging.
–
The personal opinion of
Gary G. Little
“Jon Watte” wrote in message
news:xxxxx@windbg…
>
> “Legacy USB Support” is usually an option you can turn on or off in
> your system BIOS. If you go into the setup the next time you reboot,
> and turn it OFF, then you turn off this source of interference.
>
> You will lose the ability to use USB keyboards for programs that use
> the legacy keyboard real-mode interface – which would be pretty much
> nothing these days
>
> Cheers,
>
> / h+
>
>
> Gary G. Little wrote:
>> What constitutes “Legacy USB” support? Is the tardiness of 7/8 serial
>> due to the use of “Legacy USB” on the host or the target? On the
>> target I do not use any of the available USB connections, at all.
>>
>
>
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
Yes, the initial-symbol-load break is a completely separate setting and
works in a much different way.
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Wednesday, November 09, 2005 12:33 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
There seems to be another caveat as well. If you press C+A+K twice to
request a breakpoint on first symbol load, the BP will occur.
–
The personal opinion of
Gary G. Little
“Drew Bliss” wrote in message
news:xxxxx@windbg…
We have identified a case where the Ctrl+Alt+K setting will be ignored.
1. 1394 only.
2. An initial break is not set at the time .reboot is used.
3. Ctrl+Alt+K is done after .reboot.
We were able to reproduce no break in this case. If Ctrl+Alt+K is done
before .reboot the break occurs as expected, so the workaround would be
to make sure to set the initial break before .reboot.
We’ll fix this in the next debugger drop.
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Tuesday, November 08, 2005 1:04 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] 6.5.0003.8 redux
That is all well and good, but the original question concerns the
inability of Ctl+Alt+K to inject a breakpoint at the next system start.
I use Firewire and really not interested in serial debugging.
–
The personal opinion of
Gary G. Little
“Jon Watte” wrote in message
news:xxxxx@windbg…
>
> “Legacy USB Support” is usually an option you can turn on or off in
> your system BIOS. If you go into the setup the next time you reboot,
> and turn it OFF, then you turn off this source of interference.
>
> You will lose the ability to use USB keyboards for programs that use
> the legacy keyboard real-mode interface – which would be pretty much
> nothing these days
>
> Cheers,
>
> / h+
>
>
> Gary G. Little wrote:
>> What constitutes “Legacy USB” support? Is the tardiness of 7/8 serial
>> due to the use of “Legacy USB” on the host or the target? On the
>> target I do not use any of the available USB connections, at all.
>>
>
>
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com
—
You are currently subscribed to windbg as: xxxxx@winse.microsoft.com To
unsubscribe send a blank email to xxxxx@lists.osr.com