Breakpoint angst

WINDBG
1

Is anyone else seeing strange behavior when trying to set breakpoints with
the 3.0 flavors? Specifically, I’m able to set breakpoints just fine about
fifty percent of the time. The rest of the time either the breakpoints get
set some arbitrary number of lines after the point I want to set it at or
the breakpoints are unresolved. For the former, I’ve made sure the point
where the breakpoint is to be set is “real” code (i.e., not an inline). For
the latter, this seems to happen in certain modules and other modules (in
the same driver) are fine.

I’ve seen this with 3.0.0007 and 3.0.0010. If there’s a solution, I’d like
to know about it.

–jerry

You are currently subscribed to windbg as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

2

I have had that problem when compiler optimizations confuse the debugger.
I’ve also had that problem when the PDB doesn’t match the binary. Other
than that, the 3.x 'Bags have been great.

Phil

-----Original Message-----
From: Jerry Kelley [mailto:xxxxx@NSISW.COM]
Sent: Friday, June 22, 2001 10:46 AM
To: Kernel Debugging Interest List
Subject: [windbg] Breakpoint angst

Is anyone else seeing strange behavior when trying to set breakpoints with
the 3.0 flavors? Specifically, I’m able to set breakpoints just fine about
fifty percent of the time. The rest of the time either the breakpoints get
set some arbitrary number of lines after the point I want to set it at or
the breakpoints are unresolved. For the former, I’ve made sure the point
where the breakpoint is to be set is “real” code (i.e., not an inline). For
the latter, this seems to happen in certain modules and other modules (in
the same driver) are fine.

I’ve seen this with 3.0.0007 and 3.0.0010. If there’s a solution, I’d like
to know about it.

–jerry

You are currently subscribed to windbg as: xxxxx@intel.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

3

Unresolved breakpoints can happen for 2 reasons:
The correct symbols aren’t loaded
The module is not loaded

The breakpoint on the wrong line would indicate to me that either
The correct symbols aren’t loaded
The source does not match the symbols/code being used

If you have a scenerio where you have correct symbols, etc and believe
it is a debugger bug then please mail xxxxx@microsoft.com with the
info. The info that will give us the best chance of identifing the bug
is:
Complete description of behavior & expected results
Output of the version command
A dump of the target (see the .dump command)
The relevant symbols (excluding OS symbols)
Repro steps

Thanks

-----Original Message-----
From: Barila, Phil [mailto:xxxxx@intel.com]
Sent: Friday, June 22, 2001 11:47 AM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Breakpoint angst

I have had that problem when compiler optimizations confuse the
debugger.
I’ve also had that problem when the PDB doesn’t match the binary. Other
than that, the 3.x 'Bags have been great.

Phil

-----Original Message-----
From: Jerry Kelley [mailto:xxxxx@NSISW.COM]
Sent: Friday, June 22, 2001 10:46 AM
To: Kernel Debugging Interest List
Subject: [windbg] Breakpoint angst

Is anyone else seeing strange behavior when trying to set breakpoints
with
the 3.0 flavors? Specifically, I’m able to set breakpoints just fine
about
fifty percent of the time. The rest of the time either the breakpoints
get
set some arbitrary number of lines after the point I want to set it at
or
the breakpoints are unresolved. For the former, I’ve made sure the point
where the breakpoint is to be set is “real” code (i.e., not an inline).
For
the latter, this seems to happen in certain modules and other modules
(in
the same driver) are fine.

I’ve seen this with 3.0.0007 and 3.0.0010. If there’s a solution, I’d
like
to know about it.

–jerry

You are currently subscribed to windbg as: xxxxx@intel.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: xxxxx@microsoft.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

4

Optimizations shouldn’t be the issue because I’m testing with a checked
build - all optimizations are turned off. The PDBs match the binaries.

This has really been the only issue I’ve had with the 3.0 versions.

-----Original Message-----
From: Barila, Phil [mailto:xxxxx@intel.com]
Sent: Friday, June 22, 2001 1:47 PM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Breakpoint angst

I have had that problem when compiler optimizations confuse the debugger.
I’ve also had that problem when the PDB doesn’t match the binary. Other
than that, the 3.x 'Bags have been great.

Phil

-----Original Message-----
From: Jerry Kelley [mailto:xxxxx@NSISW.COM]
Sent: Friday, June 22, 2001 10:46 AM
To: Kernel Debugging Interest List
Subject: [windbg] Breakpoint angst

Is anyone else seeing strange behavior when trying to set breakpoints with
the 3.0 flavors? Specifically, I’m able to set breakpoints just fine about
fifty percent of the time. The rest of the time either the breakpoints get
set some arbitrary number of lines after the point I want to set it at or
the breakpoints are unresolved. For the former, I’ve made sure the point
where the breakpoint is to be set is “real” code (i.e., not an inline). For
the latter, this seems to happen in certain modules and other modules (in
the same driver) are fine.

I’ve seen this with 3.0.0007 and 3.0.0010. If there’s a solution, I’d like
to know about it.

–jerry

You are currently subscribed to windbg as: xxxxx@intel.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: xxxxx@nsisw.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

5

Make sure you aren’t using the DDK with the buggy setenv.bat, which doesn’t
really disable optimizations in the checked build. From your Checked Build
environment, if you look at the MSC_OPTIMIZATION environment variable, it
should =/Oi /Od

MSC_OPTIMIZATION=/Oi /Od

If it isn’t, then optimizations are happening. Fix the setenv.bat, and the
optimizations will be turned off.

-----Original Message-----
From: Jerry Kelley [mailto:xxxxx@NSISW.COM]
Sent: Friday, June 22, 2001 12:24 PM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Breakpoint angst

Optimizations shouldn’t be the issue because I’m testing with a checked
build - all optimizations are turned off. The PDBs match the binaries.

This has really been the only issue I’ve had with the 3.0 versions.

-----Original Message-----
From: Barila, Phil [mailto:xxxxx@intel.com]
Sent: Friday, June 22, 2001 1:47 PM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Breakpoint angst

I have had that problem when compiler optimizations confuse the debugger.
I’ve also had that problem when the PDB doesn’t match the binary. Other
than that, the 3.x 'Bags have been great.

Phil

-----Original Message-----
From: Jerry Kelley [mailto:xxxxx@NSISW.COM]
Sent: Friday, June 22, 2001 10:46 AM
To: Kernel Debugging Interest List
Subject: [windbg] Breakpoint angst

Is anyone else seeing strange behavior when trying to set breakpoints with
the 3.0 flavors? Specifically, I’m able to set breakpoints just fine about
fifty percent of the time. The rest of the time either the breakpoints get
set some arbitrary number of lines after the point I want to set it at or
the breakpoints are unresolved. For the former, I’ve made sure the point
where the breakpoint is to be set is “real” code (i.e., not an inline). For
the latter, this seems to happen in certain modules and other modules (in
the same driver) are fine.

I’ve seen this with 3.0.0007 and 3.0.0010. If there’s a solution, I’d like
to know about it.

–jerry

You are currently subscribed to windbg as: xxxxx@intel.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: xxxxx@nsisw.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: xxxxx@intel.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

6

Isn’t it a known bug that setenv.bat, by default, leaves optimizations on
even in a checked build?

Check that setenv.bat is setting the environment up the way you think it
should be.

Tom Swanson
XIOtech Corporation

-----Original Message-----
From: Jerry Kelley [mailto:xxxxx@NSISW.COM]
Sent: Friday, June 22, 2001 2:24 PM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Breakpoint angst

Optimizations shouldn’t be the issue because I’m testing with
a checked
build - all optimizations are turned off. The PDBs match the binaries.

This has really been the only issue I’ve had with the 3.0 versions.

-----Original Message-----
From: Barila, Phil [mailto:xxxxx@intel.com]
Sent: Friday, June 22, 2001 1:47 PM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Breakpoint angst

I have had that problem when compiler optimizations confuse
the debugger.
I’ve also had that problem when the PDB doesn’t match the
binary. Other
than that, the 3.x 'Bags have been great.

Phil

-----Original Message-----
From: Jerry Kelley [mailto:xxxxx@NSISW.COM]
Sent: Friday, June 22, 2001 10:46 AM
To: Kernel Debugging Interest List
Subject: [windbg] Breakpoint angst

Is anyone else seeing strange behavior when trying to set
breakpoints with
the 3.0 flavors? Specifically, I’m able to set breakpoints
just fine about
fifty percent of the time. The rest of the time either the
breakpoints get
set some arbitrary number of lines after the point I want to
set it at or
the breakpoints are unresolved. For the former, I’ve made
sure the point
where the breakpoint is to be set is “real” code (i.e., not
an inline). For
the latter, this seems to happen in certain modules and other
modules (in
the same driver) are fine.

I’ve seen this with 3.0.0007 and 3.0.0010. If there’s a
solution, I’d like
to know about it.

–jerry

You are currently subscribed to windbg as: xxxxx@intel.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: xxxxx@nsisw.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: xxxxx@xiotech.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

7

You’re right. At least in the 2195 DDK, MSC_OPTIMIZATION is cleared on
checked builds - the same as in the free build. I changed it per your
suggestion and optimizations where then truly disabled. I looked at
SETENV.BAT in the 2481 DDK and it was corrected.

Thanks for the information.

–jerry

-----Original Message-----
From: Barila, Phil [mailto:xxxxx@intel.com]
Sent: Friday, June 22, 2001 3:20 PM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Breakpoint angst

Make sure you aren’t using the DDK with the buggy setenv.bat, which doesn’t
really disable optimizations in the checked build. From your Checked Build
environment, if you look at the MSC_OPTIMIZATION environment variable, it
should =/Oi /Od

MSC_OPTIMIZATION=/Oi /Od

If it isn’t, then optimizations are happening. Fix the setenv.bat, and the
optimizations will be turned off.

-----Original Message-----
From: Jerry Kelley [mailto:xxxxx@NSISW.COM]
Sent: Friday, June 22, 2001 12:24 PM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Breakpoint angst

Optimizations shouldn’t be the issue because I’m testing with a checked
build - all optimizations are turned off. The PDBs match the binaries.

This has really been the only issue I’ve had with the 3.0 versions.

-----Original Message-----
From: Barila, Phil [mailto:xxxxx@intel.com]
Sent: Friday, June 22, 2001 1:47 PM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Breakpoint angst

I have had that problem when compiler optimizations confuse the debugger.
I’ve also had that problem when the PDB doesn’t match the binary. Other
than that, the 3.x 'Bags have been great.

Phil

-----Original Message-----
From: Jerry Kelley [mailto:xxxxx@NSISW.COM]
Sent: Friday, June 22, 2001 10:46 AM
To: Kernel Debugging Interest List
Subject: [windbg] Breakpoint angst

Is anyone else seeing strange behavior when trying to set breakpoints with
the 3.0 flavors? Specifically, I’m able to set breakpoints just fine about
fifty percent of the time. The rest of the time either the breakpoints get
set some arbitrary number of lines after the point I want to set it at or
the breakpoints are unresolved. For the former, I’ve made sure the point
where the breakpoint is to be set is “real” code (i.e., not an inline). For
the latter, this seems to happen in certain modules and other modules (in
the same driver) are fine.

I’ve seen this with 3.0.0007 and 3.0.0010. If there’s a solution, I’d like
to know about it.

–jerry

You are currently subscribed to windbg as: xxxxx@intel.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: xxxxx@nsisw.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: xxxxx@intel.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: xxxxx@nsisw.com
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com

You are currently subscribed to windbg as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-windbg-$subst(‘Recip.MemberIDChar’)@lists.osr.com