This is the first piece of this thread where you indicate that
everything works okay with v6.4.7, which certainly shoots down the
theory that the dbgprint messages are what’s slowing you down.
Can you collect some additional information:
- Turn on !sym noisy
- turn on the ctrl+D noisy output
- Boot the target
There will be a bunch of transport spew (and some symbol spew inbetween
if there are symbol load attempts). Can you collect this output for
both 6.4 and 6.5 and attach them, to see if there are major differences?
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Thursday, August 18, 2005 12:47 PM
To: Kernel Debugging Interest List
Subject: Re:[windbg] Version 6.5.3.7 slow response
Then I can only say you have taken a very big JUMP backwards. Sorry, but
I
would not recommend that any one download .8 or .7 and use it for the
very
reason that it is totally unusable. With a partial check build setup
using
the HAL and Kernel check builds it took nearly 20 minutes to boot, the
fanfare banner itself taking almost 5 minutes to play. That is totally
unacceptable when you consider that the driver I am working on replaces
all
existent HBAs requiring numerous boots during the day.
The problem is that the slow down is carried over into normal system
operations after boot is complete. Case in point … the HW wizard pops
up
to install the drivers for a target HBA. Things proceed fine until I
tell it
“Have a disk” but then it taes over a minute for the cursor to respond
so I
can tell the wizard the path to the INF file. Given that Jason, would
you
continue to use .7/.8?
I just uninstalled .8 and installed 6.4.7.0, and my only conclusion, in
the
few
minutes I have been using it, is that the .7 .8 debacle is just that. A
debacle
that should not have been foisted off on us as a viable release of such
an
important tool. .7/.8 does NOT work with the checked build, at least it
does
not on my machine.
–
The personal opinion of
Gary G. Little
“Jason Shay” wrote in message
news:xxxxx@windbg…
Note that there is for the most part no difference between 6.5.3.7 and
.8. The core binaries are 100% identical, with the only change being
symsrv.dll and redist.txt (it is a hotfix for a scenario involving
downloading source from http source paths).
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Thursday, August 18, 2005 11:38 AM
To: Kernel Debugging Interest List
Subject: Re:[windbg] Version 6.5.3.7 slow response
Having a side conversation with Don Burn he pointed out to me a
discrepancy
in the versions and version numbers, so I went to the site and
downloaded
6.5.3.8, which still says it is 6.5.3.7. I’m evaluating the problem(s)
again
with this incorrectly monicered version.
I do note that the opening Windows fanfare does not sound like it is
being
played with on a very slowly spinning down record player.
–
The personal opinion of
Gary G. Little
“Jason Shay” wrote in message
news:xxxxx@windbg…
Are you seeing a lot of spew coming out of the debugger during the
slowdowns? That’s usually the cause.
Jason
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Gary G. Little
Sent: Thursday, August 18, 2005 9:35 AM
To: Kernel Debugging Interest List
Subject: [windbg] Version 6.5.3.7 slow response
There is a delay during start up that almost makes SoftIce look good. I
notice it most when I use a partial checked build because the
environment
becomes totally unusable. Mouse moves take minutes to happen, and
breaking
into the system now MIGHT happen 5 minutes from now. I suspect there may
be
a big problem between the versions of the ntoskernel files for SP2 I
have.
The fresh install of SP2 I yesterday has a build number several hundred
greater than the checked build version I down loaded today.
At least the free build seems to respond properly once you are at the
desktop. Debugging during boot may be a problem since I note the mouse
moving at cycles per minute before Winlogon is presented.
Am I alone in this observation? Or did I miss the discussion?
–
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