|
Len Goodman brother,do not know what means,I hope you can tell me,Thank you!
|
|
|
|
|
Thank you,I view next question!
|
|
|
|
|
1) Run Verifier, and check your driver for all faults except 'low resource simulation'. Reboot.
2) Set the machine up for a full kernel dump.
3) Load your driver.
4) Open memory.dmp in windbg.
5) Set the symbols path to the location of the pdb files form your driver build.
6) Set the source path to your drivers source code.
5) Type 'analyze -v' into the command line in windbg.
6) The stack will show the line that caused the error in the relevant file.
==============================
Nothing to say.
|
|
|
|
|
Thank you reply,Thank you!
|
|
|
|
|
Hey sorry to ask this question, but i have read that recharging laptop batteries too often eventually kills the battery...
Is it possible to disable battery charging while using the mains... without having the battery removed?
|
|
|
|
|
You don't. It's entirely automated by the power supply. THe only way to stop charging the battery is to remove it.
Though, since the charging circuitry won't overcharge the battery, I fail to see why you would have to do this.
|
|
|
|
|
I read that the charger will charge the battery as soon as it falls below a certain level and that can kill the battery if repeated often enough.
So that's why i wanted to know how to disable charging the battery (after it is fully charged of course) while it is still attached to the laptop to avoid that "charge recharge" circle.
But you have given me the answer already so thanks!
|
|
|
|
|
Sooo you down voted us for giving you the answer you didn't want to hear??
Sorry, but there is no API to call to disable charging the battery. It's that plain and simple.
If you don't like it, take it up with the computer hardware industry and have them make a new standard that requires exposing an API to do this. Oh, and you better come up with a tons of GOOD reasons to want this and a bunch more explaining why it should not be considered a security risk to do this.
|
|
|
|
|
Dave Kreskowiak wrote: Sooo you down voted us for giving you the answer you didn't want to hear??
Well now i have given you the vote you wanted...
Dave Kreskowiak wrote: If you don't like it, take it up with the computer hardware industry and have them make a new standard that requires exposing an API to do this.
I don't want that functionality, i was just asking if there is such a functionality.
Dave Kreskowiak wrote: Oh, and you better come up with a tons of GOOD reasons to want this and a bunch more explaining why it should not be considered a security risk to do this.
No need to go that far!
|
|
|
|
|
BupeChombaDerrick wrote: Well now i have given you the vote you wanted.
I don't really care about the votes I get, but I thought this one was a cheap shot.
BTW, how's that vision library that "outperforms the state-of-the-art" coming? You might want to define HOW it outperforms the state of the art when you write about it.
|
|
|
|
|
Glad to hear you remember the vision library. So let me take this opportunity to elaborate. First let me talk about the current systems. So far to the best of my knowledge most currently existing computer vision systems use either low or intermediate level features a good example is SIFT (Scale Invariant Feature Transform), HOG (Histogram of Oriented Gradients) both use gradient orientation information to generate sparse histograms. HOGs are usually used together with SVMs (Support Vector Machines) for classification.
From what i know the animal brains builds high-level features hierarchically from center surround receptive field types (retina and Lateral Geniculate Nucleus(LGN)) through the orientation selective receptive fields (primary visual cortex) to more complex IT (Inferior Temporal) Cortex receptive fields.
The use of high level features brings in more challenges. How do you effectively do that on a digital computer?
I have found a way of approximating the hierarchical process by making simplifying assumptions to generate high - level features.
and i have made use of the concept of foveal and peripheral vision... that is concentrating processing power on an object that appears near the foveal and less power away from the point of interest.
But my system does not provide spatial information...it will recognition an object near it's fixation point irrespective of scale and orientation and moderate distortions. But it can recognize millions of objects very fast. An object can be a face,car,plants e.t.c and it can run on a conventional computer.
so
Dave Kreskowiak wrote: You might want to define HOW it outperforms the state of the art when you write about it.
To do that i want to use a ground truth image database and publicly make the results available.
Because of it's efficiency i want to create a visual search engine to index internet based images. And will work similar to Google Goggles
|
|
|
|
|
Usually there's nothing built in that allow you to do that. I wouldn't worry about the battery that much, all you have to understand is that they have a finite lifetime that is a lot shorter than that of the device, in another words, you're going to have to change it at one point or another, so stop worrying about it.
|
|
|
|
|
Thanks for the answer...
|
|
|
|
|
(I'm sorry if this is the wrong forum, it seemed the like the best choice out of the forums)
I have a laptop that continues to randomly freeze and I am looking for help using WinDbg to find out which driver is causing this.
Specs: 2.53ghz dual core proc, Win7 Home 64bit, 8gb of memory, Dell i17r laptop.
I have WinDbg installed and it will launch and allow me to start local debugging, but I must be doing something wrong because I had it running when the machine froze this afternoon and there is no %systemroot%\memory.dmp file. Has anyone successfully used this tool, and if so can you go over some settings with me to see which one I am missing to also be able to successfully use it?
Thanks for reading.
|
|
|
|
|
turbosupramk3 wrote: I'm sorry if this is the wrong forum
Maunder wont create a specific Kernel forum so it is unsurprising you dont know where to post.
Anyway, windbg, you cant install it on the machine you are debugging. You HAVE to use a second machine and install it on that. You then connect the two machines physcally with a cable. 1394 is best but COM port too with a NULL modem cable. (Look at the docs, you have to do some fiddling with the 1394 driver some times). You then need to configvure the target to boot into debug mode. XP and 2K you do it in boot.ini, in Vista and later you do it in system config.
You then start windbg in the second machine and select debug=>kernel=> and select 1394 or COM and then boot the target into debug mode. You can then break the target, look at processes, thread stalls and so on. Not easy to do since you need to look at thread times and so on, BUT a big clue is to look at the kernel stack during the hang. If it stays the same during the hang then it look at the driver near the top and what it is calling. (Get your symbols set up, use the microsoft public symbol server, again, look in the documentaiton of windbg at 'symsrv'.
If the kernel itself isnt hung then it is normally a crss.exe process hang. This often gives a 'black screen' hang. Or a istuation where the mouse can move but the keyboard is inactive. Windbg help has some good tips on debugging hangs, both kernel and user mode.
==============================
Nothing to say.
|
|
|
|
|
Hi,
Thanks for the reply. My screen will usually hang when the screen saver is rotating, only a few times has it hung when I am using the machine actively. It does not peg the CPU, it just calmly sits there frozen.
I really would like to do a local dump, and not try and have a second laptop sitting there if I can. My laptop does not have a firewire cable, it does have virtual coms with usb. I see that Microsoft has a local option and I would like to just dump the stack to a file when it freezes, is this not possible?
|
|
|
|
|
If it freezes, how are you going to trigger the dump??
Chances are good that if the machine froze you're not going to get a dump.
|
|
|
|
|
This is my first time working with windbg, but I was hoping that it kept a running queue of what was in memory, and I could figure it out that way.
How do people solve random freezing issues if this won't work? I tried perfmon, but didn't see anything conclusive.
|
|
|
|
|
turbosupramk3 wrote: How do people solve random freezing issues if this won't work?
By hooking up the debugger from a seperate machine.
You DID read the help files and sample debugging sessions in, well, Help, didn't you??
|
|
|
|
|
Yes I did, it says the instructions are not for local debugging but for host/target debugging.
They have a local option, this is the option I would like to pursue.
|
|
|
|
|
Youc CANT use the local o;ption to debug the machine. You HAVE to use remote debugging.Period.
==============================
Nothing to say.
|
|
|
|
|
How can a machine that is frozen in the kernel runa user mode app?
Also csrss hangs also lock the UI. How can you manipulate an app with no UI?
You HAVE to use a second machine.
==============================
Nothing to say.
|
|
|
|
|
Ah, do you have power off to the USB devices when the screen saver kicks in?
You can disable power saving on the various host controlers in Device manager, try that.
You can also debug using the USB com ports by the way.
And no, you canot locally debug the kernel. Period. You have to use a second machine.
==============================
Nothing to say.
|
|
|
|
|
I'll have to find another machine then I guess, what other choice do I have
Do you know of a tutorial on the cabling? Or can I just use a standard female to female usb cable between two machines to do this? My machine does not power off usb devices at any point.
|
|
|
|
|
Did you say you have USB to COM port adapters? If so then get a second machine, the host, with a COM port, and debug the target machine with a NULL modem cable attatched to its USB to COM adaper https://www.google.co.uk/search?hl=en&gs_nf=1&cp=8&gs_id=1m&xhr=t&q=usb+com+port&bav=on.2,or.r_gc.r_pw.r_qf.,cf.osb&biw=1233&bih=606&um=1&ie=UTF-8&tbm=isch&source=og&sa=N&tab=wi&ei=kTtaT4WrHpDZ8gPguJDfDg[^].
If not it is firewire.
One can, aparently, do US to USB debuging, but since you have trouble with your USB subsys it might be affected. WHo knows.
Firewire is the best.
For a tutorial on connecting the machines look on line, in windbg help, and in the DDK (now called WDK).
==============================
Nothing to say.
|
|
|
|