bswap intrinsic?

TheMSsForum.com: The Microsoft Software Forum

The MSS Forum ‹ Drivers
- Archive
  - Biz
  - MCSE
  - CRM
  - Drivers
  - Framework
  - ADO
  - ASP
  - Compact
  - Forms
  - Dotnet
  - C#
  - VB
  - FontpageGen
  - Excel
  - WorkSheet
  - Exchange
  - Setup
  - Fox
  - Fontpage
  - ASP
  - IIS
  - Entourage
  - Money
  - Messanger
  - PocketPC
  - Powerpoint
  - Project
  - Publisher
  - Excel
  - VB
  - Security
  - Portal
  - Services
  - SQLServerDev
  - SVCS
  - SQLServer
  - VB
  - VC
  - MFC
  - ExcelGen

- Previous
  - 1
    - what io the Best method to capture all audio out? Hi, 1) I want to capture all the audio out on my machine . What i understand is that if the sound card on my machine provides "stereo Mix" capability AND I am not recording something on the physical line in of my sound card , then and only then i can capture the audio out going through the sound card using DirectSound or WaveIn API. what is the most reliable way to do this that should work on Win 9x+ and Win 2k+ , independent of whether the sound card supports stereo mix or whether something is also being recorded in the physical line in of the sound card? Is there some way to generically install an upper filter driver on top of the sound card driver and do this? what other option should help here? In the DDK i have a Microsoft Virtual Audio Driver , will that be of use ? There are multiple drivers inside the MSVAD driver directory, which one can i use with minimum amount of code change that should achieve my purpose ? and if i do need to change the code , which portion needs to be changed ? 2) I was going through the code of MSVAD and its architecture looks different from the standard WDM drivers ( particularly WDM USB client drivers ). No Read/Write Dispatchers etc. also Walter Oney's Book apparently does not delve into sound card drivers as well. What is the best resource to get more information about this , apart from the DDK ? thanks and regards Taha Tag: bswap intrinsic? Tag: 57401
  - 2
    - Handling SRB_FUNCTION_SHUTDOWN I am trying to see whether I can use SRB_FUNCTION_SHUTDOWN as a cue to start spindown of attached physical drives (both drives that are configured into virtual disks and those that are not) on a PCI RAID controller. Please clarify the following: 1) The DDK says "The system is being shut down. The request is passed to the miniport driver if CachesData was set to TRUE in the PORT_CONFIGURATION_INFORMATION data. The miniport driver can receive several of these notifications before all system activity is actually stopped; however, the last shutdown notification will occur after the last start I/O. Only the Function, PathId, TargetId and Lun fields are valid." How do I know which is the last start I/O? 2) Is there a mechanism that is available so that I can reliably spin down drives without having to spin them again because there was one more I/O? 3) In cases where I have a large number of drives, I will not be able to spinup the drives very quickly. We risk having command timeouts as the controller firmware tries to spinup drives that it just spun down. Any recommendation is appreciated. Thanks, Jacob Tag: bswap intrinsic? Tag: 57392
  - 3
    - To what device group does an intelligent frame grabber fit? Hi, I am just about to start reading the basics about windows drivers (I havent started my project...) just a question about something in general: I was wondering to which device class a pci frame grabber belongs to. I am talking about a PCI-card to which you can connect 4 or 6 industrial type CMOS cameras via some kind of LVDS bus or camera link. On the card, a dedicated processor is grabbing and evaluating the images (detection of shapes or patterns, datamatrix and OCR searches and so on). The card is supplying a video stream through the PCI bus to be displayed in a Window. Also single images can be grabbed. A user mode application communicates to the card, telling it parameters of the jobs it has to do, in which areas of the image it has to detect what and reading out the results... So I am wondering... do I have to write a driver for a PCI multifunction device? or a video streaming device? or a WIA device? or anything else? which of the samples supplied in the XP DDK would match my needs best? Anyway, I have to read a lot before starting... but maybe someone can give me a hint meanwhile. thomas Tag: bswap intrinsic? Tag: 57391
  - 4
    - (NT-)VDD blocks VDM: How to circumvent, and how to perform "secure" cleanup? Hello, I'm writing a VDD (for NT/2K/XP) for the first time, and I found the documentation is not the verbose as I would like it to be. I'm using the mechanism via DISPATCHCALL (assembler bytes 0xc4, 0xc4, 0x58, 0x02) emmitted into the source code, not via IN/OUT instruction (although, by testing, I found out the same applies for these, too). I faced the following problems: The DOS program in the DOS box calls a function of the VDD. Now, that function needs some time to perform the action. As long as the VDD function has not returned, the DOS box is completely blocked. I cannot even issue a Ctrl+C (or Ctrl+Break) to stop the program. This is very bad with programs which need to communicate very much with the VDD: I found out it is almost impossible to stop the DOS mode program at all (other than closing the whole DOS window). Another problem I faced is the following: I want to know when the DOS mode program quits, to close the "access handle" the DOS program obtained from me, that is, to do some cleanup if the DOS program forgot to do this itself (for example, if it unexpectedly crashed). Anyway, I'm not sure how to do this reliably. Currently, I use VDDInstallUserHook() to install a Block handler, that is, I am informed if the NTVDM is blocked. I found out that this happens if the DOS program quits. Anyway, is this really a reliable (and a recommended) way? First of all, I thought installing a create and a termination handler with the same function would be better, as I could find out if it is exactly the DOS program which obtained the handler which is quit. Anyway, as the VDD is started with RegisterModule (0xc4, 0xc4, 0x58, 0x00) from the DOS program, I do not "see" the creation of the DOS program, so I do not know for which PCB to wait for. Can anyone help me here? Thank you, Spiro. -- Spiro R. Trikaliotis http://www.trikaliotis.net/ Tag: bswap intrinsic? Tag: 57383
  - 5
    - VPC shot me in the ankle My development environment is crippled right now and I'm forced to use VPC 2004 instead of a target computer. I've run into a problem that has caused me a few hours of trouble and I've gotten down to the issue finally. It seems that VPC is not quite as faithful as I had hoped, or maybe there is a race in my code that simply did not show up on hardware: In the File system filter example, at the end of DriverEntry after the DO_DEVICE_INITIALIZING flag is cleared, I had assumed that it was now ok to access the file system. In my code, based on this example, after clearing the flag I created/accessed a particular directory to do some work. In the hardware version, this always worked fine, in the VPC however I reach the point of clearing DO_DEVICE_INITIALIZING long before I've seen the system drives set up and consequently my fiddling with the directory is failing. So is this a problem with VPC, or more likely have I misunderstood the significance of DO_DEVICE_INITIALIZING. If the latter, how do I know when all the file systems have been set up and are now safe to access? -- Gak - Finecats Tag: bswap intrinsic? Tag: 57374
  - 6
    - Need help in Print Driver Plug-in for Windows NT4 Hi, I'm trying to insert username, hostname, document name, printer name, driver name, custom and priority into the print job stream from my renderer plug-in dll. I realize that there are APIs available for me to gather these info if my plug-in is in user mode. That's why I get it to work in W2K. However, the APIs are not available in kernel mode, at which my renderer plug-in is invoked. I can't find any equivalent APIs supported in Kernel mode from the DDK library. I try to collect the info from the UI plug-in(which is in user mode) and pass the info to the renderer plug-in. However, if the user does not open the printer properties page, the UI plug-in will not be invoked. Are there other ways to get the info from kernel mode? Is it possible to invoke the UI plug-in from the renderer plug-in and let the plug-ins communicate? Is language monitor a possible route to explore? Suggestions are very much appreciated. Thanks, Anthony Tag: bswap intrinsic? Tag: 57372
  - 7
    - ddk where to start? i put this on earlier but not as my official msdn login and didn't get anywhere, time is getting shorter, and i am getting more confused by the day going through ddk, sdk, directx, and other mm interfaces.... so now i try again the official way: ok, i must admit i am a bit out of my normal realm here with drivers under windows... now if you wanted a dos tsr thats a different story. in any event i got the xp ddk installed and built the samples, and have been wading through some of the documentation down the audio driver path and think i want to end up in the msvad area somewhere. but i really do want some confirmation that i am in the right area and that this is really the best way to do what i need to do. Problem: take left and right channels of one or more line in or mic inputs on one or more sound cards and switch them to the left and/or right channels of one card's speaker/headphone output. this has to be done in real time and the routing has to be controllable from a user program with some kind of a control api. i can do this now with waveIn/waveOut api, but there is an annoying delay due to the filling and passing of buffers, and it seems touchy as far as the timing for setting up buffers and getting them in and out on time. so what i think i need is something like a super mux combined with a splitter that can route left and right channels and duplicate them so that i could listen to combinations like the left channel line in on card 1 in both ears, then the left channel line in on card 1 in the right ear and the left channel line in on card 2 in the right ear, or the left channel of line in on card 2 in the left ear and the right channel line in on card 2 in the right ear... etc. so, any pointers about where i should be heading? i don't want to tie it to any particular hardware which is why i am thinking the virtual audio device stuff is the way to go. but is this the only way? would directsound give the speed and control or does that work on filling and sending buffers also? with the virtual audio device how would i configure it so it gets into the audio stream the right way? too many questions, not enough reading yet i think. after more reading what i think i need is a realtime driver that impliments a ksnodetype_mux or ksnodetype_supermix. does that sound like the right route? Tag: bswap intrinsic? Tag: 57371
  - 8
    - trap joystick collection hi, I have a HID GamePad/Joystick device. I wish to have a full control on the joystick : Weather the button pressed information is passing to DInput, will be my options. How can I filter a joystick collection, so I can read buttons from a specified gamepad, and (maybe) drop it without passing to DInput? Is HIDClass upper filter ok? Will I have to write a minidriver for the full control on the gamepad? thanks. Tag: bswap intrinsic? Tag: 57370
  - 9
    - Not getting IRP_MN_STOP_DEVICE Hello all, I have a device connected over TCP. The device has a firmware running on it. I have already written a bus driver to enumerate the device and load the required function driver for it. The function driver provides a number of APIs to the application to deal with the device. One of the API is to reset the device. When this command is issued, the driver has to carry out same steps implemented in the START_DEVICE to initialize board and firmware. So, I want to stop the driver when reset command is issued and start it again. I have called IoInvalidateDeviceState(), set the flags (PNP_DEVICE_FAILED | PNP_DEVICE_RESOURCE_REQUIREMENTS_CHANGED) and I am getting IRP_MN_QUERY_STOP_DEVICE. The PnP Manager then asks for RESOURCE_REQUIREMENTS and FILTER_RESOURCE_REQUIREMETS. Here I am returning NULL since I am not using any resources. This is the same value returned by me when enumerating the device. Then I am getting IRP_MN_CANCEL_STOP_DEVICE. Why? Since there is no change in the resource requirement, PnP Manager is cancelling it? I also made sure that no application handles are opened to the driver and no driver in the stack is cancelling QUERY_STOP_DEVICE. Waiting for the replys, Aniketa K S Tag: bswap intrinsic? Tag: 57356
  - 10
    - capturing all audio out reliably : DirectSound or MSVAD? Hi, 1) I want to capture all the audio out on my machine . What i understand is that if the sound card on my machine provides "stereo Mix" capability AND I am not recording something on the physical line in of my sound card , then and only then i can capture the audio out going through the sound card using DirectSound or WaveIn API. what is the most reliable way to do this that should work on Win 9x+ and Win 2k+ , independent of whether the sound card supports stereo mix or whether something is also being recorded in the physical line in of the sound card? Is there some way to generically install an upper filter driver on top of the sound card driver and do this? what other option should help here? In the DDK i have a Microsoft Virtual Audio Driver , will that be of use ? There are multiple drivers inside the MSVAD driver directory, which one can i use with minimum amount of code change that should achieve my purpose ? and if i do need to change the code , which portion needs to be changed ? 2) I was going through the code of MSVAD and its architecture looks different from the standard WDM drivers ( particularly WDM USB client drivers ). No Read/Write Dispatchers etc. also Walter Oney's Book apparently does not delve into sound card drivers as well. What is the best resource to get more information about this , apart from the DDK ? thanks and regards Taha Tag: bswap intrinsic? Tag: 57354
  - 11
    - DeviceIOControl Hangs I make a DeviceIOControl call to my driver from within an application. This is an overlapped IO call and has worked successfully for years, however with one customer it fails after a few days under heavy load. This is an asynchronous call and it doesn't appear to even reach the driver and just hangs on the call itself. Has anyone seen or heard of this sort of error occurring before? Tag: bswap intrinsic? Tag: 57352
  - 12
    - How to Enable/Disable USB Device Remote Wakeup? Iâ??m writing a WIA driver for a USB scanner that supports remote wakeup. I have not been able to find any way through Windows to enable and disable the scannerâ??s remote wakeup feature from user mode. Is there a way? I normally deal with the scanner through USBSCAN.SYS and the DeviceIoControl() function. But using this function there seems to be no way to send a Standard Device Request type, which is the type you need to send to do a SET_FEATURE request to enable the remote wakeup. (According to the documentation, the IOCTL_SEND_USB_REQUEST control code does not let you set the bmRequestType). I notice that you can issue a SET_FEATURE request from a kernel mode driver. There is even a macro UsbBuildFeatureRequest. But please donâ??t tell me Iâ??ll need to write my own kernel mode driver merely to enable/disable the remote wakeup. Iâ??d hate to have to go to such overkill just for a simple thing. You would think there is an easier way. The USB-IF spec requires that if your device reports that it supports remote wakeup, then you must have a way to enable and disable it. (And yes, they enforce this because theyâ??ve snagged me for it already.) Sorry if this is the wrong forum for this question. If there is a more appropriate one, please let me know. Tag: bswap intrinsic? Tag: 57334
  - 13
    - GetPrinterDataFromPort() in language monitor From what I read in the MSDN, when we call GetPrinterData(), it should call GetPrinterDataFromPort() in the language monitor. I added some text to be printed in my language monitor when the function is called and although the function worked correctly, it didn't print any messages. It did print messages when the functions WritePort and StartDocPort were called (by using WritePrinter and StartDocPrinter). Any suggestions? -- John Tag: bswap intrinsic? Tag: 57329
  - 14
    - irp cloning... Hello, I am currently porting a pseudo disk driver to NT from Unix of which one aspect of its operation is to perform a mirrored write. The manner in which it performs the mirrored write is to clone the original IRP for each mirror and then issue the cloned IRPs to the primary and secondary devices. During the mirrored write the original IRP is stalled. After the cloned requests complete, the original IRP is eventually completed. If I only clone the IRP once and issue the clone to the primary device the driver seems to work well (as evidenced by a data integrity tool I'm using). However, if I instruct the driver to perform mirroring, the data sent to the primary device is always corrupted (I am currently unable to verify the secondary as it is not written in the original format). Here's is the gist of my IRP clone operation. After calling the cloning macro a cloned IRP is further modiefied to point to the correct device by directly setting the DeviceObject field in the IRP. Also, the IRPs are hand made and set up by IoInitializeIrp() rather than by IoAllocateIrp(). It's probably worth admitting that while I have programmed drivers for Unix for many years I've only been working with NT for a month or so. I would greatly appreciate any insight into how to solve this problem correctly. I have scoured this list and the web for similar examples but haven't found anything relevant thus far. Thanks! Greg #define CLONE_IRP(src, dst, zoffset) \ do { \ IO_STACK_LOCATION *srcstk, *dststk; \ srcstk = IoGetCurrentIrpStackLocation(src); \ dststk = IoGetCurrentIrpStackLocation(dst); \ *dststk = *srcstk; \ dststk->Context = (void *)0; \ (dst)->MdlAddress = IoAllocateMdl( \ MmGetMdlVirtualAddress((src)->MdlAddress), \ MmGetMdlByteCount((src)->MdlAddress), \ FALSE, FALSE, (IRP *)0); \ IoBuildPartialMdl((src)->MdlAddress, (dst)->MdlAddress, \ (char *)MmGetMdlVirtualAddress((src)->MdlAddress) + \ (zoffset), (unsigned long)0); \ (dst)->Tail.Overlay.Thread = (src)->Tail.Overlay.Thread; \ } while (0) Tag: bswap intrinsic? Tag: 57328
  - 15
    - ReadPrinter fails for TCP/IP port, works for LPT1 and USB using pjlmon XP DDK sample Hi all, I have this problem which has been eluding me for weeks. The sample pjlmon project that comes with the XP DDK works perfectly for printers that are connected via LPT and USB, but fails when trying to read from a printer that's connected via TCP/IP. The printer port we're using is a Standard TCP/IP port, over port 9100 which creates a port named IP_10.32.56.23 which is fine. Here's the kicker - when I use WritePrinter to send the data down, WritePrinter works and returns back a successful value. Using the Ethereal packet sniffer, I find that we DO get the right data back from the printer, but for some reason it doesn't "bubble up" from the port to the language monitor, and ReadPrinter returns with a 0 meaning that no data was read back, and the buffer's empty. I don't know where the data's getting blocked. Is the port monitor blocking it from the language monitor? I've tried this with both the outdated versions of pjlmon and the newer updated one (InitializePrintMonitor and InitializePrintMonitor2). Doing some digging, I found that the port type you're using uses a different DLL for filling in the MONITOR/MONITOR2 structure. Using a debugger, I found that for TCP/IP, GetPrinterDataFromPort is NULL for TCP printers, and valid for LPT and USB. It seems to me that the tcpmon.dll DLL assumes that backchannel data will never come from a TCP/IP printer. If this is true, I'm going to have to work out my own solution since I can't depend on that DLL for reading data back. In fact, I already did that, but it bypasses ReadPrinter which I really don't feel comfortable doing. According to the MSDN documentation, using SendRecvBidiDataFromPort only works on Windows XP anyway, so that solution will not be an option since we have to support 9x->XP. So, what I want to know, if anyone can answer, are the answers to these questions: 1) Why do we not get backchannel data back from a printer with pjlmon (both old and new) using TCP/IP, but works with LPT and USB? 2) Will bidi communication work on server-based system such as CUPS and Samba? HP's bidi software wouldn't work with printers connected to these types of systems, but did work with direct TCP/IP connections. Thanks in advance. Mike Tag: bswap intrinsic? Tag: 57304
  - 16
    - FIREWIRE CDR as SCSI device Hello, ist it possible to trat a firewire devive (e.g. a cdr) as a scsi device ? I have written an application which communicates with atapi and scsi cd/dvd drives using SCSI commands. Now i also want to do it with firewire drives. Can anybody help me ? Regards, Thomas Tag: bswap intrinsic? Tag: 57303
  - 17
    - Language Monitor question Hi, I try to make my own Language Monitor Dll in order to enable bi-directionnel communication with my thermal Printer. I work with Windows XP SP2 I read that I write the following functions : ClosePort EndDocPort GetPrinterDataFromPort InitializePrintMonitor2 OpenPortEx ReadPort StartDocPort WritePort. then I set to NULL the other pointer function of the MONITOR2 structure. I test it and my dll is loaded and InitializePrintMonitor2 is called but at the end of the installation of my printer mini-driver, I receive an error message : "The specified print monitor is unknow" What is wrong ? Do I write the other functions of MONITOR2 structure ? Thanks in advance Tag: bswap intrinsic? Tag: 57302
  - 18
    - How to get rid of those "delightful" "found new hardware" dialogs. Hi folks, Been working on my installation program recently, and almost all is well apart from one *really annoying* feature. The main device I have acts as a bus driver, and creates several (about 10) child devices. Initially, when the hardware is added, and the machine booted, a "found new hardware" dialog pop up (no surprise there). It's then possible to cancel this dialog, and at a later convenient moment, run my install app. The install app installs the driver for the main device, which starts up, and then creates all the child devices. What then happens is that a "found new hardware" dialog pops up for the first child device, asking if I'd like to search for drivers online. This is very irritating, as the install app is in the process of installing the drivers for the child devices. I've tried to get rid of it, but there's a catch-22: If I install the drivers for the child devices *after* the main device, then the dialog pops up, & if you leave it alone, it subsequently goes away as the install app installs the child drivers. If I install the drivers for the child devices *beforehand* the annoying dialog *still* pops up. What's even more annoying is that it's possible for the user to then respond to these popped up dialogs, at which point, my install app and the user mode PnP manager seem to have a bit of a race to see who can "install" something appropriate first. NOT what I want! This seems to happen just under XP service pack 2 - earlier OS's (2k for example) would simply allow the install app to complete, and the next time time PnP dev tree was rescanned (either next reboot, or explicitly by the install app), they'd just raise a couple of dialogs indicating things were being installed, and hey-presto, it all worked! So basically: - The install looks unprofessional. - The dialogs raised let the user faff with what should be a smooth install process. This is one situation where I feel that the security considerations of SP2 seem to have made things distinctly worse (if theoretically more secure) from the users P.O.V. How do I sort this? MH. Tag: bswap intrinsic? Tag: 57301
  - 19
    - Serial port locked I wrote a laguage monitor for our printers. When I install the printer on a serial port, it seems to lock the port so I can't use our VB or C applications to talk to our printer. Is there a way to unlock it? -- John Tag: bswap intrinsic? Tag: 57300
  - 20
    - NDIS Vs Protocol Who will drop the packets? Hello, Assume a PC with a NIC.NIC is connected to a shared network cable. Each NIC has a MAC Address.As cable is shared one it will also recieve packets destined to other PCs i.e packets with different MAC Address other than its own. So the PC will drop the packets which are not detined for it. I want to know whether NDIS driver will drop these packets or protocols above the NIC will drop these packets. Thanking you, Monica DSouza, FZI-Karlsruhe Tag: bswap intrinsic? Tag: 57295
  - 21
    - Promiscious mode with NIC driver Hello all, I am Monica from germany.We have task which is quite uncommon. So please read it slowly. We are developing some custom based PCI Ethernet Card using Realtek chips.we have downloaded the driver software(binary)from the REALTEK site. we have to use the card always in Promiscious mode.But we dont have the source code of the driver.we only have the binary code. But we have observed that softwares like ethereal are able to capture the ethernet packets in promisicous mode. So my query is,Is there any standard IOCTL function or some facility in the Windows operating systems to make the ethernet card in promiscious mode. Thanks alot in advance. Thanking you, Monica DSouza, FZI-Karlsruhe. Tag: bswap intrinsic? Tag: 57284
  - 22
    - encrypting hiberfil.sys Hello gurus, I have developed an encrypting filter driver and some int13 boot code. That combination runs so far. One of the few problems is that lacking support for hibernation. Hiberfil.sys gets written in plaintext, thats the problem. I catch SCSI_GET_DUMP_POINTERS and fill DUMP_POINTERS struct with my device pointer, then it gets called in DriverEntry. How can i continue from there ? Are there any good exampes ? Please give some answers . Tag: bswap intrinsic? Tag: 57283
  - 23
    - encrypt hibefil.sys via filter driver ? Hello gurus, I have developed an encrypting filter driver and some int13 boot code. That combination runs so far. One of the few problems is that lacking support for hibernation. Hiberfil.sys gets written in plaintext, thats the problem. I catch SCSI_GET_DUMP_POINTERS and fill DUMP_POINTERS struct with my device pointer, then it gets called in DriverEntry. How can i continue from there ? Are there any good exampes ? Please give some answers . Tag: bswap intrinsic? Tag: 57281
  - 24
    - Avstream Video capture -> AVI Mux -> File Writer Hello All, I have an AVStream video capture driver based on avshws sample code. I would try saving avi file using this driver,GraphEdit.exe and filters as follows. (1) Smart Tee Filter (2) Capture pin (Select RGB24 format) ->Video Renderer (3) Preview pin(Select RGB24 format) -> AVI Mux -> File writer I could get avi file,but the file contained only one flame image. Is there any particular reson why I could not get correct avi file. I want to know how to modify the driver source files to be able to get correct avi file. Steve Tag: bswap intrinsic? Tag: 57277
  - 25
    - Can i cancel Control IN/OUT command on Control pipe?? Sir. I have to use Control pipe to transfer some data. May i cancel Control command with IoCancelIrp durning the transmission? and if so ,what kind of problem may happened ? poly Tag: bswap intrinsic? Tag: 57274
- Next
  - 1
    - PCI Express extended Configuration Space access Hi All, The document from Microsoft "PCI Express and Windows" says: "PCI Express extended Configuration Space is not accessible on current Windows operating systems. Device drivers or the operating system must be updated to use the extended region." Could anyone please explain what updations need to be done? Can we access PCI Express extended Configuration Space by doing above discussed modification in existing PCI driver on windows. From where, i can get more information on this. Regards, Sachin sachin.b@soc-soft.com Tag: bswap intrinsic? Tag: 57265
  - 2
    - How does USBCAMD support still picture? I develop pc camera driver based on USBCAMD. Now I would support still picture feature. DDK document doesn't mention too much about still picture support in USBCAMD. In addition to expose a virtual still pin, I have no idea about how to support still picture in USBCAMD. Is it necessary to provide interrupe endpoint for snapshot hardware button in USBCAMD driver? Does I need to develop other component to support still picture? Best Regards Jackal Huang Tag: bswap intrinsic? Tag: 57263
  - 3
    - Help. How do you handle banding in a Unidrv plugin? Hi, I am having trouble getting my Unidrv plugin to handle banded bitmaps correctly. If I set the print to 100 DPI then my ImageProcessing gets 1 whole bitmap per page. However, if I set it to 300 DPI then I first get what appears to be the top 1/4 of the image. After that I get a few calls to ImageProcessing indicating blank bands and then that's it. The rest of the bitmap never appears. No matter how many pages are printed I just get a 1 partial bitmap, a few blanks and then that's it. Is there some argument that needs to be set to request the next band get sent? Also, how do I know when the last band for an image is received? Thanks, Bob Herzberg Tag: bswap intrinsic? Tag: 57257
  - 4
    - Need to write custom installer for 3rd party driver Is there a quick and easy way (using visual studio or other Win32 development environment) to install a third party driver that contains the proper .inf files, etc. We have so-so clients that have problems using the install new hardware wizard, and since we already have the driver files, I was looking for something that I can call via Win32 or shell api... Like: InstallDriverUsingThisINFFile( char *driverInfFilePath ); I'm a long time windows developer but not for drivers.... I thought I could bundle the driver files with the app and auto-install at startup if the driver is not found. Any ideas? Tag: bswap intrinsic? Tag: 57248
  - 5
    - Inconsistent URB status after IoCancelIrp between W2K and XP I have a USB function driver that implements a synchronous IRP and cancels the IPR after a fixed timeout as per Walter Oney's design pattern. It is working fine in W2K and XP but the problem is the urb->Urbheader.Status is diferrent between the two OSes. On W2K the urb status == 0x00010000 and on XP the urb status == 0xC0010000. Note that the error and stall bits are set on the XP urb status code. Why is this different? The W2K code equates to USBD_STATUS_CANCELED which is documented in the DDK headers but the XP code is not. What does the XP code signify? Thanks, SteveB, MCSD Welch Allyn, Inc. Tag: bswap intrinsic? Tag: 57247
  - 6
    - SetupDiEnumDeviceInterfaces not finding my device I have a device that supports an interface GUID_INTERFACE_XYZ. >From a Win32 app, if I call { HDEVINFO info = SetupDiGetClassDevs(NULL, NULL, NULL, DIGCF_ALLCLASSES); if (info == INVALID_HANDLE_VALUE) return; // if I use SetupDiEnumDeviceInfo(info, idx, &devdata) here to dump the ClassGuid and DevInst for all the devices in info // then I find my device at idx=0x22 DevInst=5016 and with the device class guid matching, as expected, the xxx... value in // HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{xxx...} SP_INTERFACE_DEVICE_DATA ifdata; ifdata.cbSize = sizeof(ifdata); DWORD devindex; SP_DEVINFO_DATA devdata = {sizeof(SP_DEVINFO_DATA)}; for (devindex = 0; SetupDiEnumDeviceInterfaces(info, NULL, &GUID_INTERFACE_XYZ, devindex, &ifdata); ++devindex) { // for each device DWORD needed; SetupDiGetDeviceInterfaceDetail(info, &ifdata, NULL, 0, &needed, NULL); PSP_INTERFACE_DEVICE_DETAIL_DATA detail = (PSP_INTERFACE_DEVICE_DETAIL_DATA) malloc(needed); if (!detail) continue; detail->cbSize = sizeof(SP_INTERFACE_DEVICE_DETAIL_DATA); if (!SetupDiGetDeviceInterfaceDetail(info, &ifdata, detail, needed, NULL, &devdata)) { // can't get detail info free((PVOID) detail); continue; } /*fi - get detail info? */ // get the device's friendly name if possible - this will only be set if it is not the first instance of the device // for the first instance we can simply use the device description instead BOOL bDeviceDesc = FALSE; TCHAR buffer[_MAX_PATH]; if (!(SetupDiGetDeviceRegistryProperty(info, &devdata, SPDRP_FRIENDLYNAME, NULL, (PBYTE) buffer, sizeof(buffer), NULL))) { bDeviceDesc = TRUE; SetupDiGetDeviceRegistryProperty(info, &devdata, SPDRP_DEVICEDESC, NULL, (PBYTE) buffer, sizeof(buffer), NULL); } /*fi*/ printf("%sName=%s\n", bDeviceDesc ? "Device" : "Friendly", buffer); free((PVOID) detail); } /*rof - each device */ SetupDiDestroyDeviceInfoList(info); } then the first call to SetupDiEnumDeviceInterfaces returns 0 and SetupDiGetDeviceInterfaceDetail never gets called. If, however, I use HDEVINFO info = SetupDiGetClassDevs(&GUID_INTERFACE_XYZ, NULL, NULL, DIGCF_INTERFACEDEVICE); if (info == INVALID_HANDLE_VALUE) return; the my device is found by the first call to SetupDiEnumDeviceInterfaces returns 0 and SetupDiGetDeviceInterfaceDetail gets called as I would have expected. Also, // if, in this case, I use SetupDiEnumDeviceInfo(info, idx, &devdata) here to dump the ClassGuid and DevInst for all the devices in info // then I find my device at idx=0x0 and again DevInst=5016 and with the device class guid matching, as expected, the xxx... value in // HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{xxx...} Any ideas why the device is not being found in the first case? Will Tag: bswap intrinsic? Tag: 57245
  - 7
    - USB URB Submission Times Hi, I've written a driver for a high speed USB device, everything works well, but I've noticed one major area where performance is poor. When passing an IRP down the stack, indicating that some data should be sent to the USB device using a bulk endpoint, the call to IoCallDriver takes around 50 microseconds to complete which seems like a very long time to me. The USB device isn't NACKing or delaying the transfer in any way, it appears that the USB driver just requires that amount of time to process the URB. The URB in my tests points to a buffer to be transferred which is 8 bytes in length. This has been tested on a couple of fairly high performance PCs (>3GHz, 2GB memory) with similar results. Both these PCs are running Windows XP (one with SP2, one without). I've included some of the code I'm using below (pretty standard stuff, I'd imagine). The question is, can I improve my code in some way to improve performance, or am I simply expecting too much from Windows? Thanks in advance for any help. Hereâ??s the code: irp = IoAllocateIrp(pdx->StackDeviceObject->StackSize, FALSE); nextStack = IoGetNextIrpStackLocation(irp); ASSERT(nextStack != NULL); nextStack->Parameters.DeviceIoControl.IoControlCode = IOCTL_INTERNAL_USB_SUBMIT_URB; nextStack->MajorFunction = IRP_MJ_INTERNAL_DEVICE_CONTROL; // Set up the URB ptr to pass to the USB driver stack nextStack->Parameters.Others.Argument1 = pdx->WriteUrbs[pdx->urbnum]; // Set the completion routine IoSetCompletionRoutine(irp, completionRoutine, context, TRUE, TRUE, TRUE); // Call the USB class driver to perform the operation // This call takes around 50 usecs ntStatus = IoCallDriver(pdx->StackDeviceObject, irp); The URB is built with a call to UsbBuildInterruptOrBulkTransferRequest passing in an MDL rather than a buffer (as this makes a slight improvement in performance). Tag: bswap intrinsic? Tag: 57244
  - 8
    - Frame stepping & custom KS event sets? Hi folks, I have a quick couple of queries about frame stepping & KS filters. As I understand it, frame stepping seems to work like this: 1.. FGM changes renderer state to pause. 2.. Checks filter can step, and requests step for X frames from filter. 3.. FGM changes state to play. 4.. Filter should step through the required # of frames at the normal playback rate. 5.. Once filter has done that # of frames, it should send an EC_STEP_COMPLETE to FGM. 6.. FGM gets step complete notification, and then puts the graph back in the paused state. Which means that my renderer has to implement the frame stepping property set, which it now does (many thanks to the guys on the dev driver group - got my GUID's sorted!). Strangely, the filter graph manager seems to invoke the frame step property set on the *filter*, not the pin - but that does not concern me unduly at the moment. However, the mechanism above means that my renderer also has to send the EC_STEP_COMPLETE notification. Perusal of the existing code indicates that (for example) the render sends an EC_COMPLETE notification by implementing the KSEVENTSETID_Connection event set on a stream. The problem is that I can't seem to find a suitable event set that would let me send the frame step complete notification. Perusal of newsgroups / MSDN seems to indicate that one *could* implement a custom event set, but it's unclear to me at the moment if I could tie this in to the notification that the filter graph manager expects. Any ideas anyone? MH. Tag: bswap intrinsic? Tag: 57236
  - 9
    - Can a non-Administrator install a display mirror driver? All: I have a fairly simple question. Can a regular user (without Administrator privileges) install a display mirror driver on Windows 2000 and XP? Thanks, -Bernd Tag: bswap intrinsic? Tag: 57227
  - 10
    - NDIS IM: queueing question Hello, First of all, many thanks to Pavel, Maxim, Steve and Stephan for clarifying my previous questions. As you probably remember, I am trying to limit bandwidth using NDIS IM driver. I managed to successfuly queue packets in MPSendPackets function (many thanks to Stephan!), and send them from worker thread. I do not drop any packets. All packes received by MPSendPackets are sent later from worker thread. 1) The problem is that incoming connection is dropped as soon as it was established. Why? 2) Another problem is that outgoing connection is dropped in about 1 minute as well. I can see how connection was extablished, than a lot of TCP packets with ACK bit set, and suddenly remote server stops sending any packets... Do I need to manually set PUSH bit as well? I doubt it will help, though. 3) Do I need to queue packets in PtReceive() as well? I create a copy of original packet (including NDIS_BUFFER), do not call NdisMIndicateReceivePacket(), queue my packet, and return NDIS_STATUS_RESOURCES. Later, in worker thread I call NdisMIndicateReceivePacket() but it causes D2 bugcheck... Currently I simply return NDIS_STATUS_RESOURCES and do not indicate packet if bandwidth is too high. Is that correct? Do I need to queue packet as well? The problem persists if I indicate packets immediately, though. Any advices are greatly appreciated. Best regards, Serge. Tag: bswap intrinsic? Tag: 57223
  - 11
    - OID_802_11_BSSID_LIST_SCAN Hello, I am a Win32 developer. I know nothing about driver development. Now I want to issue scan command to NDIS driver by setting OID_802_11_BSSID_LIST_SCAN. But I realised OIDs cannot be set from user-mode applications (but only from kernel mode). So, I understood that I have to write a protocol driver for that. I installed DDK and I looked at "ndisuio" example and changed all that are to be changed like making changes in INF file, and changing the name of the driver, etc., I could install it as a protocol. Now my problem is that I dont know if my driver is really in use. Can someone suggest me how to proceed from here please? Thanks, John. Tag: bswap intrinsic? Tag: 57220
  - 12
    - Memory Leak -NDIS can anyone let me know whether there are tools available to find out memory leak in NDIS driver? Thanks, Senthil Tag: bswap intrinsic? Tag: 57218
  - 13
    - Min Time Resolution on Windows Hi All, We are facing problem in calculating time stamps at driver level. We have min. of 15milli seconds resolution normally. By using keQueryPerformanceCounter ( along with ExSetTimerResolution which changes frequency at which clock interrupts )we can get resolution of min. 1millisec. But by using above two functions( which are not recommended to use frequently ) we observed substantial delay in our dispatch routine calls. Is there any efficient way to get Time Resolution in atleast 100 micro seconds. Could you throw some light on above issues. Thanks in advance. Chiranjeevi. Tag: bswap intrinsic? Tag: 57217
  - 14
    - WDF? Should I familiar with WDM now? hi, Since MS say that the WDF is the Future, Should I familiar with WDM now? (I am little doubt weather I am wasting time to deep into WDM.) Does the WDF meant for .NET paltform OS (namly LongHorn) only? Tag: bswap intrinsic? Tag: 57216
  - 15
    - an urgent issue aoubt renaming ndisuio.sys now my work is write a program to set ssid and set wep in windows xp, now I can finish this work by using the ndisuio sample driver in DDK, i have renamed it,but i have met a issue, after i change the protoname, DeviceIoControl always fail,and error code is can't find the file specified,if i no change the protoname, it run well, but It exists that the conflict between my driver with the system ndisuio driver.thanks for any help. Tag: bswap intrinsic? Tag: 57215
  - 16
    - How to cancel IRP passed down to USBD Dear developers, I want to cancel pending read/write operation on a bulk usb pipe. Can I just use IoCancelIRP() in my usb function driver to do it? Regards, Ari Tag: bswap intrinsic? Tag: 57214
  - 17
    - HBA in SCSI driver stopped before S3 does not come back (restart) after resume. I develop a SCSI miniport driver, but the HBA cannot restart after resume. If no device connect this adapter, windows will stop the HBA and remove it before standby. When the system go back to working state, the whole device driver cannot work. According to the document of SCSI miniport, windows will detect the device and restart HBA. But I do not received any request and command. If I connect device to the SCSI adapter, the interrupt do not send to my driver. Tag: bswap intrinsic? Tag: 57205
  - 18
    - ddk, where to start? ok, i must admit i am a bit out of my normal realm here with drivers under windows... now if you wanted a dos tsr thats a different story. in any event i got the xp ddk installed and built the samples, and have been wading through some of the documentation down the audio driver path and think i want to end up in the msvad area somewhere. but i really do want some confirmation that i am in the right area and that this is really the best way to do what i need to do. Problem: take left and right channels of one or more line in or mic inputs on one or more sound cards and switch them to the left and/or right channels of one card's speaker/headphone output. this has to be done in real time and the routing has to be controllable from a user program with some kind of a control api. i can do this now with waveIn/waveOut api, but there is an annoying delay due to the filling and passing of buffers, and it seems touchy as far as the timing for setting up buffers and getting them in and out on time. so what i think i need is something like a super mux combined with a splitter that can route left and right channels and duplicate them so that i could listen to combinations like the left channel line in on card 1 in both ears, then the left channel line in on card 1 in the right ear and the left channel line in on card 2 in the right ear, or the left channel of line in on card 2 in the left ear and the right channel line in on card 2 in the right ear... etc. so, any pointers about where i should be heading? i don't want to tie it to any particular hardware which is why i am thinking the virtual audio device stuff is the way to go. but is this the only way? would directsound give the speed and control or does that work on filling and sending buffers also? with the virtual audio device how would i configure it so it gets into the audio stream the right way? too many questions, not enough reading yet i think. Tag: bswap intrinsic? Tag: 57201
  - 19
    - PCI DMA Driver...Why does I/O Manager limit transfer sizes I am working on a PCI based DMA driver which is working. However the driver/user app. are only capable of transfering 63Mb at a time. When ReadFile() is called with a byte_count larger than about 63M, ReadFile() returns (Insufficient Resources Exist To Complete The Requested Service). Being new to kernel programming I am confused as to what the problem is. I do know the following: 1) The system has more than enough physical memory. 2) The driver never recieves an IRP (with transfer sizes > ~63Mb) 3) Driver uses DO_DIRECT_IO (...DMA with MDL) Could it be that the I/O Manager won't lock that much memory? Any thoughts would be greatly appreciated...even if they are regarding a work around. -Brendan Tag: bswap intrinsic? Tag: 57200
  - 20
    - Multiple DeviceObjects Hello: I have user-mode appl that through monolitic kernel-driver communicate with sensors connected to data register of parallel port. I'd like to have separate DeviceObject for each sensor. Is it right to do it by calling IoCreateDeviceObject() from DriverEntry() more then once ? thank you, dave P.S. I don't have time to accomplish it by creating bus driver. Tag: bswap intrinsic? Tag: 57198
  - 21
    - STI and Windows ME Hi, I am writing an application to go with a USB scanner (with a twain driver) that has a button on the front panel. The application responds to all button presses on the scanner (there are 9 different events) using the STI event mechanism. I have the application working correctly on Windows 98, 2000, and XP but ME does not trigger the STI event the way the other OSes do. Can anyone point me to any documentation on what the differences are between ME and all the other operating systems (with respect to STI and the registry settings)? Thanks, JM Tag: bswap intrinsic? Tag: 57197
  - 22
    - Setup Failing with Printer Driver for AMD64 Hello, Is there any way to get additional info from SETUPAPI.DLL about which "operation could not be completed"? Is there a detailed reference as to exactly what SETUPAPI.DLL is doing when installing a printer driver? I have built our (Konica Minolta) user-mode rendering driver for the AMD64, and am trying to install it into Windows XP 64-bit Edition, Version 2003, SP1, v.1069. (Because I don't have the actual printer connected to a COM or LPT port, I created a new port... how does that affect the installation?) I found and followed instructions from MS about manually removing spool/driver files and registry entries from safe mode, verified that policy settings allow unsigned drivers to be installed, ensured the .INF is decorated properly and that the AMD64 versions of the driver are being copied. SETUPAPI.LOG is complaining about unsigned drivers, of course, and plenty of times I have goofed the .INF so that Error 87 "The parameter is incorrect." is reported. But when that situation is corrected, files get copied into system32\spool\drivers\amd64\konicaminoltaXXXX , and \spool\drivers\amd64\3, but then this "operation could not be completed" message pops up, and nothing informative is in SETUPAPI.LOG. The last line is the same as all the other lines, "Error 0xe000022f: The third-party INF does not contain digital signature information." Of course, the 32-bit version has been working fine, so the problem seems to be something that has changed either with the AMD64 build of the driver, or something expected by Windows XP that I am unaware of, or maybe even that this version of Windows needs to be updated (from Windows Driver Development Conference 2003, the CD is labeled "Microsoft Windows XP 64-Bit Edition SP1 - Beta 1 for 64-bit Extended Systems, with the AMD logo on it). Thanks in advance, Bart. Tag: bswap intrinsic? Tag: 57192
  - 23
    - Problem setting WinCE RAS server user credential directly from "hashed" password Hi, We need to set the RAS server user credential information not from password in clear, but from the *hashed password* directly. The premise of our problem is that the password in clear is now known, only its hased value. Therefore, the typical approach: RASCNTL_SERVERUSERCREDENTIALS cred; (...) RasIOControl(NULL, RASCNTL_SERVER_USER_SET_CREDENTIALS, (PBYTE)&cred, sizeof(cred), NULL, 0, NULL); ... cannot be used, since the RASCNTL_SERVERUSERCREDENTIALS structure specifies password in clear. The foreseen solutions is to find the registry keys where the user credential information is stored, identify the hash algorithm used by Windows CE for RAS server user credential, and store directly the hash code. But Microsoft specifies that direct registry access to user credential information is not supported.. Any better idea? Your ideas are more than welcome. Thanks in advance. Yvan. Tag: bswap intrinsic? Tag: 57186
  - 24
    - Installing a Port monitor on a Windows 2000 Cluster I have written a cluster aware Port monitor that works fine normally but I cannot use it in cluster mode. I installed the monitor using the AddMonitor() function. This registers the monitor in both the local and cluster areas of the registry fine. I am setting the blocal parameter of PMONITORINIT to FALSE to show that the monitor is being used be a cluster spooler. I am also using the registry functions specified by MONITORREG. What I am finding happening is when I come to configfure a port using my monitor all of the information is stored in the local area of the registry and not in the cluster node. Hence when in cluster mode I cannot use my port. Is there anything else I need to be setting to force the port monitor to write to the cluster registry? Cheers for any help as this is driving me mad. Tag: bswap intrinsic? Tag: 57185
  - 25
    - multiport serial adapter and COM ports I am writing a driver for multiport serial adapter for WindowsXP. The serial ports are not any conventional asysnchronous serial ports, they are synchronous serial port. So, if any serial port is connected to our synchronous serial port then they will not work. My question is that if assign the bus driver to multiport serial adapter will there be any violation for driver signing? My second question is if I expose those ports's symbolic link as COM ports then will there be any violation of port driver signing? The customer wants to use them as COM ports in their application. I would highly appreciate if somebody from Microsoft responds to this question. Thanks, Hablu Tag: bswap intrinsic? Tag: 57184

Is there an intrinsic to wrap the bswap x86 instruction. I have written a
function to do this with inline assembler but I am wondering if an intrinsic
wouldn't be a better solution.

Regards,

George.

Top

Re: bswap intrinsic? by George

George
Sun Jan 09 13:19:54 CST 2005

Sorry, I posted to the wrong newsgroup.

Regards,

George.

"George M. Garner Jr." <gmgarner@newsgroup.nospam> wrote in message
news:uTQ3O$n9EHA.2016@TK2MSFTNGP15.phx.gbl...
> Is there an intrinsic to wrap the bswap x86 instruction. I have written a
> function to do this with inline assembler but I am wondering if an
> intrinsic wouldn't be a better solution.
>
> Regards,
>
> George.
>

Top

Re: bswap intrinsic? by Maxim

Maxim
Sun Jan 09 15:17:45 CST 2005

> Is there an intrinsic to wrap the bswap x86 instruction.

IIRC yes.

--
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
maxim@storagecraft.com
http://www.storagecraft.com

Top

Re: bswap intrinsic? by pavel_a

pavel_a
Mon Jan 10 06:09:04 CST 2005

unsigned short __cdecl _byteswap_ushort(unsigned short);
unsigned long __cdecl _byteswap_ulong (unsigned long);
unsigned __int64 __cdecl _byteswap_uint64(unsigned __int64);

in ntddk.h ... and lots of othr intrinsics.

"Maxim S. Shatskih" wrote:
> > Is there an intrinsic to wrap the bswap x86 instruction.
>
> IIRC yes.
>
> --
> Maxim Shatskih, Windows DDK MVP
> StorageCraft Corporation
> maxim@storagecraft.com
> http://www.storagecraft.com
>
>
>

Top