About DMA : Packet Based & Common Buffer

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
    - USB class For a long time, Microsoft has been mentioning in this group (and various places in the documentation) that peripherals are required not use Class=USB in their infs because that's reserved for host controllers and hubs. Dire WHQL consequences have been hinted at. I was just trying to convince one of our firmware guys of this, and he pointed out that there are several .inf files in Windows\INF that violate this rule, including many that are WHQL signed... notably many things that act like thumb drives such as media players, as well as a couple of modems (mostly this question comes up around here with respect to development devices such as FTDI/EZ-USB/etc.). So... What's the current state of this advise? What's the actual fallout if you violate this rule? -- Ray Tag: About DMA : Packet Based & Common Buffer Tag: 81859
  - 2
    - Vista TCP I'm having a problem with a 802.11n WiFi driver I am developing for Vista (build 5384). When I use a static IP address, everything works fine. However, if I use DHCP, I don't get a DHCP address allocated to me from the AP. I know DHCP in Vista 5384 works because I have a Realtek 8185 802.11g card that works OK. Here are a couple of clues to the problem: 1. The first thing I found was that the driver was transmitting the DHCP DISCOVER broadcast packet and then receiving a DHCP OFFER broadcast back from the AP. However, the driver's packets filters were setup to handle only DIRECTED (i.e. unicast) packets. This seemed a little strange, since I believe during TCP protocol initialization the driver should be called to set the the MULTICAST and BROADCAST filters. 2. Thinking that perhaps I had a problem with the packet filters, I jury-rigged the driver to accept all packets which then started indicating the DHCP OFFFER broadcasts up to the protocols. This did not work. TCP still would not transmit a DHCP REQUEST packet to the AP. 3. Thinking that I still have a problem with the packet filters, I add more debug traces to the driver and rerum the test. PRESTO! Now I have a valid DHCP address from the AP. Strange. However, now I cannot ping the AP. Anybody have any thoughts about this problem? ((&-> Tag: About DMA : Packet Based & Common Buffer Tag: 81851
  - 3
    - USB 2.0 isoch irp IoCompletionRoutine delayed on SMP machines Hi all. I'm developing a USB 2.0 driver for XP that runs in and out isoch and bulk pipes at the same time. On SMP machines I'm seeing the following issue: As long as only isoch irps are submitted, the isoch irp IoCompletionRoutine gets called immediately (or at least less than 1 ms) after the IN stream urb was filled with data. However, as soon as bulk or control requests are submitted asynchronously, there is a certain chance that the callback into the isoch IoCompletionRoutine gets delayed for up to 10 ms. This happens only on SMP machines and only with the MS USB 2.0 EHCI driver, I'm not seeing any starvation on single processor systems or when using the USB 1.1 UHCI driver on SMP systems. The starvation always occurs if the bulk request was submitted right before the isoch IoCompletionRoutine is due. I tried synchronizing the bulk requests with the isoch requests (i.e. only issue a bulk request immediately after an isoch request). This reduced the frequency of the starving issue a lot. However, I'm still seeing delays from time to time, especially if the bulk traffic is high. Another workaround might be to set the processor affinity for both, the isoch and the bulk requests to the same CPU. However, to do this, I would need to change the processor affinity of the EHCI DPC that issues the IoCompletionRoutine. Is there a way to retrieve the PKDPC from the IoCompletionRoutine? I guess not. So my question is: Is this a known issue, and if so, is there a better workaround? -- Stephan Tag: About DMA : Packet Based & Common Buffer Tag: 81848
  - 4
    - avstream pin allocator framing requirements Hi everyone, I'm in the middle of writing my first avstream-based minidriver, and have been experimenting with the behaviour of everything. I'm confused, though, when it comes to the concept of allocator framing. I gather from the DDK documentation that I should be able to declare a KSALLOCATOR_FRAMING_EX struct and point to it in my pin descriptor, and through it specify what frame sizes my pin's process dispatch can accept. DECLARE_SIMPLE_FRAMING_EX(PinFraming, STATIC_KS_TYPE_DONT_CARE, KSALLOCATOR_REQUIREMENTF_SYSTEM_MEMORY, 1024, // Frames FILE_BYTE_ALIGNMENT, 6, // MinFrameSize 6 // MaxFrameSize ); Now with these requirements specified by my pin, I'd expect to be given up to 1024 frames, each of 6 bytes in size, in one call to my process dispatch. What I actually get passed is one frame 8192 bytes long, clearly beyond the range I specified. Any ideas why the system is ignoring my framing requirements? I've tried specifying much larger frame sizes, but the system still passes me 8192 bytes. Help! Thanks, Nick Tag: About DMA : Packet Based & Common Buffer Tag: 81847
  - 5
    - USBPrint.sys on Windows XP SP2 What would be the conditions (or reasons) for usbprint.sys to stop sending the SOF to a printer device? Is there a maximum number of errors reported by the printer device that USBPrints.sys would shutdown the USB communication? Any other information would help. Mario Tag: About DMA : Packet Based & Common Buffer Tag: 81846
  - 6
    - How to set the uninstall order of drivers with DIFxApp? I've created an installation package with InstallShield 11.5 for Windows. It installs four different drivers using "ready made" DIFxApp functions ie. the device driver wizard of InstallShield. One of the drivers is the parent driver for all the remaining three drivers. The installation works fine, but when I have to uninstall the drivers there is a problem: I should be able to tell the order in which the drivers are uninstalled. Is there a possibility to tell the DIFxApp the order in which the drivers should be uninstalled? I did not found any solution to this in InstallShield help, so I wanted to ask if anybody has already had to figure this out. Tag: About DMA : Packet Based & Common Buffer Tag: 81839
  - 7
    - HELP: How to debug "stuck IRPS" in TDI filter Hi Our product contains TDI filter and File system filter driver. Lately, several applications began to stuck with no known reason. I think that one of our filters causes some of the IRPS to "get stuck" so the user mode process isn't getting a response to its request and it gets stuck too. These are the symptoms: -Two computers that sometimes have their telnet application stuck when it tries to disconnect. -One computer had it's excel stuck. I don't know if excel is creating some kind of communication so I don't know if the problem is in the TDI or the file system filter. -One computer that runs AutoIt (a QA helper app). The AutoIt got stuck and afterwards all iexplore, msnm, and outlook got stuck (the application window is freeze after I run it). The wired thing is, that mozilla firefox runs OK! All the problems occurred after SEVERAL DAYS of work while our product was running. Because most of the problems were in communications apps, we think it's the TDI filter. Any idea how to debug such a problem? Is there a way to get all the stuck IRPS for a process or a driver? (Maybe doing a manual dump and get it from there?) Thanks for any help. Tag: About DMA : Packet Based & Common Buffer Tag: 81831
  - 8
    - Serial Joystick Hello Group I'm new in driver development. And must admit that this is realy complicated for me :). My problem is like that: I have a joystick that use the serial port. My goal is to write a "driver" that use this serial port but acts like a joystick which is selectable in the game controller dialog. I've found out that I need the sermouse and the vhdmini example to to that. But could what are the steps that I have to do. Or is there any example code anywhere? I hope you guys could help me. thanks Tag: About DMA : Packet Based & Common Buffer Tag: 81829
  - 9
    - Scanner WIA Driver Hi, I'm developing a wia driver for a scanner, but I can't understand the different behavior of MS Paint and MS Office. If I scan with Paint I have a mirrored and rotated image, but in Office (Word etc) I don't have this behavior. Do you know why ? Can I modify this, to have the same behavior? Thank you Tag: About DMA : Packet Based & Common Buffer Tag: 81828
  - 10
    - Strange difference between XP and win2k driver Hi, Im am currently "porting" a XP driver to Win2k and during the process i got stuk on a wierd problem when trying to read from the USB. The USB device chip is a ftdi um245R. When reading from it in XP, that works ok i do: UsbBuildInterruptOrBulkTransferRequest( &current->dUrb, sizeof(_URB_BULK_OR_INTERRUPT_TRANSFER), pdx->myPipe, &buffer->data, NULL, pdx->bufferChunkSize, USBD_TRANSFER_DIRECTION_IN | USBD_SHORT_TRANSFER_OK, NULL ); KeInitializeEvent(&current->dEvent, NotificationEvent, FALSE); current->dIrp = IoBuildDeviceIoControlRequest( IOCTL_INTERNAL_USB_SUBMIT_URB, // IoControlCode pdx->LowerDeviceObject, // DeviceObject NULL, 0, NULL, 0, TRUE, &current->dEvent, // Event &current->dIostatus // IoStatusBlock ); if (!current->dIrp) { KdPrint((DRIVERNAME " - Unable to allocate IRP for sending URB\n")); goto endThread; } stack = IoGetNextIrpStackLocation(current->dIrp); stack->Parameters.Others.Argument1 = (PVOID) &current->dUrb; status = IoCallDriver(pdx->LowerDeviceObject, current->dIrp); KdPrint((DRIVERNAME " - IoCallDriver 1 returned: %8.8X\n",status)); This works fine in XP, status returned from IoCallDriver is 0x103. But when i do the same thing in Win2k i get a status of 0xC000000D which in my test application translates to ERROR_INVALID_PARAMETER. The test application is exactly the same code in xp and win2k. Can someone please help me with this, i have no clue on why it behaves like this since i believe that this is the correct syntax of doing a read. Regards Wilhelm Tag: About DMA : Packet Based & Common Buffer Tag: 81825
  - 11
    - SDIO: Is cmd53 tested with sdbus driver in XP, SP2 for block mode Hi, I'm doing sdio driver on XP SP2. This SDIO driver is compiled using 2003 SP1 ddk. SDIO card is coming up properly and i could do several operations using CMD52 (read/write etc). I have not enabled the interrupts yet. Now, i'm trying the CMD53. CMD53 works fine in byte mode. That is BlockMode in SD_RW_EXTENDED_ARGUMENT is set to 0 and the parameters in SDCMD_DESCRIPTOR set to SDTT_SINGLE_BLOCK and SDRT_5. But when i try the SDTT_MULTI_BLOCK mode, the command sent on the bus is not proper. I captured the CMD53 on the bus using LA, and this is what i found. My block size set was 4 bytes. I was trying to read/write 8 bytes so that two blocks of transaction should happen. But all the time, the CMD53 on the bus had 8 as Block Count. After i did some different exercise, i found that the Parameters.DeviceCommand.Length is sent as block count for CMD53 instead of .u.bits.Count in SD_RW_EXTENDED_ARGUMENT. (I made sure that the Block Mode in SD_RW_EXTENDED_ARGUMENT is set to 1 so that even on the bus i see the same for cmd53. I do not see any response from the SDIO card for this and also, the bus driver is returning back, 0xC0000185 or 0x80000011.. My sdio card supports block mode (the card capability reg- 0x08 of cccr reads 0x1F .....) Is the bus driver in XP SP2 tested with CMD53 block mode (or SDTT_MULTI_BLOCK) ? I have attached the CMD53 function i've written, along with this mail... Please let me know if there is any problem. Eliyas, please help ! Rgds Esha ------------------------------------ NTSTATUS SdioReadWriteBuffer(PDEVICE_OBJECT fdo,IN ULONG Function,IN PUCHAR PBuffer,IN ULONG Address, IN UCHAR BlockMode,IN ULONG Length,IN BOOLEAN WriteToDevice,OUT PULONG BytesRead) { PFDO_DATA fdoData; PSDBUS_REQUEST_PACKET sdrp; SD_RW_EXTENDED_ARGUMENT extendedArgument; NTSTATUS status = STATUS_SUCCESS; PMDL Mdl=NULL; UCHAR i; PUCHAR pBuf=NULL; ULONG BlockSize=0; ULONG len = Length; ULONG Data=0; SDCMD_DESCRIPTOR ReadIoExtendedDesc = {SDCMD_IO_RW_EXTENDED, SDCC_STANDARD,SDTD_READ, SDTT_MULTI_BLOCK , SDRT_5}; //{SDCMD_IO_RW_EXTENDED, SDCC_STANDARD,SDTD_READ, SDTT_MULTI_BLOCK , SDRT_1}; SDCMD_DESCRIPTOR WriteIoExtendedDesc = {SDCMD_IO_RW_EXTENDED, SDCC_STANDARD,SDTD_WRITE, SDTT_MULTI_BLOCK , SDRT_5}; fdoData = (PFDO_DATA)fdo->DeviceExtension; SdioDbgPrint(TRACE,"cmd53: fn %d len %d,add 0x%X blkmode %d WrRd %d\n",Function, Length,Address,BlockMode,WriteToDevice); if(BlockMode == BYTE_MODE) { ReadIoExtendedDesc.TransferType = SDTT_SINGLE_BLOCK; WriteIoExtendedDesc.TransferType = SDTT_SINGLE_BLOCK; //len = BlockSize; len = Length; } else { //(Function == FN0) ? (BlockSize = fdoData->BlockSizeFN0) : (BlockSize = fdoData->BlockSizeFNx); //status = SdioReadWriteByte(fdo, FN0,(PUCHAR)&BlockSize,FN0_BLOCK_SIZE_REG,FALSE); //just for testing, make sure that the same is written in FN0_BLOCK_SIZE_REG before running this BlockSize = 0x04; SdioDbgPrint(TRACE,"\nBlksize 0x%X\n",BlockSize); //Round of the length to the block size if(Length % BlockSize) len = (((Length / BlockSize) * BlockSize) + BlockSize); else len = ((Length / BlockSize) * BlockSize); } SdioDbgPrint(TRACE,"Calc len 0x%X\n",len); pBuf = ExAllocatePool(NonPagedPool,len); if (!pBuf) { return STATUS_INSUFFICIENT_RESOURCES; } RtlZeroMemory(pBuf, len); if (WriteToDevice) RtlCopyMemory(pBuf,PBuffer, Length); // first, get an MDL to map the data. Call IoAllocateMdl to // allocate an MDL and pass in a pointer to a buffer // allocated from the non-paged pool. Mdl = IoAllocateMdl(pBuf, len, FALSE, FALSE, NULL); if (Mdl == NULL) { ExFreePool(pBuf); return STATUS_INSUFFICIENT_RESOURCES; } MmBuildMdlForNonPagedPool (Mdl); // next, allocate a request packet for the arguments of the command sdrp = ExAllocatePool(NonPagedPool, sizeof(SDBUS_REQUEST_PACKET)); if (!sdrp) { IoFreeMdl(Mdl); return STATUS_INSUFFICIENT_RESOURCES; } RtlZeroMemory(sdrp, sizeof(SDBUS_REQUEST_PACKET)); sdrp->RequestFunction = SDRF_DEVICE_COMMAND; extendedArgument.u.AsULONG = 0; if(BlockMode == BYTE_MODE) { extendedArgument.u.bits.BlockMode = 0; extendedArgument.u.bits.Count = len; } else { extendedArgument.u.bits.BlockMode = 1; extendedArgument.u.bits.Count = len / BlockSize; } SdioDbgPrint(TRACE,"cmd53:set Count %d, blkmode %d\n",extendedArgument.u.bits.Count, extendedArgument.u.bits.BlockMode); extendedArgument.u.bits.Function = Function; extendedArgument.u.bits.OpCode = 1; // increment address extendedArgument.u.bits.Address = Address; if (WriteToDevice) { extendedArgument.u.bits.WriteToDevice = 1; sdrp->Parameters.DeviceCommand.CmdDesc = WriteIoExtendedDesc; } else { extendedArgument.u.bits.WriteToDevice = 0; sdrp->Parameters.DeviceCommand.CmdDesc = ReadIoExtendedDesc; } sdrp->Parameters.DeviceCommand.Argument = extendedArgument.u.AsULONG; sdrp->Parameters.DeviceCommand.Mdl = Mdl; //This send the Block Count in CMD3 as the blocksize * num of block and //makes the system crash - BSOD //sdrp->Parameters.DeviceCommand.Length = len; //With this the SdBusSubmitRequest returns 0xc0000185 status and //the Block count in CMD53 gets the actual byte read/write length //sdrp->Parameters.DeviceCommand.Length = Length; //This send the Block Count in CMD3 as proper block count, but still the //bus drivers gives status as 0xc0000185 or 0x80000011 sdrp->Parameters.DeviceCommand.Length = extendedArgument.u.bits.Count; //len; // finally, submit the request status = SdBusSubmitRequest(fdoData->SDBusInterface.Context,sdrp); SdioDbgPrint(TRACE,"cmd53: resplen: %d\n",sdrp->ResponseLength); SdioDbgPrint(TRACE,"cmd53: sts: %X \n",status); SdioDbgPrint(TRACE,"cmd53: info: %X\n",sdrp->Information); if (NT_SUCCESS(status) && !WriteToDevice) { *BytesRead = Length; RtlCopyMemory(PBuffer, pBuf, Length); } IoFreeMdl(Mdl); ExFreePool(pBuf); ExFreePool(sdrp); return status; } ---------------------------------- Tag: About DMA : Packet Based & Common Buffer Tag: 81823
  - 12
    - where can I find the definition of USB_STATUS_XXX code? where can I find the definition of USB_STATUS_XXX code on return of USB host controller driver? Thanks a lot! Best Regards Amanda Lin Tag: About DMA : Packet Based & Common Buffer Tag: 81808
  - 13
    - How to view earlier posts? Hi everyone, Every time I connect to this site, my Outlook only load a portion (I have no idea how it selects) of the posts. How can I download the rest posts (I will not waste the space and the time of the experts if similar questions have already been asked and answered)? Thanks in advance. Tag: About DMA : Packet Based & Common Buffer Tag: 81799
  - 14
    - Install KMDF on Win2k Hi everyone, I downloaded KMDF, but my Win2k professional seems not recognize the WDFv10.iso file. How can I install it? Do I need another tool for .iso file? Thanks in advance. Tag: About DMA : Packet Based & Common Buffer Tag: 81791
  - 15
    - Which miniport created \Device\HardDisk%d? I'm working on a Windows diskless boot solution (using SRP over an Infiniband adapter) and I need to know on behalf of which miniport each \Device\HardDisk%d device object was created. This is so that I can distinguish the Disk device objects that were created on behalf of my SRP SCSI miniport from those corresponding to local hard drives. This lets me choose the proper partition to mount the boot volume from. I know how to get the device objects for each \Device\HardDisk%d, but I don't know how to tell which SCSI (or whatever) miniport discovered the physical drive that the object represents. Any advice would be appreciated. Rob Netzer System Fabric Works, Inc. rob@systemfabricworks.com Tag: About DMA : Packet Based & Common Buffer Tag: 81790
  - 16
    - Cache Coherency In an earlier post[1], Stephan Wolf says "On x86, there is no need to take care of any caching issues at all because x86 always guarantees cache coherency." I may be taking him too literally, but I don't see how that can be true. I've got a PCI adapter using a PLX 9030 that presents a 1K memory buffer at PCI BAR4 which is marked as cacheable to the OS. This is mapped in at DriverLoad using MmMapIoSpace() with MmCached. Once the OS is done transferring data out of the buffer, the buffer is released back to the adapter to refill with incoming data. When the memory is done refilling, the adapter hands ownership of the memory back to the OS and an interrupt is asserted. The problem I'm trying to address is how to ensure the CPU reloads its internal cache with the new data since I can see no way that it can be aware of the changes. The PLX 9030 is a target only device, so it can't master a special cycle on to the system bus. And I didn't see any way the CPU could snoop the memory for changes. The natural (and documented) way to keep the cache coherent is to call KeFlushIoBuffers() in the IRQ DPC before reading the buffer. But, as Stephan mentions, and as I've confirmed, this is completely macrolized away for x86 builds. It doesn't do anything. I haven't seen any HAL'd abstractions in the MmXxx() or KeXxx() routines that invalidate specific cache lines, so I think I'm being forced to get a little closer to the metal. It looks like the right thing to do would be to issue an CLFLUSH[2] op sandwiched between two MFENCE ops. But, that's pretty brute force, and I'm not sure these are supported by AMD. At least I couldn't find a definitive reference. If they definitely are, or if there's a different way to flush a cache line on AMD, let me know. Of course, WBINVD will work, but that flushes the whole cache to system memory, and all I need to make coherent is a tiny chunk that's <1K. Mighty inefficient. Another alternative would be to issue an INVLPG, but that seems to get all over the Memory Manager's business. If this were the right way, it would probably be better to MmUnmapIoSpace() and then re-map them. <gack> Hopefully this wasn't too newb. I've looked around the net and here for answer, so fire away! -Mark [1] http://groups.google.com/group/microsoft.public.development.device.drivers/browse_frm/thread/7b6104ea25dfcebb/351bc138b87e61b1?q=cache+flush&rnum=4#351bc138b87e61b1 [2] Sure wish there was a MmFlushCache(base, size, read_not_write) call. Wait! There's KeFlushIoBuffers() now. :-/ Tag: About DMA : Packet Based & Common Buffer Tag: 81778
  - 17
    - Two filters in one AVStream driver Hello, All. I'm working on AVStream minidriver. And I want to create minidriver that will provides DirectShow filter to the user mode application all time while device is connected to the system and another one filter type that will be seen by system only then my device is switched in special state (looking at hardware state). And another one requirement is that these filers have different category (I mean category like "Video Capture source"). In AVStream documentation I have found two ways to implement additional filter type from same device. The first one is to implement two filter factories in one minidriver and to create filters of each type separately from driver. Does system notifies appications about new filters arrival in this case? Does there is possibility to point system that these filters have different categories (capture and decode, for example). The second way is to create AVStream child device. But it is not clear for me what driver will serves this child device. Please help me to clear these questions. Andrey Tag: About DMA : Packet Based & Common Buffer Tag: 81772
  - 18
    - Video-Capture Card Device Driver ?? 1. I want to write a device driver for a PCI video-capture card And in ddk docs related topics are : Kernel-Streming AVStream Minidrivers Streaming Minidrivers Audio Drivers Video Capture Drivers USB Camera Drivers DVD Minidrivers Broadcast Driver Architecture Drivers AV/C Client Drivers in WHDC pages it says , KS is a former model of AVStream vs. vs.. But KS is still in ddk docs For the moment , can you guide me about my video-capture-driver challenge , KS or AVS ? What about selecting framework -> streaming-mini, filter-centric or pin-centric ?? And also what about AV/C Client drivers , do i need to write it ?? (Other resources will make me very happy , if available for streaming... ) 2. DDK Docs say : "Audio/Video streaming (AVStream) client minidrivers (Microsoft® DirectX® 8.0, Microsoft® Windows® XP and later versions of the operating system) are typically written in C++. " Why typically C++ ? Is it a must or an advice for good coding or anything related with DirectX ? 3. What about needed minimum hardware knowledge for this project , what must i know about my video-capture card to write it's driver ? /* If anyone could provide any suggestions or insight, I'd be very appreciative. */ Tag: About DMA : Packet Based & Common Buffer Tag: 81770
  - 19
    - Access to the I2C_CALLBACKS data in a video miniport Is there a way for a display driver to access the subject structure in a video miniport? Tag: About DMA : Packet Based & Common Buffer Tag: 81768
  - 20
    - NdisMIndicateStatus max buffer Hi, Does any one know if there is any limitation on the buffer size given to NdisMIndicateStatus? Thanks Miki Tag: About DMA : Packet Based & Common Buffer Tag: 81767
  - 21
    - What am I doing when I read/write to LPT1 ? I am just curious, but what am I actually doing when I CreateFile to LPT1, is there a protocol which explains this? Tag: About DMA : Packet Based & Common Buffer Tag: 81765
  - 22
    - WinXP SDIO INF While making sdio inf file what has to be reffered to the Class ,class GUID directive of version section I have doubt bcos the in the devguid.h file of include folder (along with ddk) there is not any section for SD or SDIO Do I need to use Unknown section I need some idea on making an Insataller application for my SDIO. I also need an sample sdio inf file Tag: About DMA : Packet Based & Common Buffer Tag: 81763
  - 23
    - Hiding device entry Hi all I want to hide my device entry under device manager in Windows XP . I came across a post which said when we set the DEVNODE status flag to DN_NO_SHOW_IN_DM we can acheive this. But the configuration manager part has API calls only to get the status of the devnode. Does anyone know how to set this flag. Is there any other means to hide device entry. My driver is not a filter driver. Mk Tag: About DMA : Packet Based & Common Buffer Tag: 81762
  - 24
    - how to use DbgView hi all, i try install TestCap Example in DDK into my syatem and use DbgView to see output debug message, but i can not, anybody help me how to use DbgView to see debug message? Thanks, David, Tag: About DMA : Packet Based & Common Buffer Tag: 81760
  - 25
    - Stop Service Experts! I have a device conected to my computer with a USB, i want to stop this device and then start it again. i selected disable in the device manager and the device was disable successfully. But when i tried to do it programmaticatly it failed. I tried calling "net stop service_name_blabla" from the command line. and i tried using SERVICE_CONTROL_STOP,in ControlService but they both failed, reason: "The requested control is not valid for this service. " error how can i disable the service with C++ like i'm doing well in the device manager ? thanks. Tag: About DMA : Packet Based & Common Buffer Tag: 81759
- Next
  - 1
    - Get Device Descriptor from User Mode hello all, I am facing a problem, how to det the device descriptor of USB device from User Mode, I just know VID n PID of the device. It could be great if a piece of code is availble. thanks sachin Tag: About DMA : Packet Based & Common Buffer Tag: 81758
  - 2
    - Setting DefaultInputSlot in a PPD? Hello all, I build print drivers with the PostScript OEM customization tools, and I'm stuck on some driver/PPD interaction. The second tab in DocumentProperties is normally Paper/Quality, with two drop downs for Paper Source and Media. These are populated from the PPD via InputSlot and MediaType sections. The problem is that the PPD parser (or some other component) does not honor the DefaultInputSlot and DefaultMediaType directives, and it does its own thing. I'm using the German version, and the phrase "Automatisch Auswahlen" ("Automatically Select") is filled in for us. This phrase does not appear in the PPD, but does appear in PS5UI.DLL's resource section. This all means that we're not able to set a reasonable default, and this is causing us some real problems for a customer implementation. If we change the default in the UI, it "sticks", but we don't have portable access to the DEVMODE to muck with it ourselves. Looking in the MSDN docs: http://msdn.microsoft.com/library/en-us/print_d/hh/Print_d/pscript_34563439-cd1f-4418-b51d-a3250e35df28.xml.asp It seems clear that InputSlot is being handled specially, and it suggests to me that DefaultInputSlot is simply ignored entirely. Looking at the IPrintOemDriverPS::DrvGetDriverSetting helper function, it seems that I'm able to *fetch* values, and IPrintCorePS2 lets me enumerate them, but I think I cannot find any mechanism to *set* these options. Is there any way I can set a default value for this field when the driver is installed without requiring the user to click a box after install? Thanks all, Steve -- Steve Friedl / UNIX Wizard / Microsoft MVP / www.unixwiz.net Tag: About DMA : Packet Based & Common Buffer Tag: 81755
  - 3
    - SDIO multi-block write support (WinXP SP2)? I'm developmenting a SDIO WLAN client driver in WinXP SP2. When I use SdBusSubmitRequest() to send a CMD53 with multi-block. The "Byte/Block count" field in CMD53 will be overwrite with the length of the MDL (when < 64 bytes) or 64 bytes (when > 64 bytes). Does SDBus support SDIO multi-block mode? If yes, how to program it? Tag: About DMA : Packet Based & Common Buffer Tag: 81753
  - 4
    - I2C slave device address Hi all, I am in a very strange situation for the I2C bus slave device address manipulation. Basically I have some slave devices which I communicate over the I2C bus. They work fine on different machines with Intel motherboard. The problem arises when I connect a specific slave device to the Intel 82865G machine. The address I am sending over the I2C bus for the slave device is not acknowledged by the device. The same address is acknowledged by the device on another intel machine. According to the I2C specs, the slave device is supposed to send the acknowledgement when Master finsishes sending the Bits of the address byte by pulling the SDA line low. But in this case the slave device is not pulling the SDA data line low. I have tried many cases.. like increasing the delay between the SCL , SDA check, checking the video memory, assigning the different address. But nothing is working. I am not sure whats the difference between I2C bus on one motherboard and that of another. Both are Intel 82865G. I communicate with the I2C in a driver for a graphics cards and for Intel I directly get the Video memory, convert it to the Physical memory and write to the DDC registers. I am able to send the Start condition on the I2C bus , but when I send the address of the Slave device , its not acknowledging. Other slave devices of Same type work fine. It will be very helpfull If somebody can tell me whats going on over the I2C bus for this. Thanks. -- Abhijit. Tag: About DMA : Packet Based & Common Buffer Tag: 81752
  - 5
    - How long does it take to read a usb report? I am trying to check what is the maximum speed for read/writes of my pic 18f4550 usb device. It is a full speed device and an 8 byte length report, and I am using ReadFile and WriteFile. My results are that I get about 100 read write cycles in one second. Is that a normal speed for 8 byte reports? It seems quite slow ? My C loop is tight, and my device is a loopback device, i.e. it simply copies the input to the output? Tag: About DMA : Packet Based & Common Buffer Tag: 81729
  - 6
    - sscanf Have been searching around without succes on this: Is there a recommended or 'standard' alternative to swscanf for use in a driver? (I'm wanting to parse a WSTR from the registry). Tag: About DMA : Packet Based & Common Buffer Tag: 81727
  - 7
    - Driver design interogation... Hi, I am new on driver development so my question may be stupid for guru ^_^... My driver is base on passthru NDIS from DDK, I try to develop an IP filter and HTTP-DNS content filter. I want to use SQLite to store a big list of ip and rules. I want to put this code on a service. My design question is : - Is this solution is correct ( speed, complexity,... ). - How my driver can communicate with the db service. I see some info on WMI, but I have some interrogation on performance. Thanks for your help, A++ Tag: About DMA : Packet Based & Common Buffer Tag: 81726
  - 8
    - Ndisprot problem with Intel Pro/Wireless 3945ABG Hi, I have developped an application able to control the laptop wifi adapter by setting and querying the oids OID_802_11 via the Ndisprot provided in Windows DDK 3790. This application works perfectly with the chipset Intel Pro/Wireless 2200BG and Intel Pro/Wireless 2100 and many other chipsets. The problem is that when I try to use it with the chipset Intel Pro/Wireless 3945ABG the oid OID_802_11_SSID does not work. The connection is simply not established and the IP is not retrieved. All the others oids works fine. Did you have any idea about it? Is it a known bug? Did you ever try the Ndisprot with Intel Pro/Wireless 3945ABG? I've already tried also the Ndisuio approach (using the source code of the DDK 2600.1106) obtaining exactly the same result: no connection is established when the OID_802_11_SSID is set. All the others oids works fine. Thanks Davide Tag: About DMA : Packet Based & Common Buffer Tag: 81723
  - 9
    - How to save data in Display Driver? Hello, I need to save some information(for example, data in memory buffer) in Display Driver, but i don't know how to do this.(ZwXxx Routines?) Anyone can help me please? Thanks! Tag: About DMA : Packet Based & Common Buffer Tag: 81722
  - 10
    - Network Refresh slow on S3 wake up using Keyboard key press Hi all, When a XP PC is put in the sleep mode (S3), there is no network refresh happening immediately when PC comes out of S3 by means of pressing a key on the keyboard. The network refresh happens randomly from 40 seconds to 1 minute or even more sometimes. I want to get the network refresh done as soon as the PC comes out of S3. Can anyone help me ? Sri Tag: About DMA : Packet Based & Common Buffer Tag: 81520
  - 11
    - how to reset a USB device from user mode Sometimes I experienced the device failed enumeration. I need unplug and plug back that device again. Is there any solution that my user mode application can do that for me? Given the handle to a USB hub and port number. How could I reset the device connected to that port from user mode application? I noticed there is IOCTL_USB_HUB_CYCLE_PORT defined in wxp ddk. But I cannot find any documents for that. Any suggestions? Thanks zhwqiu Tag: About DMA : Packet Based & Common Buffer Tag: 81509
  - 12
    - 1394 peripheral device identifcation (1) My driver is similar to 1394diag.sys which installs above Microsoft 1394 stack in Windows 2000/XP. This is unsigned driver. DeviceID for 1394 peripheral device is 1394\VendorName&ModelName. All my peripheral devices use same VendorName and ModelName. Still when a new peripheral device is connected, Windows 2000 and XP comes up with "New Hardware found" wizard. How can I avoid this? I donâ??t mind if I get warning "unsigned driver", but I canâ??t install driver again and again. Is driver signing required here to stop Windows prompting for driver for different EUID? DeviceId is same for all my 1394 devices and EUID is unique. What is the meaning of DeviceId if EUID is considered for DeviceId. (2) Is there any auto install method for driver to 1394 peripheral devices? I am looking for a way to have unique driver installed once for VendorName and ModelName. I know there are some topics in newsgroup. But what I dont understand is document clearly states driver for device is identified by VendorName and modelName. Documentation doesnot say driver has to be installed for every device if driver is unsigned. It only says driver installation will give "Unsigned driver" warning. So I am not sure if this is happening because driver is not signed or is it because of some other error/option in my own installation. (3) I didnâ??t see any driver signing document for peripheral drivers to 1394, found only 1394 bus drivers documentation. Thanks, Raj Tag: About DMA : Packet Based & Common Buffer Tag: 81451
  - 13
    - Modem Handles not destroyed Has anyone had issues with the USB modem handles not being closed when a device is removed? Tag: About DMA : Packet Based & Common Buffer Tag: 81447
  - 14
    - Generic Volume Surprise removal In windows 2k if you have a Mass Storage device and remove it from windows without stopping the device first you get a surprise removal message. I am able to suppress the MSD surprise removal but I cannot figure out with a filter driver how to remove the Generic Volume message Can anyone help? Dan Tag: About DMA : Packet Based & Common Buffer Tag: 81437
  - 15
    - Bluetooth HCI Ioctls Is there any documentation that can help me find out how to communicate with the Dell TrueMobile Bluetooth module? I need to send HCI commands to the Bluetooth module. Is there a sample in either the SDK or DDK? Tag: About DMA : Packet Based & Common Buffer Tag: 81436
  - 16
    - WHQL, HCT, WHDC, Windows Logo, Driver signing WHQL, HCT, WHDC, Windows Logo, designed for windows, Driver signing.....are these all same. Is windows logo requirement and driver signing are same? I am not able to find how to get my driver signed. This is 1394 driver for peripheral device. This driver is on top of Microsoft 1394 stack and is designed in the lines of 1394diag.sys. All I can find is 1394 bus driver or 1394 hardware. I cant see any mention of 1394 peripheral driver. I dont understand why 1394 peripheral link is not even refered from 1394 bus driver. Someone please refer the link. After Microsft arm twisted me into driver signing with otherwise crappy driver installation, I dont find proper information on this. Thanks. Tag: About DMA : Packet Based & Common Buffer Tag: 81429
  - 17
    - MmMapIoSpace what exactly it does? MmMapIoSpace This function maps the specified physical address range to nonpaged process-dependent address space, providing a directly mapped virtual address to the device. Internally what exactly happens behing this API call. How does translation of address occurs?? suppose i have address dedicated for LCD controller or any other peripheral. In my driver i call MmMapIoSpace to access the device.. How will the processor know this address is meant for that particular device?? Or how does transaltion of address happens? Any suggestion is appreciated. Thanks Vijay PIC ,Bangalore Tag: About DMA : Packet Based & Common Buffer Tag: 81423
  - 18
    - Conflicting errors code with USB Bulk requests When our USB function driver is used in a loop or for a long transfer of data (MBs)on a Bulk endpoint, after a short period of time the I/O status (IRP) returns a failure but the URB status is successful. This is after hundreds of successful requests. The IRP errors have been STATUS_IO_DEVICE_ERROR or STATUS_IO_TIMEOUT. The URB status is USBD_STATUS_SUCCESS in either case. In what conditions would the Windows USB stack return these errors and the URB be successful? Second question, when testing for IRP error codes from a IOCTL_INTERNAL_USB_SUBMIT_URB request, which I/O status should be tested? IO_STATUS_BLOCK ioStatus = { 0, 0 }; PIRP irp = IoBuildDeviceIoControlRequest(...., &ioStatus); // Boilerplate code... ntStatus = IoCallDriver(pdx->lowerDeviceObject, irp); if (ntStatus == STATUS_PENDING) // Wait on event At this point is io.Status.Status == irp->IoStatus.Status ???? If not, which is the proper method?? Thanks, SteveB, MCSD Welch Allyn, Inc. Tag: About DMA : Packet Based & Common Buffer Tag: 81400
  - 19
    - best source address (TDI) Hello, I want to find the best source address for a given destination address in a TCP TDI filter, and bind the connection endpoint to that source address. How can I do this in TDI? Thanks, Piyush PS: Is this the right list for this question? Tag: About DMA : Packet Based & Common Buffer Tag: 81026
  - 20
    - ZwQueryInformationFile Hello I'm using ZwQueryInformationFile to retreive a file name from a file handle. So i use the FileNameInformation class. Unfortunately the filename always begins with / (for example /Windows). My question is: how am i supposed to make a difference between C:\Windows and D:\Windows? Tag: About DMA : Packet Based & Common Buffer Tag: 81022
  - 21
    - How to create a dynamic pin in video cap driver helle, everyone I'm an newbie in windows wdm driver development, i want to develop a video capture driver based testcap, can dynamic create a pin, meas when i connect the filter in graphedt, i hope i can dynamic create a new pin and dynamic delete the pin at the time of break connect like splitCam( [url]http://www.splitCamera.com[/url]). I think that is a very cool function. Can you tell me how I can get this function? Can you list some documents help me. Thank you very much for your time and consideration! -- wintop ------------------------------------------------------------------------ Posted via http://www.codecomments.com ------------------------------------------------------------------------ Tag: About DMA : Packet Based & Common Buffer Tag: 81020
  - 22
    - [Off-Topic] WDF Books & W.Oney ? I've searched and found only OSR Publication's KMDF book , do you know any other books for WDF (except MSDN Resources) , and any news from Walter Oney about a new book project ? Tag: About DMA : Packet Based & Common Buffer Tag: 81018
  - 23
    - Does the stream device have the recived/send buffer?If yes,where does it define? Hi All, When I write the MC8051 program,there is buffer to recive or send the date in the UART communication.And if there is date recived we can read the buffer directly instead of reading the bit one by one according to the timing graph. However,Does the WinCE driver program have the same mechanism as MC8051? If yes ,generally where is the buffer defined?Is it a register? For example,the driver of the AC97 ,...\WINCE420\PUBLIC\COMMON\OAK\DRIVERS\WAVEDEV\UNIFIED\ENSONIQ\ How does the chip read the media file and render?Is it one bit by one bit,or read the buffer? Thanks, Peter Tag: About DMA : Packet Based & Common Buffer Tag: 81014
  - 24
    - TDI Client vs Winsock A few questions on TDI clients that someone may be able to answer: 1. In general, should there be appreciable differences in TDI client vs. Winsock throughput and latency over LAN or wireless links? 2. Does Winsock implement SO_SNDBUF/SO_RECVBUF sizes in afd.sys or in other components? That is does a TDI Client have to reimplement buffering or is it just an option passed to tcpip.sys (like TCP_NODELAY). 3. Can TDI clients issue more than 1 send irp at once? How would that affect throughput/fairness? Thanks. Tag: About DMA : Packet Based & Common Buffer Tag: 81013
  - 25
    - page marked as read-only Hi all, is there any way to mark a piece of data as read-only? This data is previously allocated with ExAllocatePool. The purpose would be to generate a bugcheck when something tries to write unexpectedly to that part of data. Best regards, Calin Tag: About DMA : Packet Based & Common Buffer Tag: 81009

Can you give me som info about them ->

First of all which one is more common ?

Which devices use PacketBased DMA and which devices use
CommonBuffer , and also which use both of them ?

Finally which devices generally have their own DMACs ? (bus-masters)

Top

Re: About DMA : Packet Based & Common Buffer by Eliyas

Eliyas
Tue Jun 20 22:30:17 CDT 2006

Please read this whitepaper on DMA:
http://www.microsoft.com/whdc/driver/kernel/dma.mspx

-Eliyas

"Akýn Öcal" <akin_ocal@hotmail.com> wrote in message
news:ubhkPtLlGHA.4244@TK2MSFTNGP02.phx.gbl...
> Can you give me som info about them ->
>
> First of all which one is more common ?
>
> Which devices use PacketBased DMA and which devices use
> CommonBuffer , and also which use both of them ?
>
> Finally which devices generally have their own DMACs ? (bus-masters)
>
>

Top