Windows exception about amdvlk64.dll when trying to create a vulkan instance - windows

I tried using vulkan, but I can't get it to work. When I try running the first sample given (compiled with VS 2019) with the SDK (01-init_instance.cpp) I get this exception when creating the Vulkan instance:
Exception thrown at 0x00007FFFE7EDAD11 (amdvlk64.dll) in game.exe: 0xC0000005:
Access violation reading location 0xFFFFFFFFFFFFFFFF.
I've tried it with app_info.apiVersion set to VK_API_VERSION_1_0 and VK_API_VERSION_1_1. Also tried setting inst_info.pApplicationInfo to NULL but I don't get any change in the behavior.
I am using an amd gpu AMD Radeon (TM) R9 390 Series, driver version is 17.1.1 and there are some other values about vulkan which are Vulkan™ Driver Version 1.5.0 and Vulkan™ API Version 1.0.39 (all picked from the amd driver interface)
And here is the sample:
#include <iostream>
#include <cstdlib>
#include <util_init.hpp>
#define APP_SHORT_NAME "vulkansamples_instance"
int main(int, char *[]) {
VkApplicationInfo app_info = {};
app_info.sType = VK_STRUCTURE_TYPE_APPLICATION_INFO;
app_info.pNext = NULL;
app_info.pApplicationName = APP_SHORT_NAME;
app_info.applicationVersion = 1;
app_info.pEngineName = APP_SHORT_NAME;
app_info.engineVersion = 1;
app_info.apiVersion = VK_API_VERSION_1_0;
VkInstanceCreateInfo inst_info = {};
inst_info.sType = VK_STRUCTURE_TYPE_INSTANCE_CREATE_INFO;
inst_info.pNext = NULL;
inst_info.flags = 0;
inst_info.pApplicationInfo = &app_info;
inst_info.enabledExtensionCount = 0;
inst_info.ppEnabledExtensionNames = NULL;
inst_info.enabledLayerCount = 0;
inst_info.ppEnabledLayerNames = NULL;
VkInstance inst;
VkResult res;
res = vkCreateInstance(&inst_info, NULL, &inst);
if (res == VK_ERROR_INCOMPATIBLE_DRIVER) {
std::cout << "cannot find a compatible Vulkan ICD\n";
exit(-1);
} else if (res) {
std::cout << "unknown error\n";
exit(-1);
}
vkDestroyInstance(inst, NULL);
return 0;
}
Hopefully someone can help as apparently no one on the internet seems to understand why this happens.

Driver version 17.1.1 is very old (IIRC it means Jan 2017). In ideal world it should work, but as you experience there might be some compatibility issues.
Current drivers are at AMD Support site. They offer "recommended", or more current "optional" driver. Never had any problems with "optional", but it may nag to update more often.

Related

win32 (Desktop Duplication API) IDXGIOutput1::DuplicateOutput() results in Access Violation within dxgi.dll

I'm writing a small c++ app that uses Desktop Duplication API to get the display output. I've never done c programming before, and I got to where I am by staring at the win32 API documentation. https://learn.microsoft.com/en-us/windows/win32/api/dxgi1_2/
#include <iostream>
#pragma comment(lib, "windowsapp")
#include <roapi.h>
//#pragma comment(lib, "dxgi")
#include <dxgi1_2.h>
using namespace std;
int main()
{
cout << RoInitialize(RO_INIT_SINGLETHREADED);
// intermediate variables for casting
IDXGIOutput* pDisplay_old;
IDXGIFactory1* pFactory;
IDXGIAdapter1* pGPU;
IDXGIOutput1* pDisplay;
IDXGIOutputDuplication* pCapture;
DXGI_OUTDUPL_DESC captureDesc;
// create factory
if (CreateDXGIFactory1(__uuidof(IDXGIFactory1), (void**)&pFactory) != S_OK) return 1;
// get GPU
if (pFactory -> EnumAdapters1(0, &pGPU) != S_OK) return 1;
// get display
if (pGPU -> EnumOutputs(0, &pDisplay_old) != S_OK) return 1;
pDisplay = (IDXGIOutput1*)pDisplay_old;
DXGI_OUTDUPL_FRAME_INFO frameInfo;
IDXGIResource* pFrame;
HRESULT captureResult;
do
{
// create capture
// cout << pDisplay -> DuplicateOutput(pGPU, &pCapture);
//return 0;
if (pDisplay -> DuplicateOutput(pGPU, &pCapture) != S_OK) return 1;
pCapture -> GetDesc(&captureDesc);
cout << captureDesc.ModeDesc.Width << ' ' << captureDesc.ModeDesc.Height;
do
{
captureResult = pCapture -> AcquireNextFrame(2000, &frameInfo, &pFrame);
if (captureResult == S_OK)
{
cout << "HI";
captureResult = pCapture -> ReleaseFrame();
}
else if (captureResult == DXGI_ERROR_ACCESS_LOST) break;
else return 1;
}
while (true);
}
while (true);
}
I'm using visual studio 2022 witn only "desktop development with c++" enabled, on windows 11 insider build: 22623.1037 ni_release on a regular home PC with display, beyboard, mouse, etc
The code worked fine until DuplicateOutput(), when it complained E_NOINTERFACE. I'm certain there is an interface since index 0 for EnumAdapters1 and EnumOutputs are where the desktop is displayed, and I obviously have a display attached with the desktop. According to this guy https://devblogs.microsoft.com/oldnewthing/20041213-00/?p=37043, I need marshalling and apartments or something, so after more research, I tried RoInitialize() with both RO_INIT_SINGLETHREADED and RO_INIT_MULTITHREADED. Now, DuplicateOutput throws this exception
It seems to happen within the library itself, which makes me think that it's either not my fault or I really messed something up, probably the latter.
I'm really confused now, and would like some assistance, thanks!
EDIT: I replaced "pDisplay = (IDXGIOutput1*)pDisplay_old;" with "pDisplay_old -> QueryInterface(&pDisplay);", and I'm back to E_NOINTERFACE, but I think I'm on the right track, how do I fix this error?
EDIT2: I looked at a related question AcquireNextFrame not working (Desktop Duplication API & D3D11), and followed the answer to add D3D11CreateDevice to my code:
#include <iostream>
#pragma comment(lib, "dxgi")
#pragma comment(lib, "d3d11")
#include <d3d11.h>
#include <dxgi1_2.h>
using namespace std;
int main()
{
// intermediate variables for casting
IDXGIOutput* pDisplay_old;
IDXGIFactory1* pFactory;
IDXGIAdapter* pGPU;
ID3D11Device* pD3DDevice;
IDXGIDevice* pDevice;
IDXGIOutput1* pDisplay;
IDXGIOutputDuplication* pCapture;
DXGI_OUTDUPL_DESC captureDesc;
// create DXGI factory
if (CreateDXGIFactory1(__uuidof(IDXGIFactory1), (void**)&pFactory) != S_OK) return 1;
// get GPU adapter
if (pFactory -> EnumAdapters(0, &pGPU) != S_OK) return 2;
// create D3D11 device
D3D_FEATURE_LEVEL D3DFeatures [6]
{
D3D_FEATURE_LEVEL_11_0,
D3D_FEATURE_LEVEL_10_1,
D3D_FEATURE_LEVEL_10_0,
D3D_FEATURE_LEVEL_9_3,
D3D_FEATURE_LEVEL_9_2,
D3D_FEATURE_LEVEL_9_1
};
cout << D3D11CreateDevice(pGPU, D3D_DRIVER_TYPE_HARDWARE, nullptr, 0, D3DFeatures, sizeof(D3DFeatures), D3D11_SDK_VERSION, &pD3DDevice, NULL, NULL); //!= S_OK) return 3;
return 0;
// get DXGI device from that
pD3DDevice -> QueryInterface(&pDevice);
// get display
if (pGPU -> EnumOutputs(0, &pDisplay_old) != S_OK) return 4;
pDisplay_old -> QueryInterface(&pDisplay);
DXGI_OUTDUPL_FRAME_INFO frameInfo;
IDXGIResource* pFrame;
HRESULT captureResult;
do
{
// create capture
cout << pDisplay -> DuplicateOutput(pD3DDevice, &pCapture);
return 0;
if (pDisplay -> DuplicateOutput(pGPU, &pCapture) != S_OK) return 5;
pCapture -> GetDesc(&captureDesc);
cout << captureDesc.ModeDesc.Width << ' ' << captureDesc.ModeDesc.Height;
do
{
captureResult = pCapture -> AcquireNextFrame(2000, &frameInfo, &pFrame);
if (captureResult == S_OK)
{
cout << "HI";
captureResult = pCapture -> ReleaseFrame();
}
else if (captureResult == DXGI_ERROR_ACCESS_LOST) break;
else return 6;
}
while (true);
}
while (true);
D3D11CreateDevice seems like a complex function and for me it keeps complaining invalid_arg. I'm not sure how to fix that
The solution was provided in the comments:
pDisplay = (IDXGIOutput1*)pDisplay_old; is wrong, you must always use QueryInterface to get an interface from another. And you don't need RoInitialize. –
Simon Mourier Jan 1 at 9:11
I replaced it with "pDisplay_old -> QueryInterface(&pDisplay);", and I'm back to E_NOINTERFACE, but I think I'm on the right track, how do I fix this error? –
Tiger Yang Jan 4 at 16:54
I don't get E_NOINTERFACE (you shouldn't) on this QueryInterface call. What is wrong then is DuplicateOutput expects a Direct3D device, not an adapter interface reference. –
Simon Mourier Jan 4 at 17:18
I've worked on it and updated the post above –
Tiger Yang Jan 7 at 1:51
Your code is wrong again, use D3D_DRIVER_TYPE_UNKNOWN if you pass an adapter as 1st arg (or ask for hardware and pass nullptr as 1st arg) and use ARRAYSIZE(D3DFeatures), not sizeof(D3DFeatures) as 6th arg. Use DirectX Debug Layer learn.microsoft.com/en-us/windows/win32/direct3d11/… walbourn.github.io/direct3d-sdk-debug-layer-tricks to ease debugging –
Simon Mourier Jan 7 at 7:36

Has there been any change between kernel 5.15 and 5.4.0 concerning ioctl valid commands?

We have some custom driver working on 5.4.0. It's pretty old and the original developers are no longer supporting it, so we have to maintain it in our systems.
When upgrading to Ubuntu 22 (Kernel 5.15), the driver suddenly stopped working, and sending ioctl with the command SIOCDEVPRIVATE (which used to work in kernel 5.4.0, and in fact is used to get some necessary device information)now gives "ioctl: Operation not supported" error with no extra information anywhere on the logs.
So... has something changed between those two kernels? We did have to adapt some of the structures used to register the driver, but I can't see anything concerning registering valid operations there. Do I have to register valid operations somewhere now?
Alternatively, does somebody know what part of the kernel code is checking for the operation to be supported? I've been trying to find it from ioctl.c, but I can't seem to find where that particular error comes from.
The driver code that supposedly takes care of this (doesn't even reach first line on 5.15):
static int u50_dev_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) {
struct u50_priv *priv = netdev_priv(dev);
if (cmd == SIOCDEVPRIVATE) {
memcpy(&ifr->ifr_data, priv->tty->name, strlen(priv->tty->name));
}
return 0;
}
And the attempt to access it that does no longer work:
struct ifreq ifr = {0};
struct ifaddrs *ifaddr, *ifa;
getifaddrs(&ifaddr);
for (ifa = ifaddr; ifa != NULL; ifa = ifa->ifa_next) {
memcpy(ifr.ifr_name, ifa->ifa_name, IFNAMSIZ);
if (ioctl(lonsd, SIOCDEVPRIVATE, &ifr) < 0) {
perror("ioctl");
syslog(LOG_ERR, "Ioctl:%d: %s\n", __LINE__, strerror(errno));
}
...
and structure for registration
static const struct net_device_ops u50_netdev_ops = {
.ndo_init = u50_dev_init,
.ndo_uninit = u50_dev_uninit,
.ndo_open = u50_dev_open,
.ndo_stop = u50_dev_stop,
.ndo_start_xmit = u50_dev_xmit,
.ndo_do_ioctl = u50_dev_ioctl,
.ndo_set_mac_address = U50SetHWAddr,
};
If you need some code to respond to SIOCDEVPRIVATE, you used to be able to do it via ndo_do_ioctl (writing a compatible function, then linking it in a net_device_ops struct in 5.4). However, in 5.15 it was changed so now you have to implement a ndo_siocdevprivate function, rather than ndo_do_ioctl, which is no longer called, according to the kernel documentation.
source:
https://elixir.bootlin.com/linux/v5.15.57/source/include/linux/netdevice.h
Patch that did this: spinics.net/lists/netdev/msg698158.html

Install winusb driver from user application

Please can anyone give me direct for realization next functional for Windows.
I have USB device which connects to the PC (it is JTAG programmer.) I know VID and PID of this hardware. I need:
1 Check what type of driver this hardware use (detecting winusb driver or not will be enough. Maybe do I need to read registry?)
2 If driver is not winusb I need to install winusb driver for this USB device from my application.
The current driver assigned to the device is stored in the registry, so you could read it directly from there. However, it is probably better to use SetupAPI, an API provided by Microsoft. The function to call is SetupDiGetDeviceRegistryProperty, and the third argument should be SPDRP_SERVICE. This will return the name of the driver as a string. Note that you will need to call several other SetupAPI functions before you have all the pieces of info you need to call SetupDiGetDeviceRegistryProperty.
I have not tried it, but libwdi has features for installing WinUSB onto a device node. It might also have functions for getting the current driver, so you should try using it before you spend too much time learning SetupAPI. The devcon utility from Microsoft (which is open source now) might be another option.
Without knowing the details of what you are doing, I question whether you really need to do this. It might be simpler to provide a signed driver package to users and instruct them to use the "Update Driver Software..." option from the Device Manager to apply it to particular device.
I made first part of task.
#ifdef Q_OS_WIN
DEFINE_GUID(GUID_DEVCLASS_WINUSB,0x88BAE032,0x5A81,0x49f0,
0xBC,0x3D,0xA4,0xFF,0x13,0x82,0x16,0xD6);
#endif
bool WinUSB::isWinUsbDriver(quint16 vid, quint16 pid)
{
#ifndef Q_OS_WIN
Q_UNUSED(vid);
Q_UNUSED(pid);
return true;
#else
HDEVINFO deviceInfoSet;
GUID *guidDev = (GUID*) &GUID_DEVCLASS_WINUSB;
deviceInfoSet = SetupDiGetClassDevs(guidDev, NULL, NULL, DIGCF_PRESENT | DIGCF_PROFILE);
DWORD buffersize =4000;
TCHAR buffer [buffersize];
int memberIndex = 0;
bool retval = false;
QString vidPid;
vidPid = "VID_" + QString("%1").arg(vid,4,16,QChar('0')) + "&";
vidPid += "PID_" + QString("%1").arg(pid,4,16,QChar('0'));
while (true)
{
SP_DEVINFO_DATA deviceInfoData;
ZeroMemory(&deviceInfoData, sizeof(SP_DEVINFO_DATA));
deviceInfoData.cbSize = sizeof(SP_DEVINFO_DATA);
if (SetupDiEnumDeviceInfo(deviceInfoSet, memberIndex, &deviceInfoData) == FALSE) {
if (GetLastError() == ERROR_NO_MORE_ITEMS) {
break;
}
}
DWORD nSize=0 ;
SetupDiGetDeviceInstanceId (deviceInfoSet, &deviceInfoData, buffer, sizeof(buffer), &nSize);
buffer [nSize] ='\0';
QString str = QString::fromWCharArray(buffer);
if (str.indexOf(vidPid) >= 0) {
retval = true;
break;
}
memberIndex++;
}
if (deviceInfoSet) {
SetupDiDestroyDeviceInfoList(deviceInfoSet);
}
return retval;
#endif
}

GetOpenFileName fails in 64 bit, but works in 32Bit?

I have the following code, I use to Open a File Open Dialog using Win32 API. It works fine in 32bit, but fails when I use in a 64bit (In a DLL). What am I doing wrong?
char Filestring[256];
Filter = "OBJ files\0*.obj\0\0";
char* returnstring = NULL;
OPENFILENAME opf;
opf.hwndOwner = mainHWND;
opf.lpstrFilter = Filter;
opf.lpstrCustomFilter = 0;
opf.nMaxCustFilter = 0L;
opf.nFilterIndex = 1L;
opf.lpstrFile = Filestring;
opf.lpstrFile[0] = '\0';
opf.nMaxFile = 256;
opf.lpstrFileTitle = 0;
opf.nMaxFileTitle=50;
opf.lpstrInitialDir = Path;
opf.lpstrTitle = "Open Obj File";
opf.nFileOffset = 0;
opf.nFileExtension = 0;
opf.lpstrDefExt = "*.*";
opf.lpfnHook = NULL;
opf.lCustData = 0;
opf.Flags = (OFN_PATHMUSTEXIST | OFN_OVERWRITEPROMPT) & ~OFN_ALLOWMULTISELECT;
opf.lStructSize = sizeof(OPENFILENAME);
if(GetOpenFileName(&opf))
{
returnstring = opf.lpstrFile;
if (returnstring) {
result = returnstring;
}
}
EDIT: By failing, I meant that the Open File Dialog doesn't show up. The code still returns zero without any errors.
EDIT 2: I have called CommDlgExtendedError() and it returned 1. From the MSDN reference, does it mean the dialog has invalid lStructSize? I have checked the sizeof(OPENFILENAME) and it returned 140 bytes.
UPDATE: In my Project Settings, Under Code Generation the "Struct Member Alignment" is set to 4 Bytes(/Zp4). I changed this to default and it magically worked. Look for the answers and their comments below for more information.
You aren't initialising lpTemplateName and so it contains random stack noise. This in turn will lead to 'hInstance` being references which also contains stack noise.
When calling a function like this you should first of all zero out the struct and only fill in the fields that are non-zero. Something like this:
OPENFILENAME opf={0};
opf.lStructSize = sizeof(OPENFILENAME);
opf.hwndOwner = mainHWND;
opf.lpstrFilter = Filter;
opf.nFilterIndex = 1L;
opf.lpstrFile = Filestring;
opf.lpstrFile[0] = '\0';
opf.nMaxFile = 256;
opf.lpstrInitialDir = Path;
opf.lpstrTitle = "Open Obj File";
opf.lpstrDefExt = "*.*";
opf.Flags = OFN_PATHMUSTEXIST | OFN_OVERWRITEPROMPT;
There was no need to exclude OFN_ALLOWMULTISELECT explicitly since you were not including it in the first place!
EDIT
You state in a comment that this doesn't work. Calling CommDlgExtendedError is a good idea and should tell you why it fails.
You could also try to run the minimal possible GetOpenFileName which is this:
char Filestring[MAX_PATH] = "\0";
OPENFILENAME opf={0};
opf.lStructSize = sizeof(OPENFILENAME);
opf.lpstrFile = Filestring;
opf.nMaxFile = MAX_PATH;
GetOpenFileName(&opf);
I have the very same problem and a partial solution :
+ the simple following simple example (proposed abobe) was not working in x64 mode.
+ I changed the complie option "struct Member Alignment" from 1byte /Zp1 to default which solved this problem (by introducing others !!!)
char Filestring[MAX_PATH] = "\0";
OPENFILENAME opf={0};
opf.lStructSize = sizeof(OPENFILENAME);
opf.lpstrFile = Filestring;
opf.nMaxFile = MAX_PATH;
GetOpenFileName(&opf);
To find out more you should call CommDlgExtendedError to get the error code what went wrong. Besides this I would initialize all member of the struct to 0 with
ZeroMemory(&opf, sizeof(opf));
Since the file open dialog is in reality a COM component it could be worth to check out if your thread apartment state is different under 64 bit.
if( RPC_E_CHANGED_MODE == CoInitialize(NULL) )
ASSERT(FALSE); // MTA Apartment found
CoUnitialize()
Yours,
Alois Kraus
As a note in Microsoft Office 2010 64-bit we gave up and used the internal wrappers as the structure turned into 140 bytes and we were not sure how to change alignment.
Application.GetOpenFilename(FileFilter, FilterIndex, Title, ButtonText, MultiSelect)
and Application.GetSaveAsFilename(InitialFilename, FileFilter, FilterIndex, Title, ButtonText)
http://msdn.microsoft.com/en-us/library/ff834966.aspx
http://msdn.microsoft.com/en-us/library/microsoft.office.interop.excel._application.getopenfilename.aspx
Needless to say we think all individuals with fairly heavy applications in Excel should start considering other options as maintaining future versions across multiple clients and platforms may just be... insane!
I managed to get around this problem by setting the packing appropriately before including the header file. That way, for the purpose of this one function, we were using the 'default' 16 byte alignment, but did not have to change the packing alignment for the rest of our program:
#ifdef _WIN64
#pragma pack( push )
#pragma pack( 16 )
#include "Commdlg.h"
#pragma pack( pop )
#else
#include "Commdlg.h"
#endif // _WIN64

Sending a specific SCSI command to a SCSI device in Windows

Does windows have specific interface by which I can send a specific scsi command such Inquiry to scsi device ? I searched the net, found passing reference to SCSI Pass Through interface. But Its very Vague.
Is there any documentation for that API on how to use it??
#include <iostream>
#include <windows.h>
#include <winioctl.h>
#define ULONG_PTR ULONG
#include <ntddscsi.h> //from SDK
#include <spti.h> //from DDK
using namespace std;
int demo()
{
HANDLE hDisk;
SCSI_PASS_THROUGH_DIRECT_WITH_BUFFER sptdwb;
ULONG length = 0;
DWORD bytesReturn;
BYTE bufDataRead[64*1024+10];
int iRet;
hDisk = CreateFile(path,GENERIC_READ | GENERIC_WRITE,
FILE_SHARE_READ | FILE_SHARE_WRITE, NULL,OPEN_EXISTING,0,NULL
);
if (hDisk ==INVALID_HANDLE_VALUE) {
return 0;
}
ZeroMemory(&sptdwb, sizeof(SCSI_PASS_THROUGH_DIRECT_WITH_BUFFER));
sptdwb.sptd.Length = sizeof(SCSI_PASS_THROUGH_DIRECT);
sptdwb.sptd.PathId = 0;
sptdwb.sptd.TargetId = 1;
sptdwb.sptd.Lun = 0;
sptdwb.sptd.CdbLength = 6;
sptdwb.sptd.DataIn = SCSI_IOCTL_DATA_IN;
sptdwb.sptd.SenseInfoLength = 24;
sptdwb.sptd.DataTransferLength = 8;
sptdwb.sptd.TimeOutValue = 2;
sptdwb.sptd.DataBuffer = bufDataRead;
sptdwb.sptd.SenseInfoOffset = offsetof(SCSI_PASS_THROUGH_DIRECT_WITH_BUFFER,ucSenseBuf);
sptdwb.sptd.Cdb[0] = 0x12;
sptdwb.sptd.Cdb[1] = 0x00;
sptdwb.sptd.Cdb[2] = 0x00;
sptdwb.sptd.Cdb[3] = 0x00;
sptdwb.sptd.Cdb[4] = 0xFF;
sptdwb.sptd.Cdb[5] = 0x00;
length = sizeof(SCSI_PASS_THROUGH_DIRECT_WITH_BUFFER);
iRet = DeviceIoControl(hDisk,
IOCTL_SCSI_PASS_THROUGH_DIRECT,
&sptdwb,
length,
&sptdwb,
length,
&bytesReturn,
NULL);
if (0 == iRet) {
printf("inquiry fail");
return 0;
} else {
//Check returned data in sptdwb.sptd.DataBuffer.
}
return 0;
}
SCSI covers a vast amount of ground. Are you talking to a CD/DVD/Disk/Tape/Scanner or what.
For CD/DVD the best (and only) free reference for setup/read/write commands are to be found here: http://www.t10.org/drafts.htm
Re SPTI, there is some very basic documentation in the old 'Programmers guide to SCSI'. There is an article on an ASPI -> SPTI converter that can be found on the DDJ web site.
Bear in mind that SPTI is simply an API, it imposes nor knows anything about SCSI message content or format.
Brian Sawert, Addison Wesley 1998.
You can send SCSI commands to the SCSI Port driver by sending it an IRP_MJ_SCSI IRP, see http://msdn.microsoft.com/en-us/library/ff565387(VS.85).aspx. However, you will have to construct the SCSI CBD yourself, and I have yet to find a document which describes it.
Anymore, the SCSI commands are broken down into a number of specs. The INQUIRY command is in the SPC spec, while device type specific commands are broken down into several specs (i.e. block, ses, ...).

Resources