My device driver for a PCIe FPGA is based on 7600.16385.1\src\general\PLX9x5x
Upon ReadFile in the application, PLxEvtIoRead is called:
//
// Initialize this new DmaTransaction.
//
status = WdfDmaTransactionInitializeUsingRequest(
devExt->ReadDmaTransaction,
Request,
PLxEvtProgramReadDma,
WdfDmaDirectionReadFromDevice );
//
// Execute this DmaTransaction.
//
status = WdfDmaTransactionExecute( devExt->ReadDmaTransaction,
WDF_NO_CONTEXT);
....
Upon calling to WdfDmaTransactionExecute, PLxEvtProgramReadDma is called.
BOOLEAN
PLxEvtProgramReadDma(
IN WDFDMATRANSACTION Transaction,
IN WDFDEVICE Device,
IN WDFCONTEXT Context,
IN WDF_DMA_DIRECTION Direction,
IN PSCATTER_GATHER_LIST SgList
)
{
KdPrint ((???SgList->NumberOfElements = %d\n???,SgList->NumberOfElements));
}
The problem:
i want to transfer a large amount of data via this Scatter/Gather list(around 1 GB), but it seems NumberOfElements is limited by something, somehow that the larges transmition is 1MB(255 element in list, each 4k). i changed MaximumTransfecrLength in function below to 500MB:
WDF_DMA_ENABLER_CONFIG_INIT(&dmaConfig,
WdfDmaProfileScatterGatherDuplex,
deviceContext->MaximumTransferLength);
but still i can not transfer more than 1MB.
what is the thing that limits NumberOfElements and how i can solve it?
I needed to change the second parameter in WDF_DMA_ENABLER_CONFIG_INIT function to WdfDmaProfileScatterGather64, and of course we have to make sure that hardware(FPGA or anything in other side of PCIE endpoint) can support 64-bit addressing mode.
I just change my code as below:
WDF_DMA_ENABLER_CONFIG_INIT(&dmaConfig,
WdfDmaProfileScatterGather64,
deviceContext->MaximumTransferLength);
Related
Assuming a scenario where DX12 is being hooked for overlay rendering it looks like the best function to hook is the IDXGISwapChain::Present the same way it was done for DX11. Having this function hooked the swap chain is available and from that the device can be retrieved to create resources. Given those resources it’s possible to record rendering commands too. The problem arises when we are trying to execute the rendering commands as there is no option to retrieve the associated command queue from the swap chain so there is nothing like this:
CComPtr<ID3D12Device> pD3D12Device;
if (pSwapChain->GetDevice(__uuidof(ID3D12Device), (void**)(&pD3D12Device)) == S_OK)
{
pD3D12Device->GetCommandQueueForSwapChain( swapChain )->ExecuteCommandLists(…);
}
The other option would be creating a new command queue to execute on, like this:
CComPtr<ID3D12Device> pD3D12Device;
if (pSwapChain->GetDevice(__uuidof(ID3D12Device), (void**)(&pD3D12Device)) == S_OK)
{
D3D12_COMMAND_QUEUE_DESC queue_desc = {};
queue_desc.Flags = D3D12_COMMAND_QUEUE_FLAG_NONE;
queue_desc.Type = D3D12_COMMAND_LIST_TYPE_DIRECT;
HRESULT commandQueueRes = _device->CreateCommandQueue( &queue_desc, IID_PPV_ARGS( &_commandQueue ) );
_commandQueue->ExecuteCommandLists( ... );
}
This results in an error and subsequent device removal. See the error message below.
D3D12 ERROR: ID3D12CommandQueue::ExecuteCommandLists: A command list, which writes to a swap chain back buffer, may only be executed on the command queue associated with that buffer. [ STATE_SETTING ERROR #907: EXECUTECOMMANDLISTS_WRONGSWAPCHAINBUFFERREFERENCE]
The problem is not resolved even if the ID3D12CommandQueue::ExecuteCommandLists is hooked as well because there is no way to retrieve the associated swap chain from the command queue either.
So my question is what is the recommended way to deal with this problem in a scenario where the swap chain creation happens before the hooking could possibly happen?
In case anyone is looking for the answer here's what I found out.
There is no official way to do this, for overlay rendering the recommended way is to use DirectComposition but this has performance consequences which is not very nice for game overlays.
Investigating the memory a bit there is a possible solution to get the CommandQueue from the swap chain with something like this:
#ifdef _M_X64
size_t* pOffset = (size_t*)((BYTE*)swapChain + 216);
#else
size_t* pOffset = (size_t*)((BYTE*)swapChain + 132);
#endif
*(&_commandQueue) = reinterpret_cast<ID3D12CommandQueue*>(*pOffset);
Obviously this solution is not recommended but it might be useful if someone just want's to do some debugging.
My final solution is to hook into a function that uses the CommandQueue (I use ExecuteCommandLists) and get the pointer there and use it later to render the overlay. It's not completely satisfying but it works as long as there are no multiple swap chains.
So this is what I ended up with to get realtime starring/liking (of communities, in my case) working, with a Firebase datastore. It's a mess and surely I'm missing some fundamentals.
Here my element gets communities, each as a Map community stored in an observed List communities. It has to rewrite that List several times as it changes each community Map based on the results of the changed star count and the user's starred state, and some other fun:
getCommunities() {
// Since we call this method a second time after user
// signed in, clear the communities list before we recreate it.
if (communities.length > 0) { communities.clear(); }
var firebaseRoot = new db.Firebase(firebaseLocation);
var communityRef = firebaseRoot.child('/communities');
// TODO: Undo the limit of 20; https://github.com/firebase/firebase-dart/issues/8
communityRef.limit(20).onChildAdded.listen((e) {
var community = e.snapshot.val();
// snapshot.name is Firebase's ID, i.e. "the name of the Firebase location",
// so we'll add that to our local item list.
community['id'] = e.snapshot.name();
print(community['id']);
// If the user is signed in, see if they've starred this community.
if (app.user != null) {
firebaseRoot.child('/users/' + app.user.username + '/communities/' + community['id']).onValue.listen((e) {
if (e.snapshot.val() == null) {
community['userStarred'] = false;
// TODO: Add community star_count?!
} else {
community['userStarred'] = true;
}
print("${community['userStarred']}, star count: ${community['star_count']}");
// Replace the community in the observed list w/ our updated copy.
communities
..removeWhere((oldItem) => oldItem['alias'] == community['alias'])
..add(community)
..sort((m1, m2) => m1["updatedDate"].compareTo(m2["updatedDate"]));
communities = toObservable(communities.reversed.toList());
});
}
// If no updated date, use the created date.
if (community['updatedDate'] == null) {
community['updatedDate'] = community['createdDate'];
}
// Handle the case where no star count yet.
if (community['star_count'] == null) {
community['star_count'] = 0;
}
// The live-date-time element needs parsed dates.
community['updatedDate'] = DateTime.parse(community['updatedDate']);
community['createdDate'] = DateTime.parse(community['createdDate']);
// Listen for realtime changes to the star count.
communityRef.child(community['alias'] + '/star_count').onValue.listen((e) {
int newCount = e.snapshot.val();
community['star_count'] = newCount;
// Replace the community in the observed list w/ our updated copy.
// TODO: Re-writing the list each time is ridiculous!
communities
..removeWhere((oldItem) => oldItem['alias'] == community['alias'])
..add(community)
..sort((m1, m2) => m1["updatedDate"].compareTo(m2["updatedDate"]));
communities = toObservable(communities.reversed.toList());
});
// Insert each new community into the list.
communities.add(community);
// Sort the list by the item's updatedDate, then reverse it.
communities.sort((m1, m2) => m1["updatedDate"].compareTo(m2["updatedDate"]));
communities = toObservable(communities.reversed.toList());
});
}
Here we toggle the star, which again replaces the observed communities List a few times as we update the count in the affected community Maps and thus rewrite the List to reflect that:
toggleStar(Event e, var detail, Element target) {
// Don't fire the core-item's on-click, just the icon's.
e.stopPropagation();
if (app.user == null) {
app.showMessage("Kindly sign in first.", "important");
return;
}
bool isStarred = (target.classes.contains("selected"));
var community = communities.firstWhere((i) => i['id'] == target.dataset['id']);
var firebaseRoot = new db.Firebase(firebaseLocation);
var starredCommunityRef = firebaseRoot.child('/users/' + app.user.username + '/communities/' + community['id']);
var communityRef = firebaseRoot.child('/communities/' + community['id']);
if (isStarred) {
// If it's starred, time to unstar it.
community['userStarred'] = false;
starredCommunityRef.remove();
// Update the star count.
communityRef.child('/star_count').transaction((currentCount) {
if (currentCount == null || currentCount == 0) {
community['star_count'] = 0;
return 0;
} else {
community['star_count'] = currentCount - 1;
return currentCount - 1;
}
});
// Update the list of users who starred.
communityRef.child('/star_users/' + app.user.username).remove();
} else {
// If it's not starred, time to star it.
community['userStarred'] = true;
starredCommunityRef.set(true);
// Update the star count.
communityRef.child('/star_count').transaction((currentCount) {
if (currentCount == null || currentCount == 0) {
community['star_count'] = 1;
return 1;
} else {
community['star_count'] = currentCount + 1;
return currentCount + 1;
}
});
// Update the list of users who starred.
communityRef.child('/star_users/' + app.user.username).set(true);
}
// Replace the community in the observed list w/ our updated copy.
communities.removeWhere((oldItem) => oldItem['alias'] == community['alias']);
communities.add(community);
communities.sort((m1, m2) => m1["updatedDate"].compareTo(m2["updatedDate"]));
communities = toObservable(communities.reversed.toList());
print(communities);
}
There's also some other craziness where we have to get the list of communities again when app.changes because we only load app.user after the app and list initially load, and now that we have the user we need to turn on the appropriate stars. So my attached() looks like:
attached() {
app.pageTitle = "Communities";
getCommunities();
app.changes.listen((List<ChangeRecord> records) {
if (app.user != null) {
getCommunities();
}
});
}
There, it seems I could just be getting the stars and updating said each affected community Map, then repopulating the observed communities List, but that's the least of it.
The full thing: https://gist.github.com/DaveNotik/5ccdc9e74429cf87d641
How can I improve all this Map/List management, e.g. where every time I change a community Map, I have to rewrite the whole communities List? Should I be thinking of it differently?
What about all this querying Firebase? Surely, there's a better way, but it seems I need to do a lot to keep it realtime, and also the element gets attached and detached, so it seems I need to run getCommunities() each time. Unless the OOP way is objects get created, and they're always there to be observed whenever the element is attached? I'm missing those fundamentals.
This app.changes business to handle the case where we load the list before we have the app.user (which then means we want to load her stars) - is there a better way?
Other ridiculousness?
Big question, I know. Thank you for helping me get a handle on the right approach as I move forward!
I think there is two different ways to choose, if you want to keep a data of your application in real time sync with server database:
1 Polling (pull method ie. a client pulls the data from server)
Application polls ie. requests the updated data from the server. Polling can be automatic (for example with interval of 60s) or requested by user (= refresh). The short automatic interval will cause high load on server and with long interval you lose real time feeling.
2 Full-duplex (push method ie. server can push the data to the client)
An application and a server have full-duplex connection in between and server is able to send the data or a notification of the data available to the client. Then the client can decide whether or not to retrieve the data.
This method is a modern one, because it'll keep the net traffic and the server load in minimum and yet providing a real time updates.
The firebase boasts with this kind of updates, but I'm not sure is it full-dublex or just a clever way of polling. Websocket protocol is a real full-duplex connection and dart server supports it.
The updated data from a server can include:
1 A full dataset
Basically the server sends a full dataset (=initial query) and the server doesn't "know" anything about updated data. This is easiest way to go, if you have reasonable small datasets. Many times you'll have a very small datasets among the big ones, so this way can be useful.
2 A dataset including a new data only
The server can send a dataset based on modified timestamp ie. every time a record in the database changes, a timestamp for update will be saved and the query can be filtered based on this timestamp. In other words application knows when it has last updated the data and then requests newer data.
3 A changed record
A server keeps a track of updated data and sends it to the application. The data can be sent record by record when changes occurs or server can collect the data for a bigger chunks to be sent. This method requires a server to keep a track of every client connected in order to send a correct data to each client. When you add an authentication process for clients ie. not every data can be send to all, it can get quite complicated.
I think the easiest way is to use the method number 2 for updated data.
Last thing...
What to do with the data received?
1 Handle everything as a new
If application receives an updated data, it will destroy/clear all the lists and maps and recreate/refill them with the new data. Typical problems with this are a user loses a current position on a page or the data user were looking jumps around. If application has modified or extended an old data for some reason, all those modifications will be lost. This method works ok, if a user requests a refresh.
2 Update only the changed data
The application never clears initial list or maps, it just updates them with a new received data. Typically you will construct a new combined map from queried data for your specific need (for example a certain view). The combined map has already all information you want to show in the specific view (default values even if the initial queries didn't had the data for the field) and you just update a new values in it.
If the updated information needs a new member in the list you just add it in the end.
If the updated information requires a deletion from the list, it might be a good thing to use extra field "active" and filter the list/map with it. With filtering you won't lose any referencies or so.
If you need to sort a data or filter it, it should be done by a view or user request. Basically the data is stored in the application and updated as needed. When a user needs to see the data in a specific way, the view should show the data a proper way. This is called model-controller-view and the main idea is to separate the data from the view.
I'm sorry this long answer didn't answer any of your questions, but I tried to cut this challenge to a smaller chunks. Many times you can see an interface between these chunks and you can design and organize your code nicely by using these interfaces.
Basically, I just want those three things.
Color, Depth, and Skeleton. But I also want the depth and color lined up.
I know that in the official examples, there's GreenScreen.cpp, which is an example of depth/color cut-outs, but no skeleton. There's also SkeletonBasics which has a skeleton, but no image.
I've tried learning from both of them and implementing them together, but I never get a callback for when the skeleton is ready (this: m_hNextSkeletonEvent).
SkeletonBasics.cpp uses this:
hEvents[0] = m_hNextSkeletonEvent;
// Check to see if we have either a message (by passing in QS_ALLEVENTS)
// Or a Kinect event (hEvents)
// Update() will check for Kinect events individually, in case more than one are signalled
DWORD dwEvent = MsgWaitForMultipleObjects(eventCount, hEvents, FALSE, INFINITE, QS_ALLINPUT);
// Check if this is an event we're waiting on and not a timeout or message
if (WAIT_OBJECT_0 == dwEvent)
{
Update();
}
And GreenScreen.cpp uses this:
hEvents[0] = m_hNextDepthFrameEvent;
hEvents[1] = m_hNextColorFrameEvent;
// Check to see if we have either a message (by passing in QS_ALLINPUT)
// Or a Kinect event (hEvents)
// Update() will check for Kinect events individually, in case more than one are signalled
DWORD dwEvent = MsgWaitForMultipleObjects(eventCount, hEvents, FALSE, INFINITE, QS_ALLINPUT);
// Check if this is an event we're waiting on and not a timeout or message
if (WAIT_OBJECT_0 == dwEvent || WAIT_OBJECT_0 + 1 == dwEvent)
{
Update();
}
But, like I said, combining them yields no
if ( WAIT_OBJECT_0 == WaitForSingleObject(m_hNextSkeletonEvent, 0) )
happening. (The KinectBridgeWithOpenCVBasics-D2D also does Skeleton, Color, and Depth, but still is unaligned (like this example I found in a search).
Do you know why the event is not firing? Or perhaps do you know of an example somewhere that has all of these abilities? I've searched high and low, and I've tried tons of different things messing with these examples, but I just cant seem to get it.
Thanks for your help.
EDIT:
At the end of KinectBridgeWithOpenCVBasics, I'm left with a m_colorMat and an m_depthMat. Even if at that point I could align the depth to color, I think that would work out just fine.
I want to use amazon Dynamo DB with rails.But I have not found a way to implement pagination.
I will use AWS::Record::HashModel as ORM.
This ORM supports limits like this:
People.limit(10).each {|person| ... }
But I could not figured out how to implement following MySql query in Dynamo DB.
SELECT *
FROM `People`
LIMIT 1 , 30
You issue queries using LIMIT. If the subset returned does not contain the full table, a LastEvaluatedKey value is returned. You use this value as the ExclusiveStartKey in the next query. And so on...
From the DynamoDB Developer Guide.
You can provide 'page-size' in you query to set the result set size.
The response of DynamoDB contains 'LastEvaluatedKey' which will indicate the last key as per the page size. If response does't contain 'LastEvaluatedKey' it means there are no results left to fetch.
Use the 'LastEvaluatedKey' as 'ExclusiveStartKey' while fetching next time.
I hope this helps.
DynamoDB Pagination
Here's a simple copy-paste-run proof of concept (Node.js) for stateless forward/reverse navigation with dynamodb. In summary; each response includes the navigation history, allowing user to explicitly and consistently request either the next or previous page (while next/prev params exist):
GET /accounts -> first page
GET /accounts?next=A3r0ijKJ8 -> next page
GET /accounts?prev=R4tY69kUI -> previous page
Considerations:
If your ids are large and/or users might do a lot of navigation, then the potential size of the next/prev params might become too large.
Yes you do have to store the entire reverse path - if you only store the previous page marker (per some other answers) you will only be able to go back one page.
It won't handle changing pageSize midway, consider baking pageSize into the next/prev value.
base64 encode the next/prev values, and you could also encrypt.
Scans are inefficient, while this suited my current requirement it won't suit all!
// demo.js
const mockTable = [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20]
const getPagedItems = (pageSize = 5, cursor = {}) => {
// Parse cursor
const keys = cursor.next || cursor.prev || [] // fwd first
let key = keys[keys.length-1] || null // eg ddb's PK
// Mock query (mimic dynamodb response)
const Items = mockTable.slice(parseInt(key) || 0, pageSize+key)
const LastEvaluatedKey = Items[Items.length-1] < mockTable.length
? Items[Items.length-1] : null
// Build response
const res = {items:Items}
if (keys.length > 0) // add reverse nav keys (if any)
res.prev = keys.slice(0, keys.length-1)
if (LastEvaluatedKey) // add forward nav keys (if any)
res.next = [...keys, LastEvaluatedKey]
return res
}
// Run test ------------------------------------
const runTest = () => {
const PAGE_SIZE = 6
let x = {}, i = 0
// Page to end
while (i == 0 || x.next) {
x = getPagedItems(PAGE_SIZE, {next:x.next})
console.log(`Page ${++i}: `, x.items)
}
// Page back to start
while (x.prev) {
x = getPagedItems(PAGE_SIZE, {prev:x.prev})
console.log(`Page ${--i}: `, x.items)
}
}
runTest()
I faced a similar problem.
The generic pagination approach is, use "start index" or "start page" and the "page length".
The "ExclusiveStartKey" and "LastEvaluatedKey" based approach is very DynamoDB specific.
I feel this DynamoDB specific implementation of pagination should be hidden from the API client/UI.
Also in case, the application is serverless, using service like Lambda, it will be not be possible to maintain the state on the server. The other side is the client implementation will become very complex.
I came with a different approach, which I think is generic ( and not specific to DynamoDB)
When the API client specifies the start index, fetch all the keys from
the table and store it into an array.
Find out the key for the start index from the array, which is
specified by the client.
Make use of the ExclusiveStartKey and fetch the number of records, as
specified in the page length.
If the start index parameter is not present, the above steps are not
needed, we don't need to specify the ExclusiveStartKey in the scan
operation.
This solution has some drawbacks -
We will need to fetch all the keys when the user needs pagination with
start index.
We will need additional memory to store the Ids and the indexes.
Additional database scan operations ( one or multiple to fetch the
keys )
But I feel this will be very easy approach for the clients, which are using our APIs. The backward scan will work seamlessly. If the user wants to see "nth" page, this will be possible.
In fact I faced the same problem and I noticed that LastEvaluatedKey and ExclusiveStartKey are not working well especially when using Scan So I solved Like this.
GET/?page_no=1&page_size=10 =====> first page
response will contain count of records and first 10 records
retry and increase number of page until all record come.
Code is below
PS: I am using python
first_index = ((page_no-1)*page_size)
second_index = (page_no*page_size)
if (second_index > len(response['Items'])):
second_index = len(response['Items'])
return {
'statusCode': 200,
'count': response['Count'],
'response': response['Items'][first_index:second_index]
}
I'm building an nsIProtocolHandler implementation in Delphi. (more here)
And it's working already. Data the module builds gets streamed over an nsIInputStream. I've got all the nsIRequest, nsIChannel and nsIHttpChannel methods and properties working.
I've started testing and I run into something strange. I have a page "a.html" with this simple HTML:
<img src="a.png">
Both "xxm://test/a.html" and "xxm://test/a.png" work in Firefox, and give above HTML or the PNG image data.
The problem is with displaying the HTML page, the image doesn't get loaded. When I debug, I see:
NewChannel gets called for a.png, (when Firefox is processing an OnDataAvailable notice on a.html),
NotificationCallbacks is set (I only need to keep a reference, right?)
RequestHeader "Accept" is set to "image/png,image/*;q=0.8,*/*;q=0.5"
but then, the channel object is released (most probably due to a zero reference count)
Looking at other requests, I would expect some other properties to get set (such as LoadFlags or OriginalURI) and AsyncOpen to get called, from where I can start getting the request responded to.
Does anybody recognise this? Am I doing something wrong? Perhaps with LoadFlags or the LoadGroup? I'm not sure when to call AddRequest and RemoveRequest on the LoadGroup, and peeping from nsHttpChannel and nsBaseChannel I'm not sure it's better to call RemoveRequest early or late (before or after OnStartRequest or OnStopRequest)?
Update: Checked on the freshly new Firefox 3.5, still the same
Update: To try to further isolate the issue, I try "file://test/a1.html" with <img src="xxm://test/a.png" /> and still only get above sequence of events happening. If I'm supposed to add this secundary request to a load-group to get AsyncOpen called on it, I have no idea where to get a reference to it.
There's more: I find only one instance of the "Accept" string that get's added to the request headers, it queries for nsIHttpChannelInternal right after creating a new channel, but I don't even get this QueryInterface call through... (I posted it here)
Me again.
I am going to quote the same stuff from nsIChannel::asyncOpen():
If asyncOpen returns successfully, the
channel is responsible for keeping
itself alive until it has called
onStopRequest on aListener or called
onChannelRedirect.
If you go back to nsViewSourceChannel.cpp, there's one place where loadGroup->AddRequest is called and two places where loadGroup->RemoveRequest is being called.
nsViewSourceChannel::AsyncOpen(nsIStreamListener *aListener, nsISupports *ctxt)
{
NS_ENSURE_TRUE(mChannel, NS_ERROR_FAILURE);
mListener = aListener;
/*
* We want to add ourselves to the loadgroup before opening
* mChannel, since we want to make sure we're in the loadgroup
* when mChannel finishes and fires OnStopRequest()
*/
nsCOMPtr<nsILoadGroup> loadGroup;
mChannel->GetLoadGroup(getter_AddRefs(loadGroup));
if (loadGroup)
loadGroup->AddRequest(NS_STATIC_CAST(nsIViewSourceChannel*,
this), nsnull);
nsresult rv = mChannel->AsyncOpen(this, ctxt);
if (NS_FAILED(rv) && loadGroup)
loadGroup->RemoveRequest(NS_STATIC_CAST(nsIViewSourceChannel*,
this),
nsnull, rv);
if (NS_SUCCEEDED(rv)) {
mOpened = PR_TRUE;
}
return rv;
}
and
nsViewSourceChannel::OnStopRequest(nsIRequest *aRequest, nsISupports* aContext,
nsresult aStatus)
{
NS_ENSURE_TRUE(mListener, NS_ERROR_FAILURE);
if (mChannel)
{
nsCOMPtr<nsILoadGroup> loadGroup;
mChannel->GetLoadGroup(getter_AddRefs(loadGroup));
if (loadGroup)
{
loadGroup->RemoveRequest(NS_STATIC_CAST(nsIViewSourceChannel*,
this),
nsnull, aStatus);
}
}
return mListener->OnStopRequest(NS_STATIC_CAST(nsIViewSourceChannel*,
this),
aContext, aStatus);
}
Edit:
As I have no clue about how Mozilla works, so I have to guess from reading some code. From the channel's point of view, once the original file is loaded, its job is done. If you want to load the secondary items linked in file like an image, you have to implement that in the listener. See TestPageLoad.cpp. It implements a crude parser and it retrieves child items upon OnDataAvailable:
NS_IMETHODIMP
MyListener::OnDataAvailable(nsIRequest *req, nsISupports *ctxt,
nsIInputStream *stream,
PRUint32 offset, PRUint32 count)
{
//printf(">>> OnDataAvailable [count=%u]\n", count);
nsresult rv = NS_ERROR_FAILURE;
PRUint32 bytesRead=0;
char buf[1024];
if(ctxt == nsnull) {
bytesRead=0;
rv = stream->ReadSegments(streamParse, &offset, count, &bytesRead);
} else {
while (count) {
PRUint32 amount = PR_MIN(count, sizeof(buf));
rv = stream->Read(buf, amount, &bytesRead);
count -= bytesRead;
}
}
if (NS_FAILED(rv)) {
printf(">>> stream->Read failed with rv=%x\n", rv);
return rv;
}
return NS_OK;
}
The important thing is that it calls streamParse(), which looks at src attribute of img and script element, and calls auxLoad(), which creates new channel with new listener and calls AsyncOpen().
uriList->AppendElement(uri);
rv = NS_NewChannel(getter_AddRefs(chan), uri, nsnull, nsnull, callbacks);
RETURN_IF_FAILED(rv, "NS_NewChannel");
gKeepRunning++;
rv = chan->AsyncOpen(listener, myBool);
RETURN_IF_FAILED(rv, "AsyncOpen");
Since it's passing in another instance of MyListener object in there, that can also load more child items ad infinitum like a Russian doll situation.
I think I found it (myself), take a close look at this page. Why it doesn't highlight that the UUID has been changed over versions, isn't clear to me, but it would explain why things fail when (or just prior to) calling QueryInterface on nsIHttpChannelInternal.
With the new(er) UUID, I'm getting better results. As I mentioned in an update to the question, I've posted this on bugzilla.mozilla.org, I'm curious if and which response I will get there.