emporoar tv usb not working with 184 package

[PETER Verweij]

With the new 184 package and (brandy2 special build) the usb dvb-tv receiver with the emperoar program won't work. Normal i get a color on the usb tv stick when there is a signal, but with 184 package no color and thus no tv. reverted back to 183 package and (same position nothing else changed) and then perfect reception. Normally usbehcd is used by the program.

If I have/can test something let me know. (no trap or failure message on the moment to attach)

[Lars Erdmann]

Does any other USB 2.0 device work ? Like memory stick for example ?

[Lars Erdmann]

USBEHCD, USBOHCD: changed "default addr. timeout" processing

[Lars Erdmann]

try the attached.

[Lars Erdmann]

directly go for vrwp.2.zip.

[PETER Verweij]

Hi,

Teste both vrwp and vrwp2 both give the same result as 184 package, usb dvb-tv stick is recognized, but there comes no stream for tv. Reverting back to 183 gives stream. Tested also usb memory stick and usb mouse both work. I have the idea that also the usb dvb-tv stick does something, it is recognized, but that's all. In one of the earliest versions you made I had the same problem. Some additional information, ecs 2.0, lenovo t61 system, acpi 3.20.4. If you want othere information or tests let me know.

[Lars Erdmann]

There is a couple of things:

1.) attach output of pci.exe -D. I need to know what HC controllers you have exactly

2.) give the USBDock / USBResmgr output for the usb dvb-tv receiver stick. I need to know what device it is.

3.) try to find out whether USBEHCD.SYS or USBOHCD.SYS or USBUHCD.SYS fails: take the 10.183 drivers and replace one by one with the 10.184 drivers until you know which driver fails.

4.) once you find the 10.184 driver that fails, create a trace as described in the readme.txt file.

[Lars Erdmann]

USBEHCD, USBOHCD, USBUHCD: changed "default addr. timeout" processing

[Lars Erdmann]

Try vrwp.3.zip.

[PETER Verweij]

Sorry, no go. it is only the usbehcd driver which gives the problems. (using the usbehcd driver from the 183 package gives tv). Sometimes I see for a short moment the light from the reception flashing, normally it just lights if there is a good reception. One time I had reception, very strange, with the 183 drivers I have always good reception. So I don' t believe it is in the reception, rebooted within a few minutes and tested with channels which have good reception. Also tested usb stick this works with test version, tested also a canon 350d photocamera (ptp driver) works also good.

[Lars Erdmann]

Please take trace for USBEHCD.SYS 10.184

[PETER Verweij]

output from pci -d

[PETER Verweij]

output from usbresmgr

[PETER Verweij]

output trace fmt

[PETER Verweij]

I hope I created all the logs you needed, if not let me know or if you want other information.

[Lars Erdmann]

try USBEHCD.SYS 10.185. It's on Hobbes.

[PETER Verweij]

Same problem with 185. Added trace from 185 version. In same zip (183 and 185.zip) also created trace output from the 183 version.

[Lars Erdmann]

Sorry, but please redo trace WITHOUT wrapping the trace buffer (I need the beginning from boot):
specify:

TRACEBUF=512 /M=NW,Q,NDTI /D=ALL

in config.sys (/M=NW means "no wrap" whereas /M=W would mean "wrap").

[Lars Erdmann]

Telling from trace183.ftf: you have attached 2 USB devices, one USB 1.x and one USB 2.0 device.
Did you use a USB mouse or keyboard along with that dvbt stick ?

[PETER Verweij]

Will make a new trace tomorrow (I'm now in a region where there is not always reception with dvb-t) With 183 trace, nothing else connected, the only thing which is buildin, is the usb-fingerprint reader. Maybe that's the one, you see in the log.

[Lars Erdmann]

Also, try USBEHCD.SYS with the speed setting /S:8

[PETER Verweij]

trace no wrap

[PETER Verweij]

attached the trace from 183 and 185 package with no wrap.

[PETER Verweij]

tried the 185 package with the s:8 switch, as soon as I start emperoar it complains is can't find a suitable usb-stick and stops working. I made a trace, don't know if it will give good information, but it is no trouble to make it :-)

[PETER Verweij]

185 package with ehcd driver with /s:8 switch

[PETER Verweij]

Tried 186 package, same problem as with 184 and 185. Trace added. Usb memory stick is working. Strange is that on 1 usb adapter the dvb-stick isn't recognized, on the other usb adapters it is recognized

[Lars Erdmann]

Have you updated EVERYTHING, INCLUDING USBD.SYS ? You should. If it works on one USB adapter but not on the other (you have 2 EHCI controllers) then it's most likely a timing problem I have to sort out ...

By the way: will the DVB-T Stick at least show up in USBDock / USB resource manager ?
Try that for both USB EHCI controllers (change ports) ...

[Lars Erdmann]

USBEHCD: modifying end of HC reset (remove some waits, reorder schedule enablement)

[Lars Erdmann]

Try vrwp.4.zip

[PETER Verweij]

Updated everything, is most simple to do :-) tried the vrwp4 package, some improvement, but no solution. on the left side usb connector, the dvb-t stick is recognized by usb resource manager, but not by emperoar. On the right side I have 2 usb connectors, sometimes one works,sometimes the other one. ALways the dvb-t stick is recognized by usb resource manager but not always by emperoar. Once I did have a stream of a tv program, but after changing the channel (from ard to zdf) no more stream. One funny thing with this version. Normally the dvb-t stick gives a blue connection light when it picks up a signal (with constant blue connection light i normally have a good stream) I know get a continuous blue connection light, but no sound and no picture (video) in emperoar. I made a new trace, attached it. Thanks for all the work!

[Lars Erdmann]

USBEHCD: on HC reset only wait for POTPGT when ports are not permanently powered

[Lars Erdmann]

Try vrwp.5.zip.

[PETER Verweij]

I'm sorry but behavior keeps the same but a little bit better, no recognition of usb-dvb-t stick on the left side with emperoar wit usb resource manager it is recognized, on the right side it is recognized (tested a few times and seems always to be recognized by usb resource manager and emperoar). Blue light for fetching a stream is on, but no video out put on the screen and no sound. It seems as if emperoar doesn't get the data or it is in a other way presented? so that emperoar can't handle the information??

If you need other information let me know.

[Lars Erdmann]

What do you mean by "right side" and "left side" ? Are you attaching to different USB controllers ? If yes which one is "right side" and which one is "left side" ?

By the way: please post a valid email address so that I can put you on CC in that you get notified when I add something here.

[Lars Erdmann]

USBEHCD: add tracing to USBRB allocation / freeing

[Lars Erdmann]

Can you please use vrwp.6.zip and redo the trace for USBEHCD.SYS ?

[Lars Erdmann]

As an intermediate test: can you use USBD.SYS 10.162 (the original version from IBM) and see if it makes any difference ?

[PETER Verweij]

Left and right has to do with the usb adapter I use on the notebook (I have one on the left side and two on the right side) Email address added, I think ;-) will test vrwp6 The original 10.162 had a problem, with emperoar there are special patched versions delivered. can see if I can get the original 10.162 drivers (must have them somewhere). The first versions you made had the same problem no working usb-dvb-t stick. There should be an old ticket I made then. After a while all the versions you made worked perfectly up to 183. that's the last version which works.

[PETER Verweij]

Tested the new version. It gets better, not perfect and as good as 183 version. :-) the usb adapter on the left side of the notebook still refuses to work with emperoar, usb resource manager sees the dvb-t stick. (with 183 version this adapter works, strange for me) The usb adapters on the right side works. I even had sound and video output, but after changing a channel no more sound and video output (changed from ard to einsfestival) Blue connection light is on, but it takes a "long" time before an actual video stream and sound is displayed. Will try the 162 version (copied them from ecs2.1 install)

[PETER Verweij]

With original ibm-usb drivers I get the nice trap screen (picture 2012-09-27-img0178.jpg added)

[Lars Erdmann]

0) please put a valid email address here in the comments so that I can add it to the CC field.

1) I am not surprised it traps with USBEHCD.SYS 10.162. I think we had that already.

2) the new version vrwp.6.zip CANNOT be better than vrwp.5.zip. I just added more tracing and nothing else. Are you instead saying that with backlevelling USBD.SYS to 10.162 it had gotten better ? That would indicate that I'd have something to do about USBD.SYS.

[Lars Erdmann]

USBEHCD: do NOT wait for port disconnect when low speed/full speed device has been detected

[Lars Erdmann]

Try vrwp.7.zip real quick.

[Lars Erdmann]

One question: I can see that you attach a USB 1.x device (USB mouse ?) to a HC that handles 4 USB ports. Then you attach a USB 2.0 device to another port of that very same HC. I assume that this USB 2.0 device is the DVB-T stick.

Is it the very same set of USB ports (the very same HC) where you plug in your USB 1.x device and the DVB-T stick ? If you are not sure, you might want to attach your USB 1.x device to a completely different port, hopefully to one managed by the other HC.

Lars

[PETER Verweij]

The original ibm-usb drivers give a trap. With emperoar there are special versions delivered which work, that's what I meant. Version7 gives the problem that the usb-dvb t stick is recognized by usb resource manager but not by emperoar. I think I have the emailaddress now good, The usb 1 device you see is the fingerprintreader which is built-in. (can't change the connection of that device)

[Lars Erdmann]

I cannot see anything apparently wrong. Can you please backlevel USBD.SYS to 10.162 and see if that makes a difference ?

[Lars Erdmann]

By the way: can you also attach the USB Resmgr info for the built-in fingerprintreader ?

[Lars Erdmann]

can you also attach your config.sys ? Have you been specifying different /S: values for the two EHCI host controllers ?

[Lars Erdmann]

USBEHCD: set speed setting separately from starting HC (looks like it's a tming thing)

[Lars Erdmann]

Try vrwp.8.zip.

[PETER Verweij]

output from usbresource manager for fingerprint reader

[PETER Verweij]

config.sys

[PETER Verweij]

Tested vrwp8:

• on left side of the notebook usb adapter, usbresource manager sees the dvb-t stick, I hear the recognition sound if a device is inserted, but last beep isn't played. Emperoar doesn't see the stick.
• on right side of the notebook usb adapters, usb resource manager sees the dvb-t stick, I hear the recognition soudn if a device is inserted, an a last beep. Emperoar sees the usb-dvb-t stick. Sometimes it starts streaming video with sound, not always! Recognition light for stream is not always on. (but is a tv channel which gives no problem with quality with 183 version). Changing channel takes a while before something happens, mostly no video and no sound. Reverting back to the first channel mostly no sound and no video.

Will test now with usbd.sys from original ibm files. and usbehcd from vrwp8. (needs a reboot.)

[Lars Erdmann]

Please remove /S:8 from USBEHCD.SYS for the first instance. The default for /S: if unspecified is now /S:1.[[BR]] I want to see if both EHCI HCs behave exactly the same or not. They should as they are exactly the same HCs. I would also think that /S:8 is potentially too slow for operating the dvb-t stick.

[PETER Verweij]

vrwp8 usbehcd with usbd.sys from ecs2.1 install

[PETER Verweij]

usb vrwp8 tested with usbd.sys from ecs2.1 install problem keeps the same. normally on new level of usb build I only have to backlevel usbehcd.sys to 183. to get things working. see your comment about /s:8 sorry about that, is from old test, forgot that to remove, will remove and start again :-)

[PETER Verweij]

vrwp8 with sbd.sys from ecs2.1 install removed switches /s8 config.sys

[PETER Verweij]

Now usb-dvb-t stick is always recognised by usb resource manager and emperoar left or right side from notebook ! Only downside no video or sound. Blue recognition light on the usb-dvb t most of the time off. Only thing I can do at the moment is install also usbd.sys from vrwp8. I think it will not change anything, but will try that also. Thanks for all the time and effort you take for getting it to work!

[PETER Verweij]

vwp8 full package installs without switches in config.sys

[PETER Verweij]

Tested again vrwp8 but now with no switches for usbehcd.sys in config.sys. No change. no video or sound, blue connection light on usb-dvb-t stick mostly keeps off.

I don't know if a trace from usbehcd.sys from 183 usb drivers and everything from vrwp8 will help you. Up to vrwp7 that was enough to get every thing working again. (didn't test it yet with vrwp8)

[Lars Erdmann]

USBEHCD,USBOHCD,USBUHCD: using TSC for all blocking delays

[Lars Erdmann]

Try vrwp.9.zip and take trace. If relevant, let me know if it boots ok or if it hangs or traps the system on boot.

[PETER Verweij]

System boots fine, dvb-t stick recognized by usb resource manager and emperoar. Just no video and no sound with emperoar. Blue connection light most of the time stays off.

Tested also a usb-stick (memory) works fine.

[Lars Erdmann]

As a check: Is this a 2 GHz CPU with 2 cores ?

[PETER Verweij]

Lenovo ThinkPad? T61 6466 - Core 2 Duo T8100 / 2.1 GHz - Centrino Duo That's the machine I use. Have the latest acpi (betazone) works perfectly.

[Lars Erdmann]

USBEHCD: readd scheduling context hook waiting for frame rollover

[Lars Erdmann]

please try vrwp.10.zip

[PETER Verweij]

dvb-t stick recognized by usb resource manager, but reognition beep is strange, normally you hear first a kind of a melody and then a short high toned beep. Now the "melody" is played twice and then the high beep. emperoar starts, at first no video and audio, made the trace and then all of a sudden, recognition light goes on an there is video an audio. Changed to different channel, nothing, waited 6 minutes but no video. Closed and started emperoar again, but no video. If you want other information let me know!

[Lars Erdmann]

USBEHCD: readd mutex in AccNonIsoReq?, make sure that requests terminate in the order they were allocated (?)

[Lars Erdmann]

Try vrwp.11.zip.
If that works: it's not the good solution as it reverts the performance speed-up that I had achieved. I would like to find a better solution and I would need you to continue testing.

[PETER Verweij]

It's the same as with version 10, I think. Beeps are strange, but recognition is there form usb resource manager and emperoar. Started emperoar, no video no sound, changed channels nothing, waited a while nothing. started tracefmt and there it is sound and video! changing channel and no sound, video. It needs a simple reboot, but will try with trace disabled to see if there is some change.

About the testing it's in my interest that there is development ;-) So I can't and will not complain about testing. I'm just very happy to contribute :-)

[PETER Verweij]

Made no difference with trace disabled. Tried to copy 6 files each 18MB to the hard drive. After a while the usb-stick can't be accessed anymore. Tried with the usb183 no problem. I don't know if gives you some information you need.

[Lars Erdmann]

USBEHCD: free USBRB in exactly the same sequence they were allocated

[Lars Erdmann]

Try vrwp.12.zip. If that works: it's the good solution I was talking about earlier on.
Let me know if it hangs or traps. If trap, please post trap screen.
It's not a problem with having trace enabled or disabled, you don't need to worry about that.

[PETER Verweij]

No change, recognition as always, also the strange beep, I wish I could record that, maybe it gives you a clou? no sound, no video, no recognition light for a stream. Not even after waiting for a while. No trap with emperoar. DO you also want a test with a big file transfer?

[Lars Erdmann]

Yes, do a large file transfer and trace that. Also make sure the file is not corrupted ...

[Lars Erdmann]

USBEHCD: free USBRB in the REVERSE sequence with which they were allocated

[Lars Erdmann]

try vrwp.13.zip. Also do a big file transfer.

[PETER Verweij]

sorry for later reaction had a crash of the filesystem :-( took a while to restore everything. recognition ok, had even for a while video and audio. changing channel and every thing is lost. Will trace a big file transfer later in the day.

[Lars Erdmann]

Can you only use 1 CPU core by specifying /MAXCPU=1 on the ACPI.PSD command line ?
I want to find out if this is a concurrency issue amongst multiple CPU cores.

[PETER Verweij]

Tried with switch for 1 cpu. trace added. recognition by usb resource manager and emperoar. sometimes blue recognition light for a stream on the dvb stick is on, but no audio and no video.

[PETER Verweij]

Tried to trace with a big file transfer, but problem occured. during the tranfer it suddenly stopped, started tracefmt had a trace but couldn't save the file. No drive was selectable. I don't know if it has something to do with usb driver or is another problem. Worked the whole day with the laptop without a problem. Will try later again. Is there a possibility to do a trace after you have already done a trace? the only solution i know is to reboot and trace.

[PETER Verweij]

tried it 4 times, copying 6 files each 18MB. no problems for now. Just made a trace, don't know if it's helpfull. Tried it with pc on only battery, with ac adapter connected. to ramdrive and directory.

[Lars Erdmann]

As to "resume tracing":
yes you can restart trace from commandline, see "trace.exe" for help.

This will give you info on how tracing is currently set up: "trace.exe /q".

For your case, you want to clear the tracebuffer and then resume tracing which is: "trace.exe /c /r"

[Lars Erdmann]

By the way: in general, please use ALL of vrwp.x.zip.

But as a test: Can you try with usbd.sys 10.162 (and multiple CPUs as this does not seem to make any difference) ? I don't have the faintest clue if that makes a difference but ...

[Lars Erdmann]

USBEHCD: "GetQH", lower Nac counter back to 1

[Lars Erdmann]

try vrwp.14.zip real quick. Also try in conjunction with USBD.SYS 10.162.

[PETER Verweij]

added trace with usb test version vrwp13 and only usbd.sys from 10.162 version. Same results.

[PETER Verweij]

Made a trace with only dvb-t stick connected, same results (is recognized, beep is strange) no video no audio, no recognition light of a stream on the stick. Tried to make a trace with dvb-t stick an usb-stick connected, but then then the pc hung. no trap, but no reaction from the system. Message just a bit short, I'm at work,-)

[Lars Erdmann]

USBEHCD: rework USBRB management, iterate through USBRB in the sequence they were allocated

[Lars Erdmann]

Try vrwp.15.zip. If it works in general, please tell me if system response becomes jerky in particular if you operate an additional device (like a memory stick) in parallel.

[PETER Verweij]

only dvbt stick connected

[PETER Verweij]

trace with dvbt and usb stick

[PETER Verweij]

dvb stick is recognized, light for stream recognition is on, but simply no video and no sound. tried it in conjuction with usb stick connected, copied 5 files from about 18MB each, its real fast :-) no problems.

[Lars Erdmann]

Has the situation improved ? Is the strange beep gone ? Is the light for stream recognition permanently on ? Or is it just as bad as before ?

I am a bit clueless as the dvbt-stick is a bulk device just like the usb stick is.

[PETER Verweij]

Improved, yes, stream recognition light is sometimes on if it's on then permanently, sometimes like today it's off an stays off, but no video is displayed, strange but don't know what is causing it. situation is better. Only thing I can add at the moment is that Rudigher Ihle made the application, maybe he has a clue?

[PETER Verweij]

are you sure it is acting like a bulk device? usbres says it is Human Interface Device (HID), while a usb stick is recognized as a Storage Device. I know the strange thing is that changing the usbehcd driver cures the problem, but maybe it has some connection to usbd? (I'm no programmer so just guessing)

[Lars Erdmann]

"Interface 0" of your resmgr output tells me it's a bulk device. "Interface 1" tells me it can additionally act as a keyboard but I am sure for the audio and video data it's using the bulk interface.

Yes, the USBEHCD driver is the problem. It differs from USBUHCD and USBOHCD in that it tries to do a better job on accepting and queueing up requests from applications on the risk of having additional serialization issues (need to prevent requests mixing up ...). I already talked to Rüdiger Ihle and he told me it's critical that USBEHCD needs to process the requests in the order that Emperoar passes them to USBEHCD and that's what I am currently trying to straighten out as I suspect that that is the problem case.

[Lars Erdmann]

USBEHCD: adding mutex protection to "AddToQH", making "GetNextUSBRBDATA" check for qh pointer instead for status flag

[Lars Erdmann]

Try vrwp.16.zip. If possible, take trace.

[Lars Erdmann]

USBEHCD,USBOHCD,USBUHCD: kee16lib, put back in esp,ebp manip.

[Lars Erdmann]

Use vrwp.17.zip. In vrwp.16.zip I had removed stuff where I thought it was not necessary but where it turned out that it is (in order to circumvent a potential bug in the KEE API).

[PETER Verweij]

beep is still strange, the latest high beep isn't played. recognition light voor stream keeps off. no video or audio, sorry. This all for vrwp17, didn't try vrwp16

[PETER Verweij]

This is real strange, submitted the message and there it was video and audio. :-) switched channel and all is lost. Is it possible that there is some kind of interference with the ps/2 mouse (trackpoint)?

[Lars Erdmann]

USBEHCD, USBOHCD: reworked all the Mutex serialization

[Lars Erdmann]

If 10.183 USBEHCD.SYS works with PS/2 mouse then it certainly has nothing to do with PS/2 mouse.

Anyway, try vrwp.18.zip. I still have to change USBUHCD.SYS but if you don't have any USB 1.x device "interference" then it shouldn't matter for your tests.

[PETER Verweij]

tried vrwp18 usb resource manager and emperoar recognizes the usb-dvbt card. beep is more like the one with 10.183 version. only the latest high beep takes a while. started emperoar, no video no sound. closed emperoar started again, no video no sound. while i was typing this message there is was video and audio. watched a while, lost my message in the meantime, changed channel nothing no video no audio. waited 10 minutes still nothing. disconnected now from usb, reconnected and started the application again. Don't know if I get video, after 2 minutes stream recognition light is on, but still no video. after 5 minutes still nothing. Tried it this time with timing the minutes because in the latest version it mostly starts giving video after a while, but can't find out why.

[Lars Erdmann]

USBEHCD: correctly update current USBRB read pointer when a new USBRB is added while looping through the list

[Lars Erdmann]

Try vrwp.19.zip

[PETER Verweij]

trace maybe a bit short, but restarted the trace, had the device tested before. Recognition beep stays strange. DVB-t recognized by usb resource manager and emperoar. Story stays the same, starting the package gives no audio or video. After a while there is perfect audio and video. Chnaging the channel and all is gone. Tried to insert usb stick and see if that made a change, but can't conform that. If you wan't other tests let me know. Sorry can't give better news at the moment.

[Lars Erdmann]

USBEHCD: if new USBRB is added to chain, set current read pointer to BEGINNING of chain

[Lars Erdmann]

Try vrwp.20.zip

I know what the problem is but it turns out it's not all that easy to fix.
Emperoar has some special way of sending read requests to USBEHCD.SYS that Rüdiger Ihle explained to me (this is all in order to sustain the high data rate that is required for the video/audio stream) and that makes it different from let's say doing a chkdsk on a USB drive. I don't have Emperoar to test therefore I have to go blind with the traces you provide to me. They give me a clue of what's going wrong.

[Lars Erdmann]

USBEHCD: if new USBRB is added to chain, set current read pointer to BEGINNING of chain but ONLY if it is at current end

[Lars Erdmann]

Also try vrwp.21.zip. Tell me if it is better than vrwp.20.zip (if any of these work at all).

[Lars Erdmann]

USBEHCD: if new USBRB is added to chain, set current read pointer to the PREVIOUS end of the chain

[Lars Erdmann]

Also try vrwp.22.zip after you have tried the others. Let me know which of vrp.20.zip, vrwp.21.zip, vrwp.22.zip work best if at all.

What you should also do: access a USB stick, let's say open some files or start a search, whatever. At the same time, open USB resource manager and check that you don't experience a hang or failure while USB resource manager parses all the device descriptors.

[PETER Verweij]

Tried vrwp.19with usb-stick had to copy 110MB (5 105MB)after awhile nothing happend anymore. forgot to make trace. wil try the new versions.

[Lars Erdmann]

USBEHCD: rework USBRB management once more

[Lars Erdmann]

skip vrwp.20.zip,vrwp.21.zip,vrwp.22.zip, directly try vrwp.23.zip.

[PETER Verweij]

Oke, just tested vrwp.20 no video. Copying real large files 160 MB and 10 MB lead to a hang in transfer. Could start tracefmt but couldn't save the file no possibility to select a drive. Will test new version shortly.

[PETER Verweij]

tried vrwp23. recognition from both usb res manager and emperoar. beep keeps strange. started emperoar no video and audio. connected a usb stick and there it was video and audio! changed channel and nothing. disconnected and reconnected, choose refresh nothing,couldn't get video. copied 176MB with this version to the usb stick without a problem, this is better :-)

[Lars Erdmann]

USBEHCD: don't change USBRB read pointer on a USBRB allocate/unallocate

[Lars Erdmann]

Try vrwp.24.zip.

[Lars Erdmann]

USEHCD: rework "FreeClaimed?" to properly handle resource unallocation

[Lars Erdmann]

Skip vrwp.24.zip. Try vrwp.25.zip instead.

[PETER Verweij]

sorry took a bit longer, but trace from version 25 added. recognition is there both usb resource manager and emperoar. started emperoar nothing, attached usb stick, browsed directories started copying 2 large files 180 MB total, after a while video stream and audio perfectly. changed channel nothing, copying still but no more video and audio. Almost ready with copying looks good. Detaching usb-dvb-t closing emperoar and attaching, and starting again. No video, no audio. That's for me a bit funny, first time I get after a while video, but no second time. Copied the files succesfully. detached the usb-stick and reattached it but no more video, although video stream indicator is on.

[Lars Erdmann]

Please do this:

1) run "trace.exe on /M:W" from commandline: this will enable trace wrapping. That will give me the opportunity to see more data traffic once the trace buffer fills up. In short: this will enable overwriting the oldest trace buffer entries

2) run "trace.exe /C /R" from commandline: this will clear the tracebuffer and then resume tracing

3) don't bother with USB sticks, I think that this now works OK. Instead just attach dvbt stick and run Emperoar.

By the way, this is what Rüdiger explained about the beep: when you attach the dvbt stick, Emperoar will download firmware onto the stick. It will then initiate a "restart" of the dvbt stick. Once that is done, the dvbt stick will do a beep and will be operational.

[PETER Verweij]

Don't know what is going on at the moment, but no way to get the stream working. Made a trace after step1, or should i have done it after step2. I thought step 2 was after i got a working stream, but not sure. Will revert to 10.183 to test there is no hardware or other problem. if not then I will try again.

[PETER Verweij]

added the sounds connection, smplayer will play the files, made with telephone not perfect but maybe interesting to hear there is definitely another connection beep.

[PETER Verweij]

reverted back definitely no hardware problem all functions well. reinstalled vrwp25 no video, connected the usb stick, started copying a large file, notebook hangs, drives aren't accessible. rebooted, started emperoar no connection, connected usb-stick nothing, browsed folders files, no cure. started chkdsk on the usb stick and there was the stream ! mad trace.vrwp25.connection. changed channels nothing any more, connection light for stream is on. tried chkdsk on usb stick again but this time no video... let me know if this are the traces you wanted

[Lars Erdmann]

USBEHCD: reverting back to original: scanning USBRBs from back to front, no linked list

[Lars Erdmann]

Try vrwp.26.zip

[PETER Verweij]

Couldn't get a video or audio stream, connectionlight is mostly off. Attaching a usb-stick and browsing the directory structure didn't help. Connection beep stays the same as on all special builds. If you want more information let me know.

[Lars Erdmann]

USBEHCD: rework QH chaining/unchaining from physical chain

[Lars Erdmann]

try vrwp.27.zip

[PETER Verweij]

recognition from emperoar and usb resource manager ok connection beep stays the same (different from 10.183) couldn't get a video or audio stream. connection light stays off.

[PETER Verweij]

had the vrwp27 all day on the laptop, copied a few files, not that big, but after a while nothing happend on the usb stick.couldn't get the system to work further so had to reboot.

[Lars Erdmann]

As a test: can you run with 1 CPU only (ACPI.PSD /MAXCPU=1) and see if that makes a difference ? In particular using a memory stick.

When nothing happened with usb stick anymore: was the system completely hung or did USB just not work any more ? Would USB mouse continue to work (if you have such) ?

[PETER Verweij]

with 1 cpu for emperoar no change, trace added. for usb stick it seems to work, copying 2 files total of 180MB and still works (still in progress), also added trace for that. Will test tomorrow again, I know I have to copy files from and to usb stick and will leave these settings on the notebook for now.

system reacted still on alt-ctrl-del (the manager cad-popup appeared and could choose top.) it seemed as if the wps was frozen, couldn't get any icon to start. Didn't have a usb mouse at hand. It test tomoroow goes ok, will go for normal acpi setting and test again. In mean time the 2 files 180 MB copied succesfully.

[Lars Erdmann]

USBEHCD: changes to AddTDToQH,PrepareToRemove?,FreeClaimed?

[Lars Erdmann]

Try vrwp.28.zip.

[PETER Verweij]

tested vrwp27 with maxcpu=1 and copying from usb stick. (copying with a simple program I wrote in rexx) after a while it stops, searched for usbmouse and then program was gone, normally it stays active after it finished copying. added a trace, don't know if you can see the problem, trace is earlier then the problem occured.

[PETER Verweij]

normal acpi (so 2 cpu's) usb dvb t stick is recognized by usb resource manager and emperoar. started emperoar, no video, no audio, light for connection is off. Added the usb stick and there was the connection light, stayed on, but still no video or audio. Copying worked fine.

[Lars Erdmann]

You either have to manually execute "trace.exe ON /M:W" followed by "trace.exe /C /R" or you need to specify: TRACEBUF=512 /M=W,Q,NDTI /D=ALL
otherwise your traces are completely useless as they will only trace the bootup (which is obviously not the problem)
Obviously we need to catch the point where something hangs (well in case of a complete hang the trace will be lost too unless you would initiate a system dump).

[Lars Erdmann]

USBEHCD: poll if QH pointer is in use in "AddReqQH","AddPeriodicQH","AddTDToQH"

[Lars Erdmann]

Try vrwp.29.zip. For controlling trace see comment above.

[Lars Erdmann]

USBEHCD, "FreeClaimed?","PrepareToRemove?": poll HW for QH signaling inactive instead of forcing it through SW

[Lars Erdmann]

Then, try vrwp.30.zip after you have tried vrwp.29.zip

[PETER Verweij]

used new tracebuffer settings. trace made from vrwp29. beep is still different. recognition light was off, browsed some files on a fat32 partition changed channel and after a while there it was audio and video. changed channels again all lost. made the trace at that point. the recognition light is sometimes on, sometimes off.

[PETER Verweij]

vrwp30 tested, is a bit strange. as soon as I attach the dvb-t stick the cursor/mouse is gone and can't type anything. so first detached the dvb-t stick and typed tracefmt. this trace added. so can't test any further then insertion of the stick. If you want other information let me know.

[diver]

i wonder if it wouldn't be faster when Lars would get the DVB-T stick to test himself. @vrwp i don't know where you from, but eventually you could send Lars the stick?

[PETER Verweij]

I live in Germany, work in the Netherlands. Maybe I can find a second dvb-t stick. Don't know if it will make it easier for Lars.

[PETER Verweij]

I have a spare usb-dvb-t stick. My brother uses this one normally with windows, but I tested this stick also an works on OS/2. If you want to lent it let me know

[Lars Erdmann]

10.184: outdated version. No further response from reporter regarding newer versions.

[PETER Verweij]

Sorry,due to other messages I had the impression other things where tested. Tried 1.89 and stick is recognised by usb resource manager, but emperoar sees no usb-dvb stick. trace added.

[PETER Verweij]

trace with version 1.89 usb emperoar problem

[PETER Verweij]

tried version 1.90 same as with 1.89 usb-dvb-t stick isn' t recognized by emperoar.

[Lars Erdmann]

try anchieri.zip from ticket #54

[PETER Verweij]

trace with anchieri version

[PETER Verweij]

Tried with anchieri version, there are two version of anchieri, tried the first. Will try the second also.

[PETER Verweij]

trace with anchieri2 version

[PETER Verweij]

trace with anchieri2 version added. anchieri 1, usb-dvb-t stick is seen by usb widget, name msi digi vox is displayed, emperoar doens't see the usb stick. anchieri 2, usb-dvb-t stick is seen by usb widget, but as "new device", emperoar doesn't see the usb stick

[Lars Erdmann]

Try anchieri.4.zip in ticket #54.

[PETER Verweij]

trace with anchieri version 4 added, usb-dvb-t stick is seen by usb widget as "new device" emperoar doesn't recognize the usb-dvb-t stick.

[PETER Verweij]

trace withanchieri4 build

[Lars Erdmann]

Try anchieri.5.zip

[PETER Verweij]

Trace added, usb widget sees the dvb-t stick as msi digi vox, but emperoar keeps complaining no suitebla device found.

[PETER Verweij]

trace with anchieri5

[Lars Erdmann]

Try anchieri.6.zip. Don't bother with traces. I have no time to look at them.
Check ticket #54 regularly for newer versions

[PETER Verweij]

No change with anchieri6.

[PETER Verweij]

No change anchieri9

[Lars Erdmann]

specify /S:8 for usbehcd.sys.

[PETER Verweij]

Tried it, but emperoar doesn't see the usb-dvbt stick and won't start

[Lars Erdmann]

Get 10.191 from Hobbes. Then try various values for /S:, start with /S:8, then go down to /S:4, /S:2, /S:1[[BR]] Other valid values are /S:16, /S:32, /S:64 that you also might want to try out, just in case.

[PETER Verweij]

Tried all switches but emperoar keeps complaining no suitable device found....

[Lars Erdmann]

do you have access to eComStation beta zone ? If yes, get "USBHCD-eCS-10.192.zip". That should work no matter what you set for /S (in short it uses /S:1 as the default if you don't specify anything else).
If it works, report back so that I can finally close this ticket. Thanks.

[PETER Verweij]

Tried the drive from the betazone, same as before, usb-dvb-t stick is recognized by usbresmgr, but emperoar doesn't see it, complains: no suitable capture device found. Please install a device driver or attach sub device.

Config.sys is for a long time unchanged, so only logical error that keeps remaining emperoar doesn't see the usb device, don't know why.

[Lars Erdmann]

Updates to USBEHCD.SYS and USBD.SYS

[Lars Erdmann]

please try vrwp31.zip. Just back up everything and use the complete contents of this zip.

[PETER Verweij]

There is progress (regarding the dvb-t stick :-) ) Stick is seen by emperoar. Program starts, I did have a tv stream, but not as good as with old version. Some observations:

• switching to another tv channel leads to a long waiting time before a tv stream starts again.
• sometimes a perfect tv stream, but did also have situations with no pictures, but the blue connection light on the tv stick was on (normally it is on when it has good tv-stream reception. I have had this in earlier test builds

[Lars Erdmann]

USBEHCD: frame list rollover interrupt enable would be reset before the corresponding interrupt would fire

[Lars Erdmann]

try vrwp32.zip

[PETER Verweij]

Emperoar starts, blue light is on (good reception), but not always I get a good picture. Sometimes it does, but mostly the screen keeps black (no picture). made a trace. I can't say that there is for me much difference between vrwp31 and vrwp32.

[Lars Erdmann]

I have added additional tracing so that I can see if some certain error condition occurs with your dvb-t stick as I have a vague idea of what could be wrong. Make sure you update also the .TFF files and redo the trace with your dvb-t stick and the MPEGTV program.

[Lars Erdmann]

additional tracing in USBEHCD, "FinishIO" to check for certain error conditions

[PETER Verweij]

Added the new trace, had to try it twice, first time I started emperoar, switched channels, started tracefmt (while emperoar was running) then got the trap screen (trapscreen vrwp 33) Started all over again, but before I started tracefmt I stopped emperoar, that worked. If you want other information or trace let me know.

[Lars Erdmann]

SVN rev 907: USBEHCD, for completed transmission poll for all TDs to go inactive

[Lars Erdmann]

try vrwp34.zip:

1) thanks for reporting the trap screen. I found the error and hopefully fixed it. It was comnpletely unrelated to the original problem but nonetheless ...

2) take a trace as usual. I also updated the TFF file. Make sure you also update it.

[Lars Erdmann]

Before you ask: of course, with 2) I try to fix the original problem in Addition to fixing the trap.

[PETER Verweij]

Tested the new version. It takes a long time before I get a picture.

• once after a long wait I got a clear picture, blue reception light always on. had to change channels before something happend.
• tried a second time, not changing channels, waited also long, no picture, but blue reception light is on.

[Lars Erdmann]

SVN rev 908: additional tracing in "FinishIO" to check for halted QH and TD content

[Lars Erdmann]

try new version. Make sure you also replace TFF file as I need additional trace output. Take trace.

[PETER Verweij]

This version is not so good, emperoar won't start, doesn't see the usb-dvb-t stick. Made a trace, not sure if it will help. To fast in the reply, sorry. Attached the stick a second time, now emperoar starts. blue recognition light for strema is on, but no picture. system is slow in responds? tried chaning channel, no luck, for some reason I can't get another channel. will try to add a second trace vrwp35-2

[Lars Erdmann]

Thanks for the trace. I now know exactly what the problem is. Now it's time to find the right solution.

[Lars Erdmann]

SVN rev 909: USBEHCD. SYS, rework short packet processing, only do termination processing when USBINT or USBERRINT is set

[Lars Erdmann]

try new version. TFF file again updated, so make sure you replace. Take trace.

[PETER Verweij]

Emperoar starts with no problems, blue recognition light for recognition for dvb-t stream is on, but no picture. It seams easier to change channels, this goes fast, but tv-screen keeps black (no stream on the desktop.

[Lars Erdmann]

SVN rev 910: for short packet processing, check if currTD of QH is not pointing to what we think is short packet, for async advance interrupt, trigger extra ctx invocation

[Lars Erdmann]

Try vrwp37.zip and report back here. Please do me a favor and once this is done open another ticket with the same topic (but refer to 10.193 instead). This ticket is getting rather lengthy.

[Lars Erdmann]

SVN rev 910: for short packet processing, check if currTD of QH is not pointing to what we think is short packet, for async advance interrupt, trigger extra ctx invocation

[Lars Erdmann]

Sorry, use vrwp38.zip instead of vrwp37.zip.

[Lars Erdmann]

SVN rev 911: in "EHCIInterrupt" make sure that we do either termination processing or async advance proc. but not both on one ctx invocation

[Lars Erdmann]

Then try vrwp39.zip. I think that vrwp38.zip was lacking some necessary change. Just try both and we will see if it makes a difference.

[PETER Verweij]

I did manage to get sound and picture with version 38! did not test version 37, will try version 39 and then close the ticket, open a new one for new version. One thing is strange, after changing channel the reception is gone, most curious is that the channel which is the weakest will work, strong(reception) channel wont work.

[PETER Verweij]

Trace vrwp39 added, had reception from stronger channel, but not as clear as with package 38and the weaker channel. Will close this ticket and open a new one.

[Lars Erdmann]

don't forget to add your email address to the new ticket. Else you will not get any email notification.