Does anyone know of a "working" KVM switch for AmigaOS4?
I tried one and while it did work nearly 100% the mouse on AmigaOS4 side did not. It only moved upwards,downwards and when i pressed somebuttons it also moved to the right, but it was unusable.
TESmart it was, a shame, as it was really small and offered everything i'd need. I guess the reason for it to blow the mouse was, that it installs it's own hub (and probbaly emulating drivers for mouse/keyboard) as i only get signaled about new USB hardware when turning the switch on, but not when plugging something into it.
I'd be looking for a 2x HDMI KVM switch, which supports USB (keyboard/mouse), nothing fancy, no bells and whistles, it should simply switch and use the hardware thrown at it. A keyboard hotkey for switching would be nice too.
Thank you
Eit: USBInspector says it has bootmouse and bootkeyboard instead of HID, but i hvae yet to find a way to make it use HID. probably picks up the first driver it gets it's claws on
edit2:
If i remove bootmouse from kicklayout, kickstart subdir and from the USB subdirs then no driver will be loaded, crap
final edit:
Yep, it picks up bootxxx no matter what and i can't do anything about it
Edited by Raziel on 2019/5/23 18:32:52 Edited by Raziel on 2019/5/23 18:39:04 Edited by Raziel on 2019/5/23 19:24:02
If you are, could you please take a look? Maybe they have UHD versions as well?
And could you also fire up USBInspector and tell me which driver the ATEN uses?
After more tests yesterday, it seems that, due to the USB hub being initialized on bootup, the TESmart picks up the first two usbfd classes it gets it's claws on, that being bootmouse and bootkeyboard. It also seems that it never lets them go and uses them even after WB has started, which means i'm inclined to either find a mouse that works with bootmouse or find another KVM switch.
too bad i can't control the type of usb driver on a user level
PS: Did you ever get my eMail regarding the gfx card entry? I never got an answer...
Years ago, I used a KVM switch to switch keyboard, mouse and monitor between my MicroA1 computer and my HP Pavilion computer. Both used ps/2 keyboard and mouse and KVM switch worked very well.
Later on, we upgraded to a newer pc with USB keyboard and mouse. Wired mouse and keyboard seemed to work ok with both computers using the USB KVM switch.
Later on, we bought a wireless keyboard. This worked ok with the pc, but MicroA1 would loose track of the keyboard when the KVM switched the keyboard over to the pc.
@Raziel Test Rene W Olsens rmouse and rkeyboard usb drivers! I'm using an CDTV keyboard via Keyrah on my X5000 with Rene's drivers since the original Hyperion drivers where very crash prone with Keyrah.
It may also be that the X5000 (you got a X5000 right?) usb ports doesn't deliver enough current, I haven't mesured the ports.
Be aware though that you loose Amiga Input joystick support with Rene's drivers, there maybe a workaround.
If you want reverse mouse scrollwheel settings with Rene's drivers I can dig it up, don't remember how I did it at the moment.
I've tested quite a number of KVM's. The only one which worked always in CFE on the X1000 was the ATEN CS21 but it's VGA. It has two dedicated USB cables, one for mouse and one for the keyboard.
All the other ones prevented the X1000 from booting from switched off status when the switches where connected to the "keyboard-usb-port". If connected to any other USB port the computer boots from switched off status most of the times. However If I connected the usb switch after OS4 has loaded up, then a cold reboot would happen every time and I could use the keyboard in CFE for that session. Once I switched off the X1000, then I'll have to disconnect the KVMs again and reconnect later after OS4 had booted to have them working the way they should.
The other one wich works quite reliably and has not produced any grim reaper due to the USB stack not liking fast consecutive swithes is the StarTech.com SV431DVIUDDM. It kind of keeps the signatures of the monitor, keyboard and mouse in a memory and make the input devices appear to be "connected" to the OS all the time, even though the signals are redirected to the selected/desired computer/port. It is a very fast switch and the monitor does not loose the signal either. Unfortunately it has the CFE issue too, it acts as a HUB and only one USB cable carries the signals of Mouse and keyboard (and 2 additional USB ports which can be switched separately to the 4 computer using keyboard shortcuts)
oh as a side note, besides the fact that the idea o having 2 keyboards connected to X1000 is quite annoying to me (if you need to select linux in CFE boot screen) it makes os4 more unstable (linux side copes with it without problems). So it is not advisable.
I've tested quite a number of KVM's. The only one which worked always in CFE on the X1000 was the ATEN CS21 but it's VGA. It has two dedicated USB cables, one for mouse and one for the keyboard.
All the other ones prevented the X1000 from booting from switched off status when the switches where connected to the "keyboard-usb-port". If connected to any other USB port the computer boots from switched off status most of the times.
Try plugging the KVM into a different USB port. My X1000 is plugged into a 4-port ATEN DVI KVM, and it works well. I found that it works fine if I plug it in to any USB port apart from the first one. I bought a cheap keyboard to plug into the first port so that I can still use CFE's early start menu.
I don't think the problem is with the available current. Since the mouse kinda works, but it's only moving horizontal and sideways if i keep buttons pressed. I rather blame bootmouse.usbfd (and bootkeyboard.usbfd) to "normally" only support rudimentary features for keyboards and mice, while not knowing about more enhanced stuff, i guess. Apparently they were only meant to be used while booting (CFE) and for that it was probably enough.
Once the X reaches WB those drivers "should" release the hardware and hand it over to HID.usbfd. This works fine with any USB hubs on the NEMO hardware, but fails with external hubs (like the one this KVM switch uses).
Now, the problem is that, even if i remove bootmouse.usbfd and bootkeyboard.usbfd from the OS4 installation (SYS:kickstart/, SYS:kickstart/kicklayout and DEVS:USB/), with the idea that the KVM hub should immediately pick up HID.usbfd, it won't work either, because then the hub will simply use NO drivers at all (at least, that is what USBInspector tells me once the WB has finished loading). This is with another mouse/keyboard combo plugged into NEMO.
I'm out of ideas now. I also don't understand why the switch ALWAYS picks up bootmouse/bootlkeyboard...even if i plug it in after the WB has loaded completely...
Thank you for sharing your experience...browsing the StarTech products now.
wrt keeping the information about plugged in hardware stored. I guess all hubs doe that more or less nowadays and it might be why i have had problems with the TESmart. I read (i think in the manual) that it keeps the first instances of supported drivers in it's memory, uses them and never lets it go (unless unplugged). That's why it demands to work with bootmouse...why it keeps picking bootmouse up even after WB has been loaded is beyond me...probably because it looks for any driver with the "mouse" in it's name and clings onto that.
Unfortunately tricking it into thinking HID is bootmouse doesn't work either, because it crashes HID.usbfd :-/
While digging around in the settings and hardware info i found (through Ranger) that NEMO's USB ports (backside) are split up in 5 different OHCI hubs.
The first one (looking at it from the front with the cpu on the right side) on the right is hub 0 (and the only one with Header type: Normal (Multifunction) while all the others are "only" Normal) and one EHCI (pins on the board hich goes to the front). Hub 0 on the back is where the keyboard should go to be accessable in CFE boot.
All that informationn is moot though since the switch doesn't care and only wants bootmouse which prevents me from using my mouse.
I think it would be better to test mice that work with the switch than the other way round...
Have you tried plugging the KVM into a different USB port like I suggested? Yes, it won't be usable in CFE any more, but it may be enough to get it working in the OS itself. If it does, then you can always do what I do, and get a cheap additional keyboard for use with CFE.
BTW, the switch does *not* want bootmouse; it doesn't even know what that is. It's entirely an AmigaOS USB software problem. For whatever reason the handover from bootmouse to the USB stack isn't happening.
Like i said, it's ALWAYS. no matter what, if the sun is shining or rain pouring down, choosing bootmouse.usbfd and bootkeyboard.usbfd.
Quote:
BTW, the switch does *not* want bootmouse; it doesn't even know what that is. It's entirely an AmigaOS USB software problem. For whatever reason the handover from bootmouse to the USB stack isn't happening.
But that implies that bootxxx should not be available once WB has booted in...so why does the switch still use bootxxx?
Having it plugged in before i turn the system on or plugging it in after WB has booted completely, it will ALWAYS choose bootxxx
It also doesn't matter if i replug the switch, if i use the front or back ports, if i use any of the other back ports than the ones far right.
Like i said, it's ALWAYS. no matter what, if the sun is shining or rain pouring down, choosing bootmouse.usbfd and bootkeyboard.usbfd.
Sorry, I must have missed where you said that.
Quote:
But that implies that bootxxx should not be available once WB has booted in...so why does the switch still use bootxxx?
My understanding is that the bootxxx drivers are there to provide keyboard & mouse during early boot (incl. the early boot menu). Then, the HID driver takes over once the full USB stack is up and running.
No idea why it doesn't switch over when you're using the KVM. Either way, it's a software issue.
Just for information, The bootmouse.usbfd and bootkeyboard.usbfd drivers are built into kiskstart and are loaded during booting to provide support for the mouse and keyboard in the early startup screens / startup with no startup sequence etc.
They are minimal drivers, and some more modern mice in particular don;t work well with them. My Kensington mouse is quite bad, displaying very similar symptoms to those you describe.
They have setting which make them form a 'weak' attachment to the mouse of keyboard they are supporting. When the stsratup-sequance is run the USBCtrl START command is given ad the rest of the drivers are loaded. These form strong attachments and overide the boot drivers which should release the mouse / keyboard / ( or graphics tablet) to the hid.usbfd ( or uclogic.usbfd for my grahics tablet) .
So the issue must be that either the hid.usbfd is not recognising the KVM as a hid device or the bootmouse.usbfd is "refusing" to let go of it. I would think that it's more likely the first.
Curiously on the same combination of mouse / keyboard on my SAM Fkex this process is 100% robust, never ever goes wrong, on my X1000 the more mouse occasionally doesn't get picked up by the hid.usb driver and results in similar symptoms to those you see with the KVM.
This can usually befixed by unplugging and replugging the mouse cable.
My standard AmigaOne mouse never has this issue, which may be either because it always gets pucked up by the hid or is better support by the bootmouse, so the issue doesn't notice.
I don't know how transparent the KVM is to the mouse attached, does your muse work correctly when directly attached to the X1000 ?
Just for information, The bootmouse.usbfd and bootkeyboard.usbfd drivers are built into kiskstart and are loaded during booting to provide support for the mouse and keyboard in the early startup screens / startup with no startup sequence etc.
K, that explains why i can't force the system to use HID from as early on as possible...
Quote:
So the issue must be that either the hid.usbfd is not recognising the KVM as a hid device or the bootmouse.usbfd is "refusing" to let go of it. I would think that it's more likely the first.
Curiously on the same combination of mouse / keyboard on my SAM Fkex this process is 100% robust, never ever goes wrong, on my X1000 the more mouse occasionally doesn't get picked up by the hid.usb driver and results in similar symptoms to those you see with the KVM.
I faintly remember having such issues with PC mice and early versions of AmigaOS4, but that was long ago.
Quote:
This can usually befixed by unplugging and replugging the mouse cable.
Not for me unfortunately. I tried every combination possible. Plug it in before i turn on the machine and before i turn on the KVM switch Plug it in with the KVM switch running, but before i turn on the machine. Plug it in after CFE has finished booting, but WB has not. Plug it in after WB has finished booting. Remove it and replug it after WB finished booting ...you name it...
It doesn't matter, as soon as the mouse is plugged in it starts acting weird. And that is not only with the mouse i'm using right now, it's also with the original AmigaKit mouse (the one with the boingball sticker). I do not own more mice to test, though...but it looks bad nonetheless.
Quote:
My standard AmigaOne mouse never has this issue, which may be either because it always gets pucked up by the hid or is better support by the bootmouse, so the issue doesn't notice.
Probably better supported? Was this standard mouse available during development of kickstart maybe? Or maybe some similar type.
Quote:
I don't know how transparent the KVM is to the mouse attached, does your muse work correctly when directly attached to the X1000 ?
Yes, it does, i have it working for years now in any X1000 USB hub.
News feed
OS4 Feedback forum
OS4Depot Feedback forum
Amigabounty forum
OpenAmiga forum
AmiCygnix forum
ABC shell forum
AmiKit forum
Cinnamon Writer forum
CodeBench forum
Digital Universe forum
Dopus 5 forum
E-UAE forum
Gnash forum
Ibrowse forum
JAmiga forum
Odyssey forum
OWB forum
Qt forum
SmartFileSystem forum
Timberwolf forum
TouchDevice forum
TuneNet forum
Unsatisfactory Software forum
Donate
