@kas1e UBoot/CFE/whatever all seem to be very restrictive in the types of keyboards they can handle. I've found they generally don't work with keyboard + mouse combos (which is what I have), and it wouldn't surprise me if they don't handle KVMs well either.
With the X1000 I bought an extremely cheap keyboard specifically for use with CFE. I don't bother with that any more, because UBoot responds to key input over the serial port. So, I use PuTTY or Term (on AmigaOS) from another machine.
@Hans At least for me on x1000 (cfe) and x5000 (uboot) it works surely via kvm in all places, and in cfe and in uboot. Just pegasos2 and sam460 acts strange with their firmwares, while sam460 have uboot as x5000..
Anyway, will be interesting to know, from technical pov what make keyboard not works through kvm, while directly it works for sure. I mean its usb2 in end.. maybe need some additional usb combiner or something..
I will try today few different keyboards in different modes, maybe that will make difference. This kvm also have some "pass thru mode", like it should be the same like you connect it directly.
@All Tried all the ways, but still didn't able to make in pegasos2's firmware and in sam460's firmware keyboard to work (while once OS4 is loads up, everything works).
What i find intersting when experimenting with pegasos2 is this:
1. If i attach the KVM cable to pegasos2 (i.e. hdmi to hdmi-to-vga adapter in pegasos2 and usb part does not matter to which usb port) : then, no boot menu visibly on screen (but visibly on terminal). I assume that happens when no keyboard detected by peg2's firmware, and so it fallback to the "console" input.
2. If i put real keyboard to one of usb port , and usb part from KVM to another port : then, firmware showups on the real screen, and keyboard works of course.
Now, what i curious about is : what happens on the hardware level ?
I mean, pegasos2 check if real keyboard is attached to usb. If yes, then all ok, show the boot menu and you can type in. When we attach this combined usb cable pegasos2 can't detect that keyboard is attached (because it's coming from KVM, which send something, which pegasos2 firmware can't understand).
Question is : what KVM send by he's "combined usb cable" , if it usb2 protocol, and peg2 can't understand what is it.
I thinking about some kind of additional usb-splitter/combiners which maybe can somehow fix that for me, so peg2 will think it is "real" keyboard.
Even if this one for serials didn't work by any of reassons , the box itself and the serial ports placed on, so i can resoldering anything inside while have the same good box. And as i see, this one also not need additional power, USB is enough.
@TiredOfLife Yep, they all similar that they passive, have no needs for additional power, and many "in" to single out. Through the one i order (link in previous post) allow you to not swith , but have all of them at the same time without loosing of sound quality. So i can attach 4 amigas to it, and be done with it forever, all of them will works when need it. Also volume control of each "in" and a mixed "out" helps as well. Will see once it come here, currently fighting with understanding how to make firmwares of peg2 and sam460 to works over KVM.
@All About keyboard non working in Firmwares of PEgasos2 and sam460 when attached via KVM :
Interesting, if this will help : get some small simple usb switch, to use it with keyboard only, so, attach to it from pegasos2 and sam460 , and then, attach to it KVM's usb and keyboard. Maybe it will cheat peg2's and sam460's usb ports when they query for keyboard ?
I just trying to understand what differences in peg2's and sam460's usb keyboard protocol, that is not good enough for KVMs..
KVM issues on Sam460 are not confined to the firmware in my experience.
I can have the keyboard and mouse working fine initially. But after switching to another machine and back again to the Sam460, issues can then occur. Keyboard, mouse or both not working. Switching back and forth again can get them back running again. Often, you will see the message on screen flash up to show that the system recognises they are there but still don't work.
This occurs when switching between OS4.1 and other OSes.
@TiredOfLife That is probably how your KVM acts.. My one (the TESmart one) didn't have issues when switching between working machines. Everything is fine; you can use KVM mode or "pass through" mode (which simulates how things are connected). So you can have it when switches to look like the attached first time, or you can choose like they attached directly always even when you switch.
Switching is fine with the KVM I have (at least as far as I can tell); there are no loose items, no "resync" of images, and everything is as fast and accurate as it should be. That is probably what Daniel meant when he said those KVMs were good.
However, I am unable to make keyboard to work in firmwares on the Sam460 and PEG2. Even if I connect the only keyboard to KVM, and connect it not to the mouse or keyboard port on KVM but to the additional USB port they have for other devices, then I still can't cheat the firmware on Peg2 and Sam460 with that keyboard attached; it just doesn't work. When I plug in outside of KVM (i.e. directly to motherboards) it works as should.
Really, what I want to understand is what exactly happens. That is, why can't KVM handle firmware on Peg2 and Sam460? What didn't it send or receive at the USB protocol level? Because if I understand what's going on, I'll be able to solder a small device and add the missing components to a small box.
Perhaps I should get a hardware USB analyzer and see what it waits for and what differences there are when a keyboard is connected directly or via a USB2 KVM port.
It will be easier to test with Pegasos 2 because when Pegasos 2's firmware didn't detect a keyboard in the USB port attached, it displayed only the bplan logo and all bootmenus and amigaboot menus were sent via to serial. As a result, it may be simple to test.
I'm considering adding a passive USB splitter to the keyboard, attaching the keyboard to it, and connecting to KVM, PEG2, and Sam460 from this one. So firmwares can be used from the keyboard, and when OS4 boots, I can use them as well (but now through firmware or whatever).
Or maybe I just need a passive USB splitter on 4 ports and will use this one for all 4 machines in addition. I mean just to not use KVM for keyboard's switching, but something more simple and passive one, which can split the keyboard on 4 outs.
If you plug a keyboard into a USB hub and then plug that hub into the Sam460 or Peg2, does the keyboard work with the firmware?
The firmware's USB keyboard support could be very basic. On some machines, the firmware will only use the keyboard if it's plugged into the first USB port. So, it's entirely possible that it won't work behind a hub. If it doesn't work behind a hub, then that may be why it doesn't work with the KVM.
If you plug a keyboard into a USB hub and then plug that hub into the Sam460 or Peg2, does the keyboard work with the firmware?
Yeah, it works then. I do have some simple 4port hub, and when i connect keyboard to it, and connect hub to pegasos2, then it works. But then, if i connect keyboaed to a hub, and connect hub to an usb2 port of KVM, and KVM's cable to pegasos2 - then it didn't work in firmware.
Yeah, it works then. I do have some simple 4port hub, and when i connect keyboard to it, and connect hub to pegasos2, then it works. But then, if i connect keyboaed to a hub, and connect hub to an usb2 port of KVM, and KVM's cable to pegasos2 - then it didn't work in firmware.
Okay, so being behind a hub isn't a problem.
Are you plugging it in to the KVM's keyboard port? Or to one ofthe generic USB ports? The KVMs that I've used usually have a specific USB port for the keyboard, another for the mouse, and then a few generic USB ports.
1). "Logitech K120" keyboard connected to the first USB port on pegasos2 : works in firmware fine.
2). A 4 port usb hub connected to the same first usb port of Pegasos2 and to the hub connected the same logitech k120 keyboard: works too. But this hub it just this:
3). When keyboard connected to a hub, and hub connected to pegasos2 usb port, i connect also usb cable from KVM to the HUB : keyboard still works in the firmware, and when aos4 loaded up, mouse also works fine.
But then, at this point, it's the same as if i connecte keyboard to the usb1 , and one of 4 KVMs cables to second port. But at least at this point we know that HUB make no problems with for sure.
4). Now start playing with KVM.
a). connected keyboard to a hub, hub to KVM's "keyboard" port, and KVM's spare USB cable (for mouse/button) to the 1st port of the pegasos2 , as result : it didn't work in firmware, but did work in OS4.
b). the same as in 1 , but connected KVM's usb cable to second pegasos usb2 port : same as in 1, didn't works in firmware.
c). repeat a) and b) just this time in KVM's "mouse" port : firmware didn't works.
d). repeat a) and b) just this time in KVM's "additional usb2" port : firmware didnt' works.
e). Tried a) b) c) d) but with disabled "pass through" mode on KVM : same results.
So.. in any case, when we connect keyboard to KVM (any port), directly, or via HUB, with or without "pass thru" mode we have the same issue : non working Firmware at all. But when OS4 loads up, it always work in any port, with or without HUB.
Question now is what KVM do or didn't do, which make firmware didn't get that we do have keyboard attached ?
Maybe there should be something which user can change about USB in the firmware (i mean configure it somehow different, like timings, or enable/disable legacy mode if that at all present, etc).
There many hobby devices which use Uboot and all kind of other firmwares, sure they meet same issues but deal with them somehow ?
As far as i understand that what happens:
The real keyboard do send everything it should send. Pegasos2 usb handler in firmware also checking on it and if ok, handle it. But when it go through KVM (does not matter what port, with or without Hub, with or without "pass thru" mode), KVM , do send something via this spare mouse/keyboard cable what pegasos2's firmware didn't take as working keyboard. But what ? What it can send of the sort that pegasos2 (and sam460 too, so 2 different firmwares), can't detect a keyboard ?
What i also find interesting, is that if i load up os4 and everyhting working , and i plug on/plug off keyboard or mouse from their ports in KVM, then os4 offten find a vendor as a "trash symbols", while description and class ok.
Also, i found that when i attach keyboard to "usuall usb2" port of KVM, then in firmware of X5000 everything works and all ok. If attach to Keyboard USB port of KVM or to mouse port, then i have when booting lot of "Timeout poll on interrupt endpoint" on serial. Once OS4 start loads up, those outputs gone. But when it's connect to this additiona USB2 port on KVM, no those phrases.
Short story: I use this KVM to make the keyboard on the Sam460 work not only when OS4 is loaded, but also in Uboot!
Long story:
After discussing things with TESmart support, they point out that KVM has not only mouse and keyboard USB ports but also an additional USB 2.0 port (which we knew before, of course), but they at least explain the differences between the mouse and keyboard USB ports and the additional USB port.
In summary, the mouse/keyboard port is only used to control mouse gestures (which we can fine-tune in OS 4) and to make special hotkey combos from the keyboard work (such as switching machines from the keyboard, changing modes, and some other uninteresting stuff). This means that when you connect a keyboard and mouse to those ports, additional programming logic and emulation layers are involved, which only compicate things for us.
More of it was discovered today: connecting a keyboard to this "keyboard" port does not work that good on the X1000 or X5000; on serial, it throws a lot of messages about "Timeout poll on interrupt endpoint.". So while functional, still bad. That is, even for X5000 and X1000 firmware, using the dedicated keyboard port is not recommended then.
And, as I previously stated, the Sam460 and Pegasos2 didn't have working keyboard in their firmware if the keyboard is in dedicated keyboard port.
Before, I tested the additional USB 2.0 port with only Pegasos 2, and assumed Sam460 and Pegasos 2 had the same problem, but they don't! Once I connected the keyboard to this additional USB 2.0 port, it then started to work fine, not only in the X5000 and X1000 without those errors on serial, but also in the Sam460 in both firmware and OS4 itself!
Regarding the mouse, it appears that it is also preferable to use an additional USB 2.0 port with a hub (to connect both the keyboard and mouse), because while the mouse works fine in OS 4 and does not require to works in firmware, OS4 occasionally detects it with fancy characters in the vendor's name whih can sideback with some issues.
So, for the time being, I have that cheap USB hub connected to the TeSmart KVM's additional USB 2.0 port: a keyboard connected to the first port of the USB hub (or the Sam460 will not work), and a mouse connected to the second port of the USB hub.
As a result, we have three operational machines with this KVM: the X1000, X5000, and Sam460; in AOS4 itself and in their firmwares.
As for Pegasos 2, no combinations help at the moment. But (!) find out interesting details: on the KVM switch on the front, you have LEDs for the online state. And for Pegasos2, this "online" state never lights up until OS4 starts loading. That is most likely why the keyboard does not work at all.
I'm not sure how KVM switches handle this "online" state from the computer, HDMI port via some commands, or USB, but for Pegasos2, I use a DVI-TO-HDMI adapter (as the Radeon 9250 only has VGA and DVI, no HDMI), and if KVM switches take the online state from HDMI, it could be that this adapter causes it to fail to query the state correctly, and thus the device is not online at start and nothing works. When the OS4 begins to load, it may send something to some of the USB or HDMI ports, and then things begin to work.
I'll ask the TESmart guys; they might know something. But at least now 3 machines surely works.
I bought two small passive USB hubs (Dezen UH34) to try this out.
One of them is now sitting on my main KVM switch, an ATEN CS1784A which connects my X5000/40, my X1000 and a PC to my monitor, keyboard and mouse.
Until now, I've had problems using the keyboard in UBoot on the X5000, where I had to switch the KVM switch to another machine for the first 22 seconds of the boot, and then switch back, otherwise the keyboard wasn't recognized. I also couldn't do a usb reset in the UBoot command line, it would get stuck/loop indefinitely and eventually crash or lock up.
Now those problems are gone!
So now I don't have to time anything and switch back and forth. Furthermore, I can again try getting Linux and MorphOs running on the machine (when I get time). This was blocked by the usb reset failures.
The other new USB hub is now on the ATEN CS 1764A connecting my Sam440ep and my Sam460ex. I had hoped they would react to the keyboard in UBoot as well, but that doesn't seem to have happened. I'm going to have to experiment more with that.
But my one success out of two alone is worth the effort!
I checked it out, and yeah, it does have an additional USB 2.0 port, just like the TESMART one. It appears that "normal" KVM developers frequently (?) add this additional port, so that when things go wrong with their EDIT-programmed keyboard and mouse ports, you have an additional USB2 port with no logic on top as a workaround.
Basically, for myself, having KVM with just the standard USB ports for input will suffice, eliminating the need for an additional (passive, at least) USB hub, but at least we found a way to make things work!
But... so far I have had no success with Pegasos 2's firmware, whereas when OS4 loads and the USB stack sends something to the USB bus (right at the beginning of OS loading), everything has worked since.
This TESMART KVM do have "online" leds, which activate only when hardware sends or receives some unknown data via USB, which Pegasos2's firmware did not. So, if the hardware fails to send something (I didn't know what at the moment), then KVM's led indicator for online dind't light up, and KVM believes there is no alive USB.
When OS4 on Pegasos2 starts loading, I can see that when my USB mouse is "lighted" (the mouse also has some leds), then the "online" indicator begins to show that there is an working connection, and everything works since.
So far, I'm considering two extreme solutions to the "firmware in Pegasos 2 does not work in KVM" problem:
1). Buy a hardware USB analyzer (the budget models cost around 200–300 dollars). Then I can sniff all the traffic and compare what Pegasos 2 sends to what other Amiga firmware sends.
2). Tinker with the firmware, i.e., hack it, in the usual way: unpack, disassemble, find out where USB routines start, and then try to send what KVM wants to be sent.
I also think that maybe there can be some kind of adapter that can be connected to the USB port of the Pegasos 2 and pretend to be a "working USB" so KVM will be cheated ?
In any case, the first step is probably to figure out what Pegasos2 didn't send and what KVM wants, and then cheat it (either in hardware or by hacking firmware).
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
