i686 vs x86_64 and keyboards [Solved]

night0wl · 2013-06-22 17:41:31

I am using a razer blackwidow keyboard to boot the x86-64 kernel off of the dual iso install disk. But have no success. I do have an AMD64 proc. For whatever reason, if I choose the 64bit kernel, I get generic device descriptor errors (-71, -32, etc.). However, when I choose the i686 kernel, it boots normal and I get no device descriptor errors related to USB. To rule out hardware issues, I have been using this keyboard successfully for over a year on my other hdd which is win7 x64. I am curious as to why I get errors when using virtually any live disk 64bit option? What might be some troubleshooting options for me besides installing i686, which is not an option for me? I do have a ps/2 keyboard on my desk, I am looking for a way to get it off my desk and use just the razer (with 64bit kernels).

Last edited by night0wl (2013-06-28 16:46:12)

Gulver · 2013-06-24 07:12:32

I don't want to sound cocky but what you are asking is community to compile the live cd kernel just so you can use your specific keyboard while installing?

Really, why don't you look for an additional driver for that particular keyboard. I'm sure there at least is one in existence. For the lulz, I'm not even expecting intel quad cores to be supported by default

drcouzelis · 2013-06-24 13:05:12

I'm stumped. I did a quick Internet search (as I'm sure you did) and found no information on the subject. I can barely even think of any suggestions. You should be able to type with your keyboard using any x86_64 bit Linux kernels, just like you can with i686 kernels.

Let's see... Do you have a USB to PS2 adapter you can try using? How about older versions of Linux? Does your keyboard work with, say, Ubuntu LTS 64 bit? Does 64 bit Arch Linux work on your computer with a DIFFERENT USB keyboard?

Gulver wrote:

I don't want to sound cocky but what you are asking is community to compile the live cd kernel just so you can use your specific keyboard while installing? Really, why don't you look for an additional driver for that particular keyboard.

No, that's not what night0wl is asking. Also, he shouldn't need a special driver. All USB keyboards should just work on any operating system released in the past 20 years.

...Unfortunately, when something really really should "just work" and it still doesn't, my first thought is that it's a hardware problem.

night0wl · 2013-06-24 15:06:10

@Gulver

No, this is not at all what I was asking, you are way off. And I mean way off.

@drcouzelis

You are correct. Any ole keyboard should just work. x86_64 kernel or not. It would make more sense if just the macro keys or something didn't work that were specific. The problem I am having is that it does work with other operating systems, im using it right now actually; and even Linux technically. I think* the issue is the keyboard is not given enough time to be detected. I can boot into my existing Linux system with my ps/2 keyboard and my razer still won't work. Then, I can unplug the razer and plug in and it sometimes works. Very strange. At one time, oh maybe around 3.1-3.2 kernel (1 year ago'ish on stable distros), the keyboard just worked fine. Seems after some kernel it stopped functioning as it should. It does seem to work with Crunchbang live, which is roughly at that kernel level.

As an update though. The problem does scream hardware issue. But it just plain difficult to prove considering what I have mentioned and that it does work with other systems. However, there are two USB cables for this keyboard, one for the keyboard+back-lighting itself and PORT power. Removing PORT since I don't use it, and moving the keyboard plug to the 3.0 usb slot on my mobo seems to give me a functioning keyboard through kernel boot. But there end up being so many usb errors anyway, that boot fails. Weird. I've been working on this for sometime. Trying kernel options, searching, asking, etc, etc, etc......The usb to ps2 idea sounds pretty good actually. I think I may give that a whirl.

EDIT:
Things that work:
1.) Works on i686
2.) Works in Win7 64
3.) Works after unplugging and reinstalling for a few minutes on existing 64 bit distros.
4.) Works when removing port plug and moving keyboard to usb 3.0 spot on mobo.
5.) Works on older distros live disk, using older kernels, 64 bit.
6.) Other usb keyboards do work, even a wireless one.

Last edited by night0wl (2013-06-24 15:16:19)

straykat · 2013-06-25 07:44:16

night0wl,
This may have something to do with the low latency (1000Hz) of these keyboards?
The stock Arch x64 kernel has a timing (latency) of 300Hz. I have found that if I compile my kernel with a timing of 1000Hz my Logitech MX5000 keyboard &mouse are affected adversely with mouse "clicks" regularly failing to be actioned & occasionally a key stroke was getting missed. The keyboard would also regularly disconnected from the dongle & sometimes on boot I would need to replug the dongle to get both working. I put this down to too low a latency. Same as yourself this started to happen a few kernels ago.

Maybe the same is happening with your keyboard all be it the other way round, i.e. the system is not "seeing" the keyboard because of it's low latency?

With my kernel compiled with a timing (latency) of 1000Hz my system is also not able to connect to my HP Photosmart C5380 printer.

Looking at the kernel log when this happens it appears as though the system is aware that a USB device has been connected but is unable to actually make the connection so the system loops into a cycle of trying to connect to the connected USB device till it eventually gives up (times out?) or very occasionally makes the connection stick.

I have no idea why latency affects some peripherals but I'm sure someone does.

Last edited by straykat (2013-06-26 05:12:58)

Roberth · 2013-06-25 10:59:29

Shouldn't both kernels be compiled with the same settings in this regard by default?

drcouzelis · 2013-06-25 13:41:15

straykat wrote:

This may have something to do with the low latency (1000Hz) of these keyboards

@Roberth: I think the easiest way for me to test this out on a computer that doesn't have an operating system would be to try a live CD with a "media content creation" Linux distribution with low latency kernel settings, such as Ubuntu Studio 64 bit.

night0wl · 2013-06-25 20:11:27

The idea of kernel latency sounds promising. I'm actually torrenting the current Ubuntu studio 64 right now to test out. Searching for some information about latency brings much more information to the table. Let me try this out and get back to you guys on this. If we can verify this is actually what the issue is, then we can make a quick fix for it no problem. I gotta say, this is the first tidbit of information about this issue that seems to be leading somewhere. We shall see.

night0wl · 2013-06-28 16:40:10

Alas, I have found a solution. And I am amazed at how simple this is. First off, I never got around to actually testing Ubuntu Studio. It drops to an intramfs shell and the md5 does match my download, so I opted to say screw that. However, that issue is very likely solved by my solution as well. Because things that were buggy such as wifi, keyboard, any USB are now fully functional and blasting along.

My motherboard is a Gigabyte 970a. It has the American Megatrends BIOS, current. Within the peripherals tab of the BIOS admin panel, there is a option called "IOMMU". This option is NOT enabled by default. For whatever reason, when enabling this, virtually every USB related problem I have been having disappears. Searching the web for similarities, I was able to find the exact issue with other people, and they solved it the same way.

Before this, I have never heard of IOMMU. Previously I was overclocking with this motherboard and the older 870a model. Within the past year I have switched motherboards, and all I can think of is that somehow in that grey area, the IOMMU option was enabled for whatever reason. I know that since I have been troubleshooting this issue, I have been turning things on and off in test as well as resetting to BIOS default nearly daily, so IOMMU would have been off. The problem is-is that the BIOS provides no clue or description to what IOMMU is for. So I never messed with it. All I know now is that Linux USB works great for me now in with 64 bit kernels. Hopefully people will find this thread and fix this ridiculous issue.

http://ubuntuforums.org/showthread.php?t=2114055 <-----Very relevant

http://unix.stackexchange.com/questions … -uefi-bios

drcouzelis · 2013-06-28 17:06:22

Hooray!

night0wl wrote:

Within the peripherals tab of the BIOS admin panel, there is a option called "IOMMU". This option is NOT enabled by default.

Why do we even HAVE that option?? They should just label it "Fix everything?".

Arch Linux

#1 2013-06-22 17:41:31

i686 vs x86_64 and keyboards [Solved]

#2 2013-06-24 07:12:32

Re: i686 vs x86_64 and keyboards [Solved]

#3 2013-06-24 13:05:12

Re: i686 vs x86_64 and keyboards [Solved]

#4 2013-06-24 15:06:10

Re: i686 vs x86_64 and keyboards [Solved]

#5 2013-06-25 07:44:16

Re: i686 vs x86_64 and keyboards [Solved]

#6 2013-06-25 10:59:29

Re: i686 vs x86_64 and keyboards [Solved]

#7 2013-06-25 13:41:15

Re: i686 vs x86_64 and keyboards [Solved]

#8 2013-06-25 20:11:27

Re: i686 vs x86_64 and keyboards [Solved]

#9 2013-06-28 16:40:10

Re: i686 vs x86_64 and keyboards [Solved]

#10 2013-06-28 17:06:22

Re: i686 vs x86_64 and keyboards [Solved]

Board footer