osdir.com
mailing list archive

Subject: Re: Follow Up: i810/815 board not booting - msg#00741

List: os.solaris.solarisx86

Date: Prev Next Index Thread: Prev Next Index

Ben Taylor wrote:
> Dan Mick <dan.mick@xxxxxxx> wrote:
>
>
>>BK wrote:
>>
>>>Hi
>>>
>>>I have now experimented a bit and found that the reason why the
>>>i810/815 mobo doesn't boot Solaris is not the presence of USB but
>>>instead it is the absence of a PS/2 mouse.
>>>
>>>You can boot with USB mouse and keyboard attached as long as you also
>>>have a PS/2 mouse plugged in.
>>>
>>>This is rather silly and something that SUN should fix. There is no
>>>reason why a system shouldn't boot if it can't find a mouse or
>>>keyboard, or a mouse or keyboard of a certain kind.
>
>
> Have to admit, he was making a pretty good argument.
>
>
>>Trust me, it's a lot more subtle than "there's code that refuses to
>>proceed without a PS/2 mouse", as you surely must realize. There are
>>all kinds of systems with no PS/2 mouse installed.
>
>
> Yep. Though there should be some rational explaination
> why it wouldn't start without the PS/2 mouse.

Of course there should. My mail was intended to convey a sense of
frustration about diagnosing this problem, while indicating that it's
some sort of BIOS interaction, and also to convey that I clearly don't
have the answer yet, but it's far from universal. In fact, finding
systems where this fails is difficult. It's certainly not just "an i810
system". I have one system where it seems to do this about 1 in every
20 boots.


>>PS/2 and USB devices are *horrific* on PC hardware, because the USB
>>legacy emulation is done by trapping accesses to the hardware ports into
>> the SMI BIOS, which is unexaminable and unstoppable (there's no way to
>>disable it, no way to mask it, no way to examine the code, no way to
>>even tall it happened).
>
> So, as long as you're not in legacy USB mode, things
> are better, or are PS/2 and USB still pretty dodgy?

USB keyboards require legacy emulation with Solaris, because Solaris
(for various *older* BIOS issues, and for the debuggability I mentioned)
tries to run the keyboard directly thru the ports. If USB were a sane
protocol that didn't require hundreds of K of driver support, we could
do a native USB driver too, but that's what USB legacy support was
supposed to obviate. And of course it mostly does. Until it breaks.
And then it doesn't. And there's no one to blame.

>>We may end up in the Linux state soon if I don't get some more
>>cooperation with BIOS vendors.
>
>
> Thanks for the great explaination.
>
> Does running a USB keyboard on a PC require that you run
> USB legacy mode?

Unless you have a native USB stack, which we do once the kernel is up,
but not until then.


Please check the Links page before posting:
http://groups.yahoo.com/group/solarisx86/links
Post message: solarisx86@xxxxxxxxxxxxxxx
UNSUBSCRIBE: solarisx86-unsubscribe@xxxxxxxxxxxxxxx




Find Programming Jobs at git.net
(osdir sister site)

Thread at a glance:

Previous Message by Date: (click to view message preview)

Re: Re: BGE interfaces on Build 74.

John D Groenveld wrote: > In message <41E718CE.4030108@xxxxxxxxxxxxxx>, Bruce Riddle writes: > >>That doesn't answer the question, and I don't administer the network! > > > If your corner of the Dilbert-sphere is anything like mine, here's my > advice on how to proceed: > > Confirm your V20z autonegotiates with a known good switch and a known > good cable. You may be able to expense these on your purchasing card > or with petty cash. > > If it doesn't work, file a service order with Sun. > > If it does work, create a work-order for your netnazis to test the wiring. > > If it tests bad, enter the dark hell of facility/physical plant but at least > your netnazis will be on your side when they go after the electricians. > > If it test good, ask your netnazis to set the port to autonegotiate but > also confirm they're up to date on their Extreme firmware. > > If the V20z still doesn't autonegotiate, file a service order with Sun. > That's all nice. It doesn't answer the question: How to force the speed and duplex on boot? I've added a startup script to make sure it's set until I get a better answer. Please check the Links page before posting: http://groups.yahoo.com/group/solarisx86/links Post message: solarisx86@xxxxxxxxxxxxxxx UNSUBSCRIBE: solarisx86-unsubscribe@xxxxxxxxxxxxxxx

Next Message by Date: click to view message preview

Re: Follow Up: i810/815 board not booting

John D Groenveld wrote: > In message <41E71815.6030002@xxxxxxx>, Dan Mick writes: > >>We may end up in the Linux state soon if I don't get some more >>cooperation with BIOS vendors. > > > Can the HCTS be used to provide customers some leverage? > > Ie is there a way to warn customers, even if the HCTS submitter only > tested with a PS/2 mouse installed, that the BIOS is going to give > them fits if they try to use a USB mouse? I guess I didn't express sufficiently how little information we have, or indeed can have, about this problem. It's pretty limited so far. (Of course you would get a different answer from irate users like BK, but truly, we've got a lot of systems, and it's not happening very many places yet.) USB devices are really a pain during boot. The industry failed us all here. Please check the Links page before posting: http://groups.yahoo.com/group/solarisx86/links Post message: solarisx86@xxxxxxxxxxxxxxx UNSUBSCRIBE: solarisx86-unsubscribe@xxxxxxxxxxxxxxx

Previous Message by Thread: click to view message preview

Re: Re: Follow Up: i810/815 board not booting

Dan Mick <dan.mick@xxxxxxx> wrote: > BK wrote: > > Hi > > > > I have now experimented a bit and found that the reason why the > > i810/815 mobo doesn't boot Solaris is not the presence of USB but > > instead it is the absence of a PS/2 mouse. > > > > You can boot with USB mouse and keyboard attached as long as you also > > have a PS/2 mouse plugged in. > > > > This is rather silly and something that SUN should fix. There is no > > reason why a system shouldn't boot if it can't find a mouse or > > keyboard, or a mouse or keyboard of a certain kind. Have to admit, he was making a pretty good argument. > Trust me, it's a lot more subtle than "there's code that refuses to > proceed without a PS/2 mouse", as you surely must realize. There are > all kinds of systems with no PS/2 mouse installed. Yep. Though there should be some rational explaination why it wouldn't start without the PS/2 mouse. > PS/2 and USB devices are *horrific* on PC hardware, because the USB > legacy emulation is done by trapping accesses to the hardware ports into > the SMI BIOS, which is unexaminable and unstoppable (there's no way to > disable it, no way to mask it, no way to examine the code, no way to > even tall it happened). So, as long as you're not in legacy USB mode, things are better, or are PS/2 and USB still pretty dodgy? > > We may end up in the Linux state soon if I don't get some more > cooperation with BIOS vendors. Thanks for the great explaination. Does running a USB keyboard on a PC require that you run USB legacy mode? Ben Please check the Links page before posting: http://groups.yahoo.com/group/solarisx86/links Post message: solarisx86@xxxxxxxxxxxxxxx UNSUBSCRIBE: solarisx86-unsubscribe@xxxxxxxxxxxxxxx

Next Message by Thread: click to view message preview

Re: Follow Up: i810/815 board not booting

Dan Mick <dan.mick@xxxxxxx> wrote: >>>Trust me, it's a lot more subtle than "there's code that refuses to >>>proceed without a PS/2 mouse", as you surely must realize. There are >>>all kinds of systems with no PS/2 mouse installed. >> >>Yep. Though there should be some rational explaination >>why it wouldn't start without the PS/2 mouse. > > Of course there should. My mail was intended to convey a sense of > frustration about diagnosing this problem, while indicating that it's > some sort of BIOS interaction, and also to convey that I clearly don't > have the answer yet, but it's far from universal. In fact, finding > systems where this fails is difficult. It's certainly not just "an i810 > system". I have one system where it seems to do this about 1 in every > 20 boots. It seems that simply printing what is being probed would go a long way in helping problems like this out. > USB keyboards require legacy emulation with Solaris, because Solaris > (for various *older* BIOS issues, and for the debuggability I mentioned) > tries to run the keyboard directly thru the ports. If USB were a sane > protocol that didn't require hundreds of K of driver support, we could > do a native USB driver too, but that's what USB legacy support was > supposed to obviate. And of course it mostly does. Until it breaks. > And then it doesn't. And there's no one to blame. You have my sympathy, believe me. BUT, there really should be a way for modestly experienced people to diagnose WHAT is causing the problem. Too often, solaris is just mum about what is going on, which makes it very difficult to figure out something as simple as what is failing. Why there isn't a good verbose in the bootloader is beyond me. -r Please check the Links page before posting: http://groups.yahoo.com/group/solarisx86/links Post message: solarisx86@xxxxxxxxxxxxxxx UNSUBSCRIBE: solarisx86-unsubscribe@xxxxxxxxxxxxxxx

Web Hosting Reviews from OSDir.com Sister Site iBizWebHosting.com

Home | News | Patents | Sitemap | FAQ | advertise | OSDir is an Inevitable website. GBiz & git.net are too!

Advertising by