Re: OT: Why you should not use Windows : issue 1

[Tim Streater]

But you don't have to fart about *during* booting* to change those settings. Thass the point.

[John Rumm]

No, being a professional is not "why" at all - in fact is has no connection beyond suggesting you have the skills to deploy other options.

The real "why" comes down to your preferences, your workflow, the applications and platforms you run, and your budget among many other factors.

You probably don't play modern 3D games, which need full access to decent hardware rather than a VM. You probably don't have a workbench where you will install customers' PCs for work and diagnosis - where they won't be setup with any form of screen sharing, may not even boot, and where you frequently do want access to the BIOS etc. You certainly don't want them to have access to your network until you have verified they are "safe". Perhaps you are not configuring out of the box PCs every week, or maybe you have acres of desk space to waste with extra monitors and keyboards.

You may be content to throw away some of the performance of your systems by hiding them behind virtualised platforms - you are either running applications that are not demanding or prepared to take the performance hit.

You probably don't develop communications software where you want the resources of multiple physical machines for stress and performance testing. VMs are ok for some of it, but not all.

I *do* use VMs for running Linux (redhat, centos, fedora) when I want to prove or test stuff before it gets onto our "real" (i.e. virtual now) server. I also use them for access to old legacy stuff that is difficult or impossible to sensibly setup on current hardware. There are times where I use a windows VM on a windows host - either because I need a "disposable" PC, or need to for compatibility reasons.

I sometimes want to setup video presentations etc on a Raspberry Pi for use in POS applications. You can SSH to them, run a remote XServer etc, but you can't get the full graphics and video performance out of anything other than the console display connector.

The list goes on...

Why not just accept that your way is not the only way. It may be the best way for *you". Its certainly the way the big boys are going, and it makes a huge amount of sense in a cloudy server based world. That does not make it a universal solution for everything.

Remember when dribble used to bang on about how a combi boiler was the best and only solution to every heating a hot water need?

Guess how you sound?

[John Rumm]

Yup I have seen that type of kit, and it would do the job, However it does need to be in place before the fact. There are times I would like to be able to do it for a machine that I have never even seen before and has not be preconfigured. i.e. just be able to get remote control of it, and tweak some BIOS settings. Its not too bad when there is a client present and the machine is usable - if needs be you can talk them through the process. Its somewhat harder when its a non PC looking PC running headless in rack of scary looking wires.

[The Natural Philosopher]

No, the point was that *YOU* said "No need to dick with the firmware at all."

I was merely pointing out that that was false. Its the same on my lappie when I fired it up on win10 for the first and only time.

Lacking documentation as to what key to press to change boot order to 'DVD first' in order to install Linux, I used the little windows appy to change the boot order.

After fighting my way through all the windows 10 crapware to get it actually booted.

Its possible top mes with the bios setting - well the NVRAM or flash RAM setting - via an app but it has to be written FOR that bios and hardware.

[John Rumm]

Yup, CMOS I could tweak in DEBUG from DOS (accessed indirectly via a couple of IO ports)

Indeed - chances are it would need a different solution for each BIOS and possibly each MoBo vendor.

[The Natural Philosopher]

No, I am or was a professional....

(Actually I do, but its a Linux one: I would find the idea of a kvm switched game a bit weird as well).

[The Natural Philosopher]

It always was in some sort of NVRAM - used to be battery backed, now overwhelmingly flash of some sort.

[Tim Streater]

Thass right. No need to do it using the firmware. You can do it under the OS, using the gui and, I expect, from the command line too (although why bother). The BIOS is the firmware (in a generic PC) and the settings are in NVRAM or whatever.

Hopeless, eh? And if you're doing it via an app then I don't see why the BIOS would be involved at all. It's the NVRAM you're modifying. Shame no one seems to have bothered to standardise NVRAM locations and values for different functions then, if it's going to be different for every different PC on the market.

[The Natural Philosopher]

NVRAM is part of the firmware

[Vir Campestris]

That's an interesting statement.

I'd say that this particular bit of NVRAM is used by, and controlled by, the firmware.

That doesn't make it part of the firmware any more than my computer is part of me.

There are of course other types of NVRAM not owned by the system firmware at all.

Andy

[Mike Tomlinson]

En el artículo , John Rumm escribió:

Understood totally. As you said earlier, everyone's requirements are different. I offered it as an option in case you were not aware of it.

[Mike Tomlinson]

En el artículo , John Rumm escribió:

And the checksum? Checksum wrong = "CMOS checksum incorrect, defaults loaded, press F1 to continue or press F2 to run setup" or similar.

You're really going to tell me you hand calculated the checksum and poked those values into the relevant CMOS locations?

[The Natural Philosopher]

The police won't come knocking if you steal the RAM from the firmware..

when is softare and firmware only the code, and not the data?

AFAIAC 'firmware' is anything that is involdved in code that persists through power recykling

Particalarly since the same flash is used for both the data and the code in modern bioses.

[John Rumm]

You mean you no longer have the reaction times required? ;-)

Can't see why playing a game through a KVM switch would be any different from playing the same thing with the peripherals plugged directly into the machine.

[John Rumm]

It was contained in the RTC chip itself at one time - usually in battery backed SRAM. You could peek or poke its registers via I/O operations on I/O ports 0x70 and 0x71.

(In true PC tradition, it had a bunch of unrelated facilities lobbed in there, ranging from keeping the time, remembering what disk drives were installed and the CHS settings of a hard drive, along with obscure stuff like enabling or disabling the NMI interrupt line to the CPU).

[John Rumm]

ISTR one could "cheat"... tweak the value, reboot, wait for a checksum error, enter the CMOS setup, change nothing, and then save settings on exit. Handy for those occasions where you needed to set something that the manufacturers supplied setup program did not have the control available.

(remember when "setup" was not even in ROM and came supplied on the system floppy?)

Reading it was probably more practical use, so as to access the RTC directly, rather than rely on the more "fluffy" system time (especially when you know that you have pinched the system background timer tick interrupt, and not bothered to chain back to the DOS one since you don't want that overhead in your code ;-)

[T i m]

Wasn't that also seen on EISA expansion boards and slots ... or was it MCA (or both)?

(You ran a utility that showed slots and boards and would auto configure them (IRQ, DMA, I/O and DMA) for you if you wanted?)

Cheers , T i m

[RayL12]

Linux runs the Space Station. The 'United Space Alliance' decided enough was enough.

[T i m]

Linux is 'used' on the ISS, yes, however:

"International Space Station incorporating more Linux computers ..."

"[Update 5/10/13 9 p.m. EST: We heard from Kieth Chuvala below who said his comments with the Linux Foundation have been misconstrued. The ISS does use Linux as well as Windows, and has no plans to ditch Windows any time soon. This story has been updated to reflect Chuvala?s correction.]"

Or, more like the OS they had been using up to then for what, 12 years, had just gone out of support and they used that opportunity to go to several different Linux distros because:

"Chuvala said both organizations were interested in an open source operating system like Linux because it would give them in-house control. ?So if we needed to patch, adjust or adapt, we could,? "

It's like all the Linux nerds getting excited about Steam 'now being available on Linux' [1] ... again, only 10 years after being available on Windows!.

Oh well, better late than never. ;-)

Cheers, T i m

p.s. I wonder how many years it will be before NASA replaces the Microsoft HoloLens headsets on the ISS with Linux based ones? ;-)

[1] Did steam ever actually release the long awaited 'Steam Engine' Linux based games console?

[John Rumm]

Yup, they also came with config programs (which did away for the need to tweak jumpers etc on the cards).

Yup... kind of like what guys designing the Amiga managed in 1985 ;-)