Author Topic: Motherboard woes  (Read 44364 times)

Andi B.

  • Hero Member
  • *****
  • Posts: 811
  • Karma: +11/-2
    • View Profile
Re: Motherboard woes
« Reply #15 on: December 31, 2014, 10:20:32 am »
Unfortunately I can't get SNAP to run on this Asus H87-Pro. I hoped it would run in GRADD mode but this Intel HD4600 does not.

Moreover even my plan B - insert a Matrox G550 - does not work. It hangs/stops at the same place where the CPU integrated HD4600 fails. PMSHELL does not come up. After the Scitech driver messages when the screen should switch to graphic mode nothing happens.

I tried all PCIe sockets and BIOS settings I found without success. Still can not believe why an extra PCIe graphic adapter does not work but have to give up after days of frustration :( Seems to be something special in this H87 chipset.

After years working with two 1920x1200 displays I can not switch back to single monitor mode anymore. Any suggestions for a Linux distro with usable desktop? No fancy transparency windows or rounded corners or washed out fonts. I hate these W7 annoyances with thick boarders which hide the useful information behind (the reason why they must be transparent but why did MS made them so think in the first place?) and switched back to classic (pre-XP) look and fell as much as possible. No big symbols and icons only but text description in menus. Simple clear and easy to use for a long time OS/2 user. Suggestions for desktop managers I should look for? Cinnamon?

TeLLie

  • Full Member
  • ***
  • Posts: 226
  • Karma: +11/-0
    • View Profile
Re: Motherboard woes
« Reply #16 on: December 31, 2014, 03:01:49 pm »
Hi Andy,

I have a Asus M5A78L-M usb3
Also have a Matrox G550 and i can't even install snap.

Here it hangs @ same place:

After the Scitech driver messages when the screen should switch to graphic mode

Also tried a Radeon X800XT but that also hangs @ same place.
.

ivan

  • Hero Member
  • *****
  • Posts: 1557
  • Karma: +17/-0
    • View Profile
Re: Motherboard woes
« Reply #17 on: December 31, 2014, 03:10:11 pm »
Did you set the bios to switch off the internal video (is that even possible?) and/or tell it to use the external card first?

I have to say that we have never had good results with intel chipsets which is why we have used AMD processors and ATI graphics, with Nvidia graphics as a possible fall-back.

Andreas Schnellbacher

  • Hero Member
  • *****
  • Posts: 827
  • Karma: +14/-0
    • View Profile
Re: Motherboard woes
« Reply #18 on: December 31, 2014, 04:07:09 pm »
Suggestions for desktop managers I should look for? Cinnamon?
Yes, that's my recommendation. Sometimes it makes problems with installing Grub2, but sooner or later one gets it installed. You could install eCS in a VirtualBox environment then, if you still have the need for it.

The other option is to buy other hardware. No recommendations here from me. I'll try an MSI AMD board next, probably in the middle of January, depending on the availability of all parts.

OS4User

  • Sr. Member
  • ****
  • Posts: 406
  • Karma: +10/-0
    • View Profile
Re: Motherboard woes
« Reply #19 on: December 31, 2014, 05:11:33 pm »
interesting to see comport log with hung

wesco

  • Guest
Re: Motherboard woes
« Reply #20 on: January 01, 2015, 12:09:36 am »
tellie and andy b. unforunately we were at your point 18 months ago. eCS is now really marketed to veiticle app space. windows 8 deletes the areo glass giving it an OS/2 2.x look which we actually preferred to OS/2 3 and later. linux is now so dumbed down that we dumped it too.

maybe some useful insight in this exchange as to where eCS priorities lie. it is not an attack on mr. klein. he was doing the best he could. OTH you can see why we felt dismissed and not taken with due seriousness as an organization that had a lot of $$$$ invested in OS/2 since 1993.


 
222 of 340
 
Re: Demo CD pre-download comment
Inbox
   x
Roderick Klein <rwklein@mensys.nl>
   
5/19/13
      
to supportteam, me
On Sun, 19 May 2013 21:06:51 +0100, xxxxxx@gmail.com wrote:

>Requested by: wesco
>Familiar: yes
>Usage: both
>Remarks:
>---
>the problems we have involve IRQ sharing. it may not be possible to solve. if it can\'t you should say so. usb, ethernet, sata, and video adapters can never share IRQs in os/2 and work.

I keep hearing this more often that OS/2 can not share PCI IRQ's. But
let me just put simple, its a fairytale. This is the output of rmview
on my laptop.

  IRQ Level = 16  PCI Pin = A     Flg = SHARED       UHCI_2 Universal
USB Host Controller
  IRQ Level = 16  PCI Pin = B     Flg = SHARED       AHCI_0 Controller
  IRQ Level = 17  PCI Pin = NONE  Flg = SHARED       HDA Intel
  IRQ Level = 17  PCI Pin = A     Flg = SHARED       Intel 4965 a/g/n
  IRQ Level = 17  PCI Pin = B     Flg = SHARED       UHCI_3 Universal
USB Host Controller
  IRQ Level = 20  PCI Pin = A     Flg = SHARED       e1000e_0 Intel
82566MM Network Adapter
  IRQ Level = 20  PCI Pin = A     Flg = SHARED       UHCI_0 Universal
USB Host Controller
  IRQ Level = 21  PCI Pin = B     Flg = SHARED       UHCI_1 Universal
USB Host Controller
  IRQ Level = 22  PCI Pin = C     Flg = SHARED       EHCI_0 Enhanced
USB Host Controller

3 devices on IRQ 17 for example. My Intel Wireless, USB controller and
Intel audio HDA chip.

Now if you run into certain problems that please let me know and I can
see if we can pin point the problem. Some older NIC and sound drivers
had bad code when it came to sharing IRQ's, But we don't ship these
driverds with eCS 2.2 live CD,


>we bought an AsRock Mobo with 3 pci slots thinking it would be a great borad. the bios has no way to not assign irq to the video card, there is no setting to turn off \"plug -n- play\". the strides made are commendable; but not addressing more fundamental issues like this make the effort pointless.

If you could tell me what problem you run into with the demo CD ? It
could be a problem with ACPI ? Did you boot the demo CD with standard
options  ?
The only thing I can make up from your email is that you mention
something about IRQ sharing problems. But not what type of problems you
experience with the demo CD.

Regards,

Roderick Klein
Mensys B.V.
wesco <xxxxx@gmail.com>
   
5/19/13
      
to Roderick
well, new demo is a vast improvement over the first. the new realtek driver works. the usb works. however the snap with ati x850 hangs. dan caetano reported the same problem. we bought an i7 3770k processor. we wonder if the built in GPU confilcts.

i have 20 years experinece setting up os/2. i even preloaded it for resale. the often seen complaint on this board about the ethernet driver loading but no access is cured by changing IRQ to one not shared with another that also finds constant use like the hard drive or video. i had a machine where if the bios "do not allocate IRQ to video adapter" was not set to yes the shell would freeze the moment you used a network resource. we have not had problems with shared IRQs with USB as along as we weren't trying to use a scanner or others. on a machine with only one usb controller that had a bios settable IRQ the scanner with tame works flawlessly. otherwise, the system would freeze. mem sticks act as if they work in a polling mode and indeed we have had few problems with them. now,  maybe that the new drivers work; but the SNAP not working is a deal breaker. we have a video capture adapter that also MUST be working.

my personal experience is IRQs can be shared with caution. our new mobo doesn't have a ps/2 connection unfortunately. on one machine i spent 4 months rearranging ethernet adapters til i finally found what sockets to use and what to turn off in the bios. since the machine is a server i did not need sound or IDE support.

i know you are under a lot of fire and am sorry about that. i do agree with siguard that for the end user a list of known supported hardware would be nice. appearantly you have 10 machine that work well. what is their make up? are they current (less than 1 year old)? i think stating who your real intended customer are would allow many sitting on the fence technically would be a good move on the part of serenity systems.

any word on a new ecs beta. we run high end hardware and think we are candidates for beta testing.
Roderick Klein <rwklein@mensys.nl>
   
5/19/13
      
to me
On Sun, 19 May 2013 18:32:04 -0400, wesco wrote:

>well, new demo is a vast improvement over the first. the new realtek driver
>works. the usb works. however the snap with ati x850 hangs. dan caetano
>reported the same problem. we bought an i7 3770k processor. we wonder if
>the built in GPU confilcts.

Did you select SNAP in the preboot menu of the demo CD.
SNAP in modern hardware will not work for 2 reasons.
Some systems become hell unstable with SMP support.
Next to SMP support itself we have this problem.

Read these tickets:
http://svn.ecomstation.nl/snap/report/1

Thats why for now you should use SNAP on modern systems. The video card
you use
is OK. But MTRR registers also depends if you run in SMP mode.

Currently SNAP has been pushed to background to make a gigantic leap
and bound forward
with ACPI, AHCI, USB and Panorama. Later this year we hope to get USB
3.0 and SNAP updates done.
Until that time on SMP systems and SNAP should be avoided.

In other words a SNAP supported video card might not work stable a
modern systems,
especialy when using SMP (multiple CPU's switched on).

>i have 20 years experinece setting up os/2. i even preloaded it for resale.
>the often seen complaint on this board about the ethernet driver loading
>but no access is cured by changing IRQ to one not shared with another that
>also finds constant use like the hard drive or video.

Well what driver do you load for the NIC chipset ?

>i had a machine where
>if the bios "do not allocate IRQ to video adapter" was not set to yes the
>shell would freeze the moment you used a network resource. we have not had
>problems with shared IRQs with USB as along as we weren't trying to use a
>scanner or others. on a machine with only one usb controller that had a
>bios settable IRQ the scanner with tame works flawlessly. otherwise, the
>system would freeze. mem sticks act as if they work in a polling mode and
>indeed we have had few problems with them. now,  maybe that the new drivers
>work; but the SNAP not working is a deal breaker. we have a video capture
>adapter that also MUST be working.

What softwate do you use for that video capture board and what driver ?

>my personal experience is IRQs can be shared with caution. our new mobo
>doesn't have a ps/2 connection unfortunately. on one machine i spent 4
>months rearranging ethernet adapters til i finally found what sockets to
>use and what to turn off in the bios. since the machine is a server i did
>not need sound or IDE support.
>
>i know you are under a lot of fire and am sorry about that.

He its part of the job :-)  I'm the core guy that arranges all
technicall driver stuff.
I don't develop it but I do coordinate it and set the priorities.

>i do agree with
>siguard that for the end user a list of known supported hardware would be
>nice. appearantly you have 10 machine that work well. what is their make
>up? are they current (less than 1 year old)?

Sigurd is approach is nice in theory. But Rudiger Ihle  who write
Emperorar TV
also pointed at the flaws of this strategy,

The large companies that buy eCS normaly just call us and say "he we
have this model PC make it work".
Expect for NIC support we are getting in pretty good generic shape with
eCS 2.2. With the USB, ACPI and Panorama
and OS2AHCI support. Buy a box with an Intel gigabit or Realtek
chip(with the latest driver 0.31 most problems seem to be gone!) and it
will simply work.
Not every PC model (with a part number) can be bought worldwide is my
experience.

With the arrival of eCS 2.2 and the big leap forward in device support
I seriously consider things ti turn around.
We have tested even on about 50 different machines from 4 years old to
brand new servers and workstations.
The new DB2 server in the Mensys office with a Xeon 6 core CPU worked
out of the box.
A lot of other equipment we had access to seems to work just fine.
So far we received 1 or 2 bug reports of ACPI not working on native
hardware (apart from suspend/resume).
But this is a different story.

>i think stating who your real
>intended customer are would allow many sitting on the fence technically
>would be a good move on the part of serenity systems.

What do you mean by this ? The discussion of company versus private
customers ?
If that is your question I have mentioned so often we both support the
private/home user and companies.


>any word on a new ecs beta. we run high end hardware and think we are
>candidates for beta testing.

Do you have an active eCS software subscription ?
wesco <xxxxxx@gmail.com>
   
5/19/13
      
to Roderick
we tried the snap with ACPI set to 1 proc. no joy. we have never had issues with SNAP; thus our surprise. what we didn't try was the old PSD to see if intel MP support works. we have never had smp problems that weren't manageable every since the ALR revolution MP 90 we built in 1994. we even sent indelible blue a video that they could use to show off OS/2 2.11SMP.

you have 50 units working well. great!! what are they?? just tell us so we can just go buy one. how hard is that?

about intended customer. obviously there are some specific applications for os/2 such as automation, point of sale, etc. this is good business which should get priority. I however do not see much reason to update firefox, bother with flash, or even in a lot of cases the power management stuff in these applications. particularly one doesn't want any machine going to sleep that is running a nuclear power plant. these things, specially the flash, are in my mind end user products. what i am really getting at is don't you think it just simply time to call it quits. all of the people who have kept OS/2 alive and useable have done an amazing job and i would be completely understanding if they moved on to other things (i have smedley in mind). ihle's emperoar tv is an awesome program that with win TVR 350s have brought us great utility and pleasure. i notice his site hadn't updated this the video stuff; but was happy to see other ventures moving forward.

as an aside note i was able to get the own cloud iso version to run in a virtual machine under OS/2. it was neat to connect my iPad to the warp server.


for us SMP and accelarated video ARE REQUIRED. i appreciate you pointing me to the tickets. what this means is that we really can't use os/2 any more and we should look elsewhere. software subscription is active; but, year after year of getting no improvements has made os/2 much more expensive than windows. we engineering types tend to not appreciate the bean counters. yet, they really amazed us when they let us know they took microsoft up on the $30 upgrade deal. can you believe it 2 dozen win 8 pro licenses for $720. we have run XP in virt pc for os/2 on SMP machines for years. we even have win 7 on several using virtualbox. we were the ones who said that those who scoffed at the price and need for virtual pc were blind and had their heads in the sand.

we wanted to contribute to the firefox project; but, the delay in ecs 2.2 caused us to spend the money on software subscription and we are indeed pissed off about it.
Roderick Klein <rwklein@mensys.nl>
   

   

wesco

  • Guest
Re: Motherboard woes
« Reply #21 on: January 01, 2015, 12:10:50 am »
(continuation from before)

5/20/13
      
to me
Hello wesco,


On Sun, 19 May 2013 23:04:44 -0400, wesco wrote:



>we tried the snap with ACPI set to 1 proc. no joy. we have never had issues
>with SNAP; thus our surprise. what we didn't try was the old PSD to see if
>intel MP support works. we have never had smp problems that weren't
>manageable every since the ALR revolution MP 90 we built in 1994. we even
>sent indelible blue a video that they could use to show off OS/2 2.11SMP.

Mensys owns the rights to the OS/2 source code of SNAP and as I have
written we will continue development of SNAP.
Just not right now.

However priority on SNAP development was pushed down because we needed
to make a big step forward.
eCS 2.2 contains the following drivers updated:
IBMKBD.SYS, ACPI.PSD, OS2AHCI.ADD, USBEHCD.SYS, USBOHCD.SYS,
USBUHCD.SYS, USBD.SYS,  USBKBD.SYS, USBMOUSE.SYS, USBHID.SYS.
Patched DOSCALL1.DLL and TCPIP32.DLL to fix an SMP defect IBM never
traced. Mayor updated to bootable JFS code.
Fixes to the danis506.add driver. Mayor overhaul to the Panorama video
code.

As for the SNAP and SMP. The problem is much bigger then just using the
another PSD.
The old OS2APIC.PSD is COMPLETELY different from what ACPI.PSD does.
On most systems OS2APIC.PSD will not load even or it will not run the
hardware in APIC IRQ mode.
(which is required on more and more systems)

The other problem with SNAP is that SOME systems the PCI scan bus
routine does not work.
We have one system where SNAP hangs at startup. (HP Z200) thats one of
the things we need to fix.

Further down in your email you point our the importance of accelrated
video.
Maybe a weird question what define "accelrated" video. With the arrival
of the Panorama driver
a lot of systems had SLOW video. But this is not because Panorama is a
unaccelrated VESA driver.
Its because the MTRR registers on modern mainboards/CPU's are no longer
being setup correctly
on all CPU core's by the BIOS. This is now a task of the OS. The new
ACPI and Panorama driver in eCS 2.2
fix this issue.

SNAP also on some systems if they have more then 2 GB of RAM does not
setup correctly the MTRR
registers correctly. This can result in system hangs or slow video
performance.
The ring 2 video driver on OS/2 is a DLL and runs on an SMP systems on
multiple cores.
Hence the reason the MTRR registers need to be setup correctly.

In eCS 2.2 we ship a new version of Panorama that correctly setups the
MTRR registers on all CPU core's in an SMP
system. It can also handle +2GB memory sceneraio's correctly. And if
MTRR does not work Panorama automaticly switches
to PAT support.

Later this year the changes in Panarama will be backported int SNAP so
it gets the same performance and setup's
the registers on PC's properly running in SMP mode.  But having a video
driver that talks to video card with the control
API of that video chipset is not the holy grail for fast video
performance alone.

>you have 50 units working well. great!! what are they?? just tell us so we
>can just go buy one. how hard is that?

Laptops for example Lenovo T520.T500
Thinkpad W520 - 4276-37U

The fact is most equipment sold in the shop works. As along as you have
a Realtek gigabit or Intel NIC.
Other machine that works is a ltop
MSI CR61 (just bought 2 months ago). USB, network, video and SMP work
on this box and hard disc controller.
You only need to wipe the Windows 8 partition layout.

>about intended customer. obviously there are some specific applications for
>os/2 such as automation, point of sale, etc. this is good business which
>should get priority.

Vertcial applications thats what we sell eCS to for customers needing
to continue to sell/run there OS/2 application(s).

>I however do not see much reason to update firefox,
>bother with flash, or even in a lot of cases the power management stuff in
>these applications. particularly one doesn't want any machine going to
>sleep that is running a nuclear power plant. these things, specially the
>flash, are in my mind end user products. what i am really getting at is
>don't you think it just simply time to call it quits.

I guess you would be completely suprised to hear that eCS 2011 sales
where the best ever :-)
In terms of it being dead ? :)

>all of the people who
>have kept OS/2 alive and useable have done an amazing job and i would be
>completely understanding if they moved on to other things (i have smedley
>in mind).

He is still porting software, he is actively working on it (as we
speak).
http://www.os2world.com/forum/index.php?topic=109.msg1027#msg1027

> ihle's emperoar tv is an awesome program that with win TVR 350s
>have brought us great utility and pleasure. i notice his site hadn't
>updated this the video stuff; but was happy to see other ventures moving
>forward.
>
>as an aside note i was able to get the own cloud iso version to run in a
>virtual machine under OS/2. it was neat to connect my iPad to the warp
>server.
>
>
>for us SMP and accelarated video ARE REQUIRED. i appreciate you pointing me
>to the tickets.

I hear more people state that accelrated video is required and people
then think of SNAP.
The fact is even SNAP when it works on a modern box in SMP mode it not
always going to be fast.
Most of the accelration comes from SMP and proper MTRR register setup.
I'm typing this message on a Thinkpad T61
in 1680x 1050 with Panorama and video is simply snappy.

SMP support with eCS 2.2 works on all boxes I tested it on. Including a
brand new Trend Micro mainboard from 2012.
In some cases you need to switch of hyperthreading (but the performance
benefit of hyperthreading is doubtfull anyway).

>what this means is that we really can't use os/2 any more
>and we should look elsewhere.

Check down below and find the latest link to the eCS 2.2 beta and try
that!
I get the impression you have a completely wrong picture of how the
situation is unfolding.

>software subscription is active; but, year
>after year of getting no improvements has made os/2 much more expensive
>than windows. we engineering types tend to not appreciate the bean
>counters. yet, they really amazed us when they let us know they took
>microsoft up on the $30 upgrade deal. can you believe it 2 dozen win 8 pro
>licenses for $720. we have run XP in virt pc for os/2 on SMP machines for
>years. we even have win 7 on several using virtualbox. we were the ones who
>said that those who scoffed at the price and need for virtual pc were blind
>and had their heads in the sand.
>
>we wanted to contribute to the firefox project; but, the delay in ecs 2.2
>caused us to spend the money on software subscription and we are indeed
>pissed off about it.

Well sorry for the delay with getting eCS 2.2 out the door. But its by
far the biggest overhaul to OS/2 done
by us. Even somebody who used to work in the IBM PC test lab and wrote
drivers for OS/2 at IBM I'm still in touch with (Richard Jerant). He
is amazed as
to how far we pushed the platform with eCS 2.2.

>wesco

Please don't distribute this link. But the very latest beta ISO you can
download from


Let me know how it works for you.
wesco <xxxxxx@gmail.com>
   
5/21/13
      
to Roderick
short feedback for now. mixed results later.

from the point of view of using a pc as a common general purpose computer this beta is very good!! the first beta helped me learn how to install eCS with only a keyboard. never thought i'd need that knowledge. this current one works fine!!! the demo is not bad too!

most of the systems i been working on that are made in the last 3 or 4 years do not have IDE drives. we did have to turn off IDE in the pre boot menu to get our trial system working. maybe it should be off by default.

 i have followed the eCS demo thread at os2world. generally for people who have older systems working with eCS 2.o or 2.1 i think only something like the old warp up should be offered. trying the support the old and move forward is too much effort.  we consider it bad practice to upgrade an OS in place. there is also the adage: if you have it more or less working, don't %#*& with it.

t520  and w520 AWESOME!! so you can tell everybody eCS supports high end lenovo laptops which i believe should be the only ones supported. the consumer oriented l series, which we have sold several of, have been a bit disappointing quality wise. not really worse than dell (we're a reseller); certainly superior to junky HP and compaq.

home users do not have thousands of dollars to pay you to get brand what ever working. they need to know what to buy now that works. 'course my past experience as a white box pc maker was that buyers wanted os/2 systems at bargain windows prices. i wouldn't play that game because i didn't want the support headaches.

what high end desktops work well??

in short. good job. 2.2 is a major milestone. tweak here, tweak there; then may as well ship it.
Roderick Klein <rwklein@mensys.nl>
5/21/13
      
to me
On Tue, 21 May 2013 00:51:29 -0400, wesco wrote:

>short feedback for now. mixed results later.
>
>from the point of view of using a pc as a common general purpose computer
>this beta is very good!! the first beta helped me learn how to install eCS
>with only a keyboard. never thought i'd need that knowledge. this current
>one works fine!!! the demo is not bad too!

USB keyboard and mouse works ?

>most of the systems i been working on that are made in the last 3 or 4
>years do not have IDE drives. we did have to turn off IDE in the pre boot
>menu to get our trial system working. maybe it should be off by default.

What do you mean with IDE ?
You use SCSI in your systems ?

> i have followed the eCS demo thread at os2world. generally for people who
>have older systems working with eCS 2.o or 2.1 i think only something like
>the old warp up should be offered. trying the support the old and move
>forward is too much effort.  we consider it bad practice to upgrade an OS
>in place. there is also the adage: if you have it more or less working,
>don't %#*& with it.

I can certainly tell you that this is exactly the same attitude that
large company customers have that knock on eComStation door.
And that strategy f#@!# up some of the company customers beyond repair.
To many of these large customers do the same
as you describe above. Until the equipment falls apart. And then some
of them spent ages to try and get Warp 3 or 4 to run on new equipment.

What if I tell you we have company customers that still load eCS on NEW
equipment. And thats being done with eCS 2.2 beta's.
Some customers do not only replace old equipment some customers need to
ship new equipment TODAY.
eCS 2.0 and 2.1 are already 2 and 3 years old. The hardware keeps
evolving!
Thats why this renewal cycle to eCS is needed.

I hope you understand SNAP support will get back on track after release
of eCS 2.2, but stuff like ACPI, USB and storage drivers needed
a big step forward. The boot manager had to be put on scrapyard so you
can dual boot Windows Vista/7.
Primary reason to support loading on native hardware is because our
business customers are asking for this.

I don't know if you have seen that with the Panorama driver that the
video is responsive (not that bad).
Does changes will be ported in SNAP so SNAP can work stable. We will
also work on USB 3.0 support.

>t520  and w520 AWESOME!! so you can tell everybody eCS supports high end
>lenovo laptops which i believe should be the only ones supported. the
>consumer oriented l series, which we have sold several of, have been a bit
>disappointing quality wise. not really worse than dell (we're a reseller);
>certainly superior to junky HP and compaq.

You see the point is I just susggested 2 days ago to a European ca
rmanufactor who needs about 300 licenses that they should get a laptop
with a serial port from HP.
They simply expect that the Dell laptop model they buys works!

Please noet that the new laptops Wifi is currently not supported. But
we are working on porting the latest Intel wifi support to eCS.


>home users do not have thousands of dollars to pay you to get brand what
>ever working. they need to know what to buy now that works. 'course my past
>experience as a white box pc maker was that buyers wanted os/2 systems at
>bargain windows prices. i wouldn't play that game because i didn't want the
>support headaches.

Theae people do the preload
work:
http://www.ecomstation.com/where_purchase.phtml?url=nls/en/content/prelo
aders.html&title=Preloaded%20Systems

>what high end desktops work well??
>
>in short. good job. 2.2 is a major milestone. tweak here, tweak there; then
>may as well ship it.

Currently the USB stack is getting its last bugs taken out and the
OS2AHCI driver being worked on.

Sigurd Fastenrath

  • Hero Member
  • *****
  • Posts: 576
  • Karma: +27/-0
  • OS/2 Versus Hardware - Maximum Warp!
    • View Profile
Re: Motherboard woes
« Reply #22 on: January 01, 2015, 02:39:49 am »
i do agree with siguard that for the end user a list of known supported hardware would be nice. appearantly you have 10 machine that work well. what is their make up? are they current (less than 1 year old)? i think stating who your real intended customer are would allow many sitting on the fence technically would be a good move on the part of serenity systems.

5/19/13
      

On Sun, 19 May 2013 18:32:04 -0400, wesco wrote:

>i do agree with
>siguard that for the end user a list of known supported hardware would be
>nice. appearantly you have 10 machine that work well. what is their make
>up? are they current (less than 1 year old)?

Sigurd is approach is nice in theory.

>any word on a new ecs beta. we run high end hardware and think we are
>candidates for beta testing.

5/19/13
      
to Roderick
you have 50 units working well. great!! what are they?? just tell us so we can just go buy one. how hard is that?




Very interesting to read this - I am still - even if it seems outdated - of the opinion to support selected Hardware, especially regarding Laptops/Notebooks. For this one can build perfect Installation clones via DFSEE, I do have some with my Special OS/2 Edition (no parts of ecs except kernel 14.106SMP) for

- ThinkPad X41
- ThinkPad X200t (with Stylus, UMTS, SNAP, WLAN, LAN Support etc.)
- ThinkPad X230t (with Stylus, Touch, LAN Support - WLAN and Sound missing)
- Samsung Serie 7 Slate PC Tablet (with Stylus, Touch, LAN, Sound, USB Support - Sound a bit weak)
- Sony Vaio Pro 13 Core i7 Laptop (with USB, Panorama, QSINIT with 4GB RAM Disk (8GB total RAM)Support - WLAN and Sound missing)
- Maximum Warp PC (AMD 8 Core, Geforce 6600 PCIex16, SNAP (stable up to 4 cores)WLAN, LAN, Suround Sound Support, USB, SSD etc)

-> if I have new Hardware I am using my 2Gb sized Clone of the 8GB Partition of the Maximum Warp PC, changed there to Panorama Vesa Driver, the latest Arca Noae Drivers, select "no Network Card" and disable UNIAUD - this clone is booting ón almost every PC I tried with.

I am looking Forward to see the new UNIAUD and possible WLAN Intel Drviers from Arca Noae coming - this might solve the two most annoying Problems I do have on some of the Hardware mentioned above. Integrating These in my clone would give this a good Chance to work on my next Hardware platform target :-)


« Last Edit: January 01, 2015, 02:42:11 am by Sigurd Fastenrath »

David McKenna

  • Hero Member
  • *****
  • Posts: 742
  • Karma: +24/-0
    • View Profile
Re: Motherboard woes
« Reply #23 on: January 01, 2015, 06:27:25 pm »
  I also have a motherboard (MSI H55M-E33) with Intel i3, H55 chipset and 4GB RAM that otherwise works well with eCS, but I cannot get SNAP to work. I have tried X850 and X600 PCIe cards and the built in Intel video, all of which work fine with Panorama.

@OS4User:

 I installed the SMP debug kernel (supplied with eCS 2.2) to try to get a comport log for you, but I could not boot with Panorama! It would hang at the exact same spot SNAP does, right when PMSHELL is loaded. Booted to command line fine. I have this log if you want to see it. It ends with hlt in 'vman'. Not sure if it is worth trying SNAP now...
« Last Edit: January 01, 2015, 06:31:58 pm by David McKenna »

Doug Bissett

  • Hero Member
  • *****
  • Posts: 1593
  • Karma: +4/-2
    • View Profile
Re: Motherboard woes
« Reply #24 on: January 01, 2015, 07:51:55 pm »
I am going to randomly pick comments, and make a reply:

Quote
maybe some useful insight in this exchange as to where eCS priorities lie. it is not an attack on mr. klein. he was doing the best he could.

Some more information from Mensys is absolutely essential (should have happened months ago). Apparently, Mr. Klein is no longer associated with Mensys, so he is irrelevant. I don't know what has been going on, so I can't say for sure, but I suspect that Mensys has been unraveling some of the problems introduced by Mr. Klein over the years.

Quote
I keep hearing this more often that OS/2 can not share PCI IRQ's. But
let me just put simple, its a fairytale.

I fully agree with that statement. At one time (many years ago now), there were a few bad drivers that had problems, but when they are done properly, there is no problem.

Quote
The only thing I can make up from your email is that you mention
something about IRQ sharing problems. But not what type of problems you
experience with the demo CD.

I will note, that the demo CD is not really a good test. I have more than one machine that has problems with the demo CD (it can usually be made to work, if you spend enough time messing around with it), but they run eCS well, after a proper install. The problems seem to be in the default settings.

Quote
Later this year we hope to get USB
3.0 and SNAP updates done.
Until that time on SMP systems and SNAP should be avoided.

USB 3.0 is becoming more important. SNAP can be used on SOME SMP machines. I have not had a problem with it on my Asus M3A78-EM motherboard, with quad core AMD Phenom processor, and 5 GB of memory. I have been using Panorama recently, just to see if it really makes any difference. I can say that the only difference, that I see, is that Panorama will not allow Doodle's Screen Saver to turn off my monitor, while SNAP will. I also use all 5 GB of memory, by using QSINIT, and the RAMDISK that uses memory above the 4 GB mark. QSINIT also sets the MTRRs correctly, so it won't matter if SNAP doesn't do it (as long as SNAP doesn't mess with it).

Quote
you have 50 units working well. great!! what are they?? just tell us so we can just go buy one. how hard is that?

I don't know what they are using, but I suspect that every one of them is "out of production", so you can't just go out and buy one. What you may be able to do, is pick up some of them on clearance sales, but more likely, they just won't be available until the current users decide to upgrade, and sell them off. The only thing that you can do, is take a chance, and buy a current machine, then see if it can be made to work. So far, every one that I have tried, works eventually (when drivers are available). Sometimes, that takes a few months, and much messing around. The latest available eCS 2.2 b2 seems to have cured a number of problems that earlier versions of eCS 2.x had. QSINIT has fixed the problem where eCS only "sees" 512 meg out of 4 GB. The RAMDISK is a super bonus.

Quote
Maybe a weird question what define "accelrated" video. With the arrival
of the Panorama driver
a lot of systems had SLOW video. But this is not because Panorama is a
unaccelrated VESA driver.
Its because the MTRR registers on modern mainboards/CPU's are no longer
being setup correctly
on all CPU core's by the BIOS. This is now a task of the OS. The new
ACPI and Panorama driver in eCS 2.2
fix this issue.

Acceleration would seem to be somewhat important. I have a Lenovo ThinkPad L530 (UEFI BIOS). I use Panorama because SNAP won't allow turn off of the screen (Panorama doesn't either), and Panorama is marginally quicker than SNAP (VESA 2.0 mode). On the other hand, I also have an ancient IBM ThinkPad A22e (P4 900 mhz, with 256 meg), where SNAP runs in accelerated mode. That video is noticeably FASTER than the L530 (try opening a large file in the system editor, then click the mouse at the bottom to get it scrolling). I will admit that Panorama does a pretty good job, on modern, FAST hardware, which makes up for the slow driver, but acceleration would be so much better.

Quote
The fact is most equipment sold in the shop works. As along as you have
a Realtek gigabit or Intel NIC.

I will comment on this. I have found that Intel devices work pretty well. Realtek, on the other hand, are universally acknowledged to be CRAP. I use a tool netio126.zip from HOBBES. It shows that TCP works well (probably better than in windows), but UDP doesn't work, at all, in some cases. I also test to the internet, using testmy.net. That shows that both Intel, and Realtek, NICs, using the latest Multimac drivers, are much slower than in windows. I also see that Realtek is only about 70% as fast as Intel (in windows).

Quote
You only need to wipe the Windows 8 partition layout.

For many people this is a deal breaker. For those who understand what is going on, it can be done, but most users have no idea, and they certainly do not want to lose software that they paid for (whether they use it, or not). I have Win 7 Pro 64 on the L530. I used windows to shrink the C: drive (leaving all else alone). Then, I tried eCS 2.2 (probably b1 at the time). LVM complained loudly about bad partition tables, and refused to do anything. I used DFSEE to build the partitions, and basically told LVM to get off my back. LVM still complains about partition tables, but I can do everything that is required with DFSEE. Now, I have eCS 2.1, and 2.2 b2 installed (with updates), and they work fine. LVM needs to be fixed, somehow, but an option is to replace it with DFSEE.

Re: Paul Smedley:
Quote
He is still porting software, he is actively working on it (as we speak).

True, but I think he is running into problems because he is not using a real SMP machine. He is using a VBox system, and VBox seems to have a problem if you set it to run eCS with more than one processor (it eventually just stops). That means that his work is not properly tested in an SMP environment. This is a case where users can help, even if they don't know much about software.

Quote
I get the impression you have a completely wrong picture of how the situation is unfolding.

You can't get a "wrong picture" of something when there is "no picture". It seems that nobody is saying much.

Quote
>most of the systems i been working on that are made in the last 3 or 4
>years do not have IDE drives. we did have to turn off IDE in the pre boot
>menu to get our trial system working. maybe it should be off by default.

What do you mean with IDE ?
You use SCSI in your systems ?

You need to realize that SCSI is so far out of the loop, that it is ridiculous to even consider using it (not to mention that ACPI has problems with SCSI drivers). IDE no longer exists, except as an emulation for SATA. SATA is really the only disk interface, in common use, today. I have had problems with SATA (the AHCI driver) if I leave the Danis driver enabled while installing eCS (even with the latest versions), but that seems to depend on the exact implementation in the machine BIOS. I agree that the Danis driver should not be enabled by default, although a check for AHCI support should be able to make that call, and enable the proper driver.

Quote
we consider it bad practice to upgrade an OS in place.

I will agree to that, up to a point. The eCS migration feature has never worked properly, and probably never will. What does work well, is to have two boot volumes. Install (for example) eCS 2.0 on one of them, get it working, and use it. Then, install eCS 2.1 on the other volume. Get that working, at your leisure, then start using that one. The eCS 2.0 volume can be used as an alternate boot volume, when required. Okay, now eCS 2.2 becomes available. Do a clean install on the volume that had eCS 2.0 on it. Get that working at your leisure (it can still be used as an alternate boot volume, if required). Then start using that. Continue alternating the boot volumes for new versions. The main trick to doing that, is to get everything possible onto a common drive, so either boot system has access to it.

Quote
t520  and w520 AWESOME!! so you can tell everybody eCS supports high end >lenovo laptops which i believe should be the only ones supported.

Well, that works until Lenovo changes the rules on you. I have a Lenovo ThinkPad T510, which works well (after a LOT of messing around to get eCS 2.0 installed - about 3 months to sort it out). Having figured that out, and eCS 2.1 apparently fixed a number of the problems, I decided to get a Lenovo ThinkPad L530. ECS 2.1 installed pretty well, but eCS only saw 512 meg of the 4 GB memory. It took a while (6 months?), but eventually QSINIT showed up, and that fixed the problem. QSINIT also supplies a RAMDISK that uses the memory above what eCS can normally access, and now it sets the MTRRs properly, before starting to boot eCS (I don't know if SNAP or Panorama mess with them on boot, or not).

Quote
Currently the USB stack is getting its last bugs taken out and the OS2AHCI driver being worked on.

It is my opinion that the USB stack will never work properly, until it is completely rewritten. The guys who work on it have done wonderful things, but they seem to be constrained by the original design. USB 3.0 may help to get over that, but it should have been finished about 2 years ago. OS2AHCI has worked well, for me, since about the third release, but the Danis driver also works well.

It would also help the users, if regular progress reports were made, so we know what is going on, and have some idea about what to expect when the product is eventually released. Some idea about the time frame would also help. Many users are trying to hang on, wondering if eCS 2.2 will ever see the light of day, and even if it does, will it do the job that needs to be done? Far too many users have been forced to move on, simply because they can't wait forever. Once they do move on, it is highly unlikely that they would ever come back.

Quote
I am still - even if it seems outdated - of the opinion to support selected Hardware, especially regarding laptops/Notebooks.

The main "problem" with that, is that by the time somebody figures out that a device is good, or bad, it is no longer available (except possibly on the clearance shelves). Then, there is the problem with features. One model, with feature X, might work fine. The same model, with feature Y, may not work, at all. There is a good possibility that feature X may be standard in Asia, and feature Y will be standard in the US. The alternate feature may not be available in the respective countries. That leaves us with one option: Buy a whole bunch of something, all the same, then do whatever is needed to make it work. By the time that is done, the machine will be out of date, which means you can't charge as much for them, so it is impossible to do it, unless you want to lose a bunch of money. What would help, would be for the leading edge users, to post accurate reports about what does, or does not, work. Even that is a two edged sword. When I got my Lenovo T510, I had to report that it wasn't very good. 3 months later, after getting it sorted out, it was working rather well (but was also obsolete, and not available any more). At that time, I decided to change the report, and found that all I could do was add a comment, which did not change the status from "bad" to "good". Now, the average user looks at that, and see that the T510 is "bad", when it is really "good". The report is supposed to help, but it doesn't. Personally, I have quit reporting things because it causes more problems than it solves.

Dave Yeo

  • Hero Member
  • *****
  • Posts: 4787
  • Karma: +99/-1
    • View Profile
Re: Motherboard woes
« Reply #25 on: January 01, 2015, 09:38:42 pm »
About IRQ sharing. I've had continuous bad luck when it comes to IRQs causing problems.  Using the latest available from Mensys, my builtin Intel Gigabit network stops working after about a day (receives but doesn't send) and my last box on Warp V4 needed sound and the IDE controller disabled to even work at all, after which doing a warm boot to eCS allowed the Ethernet to work whereas a cold boot to eCS2.1 just didn't work (once again receiving but not sending which I've been told is due to IRQ conflict). On this box it works fine for sharing dial-up but recently I tried to move a GigaByte of info over it and it consistently died before finishing, ended up having to sneaker net.
USB isn't much better. Used a USB modem for quite a while, it had to be plugged into a specific port to work along with the mouse and keyboard being plugged into specific ports. Lots of trial and error to find the working setup. Finally replaced the system with one with a real com port.

Paul Smedley

  • Hero Member
  • *****
  • Posts: 2092
  • Karma: +159/-0
    • View Profile
Re: Motherboard woes
« Reply #26 on: January 01, 2015, 09:43:10 pm »
Re: Paul Smedley:
Quote
He is still porting software, he is actively working on it (as we speak).

True, but I think he is running into problems because he is not using a real SMP machine. He is using a VBox system, and VBox seems to have a problem if you set it to run eCS with more than one processor (it eventually just stops). That means that his work is not properly tested in an SMP environment. This is a case where users can help, even if they don't know much about software.

????? WTF? Where do you get this stuff from? I should just learn to stop reading your comments.....

Doug Bissett

  • Hero Member
  • *****
  • Posts: 1593
  • Karma: +4/-2
    • View Profile
Re: Motherboard woes
« Reply #27 on: January 02, 2015, 12:00:37 am »
Re: Paul Smedley:
Quote
He is still porting software, he is actively working on it (as we speak).

True, but I think he is running into problems because he is not using a real SMP machine. He is using a VBox system, and VBox seems to have a problem if you set it to run eCS with more than one processor (it eventually just stops). That means that his work is not properly tested in an SMP environment. This is a case where users can help, even if they don't know much about software.

????? WTF? Where do you get this stuff from? I should just learn to stop reading your comments.....

Paul, do you actually have more than one CPU working with eCS in VBox? Perhaps it does work in the Linux version. It does not work under windows. After a few minutes, it just stops, and the machine needs to be reset.

FWIW, Alex Taylor is now fighting with a problem where running in a single CPU environment is causing horrendous delays, while running in a multi CPU environment works fine. The other way around can also be true (not necessarily delays, that is just an example). If you don't get your software tested in a SMP environment, you could be missing problems that you would never see otherwise. My suggestion is that other people can do the testing for you.

Perhaps you should start READING my comments, and stop assuming what you first think you read.

Lars

  • Hero Member
  • *****
  • Posts: 1271
  • Karma: +65/-0
    • View Profile
Re: Motherboard woes
« Reply #28 on: January 04, 2015, 01:34:33 pm »

Quote
Currently the USB stack is getting its last bugs taken out and the OS2AHCI driver being worked on.

It is my opinion that the USB stack will never work properly, until it is completely rewritten. The guys who work on it have done wonderful things, but they seem to be constrained by the original design. USB 3.0 may help to get over that, but it should have been finished about 2 years ago.

I wouldn't say that the stack design is the problem. The problem is that the USB Driver code from the DDK that we have is outdated and is also full of bugs where you never know if these bugs are also in the latest drivers from IBM or not.
I think I fixed around 8 to 10 bugs in USBUHCD.SYS+USBOHCD.SYS+USBEHCD.SYS+USBD.SYS (including bugs that Rüdiger Ihle had already found). And I reworked USBEHCD.SYS to make isochronous transfer work (which Wim actually implemented, many thanks to him !) which apparently was never properly implemented and therefore not tested at all.
I am currently doing a major rework on USBMSD.ADD to allow to use USB sticks formatted under Windows and fix problems with existing devices. Again, that driver contained a couple of bugs. And it's the worst/messiest piece of SW I have seen in a tong time.
Nonetheless, rewriting the whole stack will never happen. It's much too much work. Forget it.
 

ff4

  • Guest
Re: Motherboard woes
« Reply #29 on: January 08, 2015, 04:10:00 pm »
Luckily I never hard problems  with usb in the past but when it comes to new hardware it can get interesting. Therefore a big thanks to Lars and other for improving the usb stack over the time.