ArcaOS JFS traps

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.

You should probably file an issue at the ArcaOS bug tracker, assuming
you still have an active subscription and the latest JFS.
Dave

b***@baden.nu

2019-06-03 22:27:14 UTC

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.

You should probably file an issue at the ArcaOS bug tracker, assuming
you still have an active subscription and the latest JFS.
Dave

Hi Dave:

Thanks, ArcaOS does not make it easy to address deficiencies. After purchasing 5.0 & 5.01, and a further subscription, I do not think it should be necessary for me to pay them more money to log issues that caused me considerable inconvenience, time, effort, and money to resolve.

I am hoping that usenet might be monitored by them.

thanks,
Baden

Dave Yeo

2019-06-04 06:37:06 UTC

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.

You should probably file an issue at the ArcaOS bug tracker, assuming
you still have an active subscription and the latest JFS.
Dave

Thanks, ArcaOS does not make it easy to address deficiencies. After purchasing 5.0 & 5.01, and a further subscription, I do not think it should be necessary for me to pay them more money to log issues that caused me considerable inconvenience, time, effort, and money to resolve.
I am hoping that usenet might be monitored by them.

Unluckily, they've stated they don't have time to monitor various places
such as usenet, though you might get lucky.
You don't have the cache maxed out do you?
Dave

b***@gmail.com

2019-06-11 02:55:06 UTC

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.

You should probably file an issue at the ArcaOS bug tracker, assuming
you still have an active subscription and the latest JFS.
Dave

Unluckily, they've stated they don't have time to monitor various places
such as usenet, though you might get lucky.
You don't have the cache maxed out do you?
Dave

Interestingly, the other day, I was downloading files off my camera's CF Card onto my SSD, and ... TRAP!

I thought I had it figured out by backleveling the "jfs" files. A last trap was in 'OS2KRNL', so I also backleveled all the \os2* files, but the result was the same.

I replaced the ArcaOS with the old (eCS) OS, and all the files transferred without event.

I am bewildered by this situation, and wondering what device files might precipitate a trap?

IFS=C:\OS2\JFS.IFS /LW:5,20,4 /AUTOCHECK:*
BASEDEV=OS2AHCI.ADD
BASEDEV=DANIS506.ADD /!BIOS
BASEDEV=OS2DASD.DMD
BASEDEV=OS2LVM.DMD

thanks,
Baden

Dave Yeo

2019-06-12 05:39:07 UTC

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.

You should probably file an issue at the ArcaOS bug tracker, assuming
you still have an active subscription and the latest JFS.
Dave

Unluckily, they've stated they don't have time to monitor various places
such as usenet, though you might get lucky.
You don't have the cache maxed out do you?
Dave

Interestingly, the other day, I was downloading files off my camera's CF Card onto my SSD, and ... TRAP!
I thought I had it figured out by backleveling the "jfs" files. A last trap was in 'OS2KRNL', so I also backleveled all the \os2* files, but the result was the same.
I replaced the ArcaOS with the old (eCS) OS, and all the files transferred without event.
I am bewildered by this situation, and wondering what device files might precipitate a trap?
IFS=C:\OS2\JFS.IFS /LW:5,20,4 /AUTOCHECK:*

I'd guess JFS, including the possibility of the cache.

Post by b***@gmail.com
BASEDEV=OS2AHCI.ADD
BASEDEV=DANIS506.ADD /!BIOS
BASEDEV=OS2DASD.DMD
BASEDEV=OS2LVM.DMD
thanks,
Baden

I wouldn't advise backleveling the OS2* files, the kernel has some fixes
compared to older versions and the os2ldr is quite different then the
IBM one.
As usual, you might want to open an issue at Arca Noae.
Dave

baden@baden.nu

2019-06-13 04:04:05 UTC

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.

You should probably file an issue at the ArcaOS bug tracker, assuming
you still have an active subscription and the latest JFS.
Dave

Unluckily, they've stated they don't have time to monitor various places
such as usenet, though you might get lucky.
You don't have the cache maxed out do you?
Dave

Interestingly, the other day, I was downloading files off my camera's CF Card onto my SSD, and ... TRAP!
I thought I had it figured out by backleveling the "jfs" files. A last trap was in 'OS2KRNL', so I also backleveled all the \os2* files, but the result was the same.
I replaced the ArcaOS with the old (eCS) OS, and all the files transferred without event.
I am bewildered by this situation, and wondering what device files might precipitate a trap?
IFS=C:\OS2\JFS.IFS /LW:5,20,4 /AUTOCHECK:*

I'd guess JFS, including the possibility of the cache.

Post by b***@gmail.com
BASEDEV=OS2AHCI.ADD
BASEDEV=DANIS506.ADD /!BIOS
BASEDEV=OS2DASD.DMD
BASEDEV=OS2LVM.DMD
thanks,
Baden

I wouldn't advise backleveling the OS2* files, the kernel has some fixes
compared to older versions and the os2ldr is quite different then the
IBM one.

Hi Dave:

I have found that as long as I copy over all the OS2* files in the root, the compatibility and functionality are preserved.

Since ArcaOS traps, and I back levelled all the *jfs* files and the os2* files, and I still received a trap, I surmise the problem might be in danis506.add. I see that the newest version was not hers. I back leved danis506.add, but the system would not boot at all. So, the system is now all ArcaOS with ibm1s506.add replacing danis506.add. I will see how this goes.

This issue brings forth possible problems with SMP. There have been critical corrections to ACPI.PSD since Arca 5.0. I also found my repository with newer ArcaOS Warpin fixes:

1,565,244 acpi-3.23.10.exe
98,876 ahci-2.04.exe
136,764 DaniS506-1.08.15.exe
1,047,100 JFS-1.09.07.exe
992,316 Multimac-20171019.exe
4,772,924 panorama-1.11.exe
1,562,172 Uniaud-20170625.exe
327,740 USB-11.17.exe

So, I think I will first see how IBM1S506.ADD behaves. If it fails, or not, I have the option to install the latest fixes. INterstingly, my 'old' eCS based system had most these fixes installed, which may be significant.

Post by Dave Yeo
As usual, you might want to open an issue at Arca Noae.
Dave

As far as (paying for and) opening a ticket goes, that is a sore spot with me, as I believe it exposes the current principals' principles. I do not buy the BS that monitoring the fora (~3 messages a month) is unmanageable. Many in IBM Boca spent considerable time responding to valid queries. Doug Azzarito was surprisingly responsive.

thanks,
Baden

Dave Yeo

2019-06-13 23:57:41 UTC

Post by ***@baden.nu
I have found that as long as I copy over all the OS2* files in the root, the compatibility and functionality are preserved.

As long as you're not loading DLLs high or using the ramdisk, it should
be fine. I take it you're using the ones from eCS.

Post by ***@baden.nu
Since ArcaOS traps, and I back levelled all the *jfs* files and the os2* files, and I still received a trap, I surmise the problem might be in danis506.add. I see that the newest version was not hers. I back leved danis506.add, but the system would not boot at all. So, the system is now all ArcaOS with ibm1s506.add replacing danis506.add. I will see how this goes.

Did you ever test with only one CPU? Note that there is a newer acpi.

Post by ***@baden.nu
1,565,244 acpi-3.23.10.exe
98,876 ahci-2.04.exe
136,764 DaniS506-1.08.15.exe
1,047,100 JFS-1.09.07.exe
992,316 Multimac-20171019.exe
4,772,924 panorama-1.11.exe
1,562,172 Uniaud-20170625.exe
327,740 USB-11.17.exe
So, I think I will first see how IBM1S506.ADD behaves. If it fails, or not, I have the option to install the latest fixes. INterstingly, my 'old' eCS based system had most these fixes installed, which may be significant.

Post by Dave Yeo
As usual, you might want to open an issue at Arca Noae.
Dave

Well I'm here, though I'm not paid by Arca Noae, I am listed as a
developer. Thing is is that David is basically the only one that knows
the internals of this stuff and I've never seen him active anywhere
other then on the Arca Noae lists and previously the eCS ones IIRC.
With a ticket, I could ask for a testlog and find out more about your
system rather then asking here. Which raises the question of what is
your system?
Dave

baden@baden.nu

2019-06-15 22:42:53 UTC

Post by Dave Yeo
Did you ever test with only one CPU? Note that there is a newer acpi.

That is certainly an option. My conjecture is that
there is a spinlock (SMP) error in the revised
danis506.add. So far, using IBM1s506, I have had no
problems. If I do get a trap, then I can further
investigate selecting only one CPU, and also applying
the noted fixes. I do not see much use in
troubleshooting this issue beyond getting a stable
system. As stated, this problem only originated with
ArcaOS.

A problem with no communication, is that everyone is
in the dark. This may be a known issue with known
fixes. Or the developer may be unknowingly and
continuously releasing faulty software.

Post by Dave Yeo
As usual, you might want to open an issue at Arca Noae.
Dave

Well I'm here, though I'm not paid by Arca Noae, I am listed as a

That is very much appreciated, Dave.

Post by Dave Yeo
developer. Thing is is that David is basically the only one that knows
the internals of this stuff and I've never seen him active anywhere
other then on the Arca Noae lists and previously the eCS ones IIRC.
With a ticket, I could ask for a testlog and find out more about your
system rather then asking here. Which raises the question of what is
your system?
Dave

I have a T60 with 2 GB RAM, and now a Samsung PRO
512 GB SSD.

thanks for your interest!
Baden

Lars Erdmann

2019-06-18 04:25:27 UTC

Post by Dave Yeo
Did you ever test with only one CPU? Note that there is a newer acpi.

If it were a spinlock error it would more likely hang instead of trap.

baden@baden.nu

2019-06-18 23:12:18 UTC

Post by Lars Erdmann

Post by Dave Yeo
Did you ever test with only one CPU? Note that there is a newer acpi.

If it were a spinlock error it would more likely hang instead of trap.

Hi Lars:

Interestingly, this is what it does. For about 30 s,
activity stops and hangs, and then it traps. e, then 8.

thanks,
Baden

Lars Erdmann

2019-06-18 04:32:13 UTC

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.

You should probably file an issue at the ArcaOS bug tracker, assuming
you still have an active subscription and the latest JFS.
Dave

Thanks, ArcaOS does not make it easy to address deficiencies. After
purchasing 5.0 & 5.01, and a further subscription, I do not think it
should be necessary for me to pay them more money to log issues that
caused me considerable inconvenience, time, effort, and money to
resolve.
I am hoping that usenet might be monitored by them.

Unluckily, they've stated they don't have time to monitor various places
such as usenet, though you might get lucky.
You don't have the cache maxed out do you?
Dave

Interestingly, the other day, I was downloading files off my camera's
CF Card onto my SSD, and ... TRAP!

So how did you access the CF Card ? Via a USB card reader ?
If yes, it's also possible that USBMSD.ADD fails on you.
Given its overall quality, I would not be surprised.
You can give it a try and use the one from my package after you have
backed up the existing one (you can keep the rest of the AN USB stack):
http://hobbes.nmsu.edu/download/pub/os2/system/drivers/misc/usbdrv227.zip

Make sure you reinstate everything else.

baden@baden.nu

2019-06-18 23:30:58 UTC

Post by Lars Erdmann

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.

Interestingly, the other day, I was downloading files off my camera's
CF Card onto my SSD, and ... TRAP!

So how did you access the CF Card ? Via a USB card reader ?

No, through the PCMCIA slot and drivers.

BASEDEV=PCMCIA.SYS
BASEDEV=SS2TICB.SYS
BASEDEV=AUTODRV2.SYS
DEVICE=C:\OS2\BOOT\PCMSSDIF.SYS
BASEDEV=PCM2ATA.ADD /S:1 /!DM
BASEDEV=OS2PCARD.DMD

After it trapped using the PCMCIA card, I also
'rem'ed those out, tried using SSD to SSD xfer, and it
also trapped with that.

thanks,
Baden

Post by Lars Erdmann
If yes, it's also possible that USBMSD.ADD fails on you.
Given its overall quality, I would not be surprised.
You can give it a try and use the one from my package after you have
http://hobbes.nmsu.edu/download/pub/os2/system/drivers/misc/usbdrv227.zip
Make sure you reinstate everything else.

b***@gmail.com

2021-03-25 05:07:25 UTC

Post by Lars Erdmann

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.

Interestingly, the other day, I was downloading files off my camera's
CF Card onto my SSD, and ... TRAP!

So how did you access the CF Card ? Via a USB card reader ?

No, through the PCMCIA slot and drivers.
BASEDEV=PCMCIA.SYS
BASEDEV=SS2TICB.SYS
BASEDEV=AUTODRV2.SYS
DEVICE=C:\OS2\BOOT\PCMSSDIF.SYS
BASEDEV=PCM2ATA.ADD /S:1 /!DM
BASEDEV=OS2PCARD.DMD
After it trapped using the PCMCIA card, I also
'rem'ed those out, tried using SSD to SSD xfer, and it
also trapped with that.
thanks,
Baden

b***@gmail.com

2021-03-25 05:30:34 UTC

Okay, I *finally* found what caused the traps. I was just reading on os2world how OS2AHCI.ADD is a replacement for the *506* drivers, which I never knew. I looked in my CONFIG.SYS, which had:

BASEDEV=OS2AHCI.ADD
BASEDEV=DANIS506.ADD /!BIOS
BASEDEV=DANIATAP.FLT

I subsequently removed the AHCI driver, and all worked well. I then reinstated it, and removed the DANI drivers, and all worked well. So, with a large (JFS) partition, and both drivers active, the system can trap.

Can someone let Arca know about this?

thanks, Baden

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.
=============
I have a T60, which had a 512 GB Crucial SSD (2014),
that bricked in December after some partitioning
efforts. I replaced that with a 512 GB spinning
Toshiba and (almost) identical partition setup. I
recently received a new Crucial 512 GB SSD, and when I
was transferring files, I had some write problems, and
then it continually trapped ('e' & '8' mostly in
OS2KRNL) while I was attempting to restore a 450 GB
partition. After some more errors, I determined the
SSD had write errors, and I returned and replaced it
with a Samsung 860 Pro. I partitioned that the same,
but surprisingly, it also trapped while restoring files
on the large partition.
The 454 GB partition kept trapping at 89 GB written.
I changed the partition size to different values, and
smaller partitions would allow more writing. Finally,
at a 300 GB size, I could fill the partition without
trapping. I thought it was a SSD unique problem, until
afterwards I checked the Toshiba, which I had earlier
only filled with about 50 GB onto the large partition.
That drive also trapped at about 100 GB filled.
My system is partitioned with a HPFS C: franken
OS/2 installation, which has much eCS. It ran for ten
years without issues, originally on 512 GB spinning
rust, and the last 4.5 years on a 512 GB SSD. D: Drive
is a minimal HPFS maintenance partition, and E: Drive
is JFS ArcaOS 5.0 (was eCS).
┌──────────────────────────────────────────────────────────────────────────────┐
│ Logical Volume Type Status File System Size (MB)│
├──────────────────────────────────────────────────────────────────────────────┤
│OS2 C: Compatibility Startable HPFS 502 │
│Maint D: Compatibility HPFS 15 │
│ArcaOS E: Compatibility JFS 1027 │
│Data F: LVM JFS 32002 │
│Fotos G: LVM JFS 454830 │
└──────────────────────────────────────────────────────────────────────────────┘
=============
This trap problem only started in May after I tried
restoring a large partition using ArcaOS 5.0 on SSDs.
Both the old OS (eCS) and ArcaOS trapped, and the old
OS2 mostly predates ArcaOS. I first contemplated and
checked whether heat, 'lvm' or 'format were
responsible. I further investigated, and found that
both OS use the newer JFS files, as I had installed the
"JFS - Journaled File System for OS/2 v1.09.07"
package.
44251 0 0 2016-10-24 09:48 OS2/CACHEJFS.EXE
280192 0 0 2017-01-06 17:47 OS2/DLL/UJFS.DLL
193319 0 0 2016-10-24 09:48 OS2/JFS.IFS
6495 0 0 2016-10-24 09:48 OS2/JFSCHK32.EXE
44949 0 0 2016-10-24 09:48 OS2/jfsstats.exe
I then checked an older backup, and found that the
44268 0 0 2010-04-27 12:57 OS2/CACHEJFS.EXE
99061 0 0 2010-04-27 12:57 OS2/CHKLGJFS.EXE
281594 0 0 2010-04-27 12:58 OS2/DLL/UJFS.DLL
193291 0 0 2010-04-27 12:55 OS2/JFS.IFS
6495 0 0 2010-04-27 12:57 OS2/JFSCHK32.EXE
I restored over all the old files, and the system
ran and filled the 450 GB partitions faultlessly on
both the Toshiba and SSD.
=============
It appears that the new ArcaOS JFS faults writing to
large partitions. It may be from a timing issue, as
this problem originally occurred when restoring SSD
drives. It also occurred on the spinning drive, but
the Toshiba did successfully fill 400+ GB one time out
of many traps.
thanks,
Baden

Doug Bissett

2021-03-27 17:17:19 UTC

Post by b***@gmail.com
Okay, I *finally* found what caused the traps. I was just reading on os2world how OS2AHCI.ADD is a replacement for the *506* drivers, which I never knew.

Not exactly. OS2AHCI.ADD is support for the AHCI feature of SATA. The
Dani drivers are really for IDE drives, but SATA has a feature to
operate in IDE mode. You really do need to READ the documentation.

Post by b***@gmail.com
BASEDEV=OS2AHCI.ADD
BASEDEV=DANIS506.ADD /!BIOS
BASEDEV=DANIATAP.FLT
I subsequently removed the AHCI driver, and all worked well. I then reinstated it, and removed the DANI drivers, and all worked well. So, with a large (JFS) partition, and both drivers active, the system can trap.
Can someone let Arca know about this?

Why? You are obviously using some ancient version of OS2AHCI, or you
have an invalid configuration. The first driver to load, runs the
device. Any other driver that loads, cannot operate the device, and
doesn't even try (unless you also have some ancient version of the
Dani driver).

IF you can make the latest OS2AHCI driver fail, then YOU need to
report it, along with all of the required support documentation. Look
in the Wiki for instructions on how to properly report a problem.
Nobody else can do it, because we don't have your machine to be able
to produce the supporting documentation.

If you don't have a current subscription, you don't have support.
Bitching about that here isn't going to do anybody any good.

Post by b***@gmail.com
thanks, Baden

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.
=============
I have a T60, which had a 512 GB Crucial SSD (2014),
that bricked in December after some partitioning
efforts. I replaced that with a 512 GB spinning
Toshiba and (almost) identical partition setup. I
recently received a new Crucial 512 GB SSD, and when I
was transferring files, I had some write problems, and
then it continually trapped ('e' & '8' mostly in
OS2KRNL) while I was attempting to restore a 450 GB
partition. After some more errors, I determined the
SSD had write errors, and I returned and replaced it
with a Samsung 860 Pro. I partitioned that the same,
but surprisingly, it also trapped while restoring files
on the large partition.
The 454 GB partition kept trapping at 89 GB written.
I changed the partition size to different values, and
smaller partitions would allow more writing. Finally,
at a 300 GB size, I could fill the partition without
trapping. I thought it was a SSD unique problem, until
afterwards I checked the Toshiba, which I had earlier
only filled with about 50 GB onto the large partition.
That drive also trapped at about 100 GB filled.
My system is partitioned with a HPFS C: franken
OS/2 installation, which has much eCS. It ran for ten
years without issues, originally on 512 GB spinning
rust, and the last 4.5 years on a 512 GB SSD. D: Drive
is a minimal HPFS maintenance partition, and E: Drive
is JFS ArcaOS 5.0 (was eCS).
��Ŀ
� Logical Volume Type Status File System Size (MB)�
��Ĵ
�OS2 C: Compatibility Startable HPFS 502 �
�Maint D: Compatibility HPFS 15 �
�ArcaOS E: Compatibility JFS 1027 �
�Data F: LVM JFS 32002 �
�Fotos G: LVM JFS 454830 �
��
=============
This trap problem only started in May after I tried
restoring a large partition using ArcaOS 5.0 on SSDs.
Both the old OS (eCS) and ArcaOS trapped, and the old
OS2 mostly predates ArcaOS. I first contemplated and
checked whether heat, 'lvm' or 'format were
responsible. I further investigated, and found that
both OS use the newer JFS files, as I had installed the
"JFS - Journaled File System for OS/2 v1.09.07"
package.
44251 0 0 2016-10-24 09:48 OS2/CACHEJFS.EXE
280192 0 0 2017-01-06 17:47 OS2/DLL/UJFS.DLL
193319 0 0 2016-10-24 09:48 OS2/JFS.IFS
6495 0 0 2016-10-24 09:48 OS2/JFSCHK32.EXE
44949 0 0 2016-10-24 09:48 OS2/jfsstats.exe
I then checked an older backup, and found that the
44268 0 0 2010-04-27 12:57 OS2/CACHEJFS.EXE
99061 0 0 2010-04-27 12:57 OS2/CHKLGJFS.EXE
281594 0 0 2010-04-27 12:58 OS2/DLL/UJFS.DLL
193291 0 0 2010-04-27 12:55 OS2/JFS.IFS
6495 0 0 2010-04-27 12:57 OS2/JFSCHK32.EXE
I restored over all the old files, and the system
ran and filled the 450 GB partitions faultlessly on
both the Toshiba and SSD.
=============
It appears that the new ArcaOS JFS faults writing to
large partitions. It may be from a timing issue, as
this problem originally occurred when restoring SSD
drives. It also occurred on the spinning drive, but
the Toshiba did successfully fill 400+ GB one time out
of many traps.
thanks,
Baden

--
From Doug Bissett's ArcaOS
dougb007 at telus dot net
(Please make the obvious changes, to e-mail me)

b***@gmail.com

2021-03-28 05:45:33 UTC

Hi Doug:

Thank-you very much for your prompt and informative response. I appreciate that someone with strong connections and unfaltering support for ArcaOS still monitors the usenet. Evidently, your technical knowledge, diplomacy, and business acumen might only be surpassed by Lewis Rosenthal in your endless quest for excellence.

Okay, I *finally* found what caused the traps. I was just reading on os2world how OS2AHCI.ADD is a replacement for the *506* drivers, which I never knew.

If this is the case, how do both drivers function independently and identically? Would not there be a huge performance difference?

BASEDEV=OS2AHCI.ADD
BASEDEV=DANIS506.ADD /!BIOS
BASEDEV=DANIATAP.FLT
I subsequently removed the AHCI driver, and all worked well. I then reinstated it, and removed the DANI drivers, and all worked well. So, with a large (JFS) partition, and both drivers active, the system can trap.
Can someone let Arca know about this?

I wish I knew about this more than I do, but the best obvious information I can provide is:

BldLevel: @#Arca Noae LLC:2.04#@##1## 12 Dec 2017 17:09:14 DAZAR1 ::::::@@AHCI Driver (c) 2017 Arca Noae LLC
CmdLine: OS2AHCI.ADD /N
port #0 interrupt error status: 0x40000001; restarting port
port #0 interrupt error status: 0x40000001; restarting port
Adapter 0: PCI=0:31:2 ID=8086:27c5 Intel ICH7M irq=16 addr=0xee444400 version=10100
Port 0:
Drive 0: 62260 cylinders, 255 heads, 63 sectors per track (488382MB) (LVM)
Model: Samsung SSD 860 PRO 512GB

R1.08.15 /!BIOS
Daniela's Bus Master IDE Driver for OS/2 Version 1.08.15
Controller:0 Port:01F0 IRQ:0E Status:OK BusMaster Scatter/Gather
Intel ICH7 PATA host (8086:27DF rev:02) on PCI 0:31.1#0
Unit:0 Status:OK ATAPI BusMaster UltraDMA2/PIO4
Model:DVD/CDRW UJDA775 CB03
Controller:1 Port:18D0 IRQ:10 Status:OK BusMaster Scatter/Gather
Intel ICH7 SATA host (8086:27C5 rev:02) on PCI 0:31.2#0
Unit:0 Status:OK LBA BusMaster 1.5GBit/s BPB
Model:Samsung SSD 860 PRO 512GB RVM01B6Q
OS2:log phys BPB/BIOS IDE:log phys Total Sectors
C 62260 65535 65535 16383 Avail 1000215216
H 255 16 255 16 16 OS2 1000206900
S 63 63 63 63 63 % Used 99.99

Post by Doug Bissett
IF you can make the latest OS2AHCI driver fail, then YOU need to
report it, along with all of the required support documentation. Look
in the Wiki for instructions on how to properly report a problem.
Nobody else can do it, because we don't have your machine to be able
to produce the supporting documentation.

I am amazed that a purported OS vendor does not have test machines, especially a T60 or equivalent, which is one of the few laptops that might operate ArcaOS.

Post by Doug Bissett
If you don't have a current subscription, you don't have support.
Bitching about that here isn't going to do anybody any good.

Arca should be enthusiastically receptive to receiving critical bug reports. We (customers) are doing them (vendor) a huge favour. How many customers just dropped the platform altogether after going through a fraction of the grief, time, and expense I did? Having closed subscription based fora (ostrich emulation mode) just exacerbates their hardware compatibility problems.

As soon as ArcaOS has a viable browser and supports my external 1920 monitor, I will be the first to throw some more money their way.

thanks,
Baden

Post by b***@baden.nu
It appears that the ArcaOS JFS driver is trapping on
my system.
=============
I have a T60, which had a 512 GB Crucial SSD (2014),
that bricked in December after some partitioning
efforts. I replaced that with a 512 GB spinning
Toshiba and (almost) identical partition setup. I
recently received a new Crucial 512 GB SSD, and when I
was transferring files, I had some write problems, and
then it continually trapped ('e' & '8' mostly in
OS2KRNL) while I was attempting to restore a 450 GB
partition. After some more errors, I determined the
SSD had write errors, and I returned and replaced it
with a Samsung 860 Pro. I partitioned that the same,
but surprisingly, it also trapped while restoring files
on the large partition.
The 454 GB partition kept trapping at 89 GB written.
I changed the partition size to different values, and
smaller partitions would allow more writing. Finally,
at a 300 GB size, I could fill the partition without
trapping. I thought it was a SSD unique problem, until
afterwards I checked the Toshiba, which I had earlier
only filled with about 50 GB onto the large partition.
That drive also trapped at about 100 GB filled.
My system is partitioned with a HPFS C: franken
OS/2 installation, which has much eCS. It ran for ten
years without issues, originally on 512 GB spinning
rust, and the last 4.5 years on a 512 GB SSD. D: Drive
is a minimal HPFS maintenance partition, and E: Drive
is JFS ArcaOS 5.0 (was eCS).
ЪДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДї
і Logical Volume Type Status File System Size (MB)і
ГДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДґ
іOS2 C: Compatibility Startable HPFS 502 і
іMaint D: Compatibility HPFS 15 і
іArcaOS E: Compatibility JFS 1027 і
іData F: LVM JFS 32002 і
іFotos G: LVM JFS 454830 і
АДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДДЩ
=============
This trap problem only started in May after I tried
restoring a large partition using ArcaOS 5.0 on SSDs.
Both the old OS (eCS) and ArcaOS trapped, and the old
OS2 mostly predates ArcaOS. I first contemplated and
checked whether heat, 'lvm' or 'format were
responsible. I further investigated, and found that
both OS use the newer JFS files, as I had installed the
"JFS - Journaled File System for OS/2 v1.09.07"
package.
44251 0 0 2016-10-24 09:48 OS2/CACHEJFS.EXE
280192 0 0 2017-01-06 17:47 OS2/DLL/UJFS.DLL
193319 0 0 2016-10-24 09:48 OS2/JFS.IFS
6495 0 0 2016-10-24 09:48 OS2/JFSCHK32.EXE
44949 0 0 2016-10-24 09:48 OS2/jfsstats.exe
I then checked an older backup, and found that the
44268 0 0 2010-04-27 12:57 OS2/CACHEJFS.EXE
99061 0 0 2010-04-27 12:57 OS2/CHKLGJFS.EXE
281594 0 0 2010-04-27 12:58 OS2/DLL/UJFS.DLL
193291 0 0 2010-04-27 12:55 OS2/JFS.IFS
6495 0 0 2010-04-27 12:57 OS2/JFSCHK32.EXE
I restored over all the old files, and the system
ran and filled the 450 GB partitions faultlessly on
both the Toshiba and SSD.
=============
It appears that the new ArcaOS JFS faults writing to
large partitions. It may be from a timing issue, as
this problem originally occurred when restoring SSD
drives. It also occurred on the spinning drive, but
the Toshiba did successfully fill 400+ GB one time out
of many traps.
thanks,
Baden

--
From Doug Bissett's ArcaOS
dougb007 at telus dot net
(Please make the obvious changes, to e-mail me)

Dave Yeo

2021-03-28 07:48:04 UTC

Post by b***@gmail.com
Thank-you very much for your prompt and informative response. I appreciate that someone with strong connections and unfaltering support for ArcaOS still monitors the usenet. Evidently, your technical knowledge, diplomacy, and business acumen might only be surpassed by Lewis Rosenthal in your endless quest for excellence.

There's a couple of us volunteers who monitor the newsgroups. Doug
basically said what I meant to. Note that neither Doug or I can speak
for Arca Noae, just that we have _slightly_ more access to how things go
with Arca Noae.

Okay, I *finally* found what caused the traps. I was just reading on os2world how OS2AHCI.ADD is a replacement for the *506* drivers, which I never knew.

If this is the case, how do both drivers function independently and identically? Would not there be a huge performance difference?

Think of systems that used to be common, a AHCI hard drive plugged into
a SATA port and a IDE DVD drive plugged into a IDE port. My current
computer is the first I've had with no IDE port.
Ideally the AHCI driver loads first and claims the hard drive and then
the S506 driver loads and claims the left overs, namely anything plugged
into the IDE interface, usually a DVD writer, but there has been times I
plugged in an old HD.
They should co-operate.

CmdLine: OS2AHCI.ADD /N
port #0 interrupt error status: 0x40000001; restarting port
port #0 interrupt error status: 0x40000001; restarting port
Adapter 0: PCI=0:31:2 ID=8086:27c5 Intel ICH7M irq=16 addr=0xee444400 version=10100
Drive 0: 62260 cylinders, 255 heads, 63 sectors per track (488382MB) (LVM)
Model: Samsung SSD 860 PRO 512GB
R1.08.15 /!BIOS
Daniela's Bus Master IDE Driver for OS/2 Version 1.08.15
Controller:0 Port:01F0 IRQ:0E Status:OK BusMaster Scatter/Gather
Intel ICH7 PATA host (8086:27DF rev:02) on PCI 0:31.1#0
Unit:0 Status:OK ATAPI BusMaster UltraDMA2/PIO4
Model:DVD/CDRW UJDA775 CB03
Controller:1 Port:18D0 IRQ:10 Status:OK BusMaster Scatter/Gather
Intel ICH7 SATA host (8086:27C5 rev:02) on PCI 0:31.2#0
Unit:0 Status:OK LBA BusMaster 1.5GBit/s BPB
Model:Samsung SSD 860 PRO 512GB RVM01B6Q
OS2:log phys BPB/BIOS IDE:log phys Total Sectors
C 62260 65535 65535 16383 Avail 1000215216
H 255 16 255 16 16 OS2 1000206900
S 63 63 63 63 63 % Used 99.99

As Doug says, it is old drivers which have likely been fixed. Have to
reproduce with the latest.

I am amazed that a purported OS vendor does not have test machines, especially a T60 or equivalent, which is one of the few laptops that might operate ArcaOS.

Of course they do. From what I pick up they have quite a few. Lewis
seems to have mostly T42 or T43, Andy (not sure of his involvement) has
a T60, David, who maintains the AHCI driver, seems to have lots of
changing devices, but they seem to be mostly newer as that is the focus.
Us testers can spot problems with older hardware, hopefully.

Post by Doug Bissett
If you don't have a current subscription, you don't have support.
Bitching about that here isn't going to do anybody any good.

They have to be bug reports on current drivers, otherwise it is just
wasting developers time to tell you that your bug has been fixed. As
Doug said, make a bug report with current drivers and you will get a
positive response, including a fix most of the time and otherwise a
description of why a fix is not possible. I have access to all bug
reports and follow interesting ones, Arca Noae is very good at dealing
with bug reports but they do demand that you run the most recent,
including the debugging tools such as the script to learn about your
system.

Post by b***@gmail.com
As soon as ArcaOS has a viable browser and supports my external 1920 monitor, I will be the first to throw some more money their way.

The proof of concept Simple Browser works surprisingly well, even
gracefully recovering from crashes, with a "this tab crashed, reload?"
type of thing. It is a proof of concept and very simple with no copy and
paste support or even figuring out the URL you are going to without the
whole thing typed in, but looks good remembering it is just a test
program. Bitwise is not doing the port for free and needs money to pay
for their developer, if you can afford to shoot 20+ euros their way, it
would help. Bitwise is not Arca Noae. Use OS2VOICE for donations.
And what is the problem with your 1920 monitor? ArcaOS should support
it. I know my T42 works well when plugging in an external monitor.
Dave

b***@gmail.com

2021-03-29 05:13:17 UTC

Hi Dave:

Thanks for your response.

Before I get sidetracked and involved in details, the original intent of my penultimate message was only to let other OS/2 users know about my issue. A complex problem with a trivial solution. That is what fora are for. Ideally, Arca could log my issue, and if repeated with other users, maybe look into it. I surely don't think it is prudent and wise for me to buy a new ArcaOS just to determine whether the problem endures. I even more don't think it is worth it for Arca to troubleshoot a ticket for a single licence, when that issue is trivially circumvented as noted.

If this is the case, how do both drivers function independently and identically? Would not there be a huge performance difference?

What I meant was that DANI benches the same with AHCI removed.

Post by Dave Yeo
Think of systems that used to be common, a AHCI hard drive plugged into
a SATA port and a IDE DVD drive plugged into a IDE port. My current
computer is the first I've had with no IDE port.
Ideally the AHCI driver loads first and claims the hard drive and then
the S506 driver loads and claims the left overs, namely anything plugged
into the IDE interface, usually a DVD writer, but there has been times I
plugged in an old HD.

I have now removed OS2AHCI.ADD, as everything works fine with only DANI on a T60. I will be monitoring stability.

Post by Dave Yeo
They should co-operate.

I am currently questioning that "co-operation". Examine these two OS2AHCI$ driver files, the first with no switches, and the second ignoring the '0' (IDE CDROM) adaptor. N.B. the errors in the first instance, which I do not think should be there"

BldLevel: @#Arca Noae LLC:2.04#@##1## 12 Dec 2017 17:09:14 DAZAR1 ::::::@@AHCI Driver (c) 2017 Arca Noae LLC
CmdLine: OS2AHCI.ADD /N
port #0 interrupt error status: 0x40000001; restarting port
port #0 interrupt error status: 0x40000001; restarting port
Adapter 0: PCI=0:31:2 ID=8086:27c5 Intel ICH7M irq=16 addr=0xee444400 version=10100
Port 0:
Drive 0: 62260 cylinders, 255 heads, 63 sectors per track (488382MB) (LVM)
Model: Samsung SSD 860 PRO 512GB

BldLevel: @#Arca Noae LLC:2.04#@##1## 12 Dec 2017 17:09:14 DAZAR1 ::::::@@AHCI Driver (c) 2017 Arca Noae LLC
CmdLine: OS2AHCI.ADD /N /A:0 /I
Adapter 0: PCI=0:31:2 ID=8086:27c5 Intel ICH7M irq=16 addr=0xee444400 version=10100
Port 0:
Drive 0: 62260 cylinders, 255 heads, 63 sectors per track (488382MB) (LVM)
Model: Samsung SSD 860 PRO 512GB

R1.08.15 /!BIOS
Daniela's Bus Master IDE Driver for OS/2 Version 1.08.15

As Doug says, it is old drivers which have likely been fixed. Have to
reproduce with the latest.

Is there a newer DANIS506.ADD driver? I could not find one. On a source comparison with OS2AHCI.ADD 2.04 vs. 2.08 (the latest) there are significant changes. There also remain 3 references to DANIS506.ADD.

Post by Dave Yeo
They have to be bug reports on current drivers, otherwise it is just
wasting developers time to tell you that your bug has been fixed. As
Doug said, make a bug report with current drivers and you will get a
positive response, including a fix most of the time and otherwise a
description of why a fix is not possible. I have access to all bug
reports and follow interesting ones, Arca Noae is very good at dealing

How come the public do not have access to ArcaOS bug reports? That might mitigate all this hypothesising. Note that the last documented closed ticket on Trac was in 2014. http://trac.netlabs.org/ahci/report/6

There are still several significant open tickets.

Post by Dave Yeo
with bug reports but they do demand that you run the most recent,
including the debugging tools such as the script to learn about your
system.

It would be much easier for me to compile my own AHCI driver, but there is zero value in pursuing this, as DANI suffices perfectly.

As soon as ArcaOS has a viable browser and supports my external 1920 monitor, I will be the first to throw some more money their way.

Shareware used to be "try before you buy". The new mantra seems to be "buy before you try". I am very wary from the (as expected) failed browser development attempts. I downloaded and tried the simple browser, but even though it loaded, it did not work after sucking down ~3 GB memory. I think there needs to be an impetus to visibly provide something that simply http functions, and then add later add enhanced functionally. Are there really any basic requirements for images, or especially videos? I remember that the original WebExplorer took four days from design to release.

About ten days ago, my primary bank stopped working on all my OS/2 browsers. A few days later, another bank failed. Today, Yahoo search just stopped working on my stalwart seamonkey-2.35b7. The end is nigh!

Post by Dave Yeo
And what is the problem with your 1920 monitor? ArcaOS should support
it. I know my T42 works well when plugging in an external monitor.
Dave

I have never been able to get functional 1920 x 1080 24bpp. 16 bpp, yes. Windows and Linux work superbly and additionally provide dual monitor support. The original motivation for me to buy ArcaOS was hoping for Panorama to provide 24 bpp. I found and downloaded ATOMBIOS.SYS which (only) allowed me to get 1872 x 1080 @ 24 bpp on my ATi X1300 T60. I am currently and adequately running 1680 x 1050 @ 24 bpp, but the fidelity at native resolution on Windows (XP,7) and Linux (Ubuntu) is immensely better. I assume that Linux also uses a Panorama base, so the OS/2 functionality shortfall is confusing.

thanks for your help,
Baden

Steve Wendt

2022-03-20 04:51:53 UTC

Post by b***@gmail.com
I assume that Linux also uses a Panorama base, so the OS/2
functionality shortfall is confusing.

Hardly... Panorama is just a generic VESA driver. The (now ancient)
SNAP Graphics may still work better for you.
http://os2news.warpstock.org/ArcaOS.html#VideoCards

Doug Bissett

2021-03-29 06:15:01 UTC

You are welcome. BTW, I am only a happy customer, and I contribute
time to test what Arca Naoe is building, and yes, "excellence" is the
target.

I do NOT speak for Arca Noae.

Okay, I *finally* found what caused the traps. I was just reading on os2world how OS2AHCI.ADD is a replacement for the *506* drivers, which I never knew.

If this is the case, how do both drivers function independently and identically? Would not there be a huge performance difference?

You need to READ the OS2AHCI docs, and the Dani docs. Performance
depends, a lot, on what the hardware can do.

This doesn't tell us the version of OS2AHCI.ADD. The latest is version
2.08, from March 14 2021.

Try taking the /N parameter off of the OS2AHCI line in CONFIG.SYS. It
is known that some systems do not implement ncq support properly.

Post by b***@gmail.com
R1.08.15 /!BIOS
Daniela's Bus Master IDE Driver for OS/2 Version 1.08.15

At least this is the latet Dani driver.

Post by b***@gmail.com
Controller:0 Port:01F0 IRQ:0E Status:OK BusMaster Scatter/Gather
Intel ICH7 PATA host (8086:27DF rev:02) on PCI 0:31.1#0
Unit:0 Status:OK ATAPI BusMaster UltraDMA2/PIO4
Model:DVD/CDRW UJDA775 CB03
Controller:1 Port:18D0 IRQ:10 Status:OK BusMaster Scatter/Gather
Intel ICH7 SATA host (8086:27C5 rev:02) on PCI 0:31.2#0
Unit:0 Status:OK LBA BusMaster 1.5GBit/s BPB
Model:Samsung SSD 860 PRO 512GB RVM01B6Q
OS2:log phys BPB/BIOS IDE:log phys Total Sectors
C 62260 65535 65535 16383 Avail 1000215216
H 255 16 255 16 16 OS2 1000206900
S 63 63 63 63 63 % Used 99.99

This output indicates that you have both drivers loaded, and trying to
operate a "Model: Samsung SSD 860 PRO 512GB". Perhaps you have two
drives, or the drivers are not cooperating to select which driver gets
to run the drive. READ the docs, and add parameters to select which
driver gets to run the devices. This should never happen, but older
drivers didn't know about that.

I am amazed that a purported OS vendor does not have test machines, especially a T60 or equivalent, which is one of the few laptops that might operate ArcaOS.

There is more than one tester who has similar hardware (and there are
MANY modern laptops that work well with ArcaOS). I have an IBM
ThinkPad T43, which works well. I also test on an Asus A88XM-A (A6
processor), Asus M3A78-EM (quad processor), Lenovo ThinkPad T510,
Lenovo ThinkPad L530, IBM ThinkPad A22e, Dell GX270, Dell 4300, and an
Asus P4VP. My screens include a 1920x1080, 1280x1024, 1024x768. and an
old 1366x768 TV set, with VGA adapter. They all work well, but I
cannot run a 1280x1024 screen on the oldest Dell. The video adapter
just doesn't have what it takes. It is okay with 1024x768 (even when
using a 1280x1024 screen).

The A88XM-A, and the L530 are UEFI capable, and testing with UEFI is
under way. It might be made available, along with more language
support, in Arca OS 5.1. The T510, and the L530, are using 500 GB
Seagate Baracuda SSD drives. I use /N /F on the OS2AHCI line, in
CONFIG.SYS, on systems capable of using AHCI. I don't know what
support a T60 has for AHCI.

Post by Doug Bissett
If you don't have a current subscription, you don't have support.
Bitching about that here isn't going to do anybody any good.

Are you a customer, or a freeloader? If you are a customer, update,
and then report problems, if they still exist. If you are a
freeloader, quit bitching. That is your choice. Arca Noae doesn't have
the time to deal with freeloaders, and they can't be dealing with
ancient problems, that are probably already fixed.

Post by b***@gmail.com
As soon as ArcaOS has a viable browser and supports my external 1920 monitor, I will be the first to throw some more money their way.

1920x1080 (even larger) monitors work well, as long as the video
adapter is capable, and you have a reasonably recent version of
Panorama. NVIDIA video adapters are usually not capable, and most
older adapters just don't have the capability to run a 1920x1080
screen.

The browser is up to Bitwiseworks. Arca Noae simply supports what they
do. That is taking a lot longer than anybody expected, but all of the
indications are, that it will be available for public testing soon
(soon could be a couple of months, or even longer, if they run into
more problems).

Post by b***@gmail.com
thanks,
Baden