Opened 4 years ago

Last modified 4 years ago

#22 accepted defect

Unable to reboot using SET(A)BOOT

Reported by: Doug Bissett Owned by: Ben Rietbroek
Priority: minor Milestone:
Component: Boot Manager Version: 1.0
Keywords: setboot, setaboot Cc:

Description

This is probably related to #7 and #8, but I want to make a new incident. I am using AirBoot? 1.1.0.

My setup is a little non-standard. I have an older processor, with only IDE adapters, and found a need to use a large (2 TB) disk, so I got a 4 Port PCI SATA RAID Controller Adapter Card StarTech? ID: PCISATA4R1. It works in non-RAID mode, but it sees the 2 TB disk as disk 0, with the IDE disk being disk 1 (as shown by DFSEE). Normal boot, from the IDE disk, is okay, but trying "SETBOOT /IBA:eCS21" causes the shutdown to proceed to the point where it should say that it is rebooting, but that never happens. I need to invoke CADH, and reboot from there. The indication is exactly the same as if the partition eCS21 doesn't exist. I suspect that it is only looking on disk 0, which has no bootable partitions. This becomes a little awkward because the computer is hidden under my basement stairs, and I would like to be able to reboot remotely (by command).

Thanks...

Attachments (3)

ibms506.txt (1.3 KB) - added by Doug Bissett 4 years ago.
From: copy IBMs506$ ibm506.txt
danis506_add_Build (456 bytes) - added by Doug Bissett 4 years ago.
Bldlevel from Danis506.add
Sly-config.sys (7.1 KB) - added by Doug Bissett 4 years ago.
Config.sys

Download all attachments as: .zip

Change History (16)

comment:1 in reply to:  description ; Changed 4 years ago by Ben Rietbroek

Replying to dgbisse:

This is probably related to #7 and #8, but I want to make a new incident. I am using AirBoot? 1.1.0.

My setup is a little non-standard. I have an older processor, with only IDE adapters, and found a need to use a large (2 TB) disk, so I got a 4 Port PCI SATA RAID Controller Adapter Card StarTech? ID: PCISATA4R1. It works in non-RAID mode, but it sees the 2 TB disk as disk 0, with the IDE disk being disk 1 (as shown by DFSEE).

So this is when OS/2 is up, yes ?
And which driver(s) are you using ?
AHCI or DANI or both, and if both are being loaded by CONFIG.SYS, what is the order ?

Normal boot, from the IDE disk, is okay,

But does the IDE gets booted as DISK0 or DISK1 ?
You can press TAB when the AirBoot? menu is displayed to show the disk order it found.
(AirBoot? uses BIOS disk-order and OS/2 uses disk-order imposed by drivers, which can be different)

but trying "SETBOOT /IBA:eCS21" causes the shutdown to proceed to the point where it should say that it is rebooting, but that never happens. I need to invoke CADH, and reboot from there.

Do you have eStyler Shutdown enabled ?
If so, can you try again with eStyler Shutdown disabled ?
Also, when SETBOOT does not return and you have to enter CADH, can you take a look at the process list (top), see if SETBOOT is there and try killing it. If so, you should then be able to return to the WPS.

The indication is exactly the same as if the partition eCS21 doesn't exist. I suspect that it is only looking on disk 0, which has no bootable partitions.

If the eCS21 partition could not be found SETBOOT should return. The fact that it does not seems to indicate the shutdown process is stalling. This could be caused by eStyler Shutdown.

This becomes a little awkward because the computer is hidden under my basement stairs, and I would like to be able to reboot remotely (by command).

Roger that, let's track down what goes wrong...

Ben.

comment:2 in reply to:  1 ; Changed 4 years ago by Doug Bissett

Replying to rousseau:

Replying to dgbisse:

This is probably related to #7 and #8, but I want to make a new incident. I am using AirBoot? 1.1.0.

My setup is a little non-standard. I have an older processor, with only IDE adapters, and found a need to use a large (2 TB) disk, so I got a 4 Port PCI SATA RAID Controller Adapter Card StarTech? ID: PCISATA4R1. It works in non-RAID mode, but it sees the 2 TB disk as disk 0, with the IDE disk being disk 1 (as shown by DFSEE).

So this is when OS/2 is up, yes ?

Yes.

And which driver(s) are you using ?
AHCI or DANI or both, and if both are being loaded by CONFIG.SYS, what is the order ?

DANI. There is no option to use AHCI.

Normal boot, from the IDE disk, is okay,

But does the IDE gets booted as DISK0 or DISK1 ?
You can press TAB when the AirBoot? menu is displayed to show the disk order it found.
(AirBoot? uses BIOS disk-order and OS/2 uses disk-order imposed by drivers, which can be different)

Okay, TAB shows 80H as being the old 40 GB HDD, and 81H as being the 2 TB drive. That is what I expected, but DFSEE shows them the other way around:

+---+--+--+-----------------+--------+--------+-----------+----------------------+---------+
|ID |ux|Dr|Type, description|Format  |Related |VolumeLabel|LVM Vol,Part / GPT nam| Size MiB|
+---</dev/hda    MBR disk  1>--------+--------+----------<[ D4 ]                  >--------+
|01>| 1|T:|Prim 35 Warp-LVM |JFS     |IBM 4.50|2GBJFS     |2TBJFS, 2TBJFS        |1907728.0|
+---</dev/hdb    MBR disk  2>--------+--------+----------<[ D1 ]                  >--------+
|02 |  |  |FreeSpace Primary|-- -- --|-- -- --|- - - - - -|                      |      7.8|
|02 | 5|W:|Log  07 Inst-FSys|JFS     |IBM 4.50|WDrive     |WDrive, WDrive        |  34075.3|
|03 | 6|C:|Log  07 Inst-FSys|JFS     |IBM 4.50|CDrive     |eCS21, eCS21          |   1537.4|
|04 | 7|M:|Log  07 Inst-FSys|JFS     |IBM 4.50|ECS22      |eCS22, eCS22          |   1537.4|
|05 | 8|Y:|Log  07 Inst-FSys|JFS     |IBM 4.50|MAINTENANCE|bootAble, bootAble    |   1004.0|
+---+--+--+-----------------+--------+--------+-----------+----------------------+---------+
   > = Active/Startable * = Bootable r = Removable R = Removable+Active/Bootable (possibly)

I just tried booting the DFSEE CD (freeDOS), and it shows the 40 GB disk as disk 1 and the 2 TB disk as disk 2. Exactly the opposite of what DFSEE in eCS shows. I guess I need to look in the DANI docs, to see if there is anything that it can do.

but trying "SETBOOT /IBA:eCS21" causes the shutdown to proceed to the point where it should say that it is rebooting, but that never happens. I need to invoke CADH, and reboot from there.

Do you have eStyler Shutdown enabled ?

I don't think so. It is the default setup as installed by eCS 2.1, except I installed XWorkplace (xwp-3.unofficial-1-0-10.zip) over it. I believe it uses the XWorkplace shutdown (which may be where the fault lies).

If so, can you try again with eStyler Shutdown disabled ?
Also, when SETBOOT does not return and you have to enter CADH, can you take a look at the process list (top), see if SETBOOT is there and try killing it. If so, you should then be able to return to the WPS.

SETBOOT is not in the list.

The indication is exactly the same as if the partition eCS21 doesn't exist. I suspect that it is only looking on disk 0, which has no bootable partitions.

If the eCS21 partition could not be found SETBOOT should return. The fact that it does not seems to indicate the shutdown process is stalling. This could be caused by eStyler Shutdown.

It looks like it probably returned.

This becomes a little awkward because the computer is hidden under my basement stairs, and I would like to be able to reboot remotely (by command).

Roger that, let's track down what goes wrong...

Ben.

Thanks...

comment:3 in reply to:  2 Changed 4 years ago by Ben Rietbroek

Replying to dgbisse:

Replying to rousseau:

Replying to dgbisse:

This is probably related to #7 and #8, but I want to make a new incident. I am using AirBoot? 1.1.0.

My setup is a little non-standard. I have an older processor, with only IDE adapters, and found a need to use a large (2 TB) disk, so I got a 4 Port PCI SATA RAID Controller Adapter Card StarTech? ID: PCISATA4R1. It works in non-RAID mode, but it sees the 2 TB disk as disk 0, with the IDE disk being disk 1 (as shown by DFSEE).

So this is when OS/2 is up, yes ?

Yes.

And which driver(s) are you using ?
AHCI or DANI or both, and if both are being loaded by CONFIG.SYS, what is the order ?

DANI. There is no option to use AHCI.

Normal boot, from the IDE disk, is okay,

But does the IDE gets booted as DISK0 or DISK1 ?
You can press TAB when the AirBoot? menu is displayed to show the disk order it found.
(AirBoot? uses BIOS disk-order and OS/2 uses disk-order imposed by drivers, which can be different)

Okay, TAB shows 80H as being the old 40 GB HDD, and 81H as being the 2 TB drive. That is what I expected, but DFSEE shows them the other way around:

+---+--+--+-----------------+--------+--------+-----------+----------------------+---------+
|ID |ux|Dr|Type, description|Format  |Related |VolumeLabel|LVM Vol,Part / GPT nam| Size MiB|
+---</dev/hda    MBR disk  1>--------+--------+----------<[ D4 ]                  >--------+
|01>| 1|T:|Prim 35 Warp-LVM |JFS     |IBM 4.50|2GBJFS     |2TBJFS, 2TBJFS        |1907728.0|
+---</dev/hdb    MBR disk  2>--------+--------+----------<[ D1 ]                  >--------+
|02 |  |  |FreeSpace Primary|-- -- --|-- -- --|- - - - - -|                      |      7.8|
|02 | 5|W:|Log  07 Inst-FSys|JFS     |IBM 4.50|WDrive     |WDrive, WDrive        |  34075.3|
|03 | 6|C:|Log  07 Inst-FSys|JFS     |IBM 4.50|CDrive     |eCS21, eCS21          |   1537.4|
|04 | 7|M:|Log  07 Inst-FSys|JFS     |IBM 4.50|ECS22      |eCS22, eCS22          |   1537.4|
|05 | 8|Y:|Log  07 Inst-FSys|JFS     |IBM 4.50|MAINTENANCE|bootAble, bootAble    |   1004.0|
+---+--+--+-----------------+--------+--------+-----------+----------------------+---------+
   > = Active/Startable * = Bootable r = Removable R = Removable+Active/Bootable (possibly)

I just tried booting the DFSEE CD (freeDOS), and it shows the 40 GB disk as disk 1 and the 2 TB disk as disk 2. Exactly the opposite of what DFSEE in eCS shows. I guess I need to look in the DANI docs, to see if there is anything that it can do.

but trying "SETBOOT /IBA:eCS21" causes the shutdown to proceed to the point where it should say that it is rebooting, but that never happens. I need to invoke CADH, and reboot from there.

Do you have eStyler Shutdown enabled ?

I don't think so. It is the default setup as installed by eCS 2.1, except I installed XWorkplace (xwp-3.unofficial-1-0-10.zip) over it. I believe it uses the XWorkplace shutdown (which may be where the fault lies).

If so, can you try again with eStyler Shutdown disabled ?
Also, when SETBOOT does not return and you have to enter CADH, can you take a look at the process list (top), see if SETBOOT is there and try killing it. If so, you should then be able to return to the WPS.

SETBOOT is not in the list.

The indication is exactly the same as if the partition eCS21 doesn't exist. I suspect that it is only looking on disk 0, which has no bootable partitions.

If the eCS21 partition could not be found SETBOOT should return. The fact that it does not seems to indicate the shutdown process is stalling. This could be caused by eStyler Shutdown.

It looks like it probably returned.

This becomes a little awkward because the computer is hidden under my basement stairs, and I would like to be able to reboot remotely (by command).

Roger that, let's track down what goes wrong...

Ben.

Thanks...

Ok, it is starting to look that your initial analysis was correct and that SETABOOT tries to read the AirBoot? configuration from DISK0 which is now the 2TB. But I wonder why DANI puts the SATA disk first. Can you paste the output from command:

type ibms506$

here please ?

Changed 4 years ago by Doug Bissett

Attachment: ibms506.txt added

From: copy IBMs506$ ibm506.txt

comment:4 Changed 4 years ago by Doug Bissett

IBMs506.txt attached.

Last edited 4 years ago by Doug Bissett (previous) (diff)

comment:5 in reply to:  4 ; Changed 4 years ago by Ben Rietbroek

Replying to dgbisse:

IBMs506.txt attached.

If I do a type ibms506$ I get:
R1.8.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.8.15

and you get (note the leading 0:
R1.08.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.08.15

My BLDLEVEL of DANIS506.ADD says:
Revision: 1.08.15
File Version: 1.8.15

Now, I know that BLDLEVEL does the leading 0 in Revision, but it
seems strange to me that your driver displays it when typing ibms506$
Unless its really a different version from the one I use (eCS22-beta).

Anyway:
Why does DANI put the SATA-controller first is the question...
Can you also post CONFIG.SYS and the BLDLEVEL of your DANIS506.ADD here ?
And can you check your BIOS settings if it does some strange controller
or drive order ? (Guess not coz AB would see swapped disks then too...)

This issue is probably fixable by using DANI for the IDE and DVD-drive,
prohibiting it from handling the SATA-controller and putting AHCI *after*
DANI in CONFIG.SYS to handle the SATA-controller.

Which booted system is this, eCS21 or eCS22 ?

Changed 4 years ago by Doug Bissett

Attachment: danis506_add_Build added

Bldlevel from Danis506.add

Changed 4 years ago by Doug Bissett

Attachment: Sly-config.sys added

Config.sys

comment:6 in reply to:  5 ; Changed 4 years ago by Doug Bissett

Replying to rousseau:

Replying to dgbisse:

IBMs506.txt attached.

If I do a type ibms506$ I get:
R1.8.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.8.15

and you get (note the leading 0:
R1.08.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.08.15

My BLDLEVEL of DANIS506.ADD says:
Revision: 1.08.15
File Version: 1.8.15

Bldlevel attached.

Now, I know that BLDLEVEL does the leading 0 in Revision, but it
seems strange to me that your driver displays it when typing ibms506$
Unless its really a different version from the one I use (eCS22-beta).

It is the latest from the Arca Noae driver pack. I believe that it is essentially the same as from eCS 2.2b2, with a few cosmetic changes.

Anyway:
Why does DANI put the SATA-controller first is the question...
Can you also post CONFIG.SYS and the BLDLEVEL of your DANIS506.ADD here ?
And can you check your BIOS settings if it does some strange controller
or drive order ? (Guess not coz AB would see swapped disks then too...)

Sly-config.sys added. AHCI is not used, there is no option in BIOS to use it. The machine BIOS does not show the SATA disk at all. It is visible using the BIOS plug in (resident on the card), but that doesn't show the IDE disks, and there are no configuration settings for anything like that.

This issue is probably fixable by using DANI for the IDE and DVD-drive,
prohibiting it from handling the SATA-controller and putting AHCI *after*
DANI in CONFIG.SYS to handle the SATA-controller.

I can't use AHCI. it is not supported by the hardware.

Which booted system is this, eCS21 or eCS22 ?

It is eCS 2.1, with updates. I do have eCS 2.2b2 on the machine, but it is not configured. It does configure the disks the same way though.

comment:7 in reply to:  6 ; Changed 4 years ago by Ben Rietbroek

Replying to dgbisse:

Replying to rousseau:

Replying to dgbisse:

IBMs506.txt attached.

If I do a type ibms506$ I get:
R1.8.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.8.15

and you get (note the leading 0:
R1.08.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.08.15

My BLDLEVEL of DANIS506.ADD says:
Revision: 1.08.15
File Version: 1.8.15

Bldlevel attached.

Now, I know that BLDLEVEL does the leading 0 in Revision, but it
seems strange to me that your driver displays it when typing ibms506$
Unless its really a different version from the one I use (eCS22-beta).

It is the latest from the Arca Noae driver pack. I believe that it is essentially the same as from eCS 2.2b2, with a few cosmetic changes.

Yep, I see, you are running the latest David-Build and he changed the version display here:
http://trac.netlabs.org/xata/changeset?reponame=&new=122%40%2F&old=121%40%2F
(My box was freshly installed a few days ago and I did not update all stuff yet)
Right, so the DANI level is OK.

Anyway:
Why does DANI put the SATA-controller first is the question...
Can you also post CONFIG.SYS and the BLDLEVEL of your DANIS506.ADD here ?
And can you check your BIOS settings if it does some strange controller
or drive order ? (Guess not coz AB would see swapped disks then too...)

Sly-config.sys added. AHCI is not used, there is no option in BIOS to use it. The machine BIOS does not show the SATA disk at all. It is visible using the BIOS plug in (resident on the card), but that doesn't show the IDE disks, and there are no configuration settings for anything like that.

That would be normal because the IDE-disks are connected to your on-board controller and the SATA-controller resident BIOS is only for that controller.

This issue is probably fixable by using DANI for the IDE and DVD-drive,
prohibiting it from handling the SATA-controller and putting AHCI *after*
DANI in CONFIG.SYS to handle the SATA-controller.

I can't use AHCI. it is not supported by the hardware.

Do you mean the SATA-controller does not support AHCI ?

Which booted system is this, eCS21 or eCS22 ?

It is eCS 2.1, with updates. I do have eCS 2.2b2 on the machine, but it is not configured. It does configure the disks the same way though.

OK
Experiment 1
Usually on controllers that support AHCI, the first 2 or 4 ports can be in IDE-emulation as the default. What happens if you move the SATA-disk from port 0 to port 3 ?
(This of course may not apply if AHCI is not supported, but just try it anyway)

Experiment 2
Change your CONFIG.SYS as from:

REM BASEDEV=OS2AHCI.ADD
BASEDEV=DANIS506.ADD /!BIOS

to

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=OS2AHCI.ADD

and if that does not work try:

BASEDEV=DANIS506.ADD /!BIOS /I /A:1
BASEDEV=OS2AHCI.ADD

The above will of course not work if the controller does not support AHCI.
In that case, you could try:

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=DANIS506.ADD /!BIOS /I
REM BASEDEV=OS2AHCI.ADD

or

BASEDEV=DANIS506.ADD /!BIOS /I /A:1
BASEDEV=DANIS506.ADD /!BIOS /I
REM BASEDEV=OS2AHCI.ADD

But I am not sure if DANI can be loaded twice as done above to force controller order and there may be a better way to do that.

The question remains why DANI puts the SATA-controller first and how DANI can be instructed to swap this controller order.

If none of the above tricks work we might need to see if David or Steven can be seduced to shed some more light on this...

This is a link to a PDF for your SATA-controller b.t.w., but I assume you already have something like that:
https://www.cclonline.com/content/pdfs/a9GVpvRB-2fbDrXnl33AfCbQ-3d-3d.pdf

comment:8 Changed 4 years ago by Ben Rietbroek

Owner: set to Ben Rietbroek
Status: newaccepted

comment:9 Changed 4 years ago by Ben Rietbroek

Keywords: setboot setaboot added

comment:10 in reply to:  7 ; Changed 4 years ago by Doug Bissett

Replying to rousseau:

Replying to dgbisse:

Replying to rousseau:

Replying to dgbisse:

IBMs506.txt attached.

If I do a type ibms506$ I get:
R1.8.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.8.15

and you get (note the leading 0:
R1.08.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.08.15

My BLDLEVEL of DANIS506.ADD says:
Revision: 1.08.15
File Version: 1.8.15

Bldlevel attached.

Now, I know that BLDLEVEL does the leading 0 in Revision, but it
seems strange to me that your driver displays it when typing ibms506$
Unless its really a different version from the one I use (eCS22-beta).

It is the latest from the Arca Noae driver pack. I believe that it is essentially the same as from eCS 2.2b2, with a few cosmetic changes.

Yep, I see, you are running the latest David-Build and he changed the version display here:
http://trac.netlabs.org/xata/changeset?reponame=&new=122%40%2F&old=121%40%2F
(My box was freshly installed a few days ago and I did not update all stuff yet)
Right, so the DANI level is OK.

Good...

Anyway:
Why does DANI put the SATA-controller first is the question...
Can you also post CONFIG.SYS and the BLDLEVEL of your DANIS506.ADD here ?
And can you check your BIOS settings if it does some strange controller
or drive order ? (Guess not coz AB would see swapped disks then too...)

Sly-config.sys added. AHCI is not used, there is no option in BIOS to use it. The machine BIOS does not show the SATA disk at all. It is visible using the BIOS plug in (resident on the card), but that doesn't show the IDE disks, and there are no configuration settings for anything like that.

That would be normal because the IDE-disks are connected to your on-board controller and the SATA-controller resident BIOS is only for that controller.

This issue is probably fixable by using DANI for the IDE and DVD-drive,
prohibiting it from handling the SATA-controller and putting AHCI *after*
DANI in CONFIG.SYS to handle the SATA-controller.

I can't use AHCI. it is not supported by the hardware.

Do you mean the SATA-controller does not support AHCI ?

No mention of it anywhere. It is a PCI card.

Which booted system is this, eCS21 or eCS22 ?

It is eCS 2.1, with updates. I do have eCS 2.2b2 on the machine, but it is not configured. It does configure the disks the same way though.

OK
Experiment 1
Usually on controllers that support AHCI, the first 2 or 4 ports can be in IDE-emulation as the default. What happens if you move the SATA-disk from port 0 to port 3 ?
(This of course may not apply if AHCI is not supported, but just try it anyway)

No change...

Experiment 2
Change your CONFIG.SYS as from:

REM BASEDEV=OS2AHCI.ADD
BASEDEV=DANIS506.ADD /!BIOS

to

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=OS2AHCI.ADD

and if that does not work try:

BASEDEV=DANIS506.ADD /!BIOS /I /A:1
BASEDEV=OS2AHCI.ADD

The above will of course not work if the controller does not support AHCI.

Doesn't work. No AHCI support.

In that case, you could try:

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=DANIS506.ADD /!BIOS /I
REM BASEDEV=OS2AHCI.ADD

Close, but no cigar. It needs to be:

BASEDEV=DANIS506.ADD /!BIOS /A:0 /I
BASEDEV=DANIS506.ADD /!BIOS
REM BASEDEV=OS2AHCI.ADD

Which works, with one problem <see below>.

or

BASEDEV=DANIS506.ADD /!BIOS /I /A:1
BASEDEV=DANIS506.ADD /!BIOS /I
REM BASEDEV=OS2AHCI.ADD

But I am not sure if DANI can be loaded twice as done above to force controller order and there may be a better way to do that.

Apparently, it can be loaded twice (never thought of that), and it does solve my problem. It does, however, cause a problem when I do "TYPE IBMS506$". It only shows what the second instance did (and it calls it Controller:0), it does not show the data for the first instance. So far, I have no indication of any other problem.

The question remains why DANI puts the SATA-controller first, and how DANI can be instructed to swap this controller order.

It would be good if only one instance of Dani would do it properly, but I don't see any way to do it.

If none of the above tricks work we might need to see if David or Steven can be seduced to shed some more light on this...

No need to bother them, but they should be made aware of the problem so it can be sorted out, someday.

This is a link to a PDF for your SATA-controller b.t.w., but I assume you already have something like that:
https://www.cclonline.com/content/pdfs/a9GVpvRB-2fbDrXnl33AfCbQ-3d-3d.pdf

Yes. I already have that.

Okay, my immediate problem is solved, although not ideally (and it reminded me that I wanted to install a chassis fan in that box, which also got done), so you can do what you like with this incident.

Many thanks...

comment:11 in reply to:  10 Changed 4 years ago by Ben Rietbroek

Replying to dgbisse:

Replying to rousseau:

Replying to dgbisse:

Replying to rousseau:

Replying to dgbisse:

IBMs506.txt attached.

If I do a type ibms506$ I get:
R1.8.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.8.15

and you get (note the leading 0:
R1.08.15 /!BIOS

Daniela's Bus Master IDE Driver for OS/2 Version 1.08.15

My BLDLEVEL of DANIS506.ADD says:
Revision: 1.08.15
File Version: 1.8.15

Bldlevel attached.

Now, I know that BLDLEVEL does the leading 0 in Revision, but it
seems strange to me that your driver displays it when typing ibms506$
Unless its really a different version from the one I use (eCS22-beta).

It is the latest from the Arca Noae driver pack. I believe that it is essentially the same as from eCS 2.2b2, with a few cosmetic changes.

Yep, I see, you are running the latest David-Build and he changed the version display here:
http://trac.netlabs.org/xata/changeset?reponame=&new=122%40%2F&old=121%40%2F
(My box was freshly installed a few days ago and I did not update all stuff yet)
Right, so the DANI level is OK.

Good...

Anyway:
Why does DANI put the SATA-controller first is the question...
Can you also post CONFIG.SYS and the BLDLEVEL of your DANIS506.ADD here ?
And can you check your BIOS settings if it does some strange controller
or drive order ? (Guess not coz AB would see swapped disks then too...)

Sly-config.sys added. AHCI is not used, there is no option in BIOS to use it. The machine BIOS does not show the SATA disk at all. It is visible using the BIOS plug in (resident on the card), but that doesn't show the IDE disks, and there are no configuration settings for anything like that.

That would be normal because the IDE-disks are connected to your on-board controller and the SATA-controller resident BIOS is only for that controller.

This issue is probably fixable by using DANI for the IDE and DVD-drive,
prohibiting it from handling the SATA-controller and putting AHCI *after*
DANI in CONFIG.SYS to handle the SATA-controller.

I can't use AHCI. it is not supported by the hardware.

Do you mean the SATA-controller does not support AHCI ?

No mention of it anywhere. It is a PCI card.

Which booted system is this, eCS21 or eCS22 ?

It is eCS 2.1, with updates. I do have eCS 2.2b2 on the machine, but it is not configured. It does configure the disks the same way though.

OK
Experiment 1
Usually on controllers that support AHCI, the first 2 or 4 ports can be in IDE-emulation as the default. What happens if you move the SATA-disk from port 0 to port 3 ?
(This of course may not apply if AHCI is not supported, but just try it anyway)

No change...

Experiment 2
Change your CONFIG.SYS as from:

REM BASEDEV=OS2AHCI.ADD
BASEDEV=DANIS506.ADD /!BIOS

to

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=OS2AHCI.ADD

and if that does not work try:

BASEDEV=DANIS506.ADD /!BIOS /I /A:1
BASEDEV=OS2AHCI.ADD

The above will of course not work if the controller does not support AHCI.

Doesn't work. No AHCI support.

In that case, you could try:

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=DANIS506.ADD /!BIOS /I
REM BASEDEV=OS2AHCI.ADD

Close, but no cigar. It needs to be:

BASEDEV=DANIS506.ADD /!BIOS /A:0 /I
BASEDEV=DANIS506.ADD /!BIOS
REM BASEDEV=OS2AHCI.ADD

Which works, with one problem <see below>.

or

BASEDEV=DANIS506.ADD /!BIOS /I /A:1
BASEDEV=DANIS506.ADD /!BIOS /I
REM BASEDEV=OS2AHCI.ADD

But I am not sure if DANI can be loaded twice as done above to force controller order and there may be a better way to do that.

Apparently, it can be loaded twice (never thought of that), and it does solve my problem. It does, however, cause a problem when I do "TYPE IBMS506$". It only shows what the second instance did (and it calls it Controller:0), it does not show the data for the first instance. So far, I have no indication of any other problem.

The question remains why DANI puts the SATA-controller first, and how DANI can be instructed to swap this controller order.

It would be good if only one instance of Dani would do it properly, but I don't see any way to do it.

If none of the above tricks work we might need to see if David or Steven can be seduced to shed some more light on this...

No need to bother them, but they should be made aware of the problem so it can be sorted out, someday.

I did send them a 'take a look at this ticket when you have a moment' e-mail,
and they might already be aware of this problem.

This is a link to a PDF for your SATA-controller b.t.w., but I assume you already have something like that:
https://www.cclonline.com/content/pdfs/a9GVpvRB-2fbDrXnl33AfCbQ-3d-3d.pdf

Yes. I already have that.

Okay, my immediate problem is solved, although not ideally (and it reminded me that I wanted to install a chassis fan in that box, which also got done), so you can do what you like with this incident.

Many thanks...

Ok, glad we could find a workable solution.
This ticket stays open and I submitted #23 which describes it in short.
It's on the fix-list for the v1.1.2 release.
Thanks for the nice tracking teamwork !

comment:12 Changed 4 years ago by Ben Rietbroek

This

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=DANIS506.ADD /!BIOS /I
REM BASEDEV=OS2AHCI.ADD

was a typo, I meant

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=DANIS506.ADD /!BIOS
REM BASEDEV=OS2AHCI.ADD

But even if the above would also work, it will not make type ibms506$ display both controllers because I guess only the last loaded driver handles the IOCtl providing the info.

comment:13 in reply to:  12 Changed 4 years ago by Doug Bissett

Replying to rousseau:

This

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=DANIS506.ADD /!BIOS /I
REM BASEDEV=OS2AHCI.ADD

was a typo, I meant

BASEDEV=DANIS506.ADD /!BIOS /I /A:0
BASEDEV=DANIS506.ADD /!BIOS
REM BASEDEV=OS2AHCI.ADD

Actually, that is still wrong. You need to tell it that you are working with adapter 0 before you tell it to ignore it. Using /I before the /A:0 tells it to ignore all of the adapters (and that instance of Dani doesn't take any of them). The proper incantation is:

BASEDEV=DANIS506.ADD /!BIOS /A:0 /I
BASEDEV=DANIS506.ADD /!BIOS
REM BASEDEV=OS2AHCI.ADD

But even if the above would also work, it will not make type ibms506$ display both controllers because I guess only the last loaded driver handles the IOCtl providing the info.

I don't know the details, but I suspect that each instance doesn't even know about the other, so the last one loaded supplies the information.

Thanks again, and feel free to ask me to test something.

Note: See TracTickets for help on using tickets.