Hi Finn,
On Wed, Oct 18, 2017 at 4:36 PM, Finn Thain wrote:
> Perhaps falconide_intr_lock should behave more like a ref count, so that
> falconide_release_lock() holds the stdma lock up until falconide_intr_lock
> reaches zero.
We had used that in the past, can't remember when it got changed or
On Wed, 18 Oct 2017, Michael Schmitz wrote:
> On Mon, Oct 16, 2017 at 12:24 PM, Michael Schmitz
> wrote:
> >
> > Looks like a bug in ARAnyM. Haven't seen these on actual hardware for
> > a long time (i.e. not since your NCR5380 rewrite). And the NCR5380
> > driver wasn't loaded in the emulator
Hi Finn,
On Mon, Oct 16, 2017 at 12:24 PM, Michael Schmitz wrote:
> Hi Finn,
>
> On Sun, Oct 8, 2017 at 1:09 PM, Finn Thain wrote:
>> The "hdb: lost interrupt" messages from Linux might be important. The new
>
> Looks like a bug in ARAnyM. Haven't seen these on actual hardware for
> a long time
Hi Finn,
On Sun, Oct 8, 2017 at 1:09 PM, Finn Thain wrote:
> The "hdb: lost interrupt" messages from Linux might be important. The new
Looks like a bug in ARAnyM. Haven't seen these on actual hardware for
a long time (i.e. not since your NCR5380 rewrite). And the NCR5380
driver wasn't loaded in
On Sat, 7 Oct 2017, Andreas Schwab wrote:
> On Okt 07 2017, Finn Thain wrote:
>
> > which reveals bogus sector numbers, suggesting that the emulated
> > CD-ROM device is messed up.
>
> This is very likely.
>
> > [ 258.97] hdb: Optical, ATAPI CD/DVD-ROM drive
> > [ 258.97] ide0 at 0x
On 10/07/2017 10:29 PM, John Paul Adrian Glaubitz wrote:
Which module is that supposed to be?
[ 630.06] Modules linked in: ide_cd_mod cdrom ide_gd_mod falconide
ide_core evdev mac_hid atari_scsi
So, it has to be either ide_cd_mod or ide_gd_mod.
Ok, I just confirmed that both modules a
On 10/06/2017 03:34 PM, Finn Thain wrote:
> There are some modules missing from the initrd. I copied all of the
> modules from the ISO to the initrd and now I can run 'depmod' and
> 'modprobe falconide' and all the device nodes get created automatically.
I don't see which modules those are supp
Sat, 7 Oct 2017 11:52:56 +1100 (AEDT)n Fri, 6 Oct 2017, Geert Uytterhoeven
wrote:
> On Fri, Oct 6, 2017 at 3:34 PM, Finn Thain
> wrote:
> > Sat, 7 Oct 2017 00:34:48 +1100 (AEDT)n Fri, 22 Sep 2017, Geert
> > Uytterhoeven wrote:
> >> > You don't need access to hardware. You can test the kernel j
On Fri, Oct 6, 2017 at 3:34 PM, Finn Thain wrote:
> Sat, 7 Oct 2017 00:34:48 +1100 (AEDT)n Fri, 22 Sep 2017, Geert Uytterhoeven
> wrote:
>> > You don't need access to hardware. You can test the kernel just fine
>> > on Aranym. All you need is a simple root filesystem and an initrd
>> > which I ca
Sat, 7 Oct 2017 00:34:48 +1100 (AEDT)n Fri, 22 Sep 2017, Geert Uytterhoeven
wrote:
> > You don't need access to hardware. You can test the kernel just fine
> > on Aranym. All you need is a simple root filesystem and an initrd
> > which I can provide. In fact, you can just use the kernel and ini
On 10/05/2017 04:19 PM, Stefan Niestegge wrote:
Hi i just tried the CD, and after modprobing pata_falcon it found
the CD and setup can be tried. Will report my progress!
Great. Let me know when you run into any issues.
Note that after installation finishes, you need to copy the kernel
and init
Ok, this turned out to be a problem with busybox, see [1].
Installation on Aranym (and hence on a real Atari) works now.
- Use image:
-
https://cdimage.debian.org/cdimage/ports/9.0/m68k/iso-cd/debian-9.0-m68k-NETINST-1.iso
- Use kernel and initrd from the image or grab those here:
-
On 10/04/2017 06:34 PM, John Paul Adrian Glaubitz wrote:
On 10/04/2017 03:30 PM, John Paul Adrian Glaubitz wrote:
There is an updated debian-installer image available now for m68k which
should finally fix the driver issue for Atari [1].
I have just tried it myself on both Aranym and qemu-m68k.
On 10/04/2017 03:30 PM, John Paul Adrian Glaubitz wrote:
There is an updated debian-installer image available now for m68k which
should finally fix the driver issue for Atari [1].
I have just tried it myself on both Aranym and qemu-m68k.
Unfortunately, the console code now bails out with:
/sb
Hi Stefan!
On 09/21/2017 01:33 PM, John Paul Adrian Glaubitz wrote:
Please note that this is just a test kernel, it's missing all the other
modules like NIC and so on. I will have to make that change to the actual
Debian kernel package and then we need to wait for the next kernel upload
by Debia
On 09/22/2017 11:23 AM, Finn Thain wrote:
> I know who they are. Why else would I state that drivers/ide/ has official
> maintainers?
Because the drivers are broken and no one noticed it.
>>> https://people.debian.org/~glaubitz/debian-cd/debian-installer-images_20170828_m68k.tar.gz
>>
>> Use the
On Fri, 22 Sep 2017, John Paul Adrian Glaubitz wrote:
> On 09/22/2017 03:35 AM, Finn Thain wrote:
> > The stuff in drivers/ide/ still has official maintainers, so we should
> > probably try to find the regression, if that's what it is.
>
> I have my doubts given the fact that the drivers are cur
On 09/22/2017 10:17 AM, Geert Uytterhoeven wrote:
>> You don't need access to hardware. You can test the kernel just fine
>> on Aranym. All you need is a simple root filesystem and an initrd
>> which I can provide. In fact, you can just use the kernel and initrd
>> from the debian-installer distrib
On Fri, Sep 22, 2017 at 8:12 AM, John Paul Adrian Glaubitz
wrote:
> On 09/22/2017 03:35 AM, Finn Thain wrote:
>> The stuff in drivers/ide/ still has official maintainers, so we should
>> probably try to find the regression, if that's what it is.
>
> I have my doubts given the fact that the drivers
On 09/22/2017 03:35 AM, Finn Thain wrote:
> The stuff in drivers/ide/ still has official maintainers, so we should
> probably try to find the regression, if that's what it is.
I have my doubts given the fact that the drivers are currently effectively
broken and apparently no one before me noticed
On Wed, 20 Sep 2017, Geert Uytterhoeven wrote:
>
> it's quite possibly someone broke modular drivers/ide/ lately, as most
> of the world has migrated to drivers/ata/.
>
The stuff in drivers/ide/ still has official maintainers, so we should
probably try to find the regression, if that's what i
On 09/21/2017 01:30 PM, Stefan Niestegge wrote:
IDE drives (my SSD and DVDROM) are now recognized by the kernel at boot stage
already.
Instead of the old hda / hdb system i now get my SSD partitions recognzied as
sda1-sda6
Debian-netinst-m68k-CD spins up, aaand... gets mounted by D-I
Ok. Th
Am 20.09.2017 um 17:38 schrieb John Paul Adrian Glaubitz:
On 09/20/2017 04:21 PM, John Paul Adrian Glaubitz wrote:
On 09/20/2017 04:19 PM, Geert Uytterhoeven wrote:
For Falcon IDE, you can try the new driver under drivers/ata/, just
enable CONFIG_ATA and CONFIG_PATA_FALCON.
Good idea, thank
On Mon, 18 Sep 2017, Stefan Niestegge wrote:
> /sys/block is empty
>
> /sys/devices/ide0/ contains
> 0.0 0.1 ide_portuevent
>
> the file uevent is empty
>
> 0.0 and 0.1 contain
> drivename media model subsytem unload_heads
> firmware modalias serial uevent
>
> # cd 0.0
> # cat dr
On 09/20/2017 04:21 PM, John Paul Adrian Glaubitz wrote:
> On 09/20/2017 04:19 PM, Geert Uytterhoeven wrote:
>> For Falcon IDE, you can try the new driver under drivers/ata/, just
>> enable CONFIG_ATA and CONFIG_PATA_FALCON.
>
> Good idea, thanks for the hint.
Ok, that works indeed \o/.
Stefan,
On 09/20/2017 04:19 PM, Geert Uytterhoeven wrote:
For Falcon IDE, you can try the new driver under drivers/ata/, just
enable CONFIG_ATA and CONFIG_PATA_FALCON.
Good idea, thanks for the hint.
I will give it a try.
--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaub...@debian
Hi Adrian,
On Wed, Sep 20, 2017 at 3:35 PM, John Paul Adrian Glaubitz
wrote:
> On 09/20/2017 03:29 AM, Finn Thain wrote:
>>> Maybe there is a common bug which affects all the m68k IDE drivers?
>>>
>>> Or is there any additional "IDE core" module that we need?
>>
>> The 'ide-core' module is a dep
On 09/20/2017 03:29 AM, Finn Thain wrote:
Maybe there is a common bug which affects all the m68k IDE drivers?
Or is there any additional "IDE core" module that we need?
The 'ide-core' module is a dep of the 'falconide' module. So the former
will be loaded when you modprobe the latter.
But we
On Wed, 20 Sep 2017, John Paul Adrian Glaubitz wrote:
> Btw. I posted the image for testing on the a1k.org Amiga forum and
> someone managed to boot into the installer. He also reported the issue
> that the installer was unable to detect his IDE CD-ROM drive.
>
Presumably the IDE driver module
On 09/19/2017 05:36 AM, Finn Thain wrote:
> I believe that Michael used the falconide module not so long ago. We could
> probably "git bisect" to find the regression.
>
> Michael, do you have an Atari kernel build that might be used to boot an
> installer initrd and test IDE devices?
Btw. I post
On 09/19/2017 11:11 AM, Finn Thain wrote:
falconide isn't a platform driver; maybe it will have to be converted to
one if it is ever to load automatically (likewise macide). Michael, is the
platform bus appropriate for falconide?
I'd be in favor of doing that :). My kernel coding experience is
On Tue, 19 Sep 2017, John Paul Adrian Glaubitz wrote:
> Does mac_esp already support the modern kernel interfaces? i.e., does it
> have support for MOD_ALIAS()?
>
~ # cat /sys/devices/platform/mac_esp.0/modalias
platform:mac_esp
~ #
And atari_scsi is another platform driver, so it has a simil
On 09/19/2017 05:36 AM, Finn Thain wrote:
> If I boot Adrian's ISO in the patched QEMU, this is the block device I
> get:
>
> ~ # ls -l /sys/dev/block/
> lrwxrwxrwx1 root root 0 Sep 19 02:42 11:0 ->
> ../../devices/platform/mac_esp.0/host0/target0:0:2/0:0:2:0/block/sr0
> ~ #
On Mon, 18 Sep 2017, Stefan Niestegge wrote:
> Am 18.09.2017 um 15:11 schrieb John Paul Adrian Glaubitz:
> /sys/block is empty
>
If I boot Adrian's ISO in the patched QEMU, this is the block device I
get:
~ # ls -l /sys/dev/block/
lrwxrwxrwx1 root root 0 Sep 19 02:42 11:0 -
Hi Geert!
On 09/18/2017 02:10 PM, Geert Uytterhoeven wrote:
>> https://photos.app.goo.gl/zWiC2JTrGm69fL8D3
>
> Do we have dynamic majors for IDE now? Not AFAIK.
> To be 100% sure, check out:
>
> grep . /sys/class/block/*/dev
That won't work as falconide doesn't even support the sysfs interface
Am 18.09.2017 um 15:11 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 03:03 PM, Stefan Niestegge wrote:
Trying to debug this issue with ping-pong tactics will otherwise
take forever.
Doing what i can!
To speed things up, i joined #debian-68k on irc.oftc.net (alias for
irc.debian.org) an
Am 18.09.2017 um 15:11 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 03:03 PM, Stefan Niestegge wrote:
Trying to debug this issue with ping-pong tactics will otherwise
take forever.
Doing what i can!
Inspect the block device and ide0 stuff.
Adrian
/sys/block is empty
/sys/devices/id
On 09/18/2017 03:03 PM, Stefan Niestegge wrote:
Trying to debug this issue with ping-pong tactics will otherwise
take forever.
Doing what i can!
Inspect the block device and ide0 stuff.
Adrian
--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaub...@debian.org
`. `' Freie
Am 18.09.2017 um 14:34 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 02:31 PM, Stefan Niestegge wrote:
/sys exists
Then please browse the contents and let us know what's there.
Trying to debug this issue with ping-pong tactics will otherwise
take forever.
Doing what i can!
https://pho
On 09/18/2017 02:31 PM, Stefan Niestegge wrote:
/sys exists
Then please browse the contents and let us know what's there.
Trying to debug this issue with ping-pong tactics will otherwise
take forever.
Adrian
--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaub...@debian.org
Am 18.09.2017 um 14:24 schrieb Geert Uytterhoeven:
Do we have dynamic majors for IDE now? Not AFAIK.
To be 100% sure, check out:
grep . /sys/class/block/*/dev
That gives me this output:
https://photos.app.goo.gl/AxXvF9R2usQCedJ03
Oops, sysfs not mounted.
Please run "mount x /sys -t sysfs
Hi Stefan,
On Mon, Sep 18, 2017 at 2:20 PM, Stefan Niestegge wrote:
> Am 18.09.2017 um 14:10 schrieb Geert Uytterhoeven:
>> On Mon, Sep 18, 2017 at 2:02 PM, Stefan Niestegge wrote:
>>> Am 18.09.2017 um 13:56 schrieb John Paul Adrian Glaubitz:
Are you actually sure you used the correct major
Am 18.09.2017 um 14:10 schrieb Geert Uytterhoeven:
Hi Stefan,
On Mon, Sep 18, 2017 at 2:02 PM, Stefan Niestegge wrote:
Am 18.09.2017 um 13:56 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 01:55 PM, Stefan Niestegge wrote:
Am 18.09.2017 um 13:48 schrieb John Paul Adrian Glaubitz:
On 09/
Am 18.09.2017 um 14:05 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 02:02 PM, Stefan Niestegge wrote:
Are you actually sure you used the correct major/minor IDs?
yes, i did it like this
https://photos.app.goo.gl/zWiC2JTrGm69fL8D3
Well, that's not a proof. You need to match this with
Hi Stefan,
On Mon, Sep 18, 2017 at 2:02 PM, Stefan Niestegge wrote:
> Am 18.09.2017 um 13:56 schrieb John Paul Adrian Glaubitz:
>> On 09/18/2017 01:55 PM, Stefan Niestegge wrote:
>>> Am 18.09.2017 um 13:48 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 01:44 PM, Geert Uytterhoeven wrote:
>
On 09/18/2017 02:02 PM, Stefan Niestegge wrote:
Are you actually sure you used the correct major/minor IDs?
yes, i did it like this
https://photos.app.goo.gl/zWiC2JTrGm69fL8D3
Well, that's not a proof. You need to match this with
$ cat /proc/devices
Adrian
--
.''`. John Paul Adrian Glau
Am 18.09.2017 um 13:56 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 01:55 PM, Stefan Niestegge wrote:
Am 18.09.2017 um 13:48 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 01:44 PM, Geert Uytterhoeven wrote:
Unless that has happened, it won't make much sense trying to
work-around
this
On 09/18/2017 01:55 PM, Stefan Niestegge wrote:
Am 18.09.2017 um 13:48 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 01:44 PM, Geert Uytterhoeven wrote:
Unless that has happened, it won't make much sense trying to work-around
this.
You can still load it manually using "modprobe falconide",
Am 18.09.2017 um 13:48 schrieb John Paul Adrian Glaubitz:
On 09/18/2017 01:44 PM, Geert Uytterhoeven wrote:
Unless that has happened, it won't make much sense trying to work-around
this.
You can still load it manually using "modprobe falconide", right?
Yes, but the problem is that there ar
Am 18.09.2017 um 13:44 schrieb Geert Uytterhoeven:
Hi Adrian,
On Mon, Sep 18, 2017 at 1:30 PM, John Paul Adrian Glaubitz
wrote:
On 09/18/2017 01:27 PM, Stefan Niestegge wrote:
Well, the problem is that the falconide driver has not been ported to
support
MODULE_ALIAS(), so it's unable to r
On 09/18/2017 01:44 PM, Geert Uytterhoeven wrote:
Unless that has happened, it won't make much sense trying to work-around
this.
You can still load it manually using "modprobe falconide", right?
Yes, but the problem is that there are no device nodes created.
Adrian
--
.''`. John Paul Adri
Hi Adrian,
On Mon, Sep 18, 2017 at 1:30 PM, John Paul Adrian Glaubitz
wrote:
> On 09/18/2017 01:27 PM, Stefan Niestegge wrote:
>> i am sure the IDE interface is not properly set up by the kernelmodule,
>> because the CD drive does not even spin up one time. At least when the
>> mount
>> command t
On 09/18/2017 01:27 PM, Stefan Niestegge wrote:
i am sure the IDE interface is not properly set up by the kernelmodule,
because the CD drive does not even spin up one time. At least when the mount
command tries to mount the iso9660 fliesystem, it would need to access the
actual data, spin up the
Am 17.09.2017 um 13:44 schrieb Finn Thain:
On Sun, 17 Sep 2017, Stefan Niestegge wrote:
If I have the wrong filesystem type I get the same error:
# mkdir /cdrom
# modprobe iso9660
# mount -t iso9660 /dev/sr0 /cdrom
# umount /cdrom
# mount -t foo /dev/sr0 /cdrom
mount: mounting /dev/sr0 on /
On Sun, 17 Sep 2017, Stefan Niestegge wrote:
> Am 17.09.2017 um 03:00 schrieb Finn Thain:
>
> >
> > What happens if you try this:
> >
> > # mkdir /cdrom
>
> that mountpoint already exist
>
> > # modprobe iso9660
> i did modprobe isofs, which thenn adds to the output of lsmod
>
> > # mount -t
Am 17.09.2017 um 03:00 schrieb Finn Thain:
What happens if you try this:
# mkdir /cdrom
that mountpoint already exist
# modprobe iso9660
i did modprobe isofs, which thenn adds to the output of lsmod
# mount -t iso9660 /dev/hdb /cdrom
mount: mounting /dev/hdb on /cdrom failed: No suc
On Sat, 16 Sep 2017, Stefan Niestegge wrote:
> > Which means you need a working CD-ROM. From looking at the Linux docs,
> > my guess would be something like this:
> >
> > # modprobe falconide
> > # mknod /dev/hda b 3 0
> > # mknod /dev/hdb b 3 64
>
> i did these
>
> >
> > and probably,
> >
>
Am 16.09.2017 um 03:39 schrieb Finn Thain:
On Fri, 15 Sep 2017, Stefan Niestegge wrote:
Anyway, you need to get past the "Load installer components from CD" step
if you want to load the Ext filesystem modules and mount an Ext
filesystem.
i see.
Which means you need a working CD-ROM. Fr
On Fri, 15 Sep 2017, Stefan Niestegge wrote:
> Am 06.09.2017 um 23:32 schrieb John Paul Adrian Glaubitz:
>
> - Kernel detects the drive (sda, AHDI partition table) and even its
> existing partitions (sda1 and sda2)
>
> My ZIP disc was partitioned with HDDriver BGM (TOS compatible) and LNX
>
Am 15.09.2017 um 11:03 schrieb John Paul Adrian Glaubitz:
On 09/15/2017 10:35 AM, Stefan Niestegge wrote:
I'm in for any hints :-)
Can you try adding the following line to the kernel command line?
devtmpfs.mount=1
Adrian
i did so,
makes no difference. /devtmpfs is existing with or withou
Am 15.09.2017 um 11:03 schrieb John Paul Adrian Glaubitz:
On 09/15/2017 10:35 AM, Stefan Niestegge wrote:
I'm in for any hints :-)
Can you try adding the following line to the kernel command line?
devtmpfs.mount=1
Adrian
yes, i did so.
But still, kernel only creates device nodes for SCS
On Fri, 15 Sep 2017, Geert Uytterhoeven wrote:
> CONFIG_DEVTMPFS_MOUNT=y
>
It wasn't needed with QEMU, so I doubt that it is needed with Aranym.
systemd seems to have mounted /dev regardless. I mentioned this in the
other thread:
# grep devtmpfs /proc/mounts
devtmpfs /dev devtmpfs rw,relatime
On 09/15/2017 10:35 AM, Stefan Niestegge wrote:
I'm in for any hints :-)
Can you try adding the following line to the kernel command line?
devtmpfs.mount=1
Adrian
--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaub...@debian.org
`. `' Freie Universitaet Berlin - glaub...@p
Hi Geert!
On 09/15/2017 10:51 AM, Geert Uytterhoeven wrote:
The falconide isn't a platform driver with proper MODULE_ALIAS() yet,
so its module needs to be loaded based on platform heuristics.
Recent Debian may no longer have those?
That's the hint I was hoping for :-). Thanks a lot!
I will
Hi Adrian,
On Fri, Sep 15, 2017 at 10:42 AM, John Paul Adrian Glaubitz
wrote:
> On 09/15/2017 10:35 AM, Stefan Niestegge wrote:
>> - its possible to "modprobe falconide" as well as cdrom, isofs modules
The falconide isn't a platform driver with proper MODULE_ALIAS() yet,
so its module needs to
On 09/15/2017 10:35 AM, Stefan Niestegge wrote:
- its possible to "modprobe falconide" as well as cdrom, isofs modules
- dmesg then shows that my Harddrive and my DVD Burner were found as hda, hdb
- but no device nodes are created for hda, hdb
Yes, this is what I need to debug next. I had hoped
Am 06.09.2017 um 23:32 schrieb John Paul Adrian Glaubitz:
As i read on other thread, D-I fails to detect my IDE-connected DVD ROM
drive due to missing PATA module.
I will work on this issue the next days and try to figure what we're missing.
I will let you know once I have an updated imag
>
> Ataboot.prg unpacks the kernel, unpacks the initrd and complains about
> "6815744 file | out of memory for ramdisk image" and lets me return to
> TOS desktop.
>
This is what that looks like: https://youtu.be/9PJxPiSkuvQ
I would highly apreciate get ataboot working instead of the bootstra.t
Hi Stefan!
On 09/06/2017 08:21 PM, Stefan Niestegge wrote:
> i want to try helping improve the new Debian/m68k netinstall that
> John Paul Adrian Glaubitz just released.
You can just say Adrian, that saves typing ;).
> (...)
> Now i can successfully use bootstra.tos to unpack and boot the kernel
Hello,
i want to try helping improve the new Debian/m68k netinstall that
John Paul Adrian Glaubitz just released.
Small introduction of myself and my setup:
-Stefan Niestegge, LinuxMint user, only basic linux skills.
i can install from a live CD/DVD, use apt-get and edit
config f
70 matches
Mail list logo
