Wakepup2 Aug 2008 - floppy image for booting from USB

Using applications, configuring, problems
Post Reply
Message
Author
User avatar
Crash
Posts: 453
Joined: Fri 09 Dec 2005, 06:34
Location: Melbourne, FL

#16 Post by Crash »

otropogo wrote:the worst regression is with the pcmcia_scsi
One change I made on this version was to archive some of the large files that I didn't think were used. One of them was config.exe in the \driver\cardsoft directory. My tests didn't indicate that it was needed, but I don't have the hardware to test it fully either.

Does the "August 2008" version still work? If so, please re-name the config.exe file to something else on that version and re-run it. If it fails, then I was wrong. If that is the case, I will overwrite the "Sept 2008" image with an image that has that file put back.

otropogo

Re: Sept 2008 version

#17 Post by otropogo »

John Doe wrote:
otropogo wrote:PS. it's also maddening to have to boot up Windows to write the floppy image because Puppy doesn't have Rawrite available.
like this:

Code: Select all

dd if=wakepup2.img of=/dev/fd0
double check that the value of 'of' device is correct or you can wipe out an incorrect disk.
You forgot to mention that you'll also need to either go to the directory where the image file is located,

like this

Code: Select all

cd /mnt/hda3/AllPups/Wakepup2_08
or put the path of the directory into the command (even harder).

Anyway, since I had to reburn the August floppy img, I tried it your way. I had to go online to find your post, then enter the command and the path manually, which took several corrections, due to limited typing accuracy and forgetting about case sensitivity, and then, when it was all finally ready, Puppy failed to write the image (on a perfectly good floppy I had just used several times to boot and with an image file that had just passed the md5sum test) and gave me an input output error instead.

So now I get to do what I should have done in the first place - copy the img file to a flash card, take it to my Win98 machine, and write the image with Rawrite.

If I'd done that originally, it would have taken me a tenth of the time I spent getting less than nowhere with Puppy's command line. And I say "less than nowhere" advisedly, because now, if I'm lucky, I'll have to format the disk in Windows before writing the image. And if I'm not, the disk will be toast, and I'll have to find another floppy to write to....

otropogo

#18 Post by otropogo »

Crash wrote:
otropogo wrote:the worst regression is with the pcmcia_scsi
...
Does the "August 2008" version still work?.....

I'm afraid not. IIRC, I complained about this loss of pcmcia functionality some time ago, but you didn't respond, and then I got caught up in the parallel port testing and forgot about the pcmcia.

It must have happened while we were still posting in the "Bogus PP ZIP..." thread.

Here's the last reference I could find, from Jul 09, of Wakepup recognizing the pcmcia_scsi chain:

Crash :
otropogo,

Is this the path that you took using the new pcmcia.bat to attempt to boot to the SCSI CD?

First menu:
7. PCMCIA Devices (Experimental)

Next menu, "Select PCMCIA, Zip Parallel, or SCSI":
1. PCMCIA

Next menu, "Ready to load ASPI Host Adapter Drivers":
1. All


yes

Crash :
Next menu, "Ready to load Peripherals":
6. aspicd.sys /d:USB-CD


I've generally chosen 1. All at this menu, but will try 6. alone as you suggest .

And yes, after removing your "rem", I now get the same initial result as with the original pcmcia.bat:

ASPI CD-ROM Driver of DOS Version4.01...

Host Adapter 0, Target SCSI ID = 4 - LUN 0: SONY CD-RW CRX145S 1.0c CD-ROM driver installed: I host adapter(s), 1 target(s),

Then, after pressing "any key" twice, I see the CD-ROM's access light go on for several seconds, but all wakepup reports is :

"file found on idehd, drive C:\
But I know we got a bit beyond that, to the point of having Wakepup2 recognize the version of Puppy on the CD, exactly as in the PP ZIP report just now. And then we lost it all....

One step forward, and two steps backward.

User avatar
Crash
Posts: 453
Joined: Fri 09 Dec 2005, 06:34
Location: Melbourne, FL

#19 Post by Crash »

Quote wrote:Does the "August 2008" version still work?.....

I'm afraid not.
OK, we need to go even further back to get a baseline. With the "latest official Wakepup2 version", i.e. the one from Puppy 3.01, do you still get the message:
ASPI CD-ROM Driver of DOS Version4.01...

Host Adapter 0, Target SCSI ID = 4 - LUN 0: SONY CD-RW CRX145S 1.0c CD-ROM driver installed: I host adapter(s), 1 target(s),


Gotta start somewhere...

User avatar
8-bit
Posts: 3406
Joined: Wed 04 Apr 2007, 03:37
Location: Oregon

#20 Post by 8-bit »

I had experimented with even accessing parallel port devices for install of puppy on an old laptop.
First, even though you might have the windows or dos drivers for the parallel port device installed and see the puppy files on the external device, you cannot stop there.
When control passes to the puppy kernel, that windows driver for the device is gone and if driver support for the device is not compiled into the kernel, Puppy no longer sees or is able to access the device.
I went through this when trying to install Deli Linux.
From windows, I could read the parallel port cdrom fine.
But driver support from Deli for it was not loaded so in effect the cdrom did not exist.
So... is there a way to load the drivers for the device immediately after the kernel loads?
Just my thoughts on this.

otropogo

#21 Post by otropogo »

Crash wrote:
Quote wrote:Does the "August 2008" version still work?.....

I'm afraid not.
OK, we need to go even further back to get a baseline. With the "latest official Wakepup2 version", i.e. the one from Puppy 3.01, do you still get the message:
ASPI CD-ROM Driver of DOS Version4.01...

Host Adapter 0, Target SCSI ID = 4 - LUN 0: SONY CD-RW CRX145S 1.0c CD-ROM driver installed: I host adapter(s), 1 target(s),


Gotta start somewhere...
OK. I found the post where I first reported this problem
Posted: Fri Jul 04, 2008 1:46 am
....

Have finally tested the new pcmcia.bat file with and without the new autoexec.bat and config.sys, and am sorry to report the following:

.....
2. with the pcmcia option, the situation is rather worse. Where the old pcmcia.bat loads the ASPI CD-ROM driver for DOS, v. 4.01 and succeeds in identifying the SCSI ID and model of the CDROM drive attached to my pcmcia_scsi host adapter, the new pcmcia.bat doesn't appear to detect anything except an active socket("device 1 found"). And when all of the available SCSI drivers have been tried, reports:

"no scsi controllers found, no scsi logical drives to support".

So I'm afraid there's been a significant step backward.
So there's the problem. It's in the changes you made to the pcmcia.bat file released just prior to my report on July 4. There's no need to go all the way back to the stock Wakepup2.

otropogo

#22 Post by otropogo »

8-bit wrote:... if driver support for the device is not compiled into the kernel, Puppy no longer sees or is able to access the device....
.
There must be pcmcia support in the kernel, since all the versions of Puppy I've tried support pcmcia.

The problem is providing the pcmcia support via a boot floppy. This can be done, because I've used such floppy support to install Suse 9.2 on the CF-25 laptop.

. Suse 9.2 (kernel 2.6.8.) provided a five disk floppy set, and the pcmcia disk enabled the pcmcia port and the installed slimscsi adapter. After that, the installer was able to read the installation CD in the CDROM drive attached to the pcmcia_scsi adapter.

Anyway, the pcmcia_scsi support is not that important. What's essential is the pcmcia support. If we could get that into Wakepup, we could boot from a CF card adapter in the pcmcia slot.

User avatar
Crash
Posts: 453
Joined: Fri 09 Dec 2005, 06:34
Location: Melbourne, FL

#23 Post by Crash »

Otropogo-

You still need to go back to a known good baseline. I can't work backward, only forward.

With the "latest official Wakepup2 version", i.e. the one from Puppy 3.01, do you still get the message:

ASPI CD-ROM Driver of DOS Version4.01...

Host Adapter 0, Target SCSI ID = 4 - LUN 0: SONY CD-RW CRX145S 1.0c CD-ROM driver installed: I host adapter(s), 1 target(s),

?

otropogo

#24 Post by otropogo »

Crash wrote:...
With the "latest official Wakepup2 version", i.e. the one from Puppy 3.01, do you still get the message:

ASPI CD-ROM Driver of DOS Version4.01...

Host Adapter 0, Target SCSI ID = 4 - LUN 0: SONY CD-RW CRX145S 1.0c CD-ROM driver installed: I host adapter(s), 1 target(s),

?
Yes.

Have just written a fresh official Wakepup floppy with Puppy 4.0 and booted from it on the CF-25, with a CDROM drive attached to the slimscsi pcmcia adapter and the Puppy 4 alpha 5 LiveCD in the drive:

Initial Menu - selected #7 PCMCIA

got the above message, followed by

"...file found on idehd, drive C:"

Second Menu: SELECT PUPPY BOOT MODE - selected #6 Boot Media

Third Menu: PUPPY MEDIA LOCATION - selected #5 USB CD
Wakepup reports

"....file found usbcd drive Z:"

Third Menu: selected #1 - Normal

"Searching for files in computer disk drives...."

Fourth Menu - TYPE A NUMBER TO CHOOSE WHICH PERSONAL FILE TO USE:

0 none
1. sda1 /pup_save301R.2fs
2. sda2 /dev_save.bak/pup_save-301Rnopass.2fs

I selected "0 None"

Wakepup reports:

"pup_405.sfs not found. Dropping out to initial-ramdisk console...
/bin/sh: can't access tty; job control turned off
#"

So Wakepup can clearly access the LiveCD in the pcmcia_scsi drive, otherwise, how would it know that the sfs file on it is called pup_405.sfs?

User avatar
Crash
Posts: 453
Joined: Fri 09 Dec 2005, 06:34
Location: Melbourne, FL

#25 Post by Crash »

OK. See if the attached test image works. It is derived from the Wakepup2 3.01 image with with John Doe's new autoexec.bat, config.sys, and todos.bat files inserted.

It does not support Zip Parallel Port, but if it works OK, I should be able to add it back in.

a66f0086cbafb11d4954b6f697255fce wp2test1.img

John Doe
Posts: 1681
Joined: Mon 01 Aug 2005, 04:46
Location: Michigan, US

#26 Post by John Doe »

otropogo, can you tell me what drivers you modprobe to access the ppzip once booted?

also, what device is it recognized as?

I've got a custom initrd.gz/wakepup ready to test. I just want to cut out a step if possible by getting that info first.

otropogo

#27 Post by otropogo »

Crash wrote:OK. See if the attached test image works. It is derived from the Wakepup2 3.01 image with with John Doe's new autoexec.bat, config.sys, and todos.bat files inserted.

Yes, it fails at the same place. But now there's an extra step, after Boot Media, which seems redundant.

In the original, when USBCD is selected in Boot Media, the drive is designated "Z:" immediately, now you get another menu and have to select USBCD again.


Didn't you also replace the corrupted Backpack files? I remember your reporting that they were corrupted.

otropogo

#28 Post by otropogo »

John Doe wrote:otropogo, can you tell me what drivers you modprobe to access the ppzip once booted?
no
John Doe wrote:also, what device is it recognized as?
It's not. I have no idea how to implement parallel ZIP support in Puppy. You'd think that since it claimed boot support for it in Wakepup for several years it would have it. But I don't see it, and I have no idea how to implement it.
John Doe wrote:I've got a custom initrd.gz/wakepup ready to test. I just want to cut out a step if possible by getting that info first.
Wish I could help.

User avatar
Crash
Posts: 453
Joined: Fri 09 Dec 2005, 06:34
Location: Melbourne, FL

#29 Post by Crash »

When you get to the "Select Puppy2 Boot Mode" menu with this version, you should be able to simply select "1 Normal" and continue.
Otropogo wrote wrote:Didn't you also replace the corrupted Backpack files? I remember your reporting that they were corrupted.
This is a clean restart. I'll add them back after this is tamed down a bit.

John Doe
Posts: 1681
Joined: Mon 01 Aug 2005, 04:46
Location: Michigan, US

#30 Post by John Doe »

otropogo wrote:Wish I could help.
Try 'modprobe imm' and see what happens.

Lots of info here:

http://www.faqs.org/docs/Linux-mini/ZIP-Drive.html

otropogo

#31 Post by otropogo »

Crash wrote:When you get to the "Select Puppy2 Boot Mode" menu with this version, you should be able to simply select "1 Normal" and continue.
I booted the test floppy to try this, and now it doesn't recognize my Adaptec host adapter. I tried it a second time and a third time- same thing.

I booted into Windows - no problem, it reads the LiveCD just fine. Puppy 301R mounts the drive fine too...

I suspect Puppy's floppy writes. I've always had problems making wakepups and even formatting floppies in Puppy.

Next time I'll just sneaker the image over to my Win98 machine and use rawrite.

BTW - next time you make a test img for me, could you comment out all those SCSI drivers I don't need? It takes forever to load and try all of them.

It would be nice to have the menu wait for me too. If I don't hit #7 in time on the first menu it defaults to Normal, and I have to reboot.

User avatar
8-bit
Posts: 3406
Joined: Wed 04 Apr 2007, 03:37
Location: Oregon

#32 Post by 8-bit »

I have been following along and still have some questions.
As I understand so far, drivers to detect devices are loaded by FREEDOS.
Upon selection of a boot device to boot Puppy from, your loadin program passes control to the kernel on the device and the kernel is then responsible for detecting devices and loading the drivers built into the kernel for the devices and continuing with the boot process.
If... the kernel does not have native built in support for a device, I think that is the reason that an error is given on not being able to find the remaining support files to finish the boot.
And lets just assume that the drivers loaded while in FREEDOS fail to exist after the kernel is loaded and it is the supported devices that are built into the kernel that in the end dictate the ability to recognize the device you are trying to boot from.
The key here is the driver for the device HAS to be part of the kernel code.
Am I correct in these assuptions?
Remember here that if the kernel does not recognize the device, all other files on it show up as not found.
For test purposes, would stopping after booting the kernel drop one to a command line to see if the kernel indeed has support for the device?

otropogo

#33 Post by otropogo »

8-bit wrote:...
Upon selection of a boot device to boot Puppy from, your loadin program passes control to the kernel on the device and the kernel is then responsible for detecting devices and loading the drivers built into the kernel for the devices and continuing with the boot process.
If... the kernel does not have native built in support for a device, I think that is the reason that an error is given on not being able to find the remaining support files to finish the boot.
If pcmcia support were part of the kernel, why would all the older distros, such as Red Hat, Mandrake, and Suse, need to have special pcmcia boot floppies?

The thing that puzzles me is - how can Wakepup report it can't find xx.sfs while identifying which file it is? If I switch LiveCD versions, it always identifies the one in the drive. Is Freedos passing this on and Wakepup is simply parroting the information? And if so, how does Freedos know to look for a file with an sfs extension?

User avatar
Crash
Posts: 453
Joined: Fri 09 Dec 2005, 06:34
Location: Melbourne, FL

#34 Post by Crash »

Otropogo wrote:I booted the test floppy to try this, and now it doesn't recognize my Adaptec host adapter. I tried it a second time and a third time- same thing.
Have you found out what the problem is? Here are a couple of observations from my experience:

Floppies can be temperamental, but if you have a good one, and a technique that works that you can consistently apply, they are very reliable. There is a lot of debate on this subject on the Forum, and I don't wish to start up the debate again, but once you get a good formula, you can stick to it and get good, consistent results. You may want to do all floppy writes from Win98, and start with a fresh out-of-the-box disk from a name brand manufacturer, which is sometimes easier said than done. I have good results with RawWriteWin, but also have used rawrite successfully from the DOS prompt. I use dd extensively, but it takes some getting used to.

Also, if I find a result to be successful the first time around, but unsuccessful the second, I power down the computer and come back to it later. I may just be superstitious, but I think sometimes the configuration registers in some of the peripherals get into a "can't get out of" state, and only power-down gets things back to a known default condition.

At any rate, the WP2TEST1 image should be pretty bullet proof, and I would like you to convince yourself that it is good before we move on.
8-bit wrote:Upon selection of a boot device to boot Puppy from, your loadin program passes control to the kernel on the device and the kernel is then responsible for detecting devices and loading the drivers built into the kernel for the devices and continuing with the boot process.
If... the kernel does not have native built in support for a device, I think that is the reason that an error is given on not being able to find the remaining support files to finish the boot.
Yes, this is my understanding. So as much as Otropogo may not like it, I define success as when the Kernel executes and knows enough that it can't find rest of Puppy. I realize complete success means having both Wakepup and the Kernel having the necessary hardware support, but I figure if I can get Wakepup to load the Kernel, someone else can address the Kernel issues.
Otropogo wrote:The thing that puzzles me is - how can Wakepup report it can't find xx.sfs while identifying which file it is?
The Kernel is customized to the particular version of Puppy that you are booting. The message comes from the Kernel, not Wakepup. That's why I say if you get that message, Wakepup was successful, even though the entire boot process is not.
Last edited by Crash on Sun 14 Sep 2008, 04:25, edited 1 time in total.

John Doe
Posts: 1681
Joined: Mon 01 Aug 2005, 04:46
Location: Michigan, US

#35 Post by John Doe »

8-bit wrote:As I understand so far, drivers to detect devices are loaded by FREEDOS.
Upon selection of a boot device to boot Puppy from, your loadin program passes control to the kernel on the device and the kernel is then responsible for detecting devices and loading the drivers built into the kernel for the devices and continuing with the boot process.
you got it.
8-bit wrote:If... the kernel does not have native built in support for a device, I think that is the reason that an error is given on not being able to find the remaining support files to finish the boot.
And lets just assume that the drivers loaded while in FREEDOS fail to exist after the kernel is loaded and it is the supported devices that are built into the kernel that in the end dictate the ability to recognize the device you are trying to boot from.
correct again. furthermore, the drivers from freedos are not available for the kernel (you assumed correctly).
8-bit wrote:Am I correct in these assuptions?
YES!!
8-bit wrote:For test purposes, would stopping after booting the kernel drop one to a command line to see if the kernel indeed has support for the device?
That would be one way to do it. It's not so much what's compiled in the kernel itself but rather the problem is more the modules that are included in the initr.gz image and whether or not initrd calls them. It's a pretty involved process but you seem to have gotten the gist of it.

otropogo's Parallel zipdrive is not going to boot without a modifided initrd.gz at this point (4.0.7).

I tried to post one the other night but the forum told me is was too big, although the wakepup image is bigger and it worked. That's the brakes I guess.

Post Reply