Grub2config replaces Grub4Dos

[sinc]

Thanks so much d-pupp!!! That did it!

[shinobar]

Thanks so much d-pupp!!! That did it!

Please show us your entry lines.
Mine is like this:
menuentry 'Dpupbw64 10.0.7 RAM mode(sda3/dpupbw64-10.0.7)' --class puppylinux{
search --no-floppy --set=root --fs-uuid 0c17217e-4723-4d30-a4bf-3da1636875b8
linux /dpupbw64-10.0.7/vmlinuz psubdir=/dpupbw64-10.0.7 pmedia=ataflash pfix=fsck
if [ -e /dpupbw64-10.0.7/ucode.cpio ]; then
initrd /dpupbw64-10.0.7/ucode.cpio /dpupbw64-10.0.7/initrd.gz
else
initrd /dpupbw64-10.0.7/initrd.gz
fi
}

the line
search --no-floppy --set=root --fs-uuid UUID_OF_THE_PARTISION
does't work?

[sinc]

All I added were these two statements:
pupsfs=1E74E51174E4EC89 psave=1E74E51174E4EC89

It did not work prior to this. I did not try without the psave command. Maybe this one is not necessary.

menuentry 'BWPup Linux' --class puppylinux{
search --no-floppy --set=root --fs-uuid 1E74E51174E4EC89
linux /BWPup/vmlinuz pupsfs=1E74E51174E4EC89 psave=1E74E51174E4EC89 psubdir=/BWPup pmedia=atahd pfix=fsck
if [ -e /BWPup/ucode.cpio ]; then
initrd /BWPup/ucode.cpio /BWPup/initrd.gz
else
initrd /BWPup/initrd.gz
fi
}

menuentry 'Windows uefi boot' --class puppylinux{
search --no-floppy --set=root --file /EFI/Microsoft/Boot/bootmgfw.efi
chainloader /EFI/Microsoft/Boot/bootmgfw.efi
}

if [ $grub_platform = "efi" ] ; then
menuentry 'UEFI Firmware Settings' --class puppylinux{
fwsetup
}
fi

[williwaw]

All I added were these two statements:
pupsfs=1E74E51174E4EC89psave=1E74E51174E4EC89

It did not work prior to this. I did not try without the psave command. Maybe this one is not necessary.

menuentry 'BWPup Linux' --class puppylinux{
search --no-floppy --set=root --fs-uuid 1E74E51174E4EC89

sinc, as a test, try putting your full uuid in the search --no-floppy --set=root --fs-uuid 1E74E51174E4EC89 statement and without the recently added pupsfs=1E74E51174E4EC89 in the kernel line
the command blkid should show a longer number similar to shinobars example above

[sinc]

This is all of it.

[rockedge]

Seems like the partition being formatted as NTFS is what is requiring the extra parameter(s) pupsfs=1E74E51174E4EC89 psave=1E74E51174E4EC89 being needed.

[shinobar]

Seems like the partition being formatted as NTFS is what is requiring the extra parameter(s) pupsfs=1E74E51174E4EC89 psave=1E74E51174E4EC89 being needed.

I guess this is the BookwormPup specific issue.
Fossapup64 has no problem, but the BookwormPup fails. Both on the same ntfs partition, sda2.
probepart shows:
/dev/sda1|ntfs|50M
/dev/sda2|ntfs|39.3G
/dev/sda3|ntfs|564M
/dev/sda4|ext4|10.0G
/dev/sdb1|vfat|1.2G
/dev/sdb2|ext3|6.0G
/dev/sr0|none|1023M

but the script init of the BookwormPup omits ntfs partitions.
HAVE_PART shows:
sda4|ext4
sdb1|vfat
sdb2|ext3

(grub.cfg entries)
menuentry 'Dpupbw64 10.0.7 (sda2/dpupbw64-10.0.7)' --class puppylinux{
search --no-floppy --set=root --fs-uuid FABC3FAEBC3F6473
linux /dpupbw64-10.0.7/vmlinuz psubdir=/dpupbw64-10.0.7 pmedia=atahd pfix=fsck
if [ -e /dpupbw64-10.0.7/ucode.cpio ]; then
initrd /dpupbw64-10.0.7/ucode.cpio /dpupbw64-10.0.7/initrd.gz
else
initrd /dpupbw64-10.0.7/initrd.gz
fi
}

menuentry 'Fossapup64 9.5 (sda2/fossapup649.5frugal)' --class puppylinux{
search --no-floppy --set=root --fs-uuid FABC3FAEBC3F6473
linux /fossapup649.5frugal/vmlinuz psubdir=/fossapup649.5frugal pmedia=atahd pfix=fsck
if [ -e /fossapup649.5frugal/ucode.cpio ]; then
initrd /fossapup649.5frugal/ucode.cpio /fossapup649.5frugal/initrd.gz
else
initrd /fossapup649.5frugal/initrd.gz
fi
}

[rockedge]

@shinobar Could you download the file that contains updated versions of probedisk and probepart and test these out in your BookwormPup system?

The BookwormPup init script will need a look.

[shinobar]

@shinobar Could you download the file that contains updated versions of probedisk and probepart and test these out in your BookwormPup system?

The BookwormPup init script will need a look.

Tnx. just the same as i posted before
probepart shows:
/dev/sda1|ntfs|50M
/dev/sda2|ntfs|39.3G
/dev/sda3|ntfs|564M
/dev/sda4|ext4|10.0G
/dev/sdb1|vfat|1.2G
/dev/sdb2|ext3|6.0G
/dev/sr0|none|1023M

# probedisk
/dev/sda|drive|ATA VBOX HARDDISK
/dev/sdb|usbdrv|UFD_3.0 Silicon-Power8G
/dev/sr0|optical|VBOX CD-ROM [Rev 1.0]

[shinobar]

I guess this is the BookwormPup specific issue.
Fossapup64 has no problem, but the BookwormPup fails. Both on the same ntfs partition, sda2.

Anyway, pdrv option solves.
(grub.cfg entry)
menuentry 'Dpupbw64 10.0.7 (sda2/dpupbw64-10.0.7)' --class puppylinux{
search --no-floppy --set=root --fs-uuid FABC3FAEBC3F6473
linux /dpupbw64-10.0.7/vmlinuz psubdir=/dpupbw64-10.0.7 pdrv=FABC3FAEBC3F6473 pmedia=atahd pfix=fsck
if [ -e /dpupbw64-10.0.7/ucode.cpio ]; then
initrd /dpupbw64-10.0.7/ucode.cpio /dpupbw64-10.0.7/initrd.gz
else
initrd /dpupbw64-10.0.7/initrd.gz
fi
}

[shinobar]

Major update:
grub2config-2.1
http://shinobar.server-on.net/puppy/opt ... ig-2.1.pet

#2024-09-15 v2.1: mbr stuff ext4 support no need wee nor grub4dos, pdrv option, check target space, refresh_drive_icons,

[shinobar]

grub2config-2.1 produces incorrect entry for the RAM mode. Please manually edit the file 'MOUNTPOINT/grub.cfg', where the 'MOUNTPOINT' is the mount point of the boot partition, mayby the first partition of the boot drive.

Correct 'pfix=fsck' to 'pfix=ram'

example:
menuentry 'Dpupbw64 10.0.7 RAM mode(sda3/dpupbw64-10.0.7)' --class puppylinux{
search --no-floppy --set=root --fs-uuid 0c17217e-4723-4d30-a4bf-3da1636875b8
linux /dpupbw64-10.0.7/vmlinuz psubdir=/dpupbw64-10.0.7 pmedia=ataflash pfix=fsck
if [ -e /dpupbw64-10.0.7/ucode.cpio ]; then
initrd /dpupbw64-10.0.7/ucode.cpio /dpupbw64-10.0.7/initrd.gz
else
initrd /dpupbw64-10.0.7/initrd.gz
fi
}

correct to:
menuentry 'Dpupbw64 10.0.7 RAM mode(sda3/dpupbw64-10.0.7)' --class puppylinux{
search --no-floppy --set=root --fs-uuid 0c17217e-4723-4d30-a4bf-3da1636875b8
linux /dpupbw64-10.0.7/vmlinuz psubdir=/dpupbw64-10.0.7 pmedia=ataflash pfix=ram
if [ -e /dpupbw64-10.0.7/ucode.cpio ]; then
initrd /dpupbw64-10.0.7/ucode.cpio /dpupbw64-10.0.7/initrd.gz
else
initrd /dpupbw64-10.0.7/initrd.gz
fi
}

[radky]

Seems like the partition being formatted as NTFS is what is requiring the extra parameter(s) pupsfs=1E74E51174E4EC89 psave=1E74E51174E4EC89 being needed.

I guess this is the BookwormPup specific issue.
Fossapup64 has no problem, but the BookwormPup fails. Both on the same ntfs partition, sda2.
probepart shows:
/dev/sda1|ntfs|50M
/dev/sda2|ntfs|39.3G
/dev/sda3|ntfs|564M
/dev/sda4|ext4|10.0G
/dev/sdb1|vfat|1.2G
/dev/sdb2|ext3|6.0G
/dev/sr0|none|1023M

but the script init of the BookwormPup omits ntfs partitions.
HAVE_PART shows:
sda4|ext4
sdb1|vfat
sdb2|ext3

(grub.cfg entries)
menuentry 'Dpupbw64 10.0.7 (sda2/dpupbw64-10.0.7)' --class puppylinux{
search --no-floppy --set=root --fs-uuid FABC3FAEBC3F6473
linux /dpupbw64-10.0.7/vmlinuz psubdir=/dpupbw64-10.0.7 pmedia=atahd pfix=fsck
if [ -e /dpupbw64-10.0.7/ucode.cpio ]; then
initrd /dpupbw64-10.0.7/ucode.cpio /dpupbw64-10.0.7/initrd.gz
else
initrd /dpupbw64-10.0.7/initrd.gz
fi
}

menuentry 'Fossapup64 9.5 (sda2/fossapup649.5frugal)' --class puppylinux{
search --no-floppy --set=root --fs-uuid FABC3FAEBC3F6473
linux /fossapup649.5frugal/vmlinuz psubdir=/fossapup649.5frugal pmedia=atahd pfix=fsck
if [ -e /fossapup649.5frugal/ucode.cpio ]; then
initrd /fossapup649.5frugal/ucode.cpio /fossapup649.5frugal/initrd.gz
else
initrd /fossapup649.5frugal/initrd.gz
fi
}

@shinobar , @dimkr

It seems the failed init detection of ntfs partitions may be due to the kernel definition of filesystems in /proc/filesystems.

If I understand correctly, the modern init of BookwormPup64 defines builtin filesystem support (KFILESYSTEMS) as follows:

Code: Select all

KFILESYSTEMS="$(cat /proc/filesystems | grep -v "nodev" | grep -v "fuseblk" | sed -e "s#\t##g" | tr '\n' '|' | sed -e 's#|$##g')

which returns the following for my system which has vfat and ntfs partitions:

Code: Select all

KFILESYSTEMS="ext3|ext2|ext4|squashfs|vfat|msdos|exfat|iso9660|ntfs3|udf|xfs|f2fs|btrfs#"

NOTE: the above report specifies support for 'ntfs3' filesystems

---

On my system, probepart -hr returns the following:

/dev/sda1|vfat|260M
/dev/sda2|none|16M
/dev/sda3|ntfs|117.9G
/dev/sda4|ntfs|1000M
/dev/sdb1|vfat|28.6G
/dev/sdc1|vfat|28.9G

However, the init (HAVE_PARTS) searches specifically for filesystems defined by KFILESYSTEMS.

Consequently, on my system the search for 'ntfs3' fails and the 'ntfs' partitions are not detected:

sda1|vfat
sdb1|vfat
sdc1|vfat

Maybe the init (HAVE_PARTS) could be adjusted to correctly detect and define ntfs3 as ntfs.

[dimkr]

the modern init of BookwormPup64

Nobody is maintaining it (see https://github.com/puppylinux-woof-CE/w ... nitrd/init for evidence) and I can't say I'm surprised to hear about more issues that crop up as this messy piece of code ages.

However, the init (HAVE_PARTS) searches specifically for filesystems defined by KFILESYSTEMS.

Which makes absolutely no sense.

I simplified the init script a lot in https://github.com/vanilla-dpup/woof-CE, to speed it up, make it more reliable and drop assumptions that used to be true back in the days when people actually maintained woof-CE but now break things (like the ntfs vs ntfs3 thing).

KFILESYSTEMS can be removed very easily, with zero side effects.

[ozsouth]

I still use ntfs-3g in my kernels, as I wait for ntfs3 to mature. Other than some slow transfers, seems trouble-free - I have my main storage on it.

[peebee]

Configuration problem when used in Woof-CE builds.......

All items in /usr/share/docs get put into the optional docx.sfs rather than the main puppy.sfs

This means the gnibble2.png icon is not present in installs without docx.sfs - and maybe other items....

Can the real icon be moved to /usr/share/pixmaps rather than a link? Do other items need to move?