(old)Puppy Linux Discussion Forum

Check the description of graphics under UEFI in this Wikipedia article. They have made a deliberate move away from VGA dependence. The UEFI BIOS screens I have seen so far have all been graphical. We are dealing with a new, incompatible standard.

Thanks, Prehistoric for the wiki article. I have read that article before, but missed something.

Under the services section, there is the following:

The operating system is permitted to directly write to the framebuffer provided by GOP during runtime mode. However, the ability to change video modes is lost after transitioning to runtime services mode until the OS graphics driver is loaded.

Which sounds like the problem, I and others have encountered. If there is a transition to console text mode, the ability to go back to another graphical mode with the run time services is not ensured. By probing the hardware first with coldplug, the graphical drivers (kernel frame buffer and other driver) are loaded. The kernel is not using the run time services of the UEFI BIOS.

The above is my guess. What do others think of it? I am married with children (grown), so I am use to being told I do not know what I am talking about.

Greg,

I suspect something is not initializing fast enough, or maybe more likely there's a bug in the radeon KMS driver. We'll give the 3.8.3 kernel a go when it comes out in a few days, no reason to think it will help, but there's a chance.

jamesbond wrote:
JustGreg, just saw your comment about the snow - I thought it no longer snows in March? Shocked

It sure does. But it's the snow in April that is really depressing.

What's snow?

kirk wrote:What's snow?

Ask any of the pasty-faced people showing up in your state this week.

Kirk, I hope you are correct on the new kernel. But, I do know that the coldplug option does resolve the issue. When the next kernel comes out, I will test it without coldplug option to see if it takes care of the issue.

I have tried this on my computer and it works great. None of the problems I had with 610/611.

One question though: I tried the VirtualBox .sfs. It doesn't work and gives some weird errors related to the VirtualBox kernel. I supposed it's meant to work only on 610(?)

Edit: this problem seemingly is linked, somehow, to DNS/WINS registration. One of the Live PCs has magically appeared with its name and workgroup names showing as well as a MASTER name (workgroup line) as well. Its information fields are filled when viewed from on all 3 PCs, while the other 2 PCs fields remain with empty fields no matter where the smbclient's query is run for them..

This thread's community might have an explanation to this behavior of a Live media FATDOG620 system. It occurs whether or not a save session has been done prior to reboot.

Workgroup and Server names are missing upon desktop arrival.

Assistance in understanding is requested.

I have downloaded the libreoffice sfs from control panel but am unable to get menu entries when it's loaded?

also if save file placed in fatdog folder it isn't seen?

title Fatdog64 (sda1/fatdog64620b2)
uuid 47e1e170-aa07-4dbf-a968-aacc20a806b0
kernel /fatdog64620b2/vmlinuz psubdir=fatdog64620b2 pmedia=atahd pfix=fsck
initrd /fatdog64620b2/initrd

Workgroup and Server names are missing upon desktop arrival.

Don't know, works for me:

I have downloaded the libreoffice sfs from control panel but am unable to get menu entries when it's loaded?

Might have to restart X to have those show up. Lxpanel looks for a change in /usr/share/applications before it refreshes it's menu. But when a sfs is loaded none of those files have a new time stamp. Maybe that's something we can fix.

also if save file placed in fatdog folder it isn't seen?

Fatdog64 only searches the root filesystem of each partition for a file named fd64save.ext4. Anything else will have to be specified on the kernel line. From the Boot options FAQ page:

psubdir=fatdog64620b2 pmedia=atahd pfix=fsck

This kernel line is for puppy, not fatdog64. Fatdog64 does not use the same boot parameters as Puppy. Really everything is different.

Might have to restart X to have those show up. Lxpanel looks for a change in /usr/share/applications before it refreshes it's menu. But when a sfs is loaded none of those files have a new time stamp. Maybe that's something we can fix.

I tried an X restart and also a reboot without success, tried both libre versions from ppm too

Fatdog64 only searches the root filesystem of each partition for a file named fd64save.ext4. Anything else will have to be specified on the kernel line. From the Boot options FAQ page:

thanks - will go and play, sorry for the erroneous grub details, showing my ignorance.....!

What about this new patch they send out automatically to Win8?
http://www.murga-linux.com/puppy/viewto ... 618#691618

Microsoft's latest patches squash potential USB hijack

If one have Fatdog on USB would not the UEFI see it as some alien
trying to exploit the computer unless one have shut of the protected boot?

Nooby,

The simple solution to this problem is to remove Window$ 8. There is no value added to using it. One needs to read the End User License Agreement (EULA) in full to see this. Everyone except Micro$oft has the risk and Micro$oft is not responsible if things go wrong.

With Fatdog64 UEFI version, if one has registered the Fatdog64 Grub certificate with the Machine Owner Keys (MOK) then UEFI will recognize the device as being valid and allow it to boot.

Does this patch eliminate the ability to boot from a USB stick? This would be an issue. The elimination of USB Boot also would destroy the Window$ 8 ability to restore from a USB Flash drive. Restore from a USB flash drive is a standard option.

In my case, I eliminated Widow$ 8, so it is not really an issue for me.

JustGreg wrote:In my case, I eliminated Widow$ 8, so it is not really an issue for me.

Does that mean that UEFI is not in firmware? Do you have to change your BIOS at all to revert to a conventional boot?

rcrsn51,

Does that mean that UEFI is not in firmware? Do you have to change your BIOS at all to revert to a conventional boot?

UEFI is in the BIOS firmware. I eliminated Window$ 8 by changing the hard disk to a new solid state one. I had already registered Fatdog64 Grub in the MOK. The removal of the hard drive with Window$ 8 had no effect on the UEFI BIOS and its contents (i.e. MOK). I still could boot a flash drive with the blank new drive

To do the install on the new drive, I booted a USB flash drive with Fatdog64 UEFI test version from December. All of the next steps were done with US system running in RAM. I created GPT device table with a VFAT partition for EFI components, and three other Linux partitions. Based on my testing, I suggest using ext4 Linux partitions. I recorded the UUID for each partition.

I copied the needed EFI components to the VFAT partition. I copied the needed Fatdog64 system files to the first Linux partition for a frugal installation. I modified the grub.cfg configuration file (on drive VFAT partition) to set root and boot the new system using the partition UUID values. During the reboot, I created a system save file on the first Linux partition.

The above process is a manual frugal installation.

After reboot and removing the USB flash drive, the new hard drive installation started correctly using the save file. I have been using it since for all the testing and experiments.

Please note that I did not have to change anything with the UEFI since my first boot of a CDROM that registered the Fatdog64 certificate with MOK.

UEFI BIOS is just a new type of BIOS. There is a learning curve for it, but, it is not terrible. It is not tied to Window$ 8. Window$ 8 has UEFI tools, (good luck in finding out how to use them) to manage the UEFI boot process. But, there are Linux tools that do the same.

Just like the spring snow, the FUD of UEFI will disappear in time.

For the most part, UEFI booting, even with secure boot, is not a problem. And dual booting with Windows 8 isn't too bad ether. There are some UEFI implementations that are hard to use. There are some secure boot implementations that are just broke and you can't add keys, but I think they are few. You can turn off secure boot and Windows 8 will boot normally. Most, maybe all, of the UEFI computers I've looked at also have a legacy boot option that enables BIOS/MBR type booting too. For more information:

http://distro.ibiblio.org/fatdog/web/fa ... -boot.html

and

http://distro.ibiblio.org/fatdog/web/fa ... drive.html

Please note that 620beta2's kernel isn't signed so don't try the uefi-harddrive (Windows 8 dual boot). USB/CD booting is not effected. For the next release I'll try to remember to sign it

I usually break everything I touch so I rather postpone
the whole thing until it is a must to buy a new computer.

I will instead buy hard disks to my old computers
until all of them break down

Exception would be some ARM stick like the PQ I-Stick
that has 2GB DRAM and are already rooted so one
can install Linux on it. That could maybe sway be
to buy anything like a fully functional computer ever
so first only HD and USB Flash and boot from them on
my old computers.

Thanks Kirk for responding.

Your smbclient doesn't show a "Master". Are you using a router or direct to internet link. Further, your smbclient should show others in the Server list as being detected should there be a Master. This would allow resource mounting by hostname vs IP-addresses. It may appear that your network has same issue with LAN registration as the several I looked at today. At present, this appears to be related to DNS/WINS registration not to mention Master-Browser election.

Edit: There actually may be a FATDOG solution to this (because of its integration of SAMBA product or via DNSMASQ).

P.S. Please add "Testparm" (SAMBA4) back into the distro packaging when you build the next interation. Thx.

ally wrote:

I have downloaded the libreoffice sfs from control panel but am unable to get menu entries when it's loaded?

If you delete the file in ~/.cache/menus, then restart X, lxpanel will recreate a new menu, and libreoffice entries will show up.

If you delete the file in ~/.cache/menus, then restart X, lxpanel will recreate a new menu, and libreoffice entries will show up.

hey will

nice one, all sorted

thanks

Hi ally,

I reported this as a bug at LXDE on Source Forge. Not sure if I stated the case there properly.
https://sourceforge.net/tracker/?func=d ... tid=894869