Random Failure on closing?
First screen shows a read error on the kernel load, so that would mean something went wrong with the writing of the image? Are you burning it to a CD or to a USB? So first thing I could try would to be to copy to another media and see if it will boot without that read error. Is your system 32 bit or 64 bit? Think it must be 64 bit, since kernels moved to 64bit when Fedora stopped releasing 32 bit version. I still build the iso version that uses syslinux that hasn't been updated in a long long time....
The difference in kernel size between bz6x7.8 and the added 6.7.9 is just about 93K of the 12M total size in difference. So not much there, and the lsblk was a small addition to ramdisk.lzma size. Just need to see what lines were change in config file, and add them to the bz6x8 .config. Generally keep the latest stable kernel, and the rc kernel, but 6.8 just went to released status. Usually wait for next rc to come out before changing default. Glad it works.
went ahead and make a new 31alpha and new g4lefi that have that added.
Made a new alpha30 and new g4lefi.img had added the eMMC set to y, and had hoped the make oldconfig would have set any required options, but it appears it didn't. So, set the MMC-y, and then make oldconfig prompted for a number of other things. Also, added lsblk which just required the program since all linked libraries were already thers. Need to download new files. ftp://setzco.dyndns.org/g4l-v0.66alpha30-fc39.iso ftp://setzco..dyndns.org/g4lefi-0.66.img The differences between .config is 2872,2873d2871...
ftp://setzco.dyndns.org/g4l-v0.66alpha29-fc39.iso ftp://setzco..dyndns.org/g4lefi-0.66.img Have updated bz6x7.9 kernel to add option CONFIG_PHY_INTEL_LGM_EMMC=y and rebuilt that kernel. Looked thru menuconfig, but couldn't find option in make menuconfig, so just activated it in .config Did make, and updated the above files, so both the syslinux iso and the grub4dos image files. Don't have the EMMC so having tested. Currently in Nevada taking care of mom from hospital, so remotely accessing machines...
Uploaded Version 0.65 Fedora 38
I've created updated bz6x4.15 and bz6x5.2 kernels that have it enabled. Have built a new version 0.65 as 8th so far that includes this. Send me an email to mikes@guam.net and I can send you ftp link, or can send files. The syslinux version iso, or usb efi/non-efi image or development source. Or if you just want kernel(s) files. Was just changing option in .config file and rebuilding the kernels. Added about 18K to to kernel size. Usually add new ethernet kernels, but it was in a different section....
On 8 Sep 2023 at 8:49, Mikhail wrote: To: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net From: "Mikhail" mtatarin@users.sourceforge.net Send reply to: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net Subject: [g4l:discussion] vmxnet3 ethernet driver support Date sent: Fri, 08 Sep 2023 08:49:07 -0000 Hi, Could you please create a new topic: vmxnet3 driver support The question is how to add vmxnet3 support to g4l? Any plans to add vmxnet3 support in next future g4l releases?...
Uploaded Version 0.64 working on 0.65
On 18 Jun 2023 at 23:13, KenUnix wrote: To: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net From: "KenUnix" kenmartin@users.sourceforge.net Send reply to: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net Subject: [g4l:discussion] Imageing from 931.52 SSD to 931.51 SSD possible? Date sent: Sun, 18 Jun 2023 23:13:16 -0000 Michael, Thanks for the reply. I thought I had good quality SSD's Western Digital Green 1.tb then to discover some are 931.51 and some are 931.52! Clonezilla...
On 18 Jun 2023 at 18:39, KenUnix wrote: To: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net From: "KenUnix" kenmartin@users.sourceforge.net Send reply to: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net Subject: [g4l:discussion] Imageing from 931.52 SSD to 931.51 SSD possible? Date sent: Sun, 18 Jun 2023 18:39:04 -0000 I have two laptops with WD Green 931.52GB SSD's. I wanted to make clones of them so I purchased two more WD Green 1TB SSD's. However those have a size...
On 7 Jun 2023 at 15:25, Russell Hanneken wrote: To: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net From: "Russell Hanneken" russellhanneken@users.sourceforge.net Send reply to: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net Subject: [g4l:discussion] Re: Can't boot 0.63 Date sent: Wed, 07 Jun 2023 15:25:28 -0000 Apparently the current version of Pop!_OS uses systemd-boot rather than grub. https://superuser.com/questions/1583591/systemd-permanently-add-boot-loader-options-popos...
On 6 Jun 2023 at 22:26, Russell Hanneken wrote: To: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net From: "Russell Hanneken" russellhanneken@users.sourceforge.net Send reply to: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net Subject: [g4l:discussion] Can't boot 0.63 Date sent: Tue, 06 Jun 2023 22:26:29 -0000 I just realized that my laptop has something called coreboot. From the GitHub page: coreboot is a Free Software project aimed at replacing the proprietary firmware...
On 6 Jun 2023 at 22:26, Russell Hanneken wrote: To: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net From: "Russell Hanneken" russellhanneken@users.sourceforge.net Send reply to: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net Subject: [g4l:discussion] Can't boot 0.63 Date sent: Tue, 06 Jun 2023 22:26:29 -0000 I just realized that my laptop has something called coreboot. From the GitHub page: coreboot is a Free Software project aimed at replacing the proprietary firmware...
On 6 Jun 2023 at 20:54, Russell Hanneken wrote: To: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net From: "Russell Hanneken" russellhanneken@users.sourceforge.net Send reply to: "[g4l:discussion] " 408763@discussion.g4l.p.re.sourceforge.net Subject: [g4l:discussion] Can't boot 0.63 Date sent: Tue, 06 Jun 2023 20:54:35 -0000 I'm attempting to boot Ghost for Linux 0.63 on my System76 Gazelle laptop, and each time I try I just get this message: SecureBoot is disabled. Booting from 'SanDisk...
Has 6.3.5 as default kernel and updated files. Also has memtest86+ 6.20 latest grub4dos with legacy and UEFI support. Doesn't support secure boot So must turn off secure boot before booting.
Uploaded version 0.63
Thanks for al the info. Search had mentioned something about there error might cause run time error that the compiler wouldn't catch with the Cstyle casting? That's why I was looking for solution. Seems the sort would require complete redo of the C style arrays that are created elsewhere in program. Example seems to require more than what is needed with qsort or the simple old shell sort. Interesting info. thanks again.
Tried some things, and found a solution, but is strange. sort2 version has cppcheck --enable all report the earlier style message. sort3 and sort4 don't give error, but strangely all three compiles produce identical binary files after strip?? #include <cstdio> #include <cstdlib> #include <cstring> using namespace std; int cmpstr(void const a, void const b) { // from sort2.cpp // char const aa = (char const )a; // char const bb = (char const )b; // return strcmp(aa, bb); // from sort3.cpp // char...
Tried some things, and found a solution, but is strange. sort2 version has cppcheck --enable all report the earlier style message. sort3 and sort4 don't give error, but strangely all three compiles produce identical binary files after strip?? #include <cstdio> #include <cstdlib> #include <cstring> using namespace std; int cmpstr(void const a, void const b) { // from sort2.cpp // char const aa = (char const )a; // char const bb = (char const )b; // return strcmp(aa, bb); // from sort3.cpp // char...
I must be missing something. I've look at a number of links that say they work, but compiler gives a bunch of error lines?? The old style shell sort works perfectly? Tried a number of things with various different cmp setups, but none compiled. Have fedora 37 with latest compiler. real programs have arrays of 300 to 700 strings, but did a simple example. #include <cstdio></cstdio> include <cstdlib></cstdlib> include <cstring></cstring> include <algorithm></algorithm> using namespace std; bool cmp(const...
Not clear on first one, since what I've found seems sort is for numbers? Found something with second one. Create int cmpstr(void const a, void const b) { char const aa = (char const )a; char const bb = (char const )b; return strcmp(aa, bb); } and change strcmp option to just cmpstr, but it then gives this output. style: C-style pointer casting [cstyleCast] char const aa = (char const )a; ^ style: C-style pointer casting [cstyleCast] char const bb = (char const )b; ^ So, changes style issue on each...
Program has 4 qsort options that give this. fixzzallbz.cpp:557:42: style: C-style pointer casting [cstyleCast] qsort(datax, num2, sizeof(datax[0]), (int ()(const void , const void *))strcmp); Usually find it gives a do you mean option.. Have never had a problem, but if there is a more correct way to do it, would like to update it. Thanks
Update: Was posting to the grub4dos list about image info, and the author had made a custom version for me to test that resolved the issue with the acer predator so it correctly displays the image and the text with my latest test version. Not sure if it was just an issue with the Predators UEFI setup, or was something in the UEFI for some systems.
Have been test the newer grub4dos legacy and UEFI boot options, and in virtual box and dell machine they worked fine. Did recently find on an Acer Predator notebook with the background image, the menu options got messed up completely. But without image works fine. I've also gotten the new memtest 6.10 to work on syslinux version and both grub4dos versions. Doesn't work if secure boot, unless one make addition to secure boot setting. Also note on the Acer Predator if boot option is other than AHCI...
Was sent tightvnc-2.8.77-gpl-setup-64bit.msi to test, and it fixes the issue with dropping connection if resize is requested. This doesn't resize, but not drop connection. Not listed as downloadable but hopefully will be. Sent report on my testing and full log file of connection.
Had the counter to this were tiger vncviewer would drop connection to new tightvnc server. Issue was to drop if resize failed. Did get sent an test version tightvnc-2.8.77-gpl-setup-64bit.msi and installing it fixed that issue. Isn't listed on download site yet. But perhaps soon, or even newer version.
I've found out more info? On one machine connection, found if I did vncviewer -Log *:stderr:100 -Desktopsize 1920x1080 192.168.24.3 So, that it had same size as remote screen rather than trying to resize to 1600x845 of my notebook it would connect and not crash / drop back to command prompt But just tried on another machine that has two monitors and it is showing size as 3520x1080 on remote, but that drops to command prompt. vncviewer -Log *:stderr:100 -Desktopsize 3520x1080 192.168.24.10 But vncviewer...
tightvnc server 2.8.75 will not connect with tigervnc viewer
tightvnc server 2.8.75 will not connect with tigervnc viewer
Uploaded version 0.62 and working on 0.63
Forgot to add readme file to the ISO Zip before uploading. Just added Readme for 0.62 to files directory. G4L version 0.62 has some major changes in build, but operation is mostly the same. Build Issue This Zip contains 2 ISO files First is similar to the previous ISO files uses syslinux and can be written to CD or Flash 50331648 Jan 1 03:57 g4l-v0.62.iso Second is a new setup using grub4dos as boot loader and can be written to Flash. Has both the standard USB option and supports UEFI boot as well....
The https://sourceforge.net/projects/g4l/files/g4l%20ISO%20images/g4l-v0.62-Release-ISOs.zip/download Is a zip file that contains two files. Archive: g4l-v0.62-Release-ISOs.zip Length Date Time Name --------- ---------- ----- ---- 67108864 01-01-2023 04:04 g4lefi-0.62.iso 50331648 01-01-2023 20:39 g4l-v0.62.iso --------- ------- 117440512 2 files The g4l-v0.62.iso uses syslinux to boot and is created as a hybrid version, so one can use an ISO burning program to write it to a CD or one can copy it...
Uploaded 0.61 release
A couple questions. 1. How are you trying to boot it? The g4lefi.iso is not a normal iso that is meant to be burned to a cd. It is generally copied directly to a flash with grub4dos being the boot loader. 2. Have you checked BIOS settings. I recently had a linux system that had the onboard battery die, and the BIOS reset back to default. Replace battery, but machine would not boot from hard disk. Turned out there was a strange BIOS setting that was defaulted to boot using Windows EFI, and for whatever...
The first one I would expect, since the G4L efi image does not have an Official EFI signature since that is a very expensive process as I understand it. One would have to make a signature file and manually install it in each machine to allow for secure boot to work? Don't understand what the issue is with the second? Should display the EFI grub4dos menu listing that is a basic 800x600 screen..
in cccc_tbl.cc has to add this-> to line 99? this->erase(value_iterator); The other issues were warnings about Register variables not allowed.. But those didn't have a problem compiling, and it installed. Was just running a program thru cbuilder, and was just trying the things out.
Had to look. It is under the Utilities options.
I've done similar on Linux. Used lzop to extract the compressed data image to a file that is the original size of partition. Then used the linux loop mount option? Would have to look up the process since it has been a long time. But generally just restoring to a spare disk is faster.
G4L primarily uses dd for most of the backup processes. Unfortunately, dd does not handle hard errors within disks/partitions. There are options to use ddrescue, which does have a number of methods to attempt to copy the readable data, and does try to recopy the bad areas a number of times before leaving those blocks as what it was able to get. May or may not make a useful image. You didn't mention what file systems are on disk. If it is windows, you might be able to use ntfsclone option to backup...
Agree that it isn't an easy process and it is probable more work than restoring to a spare disk. Recall I once needed a file from a smaller partition I had imaged, and didn't have a spare disk at time were I was, so tried it. Probable more just to see if it would work, and it did. So. Definitely agree that quickest option would be to just restore to a spare disk, and then search it. On issue that might be of concern. If doing this on same system with original disk, the bllkids on disk would be identicle...
That is probable mostly correct with one complex exception. It is possible to first uncompress the image file to the raw state. Then you can mount that file as a loop process. Have done it a few times, but uncompressed size is size of the original disk/partition so needs lots of space. The loop mounting process is also something I'd have to look up, since it depends if it is a partition or disk image, and OS might make difference. Usually easier to just restore to an external disk. So, mostly Not...
Did you note or can you check what was showing in cat /proc/partitions If it has something I can identify, could have it show a message to users that they might need to disable raid to be able to backup partitions? Second. Do you know what raid controller it is showing? Might be that the controller for some machines is included, but this one isn't, or it might be that the raid controller in hardware or software removes the standard drive devices if it sees them as raid? There is also a line that...
A couple things come to mind, but need more info. First, what type of drive is it? Regular hard disk or some other type? Second, on the OS, how does it report the drive as being. Third, might the drive require special firmware files. I don't include a lot of firmware files, just ones that users have mentioned. It might be some modual in the kernel build that isn't included at this point. Generally, they to add most, but occasionally it lists things that are set to default n to add. After booting...
Not a problem. I have the "G4L" sourceforge project and have the same issues with the emails not working with the addresses? But then I get spam messages that are sent to my sourceforge address?? So, guess the spammers have firgured out how to make it work?? Have a great day.
Got an email about the fix, and tested it and it works fine. Replied to email, but just got it returned with user doesn't exist?? The Postfix program giorgiotani@users.sourceforge.net: host mx.sourceforge.net[216.105.38.6] said: 550 unknown user (in reply to RCPT TO command) The body of message - Sent 11/14 Runs fine now.. Thanks. rpm -Uvh peazip-8.3.0.LINUX.GTK2-1.x86_64.rpm Verifying... ################################# [100%] Preparing... ################################# [100%] Updating / installing......
Hope that is a good sign, since usually if people have issues they let you know, but if it works don't do anything. Have started working on next version already. Upgraded 5 of my systems to Fedora 34, and upgraded the build to use the Fedora 34. In first two builds, I had missed copying on of the libraries, and the tests crashed since it was of course one of the critical libraries. But have now got 6 builds with 0.61, all seems good. Did find a few libraries I had that were not actually used by any...
G4L version 0.60 has some major changes in build, but operation is mostly the same. Build Issue This Zip contains 2 ISO files First is similar to the previous ISO files uses syslinux and can be written to CD or Flash 55574528 Oct 13 06:52 g4l-v0.60.iso Second is a new setup using grub4dos as boot loader and can be written to Flash. Has both the standard USB option and supports UEFI boot as well. 67108864 Oct 13 06:53 g4lefi-0.60.iso For most, it should just be writting either version to an usb flash....
G4L version 0.60 has some major changes in build, but operation is mostly the same. Build Issue This Zip contains 2 ISO files First is similar to the previous ISO files uses syslinux and can be written to CD or Flash 55574528 Oct 13 06:52 g4l-v0.60.iso Second is a new setup using grub4dos as boot loader and can be written to Flash. Has both the standard USB option and supports UEFI boot as well. 67108864 Oct 13 06:53 g4lefi-0.60.iso For most, it should just be writting either version to an usb flash....
Just put a new version out of the testing. Have the iso that can be copied to a flash using dd or rufus or other burner to flash. It sets up a 64M FAT32 partition with a gpt setup. Will wipe everything on flash with the process. Uses about 85% of the 64M for the 2 10M kernel files (5.14.10 and 5.15.rc4) and the about 30M ramdisk.lzma file. Uses grub4dos for the regular usb booting, and also uses grub4dos efi for the UEFI boot. Memtest86+ is available with regular usb boot, but doesn't work with UEFI...
Have been working on a few final issues. Have redone the syslinux iso setup to just have 2 kernels versus the 5, since haven't heard of people needing to run older kernels as had happened in past. Also, changed the default to use 800x600, but also have option to use vga ask to set other higher resolutions, but no big reason. 640x480 is a little tight. It also has the HDT, that I haven't found a way to get to work with grub4dos or grub2. Have the g4lefi.iso that has grub4dos to boot in regular usb...
Seems the newer grub4dos doesn't handle memtest directly, and will not use the ones from Fedora. With the syslinux it will use the memtest file from the /boot directory. With grub4dos I had to add the syslinux memdisk as the kernel file, and MEMTEST.IMG extracted from the memtest.iso file. Not sure of why, but that has it loading in the grub4dos non-UEFI boot mode. Still don't have it working on the UEFI boot, and just might not work in UEFI mode in any way. But a step closer. Don't know if there...
Just a few updates. Have made a directory for the g4lefitesting, and it contains the 64M iso image file that contains the image that boots both with regular USB and UEFI usb. The setup uses the same kernel and ramdisk.lzma files. So it contains the bz5x14.7 kernel and the bz5x15.rc2 kernel. The regular syslinux image generally has 5 kernels, but haven't heard of people needed to use older kernels for a long time, so might change the syslinux version to just have 2 kernels as well. Haven't found a...
Uploaded a g4lefi.iso and readme file in new G4LEFI-Testing
Guess everyone else is busy with other things. Have done a lot of testing, and seem to have a setup that contains both a regular boot and UEFI boot on single image. Works on my Dell 9020 machine with either regular USB or UEFI boot option. Use the following script to create it. Find that Windows 64 installation of the grub4dos bootloader doesn't work, that had been issue before. Was an exe program that did work, but now it also fails. So, install the boot loader already. Also, mention issue with...
Was doing more testing, and trying to come up with a solution that will handle both legacy and UEFI options. Have run into some strange things. For a while, I've made a g4l-lite setup available to just had the default kernel and memtest and work run from a Fat flash. Just extract the files to the flash and use the bootlace or other program to install the grub4dos boot loader. With linux the bootlace worked fine, with windows found this other program that would install it easier for usb?? So, I took...
Have discovered a number of things. I've been working to create a script that would create a USB flash with efi from a regular booted cd, syslinux usb or grub4dos usb. syslinux cd and syslinux cd was rather basic, included the g4lefitemp.xz in the isolinux directory, and the script would look for the CD-ROM. It would mount the cdrom or the usb partition having the files, and would then have user select the usb to reformat to gpt and copy the efi files, and then copy the kernel and ramdisk files....
Further update - Did the test with making the iso just 64M in size, and having it use all space. Image is then smaller at 64M versus the 144M of earlier. Shows used space at 78%. Copying it to flash with dd command works fine. Gparted does show message about not using all space, and it does fix it, and make the rest of flash usable. So steps modified to these. dd if=/dev/zero of=g4lefi.iso bs=1M count=64 fdisk g4lefi.iso press g # create GPT-Table n # new partition 1 # partition number ENTER # select...
Tried something that I didn't think would work, but it actually did. Found a page that talked about a complex process to create an iso for efi boot, but it was creating 2 partitions on the iso. One for the efi, and one for an actual linux os. So, I followed the steps, but only created the efi partition. After creating the iso, I opened it with gparted, and it gave a message about a problem with all the space not being used. Think gpt wants some information at the beginning of disk, and then at end...
On a full linux machine, I just use gparted to create a GPT flash with boot and esp flags set, and create a Fat32 partition. Then it just requires extracting the files onto partition. Windows seems to be a little more complicated to do it. Had created an xz archive of the files, but have been think about option to make it simpler for windows users. Made 2 7z self extracting files - Small one just has the EFI files, big one also has kernel and ramdisk.lzma 2.1M g4lefiyy.exe 41M Sep 15 16:23 g4lefiyyall.exe...
Like with 8.1, extracting files from rpm file, and just putting the 8.2 program in path seems to work, so just the installation problem?? ls -l /usr/bin/peazip* lrwxrwxrwx. 1 root root 24 Jun 10 07:27 /usr/bin/peazip -> /usr/share/PeaZip/peazip lrwxrwxrwx. 1 root root 27 Aug 19 21:46 /usr/bin/peazip8.1 -> /usr/share/PeaZip/peazip8.1 lrwxrwxrwx. 1 root root 27 Sep 15 09:58 /usr/bin/peazip8.2 -> /usr/share/PeaZip/peazip8.2 ls -l /usr/share/PeaZip/ total 36328 -rwxrwxrwx. 1 root root 7639 May 11 2007...
rpm -Uvh peazip-8.2.0.LINUX.GTK2-1.x86_64.rpm Verifying... ################################# [100%] Preparing... ################################# [100%] file /usr/lib/.build-id/e1/4acdf5b4474d32ca72b77479eda7e76be7b8a7 from install of peazip-8.2.0.LINUX.GTK2-1.x86_64 conflicts with file from package zstd-1.5.0-1.fc33.x86_64
These are 3 of the 6 computers I have running room. Top one on right is the G4L build machine (6 core amd. One on left is my ftp and web and mariadb server, and one below is another machine. Then have my notebook. Nothing fancy, but it gets the job done.. Well, have fun.
My Dell 9020 is a small machine, and has Windows 10, but it is configured with legacy mode, so doesn't have the UEFI setup or the Secure boot. It just has a single drive, so nothing fancy. Think I paid $185 for it on ebay. Just boot it now and then to run windows updates. Have 5 linux machines in my room. Well, glad you got it working. Got that I confirmed a process to make the flashes with basic programs that are on the g4l cd. As I mentioned there isn't enough free space for the g4lefixx.xz file...
It's now 4:30am here, here seems to be a way to do it using the G4L image booted on a machine that supports the regular boot. Boot up from the g4l and then go to the # command prompt. put an empty flash into a USB port cat /proc/partitions will show a list of the devices, and you should be able to figure out which device is the flash. On my notebook here the output is: cat /proc/partitions major minor #blocks name 11 0 1048575 sr0 8 0 976762584 sda 8 1 19922944 sda1 8 2 102400 sda2 8 3 253402112...
This is setup to load the UEFI directly from the flash partition. On my Dell 9020, it is the just inserting the flash and boot machine. Press F12 key, and the boot menu then shows the UEFI Flash as an option.
Doesn't use the windows boot manager, so have no clue on that. The esp is where the UEFI loads the files directly from the partition that is labeled on the flash as a boot and esp device. Don't have a windows machine that is setup with UEFI boot, so looks like it is trying to use some kind of shims, which I believe have something to do with the secure boot?? sfdisk -l /dev/sdb Disk /dev/sdb: 7.45 GiB, 8002732032 bytes, 15630336 sectors Disk model: Cruzer Blade Units: sectors of 1 * 512 = 512 bytes...
Just did a test on setting up a new USB flash. Tried to do it with just a 1G partition, but something didn't work. So, since the original flash was setup with the full 8G flash setup did the same. used gparted to create the partition table as gpt applied. The create a new partition Fat32 used whole disk. need to set the boot, and esp flags used label GRUB2EFI. applied. left gparted. mount /dev/sdb1 /mnt cd /mnt tar -xvf /g4lefixx.xz cd / umount /mnt Then booted the flash. So, it was the missing boot...
Have recently create a UEFI USB version of G4L
Sounds good. Would have been nice if this was all documented out there, rather than having to piece things together. Final solution wasn't that much more than the previous kernel and filesystem setup. Just having to add the framebuffers and fbcon, which was just finding settings in the kernel build menus, and the EFI files where actually copied right off the Fedora Live Install CD. About 1M of EFI file, 10M kernel (about 400K larger with the framebuffer and fbcon adds) and the 30M ramdisk.lzma.....
Made a minor changes. Added hwinfo to the g4l since the lshw didn't show info on framebuffer. hwinfo --framebuffer will show information on frame buffer. Also, change the bkey script, since it was bothering me that on the high res screen, it was throwing window in top / left. Just remove the begin option, so it now centers link the others. Haven't seen anyone download the g4lefixx.xz file, so no one has tested the UEFI boot. Someone did download the latest ISO, and would have the kernel with framebuffer...
Update: Seems the vga=ask doesn't work with the UEFI boot. Seems UEFI boot passes the video info, and that is what kernel users. When I tested booting the iso in virtual box, the vga=ask works, and also works with the grub4dos boot as well. Well, do some more looking to see if there is an option with UEFI to set resolution. Tried some options, but still getting the high resolution. But it seems to be working fine otherwise.
Contents of g4lefixx.xz archive - tar -xvf g4lefixx.xz to extract tar -Jcvf /g4lefixx.xz * bz5x14.3 bz5x14.3.config EFI/ EFI/BOOT/ EFI/BOOT/BOOTIA32.EFI EFI/BOOT/BOOTX64.EFI EFI/BOOT/grubia32.efi EFI/BOOT/grubx64.efi EFI/BOOT/mmia32.efi EFI/BOOT/mmx64.efi EFI/BOOT/fonts/ EFI/BOOT/fonts/unicode.pf2 EFI/BOOT/grub.cfg EFI/BOOT/x86_64-efi/ EFI/BOOT/x86_64-efi/efi_gop.mod EFI/BOOT/x86_64-efi/efi_uga.mod EFI/BOOT/x86_64-efi/all_video.mod ramdisk.lzma User on Fedora list provided an example boot that used...
The Option to check the EFI boot File is: ftp://setzco.dyndns.org/g4lefixx.xz You need to make a USB Flash formatted at GPT, and Fat32 partition. Then extract files to the Fat32 partition. Then boot machine, press F12, and select UEFI boot device in menu.
It's not a problem. I enjoy doing it. Just frustrating when the information isn't out there. For a while, I was getting a blank screen when I tried the efi boot, so thought something was wrong with the setup. Figured would get some error message. Wasn't till I added some testing echos to try and see if it was the kernel or ramdisk that was causing problem, but it showed that both loaded. So with the empty screen found that it was possible to get to the command prompt by just entering, and was able...
Have a possible full solution now. Had system working with the exception of video output, but telnet into system worked, but was not ideal. Did a bunch more looking, and a lot of just trying stuff. Have ended up with what requires about 100 lines added/changed in the bz5x14.2.config file to build the kernel. Had to add a lot of Framebuffer support options for various GPU cards, and make the fbconsole the primary. What I have now boots my modified kernel bz5x14.2d and loads the same ramdisk.lzma support...
Getting somewhat closer. Had a person on the fedora list give me something that was a little different than other things. Had linuxefi and initrdefi to load file. Had also setup a thing that used the fedora vmlinuz boot kernel, and he change the ramdisk.lzma to a g4l.img file?? Think that is just an uncompress version of the file since it is twice the size. With it, I was able to use the secure boot option, and it loads the g4l and it runs, but has a major issue. Since the vmlinuz kernel doesn't...
Question on the Dell 3080?? I have a Dell 9020 and in the bios it has an option under the Advanced Bios settings that talks about enable additional Rom to allow other than UEFI boots. I wonder if perhaps on the 3080, Dell might have turned this option off by default. Can you check the bios?? Again, my 9020 does support the bios boot, but not sure if this might be an older model. Don't know how the 9020 compares with the 3080 or the 3070.
Asking if anyone has any information on getting UEFI to actually work?? I've now gone thru 6 web sites I found that seem to have processes that are suppose to work, but none have actually fully worked. Some created a flash that is seen as a UEFI flash, and show the option with the G4L, but then it fails to actual boot. Others go thru the process, but then the flash doesn't show up as an UEFI flash at all?? The kernel (I believe) has all the EFI options in kernel build set?? Here are the kernel build...
Just a note: The G4L kernels are 64 bit kernels and have had the EFI options build into the kernel for some time. Fedora stopped making 32 bit version a couple full versions ago, so had to move to 64 bit. Had looked at it many years ago, but found no difference in speed between 32 bit and 64bit then. So, since 32bit would run on 64 bit hardware had no issue. Was a big change in the build and the library directories. But the actual g4l code didn't change at all that I recall. Include the .config file...
The alpha stuff is not on the public sourceforge site. I make some files occassionally on my my home ftp server for people that have an issue that I'm working on. So, would only been seen my going to ftp://setzco.dyndns.org Most recent was a user that was getting kernel crashes if he tried the G4L with some options set in bios? Turned out those options were for moduals that were not included in the kernel build since they had no use for g4l at all. Was able to find the exact options, and build them...
Reminds me of my first experience with a Dell PowerEdge Server about 20 years ago. Our College MIS director demanded at time that any new Machine that was a server had to be a Dell?? Cost about $8,000 compared to a similar $2000 regular machine. But that was how things were. Was setting up to test Windows 2000 server. Booted the brand new server up with 3 fancy SCSI drives in 3 of the 6 SCSI removable bays. Got to the installation point, and got message that no hard drives found?? Turned out, the...
First, there was no issue with the code, since all versions worked fine under linux compiler. It was just issues with messages when cross-compiling the code for windows. Seems some C functions require extra dll support files. I was able to rewrite functions to use other code like fopen to access files, and then use a loop to use fprintf to write binary data. Have tested all of those programs, and they all run fine on a real windows 10 machine with no additional dll files. Didn't try the compile that...
I went to dell's site, but it is useless without and exact machine number to ask for any help. Seems they are more interested in $$$$ than actually working with customers?? https://appuals.com/fix-system-doesnt-have-any-usb-boot-option/ site has some info on enabling legacy options. Don't know if the Dell Silverlake does completely remove legacy support or if there is some complex process that is need to enable it. Might not show up if the secure boot is enabled, so might be a multi-step process....
Had copied code and pasted into codebuilder and saved it. Assumed it was saving as .cpp, but it was saving it as .c, and that is why it was making the changes necessary. Had tried the code in linux codebuilder and it compiled with no changes needed. Then went back to the windows codebuiler, and recompiled using source files with .cpp, and they compiled with no changes. Did notice that the .cpp code produces larger exe files. from .cpp from .c % .c/.cpp changex.exe 86,263 55,921 64.83% fixf2b6.exe...
Another little update. Install codeblocks with ming under wine and recompiled the programs. Ran into a few issues. Doesn't like the include file of the linux under the wine?? //#include <cstdio> //#include <cstring> //#include <cctype> //#include <cstdlib> //#include <time.h> //using namespace std;</time.h></cstdlib></cctype></cstring></cstdio> include <stdio.h></stdio.h> include <string.h></string.h> include <stdlib.h></stdlib.h> include <time.h></time.h> include <ctype.h></ctype.h> Also, seem to...
I have never installed the ming compiler in either windows or in wine. Only have the program installed in the standard linux. That would seem strange that you would have to install the compiler on any machine that wanted to run the exe program to have the dll files available?? With the other programs I have and the one after modification I can copy the exe file to any windows machine and they run without needing other dll files to be copied. Perhaps that is why they are 4+ times the size of the g++...
Seem to have found a fix, but haven't done full test, but program compiles and runs after doing this. dnf install pthread Originally, the compat and -static files where not orginally installed, but now this is what it shows. Last metadata expiration check: 1:57:08 ago on Sun 29 Aug 2021 06:28:31 AM ChST. Package compat-libpthread-nonshared-2.32-10.fc33.x86_64 is already installed. Package mingw32-winpthreads-7.0.0-2.fc33.noarch is already installed. Package mingw32-winpthreads-static-7.0.0-2.fc33.noarch...
Went to a windows 7 machine, and searched for the file on it, and didn't find the dll files on it, so they must not be standardly installed files. Also, looked in winetricks and it has options to install dll files, but none of them is listed as an option? Since I was able to rewrite code to work in a way that doesn't have this issue will stick with it. Works great with linux compiler with no issues, and all the other programs worked fine with the ming compiler with no changes. Just this one. Have...
How would one make those libraries available? If I do the default compile, it compiles the program, but guess it then looks for run time libraries so fails. If I try the static option, it can't find linux libraries?? Was one page that talked about making a bunch of symbolic links of linux so files to the names of a bunch of windows files, but didn't think that would be useful. Was able to modify the program so that it would compile and run fine. Had to make the following changes. DEL FILE fp1; ADD...
Have a number of programs that I've written in Linux cpp, and the run fine, and all so far had compiled with ming and also do exactly the same. Now have a program that compiles, but if I try to run produces the following. 0100:err:module:import_dll Library libgcc_s_seh-1.dll (which is needed by L"H:\staffx19\CONTACTD\win\txtpmr.exe") not found 0100:err:module:import_dll Library libstdc++-6.dll (which is needed by L"H:\staffx19\CONTACTD\win\txtpmr.exe") not found 0100:err:module:LdrInitializeThunk...
Further update with a manual install of binary file from 8.1 rpm file? Did a manual test on one machine. Extracted the rpm file, and copied the peazip 8.1 file to /usr/share/PeaZip/peazip8.1 and created a link in the /usr/bin ls -l /usr/bin/peazip* lrwxrwxrwx. 1 root root 24 Jun 10 07:27 /usr/bin/peazip -> /usr/share/PeaZip/peazip lrwxrwxrwx. 1 root root 27 Aug 19 21:46 /usr/bin/peazip8.1 -> /usr/share/PeaZip/peazip8.1 ls -l /usr/share/PeaZip/peazip* -rwxr-xr-x. 1 root root 12012864 Jun 10 05:11...
To clarify? Using the force option installs the peazip 8.1, but then if try to reinstall the system zstd that it seems to show is still installed, results in a conflict error message?? With the Fedora 33 fully update and Peazip 8.0 installed rpm -qa | grep zstd libzstd-1.5.0-1.fc33.x86_64 libzstd-devel-1.5.0-1.fc33.x86_64 libzstd-1.5.0-1.fc33.i686 zstd-1.5.0-1.fc33.x86_64 Attempt to install fails [root@setzconote ~]# rpm -Uvh /root/Downloads/peazip-8.1.0.LINUX.GTK2-1.x86_64.rpm Verifying... #################################...
Did this command to extract the rpm file contents. rpm2cpio peazip-8.1.0.LINUX.GTK2-1.x86_64.rpm | cpio -idmv cd /root/Downloads/usr/share/PeaZip/res/zstd ls -l total 916 -rwxrwxrwx. 1 root root 157 Jan 6 2020 note.txt -rwxr-xr-x. 1 root root 932056 Jul 18 03:01 zstd which zstd /usr/bin/zstd ls -l /usr/bin/zstd -rwxr-xr-x. 1 root root 932056 May 17 05:59 /usr/bin/zstd diff zstd /usr/bin/zstd No differences?? With the Peazip 8.0 installed cd /usr/share/PeaZip/res/zstd/ ls -l total 1692 -rwxrwxrwx....
Did strace dnf install peazip-8.1.0.LINUX.GTK2-1.x86_64.rpm -y 2>out Then cat -n out | grep -i zstd looked at out file, but didn't see anything that stood out? 1203 openat(AT_FDCWD, "/lib64/libzstd.so.1", O_RDONLY|O_CLOEXEC) = 4 8053 pread64(8, "\0\0\0\0mlib(PayloadIsZstd)\0rtld(GNU"..., 4096, 1232896) = 4096 8246 pread64(8, "\0\0\2\210mlib(PayloadIsZstd)\0rtld(GNU"..., 4096, 2646016) = 4096 8450 pread64(8, "\0\0\0\0refix)\0rpmlib(PayloadIsZstd)"..., 4096, 3932160) = 4096 9471 pread64(8, "\0\0\0\0mlib(PayloadIsZstd)\0rtld(GNU"...,...
Thanks again for the quick response. I've done a number of upgrades over the years, usually used rpm -Uvh and it would upgrade with no issues every time. This time the error comes up. Did also try the dnf install peazip... and it gives the same conflict message. The 8.0 option works fine, so will just stick with that for now. The message under ticket 629 seemed to say to use the --force option to install 8.1, and then just copy the system zstd files on top the the ones peazip installs, but since...
Ticket 629 has a comment that mentioned using the force option to install the peazip 8.1, and it does work, but it results in downgrading the systems zstd and then trying to upgrade the system zstd results in message of conflict the other way? So, I removed peazip 8.1, and put back 8.0 and the system zstd. Never had any issue with upgrading any earlier versions of peazip? Thanks for quick response.
Install fail of 8.1 on Fedora 33