SlackwareThis Forum is for the discussion of Slackware Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
Distribution: slackware 15.0 64bit, 14.2 64 and 32bit and arm, ubuntu and rasbian
Posts: 495
Rep:
marav: My point is that if for example you update the kernel with huge kernel without removing the generic kernel, for some reason slackware actually installs both the generic and the huge kernel on inital setup/install. for me this caused my system to try to boot from the generic kernel. Although the system started, /boot/efi was not mounted, and could not be mounted because the kernel didn't support vfat, so I couldn't fix the problem in the efi partition without rebooting from the huge kernel on the install media (which obviously had a different kernel version).
luckycyborg: You can set the codepage as an option in the mount command, which over-rides the kernel default.https://www.kernel.org/doc/Documenta...stems/vfat.txt If you are using an automounter in a gui, then you would have to set the codepage in the automounter anyway. hopefully that will make the huge kernels more useful to you. I'm not sure on the rules regarding efi system partition codepages though, as that is loaded by the efi bios initially, so the bios has to understand the codepage used, and probably defaults to 437 the same as most distro's (ubuntu,arch,gentoo,lfs,fedora,debian etc.).
Last edited by timsoft; 03-21-2023 at 12:20 PM.
Reason: fix typo
marav: My point is that if for example you update the kernel with huge kernel without removing the generic kernel, for some reason slackware actually installs both the generic and the huge kernel on inital setup/install. for me this caused my system to try to boot from the generic kernel. Although the system started, /boot/efi was not mounted, and could not be mounted because the kernel didn't support vfat, so I couldn't fix the problem in the efi partition without rebooting from the huge kernel on the install media (which obviously had a different kernel version).
An UEFI system doesn’t need the partition /boot/efi to be mounted
It just needs that it exists
An UEFI system doesn’t need the partition /boot/efi to be mounted
It just needs that it exists
My /etc/fstab:
Code:
/dev/nvme0n1p1 /boot/efi vfat defaults,noauto 1 0
But of course!
However, some good people are misguided to use ancient boot technologies, which requires to have access to ESP partition again and again, for modifying files.
I for one, just like you I use modern boot technologies like GRUB2 which requires access to ESP partition only at system installation time or when the GRUB2 package is updated.
I do not know what others think, but I believe that's much simpler this way, the ESP filesystem is less prone for corruption and the system boot is much more robust. As bonus points, there is no need to learn a distinct bootloader and its configuration for the Legacy BIOS Boot - like LILO, as GRUB2 support also this mode, and it uses for Legacy BIOS Boot exactly the same configuration from /boot/grub/grub.cfg .
Yep, there is no need to fix the ESP mounting, but rather to not advertise antiquated and errors prone boot technologies.
Last edited by LuckyCyborg; 03-21-2023 at 01:13 PM.
I changed all of them anyway for other reasons, so what's your point? You've answered something I didn't ask.
It just seems redundant to set -j to same value multiple times for everything, instead of just for specific overrides like openssl which apparently does not link correctly with default $NUMJOBS.
I don't think that you understood Petris reply and I don't think you understand what this means in a bash script:
Code:
NUMJOBS=${NUMJOBS:-" -j$(expr $(nproc) + 1) "}
So let me explain that when you write:
Code:
SOMEVARIABLE=${SOMEVARIABLE:-"defaultvalue"}
Your SOMEVARIABLE will get "defaultvalue" unless your environment already has set that variable to something else before the script is done.
A default value of -j$(expr $(nproc) + 1) is a rather good default value as it will be you number of cores + 1. If you prefer to have another value you can do something like:
Code:
export NUMJOBS=-j73
./program.SlackBuild
You can set your default value in some file in /etc/profile.d or in some login file in your home directory.
No, I totally understand why he wants to indirectly break my system with bad advice.
Luckily, it's less work for me to rewrite all, than set that in ~profile and wait for builds like openssl to break at any point.
I'm certain he's aware of the consequences, just wants to have a laugh when my system starts to break apart because of what I did, according to his advice.
Forget about it. I didn't ask for solution, I just asked a bunch of questions regarding cause and effect.
I too set up my EFI Partition in /etc/fstab with the noauto option:
Code:
/dev/nvme0n1p2 /boot/efi vfat noauto 1 0
Q: is the 'defaults' option recommended for mounting vfat ?
Usually, the 'defaults' option is what you write in /etc/fstab if you don't want to change any of the default values. In this case, as you have 'noauto' you don't need 'defaults' as noauto is not the default behaviour for a line in /etc/fstab.
In your quoted example there was 'defaults,noauto'. The 'defaults' option implies 'auto', but as the option 'noauto' is given after that it will override 'noauto'.
Distribution: slackware 15.0 64bit, 14.2 64 and 32bit and arm, ubuntu and rasbian
Posts: 495
Rep:
Rather than changing the request for vfat support in generic kernel into "grub is better than elilo/lilo" which has it's points; we still have elilo and lilo for legacy reasons if nothing else. I actually use rEFInd for dualbooting with ms windows, but codepage aside, which is addressable in fstab if needed for those not in western europe or english speaking, it would still be nice to have vfat support in generic. There is support for ext2/3/4 and vfat is the most common (as mentioned re uefi system partition (and even installing via usb)) format not included. while slackware ships with elilo, and has a install from usb option, it is still useful.
luckycyborg: maybe you should request that the slackware installer gives an option to add codepage when it automatically adds /boot/efi mount point for the uefi system partition to fstab. That would be useful for both hugh and generic kernels. Also adding codepage support in automounter for the current locale, over-riding the kernel default would be also useful, based on what you have said.
it would still be nice to have vfat support in generic. There is support for ext2/3/4 and vfat is the most common (as mentioned re uefi system partition (and even installing via usb)) format not included.
The vfat file system support is similar to ext2/3/4 in generic:
Suggestion to make makepkg warn about dangling symlinks:
Code:
--- makepkg.orig 2023-03-22 14:40:25.681795009 +0100
+++ makepkg 2023-03-22 15:21:40.131577707 +0100
@@ -296,6 +296,9 @@
exit 2
fi
+# Search dangling symlinks to generate a warning later on
+dangling="$(find . -xtype l)"
+
echo
echo "Slackware package maker, version 3.14159265."
echo
@@ -451,6 +454,10 @@
echo "WARNING: site_perl directory detected (this is fine for a local package build)"
fi
+if [ -n "$dangling" ]; then
+ echo "WARNING: Dangling symlink(s) detected: $dangling"
+fi
+
# Restore the old permissions if they previously weren't chmod 755
if [ $OLDROOTPERMS -ne 755 ]; then
echo
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.