social.coop is one of the many independent Mastodon servers you can use to participate in the fediverse.
A Fediverse instance for people interested in cooperative and collective projects. If you are interested in joining our community, please apply at https://join.social.coop/registration-form.html.

Administered by:

Server stats:

488
active users

#i386

1 post1 participant0 posts today
Replied in thread

@david_chisnall I was trying to install any type of BSD on my older system that only uses #i386 and IDE CD-ROM. I cannot use a usb thumb drive with this machine. This PC was manufactured around 2003.

#MidnightBSD is the only distribution that has i386 ISOs, as far as I know -- however those boot files are only for usb thumb drives as the file size exceeds 700MB, which is the maximum storage limit for CD-Rs that are inserted into CD-ROMs.

I have #Debian already installed and the machine is slow but it works, so I'm more or less not too concerned.

I finally got motivated to buy a CF to IDE adapter and install NetBSD 10.1-RELEASE on a nostalgic old 1998 Toshiba laptop. It runs well 🙂
Here are my notes on creating a custom kernel, release and boot image. I could not boot an install image separately with no FD, CDROM, or PXE; hence a few more steps. As usual with NetBSD, I learned a few things along the way.

#netbsd #bsd #retrocomputing #sdf #i386

idatum.net/running-netbsd-101-

www.idatum.netLab Notes | Running NetBSD 10.1 on a 1998 Toshiba laptop
Continued thread

Granted, @OS1337 won't run on anything below an #i486SX due to #Linux dropping #i386 with Kernel 3.6.9 & 3.4.99 respectably and #toybox AFAICT never supported anything below #i486 and #backporting it doesn't make sense nor do I have the skills to do so nor would I expect @landley to accept the necessary patches:

  • Similar to Linux this would create a shitload of work that noone wants to do nor should do and thus violate the project goals of toybox by complicating things...

Or as @torvalds once supposedly wrote: "I'm not sentimental. Good riddance."...

GitHubOS1337/docu/linux.kernel.versions.tsv at 70f7c419db6e2ba13f720913bc4237ca165136e8 · OS-1337/OS1337OS/1337 Project . Contribute to OS-1337/OS1337 development by creating an account on GitHub.
Replied in thread

@krutonium I can recommend @bunsenlabs / #BunsenLabsLinux which runs smooth in my #Vaio #P11Z with the shitty #GMA500 in #llvmpipe.

  • If that's too straining, consider #TinyCore, which us a #BusyBox-based distro - abeit very minimalist.

Those are the options with a fancy #GUI.

#i486#linux#i386

Debian Trixie now officially withdraws support for the Linux 32-bit kernel, starting with kernel version 6.11.2. An upload of this version of the kernel was done today on October 6th, signifying that the i386 version of the kernel will no longer be available on Debian Trixie. The kernel maintainers and the release team have, therefore, concluded this support by posting this to the mailing list:

This is a new upstream version with, as usual, many changes.Some significant changes to the package are:  * [i386] Stop building kernel packages  * udeb: Fold i2c-modules, efi-modules, leds-modules, acpi-modules,    fancontrol-modules, srm-modules, and crc-modules into kernel-image  * udeb: Fold event-modules and mouse-modules into input-modulesThe udeb changes will require some changes to package lists on theinstaller side.

This means that all the i386 computers, even the most recent 32-bit only processors before the inception of 64-bit support, such as AMD Athlon XP Series and Intel Pentium 3, will no longer be able to upgrade to Debian Trixie from Debian Bookworm. However, some of the packages will remain to support i386 until either the upstream developers (that is, the real owner of the projects) or the downstream developers (that is, people that package a software for Debian) finally drop support.

A bit unrelated, the udeb packaging has undergone two significant changes, which is to migrate 7 modules – that is, i2c-modules, efi-modules, leds-modules, acpi-modules, fancontrol-modules, srm-modules, and crc-modules – into the kernel-image udeb and two modules – that is, event-modules and mouse-modules – into the input-modules udeb.

The full changelogs that lists what happened to the Linux package is as follows:

A fuller list of changes in the package:  * Properly disable common headers packages  * [x86] Enable IPU6 MIPI drivers for Intel Alder Lake laptops    (Closes: #994449, #1074441, #1060262, #1078170)  * Drop "module: Avoid ABI changes when debug info is disabled" as we no    longer try to maintain a module ABI  * d/rules.real: Unset KBUILD_HOSTCFLAGS etc. instead of overriding to be    empty  * d/rules.d/Makefile.inc: Add scripts/include to header include path  * d/config: Update with the help of kconfigeditor2  * perf tools: Pass EXTRA_CFLAGS through to libbpf build again  * d/control: Drop versions from Build-Depends that are satisfied since buster  * d/control: Fix profiles for Build-Depends on bison, cpio, flex, xz-utils  * d/control: Move bison, cpio, flex, xz-utils to Build-Depends-{Arch,Indep}  * Compile with gcc-14 on all architectures  * [arm64] Include modules for Lenovo Yoga C630 and Lenovo Miix 630 into D-I    packages.  * [riscv64] Enable STARFIVE_STARLINK_CACHE and PCIE_STARFIVE_HOST.  * [riscv64] Enable PHY_STARFIVE_JH7110_DPHY_TX as a module.  * [riscv64] Enable CLK_SOPHGO_SG2042_PLL, CLK_SOPHGO_SG2042_CLKGEN and    CLK_SOPHGO_SG2042_RPGATE as modules.  * [amd64] tools/arch/x86/intel_sdsi: Add intel-sdsi package for Intel SDSi    provisioning tool (Closes: #1059362)  * [amd64] drivers/accel/habanalabs: Enable DRM_ACCEL_HABANALABS    (Habana's AI Processors)  * [amd64] drivers/accel/ivpu: Enable DRM_ACCEL_IVPU (Intel NPU, formerly    called Intel VPU) (Closes: #1079170)  * [x86] Enable GPIO_WHISKEY_COVE, INTEL_BXT_PMIC_THERMAL as module.  * [riscv64] udeb: Add rtc-modules to Provides of kernel-image  * [riscv64] udeb: Ship mtd in kernel-image, drop mtd-core-modules and    add it to to Provides of kernel-image.  * [arm64] udeb: Add kernel modules to get USB/SATA/PCIe working on Rockchip    RK3588  * [arm64] Enable modules for MT8186 Chromebooks  * udeb: Move i2c-hid-of-goodix module to fb-modules  * [arm64] drivers/gpu/drm/panthor: Enable DRM_PANTHOR as module  * [arm64] Enable SC_CAMCC_8280XP as module, camera clock controller on    Lenovo ThinkPad X13s laptop.  * [arm64] Enable SC_GCC_8180X, SM_GPUCC_8150, INTERCONNECT_QCOM_SC8180X and    PINCTRL_SC8180X as modules in order to support Lenovo Flex 5G laptops.  * [arm64] Include modules for Lenovo Flex 5G (Snapdragon SC8180X)  * [arm64] enable Qualcomm X Elite modules  * [arm64] Include modules for Qualcomm X Elite laptops  * [arm64] Enable additional modules for rk356x devices  * [arm64] Update rk3588 platform support  * [rt] Update to 6.11-rt7  * [riscv64] Enable INPUT_MISC (Closes: #1079501).  * Revert "Make linux-libc-dev provide all cross packages".  * [amd64] drivers/crypto/intel/iaa: Enable CRYPTO_DEV_IAA_CRYPTO    (IAA Compression Accelerator Crypto Driver) (Closes: #1079272)  * [powerpc] Explicitly disable CRASH_DUMP on 32-bit (Closes: #1079755)  * [x86] ACPI: Enable ACPI_EC_DEBUGFS as module (Closes: #980555)  * [i386] Stop building kernel packages  * rtla: Switch to out-of-tree build  * rtla: Enable verbose build  * rtla: Build with dpkg's recommended compiler flags (regression in 6.9)  * rtla: Fix missing debug symbols  * rtla: Disable LTO  * rtla: Set LD correctly for cross-build  * objtool: Build with dpkg's recommended compiler flags (regression in    6.5.1-1¬exp1)  * [loong64] Enable kernel support for LBT instructions.  * [loong64] Enable KVM and para-virt support.  * [loong64] Enable USB EHCI and OHCI host support.  * mm/damon: Enable DAMON, DAMON_VADDR, DAMON_PADDR, DAMON_SYSFS,    DAMON_RECLAIM, DAMON_LRU_SORT  * [x86] linux-cpupower: Add intel-speed-select command (Closes: #1036714)  * drivers/net/wireless: Support some Wi-Fi 7 devices: enable ATH12K,    MT7925E, MT7925U, MT7996E and RTW89_8922AE as modules (Closes: #1081114)  * mm: set CONFIG_ZONE_DEVICE=y on arm64, loong64, ppc64, ppc64el, and    riscv64, not only amd64. FS_DAX depends on it.  * Revert "perf tools: Use $KBUILD_BUILD_TIMESTAMP as man page date" which is    no longer useful  * Fix some reproducibility issues (Closes: #1033663)  * d/rules.real: Try harder to set the locale to C.UTF-8  * udeb: Delete obsolete rtc-modules from kernel-image Provides  * udeb: Fold i2c-modules into kernel-image (fixes FTBFS on alpha, sparc64)  * udeb: Fold event-modules and mouse-modules into input-modules  * udeb: Fold efi-modules into kernel-image  * [arm64,armhf] udeb: Fold leds-modules into kernel-image  * [amd64] udeb: Fold acpi-modules into kernel-image  * [powerpc*] udeb: Fold fancontrol-modules into kernel-image  * [alpha] udeb: Fold srm-modules into kernel-image  * udeb: Fold crc-modules into kernel-image  * [arm64,armhf] udeb: Add all watchdog drivers to kernel-image    (Closes: #1081550)  * Remove d/b/genorig.py in favour of uscan  * net/netfilter/ipvs: Enable IP_VS_TWOS as module (Closes: #1082903)  * libcpupower: Update symbols file for changes in 6.11.2-1~exp1.  * [arm64] Enable drivers for AM64 SoC on HummingBoard-T (Closes: #1081837)  * [arm64] udeb: Add kernel modules for I2C, USB and Ethernet on TI AM64  * [arm64] udeb: Add kernel modules for RTC  * [amd64] arch/x86: Enable CONFIG_ADDRESS_MASKING    (Linear Address Masking support) (Closes: #1082296)  * [armhf] Enable support for GPIOs, i2c, spi and G-sensor for Terasic's    DE10-nano board.

In the event that the i386 packages of the Linux kernel are not removed in the final release of Trixie, the next version of Debian will not support i386, or, at least, its kernel. You can see the package here, but it will take hours for the changes to propagate.

A good alternative to Debian in your i386 systems is Arch Linux 32, but it’s better to just upgrade your systems to 64-bit if your processor allows.

https://officialaptivi.wordpress.com/2024/10/06/debian-trixie-no-longer-provides-32-bit-kernel/

@NanoRaptor sadly support discontinued with EoL of #Linux 3.6.x / 3.4.x
according to my data...

  • Not shure if it's possible to #backport #toybox to #i386 (pretty shure noone bothered because Linux already EoL'd i386 long ago and supporting it would require a lot of redundant work noone wants to do anyway!)
GitHubOS1337/docu/linux.kernel.versions.tsv at main · OS-1337/OS1337OS/1337 Project . Contribute to OS-1337/OS1337 development by creating an account on GitHub.
Replied in thread

@corbin @sigmasternchen @nixCraft @docteurklein I do however disagree extremely with the last sentece of your reply because that dismisses the reasons why #Linux went big, why #Hurd was basically dead by the time Linux was conceptualized and still is and why #BSDs only survive in nieches and why almost all #Unix|es like #Solaris are nowadays anomalies at this point, but #Apple's #macOS is gaining more and more users.

That's why more people are rocking #MSDOS on #i386 machines than #nixOS on any architecture...

So yeah, one can't just *"defy gravity"* or rather capitalism like that - even tho I wish for #SystemChangeNotClimateChange sooner than later!!!

DEF CON SocialCorbin (@corbin@defcon.social)@kkarhan@infosec.space @sigmasternchen@comfy.social @nixCraft@mastodon.social @docteurklein@mastodon.social I might have missed something in this thread, but why are "enterprise deployments" desirable? A community-produced distro that can provision its own infrastructure can deploy *itself*, and that seems like more than enough functionality for us to endorse its use. If the complaint is that large corporations aren't doing enough to adopt Nix, then...maybe don't worry about that. Let them have their own infrastructure problems and make poor choices; the Free Software ecosystem does not need their contributions in order to succeed.

@nuintari I disagree.

#SystemD is a "necessary evil" because what existed previously was bad for anything that isn't a "build once never change" server that never gets changed much...
youtube.com/watch?v=o_AIw9bGog

The only reason I "cut" systemd from @OS1337 as of now is because I can't make it - or the #GNUtils that preceded it - fit within the few hundred kB on a 1.440 kB FDD I can spare.

Noone wants to go back to #SysVinit when even the fastest systems took longer to boot that most peoole need to make a shitty coffee or dump ass on the toilet.

  • Amd dumping SysVinit - just like #Linux dumping #i386 - is somewhat inevitable given the amount of work it took to keep it workable at all...

I have created a new Image of our lightweight Linux Distribution called Moonlight. It is based on Debian Stable and available for the x86 architecture. It should run well on low-end machines like netbooks with about 1GB RAM, 20 GB HDD. You can download the ISO on Archive.org: archive.org/details/moonlight-

Internet ArchiveMoonlight Stable x86 : Lioh Möller : Free Download, Borrow, and Streaming : Internet ArchiveMoonlight is a Debian GNU/Linux Blend utilizing the IceWM Windowmanager.It contains all the necessary packages for a convenient desktop experience and still...

If you still got an old Laptop, Netbook or Computer with an i386 processor, there might still be good use for it. Try out 'Moonlight', my Debian GNU/Linux spin utilizing the IceWM windowmanager. Lean and simple. Works good with at least 1 GB RAM.

archive.org/details/moonlight-

Visit spacefun.ch if you need help and please support my work with a donation.

Internet ArchiveMoonlight Stable x86 : Lioh Möller : Free Download, Borrow, and Streaming : Internet ArchiveMoonlight is a Debian GNU/Linux Blend utilizing the IceWM Windowmanager.It contains all the necessary packages for a convenient desktop experience and still...
#Linux#i386#Laptop
Replied in thread

@apalu @phoronix What I mean is that #Xorg is #EoL'd and that all the #legacy stuff works in #Xwayland and using #Wayland for newer solutions is the way to go.

Also #Linux did already drop a lot of #Hardware #support over the years because it was either #unmaintained [a lot of older stuff], #never publicly available [#Intel #GMA500 & #GMA3000-Series...] or was deemed to cumbersome to continue forward [vanilla #i386 support for example]...

As much as I wished we could perpetually make stuff that just runs on an #i8086, that just won't work and just like #UnrealEngine5 won't support #32bit systems or 7th generation game console, so it makes sense that #SDL3 won't support #X11...

I mean, #SDL2 is still available and working so if one needs #X support on #Linux, consider using that.

And yes, I do have machines that are from the 2000s that have such hardware that is affected by it...