Arch Linux User Repository

For testing users, contributors: check out https://github.com/archlinux-jerry/nvidia-340xx
Warning: force push may happen to match the aur tree

Is there a way to:

1. backport the missing pieces with GNOME so that newer GNOME can work with the older libpavucontrol, etc, for 340x compatibility?
2. use gnome-performance and mutter-performance and send MRs for appropriate backports and fixes?
3. Make a series of AUR Packages and metapackages to "Freeze" the "last compatible GNOME + compatible packages" and have essentially a "legacy GNOME" session for nvidia 340 compatibility. Which, ironically if it preserves GLX compatibility, means that perhaps other graphics cards will be able to make use of this as well?
4. Switch to MATE or XFCE or LXDE or Cinammon
5. ULTIMATE WEAPON: Someone/s creates a GTK2 backport of the entire functionality and UI of GNOME, perhaps using MATE as a base (or Cinnamon).

@nobyte According to nvidia-304xx-utils, libGL.so.340.108 is located in /usr/lib/nvidia/libGL.so.340.108, i believe that copying the lib to another place and loading it using LD_PRELOAD may not be necessary. This command below should do the trick

sudo patchelf --add-needed /usr/lib/libpthread.so.0 /usr/lib/nvidia/libGL.so.340.108

In case anyone has problems running QT applications (segfault), there's a workaround described here: https://forums.gentoo.org/viewtopic-p-8793005.html#8793005

Run the application like this:

$ cp /usr/lib64/opengl/nvidia/lib/libGL.so.340.108 .

$ patchelf --add-needed libpthread.so.0 libGL.so.340.108

$ LD_PRELOAD=${PWD}/libGL.so.340.108 yourQTapplication

@Anakievs Yes, but who knows how long will this option be accessible.

@TheExplorer I can launch it with GSK_RENDERER=cairo pavucontrol. There is also https://archlinux.org/packages/extra/x86_64/pavucontrol-qt which I haven't tried.

@TheExplorer That is sad. Nonetheless, for some, it'll be still possible to use the Nouveau driver. On my machine, it works (not as well as the proprietary driver, but still).

Sad news: more and more apps are migrating (like pavucontrol, for example) from GTK3 to GTK4 which is incompatible with nvidia-340. As I understand, GTK4 removed support for GLX. I guess it's inevitable that we would stop using nvidia-340 drivers. It'll be simply impossible :(

I have this running now on kernel 6.9.1. I get the nasty message in dmesg below but things still seem to work. The real problem I am struggling with now though is that neither sddm nor lightdm are working. I get a very brief flicker when the display would normally switch to X11 but that is it. However I can open another tty and do

$ startx /usr/bin/startplasma-x11

and that works fine. This all happened after a full system upgrade, and I've tried going back to kernel 6.6 and have the same problem. Any ideas anyone?

Strangely enough, gdm works but it is slow to start up.

[    5.893521] ------------[ cut here ]------------
[    5.893529] Unpatched return thunk in use. This should not happen!
[    5.893533] WARNING: CPU: 0 PID: 294 at arch/x86/kernel/cpu/bugs.c:3023 __warn_thunk+0x2a/0x40
[    5.893543] Modules linked in: nvidia(OE+) videobuf2_v4l2 iTCO_wdt(+) kvm(+) snd_hda_core intel_pmc_bxt cfg80211(+) at24 snd_hwdep videodev iTCO_vendor_support e1000e(+) sha512_ssse3 videobuf2_common lg_laptop snd_pcm sha256_ssse3 wmi_bmof sparse_keymap mc psmouse ptp sha1_ssse3 snd_timer i2c_i801 acpi_cpufreq i2c_smbus snd rfkill pcspkr pps_core lpc_ich soundcore mac_hid i2c_dev crypto_user dm_mod loop nfnetlink ip_tables x_tables ext4 crc32c_generic crc16 mbcache jbd2 serio_raw atkbd libps2 vivaldi_fmap i8042 video intel_agp serio intel_gtt wmi
[    5.893601] CPU: 0 PID: 294 Comm: (udev-worker) Tainted: G           OE      6.9.1-arch1-1 #1 8721656fa781c58301f7268d475f3e6380e2b47c
[    5.893606] Hardware name: LG Electronics                   P310-K.CPW1A9                   /Montevina, BIOS ELGRSF08 10/29/2008
[    5.893608] RIP: 0010:__warn_thunk+0x2a/0x40
[    5.893613] Code: 66 0f 1f 00 0f 1f 44 00 00 80 3d 10 c9 17 02 00 74 05 c3 cc cc cc cc 48 c7 c7 f0 d1 85 8f c6 05 fb c8 17 02 01 e8 e6 0a 06 00 <0f> 0b c3 cc cc cc cc 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00
[    5.893616] RSP: 0018:ffffb22bc0a93a88 EFLAGS: 00010282
[    5.893620] RAX: 0000000000000000 RBX: ffffffffc0855790 RCX: 0000000000000027
[    5.893622] RDX: ffff89203bc219c8 RSI: 0000000000000001 RDI: ffff89203bc219c0
[    5.893624] RBP: ffffb22bc0a93ad8 R08: 0000000000000000 R09: ffffb22bc0a93908
[    5.893626] R10: ffffffff900b21a8 R11: 0000000000000003 R12: 000078be7802d376
[    5.893627] R13: ffffb22bc0a93b30 R14: ffff892005790300 R15: ffff89200142cf20
[    5.893630] FS:  000078be77712640(0000) GS:ffff89203bc00000(0000) knlGS:0000000000000000
[    5.893632] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    5.893635] CR2: 00007d75f64f95f0 CR3: 00000001070f0000 CR4: 00000000000006f0
[    5.893637] Call Trace:
[    5.893641]  <TASK>
[    5.893642]  ? __warn_thunk+0x2a/0x40
[    5.893647]  ? __warn.cold+0x8e/0xe8
[    5.893651]  ? __warn_thunk+0x2a/0x40
[    5.893657]  ? report_bug+0xff/0x140
[    5.893661]  ? console_unlock+0x84/0x130
[    5.893666]  ? handle_bug+0x3c/0x80
[    5.893670]  ? exc_invalid_op+0x17/0x70
[    5.893673]  ? asm_exc_invalid_op+0x1a/0x20
[    5.893679]  ? nv_drm_init+0x20/0x20 [nvidia 7de1dd5dfb9b75b1ce9b2afce35f97b2d0c33e27]
[    5.893918]  ? __warn_thunk+0x2a/0x40
[    5.893923]  ? __warn_thunk+0x2a/0x40
[    5.893926]  warn_thunk_thunk+0x1a/0x30
[    5.893931]  ? nv_drm_init+0x20/0x20 [nvidia 7de1dd5dfb9b75b1ce9b2afce35f97b2d0c33e27]
[    5.894103]  nvidia_init_module+0x2d/0x760 [nvidia 7de1dd5dfb9b75b1ce9b2afce35f97b2d0c33e27]
[    5.894276]  ? nv_drm_init+0x20/0x20 [nvidia 7de1dd5dfb9b75b1ce9b2afce35f97b2d0c33e27]
[    5.894447]  ? nvidia_frontend_init_module+0x50/0x870 [nvidia 7de1dd5dfb9b75b1ce9b2afce35f97b2d0c33e27]
[    5.894619]  ? nv_drm_init+0x20/0x20 [nvidia 7de1dd5dfb9b75b1ce9b2afce35f97b2d0c33e27]
[    5.894790]  ? do_one_initcall+0x5b/0x310
[    5.894795]  ? do_init_module+0x60/0x220
[    5.894800]  ? init_module_from_file+0x89/0xe0
[    5.894805]  ? idempotent_init_module+0x121/0x2b0
[    5.894810]  ? __x64_sys_finit_module+0x5e/0xb0
[    5.894814]  ? do_syscall_64+0x82/0x160
[    5.894819]  ? do_fault+0x246/0x490
[    5.894824]  ? __handle_mm_fault+0x855/0xe10
[    5.894829]  ? __count_memcg_events+0x4d/0xb0
[    5.894834]  ? count_memcg_events.constprop.0+0x1a/0x30
[    5.894837]  ? handle_mm_fault+0x1f0/0x300
[    5.894842]  ? do_user_addr_fault+0x34e/0x620
[    5.894846]  ? exc_page_fault+0x81/0x190
[    5.894851]  ? entry_SYSCALL_64_after_hwframe+0x76/0x7e
[    5.894857]  </TASK>
[    5.894858] ---[ end trace 0000000000000000 ]---

After addition of 0017-gcc-14.patch the driver successfully compiles and works. Thanks for your support.

When building it fails with the following error.

<...>/nvidia-340xx/src/NVIDIA-Linux-x86_64-340.108-no-compat32/kernel/nv->linux.h:326:2: error: #error "NV_PCI_DMA_MAPPING_ERROR() undefined!"
 326 | #error "NV_PCI_DMA_MAPPING_ERROR() undefined!"

What can be the reason of this? How the problem can be tackled?

linux-lts 6.6.30; gcc 14.1.1+r1+g43b730b9134

same here