forums.rs

Forums

Community: Linux Support 2.0 Thread is sticky

Quick find code: 409-410-916-65990459

of 71
Sir Jester
Nov Member 2011

Sir Jester

Posts: 34 Bronze Posts by user Forum Profile RuneMetrics Profile
Well, I'm pretty sure it's not just NVIDIA drivers as I don't have them and it still doesn't work. If I have some time tonight or this weekend (but I have to build my next D&D campaign so I'm not sure I will) I might try to use a previous graphic driver version. (And I still had a bunch of DXP too)

27-Feb-2021 04:49:33 - Last edited on 27-Feb-2021 04:51:17 by Sir Jester

Shudzaki

Shudzaki

Posts: 82 Iron Posts by user Forum Profile RuneMetrics Profile
I have a GT-710 using Nvidia 390 driver, which I need for other applicaitons. It worked fine before.

Switching to Nouveau long term is not an option. However I did switch.
I was able to get past the initial problem. However once Runescape tries to auto configure graphics it crashes with:

rs2client: ../nouveau/pushbuf.c:723: nouveau_pushbuf_data: Assertion `kref' failed.

Nouveau is not going to cut it for me.

27-Feb-2021 06:55:11

smithy01
Dec Member 2019

smithy01

Posts: 92 Iron Posts by user Forum Profile RuneMetrics Profile
Yesterday I was investigating the current issues everyone here seems to be experiencing and I noticed the following.

My laptop is outfitted with NVidia optimus tech which basically means there are two GPUS, an Intel GPU and an NVIDIA GPU. Since the NVIDIA GPU is pretty draining on battery life, my NVIDIA GPU is usually turned off.

I was running the Jagex RS3 client on Intel GPU when I started experiencing the issue with libopengl0. I tried installing the recommended packages: libopengl0, canberraGTK, libegl1 but this didn't help either.

So I switched to the final snap client which worked. However it did come with a itte more lag than on the normal RS3 client. To investigat this a little further I decided to try my NVIDIA GPU.

So last night I turned on my NVIDIA GPU and I was unable to start the snap client. So I tried installing the official RS3 client again and miraculously that worked. However when switching back to my Intel GPU it stopped working again (Same probem as before, stuck on the loading screen).

So as it stands I can use the snap client on Intel GPU and the RS3 official client on the NVIDIA GPU. However I would like to use the Official RS3 client on my Intel GPU which is not possible since the update.

I hope this insight helps.

27-Feb-2021 11:57:41 - Last edited on 27-Feb-2021 12:08:55 by smithy01

Shroomsssss

Shroomsssss

Posts: 11 Bronze Posts by user Forum Profile RuneMetrics Profile
Neither official RS3 client v2.2.8 works anymore, or v2.2.9, on NVIDIA GPU using up-to-date proprietary drivers. No snap.

Using strace and apitrace from what I can tell is that rs2client writes to ~/.nv/GLCache/, then around 10-16 gl frames are rendered, then SEGV_MAPERR after munmap(0x7ff74355c000,):

clock_gettime(CLOCK_MONOTONIC_RAW, {tv_sec=26805, tv_nsec=510834226}) = 0
munmap(0x7ff741cd8000, 266240) = 0
munmap(0x7ff741d19000, 200704) = 0
munmap(0x7ff741d4a000, 2101248) = 0
munmap(0x7ff741f4b000, 1052672) = 0
munmap(0x7ff74204c000, 528384) = 0
munmap(0x7ff7420cd000, 266240) = 0
munmap(0x7ff74210e000, 16781312) = 0
munmap(0x7ff74310f000, 528384) = 0
munmap(0x7ff743190000, 200704) = 0
munmap(0x7ff7431c1000, 684032) = 0
munmap(0x7ff743268000, 1052672) = 0
munmap(0x7ff743369000, 200704) = 0
munmap(0x7ff74339a000, 266240) = 0
munmap(0x7ff7433db000, 1052672) = 0
munmap(0x7ff7434dc000, 524288) = 0
munmap(0x7ff74355c000, 1572864) = 0
munmap(0x7ff7436dc000, 266240) = 0
munmap(0x7ff74371d000, 528384) = 0
munmap(0x7ff750014000, 266240) = 0
munmap(0x7ff74379e000, 397312) = 0
munmap(0x7ff750055000, 524288) = 0
munmap(0x7ff7500d5000, 524288) = 0
munmap(0x7ff750155000, 524288) = 0
munmap(0x7ff7501d5000, 262144) = 0
munmap(0x7ff750215000, 135168) = 0
munmap(0x7ff750236000, 135168) = 0
--- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x7ff74355c028} ---
futex(0x5605caad3a70, FUTEX_WAKE_PRIVATE, 2147483647) = 0
futex(0x7ff755ed69e0, FUTEX_WAKE_PRIVATE, 2147483647) = 0
--- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=NULL} ---
futex(0x5605caab26a0, FUTEX_WAIT_PRIVATE, 2, NULL) = ? ERESTARTSYS (To be restarted if SA_RESTART is set)
--- SIGWINCH {si_signo=SIGWINCH, si_code=SI_KERNEL} ---
futex(0x5605caab26a0, FUTEX_WAIT_PRIVATE, 2, NULL

27-Feb-2021 22:52:19 - Last edited on 28-Feb-2021 01:11:59 by Shroomsssss

Crossbowwzrd

Crossbowwzrd

Posts: 2 Bronze Posts by user Forum Profile RuneMetrics Profile
We use several Linux machines on Runescape regularly.

My machine has the native Radeon drivers, and I am stuck at Loading Application Resources, and the error "radeon: Failed to get PCI ID, error number -13"

Another machine has an NVidia card, and the proprietary driver (nvidia-360.39)

I was able to get it working by installing libopengl0 and libcanberra-gtk-module

sudo apt install libopengl0
sudo apt install libcanberra-gtk-module
(rebooting may be necessary.)

Yet another machine (that used to run runescape fine before) is also stuck at Loading Application Resources, but it seems to be using the basic Intel graphics, and not the NVidia card at all.

... still waiting for a solution for us lowly Radeon users.

01-Mar-2021 01:34:21

scf370

scf370

Posts: 7 Bronze Posts by user Forum Profile RuneMetrics Profile
Radeon user here, ubuntu 20.

Runescape crashes with default drivers but runs with ppa : oibaf/graphics-drivers. However, performance is horrible even on lowest settings.

01-Mar-2021 10:16:54 - Last edited on 01-Mar-2021 10:17:12 by scf370

Shroomsssss

Shroomsssss

Posts: 11 Bronze Posts by user Forum Profile RuneMetrics Profile
After looking at the patch notes and 'recent game updates' forum section, I can see the shaders were updated. Perhaps these shaders are failing to compile or run properly on some cross section of hardware, given other reports of black screens in particular areas.

Given the client on my machine fails not long after ~/.nv/GLCache is populated, NVIDIA's OpenGL Shader Disk Cache For Linux, I'm guessing the problem may have to do with shaders. However I'm not seeing any relevant error messages. With all the munmap() calls prior to a SEGV_MAPERR error, I'm guessing that the program is ungracefully exiting or returning from some subroutine. Perhaps in the error handling code, or after the ungraceful exit from a subroutine is where the invalid memory access occurs.

Perhaps consider releasing "pass-through", or "backup" shaders, which are rudimentary. Automatically fall back to their use when the flagship shaders fail to compile on particular hardware, or when use of the flagship shaders results in a dismal framerate on particular hardware (use some clock routines to determine framerate). Or otherwise print shader compile errors somewhere.

I'm not sure why apitrace isn't showing any shader compile lines like it does for other games. Just around 10-16 repeats of these:
4991 @0 glXCreateContextAttribsARB(dpy = 0x559163986e80, config = 0xf6, share_context = NULL, direct = True, attrib_list = {GLX_CONTEXT_MAJOR_VERSION_ARB, 3, GLX_CONTEXT_MINOR_VERSION_ARB, 3, GLX_CONTEXT_RESET_NOTIFICATION_STRATEGY_ARB, GLX_NO_RESET_NOTIFICATION_ARB, 0}) = 0x559163fe5df0
4992 @0 glXMakeCurrent(dpy = 0x559163986e80, drawable = 39845941, ctx = 0x559163fe5df0) = True
4993 @0 glViewport(x = 0, y = 0, width = 1024, height = 768)
4994 @0 glScissor(x = 0, y = 0, width = 1024, height = 768)
4995 @0 glXMakeCurrent(dpy = 0x559163986e80, drawable = 0, ctx = NULL) = True
4996 @0 glXDestroyContext(dpy = 0x559163986e80, ctx = 0x559163fe5df0)

Just a suggestion.

01-Mar-2021 15:18:33 - Last edited on 01-Mar-2021 18:00:30 by Shroomsssss

Quick find code: 409-410-916-65990459 Back to Top