Bug 2375851 - blender: Crashes in ELF constructor with illegal instruction
Summary: blender: Crashes in ELF constructor with illegal instruction
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Fedora
Classification: Fedora
Component: blender
Version: 45
Hardware: x86_64
OS: Linux
unspecified
urgent
Target Milestone: ---
Assignee: Luya Tshimbalanga
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2025-07-02 07:26 UTC by passymas
Modified: 2026-09-24 01:02 UTC (History)
19 users (show)

Fixed In Version: blender-5.2.2-1.fc44
Clone Of:
Environment:
Last Closed: 2026-09-24 01:02:53 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)

Description passymas 2025-07-02 07:26:05 UTC
Blender does not start... Problem in libc-start.c

Reproducible: Always

Steps to Reproduce:
1. blender
Actual Results:
Does not start.

Expected Results:
Start... Or solve illegal instruction.

Additional Information:
blender
Illegal instruction (core dumped)

#0  0x00005555558968e8 in _sub_I_65535_0.0 ()
#1  0x00007fffeb3c26e4 in call_init (argc=1, argv=0x7fffffffe078, env=<optimized out>) at ../csu/libc-start.c:145
#2  __libc_start_main_impl (main=0x55555581c240 <main(int, char const**)>, argc=1, argv=0x7fffffffe078, init=<optimized out>, fini=<optimized out>, 
    rtld_fini=<optimized out>, stack_end=0x7fffffffe068) at ../csu/libc-start.c:347
#3  0x0000555555899b95 in _start ()
(gdb)

Comment 1 Luya Tshimbalanga 2025-07-02 07:41:48 UTC
It looks like glibc issue hence "Problem in libc-start.c". I failed to reproduce the problem with Blender through Rawhide running in GNOME Boxes. Assigning to the right component.

Comment 2 passymas 2025-07-02 08:16:19 UTC
I understand that it would be good to mention if Gnome Boxes also runs from Rawhide, since this program would be reusing libc to emulate the system.

Comment 3 Florian Weimer 2025-07-02 08:19:25 UTC
That is code from the main program:

#0  0x00005555558968e8 in _sub_I_65535_0.0 ()

Are you using the blender package from Fedora?

Comment 4 passymas 2025-07-02 08:22:45 UTC
Yes:

cat /usr/bin/blender | sha256sum
a36efa1d83155ea831a53e4ba4b5098280a29dd4e012fbdccf0e2d374933f2e6  -

ldd /usr/lib64/libc.so.6 
	/lib64/ld-linux-x86-64.so.2 (0x00007f3517d25000)
	linux-vdso.so.1 (0x00007ffc059da000)

cat /usr/lib64/libc.so.6 | sha256sum
1b0f39fca8ac51d1fadfb1c7898ba8988e26941ff2ae3746fee8b310dd337d88  -

cat /lib64/ld-linux-x86-64.so.2 | sha256sum
f3a4a10869d2bde876ae1023e0565ad0b6d713dc79e352cce9be615fbe66c5aa  -

Comment 5 Florian Weimer 2025-07-02 08:28:09 UTC
Reassigning to blender then. Koji is currently down, so I can't view the build log and check if incorrect build flags were used.

Comment 6 passymas 2025-07-02 11:49:11 UTC
(gdb) print env
$2 = {{__jmpbuf = {0, 0, 0, 0, 0, 0, 0, 0}, __mask_was_saved = 0, __saved_mask = {__val = {0 <repeats 16 times>}}}}

(gdb) print stack_end
could not read '.gnu_debugaltlink' section

(gdb) print rtld_fini
could not read '.gnu_debugaltlink' section

(gdb) print main
$3 = {int (int, const char **)} 0x55555581c240 <main(int, char const**)>

(gdb) print argc
could not read '.gnu_debugaltlink' section

(gdb) print argv
could not read '.gnu_debugaltlink' section

(gdb) print init

Comment 7 passymas 2025-07-02 12:00:22 UTC
(gdb) bt full
#0  0x00005555558968e8 in _sub_I_65535_0.0 ()
No symbol table info available.
#1  0x00007fffeb3c26e4 in call_init (argc=1, argv=0x7fffffffe078, env=<optimized out>) at ../csu/libc-start.c:145
        j = 0
        jm = <optimized out>
        addrs = <optimized out>
        l = <optimized out>
        init_array = <optimized out>
#2  __libc_start_main_impl (main=0x55555581c240 <main(int, char const**)>, argc=1, argv=0x7fffffffe078, init=<optimized out>, fini=<optimized out>, 
    rtld_fini=<optimized out>, stack_end=0x7fffffffe068) at ../csu/libc-start.c:347
No locals.
#3  0x0000555555899b95 in _start ()
No symbol table info available.

Comment 8 Florian Weimer 2025-07-02 13:04:52 UTC
The output of

(gdb) disassemble

would likely be more helpful because it should show the faulting instruction.

Comment 9 passymas 2025-07-02 19:59:50 UTC
(gdb) disassemble
Dump of assembler code for function _sub_I_65535_0.0:
   0x00005555558968a0 <+0>:	endbr64
   0x00005555558968a4 <+4>:	push   %rbp
   0x00005555558968a5 <+5>:	mov    %rsp,%rbp
   0x00005555558968a8 <+8>:	push   %r13
   0x00005555558968aa <+10>:	lea    0x50c36bf(%rip),%r13        # 0x55555a959f70
   0x00005555558968b1 <+17>:	push   %r12
   0x00005555558968b3 <+19>:	movq   %r13,%xmm4
   0x00005555558968b8 <+24>:	push   %rbx
   0x00005555558968b9 <+25>:	movddup %xmm4,%xmm3
   0x00005555558968bd <+29>:	sub    $0x48,%rsp
   0x00005555558968c1 <+33>:	cmpb   $0x0,0x516a578(%rip)        # 0x55555aa00e40
   0x00005555558968c8 <+40>:	movq   0x4a5e1a0(%rip),%xmm2        # 0x55555a2f4a70
   0x00005555558968d0 <+48>:	mov    %fs:0x28,%rax
   0x00005555558968d9 <+57>:	mov    %rax,-0x28(%rbp)
   0x00005555558968dd <+61>:	lea    0x4fa861c(%rip),%rax        # 0x55555a83ef00
   0x00005555558968e4 <+68>:	movaps %xmm3,-0x60(%rbp)
=> 0x00005555558968e8 <+72>:	pinsrq $0x1,%rax,%xmm2
   0x00005555558968ef <+79>:	movaps %xmm2,-0x50(%rbp)
   0x00005555558968f3 <+83>:	jne    0x5555558968fc <_sub_I_65535_0.0+92>
   0x00005555558968f5 <+85>:	movb   $0x1,0x516a544(%rip)        # 0x55555aa00e40
   0x00005555558968fc <+92>:	call   0x5555558966c0 <_ZN7blender11robust_pred9exactinitEv>
   0x0000555555896901 <+97>:	lea    0x515a108(%rip),%rax        # 0x55555a9f0a10
   0x0000555555896908 <+104>:	lea    0x498c791(%rip),%rdx        # 0x55555a2230a0
   0x000055555589690f <+111>:	movq   $0x1,0x515a0ce(%rip)        # 0x55555a9f09e8
   0x000055555589691a <+122>:	lea    -0x30(%rax),%rsi
   0x000055555589691e <+126>:	lea    0x33c97eb(%rip),%rdi        # 0x555558c60110 <_ZNSt13unordered_mapINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEPvSt4hashIS5_ESt8equal_toIS5_ESaISt4pairIKS5_S6_EEED2Ev>
   0x0000555555896925 <+133>:	mov    %rax,0x515a0b4(%rip)        # 0x55555a9f09e0
   0x000055555589692c <+140>:	movq   $0x0,0x515a0b9(%rip)        # 0x55555a9f09f0
   0x0000555555896937 <+151>:	movq   $0x0,0x515a0b6(%rip)        # 0x55555a9f09f8
   0x0000555555896942 <+162>:	movl   $0x3f800000,0x515a0b4(%rip)        # 0x55555a9f0a00
   0x000055555589694c <+172>:	movq   $0x0,0x515a0b1(%rip)        # 0x55555a9f0a08
   0x0000555555896957 <+183>:	movq   $0x0,0x515a0ae(%rip)        # 0x55555a9f0a10
   0x0000555555896962 <+194>:	call   0x5555555656d0 <__cxa_atexit@plt>
   0x0000555555896967 <+199>:	lea    0x515a0e2(%rip),%rax        # 0x55555a9f0a50
   0x000055555589696e <+206>:	lea    0x498c72b(%rip),%rdx        # 0x55555a2230a0
   0x0000555555896975 <+213>:	movq   $0x1,0x515a0a8(%rip)        # 0x55555a9f0a28
   0x0000555555896980 <+224>:	lea    -0x30(%rax),%rsi
   0x0000555555896984 <+228>:	lea    0x33b4ed5(%rip),%rdi        # 0x555558c4b860 <_ZNSt13unordered_mapINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEESt10shared_ptrIN3aud14IDeviceFactoryEESt4hashIS5_ESt8equal_toIS5_ESaISt4pairIKS5_S9_EEED2Ev>
   0x000055555589698b <+235>:	mov    %rax,0x515a08e(%rip)        # 0x55555a9f0a20
   0x0000555555896992 <+242>:	movq   $0x0,0x515a093(%rip)        # 0x55555a9f0a30
--Type <RET> for more, q to quit, c to continue without paging--c
   0x000055555589699d <+253>:	movq   $0x0,0x515a090(%rip)        # 0x55555a9f0a38
   0x00005555558969a8 <+264>:	movl   $0x3f800000,0x515a08e(%rip)        # 0x55555a9f0a40
   0x00005555558969b2 <+274>:	movq   $0x0,0x515a08b(%rip)        # 0x55555a9f0a48
   0x00005555558969bd <+285>:	movq   $0x0,0x515a088(%rip)        # 0x55555a9f0a50
   0x00005555558969c8 <+296>:	call   0x5555555656d0 <__cxa_atexit@plt>
   0x00005555558969cd <+301>:	lea    0x498c6cc(%rip),%rdx        # 0x55555a2230a0
   0x00005555558969d4 <+308>:	lea    0x506d245(%rip),%rsi        # 0x55555a903c20
   0x00005555558969db <+315>:	lea    0x2db3e(%rip),%rdi        # 0x5555558c4520 <_ZNSt10shared_ptrIN7Alembic3Abc3v1216TypedArraySampleINS2_14Uint32TPTraitsEEEED1Ev>
   0x00005555558969e2 <+322>:	call   0x5555555656d0 <__cxa_atexit@plt>
   0x00005555558969e7 <+327>:	cmpb   $0x0,0x510748a(%rip)        # 0x55555a99de78
   0x00005555558969ee <+334>:	jne    0x5555558969f7 <_sub_I_65535_0.0+343>
   0x00005555558969f0 <+336>:	movb   $0x1,0x5107481(%rip)        # 0x55555a99de78
   0x00005555558969f7 <+343>:	lea    0x498c6a2(%rip),%rdx        # 0x55555a2230a0
   0x00005555558969fe <+350>:	lea    0x515a053(%rip),%rsi        # 0x55555a9f0a58
   0x0000555555896a05 <+357>:	lea    0x33b1624(%rip),%rdi        # 0x555558c48030 <_ZNSt10unique_ptrIN3tbb10interface914global_controlESt14default_deleteIS2_EED2Ev>
   0x0000555555896a0c <+364>:	call   0x5555555656d0 <__cxa_atexit@plt>
   0x0000555555896a11 <+369>:	cmpb   $0x0,0x5107460(%rip)        # 0x55555a99de78
   0x0000555555896a18 <+376>:	jne    0x555555896a21 <_sub_I_65535_0.0+385>
   0x0000555555896a1a <+378>:	movb   $0x1,0x5107457(%rip)        # 0x55555a99de78
   0x0000555555896a21 <+385>:	call   0x555555896600 <_Z41__static_initialization_and_destruction_0v.lto_priv.351>
   0x0000555555896a26 <+390>:	cmpb   $0x0,0x510744b(%rip)        # 0x55555a99de78
   0x0000555555896a2d <+397>:	jne    0x555555896a36 <_sub_I_65535_0.0+406>
   0x0000555555896a2f <+399>:	movb   $0x1,0x5107442(%rip)        # 0x55555a99de78
   0x0000555555896a36 <+406>:	call   0x555555896190 <_Z41__static_initialization_and_destruction_0v.lto_priv.350>
   0x0000555555896a3b <+411>:	cmpb   $0x0,0x5107436(%rip)        # 0x55555a99de78
   0x0000555555896a42 <+418>:	jne    0x555555896a4b <_sub_I_65535_0.0+427>
   0x0000555555896a44 <+420>:	movb   $0x1,0x510742d(%rip)        # 0x55555a99de78
   0x0000555555896a4b <+427>:	call   0x555555895940 <_ZN3ccl10BufferPass13register_typeIS0_EEPKNS_8NodeTypeEv>
   0x0000555555896a50 <+432>:	mov    %rax,0x5159d19(%rip)        # 0x55555a9f0770
   0x0000555555896a57 <+439>:	call   0x555555895bf0 <_ZN3ccl12BufferParams13register_typeIS0_EEPKNS_8NodeTypeEv>
   0x0000555555896a5c <+444>:	cmpb   $0x0,0x5107415(%rip)        # 0x55555a99de78
   0x0000555555896a63 <+451>:	mov    %rax,0x510781e(%rip)        # 0x55555a99e288
   0x0000555555896a6a <+458>:	jne    0x555555896a73 <_sub_I_65535_0.0+467>
   0x0000555555896a6c <+460>:	movb   $0x1,0x5107405(%rip)        # 0x55555a99de78
   0x0000555555896a73 <+467>:	call   0x5555558952d0 <_ZN3ccl4Film13register_typeIS0_EEPKNS_8NodeTypeEv>
   0x0000555555896a78 <+472>:	cmpb   $0x0,0x51073f9(%rip)        # 0x55555a99de78
   0x0000555555896a7f <+479>:	mov    %rax,0x5154afa(%rip)        # 0x55555a9eb580
   0x0000555555896a86 <+486>:	jne    0x555555896a8f <_sub_I_65535_0.0+495>
   0x0000555555896a88 <+488>:	movb   $0x1,0x51073e9(%rip)        # 0x55555a99de78
   0x0000555555896a8f <+495>:	call   0x5555558950c0 <_Z41__static_initialization_and_destruction_0v.lto_priv.486>

Comment 10 Florian Weimer 2025-07-02 20:02:59 UTC
(In reply to passymas from comment #9)

> => 0x00005555558968e8 <+72>:	pinsrq $0x1,%rax,%xmm2

That's an SSE 4.1 instruction. It's not supposed to be executed unconditionally by Fedora packages.

Koji is still down, so I can't check the build log.

Comment 11 passymas 2025-07-02 20:15:19 UTC
My computer supports sse4a, from AMD.

Chip description: 
0.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Turks XT [Radeon HD 6670/7670]

cat /proc/cpuinfo | grep flags | grep -i sse4
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt hw_pstate vmmcall npt lbrv svm_lock nrip_save

Comment 12 Florian Weimer 2025-07-02 20:23:40 UTC
(In reply to passymas from comment #11)
> My computer supports sse4a, from AMD.

That's a different ISA extension from SSE 4.1. I found some upstream discussions mentioning that SSE 4.1 is required by some parts of Blender at least. Not sure if this now a global thing that would apply to the Fedora builds as well.

Comment 13 passymas 2025-07-03 08:00:08 UTC
It would be nice if Blender would show you when it needs SSE 4.1.

I installed the old version and it works perfectly.

snap install blender --channel=3.6lts/stable --classic

If support for the computer I have can ever be fixed, I'll install the new version if you let me know.

At the moment, I'm not needing any functionality, I just wanted to see if it works. I asked the AI ​​and it told me that version 4 has these new features.

Light and Shadow Linking
AgX View Transform
Global Illumination with Screen Space Ray Tracing
Improved Material Reflections
Unlimited Light Sources
Visible Lights Through Refractive Surfaces
GPU Acceleration for OpenImageDenoise
Bake Node (Geometry Nodes)
Index Switch Node (Geometry Nodes)
Split to Instances Node (Geometry Nodes)
Sort Elements Node (Geometry Nodes)
Rotate Rotation Node (Geometry Nodes)
Bone Collections
Hierarchical Bone Collections
Motion Paths Relative to Active Camera
Redesigned Video Sequencer
New Modifier Menu
"Lock Camera to View" Gizmo Button
UV Editor Edge/Vertex Slide
UV Editor Snap to Grid
No Auto-Smooth Property
Extensions Platform
Faster Undo Operations

Comment 14 Luya Tshimbalanga 2025-07-03 19:21:28 UTC
The sad news is Blender 4.x now requires SSE4.2 and AMD GCN as minimum.
https://www.blender.org/download/requirements/

Version 3.6 is the last release for the AMD Terascale architecture which is practically outside my packaging ability. If someone is able to support 4.x series with older hardware,feel free to submit a
fix upstream.

Comment 15 Aoife Moloney 2026-08-17 13:37:50 UTC
This bug appears to have been reported against 'rawhide' during the Fedora Linux 45 development cycle.
Changing version to 45.

Comment 16 Fedora Update System 2026-09-15 21:31:42 UTC
FEDORA-2026-2361083bbf (blender-5.2.2-1.fc45) has been submitted as an update to Fedora 45.
https://bodhi.fedoraproject.org/updates/FEDORA-2026-2361083bbf

Comment 17 Fedora Update System 2026-09-16 01:32:53 UTC
FEDORA-2026-2361083bbf has been pushed to the Fedora 45 testing repository.
Soon you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2026-2361083bbf`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2026-2361083bbf

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

Comment 18 Fedora Update System 2026-09-16 01:53:41 UTC
FEDORA-2026-fbaab32755 has been pushed to the Fedora 44 testing repository.
Soon you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2026-fbaab32755`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2026-fbaab32755

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

Comment 19 passymas 2026-09-16 02:11:43 UTC
I still have the same computer, but I don't use Fedora anymore; I switched to Debian Sid Experimental.

Comment 20 Fedora Update System 2026-09-20 02:24:59 UTC
FEDORA-2026-040732b5dc has been pushed to the Fedora 45 testing repository.
Soon you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2026-040732b5dc`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2026-040732b5dc

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

Comment 21 Luya Tshimbalanga 2026-09-21 05:56:29 UTC
(In reply to passymas from comment #19)
> I still have the same computer, but I don't use Fedora anymore; I switched
> to Debian Sid Experimental.

Thanks for letting know. At least bug is fixed.

Comment 22 Fedora Update System 2026-09-22 00:19:11 UTC
FEDORA-2026-040732b5dc (alsa-plugins-1.2.12-11.fc45, amarok-3.3.3-4.fc45, and 99 more) has been pushed to the Fedora 45 stable repository.
If problem still persists, please make note of it in this bug report.

Comment 23 Fedora Update System 2026-09-24 01:02:53 UTC
FEDORA-2026-fbaab32755 (blender-5.2.2-1.fc44) has been pushed to the Fedora 44 stable repository.
If problem still persists, please make note of it in this bug report.


Note You need to log in before you can comment on or make changes to this bug.