Crashes when application tries to read memory info

[qnixsynapse]

This is the error after the crash:

Abort was called at 1060 line in file:
/usr/src/debug/intel-compute-runtime/compute-runtime-24.52.32224.5/shared/source/os_interface/linux/drm_neo.cpp
Aborted (core dumped)

When analyzing the stack:

#3  0x000077db56033582 in __GI_abort () at abort.c:73
        act = {__sigaction_handler = {sa_handler = 0x424, sa_sigaction = 0x424}, sa_mask = {__val = {131783850941600, 131783925645728, 0, 108160842766240, 108160842763584, 108160842766248,
              108160842731288, 140726044120800, 131783847489969, 140726044120952, 0, 0, 14261692118253170688, 131783926445504, 131783926435824, 140726044120800}}, sa_flags = 1443421989,
          sa_restorer = 0x8}
#4  0x000077db51418081 in NEO::abortExecution () at /usr/src/debug/intel-compute-runtime/compute-runtime-24.52.32224.5/shared/source/helpers/abort.cpp:14
No locals.
#5  NEO::abortUnrecoverable (line=<optimized out>, file=<optimized out>) at /usr/src/debug/intel-compute-runtime/compute-runtime-24.52.32224.5/shared/source/helpers/debug_helpers.cpp:27
No locals.
#6  0x000077db518c629a in NEO::Drm::queryMemoryInfo (this=<optimized out>)
    at /usr/src/debug/intel-compute-runtime/compute-runtime-24.52.32224.5/shared/source/os_interface/linux/drm_neo.cpp:1060
No locals.
#7  NEO::Drm::queryMemoryInfo (this=<optimized out>) at /usr/src/debug/intel-compute-runtime/compute-runtime-24.52.32224.5/shared/source/os_interface/linux/drm_neo.cpp:1059
No locals.
#8  0x000077db516ab36f in L0::LinuxMemoryImp::getState (this=0x625f28b2c860, pState=0x7ffd55ded9d0)
    at /usr/src/debug/intel-compute-runtime/compute-runtime-24.52.32224.5/level_zero/tools/source/sysman/memory/linux/os_memory_imp_dg1.cpp:46
        deviceRegions = std::vector of length 0, capacity 0
        memoryInfo = <optimized out>
#9  0x000077db54d4f735 in urDeviceGetInfo () from /opt/intel/oneapi/compiler/2025.0/lib/libur_adapter_level_zero.so.0
No symbol table info available.
#10 0x000077db558ee81d in ur_loader::urDeviceGetInfo(ur_device_handle_t_*, ur_device_info_t, unsigned long, void*, unsigned long*) ()
   from /opt/intel/oneapi/compiler/2025.0/lib/libur_loader.so.0
No symbol table info available.
#11 0x000077db558fe09d in urDeviceGetInfo () from /opt/intel/oneapi/compiler/2025.0/lib/libur_loader.so.0
No symbol table info available.
#12 0x000077db567ed9eb in sycl::_V1::detail::device_impl::has(sycl::_V1::aspect) const () from /opt/intel/oneapi/compiler/2025.0/lib/libsycl.so.8
No symbol table info available.
#13 0x0000625ee8d0738c in printDeviceInfo(sycl::_V1::device const&, bool, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) ()
No symbol table info available.
#14 0x0000625ee8d05097 in main ()

Sysinfo:
Intel Compute Runtime: 24.52.32224.5-1
Linux: 6.13.1
OS: Arch Linux

[saik-intel]

@qnixsynapse do we have reproduction steps? Ideally issue shouldn't crash and will try internally

[qnixsynapse]

@saik-intel Running sycl-ls --verbose triggers it on my system.

This env need to be set in some cases: export ZES_ENABLE_SYSMAN=1
It started happening all of a sudden and I am unable to figure out why.

[JablonskiMateusz]

@qnixsynapse could you share more details about the device you are using?

[qnixsynapse]

@JablonskiMateusz GPU is Intel Arc A750 LE 8GB VRAM, CPU is Intel i3 Alder Lake.

[qnixsynapse]

Fixed by recompiling from source with -DNEO_ENABLE_I915_PRELIM_DETECTION=1.
Looks like a downstream issue to me. Arch Linux Packaging do not use this flag.
Will close it once I get a comment from a maintainer about this flag.

[JablonskiMateusz]

i915 prelim detection is non-default feature and our driver should work fine without this setting. However, based on this issue it looks like something went wrong on our side. @saik-intel we need to fix this on our side, Arch Linux packaging does nothing wrong.
@qnixsynapse please keep this issue open until we provide proper fix for that

[qnixsynapse]

@JablonskiMateusz Okay but it is still weird that recompiling with the flag fixed the issue, albeit with little lower performance.

[JablonskiMateusz]

@JablonskiMateusz Okay but it is still weird that recompiling with the flag fixed the issue, albeit with little lower performance.

"weird" part is the logic in our cmake file where different cpp file is included in case of i915 prelim support and another one in case of no i915 prelim support.

i915 prelim support enabling should just extend the support

[ikubetoomuzik]

I don't know if you need more information at this point but I am experiencing the same issue trying to follow this guide from intel on using Ollama with Intel GPUs.

I am using the same kernel, distro, and Basekit version as OP, however I am using an Intel ARC A770.
Intel Compute Runtime: 24.52.32224.5-1
Linux: 6.13.1
OS: Arch Linux

I see the following message in the ollama serve logs:

="0 B" memory.required.full="3.3 GiB" memory.required.partial="0 B" memory.required.kv="896.0 MiB" memory.required.allocations="[3.3 GiB]" memory.weights.total="2.4 GiB" memory.weights.repeating="2.1 GiB" memory.weights.nonrepeating="308.2 MiB" memory.graph.full="424.0 MiB" memory.graph.partial="570.7 MiB"
time=2025-02-07T10:54:52.965-05:00 level=INFO source=server.go:380 msg="starting llama server" cmd="/home/<user>/miniforge3/envs/llm-cpp/lib/python3.11/site-packages/bigdl/cpp/libs/ollama runner --model /home/<user>/.ollama/models/blobs/sha256-dde5aa3fc5ffc17176b5e8bdc82f587b24b2678c6c66101bf7da77af9f7ccdff --ctx-size 8192 --batch-size 512 --n-gpu-layers 999 --threads 4 --no-mmap --parallel 4 --port 38585"
time=2025-02-07T10:54:52.965-05:00 level=INFO source=sched.go:449 msg="loaded runners" count=1
time=2025-02-07T10:54:52.965-05:00 level=INFO source=server.go:559 msg="waiting for llama runner to start responding"
time=2025-02-07T10:54:52.965-05:00 level=INFO source=server.go:593 msg="waiting for server to become available" status="llm server error"
ggml_sycl_init: GGML_SYCL_FORCE_MMQ:   no
ggml_sycl_init: SYCL_USE_XMX: yes
ggml_sycl_init: found 1 SYCL devices:
time=2025-02-07T10:54:53.034-05:00 level=INFO source=runner.go:967 msg="starting go runner"
time=2025-02-07T10:54:53.034-05:00 level=INFO source=runner.go:968 msg=system info="CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX_VNNI = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | LLAMAFILE = 1 | OPENMP = 1 | AARCH64_REPACK = 1 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX_VNNI = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | LLAMAFILE = 1 | OPENMP = 1 | AARCH64_REPACK = 1 | cgo(gcc)" threads=4
Abort was called at 1060 line in file:
/usr/src/debug/intel-compute-runtime/compute-runtime-24.52.32224.5/shared/source/os_interface/linux/drm_neo.cpp
SIGABRT: abort
PC=0x753af86a5624 m=7 sigcode=18446744073709551610
signal arrived during cgo execution

goroutine 13 gp=0xc000604700 m=7 mp=0xc000580008 [syscall]:
runtime.cgocall(0x558a218584e0, 0xc000099b68)
        runtime/cgocall.go:167 +0x4b fp=0xc000099b40 sp=0xc000099b08 pc=0x558a20cb752b
ollama/llama/llamafile._Cfunc_llama_load_model_from_file(0x753a78000de0, {0x0, 0x3e7, 0x1, 0x0, 0x0, 0x0, 0x558a21857ef0, 0xc00059c2a0, 0x0, ...})
        _cgo_gotypes.go:702 +0x50 fp=0xc000099b68 sp=0xc000099b40 pc=0x558a2107a1f0
ollama/llama/llamafile.LoadModelFromFile.func1({0x7ffc0d52b6dc?, 0x0?}, {0x0, 0x3e7, 0x1, 0x0, 0x0, 0x0, 0x558a21857ef0, 0xc00059c2a0, ...})
        ollama/llama/llamafile/llama.go:247 +0x127 fp=0xc000099c68 sp=0xc000099b68 pc=0x558a2107d627
ollama/llama/llamafile.LoadModelFromFile({0x7ffc0d52b6dc, 0x69}, {0x3e7, 0x0, 0x0, 0x0, {0x0, 0x0, 0x0}, 0xc00047b970, ...})
        ollama/llama/llamafile/llama.go:247 +0x2d6 fp=0xc000099db8 sp=0xc000099c68 pc=0x558a2107d316
ollama/llama/runner.(*Server).loadModel(0xc0001a75f0, {0x3e7, 0x0, 0x0, 0x0, {0x0, 0x0, 0x0}, 0xc00047b970, 0x0}, ...)
        ollama/llama/runner/runner.go:859 +0xc5 fp=0xc000099f10 sp=0xc000099db8 pc=0x558a2108a4a5
ollama/llama/runner.Execute.gowrap1()
        ollama/llama/runner/runner.go:1001 +0xda fp=0xc000099fe0 sp=0xc000099f10 pc=0x558a2108c03a
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc000099fe8 sp=0xc000099fe0 pc=0x558a20cc6001
created by ollama/llama/runner.Execute in goroutine 1
        ollama/llama/runner/runner.go:1001 +0xd0d

goroutine 1 gp=0xc0000061c0 m=nil [runnable]:
log.formatHeader(0xc000556570, {0xc1e1a9134214285f?, 0x1e90248?, 0x558a2259eb60?}, {0x0, 0x0}, 0x0, {0x0, 0x0}, 0x0)
        log/log.go:114 +0x159f fp=0xc0004af748 sp=0xc0004af740 pc=0x558a20ec5f3f
log.(*Logger).output(0xc0000c62d0, 0x0, 0x2, 0xc0004afc60)
        log/log.go:237 +0x330 fp=0xc0004af908 sp=0xc0004af748 pc=0x558a20ec6470
log.Println(...)
        log/log.go:405
ollama/llama/runner.Execute({0xc000036130?, 0x0?, 0x0?})
        ollama/llama/runner/runner.go:1026 +0x118f fp=0xc0004afca8 sp=0xc0004af908 pc=0x558a2108bbef
ollama/cmd.NewCLI.func2(0xc0001cb200?, {0x558a2185cdfd?, 0x4?, 0x558a2185ce01?})
        ollama/cmd/cmd.go:1430 +0x45 fp=0xc0004afcd0 sp=0xc0004afca8 pc=0x558a218575c5
github.com/spf13/cobra.(*Command).execute(0xc0004c0908, {0xc00051cc30, 0xf, 0xf})
        github.com/spf13/cobra@v1.8.1/command.go:985 +0xaaa fp=0xc0004afe58 sp=0xc0004afcd0 pc=0x558a20e4b3ca
github.com/spf13/cobra.(*Command).ExecuteC(0xc000546f08)
        github.com/spf13/cobra@v1.8.1/command.go:1117 +0x3ff fp=0xc0004aff30 sp=0xc0004afe58 pc=0x558a20e4bc9f
github.com/spf13/cobra.(*Command).Execute(...)
        github.com/spf13/cobra@v1.8.1/command.go:1041
github.com/spf13/cobra.(*Command).ExecuteContext(...)
        github.com/spf13/cobra@v1.8.1/command.go:1034
main.main()
        ollama/main.go:12 +0x4d fp=0xc0004aff50 sp=0xc0004aff30 pc=0x558a21857c2d
runtime.main()
        runtime/proc.go:272 +0x29d fp=0xc0004affe0 sp=0xc0004aff50 pc=0x558a20c88f3d
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc0004affe8 sp=0xc0004affe0 pc=0x558a20cc6001

goroutine 2 gp=0xc000006c40 m=nil [force gc (idle)]:
runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc000084fa8 sp=0xc000084f88 pc=0x558a20cbdc2e
runtime.goparkunlock(...)
        runtime/proc.go:430
runtime.forcegchelper()
        runtime/proc.go:337 +0xb8 fp=0xc000084fe0 sp=0xc000084fa8 pc=0x558a20c89278
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc000084fe8 sp=0xc000084fe0 pc=0x558a20cc6001
created by runtime.init.7 in goroutine 1
        runtime/proc.go:325 +0x1a

goroutine 3 gp=0xc000007180 m=nil [GC sweep wait]:
runtime.gopark(0x1?, 0x0?, 0x0?, 0x0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc000085780 sp=0xc000085760 pc=0x558a20cbdc2e
runtime.goparkunlock(...)
        runtime/proc.go:430
runtime.bgsweep(0xc0000b2000)
        runtime/mgcsweep.go:317 +0xdf fp=0xc0000857c8 sp=0xc000085780 pc=0x558a20c7391f
runtime.gcenable.gowrap1()
        runtime/mgc.go:204 +0x25 fp=0xc0000857e0 sp=0xc0000857c8 pc=0x558a20c67f65
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc0000857e8 sp=0xc0000857e0 pc=0x558a20cc6001
created by runtime.gcenable in goroutine 1
        runtime/mgc.go:204 +0x66

goroutine 4 gp=0xc000007340 m=nil [GC scavenge wait]:
runtime.gopark(0x10000?, 0x558a21a02e80?, 0x0?, 0x0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc000085f78 sp=0xc000085f58 pc=0x558a20cbdc2e
runtime.goparkunlock(...)
        runtime/proc.go:430
runtime.(*scavengerState).park(0x558a2259ed80)
        runtime/mgcscavenge.go:425 +0x49 fp=0xc000085fa8 sp=0xc000085f78 pc=0x558a20c712e9
runtime.bgscavenge(0xc0000b2000)
        runtime/mgcscavenge.go:658 +0x59 fp=0xc000085fc8 sp=0xc000085fa8 pc=0x558a20c71879
runtime.gcenable.gowrap2()
        runtime/mgc.go:205 +0x25 fp=0xc000085fe0 sp=0xc000085fc8 pc=0x558a20c67f05
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc000085fe8 sp=0xc000085fe0 pc=0x558a20cc6001
created by runtime.gcenable in goroutine 1
        runtime/mgc.go:205 +0xa5

goroutine 5 gp=0xc000007c00 m=nil [finalizer wait]:
runtime.gopark(0xc000084648?, 0x558a20c5e465?, 0xb0?, 0x1?, 0xc0000061c0?)
        runtime/proc.go:424 +0xce fp=0xc000084620 sp=0xc000084600 pc=0x558a20cbdc2e
runtime.runfinq()
        runtime/mfinal.go:193 +0x107 fp=0xc0000847e0 sp=0xc000084620 pc=0x558a20c66fe7
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc0000847e8 sp=0xc0000847e0 pc=0x558a20cc6001
created by runtime.createfing in goroutine 1
        runtime/mfinal.go:163 +0x3d

goroutine 6 gp=0xc0001f6e00 m=nil [chan receive]:
runtime.gopark(0xc000086760?, 0x558a20d99665?, 0x40?, 0x88?, 0x558a21e19ce0?)
        runtime/proc.go:424 +0xce fp=0xc000086718 sp=0xc0000866f8 pc=0x558a20cbdc2e
runtime.chanrecv(0xc000050310, 0x0, 0x1)
        runtime/chan.go:639 +0x41c fp=0xc000086790 sp=0xc000086718 pc=0x558a20c5765c
runtime.chanrecv1(0x0?, 0x0?)
        runtime/chan.go:489 +0x12 fp=0xc0000867b8 sp=0xc000086790 pc=0x558a20c57212
runtime.unique_runtime_registerUniqueMapCleanup.func1(...)
        runtime/mgc.go:1781
runtime.unique_runtime_registerUniqueMapCleanup.gowrap1()
        runtime/mgc.go:1784 +0x2f fp=0xc0000867e0 sp=0xc0000867b8 pc=0x558a20c6afcf
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc0000867e8 sp=0xc0000867e0 pc=0x558a20cc6001
created by unique.runtime_registerUniqueMapCleanup in goroutine 1
        runtime/mgc.go:1779 +0x96

goroutine 7 gp=0xc0001f76c0 m=nil [GC worker (idle)]:
runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc000086f38 sp=0xc000086f18 pc=0x558a20cbdc2e
runtime.gcBgMarkWorker(0xc0000518f0)
        runtime/mgc.go:1412 +0xe9 fp=0xc000086fc8 sp=0xc000086f38 pc=0x558a20c6a2c9
runtime.gcBgMarkStartWorkers.gowrap1()
        runtime/mgc.go:1328 +0x25 fp=0xc000086fe0 sp=0xc000086fc8 pc=0x558a20c6a1a5
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc000086fe8 sp=0xc000086fe0 pc=0x558a20cc6001
created by runtime.gcBgMarkStartWorkers in goroutine 1
        runtime/mgc.go:1328 +0x105

goroutine 8 gp=0xc0001f7880 m=nil [GC worker (idle)]:
runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc000087738 sp=0xc000087718 pc=0x558a20cbdc2e
runtime.gcBgMarkWorker(0xc0000518f0)
        runtime/mgc.go:1412 +0xe9 fp=0xc0000877c8 sp=0xc000087738 pc=0x558a20c6a2c9
runtime.gcBgMarkStartWorkers.gowrap1()
        runtime/mgc.go:1328 +0x25 fp=0xc0000877e0 sp=0xc0000877c8 pc=0x558a20c6a1a5
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc0000877e8 sp=0xc0000877e0 pc=0x558a20cc6001
created by runtime.gcBgMarkStartWorkers in goroutine 1
        runtime/mgc.go:1328 +0x105

goroutine 9 gp=0xc0001f7a40 m=nil [GC worker (idle)]:
runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc000087f38 sp=0xc000087f18 pc=0x558a20cbdc2e
runtime.gcBgMarkWorker(0xc0000518f0)
        runtime/mgc.go:1412 +0xe9 fp=0xc000087fc8 sp=0xc000087f38 pc=0x558a20c6a2c9
runtime.gcBgMarkStartWorkers.gowrap1()
        runtime/mgc.go:1328 +0x25 fp=0xc000087fe0 sp=0xc000087fc8 pc=0x558a20c6a1a5
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc000087fe8 sp=0xc000087fe0 pc=0x558a20cc6001
created by runtime.gcBgMarkStartWorkers in goroutine 1
        runtime/mgc.go:1328 +0x105

goroutine 10 gp=0xc0001f7c00 m=nil [GC worker (idle)]:
runtime.gopark(0x114a6e31383?, 0x0?, 0x0?, 0x0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc000080738 sp=0xc000080718 pc=0x558a20cbdc2e
runtime.gcBgMarkWorker(0xc0000518f0)
        runtime/mgc.go:1412 +0xe9 fp=0xc0000807c8 sp=0xc000080738 pc=0x558a20c6a2c9
runtime.gcBgMarkStartWorkers.gowrap1()
        runtime/mgc.go:1328 +0x25 fp=0xc0000807e0 sp=0xc0000807c8 pc=0x558a20c6a1a5
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc0000807e8 sp=0xc0000807e0 pc=0x558a20cc6001
created by runtime.gcBgMarkStartWorkers in goroutine 1
        runtime/mgc.go:1328 +0x105

goroutine 11 gp=0xc0001f7dc0 m=nil [GC worker (idle)]:
runtime.gopark(0x114a6e33f1f?, 0x3?, 0x2e?, 0x3e?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc000080f38 sp=0xc000080f18 pc=0x558a20cbdc2e
runtime.gcBgMarkWorker(0xc0000518f0)
        runtime/mgc.go:1412 +0xe9 fp=0xc000080fc8 sp=0xc000080f38 pc=0x558a20c6a2c9
runtime.gcBgMarkStartWorkers.gowrap1()
        runtime/mgc.go:1328 +0x25 fp=0xc000080fe0 sp=0xc000080fc8 pc=0x558a20c6a1a5
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc000080fe8 sp=0xc000080fe0 pc=0x558a20cc6001
created by runtime.gcBgMarkStartWorkers in goroutine 1
        runtime/mgc.go:1328 +0x105

goroutine 12 gp=0xc0004a6000 m=nil [GC worker (idle)]:
runtime.gopark(0x114a6e2d8b9?, 0x0?, 0x0?, 0x0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc000081738 sp=0xc000081718 pc=0x558a20cbdc2e
runtime.gcBgMarkWorker(0xc0000518f0)
        runtime/mgc.go:1412 +0xe9 fp=0xc0000817c8 sp=0xc000081738 pc=0x558a20c6a2c9
runtime.gcBgMarkStartWorkers.gowrap1()
        runtime/mgc.go:1328 +0x25 fp=0xc0000817e0 sp=0xc0000817c8 pc=0x558a20c6a1a5
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc0000817e8 sp=0xc0000817e0 pc=0x558a20cc6001
created by runtime.gcBgMarkStartWorkers in goroutine 1
        runtime/mgc.go:1328 +0x105

goroutine 18 gp=0xc000504000 m=nil [GC worker (idle)]:
runtime.gopark(0x558a225c8900?, 0x1?, 0x48?, 0xb0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc00050a738 sp=0xc00050a718 pc=0x558a20cbdc2e
runtime.gcBgMarkWorker(0xc0000518f0)
        runtime/mgc.go:1412 +0xe9 fp=0xc00050a7c8 sp=0xc00050a738 pc=0x558a20c6a2c9
runtime.gcBgMarkStartWorkers.gowrap1()
        runtime/mgc.go:1328 +0x25 fp=0xc00050a7e0 sp=0xc00050a7c8 pc=0x558a20c6a1a5
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc00050a7e8 sp=0xc00050a7e0 pc=0x558a20cc6001
created by runtime.gcBgMarkStartWorkers in goroutine 1
        runtime/mgc.go:1328 +0x105

goroutine 19 gp=0xc0005041c0 m=nil [GC worker (idle)]:
runtime.gopark(0x114a6e2873c?, 0x0?, 0x0?, 0x0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc00050af38 sp=0xc00050af18 pc=0x558a20cbdc2e
runtime.gcBgMarkWorker(0xc0000518f0)
        runtime/mgc.go:1412 +0xe9 fp=0xc00050afc8 sp=0xc00050af38 pc=0x558a20c6a2c9
runtime.gcBgMarkStartWorkers.gowrap1()
        runtime/mgc.go:1328 +0x25 fp=0xc00050afe0 sp=0xc00050afc8 pc=0x558a20c6a1a5
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc00050afe8 sp=0xc00050afe0 pc=0x558a20cc6001
created by runtime.gcBgMarkStartWorkers in goroutine 1
        runtime/mgc.go:1328 +0x105

goroutine 14 gp=0xc0006048c0 m=nil [semacquire]:
runtime.gopark(0x0?, 0x0?, 0x60?, 0xe0?, 0x0?)
        runtime/proc.go:424 +0xce fp=0xc00050d618 sp=0xc00050d5f8 pc=0x558a20cbdc2e
runtime.goparkunlock(...)
        runtime/proc.go:430
runtime.semacquire1(0xc0001a75f8, 0x0, 0x1, 0x0, 0x12)
        runtime/sema.go:178 +0x22c fp=0xc00050d680 sp=0xc00050d618 pc=0x558a20c9bfec
sync.runtime_Semacquire(0x0?)
        runtime/sema.go:71 +0x25 fp=0xc00050d6b8 sp=0xc00050d680 pc=0x558a20cbf445
sync.(*WaitGroup).Wait(0x0?)
        sync/waitgroup.go:118 +0x48 fp=0xc00050d6e0 sp=0xc00050d6b8 pc=0x558a20cd4828
ollama/llama/runner.(*Server).run(0xc0001a75f0, {0x558a21e08aa0, 0xc00052dd10})
        ollama/llama/runner/runner.go:315 +0x47 fp=0xc00050d7b8 sp=0xc00050d6e0 pc=0x558a21086b47
ollama/llama/runner.Execute.gowrap2()
        ollama/llama/runner/runner.go:1006 +0x28 fp=0xc00050d7e0 sp=0xc00050d7b8 pc=0x558a2108bf28
runtime.goexit({})
        runtime/asm_amd64.s:1700 +0x1 fp=0xc00050d7e8 sp=0xc00050d7e0 pc=0x558a20cc6001
created by ollama/llama/runner.Execute in goroutine 1
        ollama/llama/runner/runner.go:1006 +0xde5

[JablonskiMateusz]

Unification fix is merged: e91480c
@qnixsynapse could you confirm that it resolves your isssue (when not setting i915 prelim support at cmake level)?

[qnixsynapse]

Yes, it seems to fix the issue.

I had to remove the diff of the unit tests to get the patch to apply successfully.

BTW, I also noticed that some Windows stuffs are also getting built on Linux. Not sure if that's intentional or not but anyways, thanks.

[JablonskiMateusz]

@qnixsynapse

Yes, it seems to fix the issue.

Great to hear that!

BTW, I also noticed that some Windows stuffs are also getting built on Linux. Not sure if that's intentional or not but anyways, thanks.

It could be related to WSL (WDDM on Linux) support

[qnixsynapse]

Ah. Okay. Closing this.