This is a known behavior. The opencl binary exists in the tree but this feature is not supported at the moment (see: https://git.yoctoproject.org/cgit/cgit.cgi/meta-amd/tree/FEATURES.md?h=dunfell#n100). We might implement this later this year, but in the mean time, you could try the opensource solution available here: https://github.com/RadeonOpenCompute/ROCm-OpenCL-Runtime.git.
Also, the binary solution will most probably not be supported moving forward. AMD was interested in only having an open source solution for the ROCm/OpenCL feature. Unfortunately, due to lack of time and priority for other features, it was not supported in the warrior branch either. It was only supported/tested in the sumo release. Regards - Arsalan ________________________________ From: [email protected] <[email protected]> on behalf of Oleksii Moisieiev via lists.yoctoproject.org <[email protected]> Sent: Wednesday, June 24, 2020 4:15 PM To: [email protected] Subject: [meta-amd] Segfault during clinfo call. Hello. I'm getting SEGFAULT when trying to start clinfo on v1000 board. Error in dmesg is the following: ----------------------------------------- [ 6225.957611] clinfo[27061]: segfault at 10 ip 00007f6e381a8aa0 sp 00007ffeb6b69818 error 4 in libamdocl64.so[7f6e37ca7000+37cc000] [ 6225.968035] Code: cf 74 17 48 8b 01 48 3b 70 70 75 ee 48 3b 90 88 00 00 00 77 e5 f3 c3 0f 1f 00 31 c0 c3 90 66 66 66 2e 0f 1f 84 00 00 00 00 00 <48> 8b 47 10 c6 40 0c 01 c3 90 66 0f 1f 44 00 00 55 53 48 89 fb 48 ----------------------------------------- If I understand correctly - meta-amd provides opencl as prebuilt binaries. Have you experienced such kind of errors before? I'm working meta-amd dunfell release on v1000 board. Added OpenCL support by setting INCLUDE_OPENCL = "yes" in local.conf. Information from dmesg that may be useful: ------------------------------------------- [ 0.076952] Kernel command line: BOOT_IMAGE=(hd0,gpt5)/boot/bzImage root=/dev/nvme0n1p5 rootwait rootfstype=ext4 console=ttyS0,115200n8 console=tty0 amdgpu.hw_i2c=1 amd_iommu_dump=1 panic=60 NUANCE.bootIndex=0 [ 3.460132] amd_uncore: AMD NB counters detected [ 3.463457] amd_uncore: AMD LLC counters detected [ 3.711244] [drm] amdgpu kernel modesetting enabled. [ 3.728727] amdgpu 0000:05:00.0: remove_conflicting_pci_framebuffers: bar 0: 0xe0000000 -> 0xefffffff [ 3.737087] amdgpu 0000:05:00.0: remove_conflicting_pci_framebuffers: bar 2: 0xf0000000 -> 0xf01fffff [ 3.745433] amdgpu 0000:05:00.0: remove_conflicting_pci_framebuffers: bar 5: 0xfe500000 -> 0xfe57ffff [ 3.753781] fb0: switching to amdgpudrmfb from simple [ 3.762350] amdgpu 0000:05:00.0: vgaarb: deactivate vga console [ 3.767000] amdgpu 0000:05:00.0: enabling device (0006 -> 0007) [ 3.836682] amdgpu 0000:05:00.0: VRAM: 1024M 0x000000F400000000 - 0x000000F43FFFFFFF (1024M used) [ 3.844243] amdgpu 0000:05:00.0: GART: 1024M 0x0000000000000000 - 0x000000003FFFFFFF [ 3.850679] amdgpu 0000:05:00.0: AGP: 267419648M 0x000000F800000000 - 0x0000FFFFFFFFFFFF [ 3.880939] [drm] amdgpu: 1024M of VRAM memory ready [ 3.884609] [drm] amdgpu: 3072M of GTT memory ready. [ 3.905167] amdgpu: [powerplay] hwmgr_sw_init smu backed is smu10_smu [ 4.228335] fbcon: amdgpudrmfb (fb0) is primary device [ 4.290124] amdgpu 0000:05:00.0: fb0: amdgpudrmfb frame buffer device [ 4.295684] amdgpu 0000:05:00.0: ring gfx uses VM inv eng 0 on hub 0 [ 4.301068] amdgpu 0000:05:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0 [ 4.307162] amdgpu 0000:05:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0 [ 4.313370] amdgpu 0000:05:00.0: ring comp_1.2.0 uses VM inv eng 5 on hub 0 [ 4.319357] amdgpu 0000:05:00.0: ring comp_1.3.0 uses VM inv eng 6 on hub 0 [ 4.325331] amdgpu 0000:05:00.0: ring comp_1.0.1 uses VM inv eng 7 on hub 0 [ 4.331333] amdgpu 0000:05:00.0: ring comp_1.1.1 uses VM inv eng 8 on hub 0 [ 4.337582] amdgpu 0000:05:00.0: ring comp_1.2.1 uses VM inv eng 9 on hub 0 [ 4.343761] amdgpu 0000:05:00.0: ring comp_1.3.1 uses VM inv eng 10 on hub 0 [ 4.349818] amdgpu 0000:05:00.0: ring kiq_2.1.0 uses VM inv eng 11 on hub 0 [ 4.355811] amdgpu 0000:05:00.0: ring sdma0 uses VM inv eng 0 on hub 1 [ 4.361338] amdgpu 0000:05:00.0: ring vcn_dec uses VM inv eng 1 on hub 1 [ 4.367062] amdgpu 0000:05:00.0: ring vcn_enc0 uses VM inv eng 4 on hub 1 [ 4.372832] amdgpu 0000:05:00.0: ring vcn_enc1 uses VM inv eng 5 on hub 1 [ 4.378575] amdgpu 0000:05:00.0: ring vcn_jpeg uses VM inv eng 6 on hub 1 [ 4.396137] [drm] Initialized amdgpu 3.35.0 20150101 for 0000:05:00.0 on minor 0 [ 7.602513] EDAC amd64: Node 0: DRAM ECC enabled. [ 7.607008] EDAC amd64: F17h_M10h detected (node 0). [ 7.612016] EDAC amd64: MC: 0: 8192MB 1: 0MB [ 7.616887] EDAC amd64: MC: 2: 0MB 3: 0MB [ 7.621701] EDAC amd64: MC: 0: 8192MB 1: 0MB [ 7.626081] EDAC amd64: MC: 2: 0MB 3: 0MB [ 7.630274] EDAC amd64: using x8 syndromes. [ 7.630276] EDAC amd64: MCT channel count: 2 [ 7.630627] EDAC MC0: Giving out device to module amd64_edac controller F17h_M10h: DEV 0000:00:18.3 (INTERRUPT) [ 7.648138] EDAC PCI0: Giving out device to module amd64_edac controller EDAC PCI controller: DEV 0000:00:18.0 (POLLED) ------------------------ Best regards, Oleksii.
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#116): https://lists.yoctoproject.org/g/meta-amd/message/116 Mute This Topic: https://lists.yoctoproject.org/mt/75079349/21656 Group Owner: [email protected] Unsubscribe: https://lists.yoctoproject.org/g/meta-amd/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
