Jump to content

mue473

Members
  • Posts

    6
  • Joined

  • Last visited

Everything posted by mue473

  1. Hello, at BananaPi my investigations show a difference between the layout of the "struct module" in module.h used for kernel production on one hand and provided for user module production on the other. The layout could be influenced by a lot of configuration switches. For modules coming with kernel, the objdump tool shows in the kernel object file the section for relocating the module structure in this way: RELOCATION RECORDS FOR [.gnu.linkonce.this_module]: OFFSET TYPE VALUE 000000cc R_ARM_ABS32 init_module 000002cc R_ARM_ABS32 cleanup_module and for an externally produced (failing) module: RELOCATION RECORDS FOR [.gnu.linkonce.this_module]: OFFSET TYPE VALUE 000000cc R_ARM_ABS32 init_module 000002c4 R_ARM_ABS32 cleanup_module So the entry point of the exit procedure is stored at offset 2c4 instead of 2cc, overwriting pointer of the linked list of modules. Because after an updated the problem had disappeared, I stopped my work to find out more details, e.g. which switch was not consistent. Regards, Rainer
  2. Hello, I updated my BananaPi-M1 to 6.6.75-current-sunxi-25.5.0-trunk.352 - and the issue has disappeared. Generating and inserting user modules works again, thanks a lot. There are only two cosmetic problems: - "CONFIG_DEBUG_INFO_DWARF5=y" and "CONFIG_DEBUG_INFO=y" in the config file let gcc produce a lot of debug info, the object files are about twenty times as big as in the past. - "CONFIG_DEBUG_INFO_BTF_MODULES=y" triggers an aditional step which will not work: BTF [M] /usr/src/drivers/i2card/i2card.ko Skipping BTF generation for /usr/src/drivers/i2card/i2card.ko due to unavailability of vmlinux
  3. Hello all, the same issue exists on BananaPi-M1 since upgrade to kernel 6.6.75. Originally "Armbian_community_24.5.0-trunk.168_Bananapi_bookworm_current_6.6.21_minimal.img" had been installed and upgraded several times. Because driver module was automatically loaded at start-up, the kernel crashed during init: Message from syslogd@bananapi at Mar 19 15:42:37 ... kernel:[ 3508.455002] 9fe0: bef4b400 bef4b3f0 0044f149 b6bedf72 40030030 00000003 00000000 00000000 [ 3508.083886] i2card: loading out-of-tree module taints kernel. [ 3508.084647] 8<--- cut here --- [ 3508.084672] Unable to handle kernel paging request at virtual address 7f2fe999 when read [ 3508.084700] [7f2fe999] *pgd=00000000 [ 3508.084736] Internal error: Oops: 5 [#1] SMP THUMB2 [ 3508.089676] Modules linked in: i2card(O+) rfkill ir_rc5_decoder lima gpu_sched can_dev drm_shmem_helper axp20x_adc sun4i_ts sunxi_cir sunxi_cedrus(C) v4l2_mem2mem videobuf2_dma_contig lz4hc videobuf2_memops videobuf2_v4l2 videodev videobuf2_common mc lz4 display_connector cpufreq_dt evdev uio_pdrv_genirq uio zram binfmt_misc dm_mod ip_tables x_tables autofs4 axp20x_usb_power pinctrl_axp209 industrialio sunxi phy_generic spidev realtek [ 3508.128583] CPU: 1 PID: 2294 Comm: insmod Tainted: G C O 6.6.75-current-sunxi #2 [ 3508.137241] Message from syslogd@bananapi at Mar 19 15:42:37 ... kernel:[ 3508.446812] 9fc0: 00000000 00000002 46c09400 0000017b 019991d0 00454a8c 00000002 0046fda4 Hardware name: Allwinner sun7i (A20) Family [ 3508.142495] PC is at mod_sysfs_setup+0x194/0x500 [ 3508.147165] LR is at mod_sysfs_setup+0x179/0x500 [ 3508.151815] pc : [<c018a548>] lr : [<c018a52d>] psr: 30030033 [ 3508.158110] sp : f0f49d98 ip : 00000000 fp : 00000008 [ 3508.163362] r10: bf9800cc r9 : bf980384 r8 : f0f49ee8 [ 3508.168614] r7 : 7f2fe985 r6 : bf980108 r5 : bf9800c0 r4 : 00000000 [ 3508.175171] r3 : 7f2fe98d r2 : 00000000 r1 : 00000000 r0 : c157c5d0 [ 3508.181728] Flags: nzCV IRQs on FIQs on Mode SVC_32 ISA Thumb Segment none [ 3508.189074] Control: 50c5387d Table: 44ac006a DAC: 00000051
×
×
  • Create New...

Important Information

Terms of Use - Privacy Policy - Guidelines