# 20220522_multiarch-kdump-on-coreos ## aarch64 - Fedora (FCOS) - aarch64 qemu passes - aarch64 on AWS fails without workaround () - Fedora Bug: https://bugzilla.redhat.com/show_bug.cgi?id=2080468 - workaround to remove irqpoll - working with Pingfan on this one already - RHEL (RHCOS) - aarch64 qemu passes - aarch64 on AWS fails without workaround - workaround to remove irqpoll ## ppc64le - RHEL (RHCOS) - panic after sysrq trigger of kdump - RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=2082712 - adding `initcall_blacklist=ibmvtpm_module_init` to KDUMP_CMDLINE_APPEND gets us farther - Fedora (FCOS) - panic after sysrq trigger of kdump (see below) ``` [root@cosa-devsh ~]# echo 1 > /proc/sys/kernel/sysrq [root@cosa-devsh ~]# echo c > /proc/sysrq-trigger [ 68.494076] sysrq: Trigger a crash [ 68.494177] Kernel panic - not syncing: sysrq triggered crash [ 68.494313] CPU: 6 PID: 5897 Comm: bash Kdump: loaded Not tainted 5.17.9-300.fc36.ppc64le #1 [ 68.494494] Call Trace: [ 68.494555] [c00000001240bad0] [c000000000a65d00] dump_stack_lvl+0x74/0xa8 (unreliable) [ 68.494722] [c00000001240bb10] [c00000000014facc] panic+0x150/0x404 [ 68.494860] [c00000001240bba0] [c000000000b749c8] sysrq_handle_crash+0x28/0x30 [ 68.495023] [c00000001240bc00] [c000000000b75310] __handle_sysrq+0xd0/0x230 [ 68.495165] [c00000001240bca0] [c000000000b75c3c] write_sysrq_trigger+0xdc/0x198 [ 68.495330] [c00000001240bce0] [c00000000062a49c] proc_reg_write+0xfc/0x160 [ 68.495470] [c00000001240bd10] [c00000000052bf5c] vfs_write+0xec/0x390 [ 68.495614] [c00000001240bd60] [c00000000052c4cc] ksys_write+0x7c/0x140 [ 68.495727] [c00000001240bdb0] [c00000000002c448] system_call_exception+0x188/0x360 [ 68.495840] [c00000001240be10] [c00000000000c0e8] system_call_vectored_common+0xe8/0x278 [ 68.495953] --- interrupt: 3000 at 0x7fff7f5bd6b8 [ 68.496175] NIP: 00007fff7f5bd6b8 LR: 0000000000000000 CTR: 0000000000000000 [ 68.496285] REGS: c00000001240be80 TRAP: 3000 Not tainted (5.17.9-300.fc36.ppc64le) [ 68.496395] MSR: 800000000280f033 <SF,VEC,VSX,EE,PR,FP,ME,IR,DR,RI,LE> CR: 48222488 XER: 00000000 [ 68.496540] IRQMASK: 0 [ 68.496540] GPR00: 0000000000000004 00007fffe655ada0 00007fff7f6d6e00 0000000000000001 [ 68.496540] GPR04: 000000013df316c0 0000000000000002 0000000000000010 0000000000000063 [ 68.496540] GPR08: 0000000137f156c8 0000000000000000 0000000000000000 0000000000000000 [ 68.496540] GPR12: 0000000000000000 00007fff7f81ac60 0000000137f68818 0000000137f69534 [ 68.496540] GPR16: 0000000137f69400 0000000020000000 0000000000000000 0000000137e6a190 [ 68.496540] GPR20: 000000013de50670 0000000000000000 00007fff7f6d19b0 0000000000000000 [ 68.496540] GPR24: 0000000137f68ac0 0000000000000001 000000013df316c0 0000000000000002 [ 68.496540] GPR28: 0000000000000002 00007fff7f6d19b0 000000013df316c0 0000000000000002 [ 68.497437] NIP [00007fff7f5bd6b8] 0x7fff7f5bd6b8 [ 68.497514] LR [0000000000000000] 0x0 [ 68.497570] --- interrupt: 3000 ``` ## s390x - Fedora (FCOS) - bug in kdump start - works after workaround (mentioned below) - RHEL (RHCOS) - bug in kdump start - RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=2089459 - works after workaround (mentioned below) Workaound: ``` cp /usr/lib/kdump/kdump-lib.sh /tmp/ # edit to be: boot_img=/ostree/fedora-coreos-24852652ec86ea249d3dcd0851dce98f0e26bd65d806c6ed7a35c70f220001a0/vmlinuz-5.17.9-300.fc36.s390x mount --bind /tmp/kdump-lib.sh /usr/lib/kdump/kdump-lib.sh ```