[Crash-utility] Re: [PATCH v6 00/14] gdb stack unwinding support for crash utility

On Mon, Aug 26, 2024 at 11:55 AM <devel-request(a)lists.crash-utility.osci.io&gt; wrote: Thank you for the update, Tao. I have a few more comments here: [1] [PATCH v6 04/14]: + * Task is active, grab CPU's registers + */ + if (is_task_active(tc->task) && VMSS_DUMPFILE()) + return vmware_vmss_get_cpu_reg(tc->processor, regno, name, size, value); Can you help confirm that it only works for the active task? Is this expected behavior? [2] [PATCH v6 07/14] #ifdef CRASH_MERGE CORE_ADDR pc = 0; get_frame_pc_if_available (fi, &pc); if (!is_kvaddr(pc)) { printf_filtered (_("Backtrace stopped due to non-kernel addr: %lx\n"),pc); fi = NULL; break; } #endif I would suggest removing the above printf_filtered(...), otherwise it will be always displayed. [3] [PATCH v6 08/14] a. warning cc -c -g -DPPC64 -m64 -DLZO -DGDB_10_2 ppc64.c -Wall -O2 -Wstrict-prototypes -Wmissing-prototypes -fstack-protector -Wformat-security ppc64.c: In function ‘ppc64_get_current_task_reg’: ppc64.c:2519:15: warning: unused variable ‘task’ [-Wunused-variable] 2519 | ulong task; | ^~~~ b. There are different results between 'bt' and 'gdb bt' commands. Can you help double check? crash> bt PID: 6298 TASK: c000000050bbcb00 CPU: 2 COMMAND: "bash" R0: c0000000002be63c R1: c00000005102b710 R2: c000000001bf8f00 R3: c00000005102b728 R4: 0000000000000000 R5: 0000000000000000 R6: c00000005102b8d8 R7: c00000000f5d0000 R8: 0000000000000000 R9: c0000000557d3c00 R10: 0000000000000001 R11: 0000000000002000 R12: c0000000027634a8 R13: c00000000f6cdf00 R14: 0000000000000000 R15: 0000000000000000 R16: 0000000000000000 R17: 0000000000000000 R18: 0000000000000000 R19: 0000000000000000 R20: 0000000000000000 R21: 0000000000000000 R22: 0000000000000000 R23: 0000000000000000 R24: 0000000000000007 R25: 0000000000000000 R26: 0000000000000000 R27: c00000000274d898 R28: c000000002d0a8d0 R29: c000000002e23df0 R30: 0000000000000000 R31: c000000002e23ddc NIP: c0000000002bdf64 MSR: 8000000000009033 OR3: 0000000000000000 CTR: 0000000000000000 LR: c0000000002be63c XER: 0000000020040004 CCR: 0000000024222280 MQ: 0000000000000001 DAR: 0000000000000000 DSISR: 0000000000000000 Syscall Result: 0000000000000000 [NIP : crash_setup_regs+68] [LR : __crash_kexec+156] #0 [c00000005102b710] crash_setup_regs at c0000000002bdf64 crash> gdb bt #0 0xc0000000002bdf64 in crash_setup_regs (newregs=0xc00000005102b728, oldregs=0x0) at ./arch/powerpc/include/asm/kexec.h:133 #1 0xc0000000002be658 in __crash_kexec (regs=0x0) at kernel/crash_core.c:122 #2 0xc00000000016c284 in panic (fmt=0xc0000000015dd018 "sysrq triggered crash\n") at kernel/panic.c:367 #3 0xc000000000a66e78 in sysrq_handle_crash (key=<optimized out>) at drivers/tty/sysrq.c:154 #4 0xc000000000a67994 in __handle_sysrq (key=key@entry=99 'c', check_mask=check_mask@entry=false) at drivers/tty/sysrq.c:612 #5 0xc000000000a68454 in write_sysrq_trigger (file=<optimized out>, buf=<optimized out>, count=2, ppos=<optimized out>) at drivers/tty/sysrq.c:1181 #6 0xc00000000072868c in pde_write (pde=0xc000000003671f80, file=<optimized out>, buf=<optimized out>, count=<optimized out>, ppos=<optimized out>) at fs/proc/inode.c:334 #7 proc_reg_write (file=<optimized out>, buf=<optimized out>, count=<optimized out>, ppos=<optimized out>) at fs/proc/inode.c:346 #8 0xc00000000063bf20 in vfs_write (file=0xc000000009cb5b00, buf=0x100123a71a0 <error: Cannot access memory at address 0x100123a71a0>, count=2, pos=0xc00000005102bc00) at fs/read_write.c:588 #9 vfs_write (file=0xc000000009cb5b00, buf=0x100123a71a0 <error: Cannot access memory at address 0x100123a71a0>, count=<optimized out>, pos=0xc00000005102bc00) at fs/read_write.c:570 #10 0xc00000000063c4d0 in ksys_write (fd=<optimized out>, buf=0x100123a71a0 <error: Cannot access memory at address 0x100123a71a0>, count=2) at fs/read_write.c:643 #11 0xc000000000031a28 in system_call_exception (regs=0xc00000005102be80, r0=<optimized out>) at arch/powerpc/kernel/syscall.c:153 #12 0xc00000000000d05c in system_call_vectored_common () at arch/powerpc/kernel/interrupt_64.S:198 Backtrace stopped: previous frame inner to this frame (corrupt stack?) crash> [4] [PATCH v6 10/14] diff --git a/xen_hyper.c b/xen_hyper.c index 32e56fa..54c7f57 100644 --- a/xen_hyper.c +++ b/xen_hyper.c @@ -52,7 +52,7 @@ xen_hyper_init(void) */ xht->xen_virt_start &= 0xffffffffc0000000; #endif - + STRUCT_SIZE_INIT(cpumask_t, "cpumask_t"); if (machine_type("X86_64") && symbol_exists("xen_phys_start") && !xen_phys_start()) error(WARNING, Could you please explain why the above changes are needed? [5] [PATCH v6 14/14] a. warning cc -c -g -DARM64 -DLZO -DGDB_10_2 arm64.c -Wall -O2 -Wstrict-prototypes -Wmissing-prototypes -fstack-protector -Wformat-security arm64.c: In function ‘arm64_get_stack_frame’: arm64.c:4138:13: warning: variable ‘ret’ set but not used [-Wunused-but-set-variable] 4138 | int ret; | ^~~ b. indentation issue - bt->task); + ur_bitmap = (struct user_regs_bitmap_struct *)GETBUF(sizeof(*ur_bitmap)); + memset(ur_bitmap, 0, sizeof(*ur_bitmap)); + ur_bitmap->ur.pc = stackframe.pc; [6] The 'info threads' command only displays the current task, not all known tasks. And this looks different behavior from the gdb gdb> info threads Id Target Id Frame * 1 6298 bash 0xc0000000002bdf64 in crash_setup_regs (newregs=0xc00000005102b728, oldregs=0x0) at ./arch/powerpc/include/asm/kexec.h:133 [7] the 'thread' command can not switch to another task. crash> ps ... 6297 6287 0 c000000057517e80 IN 0.0 22784 11776 sshd-session ... gdb> thread 6297 gdb: gdb request failed: thread 6297 gdb> Finally I have to do it with 'set' command as below: crash> set 6297 PID: 6297 COMMAND: "sshd-session" TASK: c000000057517e80 [THREAD_INFO: c000000057517e80] CPU: 0 STATE: TASK_INTERRUPTIBLE crash> set gdb on gdb: on gdb> bt #0 0xc00000009cd57650 in ?? () gdb: gdb request failed: bt gdb> In addition, the code is changed, but its patch log is not updated accordingly. Could you please double check? I won't list them here. Thanks. Lianbo v5 -> v4: