On 2023/02/22 18:36, lijiang wrote:
>>> On Wed, Feb 22, 2023 at 1:17 PM HAGIO KAZUHITO(萩尾 一仁) <k-hagio-ab@nec.com <mailto:k-hagio-ab@nec.com>> wrote:
>>>
>>> On 2023/02/21 12:03, Lianbo Jiang wrote:
>>> > For gdb-10.2, the disassembly code may start with "=>", which needs to
>>> > be stripped when calculating the address. Otherwise, parsing the address
>>> > will fail because the current code always assumes that it starts with the
>>> > "0x". For example:
>>> >
>>> > crash> gdb disassemble 0xffffffffa2317add
>>> > Dump of assembler code for function native_queued_spin_lock_slowpath:
>>> > 0xffffffffa2317ab0 <+0>: nopl 0x0(%rax,%rax,1)
>>> > 0xffffffffa2317ab5 <+5>: push %rbp
>>> > 0xffffffffa2317ab6 <+6>: mov %rsp,%rbp
>>> > ...
>>> > 0xffffffffa2317ad3 <+35>: mov %edx,%eax
>>> > 0xffffffffa2317ad5 <+37>: lock cmpxchg %ecx,(%rdi)
>>> > => 0xffffffffa2317ad9 <+41>: cmp %eax,%edx
>>> > 0xffffffffa2317adb <+43>: jne 0xffffffffa2317ac0 <native_queued_spin_lock_slowpath+16>
>>> > 0xffffffffa2317add <+45>: pop %rbp
>>> > 0xffffffffa2317ade <+46>: xchg %ax,%ax
>>> > ...
>>> Probably the following prints the "=> ".
>>> Out of curiosity, what did you do before the gdb disas?
>>>
>>> I didn't run any crash commands or gdb commands before the gdb disass, the current command(gdb disassemble 0xffffffffa2317add) is the first one.
oh, so is it a VMware memory dump or something?
Yes. This is a vmss vmcore. And this issue was observed on the kernel-3.10.
Anyway, the patch was applied.
https://github.com/crash-utility/crash/commit/59c19818190dd4b7ae0dc2221586a4ad6f4fe905
Thank you, Kazu.
Lianbo
Thanks,
Kazu