Am 01.04.21 um 17:39 schrieb Linus Torvalds:
> On Thu, Apr 1, 2021 at 7:58 AM Stefan Metzmacher <metze@xxxxxxxxx> wrote:
>>
>>>
>>> Ok, the following makes gdb happy again:
>>>
>>> --- a/arch/x86/kernel/process.c
>>> +++ b/arch/x86/kernel/process.c
>>> @@ -163,6 +163,8 @@ int copy_thread(unsigned long clone_flags, unsigned long sp, unsigned long arg,
>>> /* Kernel thread ? */
>>> if (unlikely(p->flags & (PF_KTHREAD | PF_IO_WORKER))) {
>>> memset(childregs, 0, sizeof(struct pt_regs));
>>> + if (p->flags & PF_IO_WORKER)
>>> + childregs->cs = current_pt_regs()->cs;
>>> kthread_frame_init(frame, sp, arg);
>>> return 0;
>>> }
>>
>> Would it be possible to fix this remaining problem before 5.12 final?
>
> Please not that way.
>
> But doing something like
>
> childregs->cs = __USER_CS;
> childregs->ss = __USER_DS;
> childregs->ds = __USER_DS;
> childregs->es = __USER_DS;
>
> might make sense (just do it unconditionally, rather than making it
> special to PF_IO_WORKER).
>
> Does that make gdb happy too?
I haven't tried it, but it seems gdb tries to use PTRACE_PEEKUSR
against the last thread tid listed under /proc/<pid>/tasks/ in order to
get the architecture for the userspace application, so my naive assumption
would be that it wouldn't allow the detection of a 32-bit application
using a 64-bit kernel.
metze
