Re: [PATCH 0/6] Allow signals for IO threads

[Jens Axboe]

On 3/26/21 7:46 PM, Stefan Metzmacher wrote:
> 
> Hi Jens,
> 
>> root@ub1704-166:~# LANG=C gdb --pid 1320
>> GNU gdb (Ubuntu 9.2-0ubuntu1~20.04) 9.2
>> Copyright (C) 2020 Free Software Foundation, Inc.
>> License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
>> This is free software: you are free to change and redistribute it.
>> There is NO WARRANTY, to the extent permitted by law.
>> Type "show copying" and "show warranty" for details.
>> This GDB was configured as "x86_64-linux-gnu".
>> Type "show configuration" for configuration details.
>> For bug reporting instructions, please see:
>> <http://www.gnu.org/software/gdb/bugs/>.
>> Find the GDB manual and other documentation resources online at:
>>     <http://www.gnu.org/software/gdb/documentation/>.
>>
>> For help, type "help".
>> Type "apropos word" to search for commands related to "word".
>> Attaching to process 1320
>> [New LWP 1321]
>> [New LWP 1322]
>>
>> warning: Selected architecture i386:x86-64 is not compatible with reported target architecture i386
>>
>> warning: Architecture rejected target-supplied description
>> syscall () at ../sysdeps/unix/sysv/linux/x86_64/syscall.S:38
>> 38      ../sysdeps/unix/sysv/linux/x86_64/syscall.S: No such file or directory.
>> (gdb)
> 
> 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;
>         }

Confirmed, it stops complaining about the arch at that point.

> I'm wondering if we should decouple the PF_KTHREAD and PF_IO_WORKER
> cases even more and keep as much of a userspace-like copy_thread as
> possible.

Probably makes sense, the only thing they really share is the func+arg
setup. Hence PF_IO_WORKER threads likely just use the rest of the init,
where it doesn't conflict with the frame setup.

-- 
Jens Axboe