Re: riscv32, qemu-user: semi-functional, what goes wrong?

[Andreas K. Huettel]

> > --- when I run a console program and try to background it with ctl-z, it 
> > hangs
> >      (maybe only the first time, not sure)
> >      repeated ctl-c gets me back to the shell, then the program is in the 
> > background
> > 
> > riscv32 ~ # python
> > Python 3.11.5 (main, Aug 31 2023, 21:56:30) [GCC 13.2.1 20230826] on linux
> > Type "help", "copyright", "credits" or "license" for more information.
> >>>>
> > [1]+  Stopped                 python
> > ^C^C^C^C^C^C^C
> > riscv32 ~ # ^C
> > riscv32 ~ #
> > riscv32 ~ # jobs
> > [1]+  Stopped                 python
> > riscv32 ~ # fg
> > python
> 
> It would help to see where bash is hanging, if it is -- attach to it with
> gdb and run a backtrace.
> 

So I rebuilt bash with debug info, and network-attached host gdb to the debug 
facility of qemu-user. 
gdb only managed to get control when I pressed after ctl-z ctl-c.

==============riscv32 terminal===============================================
riscv32 ~ # 
riscv32 ~ # python
Python 3.11.5 (main, Aug 31 2023, 21:56:30) [GCC 13.2.1 20230826] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> 
[1]+  Stopped                 python
^C
==============gdb terminal===================================================
[...]
^CThe target is not responding to interrupt requests.
Stop debugging it? (y or n) n

Program received signal SIGINT, Interrupt.
0x5558783c in waitchld (block=block@entry=0, wpid=-1) at jobs.c:3883
3883    jobs.c: Datei oder Verzeichnis nicht gefunden.
(gdb) bt
#0  0x5558783c in waitchld (block=block@entry=0, wpid=-1) at jobs.c:3883
#1  0x55587b48 in sigchld_handler (sig=<optimized out>) at jobs.c:3751
#2  <signal handler called>
#3  0x2b2d2540 in ?? ()
(gdb) 
==============gdb terminal===================================================

jobs.c: 
   3878 #if defined (PROCESS_SUBSTITUTION)
   3879       /* Only manipulate the list of process substitutions while SIGCHLD
   3880          is blocked. We only use this as a hint that we can remove FIFOs
   3881          or close file descriptors corresponding to terminated process
   3882          substitutions. */
   3883       if ((ind = find_procsub_child (pid)) >= 0)
   3884         set_procsub_status (ind, pid, WSTATUS (status));
   3885 #endif


> 
> > --- make, when building something, seems to always start only one job in 
> > parallel
> 
> That sounds like make's autoconf found something it didn't like.
>

We found another oddity in make [1,2], where apparently wait() returns an
undocumented exit code.

Since wait seems to also be involved above, I tend to blame this in qemu now...



[1] https://savannah.gnu.org/bugs/?64664
[2] https://lists.gnu.org/archive/html/qemu-devel/2023-09/msg03207.html

-- 
Andreas K. Hüttel
dilfridge@gentoo.org
Gentoo Linux developer
(council, toolchain, base-system, perl, libreoffice)

signature.asc
Description: This is a digitally signed message part.