Re: [PATCH v15 8/9] target/loongarch: Adjust functions and structure to support user-mode

[gaosong]

On 2022/6/11 上午6:45, Richard Henderson wrote:

> On
>       6/9/22 23:53, gaosong wrote:
>       
> > Hi Richard,
> >         
> >         
> >         On 2022/6/10 上午2:42, Richard Henderson wrote:
> >         
> > > >   void
> > > >             helper_asrtle_d(CPULoongArchState *env, target_ulong rj,
> > > >             target_ulong rk)
> > > >             
> > > >               {
> > > >             
> > > >                   if (rj > rk) {
> > > >             
> > > >             +#ifdef CONFIG_USER_ONLY
> > > >             
> > > >             +        cpu_loop_exit_sigsegv(env_cpu(env), GETPC(),
> > > >             
> > > >             +                              MMU_DATA_LOAD, true,
> > > >             GETPC());
> > > >             
> > > >             +#else
> > > >             
> > > >                       do_raise_exception(env, EXCCODE_ADEM, GETPC());
> > > >             
> > > >             +#endif
> > > >             
> > > >           
> > >           
> > >           This change is wrong.  First, the kernel's do_ade raises
> > >           SIGBUS. Second, GETPC() is a host address, not a guest
> > >           address.  Third, this highlights the fact that the existing
> > >           system code is wrong, and should be setting badvaddr.
> > >           
> > >           
> > >           You need to
> > >           
> > >           (1) set badvaddr here, and then
> > >           
> > >           (2) handle EXCCODE_ADEM in linux-user/loongarch/cpu_loop.c to
> > >           force_fix_fault(TARGET_SIGBUS, TARGET_BUS_ADRERR,
> > >           env->badvaddr). 
> >         
> >         badvaddr is env->pc or base->pc_next?
> >         
> >       
>       
>       I don't know.  The documentation for the ASRT{LE,GD}.D
>       instructions is incomplete.
>       
>       
>       However, from the kernel code,
>       
>       
>       https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/loongarch/kernel/traps.c#n360
>       
>       
>       I can see that the kernel expects *some*
>       value to be set there.  Given that this is the same trap used by
>       the bound check memory accesses, I presume that badvaddr is
>       related to the memory access not the PC.  So I would expect
>       badvaddr to be RJ.
>       
>       
>       But that is a guess.  Please check with your hardware engineers.
>       
>

Thanks you,

I had tested asrtgt.d on hardware.  the log is :

the gdb log

(gdb) info registers
                  zero               ra               tp               sp
R0   0000000000000000 000000fff7e7c774 000000fff7ffef20 000000ffffff6940
                    a0               a1               a2               a3
R4   0000000000000001 000000ffffff6a88 000000ffffff6a98 000000fff7fbc4b0
                    a4               a5               a6               a7
R8   0000000000000000 000000fff7fe6f70 000000ffffff6a80 0000000000000080
                    t0               t1               t2               t3
R12  0000000000000000 0000000000000000 0000000000000000 000000fff7fbeeb8
                    t4               t5               t6               t7
R16  000000fff7fbdd40 000000fff7fbdd40 7f7f7f7f7f7f7f7f 0000000000000000
                    t8                x               fp               s0
R20  ffffff0000000000 0000000000000000 000000ffffff6960 0000000000000000
                    s1               s2               s3               s4
R24  0000000120000658 000000aaaabd9c60 0000000000000000 000000aaaabe9b50
                    s5               s6               s7               s8
R28  000000aaaabd9c60 0000000000000000 0000000000000000 000000aaaabcfa08
pc             0x120000638         0x120000638 <main+32>
badvaddr       0xfff7e935c8        0xfff7e935c8 <__cxa_atexit>
(gdb) stepi

Program received signal SIGSYS, Bad system call.
0x0000000120000638 in main () at asrtle.c:8
8        asm volatile("asrtgt.d  %0,%1\n\t"
(gdb) info registers
                  zero               ra               tp               sp
R0   0000000000000000 000000fff7e7c774 000000fff7ffef20 000000ffffff6940
                    a0               a1               a2               a3
R4   0000000000000001 000000ffffff6a88 000000ffffff6a98 000000fff7fbc4b0
                    a4               a5               a6               a7
R8   0000000000000000 000000fff7fe6f70 000000ffffff6a80 0000000000000080
                    t0               t1               t2               t3
R12  0000000000000000 0000000000000000 0000000000000000 000000fff7fbeeb8
                    t4               t5               t6               t7
R16  000000fff7fbdd40 000000fff7fbdd40 7f7f7f7f7f7f7f7f 0000000000000000
                    t8                x               fp               s0
R20  ffffff0000000000 0000000000000000 000000ffffff6960 0000000000000000
                    s1               s2               s3               s4
R24  0000000120000658 000000aaaabd9c60 0000000000000000 000000aaaabe9b50
                    s5               s6               s7               s8
R28  000000aaaabd9c60 0000000000000000 0000000000000000 000000aaaabcfa08
pc             0x120000638         0x120000638 <main+32>
badvaddr       0x120000638         0x120000638 <main+32>
(gdb) disas
Dump of assembler code for function main:
   0x0000000120000618 <+0>:    addi.d    $r3,$r3,-32(0xfe0)
   0x000000012000061c <+4>:    st.d    $r22,$r3,24(0x18)
   0x0000000120000620 <+8>:    addi.d    $r22,$r3,32(0x20)
   0x0000000120000624 <+12>:    addi.w    $r12,$r0,23(0x17)
   0x0000000120000628 <+16>:    st.d    $r12,$r22,-24(0xfe8)
   0x000000012000062c <+20>:    st.d    $r0,$r22,-32(0xfe0)
   0x0000000120000630 <+24>:    ld.d    $r12,$r22,-32(0xfe0)
   0x0000000120000634 <+28>:    ld.d    $r13,$r22,-32(0xfe0)
=> 0x0000000120000638 <+32>:    asrtgt.d    $r12,$r12
   0x000000012000063c <+36>:    st.d    $r12,$r22,-24(0xfe8)
   0x0000000120000640 <+40>:    move    $r12,$r0
   0x0000000120000644 <+44>:    move    $r4,$r12
   0x0000000120000648 <+48>:    ld.d    $r22,$r3,24(0x18)
   0x000000012000064c <+52>:    addi.d    $r3,$r3,32(0x20)
   0x0000000120000650 <+56>:    jirl    $r0,$r1,0
End of assembler dump.

dmesg :

[151125.931122] pid:32782 [a.out] Caught reserved exception 10 - should not happen

So

badvaddr is the PC, 

exitsting system code is BCE, (10)

And I think the change like this:

void helper_asrtle_d(CPULoongArchState *env,  target_ulong rj, target_ulong  rk)

{

     if (rj > rk) {

        env->badvaddr = env->pc;

        do_raise_exception(env, EXCCODE_BCE,  env->badvaddr);

     }

}

cpu_loop.c

case EXCCODE_BCE:

    force_sig_fault(TARGET_SIGSYS,  TARGET_SI_KERNEL,  env->badvaddr)

    Thanks.
    Song Gao