Re: [Qemu-devel] [PATCH v8] target-ppc: gdbstub allow byte swapping for reading/writing registers

["Andreas Färber" <afaerber@suse.de>]

Am 04.04.2014 22:23, schrieb Thomas Falcon:
> This patch allows registers to be properly read from and written to
> when using the gdbstub to debug a ppc guest running in little
> endian mode.  It accomplishes this goal by byte swapping the values of
> any registers if the MSR:LE value is set.
> 
> Signed-off-by: Thomas Falcon <tlfalcon@linux.vnet.ibm.com>
> ---
> Differences for v7:
> 
> Inlined the register_read() and register_write() wrapper functions
> ---
>  target-ppc/cpu-qom.h |   1 +
>  target-ppc/gdbstub.c | 125 +++++++++++++++++++++++++++++++++++++--------------
>  2 files changed, 92 insertions(+), 34 deletions(-)
> 
> diff --git a/target-ppc/cpu-qom.h b/target-ppc/cpu-qom.h
> index 47dc8e6..aab4977 100644
> --- a/target-ppc/cpu-qom.h
> +++ b/target-ppc/cpu-qom.h
> @@ -111,6 +111,7 @@ void ppc_cpu_dump_state(CPUState *cpu, FILE *f, fprintf_function cpu_fprintf,
>                          int flags);
>  void ppc_cpu_dump_statistics(CPUState *cpu, FILE *f,
>                               fprintf_function cpu_fprintf, int flags);
> +void ppc_cpu_gdb_swap_register(uint8_t *buf, int reg, int len);

This is only ever used in gdbstub.c. Can we please keep it static there
to avoid a full ppc*-softmmu rebuild?

>  hwaddr ppc_cpu_get_phys_page_debug(CPUState *cpu, vaddr addr);
>  int ppc_cpu_gdb_read_register(CPUState *cpu, uint8_t *buf, int reg);
>  int ppc_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg);
> diff --git a/target-ppc/gdbstub.c b/target-ppc/gdbstub.c
> index 1c91090..594dd08 100644
> --- a/target-ppc/gdbstub.c
> +++ b/target-ppc/gdbstub.c
> @@ -21,6 +21,58 @@
>  #include "qemu-common.h"
>  #include "exec/gdbstub.h"
>  
> +static int ppc_cpu_gdb_register_len(int n)

Nitpick: Since these two functions do not operate on the CPU, you could
just use ppc_gdb_* rather than ppc_cpu_gdb_*.

> +{
> +    switch (n) {
> +    case 0 ... 31:
> +        /* gprs */
> +        return sizeof(target_ulong);
> +    case 32 ... 63:
> +        /* fprs */
> +        if (gdb_has_xml) {
> +            return 0;
> +        }
> +        return 8;
> +    case 66:
> +        /* cr */
> +        return 4;
> +    case 64:
> +        /* nip */
> +    case 65:
> +        /* msr */
> +    case 67:
> +        /* lr */
> +    case 68:
> +        /* ctr */
> +    case 69:
> +        /* xer */
> +        return sizeof(target_ulong);
> +    case 70:
> +        /* fpscr */
> +        if (gdb_has_xml) {
> +            return 0;
> +        }
> +        return sizeof(target_ulong);
> +    default:
> +        return 0;
> +    }
> +}
> +
> +
> +/* The following functions are used to ensure the correct
> + * transfer of registers between a little endian ppc target
> + * and a big endian host by checking the LE bit in the Machine State Register
> + */
> +
> +void ppc_cpu_gdb_swap_register(uint8_t *mem_buf, int n, int len)
> +{
> +    if (len == 4) {
> +        bswap32s((uint32_t *)mem_buf);
> +    } else {
> +        bswap64s((uint64_t *)mem_buf);
> +    }

This logic assumes that len can only be either 4 or 8. Please use an
explicit len == 8 comparison and g_assert_not_reached() on unhandled len
values.

> +}
> +
>  /* Old gdb always expects FP registers.  Newer (xml-aware) gdb only
>   * expects whatever the target description contains.  Due to a
>   * historical mishap the FP registers appear in between core integer
> @@ -32,23 +84,26 @@ int ppc_cpu_gdb_read_register(CPUState *cs, uint8_t *mem_buf, int n)
>  {
>      PowerPCCPU *cpu = POWERPC_CPU(cs);
>      CPUPPCState *env = &cpu->env;
> +    int r = ppc_cpu_gdb_register_len(n);
> +
> +    if (!r) {
> +        return r;
> +    }
>  
>      if (n < 32) {
>          /* gprs */
> -        return gdb_get_regl(mem_buf, env->gpr[n]);
> +        gdb_get_regl(mem_buf, env->gpr[n]);
>      } else if (n < 64) {
>          /* fprs */
> -        if (gdb_has_xml) {
> -            return 0;
> -        }

I stumbled over dropping this not being related to Little Endian or
being mentioned in the commit message. Maybe mention that this is
replaced by ..._register_len() and returning early?

Otherwise looking much more to my taste now. :)

Regards,
Andreas

>          stfq_p(mem_buf, env->fpr[n-32]);
> -        return 8;
>      } else {
>          switch (n) {
>          case 64:
> -            return gdb_get_regl(mem_buf, env->nip);
> +            gdb_get_regl(mem_buf, env->nip);
> +            break;
>          case 65:
> -            return gdb_get_regl(mem_buf, env->msr);
> +            gdb_get_regl(mem_buf, env->msr);
> +            break;
>          case 66:
>              {
>                  uint32_t cr = 0;
> @@ -56,50 +111,55 @@ int ppc_cpu_gdb_read_register(CPUState *cs, uint8_t *mem_buf, int n)
>                  for (i = 0; i < 8; i++) {
>                      cr |= env->crf[i] << (32 - ((i + 1) * 4));
>                  }
> -                return gdb_get_reg32(mem_buf, cr);
> +                gdb_get_reg32(mem_buf, cr);
> +                break;
>              }
>          case 67:
> -            return gdb_get_regl(mem_buf, env->lr);
> +            gdb_get_regl(mem_buf, env->lr);
> +            break;
>          case 68:
> -            return gdb_get_regl(mem_buf, env->ctr);
> +            gdb_get_regl(mem_buf, env->ctr);
> +            break;
>          case 69:
> -            return gdb_get_regl(mem_buf, env->xer);
> +            gdb_get_regl(mem_buf, env->xer);
> +            break;
>          case 70:
> -            {
> -                if (gdb_has_xml) {
> -                    return 0;
> -                }
> -                return gdb_get_reg32(mem_buf, env->fpscr);
> -            }
> +            gdb_get_reg32(mem_buf, env->fpscr);
> +            break;
>          }
>      }
> -    return 0;
> +    if (msr_le) {
> +        ppc_cpu_gdb_swap_register(mem_buf, n, r);
> +    }
> +    return r;
>  }
>  
>  int ppc_cpu_gdb_write_register(CPUState *cs, uint8_t *mem_buf, int n)
>  {
>      PowerPCCPU *cpu = POWERPC_CPU(cs);
>      CPUPPCState *env = &cpu->env;
> +    int r = ppc_cpu_gdb_register_len(n);
>  
> +    if (!r) {
> +        return r;
> +    }
> +    if (msr_le) {
> +        ppc_cpu_gdb_swap_register(mem_buf, n, r);
> +    }
>      if (n < 32) {
>          /* gprs */
>          env->gpr[n] = ldtul_p(mem_buf);
> -        return sizeof(target_ulong);
>      } else if (n < 64) {
>          /* fprs */
> -        if (gdb_has_xml) {
> -            return 0;
> -        }
>          env->fpr[n-32] = ldfq_p(mem_buf);
> -        return 8;
>      } else {
>          switch (n) {
>          case 64:
>              env->nip = ldtul_p(mem_buf);
> -            return sizeof(target_ulong);
> +            break;
>          case 65:
>              ppc_store_msr(env, ldtul_p(mem_buf));
> -            return sizeof(target_ulong);
> +            break;
>          case 66:
>              {
>                  uint32_t cr = ldl_p(mem_buf);
> @@ -107,25 +167,22 @@ int ppc_cpu_gdb_write_register(CPUState *cs, uint8_t *mem_buf, int n)
>                  for (i = 0; i < 8; i++) {
>                      env->crf[i] = (cr >> (32 - ((i + 1) * 4))) & 0xF;
>                  }
> -                return 4;
> +                break;
>              }
>          case 67:
>              env->lr = ldtul_p(mem_buf);
> -            return sizeof(target_ulong);
> +            break;
>          case 68:
>              env->ctr = ldtul_p(mem_buf);
> -            return sizeof(target_ulong);
> +            break;
>          case 69:
>              env->xer = ldtul_p(mem_buf);
> -            return sizeof(target_ulong);
> +            break;
>          case 70:
>              /* fpscr */
> -            if (gdb_has_xml) {
> -                return 0;
> -            }
>              store_fpscr(env, ldtul_p(mem_buf), 0xffffffff);
> -            return sizeof(target_ulong);
> +            break;
>          }
>      }
> -    return 0;
> +    return r;
>  }
> 


-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg