[PATCH 3/11] Add MIPS_MAX_REGISTER_SIZE
Yao Qi
qiyaoltc@gmail.com
Tue Apr 11 15:37:00 GMT 2017
Alan Hayward <Alan.Hayward@arm.com> writes:
Hi Alan,
There are different ways of getting rid of MAX_REGISTER_SIZE, let us try
some simple approaches first. Some uses of MAX_REGISTER_SIZE still
can't be removed, but let us start from easy part.
> diff --git a/gdb/mips-linux-tdep.c b/gdb/mips-linux-tdep.c
> index 57e75b5343e1b927e9fe28dea16759f769cf4506..ce2f378854f4b66c426fd9d6683831e8795c58a6 100644
> --- a/gdb/mips-linux-tdep.c
> +++ b/gdb/mips-linux-tdep.c
> @@ -118,7 +118,7 @@ supply_32bit_reg (struct regcache *regcache, int regnum, const void *addr)
> {
> struct gdbarch *gdbarch = get_regcache_arch (regcache);
> enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
> - gdb_byte buf[MAX_REGISTER_SIZE];
> + gdb_byte buf[MIPS_MAX_REGISTER_SIZE];
Given the function name supply_32bit_reg, all registers are 32-bit here,
so we can do "buf[4]"?
> @@ -470,7 +470,7 @@ mips64_fill_gregset (const struct regcache *regcache,
>
> if (regaddr != -1)
> {
> - gdb_byte buf[MAX_REGISTER_SIZE];
> + gdb_byte buf[MIPS_MAX_REGISTER_SIZE];
> LONGEST val;
>
> regcache_raw_collect (regcache, regno, buf);
The "buf" is used by regcache_raw_collect + extract_signed_integer,
regcache_raw_collect (regcache, regno, buf);
val = extract_signed_integer (buf, register_size (gdbarch, regno),
byte_order);
this pattern exists in some places, so can we add a new regcache api,
LONGEST
regcache_raw_collect_signed (const struct regcache *regcache, int regnum)
{
struct gdbarch *gdbarch = get_regcache_arch (regcache);
enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
return extract_signed_integer (register_buffer (regcache, regnum),
register_size (gdbarch, regnum),
byte_order);
}
this will remove one memory copy, as well as the local buffer "buf".
> diff --git a/gdb/mips-tdep.c b/gdb/mips-tdep.c
> index 41cb9d82c6ef473c1fbbf86601914f9a4f462411..51a22ba29a520639bdeb95c235c00c74ad40435b 100644
> --- a/gdb/mips-tdep.c
> +++ b/gdb/mips-tdep.c
> @@ -4527,7 +4527,7 @@ mips_eabi_push_dummy_call (struct gdbarch *gdbarch, struct value *function,
> for (argnum = 0; argnum < nargs; argnum++)
> {
> const gdb_byte *val;
> - gdb_byte valbuf[MAX_REGISTER_SIZE];
> + gdb_byte valbuf[MIPS_MAX_REGISTER_SIZE];
> struct value *arg = args[argnum];
> struct type *arg_type = check_typedef (value_type (arg));
> int len = TYPE_LENGTH (arg_type);
> @@ -5758,7 +5758,7 @@ mips_o32_return_value (struct gdbarch *gdbarch, struct value *function,
> /* A struct that contains one or two floats. Each value is part
> in the least significant part of their floating point
> register.. */
> - gdb_byte reg[MAX_REGISTER_SIZE];
> + gdb_byte reg[MIPS_MAX_REGISTER_SIZE];
"reg" is not used at all, we can remove it.
> int regnum;
> int field;
> for (field = 0, regnum = mips_regnum (gdbarch)->fp0;
> @@ -6473,7 +6473,7 @@ print_gp_register_row (struct ui_file *file, struct frame_info *frame,
> {
> struct gdbarch *gdbarch = get_frame_arch (frame);
> /* Do values for GP (int) regs. */
> - gdb_byte raw_buffer[MAX_REGISTER_SIZE];
> + gdb_byte raw_buffer[MIPS_MAX_REGISTER_SIZE];
> int ncols = (mips_abi_regsize (gdbarch) == 8 ? 4 : 8); /* display cols
> per row. */
"raw_buffer" is used in deprecated_frame_register_read, so can we use
get_frame_register_value instead? so that "raw_buffer" is not needed.
--
Yao (齐尧)
More information about the Gdb-patches
mailing list