Do not dump literal pool when a .section .line directive is encountered.
This commit is contained in:
Nick Clifton
parent
c5089c8475
commit
bcef92fa23
@ -1,3 +1,8 @@
|
||||
1999-09-02 Nick Clifton <nickc@cygnus.com>
|
||||
|
||||
* config/tc-mcore.c (mcore_s_section): Do not dump literals if a
|
||||
.section .line directive is encountered.
|
||||
|
||||
1999-09-01 Nick Clifton <nickc@cygnus.com>
|
||||
|
||||
* config/tc-arm.c (md_section_align): Do not align sections in ELF
|
||||
|
||||
@ -40,6 +40,7 @@ static void mcore_s_literals PARAMS ((int));
|
||||
static void mcore_cons PARAMS ((int));
|
||||
static void mcore_float_cons PARAMS ((int));
|
||||
static void mcore_stringer PARAMS ((int));
|
||||
static void mcore_fill PARAMS ((int));
|
||||
static int log2 PARAMS ((unsigned int));
|
||||
static char * parse_reg PARAMS ((char *, unsigned *));
|
||||
static char * parse_creg PARAMS ((char *, unsigned *));
|
||||
@ -178,11 +179,11 @@ const pseudo_typeS md_pseudo_table[] =
|
||||
{ "byte", mcore_cons, 1 },
|
||||
{ "dc", mcore_cons, 2 },
|
||||
{ "dc.b", mcore_cons, 1 },
|
||||
{ "dc.d", mcore_float_cons, 'd' },
|
||||
{ "dc.d", mcore_float_cons, 'd'},
|
||||
{ "dc.l", mcore_cons, 4 },
|
||||
{ "dc.s", mcore_float_cons, 'f' },
|
||||
{ "dc.s", mcore_float_cons, 'f'},
|
||||
{ "dc.w", mcore_cons, 2 },
|
||||
{ "dc.x", mcore_float_cons, 'x' },
|
||||
{ "dc.x", mcore_float_cons, 'x'},
|
||||
{ "double", mcore_float_cons, 'd'},
|
||||
{ "float", mcore_float_cons, 'f'},
|
||||
{ "hword", mcore_cons, 2 },
|
||||
@ -194,6 +195,7 @@ const pseudo_typeS md_pseudo_table[] =
|
||||
{ "single", mcore_float_cons, 'f'},
|
||||
{ "string", mcore_stringer, 1 },
|
||||
{ "word", mcore_cons, 2 },
|
||||
{ "fill", mcore_fill, 0 },
|
||||
|
||||
/* Allow for the effect of section changes. */
|
||||
{ "text", mcore_s_text, 0 },
|
||||
@ -307,6 +309,40 @@ mcore_stringer (append_zero)
|
||||
check_literals (2, 0);
|
||||
}
|
||||
|
||||
static void
|
||||
mcore_fill (unused)
|
||||
int unused;
|
||||
{
|
||||
if (now_seg == text_section)
|
||||
{
|
||||
char * str = input_line_pointer;
|
||||
int size = 1;
|
||||
int repeat;
|
||||
|
||||
repeat = atoi (str);
|
||||
|
||||
/* Look to see if a size has been specified. */
|
||||
while (*str != '\n' && *str != 0 && *str != ',')
|
||||
++ str;
|
||||
|
||||
if (* str == ',')
|
||||
{
|
||||
size = atoi (str + 1);
|
||||
|
||||
if (size > 8)
|
||||
size = 8;
|
||||
else if (size < 0)
|
||||
size = 0;
|
||||
}
|
||||
|
||||
poolspan += size * repeat;
|
||||
}
|
||||
|
||||
s_fill (unused);
|
||||
|
||||
check_literals (2, 0);
|
||||
}
|
||||
|
||||
/* Handle the section changing pseudo-ops. These call through to the
|
||||
normal implementations, but they dump the literal pool first. */
|
||||
|
||||
@ -340,7 +376,22 @@ static void
|
||||
mcore_s_section (ignore)
|
||||
int ignore;
|
||||
{
|
||||
dump_literals (0);
|
||||
/* Scan forwards to find the name of the section. If the section
|
||||
being switched to is ".line" then this is a DWARF1 debug section
|
||||
which is arbitarily placed inside generated code. In this case
|
||||
do not dump the literal pool because it is a) inefficient and
|
||||
b) would require the generation of extra code to jump around the
|
||||
pool. */
|
||||
char * ilp = input_line_pointer;
|
||||
|
||||
while (*ilp != 0 && isspace(*ilp))
|
||||
++ ilp;
|
||||
|
||||
if (strncmp (ilp, ".line", 5) == 0
|
||||
&& (isspace (ilp[5]) || *ilp == '\n' || *ilp == '\r'))
|
||||
;
|
||||
else
|
||||
dump_literals (0);
|
||||
|
||||
#ifdef OBJ_ELF
|
||||
obj_elf_section (ignore);
|
||||
@ -371,7 +422,7 @@ mcore_s_comm (needs_align)
|
||||
#endif
|
||||
|
||||
/* This function is called once, at assembler startup time. This should
|
||||
set up all the tables, etc that the MD part of the assembler needs. */
|
||||
set up all the tables, etc that the MD part of the assembler needs. */
|
||||
void
|
||||
md_begin ()
|
||||
{
|
||||
@ -1466,7 +1517,7 @@ md_assemble (str)
|
||||
case JL:
|
||||
inst = MCORE_INST_JSRI; /* jsri */
|
||||
input_line_pointer = parse_rt (op_end + 1, & output, 1, & e);
|
||||
/* parse_rt() calls frag_more for us */
|
||||
/* parse_rt() calls frag_more for us. */
|
||||
|
||||
/* Only do this if we know how to do it ... */
|
||||
if (e.X_op != O_absent && do_jsri2bsr)
|
||||
@ -1725,7 +1776,7 @@ md_convert_frag (abfd, sec, fragP)
|
||||
case C (COND_JUMP, COND12):
|
||||
case C (UNCD_JUMP, UNCD12):
|
||||
{
|
||||
/* Get the address of the end of the instruction */
|
||||
/* Get the address of the end of the instruction. */
|
||||
int next_inst = fragP->fr_fix + fragP->fr_address + 2;
|
||||
unsigned char t0;
|
||||
int disp = targ_addr - next_inst;
|
||||
@ -1847,7 +1898,7 @@ md_convert_frag (abfd, sec, fragP)
|
||||
buffer[6] = 0;
|
||||
buffer[7] = 0;
|
||||
|
||||
/* Make reloc for the long disp */
|
||||
/* Make reloc for the long disp. */
|
||||
fix_new (fragP, fragP->fr_fix + 4, 4,
|
||||
fragP->fr_symbol, fragP->fr_offset, 0, BFD_RELOC_32);
|
||||
|
||||
@ -1861,7 +1912,7 @@ md_convert_frag (abfd, sec, fragP)
|
||||
buffer[4] = 0;
|
||||
buffer[5] = 0;
|
||||
|
||||
/* Make reloc for the long disp */
|
||||
/* Make reloc for the long disp. */
|
||||
fix_new (fragP, fragP->fr_fix + 2, 4,
|
||||
fragP->fr_symbol, fragP->fr_offset, 0, BFD_RELOC_32);
|
||||
fragP->fr_fix += U32_LEN;
|
||||
@ -2071,7 +2122,7 @@ md_estimate_size_before_relax (fragP, segment_type)
|
||||
return fragP->fr_var;
|
||||
}
|
||||
|
||||
/* Put number into target byte order */
|
||||
/* Put number into target byte order. */
|
||||
void
|
||||
md_number_to_chars (ptr, use, nbytes)
|
||||
char * ptr;
|
||||
|
||||
Loading…
Reference in New Issue
Block a user