* [Patch] New ia64 @slotcount pseudo func (for VMS)
@ 2010-02-10 13:24 Tristan Gingold
2010-02-12 11:42 ` Nick Clifton
0 siblings, 1 reply; 6+ messages in thread
From: Tristan Gingold @ 2010-02-10 13:24 UTC (permalink / raw)
To: binutils; +Cc: Douglas B Rupp
Hi,
with Doug we finally wrote this patch to implement @slotcount pseudo function for ia64 VMS. This pseudo
is available only on VMS to avoid pseudo namespace pollution.
There is also a testcase.
No regressions for ia64 linux.
Tristan.
gas/
2010-02-10 Tristan Gingold <gingold@adacore.com>
Douglas B Rupp <rupp@gnat.com>
* config/tc-ia64.c (enum reloc_func): Add FUNC_SLOTCOUNT_RELOC.
(DUMMY_RELOC_IA64_SLOTCOUNT): Added.
(pseudo_func): Add an entry for slotcount.
(md_begin): Initialize slotcount pseudo symbol.
(ia64_parse_name): Handle @slotcount parameter.
(ia64_gen_real_reloc_type): Handle slotcount.
(md_apply_fix): Ditto.
gas/testsuite
2010-02-10 Tristan Gingold <gingold@adacore.com>
* gas/ia64/slotcount.s, gas/ia64/slotcount.s: New test.
* gas/ia64/ia64.exp: Add slotcount test (vms only).
diff --git a/gas/config/tc-ia64.c b/gas/config/tc-ia64.c
index e4563a7..73e31aa 100644
--- a/gas/config/tc-ia64.c
+++ b/gas/config/tc-ia64.c
@@ -103,6 +103,9 @@ enum reloc_func
FUNC_LT_DTP_RELATIVE,
FUNC_LT_TP_RELATIVE,
FUNC_IPLT_RELOC,
+#ifdef TE_VMS
+ FUNC_SLOTCOUNT_RELOC,
+#endif
};
enum reg_symbol
@@ -163,6 +166,11 @@ struct label_fix
bfd_boolean dw2_mark_labels;
};
+#ifdef TE_VMS
+/* An internally used relocation. */
+#define DUMMY_RELOC_IA64_SLOTCOUNT (BFD_RELOC_UNUSED + 1)
+#endif
+
/* This is the endianness of the current section. */
extern int target_big_endian;
@@ -575,6 +583,9 @@ pseudo_func[] =
{ NULL, 0, { 0 } }, /* placeholder for FUNC_LT_DTP_RELATIVE */
{ NULL, 0, { 0 } }, /* placeholder for FUNC_LT_TP_RELATIVE */
{ "iplt", PSEUDO_FUNC_RELOC, { 0 } },
+#ifdef TE_VMS
+ { "slotcount", PSEUDO_FUNC_RELOC, { 0 } },
+#endif
/* mbtype4 constants: */
{ "alt", PSEUDO_FUNC_CONST, { 0xa } },
@@ -7158,6 +7169,12 @@ md_begin (void)
symbol_new (".<iplt>", undefined_section, FUNC_IPLT_RELOC,
&zero_address_frag);
+#ifdef TE_VMS
+ pseudo_func[FUNC_SLOTCOUNT_RELOC].u.sym =
+ symbol_new (".<slotcount>", undefined_section, FUNC_SLOTCOUNT_RELOC,
+ &zero_address_frag);
+#endif
+
if (md.tune != itanium1)
{
/* Convert MFI NOPs bundles into MMI NOPs bundles. */
@@ -7744,6 +7761,15 @@ ia64_parse_name (char *name, expressionS *e, char *nextcharP)
}
/* Skip ')'. */
++input_line_pointer;
+#ifdef TE_VMS
+ if (idx == FUNC_SLOTCOUNT_RELOC)
+ {
+ /* @slotcount can accept any expression. Canonicalize. */
+ e->X_add_symbol = make_expr_symbol (e);
+ e->X_op = O_symbol;
+ e->X_add_number = 0;
+ }
+#endif
if (e->X_op != O_symbol)
{
if (e->X_op != O_pseudo_fixup)
@@ -11109,6 +11135,11 @@ ia64_gen_real_reloc_type (struct symbol *sym, bfd_reloc_code_real_type r_type)
}
break;
+#ifdef TE_VMS
+ case FUNC_SLOTCOUNT_RELOC:
+ return DUMMY_RELOC_IA64_SLOTCOUNT;
+#endif
+
default:
abort ();
}
@@ -11278,12 +11309,48 @@ md_apply_fix (fixS *fix, valueT *valP, segT seg ATTRIBUTE_UNUSED)
S_SET_THREAD_LOCAL (fix->fx_addsy);
break;
+#ifdef TE_VMS
+ case DUMMY_RELOC_IA64_SLOTCOUNT:
+ as_bad_where (fix->fx_file, fix->fx_line,
+ _("cannot resolve @slotcount parameter"));
+ fix->fx_done = 1;
+ return;
+#endif
+
default:
break;
}
}
else if (fix->tc_fix_data.opnd == IA64_OPND_NIL)
{
+#ifdef TE_VMS
+ if (fix->fx_r_type == DUMMY_RELOC_IA64_SLOTCOUNT)
+ {
+ /* For @slotcount, convert an addresses difference to a slots
+ difference. */
+ valueT v;
+
+ v = (value >> 4) * 3;
+ switch (value & 0x0f)
+ {
+ case 0:
+ case 1:
+ case 2:
+ v += value & 0x0f;
+ break;
+ case 0x0f:
+ v += 2;
+ break;
+ case 0x0e:
+ v += 1;
+ break;
+ default:
+ as_bad (_("invalid @slotcount value"));
+ }
+ value = v;
+ }
+#endif
+
if (fix->tc_fix_data.bigendian)
number_to_chars_bigendian (fixpos, value, fix->fx_size);
else
diff --git a/gas/testsuite/gas/ia64/ia64.exp b/gas/testsuite/gas/ia64/ia64.exp
index 3dfd10b..221b8d4 100644
--- a/gas/testsuite/gas/ia64/ia64.exp
+++ b/gas/testsuite/gas/ia64/ia64.exp
@@ -89,4 +89,9 @@ if [istarget "ia64-*"] then {
run_dump_test "operand-or"
run_list_test "hint.b-err" ""
run_list_test "hint.b-warn" "-mhint.b=warning"
+
+ if [istarget "ia64-*-*vms*"] then {
+ run_dump_test "slotcount"
+ }
+
}
diff --git a/gas/testsuite/gas/ia64/slotcount.d b/gas/testsuite/gas/ia64/slotcount.d
new file mode 100644
index 0000000..3548266
--- /dev/null
+++ b/gas/testsuite/gas/ia64/slotcount.d
@@ -0,0 +1,10 @@
+#objdump: -s -j .slot_test
+#name: ia64 slotcount
+
+.*: +file format .*
+
+Contents of section .slot_test:
+ 0000 04000000 01000000 02000000 03000000 ................
+ 0010 04000000 05000000 06000000 07000000 ................
+ 0020 08000000 02000000 06000000 03000000 ................
+ 0030 02000000 ....
diff --git a/gas/testsuite/gas/ia64/slotcount.s b/gas/testsuite/gas/ia64/slotcount.s
new file mode 100644
index 0000000..bdbc318
--- /dev/null
+++ b/gas/testsuite/gas/ia64/slotcount.s
@@ -0,0 +1,51 @@
+ .section .slot_test0,"",@progbits
+ data4.ua @slotcount(.L1-.L0)
+
+ .text
+ .align 16
+foo:
+[.L0:]
+ mov r2 = r12
+[.L1:]
+ mov r8 = r14
+[.L2:]
+ ;;
+ mov r12 = r2
+[.L3:]
+ {
+ .mii
+ nop 0
+[.L4:]
+ nop 0
+[.L5:]
+ nop 0
+ }
+ {
+[.L6:]
+ nop 0
+[.L7:]
+ nop 0
+[.L8:]
+ br.ret.sptk.many b0
+ ;;
+ }
+
+ .section .slot_test,"",@progbits
+// data4.ua @slotcount(.Lundef)
+
+ data4.ua @slotcount(17)
+
+ data4.ua @slotcount(.L1-.L0) // 1
+ data4.ua @slotcount(.L2-.L0) // 2
+ data4.ua @slotcount(.L3-.L0) // 3
+ data4.ua @slotcount(.L4-.L0) // 4
+ data4.ua @slotcount(.L5-.L0) // 5
+ data4.ua @slotcount(.L6-.L0) // 6
+ data4.ua @slotcount(.L7-.L0) // 7
+ data4.ua @slotcount(.L8-.L0) // 8
+
+ data4.ua @slotcount(.L3-.L1) // 2
+ data4.ua @slotcount(.L8-.L2) // 6
+ data4.ua @slotcount(.L4-.L1) // 3
+ data4.ua @slotcount(.L4-.L2) // 2
+// data4.ua @slotcount(.L2-.Lundef)
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Patch] New ia64 @slotcount pseudo func (for VMS)
2010-02-10 13:24 [Patch] New ia64 @slotcount pseudo func (for VMS) Tristan Gingold
@ 2010-02-12 11:42 ` Nick Clifton
2010-02-12 13:44 ` Tristan Gingold
2010-02-12 14:01 ` Tristan Gingold
0 siblings, 2 replies; 6+ messages in thread
From: Nick Clifton @ 2010-02-12 11:42 UTC (permalink / raw)
To: Tristan Gingold; +Cc: binutils, Douglas B Rupp
Hi Tristan,
> with Doug we finally wrote this patch to implement @slotcount pseudo function for ia64 VMS. This pseudo
> is available only on VMS to avoid pseudo namespace pollution.
I think that this really ought to have some documentation written for it
in gas/doc/c-ia64.texi...
> gas/
> 2010-02-10 Tristan Gingold<gingold@adacore.com>
> Douglas B Rupp<rupp@gnat.com>
>
> * config/tc-ia64.c (enum reloc_func): Add FUNC_SLOTCOUNT_RELOC.
> (DUMMY_RELOC_IA64_SLOTCOUNT): Added.
> (pseudo_func): Add an entry for slotcount.
> (md_begin): Initialize slotcount pseudo symbol.
> (ia64_parse_name): Handle @slotcount parameter.
> (ia64_gen_real_reloc_type): Handle slotcount.
> (md_apply_fix): Ditto.
>
> gas/testsuite
> 2010-02-10 Tristan Gingold<gingold@adacore.com>
>
> * gas/ia64/slotcount.s, gas/ia64/slotcount.s: New test.
> * gas/ia64/ia64.exp: Add slotcount test (vms only).
Approved - please apply - but please do consider adding some
documentation as well.
Cheers
Nick
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Patch] New ia64 @slotcount pseudo func (for VMS)
2010-02-12 11:42 ` Nick Clifton
@ 2010-02-12 13:44 ` Tristan Gingold
2010-02-12 14:01 ` Tristan Gingold
1 sibling, 0 replies; 6+ messages in thread
From: Tristan Gingold @ 2010-02-12 13:44 UTC (permalink / raw)
To: Nick Clifton; +Cc: binutils, Douglas B Rupp
On Feb 12, 2010, at 12:41 PM, Nick Clifton wrote:
> Hi Tristan,
>
>> with Doug we finally wrote this patch to implement @slotcount pseudo function for ia64 VMS. This pseudo
>> is available only on VMS to avoid pseudo namespace pollution.
>
> I think that this really ought to have some documentation written for it in gas/doc/c-ia64.texi...
Ok, will write it before committing.
Thanks,
Tristan.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Patch] New ia64 @slotcount pseudo func (for VMS)
2010-02-12 11:42 ` Nick Clifton
2010-02-12 13:44 ` Tristan Gingold
@ 2010-02-12 14:01 ` Tristan Gingold
2010-02-12 14:10 ` Nick Clifton
1 sibling, 1 reply; 6+ messages in thread
From: Tristan Gingold @ 2010-02-12 14:01 UTC (permalink / raw)
To: Nick Clifton; +Cc: binutils, Douglas B Rupp
On Feb 12, 2010, at 12:41 PM, Nick Clifton wrote:
> Hi Tristan,
>
>> with Doug we finally wrote this patch to implement @slotcount pseudo function for ia64 VMS. This pseudo
>> is available only on VMS to avoid pseudo namespace pollution.
>
> I think that this really ought to have some documentation written for it in gas/doc/c-ia64.texi...
What about this ?
(checked with make info).
Tristan.
gas/
2010-02-12 Tristan Gingold <gingold@adacore.com>
* doc/c-ia64.texi (IA-64-Relocs): Document @slotcount.
diff --git a/gas/doc/c-ia64.texi b/gas/doc/c-ia64.texi
index 6b0f3a9..f684171 100644
--- a/gas/doc/c-ia64.texi
+++ b/gas/doc/c-ia64.texi
@@ -120,7 +120,7 @@ Reference Guide.
* IA-64-Chars:: Special Characters
* IA-64-Regs:: Register Names
* IA-64-Bits:: Bit Names
-@c * IA-64-Relocs:: Relocations // to be written
@node IA-64-Chars
@@ -170,6 +170,21 @@ the @samp{ssm}/@samp{sum} and @samp{rsm}/@samp{rum}
instructions, but they can be used anywhere else where an integer
constant is expected.
+@node IA-64-Relocs
+@subsection Relocations
+@cindex IA-64 relocations
+
+In addition to the standard IA-64 relocations, the following relocations are
+implemented by @code{@value{AS}}:
+
+@table @code
+@item @@slotcount(@var{V})
+Convert the address offset @var{V} into a slot count. This pseudo
+function is available only on VMS. The expression @var{V} must be
+known at assembly time: it can't reference undefined symbols or symbols in
+different sections.
+@end table
+
@node IA-64 Opcodes
@section Opcodes
For detailed information on the IA-64 machine instruction set, see the
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Patch] New ia64 @slotcount pseudo func (for VMS)
2010-02-12 14:01 ` Tristan Gingold
@ 2010-02-12 14:10 ` Nick Clifton
2010-02-12 14:35 ` Tristan Gingold
0 siblings, 1 reply; 6+ messages in thread
From: Nick Clifton @ 2010-02-12 14:10 UTC (permalink / raw)
To: Tristan Gingold; +Cc: binutils, Douglas B Rupp
Hi Tristan,
> What about this ?
Great - thanks for writing this.
> gas/
> 2010-02-12 Tristan Gingold<gingold@adacore.com>
>
> * doc/c-ia64.texi (IA-64-Relocs): Document @slotcount.
Approved - please apply.
Cheers
Nick
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Patch] New ia64 @slotcount pseudo func (for VMS)
2010-02-12 14:10 ` Nick Clifton
@ 2010-02-12 14:35 ` Tristan Gingold
0 siblings, 0 replies; 6+ messages in thread
From: Tristan Gingold @ 2010-02-12 14:35 UTC (permalink / raw)
To: Nick Clifton; +Cc: binutils, Douglas B Rupp
On Feb 12, 2010, at 3:09 PM, Nick Clifton wrote:
> Hi Tristan,
>
>> What about this ?
>
> Great - thanks for writing this.
>
>> gas/
>> 2010-02-12 Tristan Gingold<gingold@adacore.com>
>>
>> * doc/c-ia64.texi (IA-64-Relocs): Document @slotcount.
>
> Approved - please apply.
Thanks. The whole patch set has been committed.
Tristan.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2010-02-12 14:35 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-02-10 13:24 [Patch] New ia64 @slotcount pseudo func (for VMS) Tristan Gingold
2010-02-12 11:42 ` Nick Clifton
2010-02-12 13:44 ` Tristan Gingold
2010-02-12 14:01 ` Tristan Gingold
2010-02-12 14:10 ` Nick Clifton
2010-02-12 14:35 ` Tristan Gingold
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).