[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:a0a2f5b811afe2e43d32336208babc2a27efbea3

commit a0a2f5b811afe2e43d32336208babc2a27efbea3
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Tue Jun 28 16:49:03 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index 2e5d06341cf..38cfc1ed6a7 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,24 @@
+==================== Work092, patch #10
+
+Make float128 test builtin's overloaded.
+
+2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000-builtins.def
+	(__builtin_vsx_scalar_test_data_class_qp_kf): Make rename builtin
+	to add _kf suffix.
+	(__builtin_vsx_scalar_test_neg_qp_kf): Likewise.
+	(__builtin_vsx_scalar_test_data_class_qp_tf): New builtin for IEEE
+	128-bit long double.
+	(__builtin_vsx_scalar_test_neg_qp_tf): Likewise.
+	* config/rs6000/rs6000-overload.def
+	(__builtin_vec_scalar_test_data_class): Add KF/TF overloads.
+	(__builtin_vec_scalar_test_neg): Likewise.
+	(__builtin_vec_scalar_test_neg_qp): Likewise.
+	(__builtin_vec_scalar_test_data_class_qp): Likewise.
+
 ==================== Work092, patch #9
 
 Add builtin ieeeld for long double is IEEE 128-bit.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:aa7aacfaa1525d19d9952340249f93754fea5129

commit aa7aacfaa1525d19d9952340249f93754fea5129
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Tue Jun 28 18:27:47 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index d1fae6ff79e..a3db44b3560 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,18 @@
+==================== Work092, patch #13
+
+Fix typos.
+
+2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000-builtins.def
+	(__builtin_vsx_scalar_extract_expq_tf): Fix argument to be long
+	double, add ieeeld condition.
+	(__builtin_vsx_scalar_extract_sigq_tf): Likewise.
+	(__builtin_vsx_scalar_test_data_class_qp_tf): Add ieeeld condition.
+	(__builtin_vsx_scalar_test_neg_qp): Likewise.
+
 ==================== Work092, patch #12
 
 Make float128 extract builtins overloaded.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:0fc27ccd4c465e76f87e0e7910dba39ae3380b1d

commit 0fc27ccd4c465e76f87e0e7910dba39ae3380b1d
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Tue Jun 28 18:10:30 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index 7dc86881975..d1fae6ff79e 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,22 @@
+==================== Work092, patch #12
+
+Make float128 extract builtins overloaded.
+
+2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000-builtins.def
+	(__builtin_vsx_scalar_extract_expq_kf): Rename builtin to add _kf
+	suffix.
+	(__builtin_vsx_scalar_extract_sigq_kf): Likewise.
+	(__builtin_vsx_scalar_extract_expq_tf): New builtin for IEEE
+	128-bit long double.
+	(__builtin_vsx_scalar_extract_sigq_kf): Likewise.
+	* config/rs6000/rs6000-overload.def
+	(__builtin_vec_scalar_extract): Add KF/TF overloads.
+	(__builtin_vec_scalar_extract_sig): Likewise.
+
 ==================== Work092, patch #11
 
 Remove support for renaming KF builtins to TF.
@@ -12,15 +31,15 @@ gcc/
 
 ==================== Work092, patch #10
 
-Make float128 test builtin's overloaded.
+Make float128 test builtins overloaded.
 
 2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
 
 gcc/
 
 	* config/rs6000/rs6000-builtins.def
-	(__builtin_vsx_scalar_test_data_class_qp_kf): Make rename builtin
-	to add _kf suffix.
+	(__builtin_vsx_scalar_test_data_class_qp_kf): Rename builtin to
+	add _kf suffix.
 	(__builtin_vsx_scalar_test_neg_qp_kf): Likewise.
 	(__builtin_vsx_scalar_test_data_class_qp_tf): New builtin for IEEE
 	128-bit long double.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:7e463b715440618e3903b0a69337dc8f534a5bb6

commit 7e463b715440618e3903b0a69337dc8f534a5bb6
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Tue Jun 28 16:53:53 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index 38cfc1ed6a7..7dc86881975 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,15 @@
+==================== Work092, patch #11
+
+Remove support for renaming KF builtins to TF.
+
+2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000-builtin.cc (rs6000_expand_builtin): Do not
+	convert various KFmode builtins to TFmode builtins if long double
+	is IEEE 128-bit.
+
 ==================== Work092, patch #10
 
 Make float128 test builtin's overloaded.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:8a87882eef9777d7b11785f30048d109ef2781e3

commit 8a87882eef9777d7b11785f30048d109ef2781e3
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Tue Jun 28 16:06:20 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index a918d6fa50c..2e5d06341cf 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,27 @@
+==================== Work092, patch #9
+
+Add builtin ieeeld for long double is IEEE 128-bit.
+
+2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000-builtin.cc (rs6000_expand_builtin): Add
+	check for ieeeld.
+	* config/rs6000/rs6000-builtins.def (ieeeld): New attribute.
+	(__builtin_addf128_round_to_odd_tf): Add ieeeld.
+	(__builtin_divf128_round_to_odd_tf): Likewise.
+	(__builtin_fmaf128_round_to_odd_tf): Likewise.
+	(__builtin_mulf128_round_to_odd_tf): Likewise.
+	(__builtin_sqrtf128_round_to_odd_tf): Likewise.
+	(__builtin_subf128_round_to_odd_tf): Likewise.
+	(__builtin_truncf128_round_to_odd_tf): Likewise.
+	* config/rs6000/rs6000-gen-builtins.c (ieeeld): New attribute.
+	(struct attrinfo): Add ieeeld.
+	(parse_bif_attrs): Likewise.
+	(write_decls): Likewise.
+	(write_bif_static_init): Likewise.
+
 ==================== Work092, patch #8
 
 Use overloading for ODD KF/TF functions.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:80de8201abc32b688885843bb365cdd050a2c9cc

commit 80de8201abc32b688885843bb365cdd050a2c9cc
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Tue Jun 28 01:19:32 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index b3c278b4bf9..a918d6fa50c 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,29 @@
+==================== Work092, patch #8
+
+Use overloading for ODD KF/TF functions.
+
+2022-06-28   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000-builtin.cc (rs6000_invalid_builtin): Add
+	support for ieee128-hw-ld.
+	(rs6000_builtin_is_supported): Likewise.
+	(rs6000_init_builtins): Fix comment from previous change.  Add
+	support of ieee128-hd-ld.
+	(rs6000_expand_builtin): Add support for ieee128-hw-ld.
+	* config/rs6000/rs6000-builtins.def (IEEE 128-bit odd builtins):
+	Move the IEEE 128-bit odd builtins to being overloaded.
+	* config/rs6000/rs6000-gen-builtins.cc (bif_stanza): Add
+	stanza for IEEE 128-bit floating point where long double is IEEE
+	128-bit.
+	(stanza_map): Likewise.
+	(enable_string): Likewise.
+	(write_decls): Likewise.
+	* config/rs6000/rs6000-overload.def (IEEE 128-bit odd builtins):
+	Add the various IEEE 128-bit odd built-in functions for both
+	KFmode and TFmode.
+
 ==================== Work092, patch #7
 
 Only change IEEE 128-bit built-ins if passed IEEE 128-bit long double.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:6da8385a64afaec6266d48c3cc4ab514566fe2fd

commit 6da8385a64afaec6266d48c3cc4ab514566fe2fd
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Mon Jun 27 17:30:31 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-27   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index 721739a86f1..b3c278b4bf9 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,15 @@
+==================== Work092, patch #7
+
+Only change IEEE 128-bit built-ins if passed IEEE 128-bit long double.
+
+2022-06-27   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000-builtin.cc (rs6000_expand_builtin): Restore
+	code from original source, but do the conversion only if the
+	argument is a long double that is IEEE 128-bit.
+
 ==================== Work092, patch #6
 
 Allow __ibm128 on systems that support 128-bit long double.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:472f1b3a5c82418a26f0983ec2ea4bd7a51a7792

commit 472f1b3a5c82418a26f0983ec2ea4bd7a51a7792
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Mon Jun 27 14:08:08 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-27   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index a189a9a699d..721739a86f1 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,20 @@
+==================== Work092, patch #6
+
+Allow __ibm128 on systems that support 128-bit long double.
+
+2022-06-27   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000.cc (init_float128_ibm): Allow -msoft-float
+	with __ibm128.
+	(rs6000_scalar_mode_supported_p): Allow __ibm128 on systems that
+	support 128-bit long double.
+	* config/rs6000/rs6000.h (FLOAT128_IBM_P): Allow -msoft-float with
+	__ibm128.
+	* config/rs6000/rs6000.md (unpack<mode>_dm): Likewise.
+	(unpack<mode>_nodm): Likewise.
+
 ==================== Work092, patch #5
 
 Remove code that depends on __float128 or __ibm128 using the long double type.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:404b66b0a118e414abe977f5cb61249ed971ae1a

commit 404b66b0a118e414abe977f5cb61249ed971ae1a
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Fri Jun 24 17:27:23 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-24   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index 3cbfd884cbf..a189a9a699d 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,16 @@
+==================== Work092, patch #5
+
+Remove code that depends on __float128 or __ibm128 using the long double type.
+
+2022-06-24   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000-builtin.cc (rs6000_expand_builtin): Remove
+	code that assumed __float128 and __ibm128 used the long double
+	type if long double uses 128-bit IEEE or IBM encodings
+	respectively.
+
 ==================== Work092, patch #4
 
 Add __float128 to __ibm128 conversations.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:0a33843a2e80750ca6c72bef6b8c097c20732e3d

commit 0a33843a2e80750ca6c72bef6b8c097c20732e3d
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Fri Jun 24 11:50:32 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-24   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index 6ea7694d01d..3cbfd884cbf 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,18 @@
+==================== Work092, patch #4
+
+Add __float128 to __ibm128 conversations.
+
+In the past, GCC used the long double type for __ibm128 if long double
+used the IBM 128-bit encoding.  Now that we always use a separate internal
+type for __ibm128, it exposed that some conversions were missing.
+
+2022-06-24   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000.md (extendkfif2): New insn.
+	(trunckfif2): Likewise.
+
 ==================== Work092, patch #3
 
 Fix IEEE 128-bit long double Complex multiply and divide tests.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:343688c2a5e8af39634334bca61d695d1df0233d

commit 343688c2a5e8af39634334bca61d695d1df0233d
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Thu Jun 23 14:54:02 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-23   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index 76d9bb695d9..6ea7694d01d 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,14 @@
+==================== Work092, patch #3
+
+Fix IEEE 128-bit long double Complex multiply and divide tests.
+
+2022-06-23   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/testsuite/
+
+	* gcc.target/powerpc/mulkd3-2.c: Update test.
+	* gcc.target/powerpc/divkd3-2.c: Likewise.
+
 ==================== Work092, patch #2
 
 Iterate on getting long double complex multiply and divide working.

[gcc(refs/users/meissner/heads/work092)] Update ChangeLog.meissner.

[Michael Meissner]

https://gcc.gnu.org/g:035103b88ab819c60160f42fd1b8bda4fae67040

commit 035103b88ab819c60160f42fd1b8bda4fae67040
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Wed Jun 22 17:43:22 2022 -0400

    Update ChangeLog.meissner.
    
    2022-06-22   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index 7108cf4f921..be6ca55473e 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,32 @@
+==================== Work092, patch #1
+
+__float128 and __ibm128 always gets different internal types.
+
+2022-06-22   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+	* config/rs6000/rs6000-builtin.cc (rs6000_init_builtins): Always
+	create a new tree node for __ibm128, even if long double uses the
+	IBM 128-bit format.  Always create a new tree node for __float128,
+	even if long double uses the IEEE 128-bit format.  Change the
+	keyword create for IEEE 128-bit to be '__float128', not
+	'__ieee128'.
+	* config/rs6000/rs6000-c.cc (rs6000_target_modify_macros): Define
+	__ieee128 compatibility macro.
+	* config/rs6000/rs600.cc (reg_offset_addressing_ok_p): Allow
+	IFmode to use offset addresses.
+	(create_complex_muldiv): Delete.
+	(init_float128_ieee): Do not create __mulkc3 and __divkc3 since
+	__float128 always creates these functions.
+	(rs6000_libgcc_floating_mode_supported_p): Always allow KFmode.
+	(rs6000_floatn_mode): Always return KFmode for IEEE 128-bit.  Do
+	not return TFmode, even if long double uses the IEEE 128-bit
+	format.
+	(rs6000_c_mode_for_suffix): The 'q' suffix always uses KFmode,
+	even if long double uses the IEEE 128-bit encoding.
+
+==================== Branch creation
+
 2022-06-22   Michael Meissner  <meissner@linux.ibm.com>
 
 	Clone branch