Sun Mar 27 00:07:20 2016 UTC ()
- figured out the ldap/ssh problem.  was miscompiled crtbeginS.o which
  has been worked around.
- update some testing info.
- merge or1k/riscv descriptions that are actually the same problem.


(mrg)
diff -r1.14 -r1.15 src/external/gpl3/gcc/README.gcc53
cvs diff -r1.14 -r1.15 src/external/gpl3/gcc/Attic/README.gcc53 (expand / switch to unified diff)

--- src/external/gpl3/gcc/Attic/README.gcc53 2016/03/26 09:01:20 1.14
+++ src/external/gpl3/gcc/Attic/README.gcc53 2016/03/27 00:07:20 1.15
@@ -1,14 +1,14 @@ @@ -1,14 +1,14 @@
1$NetBSD: README.gcc53,v 1.14 2016/03/26 09:01:20 mrg Exp $ 1$NetBSD: README.gcc53,v 1.15 2016/03/27 00:07:20 mrg Exp $
2 2
3lib: 3lib:
4 libasan is disabled currently (haven't tried yet) 4 libasan is disabled currently (haven't tried yet)
5 5
6new stuff: 6new stuff:
7 cc1objcplus 7 cc1objcplus
8 collect2 -- need to re-introduce? 8 collect2 -- need to re-introduce?
9 libcc1 9 libcc1
10 libcilkrts 10 libcilkrts
11 libmpx 11 libmpx
12 liboffloadmic 12 liboffloadmic
13 libquadmath 13 libquadmath
14 libvtv 14 libvtv
@@ -59,61 +59,54 @@ m68000 y ? y y y @@ -59,61 +59,54 @@ m68000 y ? y y y
59m68k y ? y y y[13] 59m68k y ? y y y[13]
60mipseb y ? y y y 60mipseb y ? y y y
61mipsel y y y y y y 61mipsel y y y y y y
62mips64eb y ? n y n[8] 62mips64eb y ? n y n[8]
63mips64el y ? n y y[8] 63mips64el y ? n y y[8]
64powerpc y y y y y y[6]  64powerpc y y y y y y[6]
65sh3eb y ? y y y[5] 65sh3eb y ? y y y[5]
66sh3el y ? y y y 66sh3el y ? y y y
67sparc y y y y y y[6] 67sparc y y y y y y[6]
68sparc64 y y y y y y[6,2] y 68sparc64 y y y y y y[6,2] y
69vax y y y y y y[15] n 69vax y y y y y y[15] n
70x86_64 y y y y y y y 70x86_64 y y y y y y y
71-- 71--
72or1k n[11] 72or1k n[10]
73riscv32 n[10] 73riscv32 n[10]
74riscv64 n[10] 74riscv64 n[10]
75ia64 75ia64
76powerpc64 76powerpc64
77------------ ----- ------- ------ ---------- ------------ ---- --- 77------------ ----- ------- ------ ---------- ------------ ---- ---
78architecture tools kernels libgcc native-gcc make release runs atf 78architecture tools kernels libgcc native-gcc make release runs atf
79 79
80 80
81[1] -  81[1] -
82[2] - -O2 and -Os kernels hang, -O1 SIR reset 82[2] - -O2 and -Os kernels hang, -O1 SIR reset
83 -- may be a SMP issue; ultra10 works, ultra45 SMP does not 83 -- may be a SMP issue; ultra10 works, ultra45 SMP does not
84 -- try sb2000 (particularly with UP kernel. u45 UP is shitty.) 84 -- try sb2000 (particularly with UP kernel. u45 UP is shitty.)
85[3] - MKCOMPAT=no enabled currently, infact OABI seems to be broken. the docs say it was only deprecated, but the options make broken crap. 85[3] - MKCOMPAT=no enabled currently, infact OABI seems to be broken. the docs say it was only deprecated, but the options make broken crap.
86[4] - bin/expr/t_expr:overflow fails on shark, only new failure. fails on other ports too. 86[4] - bin/expr/t_expr:overflow fails on shark, only new failure. fails on other ports too.
87[5] - sh3eb tries to use sh specific atomic config which doesn't work, has hand edited mknative output 87[5] - sh3eb tries to use sh specific atomic config which doesn't work, has hand edited mknative output
88[6] - sshd is broken, regardless of compiler options 88[6] - crtbeginS.o builds incorrectly with GCC 5.3 and -O2. a hack for -O1 has been added.
89 -- libldap.so.4.3 is broken on sparc64; replacing just this fixes sshd. 89[7] -
90 -- very very odd -- linking the GCC 5.3 created libldap *.pico flies with the GCC 4.8 compiler works 
91 -- it is not binutils 2.23 vs 2.26 
92[7] - GCC fails to build 
93[8] - sgimips64 release build fails, mdsetimage'd gz'd kernels don't get built? 90[8] - sgimips64 release build fails, mdsetimage'd gz'd kernels don't get built?
94[9] -  91[9] -
95[10] - fails to build tools with: 92[10] - needs to be re-merged with GCC 5.3 versions of these not-merged-to-mainline-gcc ports.
96 /usr/src8/tools/gcc/../../external/gpl3/gcc/dist/gcc/config/riscv/riscv.h:1087:9: error: attempt to use poisoned "MOVE_BY_PIECES_P" 93[11] -
97 #define MOVE_BY_PIECES_P(SIZE, ALIGN) \ 
98[11] - diagnostic.c fails to bulid -- perhaps some bad merge? 
99 /usr/src8/tools/gcc/../../external/gpl3/gcc/dist/gcc/cp/error.c:3384:40: warning: too many arguments for format [-Wformat-extra-args] 
100 .... 
101[13] - builds tested: 94[13] - builds tested:
102 - m68k: mvm68k amiga atari 95 - m68k: mvm68k amiga atari
103 - earm : shark evbarm* 96 - earm : shark evbarm*
104 - mipsel: sgimips 97 - mipsel: sgimips
105 - mips64el: sgimips[fail] evbmips 98 - mips64el: sgimips[fail] evbmips
106[14] - machines tested: 99[14] - machines tested:
107 - sparc ss20 (OK) 100 - sparc ss20 (OK)
108 - sparc64 ultra45 (FAIL), ultra10 (OK) 101 - sparc64 ultra45 (FAIL), ultra10 (OK), sb2000 (OK UP, FAIL SMP.)
109 - powerpc pegasosII (OK) 102 - powerpc pegasosII (OK)
110 - alpha UP1000 (OK) 103 - alpha UP1000 (OK)
111 - arm shark (OK) 104 - arm shark (OK)
112 - evbmips gxemul MALTA (FAIL - not gcc problem) 105 - evbmips gxemul MALTA (FAIL - not gcc problem)
113 - pmax gxemul (OK) 106 - pmax gxemul (OK)
114 - i386/amd64 (OK) 107 - i386/amd64 (OK)
115 - hppa (OK) 108 - hppa (OK)
116[15] - vax 109[15] - vax
117 - ssh is broken, was broken with 4.8 110 - ssh is broken, was broken with 4.8
118 - there are too many -O0's we added because of mis-handled rtl 111 - there are too many -O0's we added because of mis-handled rtl
119 - the eh_frame stuff is probably busted, but it never worked? 112 - the eh_frame stuff is probably busted, but it never worked?