commit | author | age
|
a7b936
|
1 |
This is originally derived from the pkgsrc-joyent set for openjdk11, |
NP |
2 |
considerably modified. Cut from the jdk15 patches as of jdk15+32. |
|
3 |
|
|
4 |
See also README-zero.txt for patches to build a project zero variant. |
|
5 |
|
|
6 |
Most patches -p0 |
|
7 |
|
|
8 |
JDK17: |
9a6458
|
9 |
|
AW |
10 |
17.0.10 |
|
11 |
|
|
12 |
Drop the make/modules/jdk.net/Lib.gmk patch, as it's guarded in such a |
|
13 |
way as will never be used on solaris. |
|
14 |
|
|
15 |
Need to add illumos-port-18.patch (derived from jdk21's |
|
16 |
illumos-port-24.patch) so that test/jdk/java/io/File/libGetXSpace.c |
|
17 |
will compile. |
|
18 |
|
|
19 |
17.0.9 |
|
20 |
|
|
21 |
Much rework around safefetch. Removed illumos-port-12.patch, and |
|
22 |
removed the is_safefetch_fault() block entirely. (Looked at the |
|
23 |
Windows port, and that block was removed there in this release. It |
|
24 |
wasn't present at all in other platforms.) Copied the new assembler |
|
25 |
src/hotspot/os_cpu/linux_x86/safefetch_linux_x86_64.S to |
|
26 |
src/hotspot/os_cpu/solaris_x86/safefetch_solaris_x86_64.S, see |
|
27 |
illumos-port-17.patch |
|
28 |
|
|
29 |
In os_solaris_x86.cpp, the end of print_context() is now split off |
|
30 |
into print_tos_pc() |
|
31 |
|
|
32 |
We need an implementation of os::can_trim_native_heap() and |
|
33 |
trim_native_heap(); simply add a stub in os_solaris.inline.hpp to |
|
34 |
return false like pretty much every other platform except linux does. |
|
35 |
|
56491d
|
36 |
17.0.8 |
NP |
37 |
|
|
38 |
Reinstate make/data/charsetmapping/stdcs-solaris, removal broke the |
|
39 |
build. |
|
40 |
See illumos-port-16.patch |
|
41 |
|
|
42 |
Cleanup: remove TAR_CREATE_FILE_PARAM TAR_CREATE_EXTRA_PARAM |
|
43 |
|
|
44 |
Cleanup: Removed another STLPORT reference |
a7b936
|
45 |
|
21d27d
|
46 |
17.0.7 |
NP |
47 |
|
|
48 |
Minor patch noise. Removed an STLPORT patch. |
|
49 |
|
a7b936
|
50 |
17.0.6 |
NP |
51 |
|
|
52 |
Minor patch noise. |
|
53 |
|
|
54 |
17.0.5 |
|
55 |
|
|
56 |
A number of functions centralised into os_posix |
|
57 |
|
|
58 |
17.0.4 |
|
59 |
|
|
60 |
Build broken by https://www.illumos.org/issues/14418. That did 2 |
|
61 |
things - (1) exposed memcntl and meminfo by default, and (2) changed |
|
62 |
the signature for memcntl from caddr_t to void so there's a |
|
63 |
mismatch. The fix adopted is to modify the internal java signature for |
|
64 |
memcntl to the new version, which still allows builds on older |
|
65 |
releases as the old definition in sys/mman.h was effectively invisible |
|
66 |
there. |
|
67 |
|
|
68 |
|
|
69 |
Remove some unnecessary patches |
|
70 |
src/java.base/unix/native/libjli/java_md.h |
|
71 |
- new comment is good |
|
72 |
src/java.desktop/unix/native/common/awt/fontpath.c |
|
73 |
- shouldn't need to check for SunOS 5.8/5.9 |
|
74 |
Remove HS_DTRACE_WORKAROUND_TAIL_CALL_BUG, the bug it works around |
|
75 |
was fixed in 2008 |
|
76 |
Remove remaining SUNPRO and related MLIB_NO_LIBSUNMATH checks |
|
77 |
src/java.desktop/unix/native/common/awt/X11Color.c |
|
78 |
- the complexity is over 20 years old and no longer relevant |
|
79 |
make/autoconf/flags-cflags.m4 |
|
80 |
- adding -DTRIMMED is useless |
|
81 |
src/java.desktop/unix/native/libawt_xawt/awt/awt_InputMethod.c |
|
82 |
- unnecessary on illumos and current Solaris (and my S10 system too) |
|
83 |
use the 64-bit rdtsc variant |
|
84 |
|
|
85 |
17.0.2 |
|
86 |
|
|
87 |
illumos-port-15.patch to fix the broken ld check |
|
88 |
|
|
89 |
17-35 |
|
90 |
|
|
91 |
RC1, no changes |
|
92 |
|
|
93 |
17-29,17-30,17-31,17-32,17-33,17-34 |
|
94 |
|
|
95 |
No changes |
|
96 |
|
|
97 |
17-27,17-28 |
|
98 |
|
|
99 |
No changes. Packaged 17-28. |
|
100 |
|
|
101 |
17-26 |
|
102 |
|
|
103 |
First from the jdk17 stabilisation repository. Same changes as in the |
|
104 |
first jdk18 tag. |
|
105 |
|
|
106 |
Remove extended file io workaround, it's 32-bit only |
|
107 |
|
|
108 |
17-25 |
|
109 |
|
|
110 |
SecurityManager deprecation. Added suppression annotations (the way it |
|
111 |
was done for Windows), and cleaned up the line number noise. |
|
112 |
|
|
113 |
17-24 |
|
114 |
|
|
115 |
Line number noise. |
|
116 |
|
|
117 |
Clean up some of the compiler warnings. |
|
118 |
|
|
119 |
17-23 |
|
120 |
|
|
121 |
Line number noise. |
|
122 |
|
|
123 |
os::os_exception_wrapper signature; last argument is a JavaThread not |
|
124 |
a Thread. No impact as we don't actually use it. |
|
125 |
|
|
126 |
17-22 |
|
127 |
|
|
128 |
Signature of reserve_memory_special and pd_reserve_memory_special has |
|
129 |
changed to pass the page size. All we do is change the signature as |
|
130 |
it's a no-op on solaris. |
|
131 |
|
|
132 |
17-21 |
|
133 |
|
|
134 |
Line number noise. |
|
135 |
|
|
136 |
17-20 |
|
137 |
|
|
138 |
Quite a lot of changes this time around. |
|
139 |
|
|
140 |
No more aot, so remove patches for that. |
|
141 |
|
|
142 |
src/hotspot/share/cds/classListParser.cpp has moved to |
|
143 |
src/hotspot/share/cds/classListParser.cpp |
|
144 |
|
|
145 |
suspend/resume has been essentially gutted |
|
146 |
|
|
147 |
Assembler files are all .S not .s now; and the way it's called has |
|
148 |
changed, so we need to just forget about forcing the studio assembler |
|
149 |
|
|
150 |
Other changes made: |
|
151 |
|
|
152 |
Also removed solstudio, which also cleans up C_FLAG_REORDER |
|
153 |
|
|
154 |
Cleared out some of the sparc pieces |
|
155 |
|
|
156 |
Gutted ucrypto |
|
157 |
|
|
158 |
17-19 |
|
159 |
|
|
160 |
.mx.jvmci has gone, we don't need it anyway |
|
161 |
|
|
162 |
Boot jdk requirement bumped to 16. |
|
163 |
This triggers a lot of noisy |
|
164 |
"SIGSEGV happened inside stack but outside yellow and red zone." |
|
165 |
warnings during the build, but they appear harmless. |
|
166 |
|
|
167 |
17-18 |
|
168 |
|
|
169 |
Don't patch jib-profiles.js. We don't use this, there aren't any AIX |
|
170 |
entries, so putting ours in is a waste of effort. |
|
171 |
|
|
172 |
Remove the nbproject patches; they're either wrong or irrelevant, or both. |
|
173 |
|
|
174 |
The boolAtPut flag stuff is no more. The whole JVMFlag machinery has |
|
175 |
been reworked. Renamed to set_bool. |
|
176 |
|
|
177 |
Remove patching of JvmOverrideFiles.gmk when we then disable the |
|
178 |
patch later. |
|
179 |
|
|
180 |
17-17 |
|
181 |
|
|
182 |
No significant changes |
|
183 |
|
|
184 |
17-16 |
|
185 |
|
|
186 |
Some patch noise this time. Some of the SPARC-related noise simply |
|
187 |
removed. |
|
188 |
|
|
189 |
17-15 |
|
190 |
|
|
191 |
No changes this time |
|
192 |
|
|
193 |
17-14 |
|
194 |
|
|
195 |
Remove patch to c1_LIRAssembler.cpp as it's only for SPARC |
|
196 |
Maybe lir_pack/lir_unpack could go too? |
|
197 |
|
|
198 |
Tidied up Awt2dLibraries.gmk |
|
199 |
|
|
200 |
More centralization into os_posix: os::die() os::shutdown() os::abort() |
|
201 |
|
|
202 |
Tidy up some of the format noise |
|
203 |
|
|
204 |
17-13 |
|
205 |
|
|
206 |
Remove the patches for lib-X11.m4; they're actually for S10, so we |
|
207 |
don't want them anyway. And remove the patch to XRBackendNative.c |
|
208 |
related to it which is also an S10 fixup. |
|
209 |
|
|
210 |
os::fork_and_exec has been centralized |
|
211 |
|
|
212 |
MetaspaceShared::is_in_trampoline_frame() has been removed |
|
213 |
|
|
214 |
17-12 |
|
215 |
|
|
216 |
Remove saved_rbp_address_offset() from thread_solaris_x86.hpp |
|
217 |
|
|
218 |
There's also evidence of xattr support on some platforms having been |
|
219 |
added; it's not needed for the build so I've ignored it for now. |
|
220 |
|
|
221 |
17-11 |
|
222 |
|
|
223 |
Fix memTracker.hpp; signature changed |
|
224 |
Don't patch deoptimization.cpp, it was SPARC-specific |
|
225 |
|
|
226 |
17-10 |
|
227 |
|
|
228 |
No significant changes |
|
229 |
|
|
230 |
17-9 |
|
231 |
|
|
232 |
No significant changes |
|
233 |
|
|
234 |
17-8 |
|
235 |
|
|
236 |
Man os:: functions were centralized into os_posix.inline.hpp, so |
|
237 |
needed to remove our duplicates from os_solaris.inline.hpp and |
|
238 |
os_solaris.cpp. Mostly the central copy was identical, the only odd |
|
239 |
one was that the Solaris os::connect had some funky error handling |
|
240 |
|
|
241 |
17-7 |
|
242 |
|
|
243 |
CLOCK_MONOTONIC is now just on. |
|
244 |
Needed to remove os::supports_monotonic_clock() from os_solaris.inline.hpp |
|
245 |
Remove javaTimeMillis() and javaTimeSystemUTC() and javaTimeNanos() |
|
246 |
and javaTimeNanos_info() from os_solaris.cpp |
|
247 |
Parker::park and Parker::unpark no longer compile, park.hpp has been |
|
248 |
modified so we need to make sure _counter is present in |
|
249 |
os_solaris.hpp; that makes it compile but it then crashes the jvm in there |
|
250 |
The posix variant removed the guarantee in PlatformParker; we need to |
|
251 |
do the same, fix the signature, and provide an implementation. |
|
252 |
(The PlatformParker piece took a lot of effort.) |
|
253 |
|
|
254 |
17-6 |
|
255 |
|
|
256 |
we now have perfMemory_posix.cpp, so perfMemory_solaris.ccp should be |
|
257 |
removed or the build fails with multiply defined symbols |
|
258 |
|
|
259 |
17-5 |
|
260 |
|
|
261 |
make/CompileJavaModules.gmk has mostly been gutted; ignore our patch |
|
262 |
as it seem to build fine without |
|
263 |
|
|
264 |
os.cpp needs to include alloca.h, see illumos-port-11.patch |
|
265 |
|
|
266 |
17-4 |
|
267 |
|
|
268 |
src/jdk.hotspot.agent/share/classes/sun/jvm/hotspot/asm/Disassembler.java |
|
269 |
has been refactored, but the new version appears to do the same thing |
|
270 |
as we need in a platform-independent manner, so drop our patch |
|
271 |
|
|
272 |
17-3 |
|
273 |
|
|
274 |
make/common/Modules.gmk has been completely redone; remove our patch |
|
275 |
|
|
276 |
17-2 |
|
277 |
|
|
278 |
More memory stuff; pd_uncommit_memory pd_reserve_memory |
|
279 |
pd_attempt_reserve_memory_at have changed signatures, with an extra |
|
280 |
bool argument (that appears to be unused, at least linux and aix |
|
281 |
don't make any use of it; this appears to be a Mac thing) |
|
282 |
|
|
283 |
17-0/17-1 |
|
284 |
|
|
285 |
Quite a lot of refactoring of the toolchain stuff. |
|
286 |
|
|
287 |
The page_size rework is really quite massive. The old _page_sizes was |
|
288 |
an array you manipulated directly. It's now a set. This needs various |
|
289 |
changes: |
|
290 |
|
|
291 |
os::Solaris::is_valid_page_size() is just _page_sizes.contains() |
|
292 |
|
|
293 |
Not strictly broken, but easy to fix: in mpss_sanity_check, we can |
|
294 |
simplify the getpagesizes() stuff as we know we're modern |
|
295 |
|
|
296 |
Rework listing of valid page sizes |
|
297 |
|
|
298 |
This is a bit ugly because we don't really end up using the new |
|
299 |
_page_sizes, but instead emulate the old array. Still, I can't find |
|
300 |
examples of how the new way is supposed to work on any other platform. |
|
301 |
|
|
302 |
Also need to implement print_memory_mappings, as a no-op (like AIX is) |
|
303 |
|
|
304 |
That's enough to make it compile; it blows up with an arithmetic |
|
305 |
exception, apparently inside apply_ergo(). In mpss_sanity_check, we |
|
306 |
need to make sure page_size (which is really a pointer to |
|
307 |
_large_page_size) is initialized to the largest valid page size. |
|
308 |
|
|
309 |
Build: |
|
310 |
|
|
311 |
env PATH=/usr/bin:/usr/sbin:/usr/sfw/bin:/usr/gnu/bin bash ./configure \ |
|
312 |
--enable-unlimited-crypto --with-boot-jdk=/usr/jdk/instances/jdk16 \ |
|
313 |
--with-native-debug-symbols=none \ |
|
314 |
--with-toolchain-type=gcc \ |
|
315 |
--disable-dtrace \ |
|
316 |
--disable-warnings-as-errors \ |
|
317 |
--enable-deprecated-ports=yes \ |
|
318 |
--with-jobs=3 |
|
319 |
|
|
320 |
env PATH=/usr/bin:/usr/sbin:/usr/sfw/bin:/usr/gnu/bin gmake all |