-
Fri Jan 17 2020 Qing Zhao <qing.zhao@oracle.com> 7.3.0-7
- Add 7 patches from gcc9 to support Arm Ares and Neoverse-N1 for Aarch64
gcc9-add-rcpc-extersion.patch
gcc9-add-dotprod.patch
gcc9-add-mtune-support-for-arm-ares.patch
gcc9-add-vec-reverse.patch
gcc9-multiple-changes-align.patch
gcc9-initial-mcpu-ares-tuning.patch
gcc9-add-support-for-neoverse-n1.patch
-
Mon Jan 13 2020 Qing Zhao <qing.zhao@oracle.com> 7.3.0-6
- Add 5 more marvell patches:Patch4092 to Patch4096.
These 5 patches are new patches from their build 249.
-
Wed Oct 16 2019 Qing Zhao <qing.zhao@oracle.com> 7.3.0-5
- Add 3 more marvell patches:Patch4089 to Patch4091.
patch 4089 and 4090 are new patches from their build 238;
patch 4091 is the fix to Orabug 30424607: CIE with armtoolset-1 gcc with Marvell patch.
-
Tue Sep 24 2019 Qing Zhao <qing.zhao@oracle.com> 7.3.0-4
- Add 88 marvell patches:Patch4001 to Patch4088.
-
Thu Jul 26 2018 Qing Zhao <qing.zhao@oracle.com> 7.3.0-3
- Move the source files added by the debuginfo package from /usr/src/debug/
to /usr/src/debug/oracle-armtoolset-1/
-
Thu May 31 2018 Qing Zhao <qing.zhao@oracle.com> 7.3.0-2
- Backport the following patches from https://git.theobroma-systems.com/ampere-computing/gcc.git:
16d3dddc0913148d8f7b3dc5b0d560934f853031: aarch64: Optimize retpoline (Spectre-V2 mitigation) for aarch64.
ecdb97168b5ba58d1335fd28e4aed2d6ef76d56f: aarch64: Fixing code style issues of retpoline changes.
-
Fri May 11 2018 Qing Zhao <qing.zhao@oracle.com> 7.3.0-1
- Backport the following cavium patches from upstream:
c44b93a4101f6e626e14bb9017661645e2bfea4e: Thunderx shift tuning.
6e0b4434d71790df7b0b099906d33ddd6b9df242: copysign new pattern.
b920c9b17081d65a77ef9ee7cc1699f693888592: CC usage.
1f836a9fa3dd9ea80a323f6a5da3faeec40c7d66: aarch64 fusion tuning.
483d5f69d7d371f39623eabe5e710461a861cb8d: copysign new pattern bug fix.
2988c5dac284abf7d3c7725daf481c69fe4452ad: aarch64 fuse alu branch.
2c42281526deda5e19b6b72c979e8f24a8c0c8ec: type precision.
e3126fb7cacbc050b225ab620a612130cac962ee: thunderx vector cost fix.
016b4ff4475c37b61c88b48373dc56eb3df7828f: thunderx vector cost fix 2.
5e09c9baaf432dc7be445e25acf0e8866badba6e: type precision 2.
4fc06d6a909011a75c81e39e7ffcb16d60f4f061: skip float16 tests.
c526c5b43a13b9ccb5aac931276ddb5869c3a3ca: target attr tests fix.
74956124779123d700efa4f9d2cd1aabc55a3986: document change on common function attribute.
7a3c63c6f6370bca81e9775671a631fdcb8b00da: gnu indirect function.
7ba254509add68b209e8b73e9bb5edf1e6cb870a: Fix PR81356.
5dfed983bb76ec0b898dd0661d7b302ab5e71f00: Fix PR81356 test cases.
fae12d432888719b697b7ebee6db10b1a6569071: Move neg operator to canonical location.
69eab567dcc51ec88d4f482c82d80aa01c3fcdd7: Fix PR81356 test cases 2.
0d17f1941b68351df07e6bef98936a22b96ea989: Ifunction.
4c7a29d5c9cb1c90ce5093b411735d850569586e: add insns reservation.
f59d67f6abf2d55e69c154a778261d2b334792bf: Change aarch64 option from -march=armv8.1-a to -march=armv8-a+lse.
3bf398e12218ab188f82d66c5d78c0d30a966168: aarch64: add prefetch configure.
067e12015409823c2149324fd5c05c0dee2a4167: aarch64: enable -fprefetch-loop-arrays at given opt level
404fe869d0d62e2d6daa7b9bb23fa15f7513d8db: aarch64: Update prefetch tuning parameters for qdf24xx.
000d2de6b7ad01558ced27d5a60e7cee73c9b081: aarch64: use thunderxt88 tuning for prefetch.
- Backport the following Oracle patches from upstream:
d0071aec2a6d677ae308e62ca6e1ccf79d89d6e9: Fix PR81422
d29e4e8c5dad17dc1f22c5acf950b13820c7876f: Fix PR81942
- Wed May 02 2018 Qing Zhao <qing.zhao@oracle.com>
- Backport the following patches from https://git.theobroma-systems.com/ampere-computing/gcc.git:
d10b48c9527fe0bcd5f35cb4086712fdb74056d4: aarch64: Retpoline (Spectre-V2 mitigation) for aarch64.
00ea0fbb71a348505b2ba0b9ba073962da5383d3: aarch64: Correct the maximum shift amount for shifted operands.
d0145f7336fe98a040938f89eb045115207ccd00: aarch64: Extend '*tb<optab><mode>1'
d7e5332c0809d246afdbe6e27c65ed4aae7005ac: aarch64: Optimize and(s) patterns for HI/QI operands.
ba5a3780dd9937d0f3adee051239ea3aa65dd148: aarch64: Fix and<mode>3_zeroextend case (20040709-1.s regression)
43881c3d5e9a4fb1e68481d6a4fa3bfa2489f2f0: aarch64: Xgene: Procedural cost-model for X-Gene processors.
8bf8a9909e5fb3864254494cca6a76f72058c1fe: aarch64: X-Gene: Updated scheduling model for GCC 7
36989845c3814cb0fbb441eaca21039785fd54bc: aarch64: X-Gene: Adapt tuning struct for GCC 7.
7c9039cf2c981fe243d2807ce4305f30555c5da3: Adding -fnoalias to delcare non-aliasing on function level
a3ba4a0362e6707ca4ac86ee8590faecc10ae5d4: tree-vect-slp: Enable vectorization when -fnoalias is given.
942940be0f81b83037da18b0a0c2c8fb7a26f29f: tree-vect-slp: Adapt calculation of scalar cost.
3df45c222875f9edb98342cf1fcdf92edb2850b2: tree-predcom: Do more unrolling if -fnoalias flag is set.
a7c5fae2afcee786dcfac1c5443606b14479e893: tree-ssa-list-find-pipeline: Add pipelining loads for list finds.
017ea04488be91bab4952859b83e5a6f5f300e36: uncse: Added pass to undo common subexpression elimination.
6fec8836a472fbf6306ee4db5291dc85f91944f3: Added clone functions to duplicated passes.
d28b83f9d0284a6a7c55638ff25715d650b8012a: tree-cfg: Store gimple_build_nop () instead of NULL.
aebbf1f37938b6143d62be77275cd1744551d6bb: Avoid assembler warnings from AArch64 constructor/destructor priorities.
050f1d28b1de094b3112c0d5331322f109561713: Revert "2017-10-24 Richard Biener <rguenther@suse.de>"
-
Mon Apr 23 2018 Qing Zhao <qing.zhao@oracle.com> 7.3.0-0
- Initial version for Oracle ARM Toolset.