linaro-toolchain August 2019

linaro-toolchain@lists.linaro.org

11 participants
27 discussions

New IRC Channel for Linaro Toolchain Working Group

by Matthew Gretton-Dann

All, During Connect the suggestion was made that each working group should have its own IRC Channel for discussions and topics relating to the group in particular (as opposed to #linaro which is 'generic' Linaro conversations). Therefore I have just set up #linaro-tcwg on Freenode for the Toolchain Working Group. This channel is public and open to anyone who wants to talk with the TCWG group about anything toolchain related. Thanks, Matt -- Matthew Gretton-Dann Toolchain Working Group, Linaro

2 years, 9 months

ARM GCC 8.x Performance Dropping Compared to Linaro GCC 7.x

by Yupeng Chang

Hi Dear Linaro Team, I recently found a very strange issue regarding the code performance. I have a loop written in GCC NEON. The binary of this coded generated by Linaro GCC 7.x is much faster than it generated by ARM GCC 8.x My CPU is ARM Cortex-A53 AARCH64. The compile option is: -Wall -O3 -mcpu=cortex-a53+crypto the code is like below: for (uint32 c = 0; c < channels; c += 16, roi_result += 16) { int32x4_t S1, S2, S3, S4; int16x4_t DT; DT = vld1_s16(feature1 + c + 0); S1 = vmull_lane_s16(DT, SZ, 0); DT = vld1_s16(feature1 + c + 4); S2 = vmull_lane_s16(DT, SZ, 0); DT = vld1_s16(feature1 + c + 8); S3 = vmull_lane_s16(DT, SZ, 0); DT = vld1_s16(feature1 + c + 12); S4 = vmull_lane_s16(DT, SZ, 0); DT = vld1_s16(feature2 + c + 0); S1 = vmlal_lane_s16(S1, DT, SZ, 1); DT = vld1_s16(feature2 + c + 4); S2 = vmlal_lane_s16(S2, DT, SZ, 1); DT = vld1_s16(feature2 + c + 8); S3 = vmlal_lane_s16(S3, DT, SZ, 1); DT = vld1_s16(feature2 + c + 12); S4 = vmlal_lane_s16(S4, DT, SZ, 1); DT = vld1_s16(feature3 + c + 0); S1 = vmlal_lane_s16(S1, DT, SZ, 2); DT = vld1_s16(feature3 + c + 4); S2 = vmlal_lane_s16(S2, DT, SZ, 2); DT = vld1_s16(feature3 + c + 8); S3 = vmlal_lane_s16(S3, DT, SZ, 2); DT = vld1_s16(feature3 + c + 12); S4 = vmlal_lane_s16(S4, DT, SZ, 2); DT = vld1_s16(feature4 + c + 0); S1 = vmlal_lane_s16(S1, DT, SZ, 3); DT = vld1_s16(feature4 + c + 4); S2 = vmlal_lane_s16(S2, DT, SZ, 3); DT = vld1_s16(feature4 + c + 8); S3 = vmlal_lane_s16(S3, DT, SZ, 3); DT = vld1_s16(feature4 + c + 12); S4 = vmlal_lane_s16(S4, DT, SZ, 3); DT = vrshrn_n_s32(S1, Q_VALUE); vst1_s16(roi_result + 0, DT); DT = vrshrn_n_s32(S2, Q_VALUE); vst1_s16(roi_result + 4, DT); DT = vrshrn_n_s32(S3, Q_VALUE); vst1_s16(roi_result + 8, DT); DT = vrshrn_n_s32(S4, Q_VALUE); vst1_s16(roi_result + 12, DT); } Code generated by GCC7: 294: 6b10031f cmp w24, w16 298: fc606959 ldr d25, [x10, x0] 29c: fc686922 ldr d2, [x9, x8] 2a0: fc676921 ldr d1, [x9, x7] 2a4: fc666920 ldr d0, [x9, x6] 2a8: fc686958 ldr d24, [x10, x8] 2ac: fc676957 ldr d23, [x10, x7] 2b0: fc666956 ldr d22, [x10, x6] 2b4: fc606855 ldr d21, [x2, x0] 2b8: fc686854 ldr d20, [x2, x8] 2bc: fc676853 ldr d19, [x2, x7] 2c0: fc666852 ldr d18, [x2, x6] 2c4: fc606891 ldr d17, [x4, x0] 2c8: fc686890 ldr d16, [x4, x8] 2cc: fc676887 ldr d7, [x4, x7] 2d0: fc666885 ldr d5, [x4, x6] 2d4: 0f44a063 smull v3.4s, v3.4h, v4.h[0] 2d8: 0f44a042 smull v2.4s, v2.4h, v4.h[0] 2dc: 0f44a021 smull v1.4s, v1.4h, v4.h[0] 2e0: 0f44a000 smull v0.4s, v0.4h, v4.h[0] 2e4: 0f542323 smlal v3.4s, v25.4h, v4.h[1] 2e8: 0f542302 smlal v2.4s, v24.4h, v4.h[1] 2ec: 0f5422e1 smlal v1.4s, v23.4h, v4.h[1] 2f0: 0f5422c0 smlal v0.4s, v22.4h, v4.h[1] 2f4: 0f6422a3 smlal v3.4s, v21.4h, v4.h[2] 2f8: 0f642282 smlal v2.4s, v20.4h, v4.h[2] 2fc: 0f642261 smlal v1.4s, v19.4h, v4.h[2] 300: 0f642240 smlal v0.4s, v18.4h, v4.h[2] 304: 0f742223 smlal v3.4s, v17.4h, v4.h[3] 308: 0f742202 smlal v2.4s, v16.4h, v4.h[3] 30c: 0f7420e1 smlal v1.4s, v7.4h, v4.h[3] 310: 0f7420a0 smlal v0.4s, v5.4h, v4.h[3] 314: 0f138c63 rshrn v3.4h, v3.4s, #13 318: 0f138c42 rshrn v2.4h, v2.4s, #13 31c: 0f138c21 rshrn v1.4h, v1.4s, #13 320: 0f138c00 rshrn v0.4h, v0.4s, #13 324: 6d3e0a63 stp d3, d2, [x19, #-32] 328: 6d3f0261 stp d1, d0, [x19, #-16] Code generated by GCC8: 26c: 6b0b02ff cmp w23, w11 270: fc606922 ldr d2, [x9, x0] 274: fc666941 ldr d1, [x10, x6] 278: fc666920 ldr d0, [x9, x6] 27c: 0f44a000 smull v0.4s, v0.4h, v4.h[0] 280: 0f542020 smlal v0.4s, v1.4h, v4.h[1] 284: fc6668e1 ldr d1, [x7, x6] 288: 0f642020 smlal v0.4s, v1.4h, v4.h[2] 28c: fc646945 ldr d5, [x10, x4] 290: fc666901 ldr d1, [x8, x6] 294: 0f742020 smlal v0.4s, v1.4h, v4.h[3] 298: fc646921 ldr d1, [x9, x4] 29c: 0f44a021 smull v1.4s, v1.4h, v4.h[0] 2a0: 0f5420a1 smlal v1.4s, v5.4h, v4.h[1] 2a4: fc626945 ldr d5, [x10, x2] 2a8: 0f138c03 rshrn v3.4h, v0.4s, #13 2ac: fc626920 ldr d0, [x9, x2] 2b0: 0f44a000 smull v0.4s, v0.4h, v4.h[0] 2b4: 0f5420a0 smlal v0.4s, v5.4h, v4.h[1] 2b8: fc606945 ldr d5, [x10, x0] 2bc: 0f44a042 smull v2.4s, v2.4h, v4.h[0] 2c0: 0f5420a2 smlal v2.4s, v5.4h, v4.h[1] 2c4: fc6468e5 ldr d5, [x7, x4] 2c8: 0f6420a1 smlal v1.4s, v5.4h, v4.h[2] 2cc: fc6268e5 ldr d5, [x7, x2] 2d0: 0f6420a0 smlal v0.4s, v5.4h, v4.h[2] 2d4: fc6068e5 ldr d5, [x7, x0] 2d8: 0f6420a2 smlal v2.4s, v5.4h, v4.h[2] 2dc: fc646905 ldr d5, [x8, x4] 2e0: 0f7420a1 smlal v1.4s, v5.4h, v4.h[3] 2e4: fc626905 ldr d5, [x8, x2] 2e8: 0f138c21 rshrn v1.4h, v1.4s, #13 2ec: 0f7420a0 smlal v0.4s, v5.4h, v4.h[3] 2f0: 0f138c00 rshrn v0.4h, v0.4s, #13 2f4: fc606905 ldr d5, [x8, x0] 2f8: 0f7420a2 smlal v2.4s, v5.4h, v4.h[3] 2fc: 0f138c42 rshrn v2.4h, v2.4s, #13 300: 6d000e62 stp d2, d3, [x19] 304: 6d010261 stp d1, d0, [x19, #16] 308: 91008273 add x19, x19, #0x20 I did some tests on different compile options, and found that option "-fschedule-insns" on GCC 7 will generate code that runs faster, if I disable schedule-insns, GCC7 will generate the same code as GCC8. However, this option seems don't work on GCC8, if I enable "-fschedule-insns" with GCC8, the code generated by GCC8 is even slower. If I disable "-fschedule-insns" with GCC8, the generated code is just like the sequence as in C code. I compiled my code with -O3, which means -fschedule-insns will be enabled by default. With this option enabled, GCC7 will reschedule instructions, and it seems that GCC7 will arrange the same instructions all together, but GCC8 doesn't do that, or GCC8 will reschedule instructions in a worse way. My question is, is this behavior expected in GCC8, GCC9 and the future version? Is this change in GCC code scheduling related to the fix of "spectre and mitigation" ? If I want the same instruction scheduling mechanism in GCC8, what can I do ? Thank you for looking into this. Looking forward to your reply! Tomas Chang Aug 19, 2019

4 years, 7 months

[ACTIVITY] 26 - 30 August 2019

by Prathamesh Kulkarni

== This Week == * GCC: (i) PR86753 - Addressing upstream suggestions, pivoted to another approach. (ii) PR91272 - Patch fixes issue but not correct approach, need to rework it. (iii) PR78736 - Submitted patch upstream. == Next Week == - Continue ongoing tasks.

4 years, 7 months

[ACTIVITY] week ending Sep. 1 2019

by Alex Bennée

QEMU Tooling ([VIRT-252]) ========================= QEMU plugin support ([VIRT-280]) - continued working through v4 comments on [next version branch] - split off {PATCH v1 0/9} arm semih-hosting cleanups and other misc cleanups Message-Id: <20190828165307.18321-1-alex.bennee(a)linaro.org> [VIRT-280] https://projects.linaro.org/browse/VIRT-280 [next version branch] https://github.com/stsquad/qemu/tree/plugins/plugins-v5 GSoC Mentoring ([VIRT-348]) - final wrap up reviews for GSoC [VIRT-348] https://projects.linaro.org/browse/VIRT-384 Upstream Work ([VIRT-109]) ========================== - more work on [my docker fixup branch] - now can run make docker-test-build on qemu-test - posted {PATCH v3 00/13} softfloat updates (include tweaks, rm LIT64) Message-Id: <20190813124946.25322-1-alex.bennee(a)linaro.org> - posted {PATCH v1 0/8} various gitdm and mailmap updates Message-Id: <20190828120832.9208-1-alex.bennee(a)linaro.org> - posted {PATCH} linux-user: fail and report on bad dfilter specs Message-Id: <20190830143648.2967-1-alex.bennee(a)linaro.org> - posted {PATCH} tests/docker: upgrade docker.py to python3 Message-Id: <20190829144120.12701-1-alex.bennee(a)linaro.org> - restored shippable cross builds to working order [VIRT-109] https://projects.linaro.org/browse/VIRT-109 Completed Reviews [9/9] ======================= {Qemu-devel} {PATCH v5 00/10} Measure Tiny Code Generation Quality Message-Id: <20190815021857.19526-1-vandersonmr2(a)gmail.com> - CLOSING NOTE [2019-08-15 Thu 20:45] Looks pretty good but some racy crashes need to be fixed. {Qemu-devel} {PATCH v1 0/2} Integrating qemu to Linux Perf Message-Id: <20190815023725.2659-1-vandersonmr2(a)gmail.com> - CLOSING NOTE [2019-08-15 Thu 20:46] Needs a v2 with compile fixes {PATCH} gdbstub: Fix handler for 'F' packet Message-Id: <20190827194849.7076-1-sandra(a)codesourcery.com> - CLOSING NOTE [2019-08-28 Wed 11:30] Queued to my tree {Qemu-devel} {PATCH} linux-user: Support gdb 'qOffsets' query for ELF Message-Id: <20190816233422.16715-1-jkz(a)google.com> - CLOSING NOTE [2019-08-28 Wed 11:31] Laurent had a comment, awaiting v2 {Qemu-devel} {PATCH 0/3} mailmap: Clean up Message-Id: <20190822230916.576-1-philmd(a)redhat.com> - CLOSING NOTE [2019-08-29 Thu 16:29] Queued to my tree {PATCH v3 0/6} tests/docker: add podman support Message-Id: <20190713143311.17620-1-marcandre.lureau(a)redhat.com> - CLOSING NOTE [2019-08-30 Fri 16:12] Acked, it got merged {PATCH 0/4} docs: add docs about use of automatic cleanup functions Message-Id: <20190823163931.7442-4-berrange(a)redhat.com> {PATCH 0/2} contrib/gitdm: Minor adjustments Message-Id: <20190822122350.29852-1-armbru(a)redhat.com> - CLOSING NOTE [2019-08-29 Thu 16:29] Queued to my tree {Qemu-devel} {PATCH v8 00/11} Measure Tiny Code Generation Quality Message-Id: <20190829173437.5926-1-vandersonmr2(a)gmail.com> Absences ======== - Connect Sep 23-27th Current Review Queue ==================== * {Qemu-devel} {PATCH v6 0/6} Allow Valgrind checking all QEMU processes Message-Id: <1566834628-485525-1-git-send-email-andrey.shinkevich(a)virtuozzo.com> * {Qemu-devel} {PATCH 0/6} exec: Cleanup watchpoints Message-Id: <20190824213451.31118-1-richard.henderson(a)linaro.org> * {PATCH 0/4} docs: add docs about use of automatic cleanup functions Message-Id: <20190823163931.7442-1-berrange(a)redhat.com> * {PATCH 0/4} docker: Update Travis-CI image to run acceptance tests Message-Id: <20190818231827.27573-1-philmd(a)redhat.com> * {PATCH v3 00/34} target/arm: Implement ARMv8.1-VHE Message-Id: <20190803184800.8221-1-richard.henderson(a)linaro.org> * {Qemu-devel} {PATCH v4 00/22} target/arm: Implement ARMv8.5-MemTag, system mode Message-Id: <20190307170440.3113-1-richard.henderson(a)linaro.org> -- Alex Bennée

4 years, 7 months

[ACTIVITY] 26 - 30 August 2019

by Diana Picus

== Progress == * Investigate running benchmarks in containers [TCWG-1513] - Cleaned and uploaded scripts for comments - mcf still hangs, but only when run with clang; investigating root cause * SVE IR fuzzer [LLVM-586] - Made some progress with the prototype but it still needs work * Re-running 8.0.1 release on AArch64 since the archive on llvm.org/releases seems to be broken

4 years, 7 months

[ACTIVITY] week ending 25 Aug 2019

by Richard Henderson

[VIRT-263 # ARMv8.1-VHE Virtual Host Extensions ] Reorg ARMMMUIdx again; this time, do not overlap EL1&0 and EL2&0 mmu_idx. This makes debugging a bit easier. Fix one more bug in EL2&0 selection. This was not the last, because a nested kernel does not yet boot. [VIRT-327 # Richard's upstream QEMU work ] Another round of aa32 decodetree patches. Another round of arm hflags patches. Review of and some patches for cpu watchpoints. r~

4 years, 7 months

[ACTIVITY] report week ending 23 Aug

by Peter Maydell

Progress: * VIRT-65 [QEMU upstream maintainership] + sent fix for LP:1840922 (incorrect handling of branches to M-profile exception-return magic addresses in linux-user mode) + some more work on requirements for JIRA KVM related issues + started reviewing RTH's monster decodetree conversion patchset (and made it nearly halfway through it) thanks -- PMM

4 years, 7 months

[ACTIVITY] 19-23 August 2019

by Christophe Lyon

== Progress == * GCC: - FDPIC: handling feedback. * GCC upstream validation: - no issue to report this week. * Binutils: - Non-contiguous memory regions support in the BFD linker: not started yet. * misc: - infra fixes / troubleshooting - qemu-4.1 confirmed OK for GCC validations - trying validations for cortex-m33, found a few issues in GCC and a small one in QEMU - a bit of Jira == Next == GCC: - handle feedback on FDPIC patches - binutils/linker support for non-contiguous memory regions - GNU-583 - GCC upstream validation: Add a config for cortex-m33 (v8-m)

4 years, 7 months

[Activity] week ending 23rd August 2019

by Peter Smith

- Reorganised buildbots to remove some of the duplication between v7 and v8 AArch32 to free up more v7 hardware. - Got final code-size information and wrote my slides for Linaro Connect - Track down an AArch64 specific LLD problem due to a recent change, now resolved. - Wrote worst case thunk/veneer convergence patch to help move through an iterative symbol assignment algorithm. UK Bank Holiday Monday

4 years, 7 months

[ACTIVITY] 19 - 23 August 2019

by Prathamesh Kulkarni

== This Week == * GCC : (i) PR83756: Addressing feedback from upstream. (ii) PR90724: Committed to trunk. (iii) PR91272: Started working on patch. (iv) PR88839: Merged fix from sve-acle-branch to trunk. * Validation (i) Patch to separate build and test steps in tcwg_gnu. (ii) SVE validation == Next Week == - Continue ongoing tasks

4 years, 7 months

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-toolchain August 2019