Bug 282663
| Summary: | lang/rust: building 1.82.0 on/for armv7 fails for multiple definitions and undefined field name | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Mark Millard <marklmi26-fbsd> | ||||
| Component: | Individual Port(s) | Assignee: | FreeBSD Rust Team <rust> | ||||
| Status: | Closed FIXED | ||||||
| Severity: | Affects Some People | CC: | asomers, fuz, jonc, marklmi26-fbsd, mikael, rcm | ||||
| Priority: | --- | Flags: | bugzilla:
maintainer-feedback?
(rust) fuz: merge-quarterly? |
||||
| Version: | Latest | ||||||
| Hardware: | Any | ||||||
| OS: | Any | ||||||
| Attachments: |
|
||||||
(In reply to Mark Millard from comment #0) It may well be that this submittal is mostly a reminder to pick up an up stream fix if/when such becomes available. I make no claim that the issue is FreeBSD specific, though it could be. (A FreeBSD vs. gnu confusion, where it appears to meet both criteria at the same time?) Notably, this is a different error than the one I got in (and still get) bug #280027. Mark said he has a bunch of patches applied to his ports tree when he got this error. Maybe that changed things? What kernel version with what configuration are you testing on? (In reply to Robert Clausecker from comment #2) I already reported some kernel identification information, repeated below: # uname -apKU FreeBSD aarch64-main-pbase 15.0-CURRENT FreeBSD 15.0-CURRENT main-n273174-8b2e7da70855 GENERIC-NODEBUG arm64 aarch64 1500026 1500026 That is from an installation of an official PkgBase kernel. As such the configuration is as it was in the official PkgBase build. Any /etc/src.conf or /etc/make.conf content, for example, came from the PkgBase build server, not my context. I expect that /usr/src/sys/arm64/conf/* related files involved are as they were provided in the pkg that supplies /usr/src/sys/ content. [I'll note that I also have the debug (WITNESS: just "GENERIC") kernel from that PkgBase as well and could run while booted from that.] But PkgBase is a bit of a problem for well identifying its commit(s) and the matching source code, especially because sys/ has its own package that need not be from the same commit that the rest of the source tree is from. Converting back to a pair of Commit-IDs is not obvious as well. So I provide the following summary based on snapshot IDs (snapshot date and time information): # ~/pkgbase-snapshot-list.sh Via pkg-static info -C -x '^FreeBSD-' . . . 352 FreeBSD-*-15.snap20241023235252 1 FreeBSD-*-15.snap20241022122410 1 FreeBSD-*-15.snap20241020224518 1 FreeBSD-*-15.snap20241020094723 7 FreeBSD-*-15.snap20241015203742 1 FreeBSD-*-15.snap20241015145839 1 FreeBSD-*-15.snap20241015120827 1 FreeBSD-*-15.snap20241014101436 34 FreeBSD-*-15.snap20241011184813 129 FreeBSD-*-15.snap20241009162208 Instead via /var/cache/pkg/*.snap*.pkg . . . 352 FreeBSD-*-15.snap20241023235252 1 FreeBSD-*-15.snap20241022122410 1 FreeBSD-*-15.snap20241020224518 1 FreeBSD-*-15.snap20241020094723 7 FreeBSD-*-15.snap20241015203742 1 FreeBSD-*-15.snap20241015145839 1 FreeBSD-*-15.snap20241015120827 1 FreeBSD-*-15.snap20241014101436 34 FreeBSD-*-15.snap20241011184813 129 FreeBSD-*-15.snap20241009162208 (A full list of the individual packages would be over 500 packages long.) As for kern.conftxt as seen via sysctl : # sysctl kern.conftxt kern.conftxt: options CONFIG_AUTOGENERATED ident GENERIC-NODEBUG machine arm64 cpu ARM64 makeoptions MODULES_EXTRA=dtb/allwinner makeoptions WITH_CTF=1 makeoptions DEBUG=-g options SOC_XILINX_ZYNQ options SOC_ROCKCHIP_RK3568 options SOC_ROCKCHIP_RK3399 options SOC_ROCKCHIP_RK3328 options SOC_NXP_LS options SOC_NVIDIA_TEGRA210 options SOC_MARVELL_8K options SOC_FREESCALE_IMX8 options SOC_HISI_HI6220 options THUNDERX_PASS_1_1_ERRATA options SOC_CAVM_THUNDERX options SOC_BRCM_NS2 options SOC_BRCM_BCM2838 options SOC_BRCM_BCM2837 options RATELIMIT options COMPAT_LINUXKPI options SOC_APPLE_T8103 options SOC_INTEL_STRATIX10 options SOC_ALLWINNER_H6 options SOC_ALLWINNER_H5 options SOC_ALLWINNER_A64 options FDT options IEEE80211_SUPPORT_MESH options USB_HOST_ALIGN=64 options EVDEV_SUPPORT options NVME_USE_NVD=0 options PCI_IOV options PCI_HP options PPS_SYNC options SMP options NETDUMP options DEBUGNET options ZSTDIO options GZIO options EKCD options ALT_BREAK_TO_DEBUGGER options VERBOSE_SYSINIT=0 options GDB options DDB options KDB_TRACE options KDB options PERTHREAD_SSP options LINUX_BOOT_ABI options RCTL options RACCT_DEFAULT_TO_DISABLED options RACCT options VFP options INCLUDE_CONFIG_FILE options DDB_CTF options KDTRACE_HOOKS options KDTRACE_FRAME options MAC options CAPABILITIES options CAPABILITY_MODE options AUDIT options HWPMC_HOOKS options KBD_INSTALL_CDEV options PRINTF_BUFR_SIZE=128 options _KPOSIX_PRIORITY_SCHEDULING options SYSVSEM options SYSVMSG options SYSVSHM options STACK options KTRACE options SCSI_DELAY=5000 options COMPAT_FREEBSD14 options COMPAT_FREEBSD13 options COMPAT_FREEBSD12 options COMPAT_FREEBSD11 options COMPAT_FREEBSD32 options EFIRT options GEOM_LABEL options GEOM_RAID options TMPFS options PSEUDOFS options PROCFS options CD9660 options MSDOSFS options NFS_ROOT options NFSLOCKD options NFSD options NFSCL options MD_ROOT options QUOTA options UFS_GJOURNAL options UFS_DIRHASH options UFS_ACL options SOFTUPDATES options FFS options KERN_TLS options SCTP_SUPPORT options TCP_RFC7413 options TCP_HHOOK options TCP_BLACKBOX options TCP_OFFLOAD options FIB_ALGO options ROUTE_MPATH options IPSEC_OFFLOAD options IPSEC_SUPPORT options INET6 options INET options VIMAGE options PREEMPTION options NUMA options SCHED_ULE options NETLINK options INTRNG options CC_CUBIC options GEOM_PART_GPT options GEOM_PART_MBR options GEOM_PART_BSD options CC_CUBIC options FDT options FDT options FDT options FDT options FDT options KERN_TLS options FDT options FDT options FDT options FDT options FDT options FDT options FDT options FDT options FDT options FDT options FDT device mem device efidev device efirtc device smbios device pci device cpufreq device ahci device scbus device da device cd device pass device nvme device nvd device gpio device gpioled device fdt_pinctrl device gpioregulator device iicbus device iicmux device iic device icee device armv8crypto device spibus device pwm device uart device vt device kbdmux device vt_efifb device vt_simplefb device crypto device armv8_rng device loop device ether device vlan device tuntap device md device gif device firmware device clk device phy device hwreset device nvmem device regulator device syscon device evdev device uinput device iflib device em device ix device mdio device mii device miibus device bpf device netmap device ohci device uhci device ehci device xhci device usb device usbhid device hkbd device ukbd device umass device sound device mmc device mmcsd device hid device hidbus device mmio_sram device al_ccu device al_nb_service device al_iofic device al_serdes device al_udma device al_pci device uart_ns8250 device al_eth device acpi device a10_timer device a31_dmac device aw_gpio device aw_rsb device twsi device sy8106a device aw_ccu device aw_r_intc device aw_nmi device aw_rtc device aw_wdog device aw_syscon device axp81x device aw_sid device aw_thermal device uart_snps device awg device aw_usbphy device musb device dwc3 device aw_dwc3 device a10_codec device aw_i2s device a33_codec device a64_codec device aw_mmc device dwgpio device dwc device dwc_socfpga device dwmmc device dwmmc_altera device pl061 device pl011 device axa device msk device bge device pci_n1sdp device pl031 device scmi device arm_doorbell device hyperv device xz device mlx5 device mlxfw device mlx5en device bcm2835_bsc device bcm2835_spi device uart_mu device genet device dwcotg device muge device smsc device sdhci device vnic device virtio device virtio_pci device virtio_mmio device virtio_blk device vtnet device dwmmc_hisi device vf_i2c device fsliic device uart_imx device ffec device a37x0_gpio device mv_gpio device mvebu_pinctrl device a37x0_iic device tca64xx device mv_cp110_icu device mv_ap806_gicp device mv_ap806_sei device mv_rtc device safexcel device mv_thermal device a37x0_spi device uart_mvebu device neta device etherswitch device miiproxy device ehci_mv device sdhci_xenon device a37x0_xtal device a37x0_tbg device a37x0_nb_periph device a37x0_sb_periph device re device tegra210_xusb_fw device ure device pca954x device pcf8563 device pcf85063 device dpaa2 device enetc device sff device qcom_gcc device qcom_mdio device uart_msm device rk_gpio device rk_pinctrl device rk_combphy device rk_i2c device fan53555 device rk805 device rk817 device syr827 device tcs4525 device rk_spi device rk_pwm device dwc_rk device eqos device rk_iodomain device rk_usb2phy device rk_typec_phy device rk_dwc3 device rk_dwmmc device rk_emmcphy device pvscsi device vmx device virtio_gpu device virtio_scmi device virtio_scsi device acpi_ged device cgem device cdnc_i2c (In reply to Robert Clausecker from comment #2) FYI: My ports activity is biased to testing the recent devel/llvm[0-9][0-9] use (that is not a -devel one). There are other tailored aspects as well. I tend to prefer to replicate problems via official FreeBSD builds. But in this context for the world code, I do not have a reasonable context to do so. So my personal environment is involved. Sorry. Also, I've had no problem with building 1.81.0 of lang/rust for armv7. My problem started with 1.82.0 (which internally upgraded to LLVM 19). But Robert's reference to bug #280027 indicates a longer history for his problems, not specific to 1.82.0 of rust or to LLVM 19 being involved. And it makes a claim that contradicts the historical behavior on the official armv7 build server (ampere2 using armv7 jails via poudriere): QUOTE Seems like it's a port bug after all. To reproduce, build with Poudriere . . . END QUOTE So far as I know, the ampere2 has never shown the behaviors Robert has reported for its poudriere based builds. The recent: https://pkg-status.freebsd.org/ampere2/data/main-armv7-default/p711a2f7f1ca4_s0d965bc034/logs/rust-1.81.0.log is another example of a successful rust 1.81.0 build on ampere2. =>> Building lang/rust build started at Fri Nov 1 05:01:04 UTC 2024 port directory: /usr/ports/lang/rust package name: rust-1.81.0 building for: FreeBSD main-armv7-default-job-10 15.0-CURRENT FreeBSD 15.0-CURRENT 1500026 arm maintained by: rust@FreeBSD.org Makefile datestamp: -rw-r--r-- 1 root wheel 11307 Oct 26 01:05 /usr/ports/lang/rust/Makefile Ports top last git commit: 711a2f7f1ca Ports top unclean checkout: no Port dir last git commit: 2501f5298c3 Port dir unclean checkout: no Poudriere version: poudriere-git-3.4.2 Host OSVERSION: 1500023 Jail OSVERSION: 1500026 Job Id: 10 . . . So I think what I've reported and what Robert's reporting are separate issues. Also, based on ampere2 build behavior, the problem seems specific to Robert's context somehow. (In reply to Mark Millard from comment #4) While the evidence is still for separate problems for my report vs. Robert's reports, below I correct a sampling bias in my comparisons. CORRECTION for the ampere* status for building lang/rust . . . My comments were based on a sampling bias: I use and pay attention to main [so: 15 for now] for the most part. This mean ampere2 only as well. https://portsfallout.com/fallout?port=lang%2Frust&maintainer=&env=releng-armv7-&category=&flavor= does indicate that *releng-armv7-* (so 133 and 141, default and quarterly, as things are now) are failing to build lang/rust on the ampere[13] buildlers and have been that way for some time. (I've not reviewed the log details vs. Robert's reports in the other bugzilla submittal.) I'll note that port-package builders (ampere* included) run a main [so: 15] kernel but a poudriere/world-jail that is for the target system type and version. So: only main-armv7-default seems to be well behaved for lang/rust. None of the *releng-armv7-* seems to be well behaved for that. lang/rust armv7 package building for releases is broken on FreeBSD and has been for some time. https://portsfallout.com/ does not present a long term history as far as I can tell, currently going back to early 2024-Jul. So I've no evidence of when before then the lang/rust build problems for releases started. I'll note that the FreeBSD context here is main [so: 15 as stands, both kernel and world] aarch64 with aarch32/armv7 supported for use code as well. Building lang/rust on stable/14 (kernel and world) on an aarch64 with aarch32/armv7 support is broken other ways. Similarly the official package build server records show failures for that main-kernel [so: 15 as stands] but releng/1[34].* poudriere jail world --that match my attempted stable/14 test that did not involve FreeBSD main at all. Also, back when it was 1.81.0, I had no problem for main [so: 15 as stands, both kernel and world] But the historical official package builder logs show failures for all the 1.[78]*.0 versions that still have log files available. So I seem to be reporting something new here. I think the bootstrap compiler still has 'env: "gnu".into(),' in its target definition. I'll regen a working bootstrap "soon". (In reply to Mikael Urankar from comment #7) Yes, that sounds likely to me. But if that proves difficult, an alternative would be to reintroduce the patch-vendor_nix-0.28.0_src_sys_signal.rs file that was removed in 05961664b7f01ac5ed3e9352d43c2aa1ae3028b5 (PR #282518). It should be harmless, but it might help to regenerate a newer rust-bootstrap As for 1.81.0 builds, the official main-armv7-default overall build on ampere2 completed and was successful for its 1.81.0 lang/rust build: =>> Building lang/rust build started at Mon Nov 11 06:37:44 UTC 2024 port directory: /usr/ports/lang/rust package name: rust-1.81.0 building for: FreeBSD main-armv7-default-job-02 15.0-CURRENT FreeBSD 15.0-CURRENT 1500026 arm maintained by: rust@FreeBSD.org Makefile datestamp: -rw-r--r-- 1 root wheel 11307 Oct 26 01:05 /usr/ports/lang/rust/Makefile Ports top last git commit: 380be9c7980 Ports top unclean checkout: no Port dir last git commit: 2501f5298c3 Port dir unclean checkout: no Poudriere version: poudriere-git-3.4.2 Host OSVERSION: 1500023 Jail OSVERSION: 1500026 Job Id: 02 . . . =======================<phase: package >============================ ===== env: 'PKG_NOTES=build_timestamp ports_top_git_hash ports_top_checkout_unclean port_git_hash port_checkout_unclean built_by' 'PKG_NOTE_build_timestamp=2024-11-11T06:37:44+0000' 'PKG_NOTE_ports_top_git_hash=380be9c7980' 'PKG_NOTE_ports_top_checkout_unclean=no' 'PKG_NOTE_port_git_hash=2501f5298c3' 'PKG_NOTE_port_checkout_unclean=no' 'PKG_NOTE_built_by=poudriere-git-3.4.2' NO_DEPENDS=yes USER=root UID=0 GID=0 ===> Building packages for rust-1.81.0 ===> Building rust-1.81.0 =========================================================================== =>> Cleaning up wrkdir ===> Cleaning for rust-1.81.0 build of lang/rust | rust-1.81.0 ended at Mon Nov 11 16:27:47 UTC 2024 build time: 09:50:04 Note that "Jail OSVERSION: 1500026" means the poudriere jail had the LLVM 19 system clang toolchain. [Note: official builds for recent 13.* and 14.* do not work, just main. This had been true for a long time.] A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/ports/commit/?id=560fd669d24aafad437789954800d57e7cfeafa9 commit 560fd669d24aafad437789954800d57e7cfeafa9 Author: Mikael Urankar <mikael@FreeBSD.org> AuthorDate: 2024-11-14 14:59:30 +0000 Commit: Mikael Urankar <mikael@FreeBSD.org> CommitDate: 2024-11-14 15:30:31 +0000 lang/rust: Regen bootstrap for armv7 PR: 282663 lang/rust/distinfo | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) Created attachment 255174 [details]
rust-1.82.0_1 build log on armv7 FreeBSD 14.1
Build still fails for me with the recent bootstrap toolchain regen.
(In reply to Robert Clausecker from comment #11) That looks like a new error. Or have you seen it before? thread 'rustc' panicked at compiler/rustc_middle/src/ty/adt.rs:155:20: already borrowed: BorrowMutError stack backtrace: 0: rust_begin_unwind 1: core::panicking::panic_fmt 2: core::cell::panic_already_borrowed 3: <rustc_middle::ty::adt::AdtDefData as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable 4: <rustc_type_ir::ty_info::WithCachedTypeInfo<rustc_type_ir::ty_kind::TyKind<rustc_middle::ty::context::TyCtxt>> as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable 5: <rustc_type_ir::ty_info::WithCachedTypeInfo<rustc_type_ir::ty_kind::TyKind<rustc_middle::ty::context::TyCtxt>> as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable 6: <&rustc_middle::ty::list::RawList<(), rustc_middle::ty::Ty> as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable 7: rustc_symbol_mangling::legacy::mangle 8: rustc_symbol_mangling::symbol_name_provider [... omitted 2 frames ...] 9: <hashbrown::raw::RawIterRange<(rustc_span::def_id::LocalDefId, ())>>::fold_impl::<<hashbrown::map::Iter<rustc_span::def_id::LocalDefId, ()> as core::iter::traits::iterator::Iterator>::fold<(), <hashbrown::map::Keys<rustc_span::def_id::LocalDefId, ()> as core::iter::traits::iterator::Iterator>::fold<(), core::iter::adapters::filter_map::filter_map_fold<&rustc_span::def_id::LocalDefId, rustc_span::def_id::LocalDefId, (), rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider::{closure#0}, core::iter::adapters::map::map_fold<rustc_span::def_id::LocalDefId, (rustc_span::def_id::DefId, rustc_middle::middle::exported_symbols::SymbolExportInfo), (), rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider::{closure#1}, core::iter::traits::iterator::Iterator::for_each::call<(rustc_span::def_id::DefId, rustc_middle::middle::exported_symbols::SymbolExportInfo), <hashbrown::map::HashMap<rustc_span::def_id::DefId, rustc_middle::middle::exported_symbols::SymbolExportInfo, core::hash::BuildHasherDefault<rustc_hash::FxHasher>> as core::iter::traits::collect::Extend<(rustc_span::def_id::DefId, rustc_middle::middle::exported_symbols::SymbolExportInfo)>>::extend<core::iter::adapters::map::Map<core::iter::adapters::filter_map::FilterMap<std::collections::hash::set::Iter<rustc_span::def_id::LocalDefId>, rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider::{closure#0}>, rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider::{closure#1}>>::{closure#0}>::{closure#0}>::{closure#0}>::{closure#0}>::{closure#0}>::{closure#0}, ()> 10: rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider [... omitted 2 frames ...] 11: rustc_codegen_ssa::back::symbol_export::exported_symbols_provider_local [... omitted 2 frames ...] 12: <rustc_metadata::rmeta::encoder::EncodeContext>::encode_crate_root 13: rustc_metadata::rmeta::encoder::encode_metadata 14: rustc_metadata::fs::encode_and_write_metadata 15: rustc_interface::passes::start_codegen 16: <rustc_interface::queries::Linker>::codegen_and_build_linker 17: <rustc_middle::ty::context::GlobalCtxt>::enter::<rustc_driver_impl::run_compiler::{closure#0}::{closure#1}::{closure#6}, core::result::Result<core::option::Option<rustc_interface::queries::Linker>, rustc_span::ErrorGuaranteed>> 18: <rustc_interface::queries::QueryResult<&rustc_middle::ty::context::GlobalCtxt>>::enter::<core::result::Result<core::option::Option<rustc_interface::queries::Linker>, rustc_span::ErrorGuaranteed>, rustc_driver_impl::run_compiler::{closure#0}::{closure#1}::{closure#6}> 19: <rustc_interface::interface::Compiler>::enter::<rustc_driver_impl::run_compiler::{closure#0}::{closure#1}, core::result::Result<core::option::Option<rustc_interface::queries::Linker>, rustc_span::ErrorGuaranteed>> note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace. error: the compiler unexpectedly panicked. this is a bug. (In reply to Alan Somers from comment #12) Note that this same error was reported by somebody attempting to cross-build Rust 1.77.1 from amd64 to aarch64 on NetBSD. https://github.com/rust-lang/rust/issues/123551 clippy fails to build on armv7 due to memory exhaustion. Otherwise rust 1.82.0 builds fine on my armv7 board (with tools=["analysis","cargo","rust-analyzer","rustdoc","rustfmt","rls"] in config.toml) (In reply to Mikael Urankar from comment #14) On 14.1 or on 15-CURRENT? The problems so far only reproduce on 14.1 and 13.3. (In reply to Robert Clausecker from comment #15) 15-current clippy builds if I put codegen-units=1 in config.toml (In reply to Robert Clausecker from comment #15) My non-FreeBSD-main test context is: stable/14 . I get the failures for that context. I've not yet updated a FreeBSD-main [so: 15 as stands] test context to test building there after the recent "gnu" removal related updates. (In reply to Mark Millard from comment #17) I updated on a personally configured FreeBSD-main's port tree and started a 1.82.0 lang/rust build. It is past where the stable/14 (my testing) and releases (Robert's testing) fail. (I've never seen the specific problem on FreeBSD main, despite use of armv7 poudriere jails on aarch64.) On stable/14 I now have a PORTS_LLVM based build of 1.82.0 of lang/rust in process. But it has to build the LLVM 17 first. (I turned off various options for the LLVM 17 build that are not needed in order to cut the time involved. But it will be hours before it starts to build lang/rust itself.) (In reply to Mikael Urankar from comment #14) I've a question, generated on/for my stable/14 test context (I split the lines before "ELF"): # file /usr/obj/DESTDIRs/PBstable-armv7/bin/sh /usr/obj/DESTDIRs/PBstable-armv7/bin/sh: ELF 32-bit LSB executable, ARM, EABI5 version 1 (FreeBSD), dynamically linked, interpreter /libexec/ld-elf.so.1, FreeBSD-style, for FreeBSD 14.2 (1402500), stripped vs. # file /usr/obj/DESTDIRs/PBstable-armv7/wrkdirs/usr/ports/lang/rust/work/bootstrap/bin/rustc /usr/obj/DESTDIRs/PBstable-armv7/wrkdirs/usr/ports/lang/rust/work/bootstrap/bin/rustc: ELF 32-bit LSB pie executable, ARM, EABI5 version 1 (FreeBSD), dynamically linked, interpreter /libexec/ld-elf.so.1, FreeBSD-style, with debug_info, not stripped Or, for easier comparison just the content that I'm worried about: ELF 32-bit LSB executable, ARM, EABI5 version 1 (FreeBSD), dynamically linked, interpreter /libexec/ld-elf.so.1, FreeBSD-style, for FreeBSD 14.2 (1402500), stripped vs. ELF 32-bit LSB pie executable, ARM, EABI5 version 1 (FreeBSD), dynamically linked, interpreter /libexec/ld-elf.so.1, FreeBSD-style, with debug_info, not stripped Why does work/bootstrap/bin/rustc not show ", for FreeBSD 14.2 (1402500)" or anything indicating the FreeBSD version that is is for? Is work/bootstrap/bin/rustc trying to be the same executable by content across all of 1[34].*-RELELASE and stable/1[34] and main [so: 15 as stands]? (In reply to Mark Millard from comment #19) Looks like some involved .so files are explicit (FreeBSD 13.3 here): # file /usr/obj/DESTDIRs/PBstable-armv7/wrkdirs/usr/ports/lang/rust/work/bootstrap/bin/../lib/librustc_driver-0609d770979c2845.so /usr/obj/DESTDIRs/PBstable-armv7/wrkdirs/usr/ports/lang/rust/work/bootstrap/bin/../lib/librustc_driver-0609d770979c2845.so: ELF 32-bit LSB shared object, ARM, EABI5 version 1 (FreeBSD), dynamically linked, for FreeBSD 13.3, with debug_info, not stripped # file /usr/obj/DESTDIRs/PBstable-armv7/wrkdirs/usr/ports/lang/rust/work/bootstrap/bin/../lib/libstd-0a46fd5bc285334c.so /usr/obj/DESTDIRs/PBstable-armv7/wrkdirs/usr/ports/lang/rust/work/bootstrap/bin/../lib/libstd-0a46fd5bc285334c.so: ELF 32-bit LSB shared object, ARM, EABI5 version 1 (FreeBSD), dynamically linked, for FreeBSD 13.3, with debug_info, not stripped I looked at ztop: same sort of lack of indication. So such may be normal for rust. (In reply to Mark Millard from comment #18) The lang/rust 1.82.0_1 with the updated boostrap built on a aarch32/armv7 capable aarch64 in an armv7 poudriere jail built just fine. This contrasts with my testing on stable/14 and Robert's tests on 1[34].*-RELEASE . So far as I can tell, native armv7 vs. aarch64 with aarch32/armv7 support is not the controlling aspect at all. Unfortunately, I've still no clue why stable/14 and 1[34].*-RELEASE do the odd things that they do. (In reply to Mikael Urankar from comment #14) I've put a PkgBase based stable/14 on USB media and used it on a Native armv7 (OrangePi+ 2ed) and it replicated the newer "already borrowed" failure in this context: thread 'rustc' panicked at compiler/rustc_middle/src/ty/adt.rs:155:20: already borrowed: BorrowMutError stack backtrace: thread 'rustc' panicked at compiler/rustc_middle/src/ty/adt.rs:155:20: already borrowed: BorrowMutError stack backtrace: thread 'rustc' panicked at compiler/rustc_middle/src/ty/adt.rs:155:20: already borrowed: BorrowMutError stack backtrace: 0: rust_begin_unwind 1: core::panicking::panic_fmt 2: core::cell::panic_already_borrowed 3: <rustc_middle::ty::adt::AdtDefData as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable 4: <rustc_type_ir::ty_info::WithCachedTypeInfo<rustc_type_ir::ty_kind::TyKind<rustc_middle::ty::context::TyCtxt>> as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable 5: rustc_symbol_mangling::legacy::mangle 6: rustc_symbol_mangling::symbol_name_provider . . . native armv7 vs. aarch32/armv7 on aarch64 is not what makes the difference here. What makes the difference is main [so: 15 as stands] vs. stable/1[34] and 1[34].*-RELEASE : it only works on main . For reference: # uname -apKU FreeBSD armv7-OfficialStable 14.2-STABLE FreeBSD 14.2-STABLE stable/14-n269609-9ee0b40e4a51 GENERIC arm armv7 1402500 1402500 # poudriere jail -jstable-armv7 -i Jail name: stable-armv7 Jail version: 14.2-STABLE Jail arch: armv7 Jail method: pkgbase Jail mount: /usr/local/poudriere/jails/stable-armv7-poud Jail fs: Jail updated: 2024-11-16 09:59:48 Jail pkgbase: disabled # poudriere ports -l PORTSTREE METHOD TIMESTAMP PATH default git+https 2024-11-16 10:15:57 /usr/local/poudriere/ports/default # ~/fbsd-based-on-what-commit.sh -C /usr/local/poudriere/ports/default 10098268c (grafted, HEAD -> main, origin/main, origin/HEAD) x11-wm/awesome-vicious: Update to 2.7.1 Author: Nuno Teixeira <eduardo@FreeBSD.org> Commit: Nuno Teixeira <eduardo@FreeBSD.org> CommitDate: 2024-11-16 09:29:12 +0000 branch: main fatal: Not a valid object name freebsd/main (In reply to Alan Somers from comment #13) I've rebuilt a bootstrap based on rust 1.82 with "has_thread_local: false" in compiler/rustc_target/src/spec/base/freebsd.rs It seems to work (at least it doesn't "crash" straight away). (In reply to Mikael Urankar from comment #23) If someone wants to confirm, the bootstrap are available here: https://cocyte.westeurope.cloudapp.azure.com/rust/ you'll have to rename them to 1.81 and regen distinfo It'd be nice to know which commit on head fixed this. (In reply to Mikael Urankar from comment #24) Started a test build with that bootstrap toolchain. Seems like it doesn't fail immediately anymore, which is definitely a big plus. I'll let you know if it runs through! The build went almost to completion and finally crashed as follows:
thread 'rustc' panicked at compiler/rustc_middle/src/ty/adt.rs:154:20:
already borrowed: BorrowMutError
stack backtrace:
0: rust_begin_unwind
1: core::panicking::panic_fmt
2: core::cell::panic_already_borrowed
3: <rustc_middle::ty::adt::AdtDefData as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable
4: <rustc_type_ir::ty_info::WithCachedTypeInfo<rustc_type_ir::ty_kind::TyKind<rustc_middle::ty::context::TyCtxt>> as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable
5: <rustc_type_ir::ty_info::WithCachedTypeInfo<rustc_type_ir::ty_kind::TyKind<rustc_middle::ty::context::TyCtxt>> as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable
6: <&rustc_middle::ty::list::RawList<(), rustc_middle::ty::Ty> as rustc_data_structures::stable_hasher::HashStable<rustc_query_system::ich::hcx::StableHashingContext>>::hash_stable
7: rustc_symbol_mangling::legacy::mangle
8: rustc_symbol_mangling::symbol_name_provider
[... omitted 2 frames ...]
9: <hashbrown::raw::RawIterRange<(rustc_span::def_id::LocalDefId, ())>>::fold_impl::<<hashbrown::map::Iter<rustc_span::def_id::LocalDefId, ()> as core::iter::traits::iterator::Iterator>::fold<(), <hashbrown::map::Keys<rustc_span::def_id::LocalDefId, ()> as core::iter::traits::iterator::Iterator>::fold<(), core::iter::adapters::filter_map::filter_map_fold<&rustc_span::def_id::LocalDefId, rustc_span::def_id::LocalDefId, (), rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider::{closure#0}, core::iter::adapters::map::map_fold<rustc_span::def_id::LocalDefId, (rustc_span::def_id::DefId, rustc_middle::middle::exported_symbols::SymbolExportInfo), (), rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider::{closure#1}, core::iter::traits::iterator::Iterator::for_each::call<(rustc_span::def_id::DefId, rustc_middle::middle::exported_symbols::SymbolExportInfo), <hashbrown::map::HashMap<rustc_span::def_id::DefId, rustc_middle::middle::exported_symbols::SymbolExportInfo, core::hash::BuildHasherDefault<rustc_hash::FxHasher>> as core::iter::traits::collect::Extend<(rustc_span::def_id::DefId, rustc_middle::middle::exported_symbols::SymbolExportInfo)>>::extend<core::iter::adapters::map::Map<core::iter::adapters::filter_map::FilterMap<std::collections::hash::set::Iter<rustc_span::def_id::LocalDefId>, rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider::{closure#0}>, rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider::{closure#1}>>::{closure#0}>::{closure#0}>::{closure#0}>::{closure#0}>::{closure#0}>::{closure#0}, ()>
10: rustc_codegen_ssa::back::symbol_export::reachable_non_generics_provider
[... omitted 2 frames ...]
11: rustc_codegen_ssa::back::symbol_export::exported_symbols_provider_local
[... omitted 2 frames ...]
12: <rustc_metadata::rmeta::encoder::EncodeContext>::encode_crate_root
13: rustc_metadata::rmeta::encoder::encode_metadata
14: rustc_metadata::fs::encode_and_write_metadata
15: rustc_interface::passes::start_codegen
16: <rustc_interface::queries::Linker>::codegen_and_build_linker
17: <rustc_middle::ty::context::GlobalCtxt>::enter::<rustc_driver_impl::run_compiler::{closure#0}::{closure#1}::{closure#6}, core::result::Result<core::option::Option<rustc_interface::queries::Linker>, rustc_span::ErrorGuaranteed>>
18: <rustc_interface::queries::QueryResult<&rustc_middle::ty::context::GlobalCtxt>>::enter::<core::result::Result<core::option::Option<rustc_interface::queries::Linker>, rustc_span::ErrorGuaranteed>, rustc_driver_impl::run_compiler::{closure#0}::{closure#1}::{closure#6}>
19: <rustc_interface::interface::Compiler>::enter::<rustc_driver_impl::run_compiler::{closure#0}::{closure#1}, core::result::Result<core::option::Option<rustc_interface::queries::Linker>, rustc_span::ErrorGuaranteed>>
20: rustc_span::create_session_globals_then::<core::result::Result<(), rustc_span::ErrorGuaranteed>, rustc_interface::util::run_in_thread_with_globals<rustc_interface::interface::run_compiler<core::result::Result<(), rustc_span::ErrorGuaranteed>, rustc_driver_impl::run_compiler::{closure#0}>::{closure#1}, core::result::Result<(), rustc_span::ErrorGuaranteed>>::{closure#0}::{closure#0}::{closure#0}>
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
error: the compiler unexpectedly panicked. this is a bug.
note: we would appreciate a bug report: https://github.com/rust-lang/rust/issues/new?labels=C-bug%2C+I-ICE%2C+T-compiler&template=ice.md
note: please make sure that you have updated to the latest nightly
note: please attach the file at `/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/library/rustc-ice-2024-11-16T20_44_49-2364.txt` to your bug report
note: compiler flags: --crate-type lib -C opt-level=3 -C embed-bitcode=no -C linker=cc -C symbol-mangling-version=legacy -Z unstable-options -Z macro-backtrace -C split-debuginfo=off -C prefer-dynamic -Z inline-mir -Z inline-mir-preserve-debug -C embed-bitcode=yes -C force-frame-pointers=yes -Z crate-attr=doc(html_root_url="https://doc.rust-lang.org/1.82.0/") -Z binary-dep-depinfo -Z force-unstable-if-unmarked
note: some of the compiler flags provided by cargo are hidden
query stack during panic:
#0 [symbol_name] computing the symbol for `crate::fmt::num::imp::<impl at core/src/fmt/num.rs:480:13: 480:38>::fmt`
#1 [reachable_non_generics] looking up the exported symbols of a crate
#2 [exported_symbols] collecting exported symbols for crate `0`
end of query stack
Did not run successfully: exit status: 101
LD_LIBRARY_PATH="/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2/lib:/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2-std/release/deps:/usr/home/fuz/lib" "/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2/bin/rustc" "--crate-name" "core" "--edition=2021" "core/src/lib.rs" "--error-format=json" "--json=diagnostic-rendered-ansi,artifacts,future-incompat" "--diagnostic-width=119" "--crate-type" "lib" "--emit=dep-info,metadata,link" "-C" "opt-level=3" "-C" "embed-bitcode=no" "--warn=unexpected_cfgs" "--check-cfg" "cfg(docsrs)" "--check-cfg" "cfg(feature, values(\"debug_refcell\", \"optimize_for_size\", \"panic_immediate_abort\"))" "--check-cfg" "cfg(bootstrap)" "--check-cfg" "cfg(no_fp_fmt_parse)" "--check-cfg" "cfg(stdarch_intel_sde)" "--check-cfg" "cfg(feature, values(any()))" "-C" "metadata=4122020641cb3408" "-C" "extra-filename=-4122020641cb3408" "--out-dir" "/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2-std/wasm32-unknown-unknown/release/deps" "--target" "wasm32-unknown-unknown" "-C" "linker=cc" "-L" "dependency=/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2-std/wasm32-unknown-unknown/release/deps" "-L" "dependency=/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2-std/release/deps" "-Csymbol-mangling-version=legacy" "--check-cfg=cfg(feature,values(any()))" "-Zunstable-options" "--check-cfg=cfg(bootstrap)" "-Zmacro-backtrace" "-Csplit-debuginfo=off" "-Cprefer-dynamic" "-Zinline-mir" "-Zinline-mir-preserve-debug" "-Cembed-bitcode=yes" "-Cforce-frame-pointers=yes" "-Zcrate-attr=doc(html_root_url=\"https://doc.rust-lang.org/1.82.0/\")" "-Z" "binary-dep-depinfo" "-Wrust_2018_idioms" "-Wunused_lifetimes" "--sysroot" "/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2" "--remap-path-prefix" "/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src=/rustc/f6e511eec7342f59a25f7c0534f1dbea00d01b14" "--remap-path-prefix" "/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/vendor=/rust/deps" "-Z" "force-unstable-if-unmarked"
-------------
warning: `core` (lib) generated 1 warning (1 duplicate)
error: could not compile `core` (lib); 1 warning emitted
Caused by:
process didn't exit successfully: `CARGO=/usr/home/ports/main.ports/lang/rust/work/bootstrap/bin/cargo CARGO_CRATE_NAME=core CARGO_MANIFEST_DIR=/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/library/core CARGO_PKG_AUTHORS='' CARGO_PKG_DESCRIPTION='The Rust Core Library' CARGO_PKG_HOMEPAGE='' CARGO_PKG_LICENSE='MIT OR Apache-2.0' CARGO_PKG_LICENSE_FILE='' CARGO_PKG_NAME=core CARGO_PKG_README='' CARGO_PKG_REPOSITORY='https://github.com/rust-lang/rust.git' CARGO_PKG_RUST_VERSION='' CARGO_PKG_VERSION=0.0.0 CARGO_PKG_VERSION_MAJOR=0 CARGO_PKG_VERSION_MINOR=0 CARGO_PKG_VERSION_PATCH=0 CARGO_PKG_VERSION_PRE='' CARGO_RUSTC_CURRENT_DIR=/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/library LD_LIBRARY_PATH='/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2-std/release/deps:/usr/home/fuz/lib' /usr/home/ports/main.ports/lang/rust/work/_build/bootstrap/debug/rustc /usr/home/ports/main.ports/lang/rust/work/_build/bootstrap/debug/rustc --crate-name core --edition=2021 core/src/lib.rs --error-format=json --json=diagnostic-rendered-ansi,artifacts,future-incompat --diagnostic-width=119 --crate-type lib --emit=dep-info,metadata,link -C opt-level=3 -C embed-bitcode=no --warn=unexpected_cfgs --check-cfg 'cfg(docsrs)' --check-cfg 'cfg(feature, values("debug_refcell", "optimize_for_size", "panic_immediate_abort"))' --check-cfg 'cfg(bootstrap)' --check-cfg 'cfg(no_fp_fmt_parse)' --check-cfg 'cfg(stdarch_intel_sde)' --check-cfg 'cfg(feature, values(any()))' -C metadata=4122020641cb3408 -C extra-filename=-4122020641cb3408 --out-dir /usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2-std/wasm32-unknown-unknown/release/deps --target wasm32-unknown-unknown -C linker=cc -L dependency=/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2-std/wasm32-unknown-unknown/release/deps -L dependency=/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2-std/release/deps -Csymbol-mangling-version=legacy '--check-cfg=cfg(feature,values(any()))' -Zunstable-options '--check-cfg=cfg(bootstrap)' -Zmacro-backtrace -Csplit-debuginfo=off -Cprefer-dynamic -Zinline-mir -Zinline-mir-preserve-debug -Cembed-bitcode=yes -Cforce-frame-pointers=yes '-Zcrate-attr=doc(html_root_url="https://doc.rust-lang.org/1.82.0/")' -Z binary-dep-depinfo` (exit status: 101)
command did not execute successfully: cd "/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src" && env -u MAKEFLAGS -u MFLAGS AR_wasm32_unknown_unknown="ar" CARGO_INCREMENTAL="0" CARGO_PROFILE_RELEASE_DEBUG="0" CARGO_PROFILE_RELEASE_DEBUG_ASSERTIONS="false" CARGO_PROFILE_RELEASE_OVERFLOW_CHECKS="false" CARGO_PROFILE_RELEASE_STRIP="false" CARGO_TARGET_DIR="/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2-std" CARGO_TARGET_WASM32_UNKNOWN_UNKNOWN_LINKER="cc" CC_wasm32_unknown_unknown="cc" CFG_COMPILER_BUILD_TRIPLE="armv7-unknown-freebsd" CFG_COMPILER_HOST_TRIPLE="wasm32-unknown-unknown" CFG_DISABLE_UNSTABLE_FEATURES="1" CFG_RELEASE_CHANNEL="stable" CFG_VIRTUAL_RUST_SOURCE_BASE_DIR="/rustc/f6e511eec7342f59a25f7c0534f1dbea00d01b14" CFLAGS_wasm32_unknown_unknown="-ffunction-sections -fdata-sections -fPIC -pipe -fstack-protector-strong -fno-strict-aliasing" CXXFLAGS_wasm32_unknown_unknown="-ffunction-sections -fdata-sections -fPIC -pipe -fstack-protector-strong -fno-strict-aliasing" CXX_wasm32_unknown_unknown="c++" LIBC_CHECK_CFG="1" RANLIB_wasm32_unknown_unknown="ar s" REAL_LIBRARY_PATH="/usr/home/fuz/lib" REAL_LIBRARY_PATH_VAR="LD_LIBRARY_PATH" RUSTBUILD_NATIVE_DIR="/usr/home/ports/main.ports/lang/rust/work/_build/wasm32-unknown-unknown/native" RUSTC="/usr/home/ports/main.ports/lang/rust/work/_build/bootstrap/debug/rustc" RUSTC_BOOTSTRAP="1" RUSTC_BREAK_ON_ICE="1" RUSTC_CARGO_REGISTRY_SRC_TO_REMAP="/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/vendor=/rust/deps" RUSTC_DEBUGINFO_MAP="/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src=/rustc/f6e511eec7342f59a25f7c0534f1dbea00d01b14" RUSTC_ERROR_METADATA_DST="/usr/home/ports/main.ports/lang/rust/work/_build/tmp/extended-error-metadata" RUSTC_FORCE_UNSTABLE="1" RUSTC_HOST_FLAGS="-Zunstable-options --check-cfg=cfg(bootstrap) -Clinker=cc" RUSTC_INSTALL_BINDIR="bin" RUSTC_LIBDIR="/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2/lib" RUSTC_LINK_STD_INTO_RUSTC_DRIVER="1" RUSTC_LINT_FLAGS="-Wrust_2018_idioms -Wunused_lifetimes" RUSTC_REAL="/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2/bin/rustc" RUSTC_SNAPSHOT="/usr/home/ports/main.ports/lang/rust/work/bootstrap/bin/rustc" RUSTC_SNAPSHOT_LIBDIR="/usr/home/ports/main.ports/lang/rust/work/bootstrap/lib" RUSTC_STAGE="2" RUSTC_SYSROOT="/usr/home/ports/main.ports/lang/rust/work/_build/armv7-unknown-freebsd/stage2" RUSTC_VERBOSE="2" RUSTC_WRAPPER="/usr/home/ports/main.ports/lang/rust/work/_build/bootstrap/debug/rustc" RUSTDOC="/usr/home/ports/main.ports/lang/rust/work/_build/bootstrap/debug/rustdoc" RUSTDOCFLAGS="-Csymbol-mangling-version=legacy --check-cfg=cfg(feature,values(any())) -Zunstable-options --check-cfg=cfg(bootstrap) -Wrustdoc::invalid_codeblock_attributes --crate-version 1.82.0\t(f6e511eec\t2024-10-15)\t(built\tfrom\ta\tsource\ttarball) -Clinker=cc -Zcrate-attr=doc(html_root_url=\"https://doc.rust-lang.org/1.82.0/\") -Zcrate-attr=warn(rust_2018_idioms)" RUSTDOC_REAL="/path/to/nowhere/rustdoc/not/required" RUSTFLAGS="-Csymbol-mangling-version=legacy --check-cfg=cfg(feature,values(any())) -Zunstable-options --check-cfg=cfg(bootstrap) -Zmacro-backtrace -Csplit-debuginfo=off -Cprefer-dynamic -Zinline-mir -Zinline-mir-preserve-debug -Cembed-bitcode=yes -Cforce-frame-pointers=yes -Zcrate-attr=doc(html_root_url=\"https://doc.rust-lang.org/1.82.0/\")" RUST_COMPILER_RT_ROOT="/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/src/llvm-project/compiler-rt" RUST_TEST_THREADS="8" WINAPI_NO_BUNDLED_LIBRARIES="1" __CARGO_DEFAULT_LIB_METADATA="stablestd" "/usr/home/ports/main.ports/lang/rust/work/bootstrap/bin/cargo" "build" "--target" "wasm32-unknown-unknown" "--release" "-Zbinary-dep-depinfo" "-j" "8" "-v" "-v" "--frozen" "--features" " panic-unwind backtrace compiler-builtins-c" "--manifest-path" "/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/library/sysroot/Cargo.toml" "--message-format" "json-render-diagnostics"
expected success, got: exit status: 101
Traceback (most recent call last):
File "/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/x.py", line 50, in <module>
bootstrap.main()
File "/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/src/bootstrap/bootstrap.py", line 1208, in main
bootstrap(args)
File "/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/src/bootstrap/bootstrap.py", line 1184, in bootstrap
run(args, env=env, verbose=build.verbose, is_bootstrap=True)
File "/usr/home/ports/main.ports/lang/rust/work/rustc-1.82.0-src/src/bootstrap/bootstrap.py", line 195, in run
raise RuntimeError(err)
RuntimeError: failed to run: /usr/home/ports/main.ports/lang/rust/work/_build/bootstrap/debug/bootstrap dist --jobs=8
*** Error code 1
Stop.
make[1]: stopped in /usr/home/ports/main.ports/lang/rust
*** Error code 1
(In reply to Robert Clausecker from comment #26) Can you try to create a patch for has_thread_local = no and start over? I suppose the stage1 compiler has has_thread_local = yes (In reply to Mikael Urankar from comment #24) The following ends up reporting that the main vs. other distinction for building looks to go back to at least 2024-Apr-30 . Going in the direction of finding when the problem started, the first (earliest) fallout log for *releng-armv7-default for lang/rust consistently failing looks too be: Maintainer: rust@FreeBSD.org Log URL: https://pkg-status.freebsd.org/ampere3/data/132releng-armv7-default/26c5fe5d2fe6/logs/rust-1.77.0.log Build URL: https://pkg-status.freebsd.org/ampere3/build.html?mastername=132releng-armv7-default&build=26c5fe5d2fe6 Log: =>> Building lang/rust build started at Mon Apr 1 13:42:10 UTC 2024 port directory: /usr/ports/lang/rust package name: rust-1.77.0 building for: FreeBSD 132releng-armv7-default-job-02 13.2-RELEASE-p5 FreeBSD 13.2-RELEASE-p5 1302001 arm maintained by: rust@FreeBSD.org Makefile datestamp: -rw-r--r-- 1 root wheel 11562 Mar 28 01:02 /usr/ports/lang/rust/Makefile Ports top last git commit: 26c5fe5d2fe Ports top unclean checkout: no Port dir last git commit: 6f156f501ce Port dir unclean checkout: no Poudriere version: poudriere-git-3.4.1-1-g1e9f97d6 Host OSVERSION: 1500006 Jail OSVERSION: 1302001 Job Id: 02 . . . The details are different for what the logs look like for 1.77.0 vs. 1.78.0 and later. I'll note that we know the failures happen both on the release kernels as well as the 15????? kernels used on the package builders. This is despite main-armv7-default not getting the problem for its 15????? kernel use. Note the 13.2-RELEASE-p5 and 1302001, making it looks like that 13.2-RELEASE-p4 worked. So looking at the first 14.0 example of the armv7 failure: Maintainer: rust@FreeBSD.org Log URL: https://pkg-status.freebsd.org/ampere3/data/140releng-armv7-default/26c5fe5d2fe6/logs/rust-1.77.0.log Build URL: https://pkg-status.freebsd.org/ampere3/build.html?mastername=140releng-armv7-default&build=26c5fe5d2fe6 Log: =>> Building lang/rust build started at Mon Apr 8 03:40:28 UTC 2024 port directory: /usr/ports/lang/rust package name: rust-1.77.0 building for: FreeBSD 140releng-armv7-default-job-10 14.0-RELEASE-p5 FreeBSD 14.0-RELEASE-p5 1400097 arm maintained by: rust@FreeBSD.org Makefile datestamp: -rw-r--r-- 1 root wheel 11562 Mar 28 01:02 /usr/ports/lang/rust/Makefile Ports top last git commit: 26c5fe5d2fe Ports top unclean checkout: no Port dir last git commit: 6f156f501ce Port dir unclean checkout: no Poudriere version: poudriere-git-3.4.1-1-g1e9f97d6 Host OSVERSION: 1500006 Jail OSVERSION: 1400097 Job Id: 10 So: 14.0-RELEASE-p5 and 1400097 These lead to checking when 1.77.0 happend: author Mikael Urankar <mikael@FreeBSD.org> 2024-03-18 13:14:34 +0000 committer Mikael Urankar <mikael@FreeBSD.org> 2024-03-23 09:41:45 +0000 . . lang/rust: Update to 1.77.0 Announce: https://blog.rust-lang.org/2024/03/21/Rust-1.77.0.html ChangeLog: https://github.com/rust-lang/rust/releases/tag/1.77.0 PR: 277786 Tested by: mikael 1.77.00 log failure details vs. 1.78.0 and later are distinct. 1.78.0 stops sooner and lagnuage chnages seem to be involved in 1.77.0 reporting additional output. Unfortunately, main-armv7-default from 2024-Feb-28 until 2024-Jun-02 was broken by a system hangup problem unless the build happened to avoid the ports that happend to hit the problem. That happend once and main-armv7-default lang/rust 1.77.0_1 built okay for that one example: =>> Building lang/rust build started at Tue Apr 30 02:32:22 UTC 2024 port directory: /usr/ports/lang/rust package name: rust-1.77.0_1 building for: FreeBSD main-armv7-default-job-11 15.0-CURRENT FreeBSD 15.0-CURRENT 1500018 arm maintained by: rust@FreeBSD.org Makefile datestamp: -rw-r--r-- 1 root wheel 11241 Apr 28 01:05 /usr/ports/lang/rust/Makefile Ports top last git commit: 5a022ef6a46 Ports top unclean checkout: no Port dir last git commit: 54f7b9d654d Port dir unclean checkout: no Poudriere version: poudriere-git-3.4.1-30-g79e3edcd Host OSVERSION: 1500018 Jail OSVERSION: 1500018 Job Id: 11 In the general time frame, the non-main builds were failing. I do not find armv7 fallout reports from the months prior to 2024-Apr. This might be getting back to before armv7 was moved to the ampere*'s. Also, it might be that main was moved first and the other later. The available records do not seem to span any pre-ampere* armv7 build activity any more. This means no driect evidence for 1.76.0 or before of lang/rust as far as I can tell: both system and rust evidence is lacking. It's great that people finally believe me about lang/rust not building on armv7, when I was previously told that this can't possibly be the case. It would be great if the rust team could in the future conduct more comprehensive build tests and actually look at fallouts to catch these failures. I'm currently redoing the build test with a patch to disable TLS as requested, but that'll take a couple more hours. (In reply to Mikael Urankar from comment #24) I think found a potential distinction for main vs. not. It only has been done in main and yet was not recent at all. The removed #if . . . #endif (where the . . . was not removed) is still present in stable/14 . It is a change to the default thread-local storage module used: author R. Christian McDonald <rcm@rcm.sh> 2023-11-09 20:22:21 +0000 committer Kristof Provost <kp@FreeBSD.org> 2023-11-09 20:24:23 +0000 commit 6e5b1ff71e01bd48172483cb6df921f84300ea3a (patch) tree f68e0f494c8eca8fe83118e648563b6a9b64db26 parent ede4c412b3ea9289ef42c664b01b6b5ff7eac434 (diff) download src-6e5b1ff71e01bd48172483cb6df921f84300ea3a.tar.gz src-6e5b1ff71e01bd48172483cb6df921f84300ea3a.zip libc: enable initial-exec (IE) as default thread-local storage model on arm As suggested by jrtc27@ in https://reviews.freebsd.org/D42415, this patch enables IE as default thread-local storage model in libc on arm. Reviewed by: kib Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D42445 Diffstat -rw-r--r-- lib/libc/Makefile 4 1 files changed, 0 insertions, 4 deletions diff --git a/lib/libc/Makefile b/lib/libc/Makefile index 7540eb8c21ad..e817104642b8 100644 --- a/lib/libc/Makefile +++ b/lib/libc/Makefile @@ -54,11 +54,7 @@ CFLAGS+=${CANCELPOINTS_CFLAGS} # Use a more efficient TLS model for libc since we can reasonably assume that # it will be loaded during program startup. -.if ${LIBC_ARCH} == "aarch64" || ${LIBC_ARCH} == "amd64" || \ - ${LIBC_ARCH} == "i386" || ${LIBC_ARCH} == "riscv" || \ - ${LIBC_ARCH:Mpowerpc*} != "" CFLAGS+= -ftls-model=initial-exec -.endif # # Link with static libcompiler_rt.a. (In reply to Mark Millard from comment #30) Looks like the https://reviews.freebsd.org/D42415 that is mentioned might be related work done that helped enable the change to be supported. I've not checked the details for if those changes made to outside main. (In reply to Mark Millard from comment #31) Looks like at least: https://cgit.freebsd.org/src/commit/?id=98fd69f0090da73d9d0451bd769d7752468284c6 is part of the changes needed to have -ftls-model=initial-exec work for armv7. For reference, the diff it lists is: diff --git a/libexec/rtld-elf/arm/reloc.c b/libexec/rtld-elf/arm/reloc.c index c3e95940be74..6efc9f499761 100644 --- a/libexec/rtld-elf/arm/reloc.c +++ b/libexec/rtld-elf/arm/reloc.c @@ -280,10 +280,13 @@ reloc_nonplt_object(Obj_Entry *obj, const Elf_Rel *rel, SymCache *cache, return -1; tmp = (Elf_Addr)def->st_value + defobj->tlsoffset; - if (__predict_true(RELOC_ALIGNED_P(where))) + if (__predict_true(RELOC_ALIGNED_P(where))) { + tmp += *where; *where = tmp; - else + } else { + tmp += load_ptr(where); store_ptr(where, tmp); + } dbg("TLS_TPOFF32 %s in %s --> %p", obj->strtab + obj->symtab[symnum].st_name, obj->path, (void *)tmp); (In reply to Mikael Urankar from comment #27) I can confirm that the updated bootstrap plus the requested patch lead to a succesfull lang/rust build on armv7 FreeBSD 14.1! Finally! Could you commit a patch to this effect and MFH it for good measure? *** Bug 280027 has been marked as a duplicate of this bug. *** (In reply to Mark Millard from comment #32) I created a poudriere-devel jail based on the 2 commits missing from all but main and the updated bootstrap that had the "gnu" removal. This is on the stable/14 boot media. This does not have your thread local storage removals or the like. The build is still on-going: it did not stop quickly. URL: https://cgit.FreeBSD.org/ports/commit/?id=560fd669d24aafad437789954800d57e7cfeafa9 commit 560fd669d24aafad437789954800d57e7cfeafa9 Author: Mikael Urankar <mikael@FreeBSD.org> AuthorDate: 2024-11-14 14:59:30 +0000 Commit: Mikael Urankar <mikael@FreeBSD.org> CommitDate: 2024-11-14 15:30:31 +0000 lang/rust: Regen bootstrap for armv7 PR: 282663 lang/rust/distinfo | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) [Note: Previously, with just that change it still got errors, just not the multiple definition errors and the like.] What is new is the following changes to the world build: # git -C /usr/official-src/ diff libexec/rtld-elf/arm/reloc.c diff --git a/libexec/rtld-elf/arm/reloc.c b/libexec/rtld-elf/arm/reloc.c index c3e95940be74..6efc9f499761 100644 --- a/libexec/rtld-elf/arm/reloc.c +++ b/libexec/rtld-elf/arm/reloc.c @@ -280,10 +280,13 @@ reloc_nonplt_object(Obj_Entry *obj, const Elf_Rel *rel, SymCache *cache, return -1; tmp = (Elf_Addr)def->st_value + defobj->tlsoffset; - if (__predict_true(RELOC_ALIGNED_P(where))) + if (__predict_true(RELOC_ALIGNED_P(where))) { + tmp += *where; *where = tmp; - else + } else { + tmp += load_ptr(where); store_ptr(where, tmp); + } dbg("TLS_TPOFF32 %s in %s --> %p", obj->strtab + obj->symtab[symnum].st_name, obj->path, (void *)tmp); # git -C /usr/official-src/ diff lib/libc/Makefile diff --git a/lib/libc/Makefile b/lib/libc/Makefile index fbfd6619784d..d887e0808ceb 100644 --- a/lib/libc/Makefile +++ b/lib/libc/Makefile @@ -54,11 +54,7 @@ CFLAGS+=${CANCELPOINTS_CFLAGS} # Use a more efficient TLS model for libc since we can reasonably assume that # it will be loaded during program startup. -.if ${LIBC_ARCH} == "aarch64" || ${LIBC_ARCH} == "amd64" || \ - ${LIBC_ARCH} == "i386" || ${LIBC_ARCH} == "riscv" || \ - ${LIBC_ARCH:Mpowerpc*} != "" CFLAGS+= -ftls-model=initial-exec -.endif # # Link with static libcompiler_rt.a. So far the build is suggestive of being able to complete but we will just have to wait and see. If all goes well: Too bad about the timing relative to 14.2-STABLE. (In reply to Mark Millard from comment #35) Dumb typo. Fix: ". . . relative to 14.2-RELEASE." (In reply to Mikael Urankar from comment #24) (In reply to Mark Millard from comment #36) lang/rust 1.82.0_1 built just fine on/for stable/14 based on the 3 points: ) The lack of the "gnu" target_env issue for the bootstrap rustc. ) The libexec/rtld-elf/arm/reloc.c fix for -ftls-model=initial-exec support. ) The change to when 'CFLAGS+= -ftls-model=initial-exec' is used in lib/libc/Makefile . For reference: [00:29:39] [01] [00:00:00] Building lang/rust | rust-1.82.0_1 [02:59:41] [01] [02:30:02] Finished lang/rust | rust-1.82.0_1: Success # pkg-static -r/stable-armv7-alt-chroot/ add /usr/local/poudriere/data/packages/stable-armv7-alt-default/All/rust-1.82.0_1.pkg Installing rust-1.82.0_1... `-- Installing curl-8.10.1... | `-- Installing libnghttp2-1.64.0... | `-- Extracting libnghttp2-1.64.0: 100% | `-- Installing libpsl-0.21.5_1... | | `-- Installing libidn2-2.3.7... | | `-- Installing indexinfo-0.3.1... | | `-- Extracting indexinfo-0.3.1: 100% | | `-- Installing libunistring-1.2... | | `-- Extracting libunistring-1.2: 100% | | `-- Extracting libidn2-2.3.7: 100% | `-- Extracting libpsl-0.21.5_1: 100% | `-- Installing libssh2-1.11.1,3... | `-- Extracting libssh2-1.11.1,3: 100% `-- Extracting curl-8.10.1: 100% Extracting rust-1.82.0_1: 100% # ~/do-chroot-stable-armv7-alt-chroot.sh # rustc -vV rustc 1.82.0 (f6e511eec 2024-10-15) (built from a source tarball) binary: rustc commit-hash: f6e511eec7342f59a25f7c0534f1dbea00d01b14 commit-date: 2024-10-15 host: armv7-unknown-freebsd release: 1.82.0 LLVM version: 19.1.1 # rustc --print=cfg debug_assertions panic="unwind" target_abi="eabihf" target_arch="arm" target_endian="little" target_env="" target_family="unix" target_has_atomic="16" target_has_atomic="32" target_has_atomic="64" target_has_atomic="8" target_has_atomic="ptr" target_os="freebsd" target_pointer_width="32" target_vendor="unknown" unix (In reply to Mark Millard from comment #37) Note, both: /stable-armv7-alt-chroot /stable/armv7-alt-poud (the poudriere jail stable-armv7-alt 's world) are populated with installs of the world built with the 2 patches involved. (In reply to Mark Millard from comment #38) [Just correcting the mistaken "/" to be a "-".] Note, both: /stable-armv7-alt-chroot /stable-armv7-alt-poud (the poudriere jail stable-armv7-alt 's world) are populated with installs of the world built with the 2 patches involved. (In reply to Mark Millard from comment #37) Thank you! I'll note that the main-armv7-default official package builder run that recently started had lang/rust fail while fetching. https://pkg-status.freebsd.org/ampere2/data/main-armv7-default/pdfee61567d9e_s5036d9652a/logs/errors/rust-1.82.0_1.log reports: . . . =======================<phase: fetch >============================ ===== env: NO_DEPENDS=yes USER=root UID=0 GID=0 ===> License APACHE20 MIT accepted by the user => 2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz doesn't seem to exist in /portdistfiles/rust. => Attempting to fetch http://distcache.FreeBSD.org/local-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz fetch: http://distcache.FreeBSD.org/local-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz: size mismatch: expected 61790324, actual 61737980 => Attempting to fetch http://distcache.us-east.FreeBSD.org/local-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz fetch: http://distcache.us-east.FreeBSD.org/local-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz: size mismatch: expected 61790324, actual 61737980 => Attempting to fetch http://distcache.eu.FreeBSD.org/local-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz fetch: http://distcache.eu.FreeBSD.org/local-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz: size mismatch: expected 61790324, actual 61737980 => Attempting to fetch http://distcache.us-west.FreeBSD.org/local-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz fetch: http://distcache.us-west.FreeBSD.org/local-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz: size mismatch: expected 61790324, actual 61737980 => Attempting to fetch https://static.rust-lang.org/dist/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz fetch: https://static.rust-lang.org/dist/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz: Not Found => Attempting to fetch http://distcache.FreeBSD.org/ports-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz fetch: http://distcache.FreeBSD.org/ports-distfiles/rust/2024-09-05/rustc-1.81.0-armv7-unknown-freebsd.tar.xz: Not Found => Couldn't fetch it - please try to retrieve this => port manually into /portdistfiles/rust and try again. *** Error code 1 . . . (In reply to Mark Millard from comment #41) May be that is just the ports snapshot it was working from being older than, for example, the bootstrap update. (main-arm64-default and then main-armv7-default use the same snapshot as a pair but bulk -a times are longer than for amd64 so the time frame ends up being older for main-armv7-default by the time it actually starts its build.) (In reply to Mikael Urankar from comment #27) Now that we've confirmed that this is the fix, could you commit something to lang/rust that updates to your regenerated bootstrap and adds the necessary patch to make the build succeed on armv7 FreeBSD 14.1? Please MFH if possible. A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/ports/commit/?id=c348a547923c25a771d0d2a5db796103e25511ec commit c348a547923c25a771d0d2a5db796103e25511ec Author: Mikael Urankar <mikael@FreeBSD.org> AuthorDate: 2024-11-22 08:40:03 +0000 Commit: Mikael Urankar <mikael@FreeBSD.org> CommitDate: 2024-11-22 08:40:03 +0000 lang/rust: Fix build on armv7 PR: 282663 lang/rust/distinfo | 12 ++++++------ ...tch-compiler_rustc__target_src_spec_base_freebsd.rs (new) | 11 +++++++++++ 2 files changed, 17 insertions(+), 6 deletions(-) FYI: Today stable/14 and stable/13 got the updates that are what had main [so: 15 as stands] working: git: d39e0bdc6b76 - stable/14 - rtld/arm: fix initial-exec (IE) thread-local storage relocation R. Christian McDonald git: f424c4a907e1 - stable/14 - libc: enable initial-exec (IE) as default thread-local storage model on arm R. Christian McDonald git: e10cca68cf34 - stable/13 - rtld/arm: fix initial-exec (IE) thread-local storage relocation R. Christian McDonald git: c364608261d1 - stable/13 - libc: enable initial-exec (IE) as default thread-local storage model on arm R. Christian McDonald So: at least 14.3-RELEASE and 13.5-RELEASE (if there is such) should end up fixed for such issues. (lang/rust need not be the only example of an issue.) It may be too late for 14.2-RELEASE . (In reply to commit-hook from comment #44) Could you MFH these updates if possible? |
This was a poudriere-devel based build after updating my /usr/ports/ via git. I did this because someone asked if I would also get build errors. The context is aarch64 with an armv7 poudriere jail on a Windows Dev Kit 2023. It supports user armv7 code. FreeBSD is main [so: 15]. The build log file reports . . . error[E0428]: the name `type_of_thread_id` is defined multiple times --> /rust/deps/nix-0.29.0/src/sys/signal.rs:1110:1 | 1107 | pub type type_of_thread_id = libc::lwpid_t; | ------------------------------------------- previous definition of the type `type_of_thread_id` here ... 1110 | pub type type_of_thread_id = libc::pid_t; | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ `type_of_thread_id` redefined here | = note: `type_of_thread_id` must be defined only once in the type namespace of this module . . . error[E0609]: no field `sigev_notify_thread_id` on type `signal::sigevent::ffi::sigevent` --> /rust/deps/nix-0.29.0/src/sys/signal.rs:1364:25 | 1364 | sev.sigev_notify_thread_id = thread_id; | ^^^^^^^^^^^^^^^^^^^^^^ unknown field | = note: available fields are: `sigev_notify`, `sigev_signo`, `sigev_value`, `_sigev_un` Some errors have detailed explanations: E0428, E0609. For more information about an error, try `rustc --explain E0428`. . . . ------------- warning: `nix` (lib) generated 9 warnings error: could not compile `nix` (lib) due to 2 previous errors; 9 warnings emitted [I ignore later reporting that might be consequences of the above earlier reports of errors.] For reference: # ~/fbsd-based-on-what-commit.sh -C /usr/ports/ 716a5651faed (HEAD -> main, freebsd/main, freebsd/HEAD) net-mgmt/rubygem-oxidized-web: Fix dependency on www/rubygem-haml Author: Einar Bjarni Halldórsson <einar@isnic.is> Commit: Juraj Lutter <otis@FreeBSD.org> CommitDate: 2024-11-09 20:49:29 +0000 branch: main merge-base: 716a5651faedef308ee5c267c381d6dd4483d88c merge-base: CommitDate: 2024-11-09 20:49:29 +0000 n685665 (--first-parent --count for merge-base) # uname -apKU FreeBSD aarch64-main-pbase 15.0-CURRENT FreeBSD 15.0-CURRENT main-n273174-8b2e7da70855 GENERIC-NODEBUG arm64 aarch64 1500026 1500026 (That is an install of an official PkgBase kernel that was in use.)