entity-manager terminated by 'std::filesystem::__cxx11::filesystem_error'

[leiyu-bytedance]

The issue is found by g220a's QEMU, that the entity-manager is terminated by 'std::filesystem::__cxx11::filesystem_error':

Jan 01 00:00:47 g220a entity-manager[312]: Clearing previous configuration
Jan 01 00:01:35 g220a entity-manager[312]: configuration file missing probe:
Jan 01 00:01:35 g220a entity-manager[312]:  {"buses":[]}
Jan 01 00:01:36 g220a entity-manager[312]: Inventory Added
Jan 01 00:01:36 g220a entity-manager[312]: terminate called after throwing an instance of 'std::filesystem::__cxx11::filesystem_error'
Jan 01 00:01:36 g220a entity-manager[312]:   what():  filesystem error: recursive directory iterator cannot open directory: No such file or directory [/sys/bus/i2c/devices/i2c-26]

The issue is introduced by 9b86787, if we revert the commit, the issue is gone.

[Jayashree-D]

I am also facing this issue in tiogapass.

root@tiogapass:~# systemctl status xyz.openbmc_project.EntityManager.service -l
● xyz.openbmc_project.EntityManager.service - Entity Manager
Loaded: loaded (�]8;;file://tiogapass/lib/systemd/system/xyz.openbmc_project.EntityManager.service/lib/systemd/system/xyz.openbmc_project.EntityManager.service�]8;;; enabled; vendor preset: enabled)
Active: active (running) since Thu 1970-01-01 00:04:10 UTC; 25s ago
Process: 851 ExecStartPre=/bin/mkdir -p /var/configuration (code=exited, status=0/SUCCESS)
Process: 852 ExecStartPre=/bin/mkdir -p /tmp/overlays (code=exited, status=0/SUCCESS)
Main PID: 853 (entity-manager)
CGroup: /system.slice/xyz.openbmc_project.EntityManager.service
└─853 /usr/bin/entity-manager

Jan 01 00:04:09 tiogapass systemd[1]: Starting Entity Manager...
Jan 01 00:04:10 tiogapass systemd[1]: Started Entity Manager.
Jan 01 00:04:27 tiogapass entity-manager[853]: Inventory Added
Jan 01 00:04:27 tiogapass entity-manager[853]: terminate called after throwing an instance of 'std::filesystem::__cxx11::filesystem_error'
Jan 01 00:04:27 tiogapass entity-manager[853]: what(): filesystem error: recursive directory iterator cannot open directory: No such file or directory [/sys/bus/i2c/devices/i2c-16]

[leiyu-bytedance]

Here is the full stack back trace:

#0  __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:49
#1  0x76b26f10 in __GI_abort () at abort.c:79
#2  0x76dabf60 in __gnu_cxx::__verbose_terminate_handler () at ../../../../../../../../../work-shared/gcc-10.2.0-r0/gcc-10.2.0/libstdc++-v3/libsupc++/vterminate.cc:95
#3  0x76da9aec in __cxxabiv1::__terminate (handler=<optimized out>) at ../../../../../../../../../work-shared/gcc-10.2.0-r0/gcc-10.2.0/libstdc++-v3/libsupc++/eh_terminate.cc:48
#4  0x76da9b78 in std::terminate () at ../../../../../../../../../work-shared/gcc-10.2.0-r0/gcc-10.2.0/libstdc++-v3/libsupc++/eh_terminate.cc:58
#5  0x76da9f30 in __cxxabiv1::__cxa_throw (obj=obj@entry=0x1e32ae0, tinfo=0x76eae188 <typeinfo for std::filesystem::__cxx11::filesystem_error>,
    dest=0x76e643bc <std::filesystem::__cxx11::filesystem_error::~filesystem_error()>) at ../../../../../../../../../work-shared/gcc-10.2.0-r0/gcc-10.2.0/libstdc++-v3/libsupc++/eh_throw.cc:95
#6  0x76e591bc in std::filesystem::__cxx11::recursive_directory_iterator::recursive_directory_iterator (this=0x7eb50424, p=..., options=<optimized out>, ecptr=<optimized out>)
    at /usr/src/debug/gcc-runtime/10.2.0-r0/arm-openbmc-linux-gnueabi/libstdc++-v3/include/ext/new_allocator.h:89
#7  0x00525b50 in std::filesystem::__cxx11::recursive_directory_iterator::recursive_directory_iterator (__p=..., this=0x7eb5041c) at /usr/include/c++/10.2.0/bits/fs_dir.h:473
#8  deviceIsCreated (devicePath=..., bus=..., address=..., retrying=retrying@entry=false) at ../git/src/Overlay.cpp:172
#9  0x00528184 in buildDevice (devicePath=..., parameters=..., bus=..., address=..., constructor=..., destructor=..., createsHWMon=true, createsHWMon@entry=156, retries=retries@entry=5)
    at ../git/src/Overlay.cpp:236
#10 0x0052986c in exportDevice (type=..., exportTemplate=..., configuration=...) at ../git/src/Overlay.cpp:322
#11 0x0052a91c in loadOverlays (systemConfiguration=...) at ../git/src/Overlay.cpp:363
#12 0x004a7670 in operator() (__closure=0x7eb50a78) at ../git/src/EntityManager.cpp:1830
#13 boost::asio::asio_handler_invoke<propertiesChangedCallback(nlohmann::json&, sdbusplus::asio::object_server&)::<lambda(const boost::system::error_code&)>::<lambda()>::<lambda()> > (
    function=...) at /usr/include/boost/asio/handler_invoke_hook.hpp:88
#14 boost_asio_handler_invoke_helpers::invoke<propertiesChangedCallback(nlohmann::json&, sdbusplus::asio::object_server&)::<lambda(const boost::system::error_code&)>::<lambda()>::<lambda()>, propertiesChangedCallback(nlohmann::json&, sdbusplus::asio::object_server&)::<lambda(const boost::system::error_code&)>::<lambda()>::<lambda()> > (context=..., function=...)
    at /usr/include/boost/asio/detail/handler_invoke_helpers.hpp:54
#15 boost::asio::detail::handler_work<propertiesChangedCallback(nlohmann::json&, sdbusplus::asio::object_server&)::<lambda(const boost::system::error_code&)>::<lambda()>::<lambda()>, boost::asio::io_context::basic_executor_type<std::allocator<void>, 0>, void>::complete<propertiesChangedCallback(nlohmann::json&, sdbusplus::asio::object_server&)::<lambda(const boost::system::error_code&)>::<lambda()>::<lambda()> > (handler=..., function=..., this=<synthetic pointer>) at /usr/include/boost/asio/detail/handler_work.hpp:425
#16 boost::asio::detail::completion_handler<propertiesChangedCallback(nlohmann::json&, sdbusplus::asio::object_server&)::<lambda(const boost::system::error_code&)>::<lambda()>::<lambda()>, boost::asio::io_context::basic_executor_type<std::allocator<void>, 0> >::do_complete(void *, boost::asio::detail::operation *, const boost::system::error_code &, std::size_t) (
    owner=owner@entry=0x1da7150, base=0x1da8d58) at /usr/include/boost/asio/detail/completion_handler.hpp:74
#17 0x0049da0c in boost::asio::detail::scheduler_operation::complete (bytes_transferred=0, ec=..., owner=0x1da7150, this=<optimized out>)
    at /usr/include/boost/asio/detail/scheduler_operation.hpp:40
#18 boost::asio::detail::scheduler::do_run_one (ec=..., this_thread=..., lock=..., this=0x1da7150) at /usr/include/boost/asio/detail/impl/scheduler.ipp:481
#19 boost::asio::detail::scheduler::run (ec=..., this=0x1da7150) at /usr/include/boost/asio/detail/impl/scheduler.ipp:204
#20 boost::asio::io_context::run (this=0x5c326c <io>, this=0x5c326c <io>) at /usr/include/boost/asio/impl/io_context.ipp:63
#21 main () at ../git/src/EntityManager.cpp:1970

[edtanous]

I suspect this is the fix, although to hit this failure, your config files would need to reference a bus that doesn't exist, so I'm guessing this wont fix your system itself, it will just avoid the crash.
https://gerrit.openbmc-project.xyz/c/openbmc/entity-manager/+/42813

[edtanous]

I'm closing as this has been merged. If this is still an issue, please reopen.