Weibing Wang comments

Results 115 comments of


                                            Weibing Wang

构建失败

可以试试这个PR：https://github.com/apache/incubator-brpc/pull/1657

Sometimes SetFailed() is called with errno=0 in InputMessenger::OnNewMessages()

> Another piece of debugging log supports this scenario. Could you show this log?

Sometimes SetFailed() is called with errno=0 in InputMessenger::OnNewMessages()

This issue may be related to https://github.com/apache/incubator-brpc/issues/1693. Although we have removed the `__const__` attribute of __errno_location(), but powerful link time optimization may still cache the tls errno location ?

Sometimes SetFailed() is called with errno=0 in InputMessenger::OnNewMessages()

> I observed a different coredump saying something like "bthread sched_to itself" I encountered this error before, that is related to tls cache. Compiler options: -fno-gcse, -fno-cse-follow-jumps, -fno-move-loop-invariants may fix...

Sometimes SetFailed() is called with errno=0 in InputMessenger::OnNewMessages()

> BTW-3: About half a year ago, when we are using 0.9.7, I roughly tested LTO capabilities once and BRPC wasn't a blocker at that time (the service behaves correctly...

Sometimes SetFailed() is called with errno=0 in InputMessenger::OnNewMessages()

We are using gcc 8 and gcc 10 with lto to compile brpc and everything works well. It seems that some new optimization in clang is not compatible with bthread,...

thrift协议支持多参数-讨论

之前有个PR，可以看看 #495

Why switch bthread cost so much time? (2us), both libgo and golang coroutine switch cost ~200ns?

The work stealing in libgo is implemented by a global scheduler: https://github.com/yyzybb537/libgo/blob/master/libgo/scheduler/scheduler.cpp#L340 It runs work stealing at an interval of 1000us by default. When a task is ready to run...

有支持io_uring的计划吗？

bthread里面没有读磁盘的操作吧，你是指在bthread中运行的业务代码中进行了读磁盘的操作吗？如果想减少阻塞，可以用异步的读，可以使用bthread_fd_wait来等待可读事件，等待过程中bthread会被挂起。