Date: Thu, 6 Jun 2013 19:47:18 +1000 (EST) From: Bruce Evans <brde@optusnet.com.au> To: Ryan Stone <rysto32@gmail.com> Cc: Navdeep Parhar <np@FreeBSD.org>, freebsd-arch@FreeBSD.org Subject: Re: missing DTrace FBT return probes Message-ID: <20130606191306.P2408@besplex.bde.org> In-Reply-To: <CAFMmRNzj2LhsdxankJV5CT6yHy8J%2BzyL-YBzFRUSKz%2Bq-ZTSNw@mail.gmail.com> References: <51AFB2B3.5050105@FreeBSD.org> <CAFMmRNzj2LhsdxankJV5CT6yHy8J%2BzyL-YBzFRUSKz%2Bq-ZTSNw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 5 Jun 2013, Ryan Stone wrote: > On Wed, Jun 5, 2013 at 5:50 PM, Navdeep Parhar <np@freebsd.org> wrote: > >> A large number of kernel functions have an FBT entry probe but no return >> probe. I believe this is due to tail call optimization by the compiler. >> Should we disable this optimization for kernel configs that have DTrace >> support? The missing return probes make it very difficult to write >> DTrace scripts that want to set flags etc. at function entry and then >> clean them up on return. >> ... > > I would be in favour of turning this on unconditionally, along with > -fno-inline-functions-called-once and -fno-omit-frame-pointer. Also -O2. But -fno-inline-functions-called-once isn't even supported by clang, and -O for clang is more like -O3 for gcc (it does excessive inlining of even more than functions called once). -fno-omit-frame-pointer is the default for gcc by apparently not for clang. > All of the > optimizations are of dubious value and significantly impact debugging tools > like dtrace and pmc. Also stack traces in panics and debuggers, debuggers generally (they can rarely find variables in inline functions, or even step over an inline function like a non-inline function), and profiling. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20130606191306.P2408>