Messages in this thread | | | Date | Mon, 3 Nov 2014 12:12:18 -0500 (EST) | From | Vince Weaver <> | Subject | Re: [RFD] perf syscall error handling |
| |
On Mon, 3 Nov 2014, Arnaldo Carvalho de Melo wrote:
> Em Mon, Nov 03, 2014 at 05:50:19PM +0100, Peter Zijlstra escreveu: > > On Mon, Nov 03, 2014 at 02:25:48PM -0200, Arnaldo Carvalho de Melo wrote: > > > > The way that peterz suggested, i.e. returning information about which > > > perf_event_attr and which of the parameters was invalid/had issues could > > > help with fallbacking/capability querying, i.e. tooling may want to use > > > some features if available automagically, fallbacking to something else > > > when that fails. > > > > We already do that to some degree in various cases, but for some if the > > > only way that becomes available to disambiguate some EINVAL return is a > > > string, code will start having strcmps :-\ > > > OK, so how about we do both, the offset+mask for the tools and the > > string for the humans? > > Yeah, tooling tries to provide the best it can with the offset+mask, and > if doesn't manage to do anything smart with it, just show the string and > hope that helps the user to figure out what is happening.
I don't know if having an offset/mask helps much. Knowing your EINVAL comes from ->config is nice to know, but if there's 30 different ways to get an EINVAL from an improper config then you still can waste a lot of time narrowing things down.
The string solution might be nice, but it is going to take major changes to the code and increase the size a bit. For example:
$ cat arch/x86/kernel/cpu/perf* kernel/events/* | grep EINVAL | wc -l 100
And some of the code is passing the return values back through various long callchains (and overloaded pointers via casts) where it's not clear how you could also pass a string value.
Vince
| |