Messages in this thread | | | Subject | Re: [PATCHv4 0/9] perf tool: parser generator for events parsing | From | Peter Zijlstra <> | Date | Wed, 15 Feb 2012 12:18:32 +0100 |
| |
On Wed, 2012-02-15 at 10:24 +0100, Jiri Olsa wrote: > On Tue, Feb 14, 2012 at 10:03:11PM +0100, Peter Zijlstra wrote:
> > These would be: period, config, config1, config2 and stephane's new > > branch_sample_type, although that might want to get a shorter name. > > right, for pmu/.../ syntax there are allowed only fields from 'format' > directory are so far.. with exceptions like: cycles/period=100000/ > > so I'll hardcode following config fields: > config > config1 > config2 > period > branch_sample_type (branch, branch_type, branch_st ???)
Uhmm,.. Stephane any particular preference on this?
> to be used in xxx/.../ syntax > > also we need to choose some strategy of format field name shadowing > with hardcoded fields: > > - BUILD_BUG in kernel PMU_FORMAT_ATTR > - not allowed.. report error in perf runtime > - allowed - hardcoded field have precedence > - allowed - format field values have precedence > > I'd say either allow shadowing(with whatever precedence we this is better), > or have the BUILD_BUG line in kernel..
I agree, if we can get the BUILD_BUG thing working that might be the best option, otherwise we can do the precedence thing. We could even add a syntax to resolve the namespace conflict in the latter case (eg. use '$' PE_NAME to mean the hardcoded in case of conflict).
| |