lkml.org 
[lkml]   [2019]   [Oct]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH v3 1/9] perf tools: add parse events append error
On Fri, Oct 25, 2019 at 08:14:36AM -0700, Ian Rogers wrote:
> On Fri, Oct 25, 2019 at 12:58 AM Jiri Olsa <jolsa@redhat.com> wrote:
> >
> > On Thu, Oct 24, 2019 at 12:01:54PM -0700, Ian Rogers wrote:
> > > Parse event error handling may overwrite one error string with another
> > > creating memory leaks and masking errors. Introduce a helper routine
> > > that appends error messages and avoids the memory leak.
> > >
> > > A reproduction of this problem can be seen with:
> > > perf stat -e c/c/
> > > After this change this produces:
> > > event syntax error: 'c/c/'
> > > \___ unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: unknown term (previous error: Cannot find PMU `c'. Missing kernel support?)(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,pc,in_tx,edge,any,offcore_rsp,in_tx_cp,ldlat,inv,umask,frontend,cmask,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))(help: valid terms: event,filter_rem,filter_opc0,edge,filter_isoc,filter_tid,filter_loc,filter_nc,inv,umask,filter_opc1,tid_en,thresh,filter_all_op,filter_not_nm,filter_state,filter_nm,config,config1,config2,name,period,percore))
> >
> >
> > hum... I'd argue that the previous state was better:
> >
> > [jolsa@krava perf]$ ./perf stat -e c/c/
> > event syntax error: 'c/c/'
> > \___ unknown term
> >
> >
> > jirka
>
> I am agnostic. We can either have the previous state or the new state,
> I'm keen to resolve the memory leak. Another alternative is to warn
> that multiple errors have occurred before dropping or printing the
> previous error. As the code is shared in memory places the approach
> taken here was to try to not conceal anything that could potentially
> be useful. Given this, is the preference to keep the status quo
> without any warning?

if the other alternative is string above, yes.. but perhaps
keeping just the first error would be the best way?

here it seems to be the:
"Cannot find PMU `c'. Missing kernel support?)(help: valid..."

jirka

\
 
 \ /
  Last update: 2019-10-28 20:33    [W:0.092 / U:2.080 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site