Re: [PATCH v2 11/13] selftests/resctrl: Convert ctrlgrp & mongrp to pointers

[Ilpo Järvinen]

On Fri, 22 Mar 2024, Reinette Chatre wrote:
> On 3/22/2024 5:30 AM, Ilpo Järvinen wrote:
> > On Wed, 20 Mar 2024, Reinette Chatre wrote:
> >> On 3/11/2024 6:52 AM, Ilpo Järvinen wrote:
> >>> The struct resctrl_val_param has control and monitor groups as char
> >>> arrays but they are not supposed to be mutated within resctrl_val().
> >>>
> >>> Convert the ctrlgrp and mongrp char array within resctrl_val_param to
> >>> plain const char pointers and adjust the strlen() based checks to
> >>> check NULL instead.
> >>>
> >>> Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>
> >>> ---
> >>>  tools/testing/selftests/resctrl/resctrl.h   | 4 ++--
> >>>  tools/testing/selftests/resctrl/resctrlfs.c | 8 ++++----
> >>>  2 files changed, 6 insertions(+), 6 deletions(-)
> >>>
> >>> diff --git a/tools/testing/selftests/resctrl/resctrl.h b/tools/testing/selftests/resctrl/resctrl.h
> >>> index 52769b075233..54e5bce4c698 100644
> >>> --- a/tools/testing/selftests/resctrl/resctrl.h
> >>> +++ b/tools/testing/selftests/resctrl/resctrl.h
> >>> @@ -89,8 +89,8 @@ struct resctrl_test {
> >>>   */
> >>>  struct resctrl_val_param {
> >>>  	char		*resctrl_val;
> >>> -	char		ctrlgrp[64];
> >>> -	char		mongrp[64];
> >>> +	const char	*ctrlgrp;
> >>> +	const char	*mongrp;
> >>>  	char		filename[64];
> >>>  	unsigned long	mask;
> >>>  	int		num_of_runs;
> >>> diff --git a/tools/testing/selftests/resctrl/resctrlfs.c b/tools/testing/selftests/resctrl/resctrlfs.c
> >>> index 79cf1c593106..dbe0cc6d74fa 100644
> >>> --- a/tools/testing/selftests/resctrl/resctrlfs.c
> >>> +++ b/tools/testing/selftests/resctrl/resctrlfs.c
> >>> @@ -469,7 +469,7 @@ static int create_grp(const char *grp_name, char *grp, const char *parent_grp)
> >>>  	 * length of grp_name == 0, it means, user wants to use root con_mon
> >>>  	 * grp, so do nothing
> >>>  	 */
> >>
> >> Could you please confirm that the comments are still accurate?
> > 
> > It's not, I missed it.
> > 
> >>> -	if (strlen(grp_name) == 0)
> >>> +	if (!grp_name)
> >>>  		return 0;
> > 
> > But now when looking into the surrounding code, to me it looks the correct 
> > action here is to remove the comment and return -1 instead of 0. It makes
> > this just an internal sanity check that grp_name is provided by the 
> > caller.
> > 
> 
> hmmm ... this should not be an error because the caller is not required
> to provide grp_name. Not providing grp_name has a specific meaning
> of this operating on the CON_MON group and a failure would break flows
> operating on the CON_MON group.

write_bm_pid_to_resctrl() checks for non-NULL mongrp before it calls into 
create_grp() so with current code, I don't think it changes anything. And 
param->ctrlgrp is always non-NULL too so I don't think the return ever 
triggers with the current codebase.

However, I was more talking from API point of view. It feels more natural 
for "create group" function to return error if the caller is inconsistent
with itself by asking to create a group but doesn't want to create a 
group.

-- 
 i.