Hi Babu,
On 7/17/25 8:54 PM, Reinette Chatre wrote:
> Hi Babu,
>
> On 7/8/25 3:17 PM, Babu Moger wrote:
>> diff --git a/fs/resctrl/rdtgroup.c b/fs/resctrl/rdtgroup.c
>> index 15d10c346307..bb28ef7e4600 100644
>> --- a/fs/resctrl/rdtgroup.c
>> +++ b/fs/resctrl/rdtgroup.c
>> @@ -86,6 +86,8 @@ enum resctrl_event_id mba_mbps_default_event;
>>
>> static bool resctrl_debug;
>>
>> +extern struct mbm_transaction mbm_transactions[NUM_MBM_TRANSACTIONS];
>> +
>
> Please move this extern to fs/resctrl/internal.h.
>
>> void rdt_last_cmd_clear(void)
>> {
>> lockdep_assert_held(&rdtgroup_mutex);
>> @@ -1895,6 +1897,25 @@ static int resctrl_available_mbm_cntrs_show(struct kernfs_open_file *of,
>> return ret;
>> }
>>
>> +static int event_filter_show(struct kernfs_open_file *of, struct seq_file *seq, void *v)
>> +{
>> + struct mon_evt *mevt = rdt_kn_parent_priv(of->kn);
>> + bool sep = false;
>> + int i;
>> +
>> + for (i = 0; i < NUM_MBM_TRANSACTIONS; i++) {
>> + if (mevt->evt_cfg & mbm_transactions[i].val) {
>
> mevt->evt_cfg could possibly be changed concurrently. This should
> be protected with the rdtgroup_mutex.
>
>> + if (sep)
>> + seq_putc(seq, ',');
>> + seq_printf(seq, "%s", mbm_transactions[i].name);
>> + sep = true;
>> + }
>> + }
>> + seq_putc(seq, '\n');
>> +
>> + return 0;
>> +}
>> +
>> /* rdtgroup information files for one cache resource. */
>> static struct rftype res_common_files[] = {
>> {
>> @@ -2019,6 +2040,12 @@ static struct rftype res_common_files[] = {
>> .seq_show = mbm_local_bytes_config_show,
>> .write = mbm_local_bytes_config_write,
>> },
>> + {
>> + .name = "event_filter",
>> + .mode = 0444,
>> + .kf_ops = &rdtgroup_kf_single_ops,
>> + .seq_show = event_filter_show,
>> + },
>> {
>> .name = "mbm_assign_mode",
>> .mode = 0444,
I am looking at the SDCIAE series now and it really looks like rdtgroup.c is
becoming very large with code from various areas mixed in. In this series I
requested several functions to be moved to monitor.c but generally did not
highlight the main handlers added by this work. For example event_filter_show()
added by this patch. For this work, could you please take a look if the
handlers can also be moved to monitor.c? I think this will make things cleaner
since the same helpers and data structures no longer need to be shared by multiple
resctrl files. For example here, mbm_transactions[] can remain private
to monitor.c where it is most relevant. In retrospect I think this is what
we should have done for the BMEC work also.
Reinette
