Re: [PATCH v6 04/13] coresight: etm4x: exclude ss_status from drvdata->config

[Yeoreum Yun <yeoreum.yun@arm.com>]

Hi Leo,

> On Wed, Apr 22, 2026 at 02:21:54PM +0100, Yeoreum Yun wrote:
> 
> [...]
> 
> > @@ -573,11 +573,9 @@ static int etm4_enable_hw(struct etmv4_drvdata *drvdata)
> >  		etm4x_relaxed_write32(csa, config->res_ctrl[i], TRCRSCTLRn(i));
> >  
> >  	for (i = 0; i < caps->nr_ss_cmp; i++) {
> > -		/* always clear status bit on restart if using single-shot */
> > -		if (config->ss_ctrl[i] || config->ss_pe_cmp[i])
> > -			config->ss_status[i] &= ~TRCSSCSRn_STATUS;
> >  		etm4x_relaxed_write32(csa, config->ss_ctrl[i], TRCSSCCRn(i));
> > -		etm4x_relaxed_write32(csa, config->ss_status[i], TRCSSCSRn(i));
> > +		/* always clear status and pending bits on restart if using single-shot */
> > +		etm4x_relaxed_write32(csa, 0x0, TRCSSCSRn(i));
> 
> After confirmed with hardware team, we should preserve status bits
> (including STATUS and PENDING bits) during a session. So here we should
> set drvdata->ss_status to TRCSSCSRn.
> 
> > @@ -1503,8 +1501,9 @@ static void etm4_init_arch_data(void *info)
> >  	 */
> >  	caps->nr_ss_cmp = FIELD_GET(TRCIDR4_NUMSSCC_MASK, etmidr4);
> >  	for (i = 0; i < caps->nr_ss_cmp; i++) {
> > -		drvdata->config.ss_status[i] =
> > -			etm4x_relaxed_read32(csa, TRCSSCSRn(i));
> > +		drvdata->ss_status[i] = etm4x_relaxed_read32(csa, TRCSSCSRn(i));
> > +		drvdata->ss_status[i] &= (TRCSSCSRn_PC | TRCSSCSRn_DV |
> > +					  TRCSSCSRn_DA | TRCSSCSRn_INST);
> 
> It is fine for read these capacity bits when probe, but we need to clear
> status when a session is starting to avoid the stale value left from
> previous session:
> 
>   drvdata->ss_status[idx] &= ~(TRCSSCSRn_STATUS | TRCSSCSRn_PENDING);
> 
> We can do this in etm4_parse_event_config() for perf mode, and might
> create a new function (say etm4_parse_sysfs_config()) for preparing
> config for sysfs mode?
> 

As we discussed in offline, those bits should be cleared at the begining
of the session. so clearing drvdata->ss_status at start of session
in each mode is fine for me and for future integration for cpu suspend/resume.

But, I want to clarify that the perf is one of exceptional case
since the "etm4_parse_event_config()" is called at the "resume" of session
for per-thread mode event.

TBH, We don't have some specific usage how STATUS or PENDING bit could
be used with perf session and until now those bits are always cleared 
at the time of "sched-in" though we need to keep those bits theoretically.

Anyway as we discussed, since now there have been no issue
relavant for those bits, let the clear drvdata->ss_status at the
etm4_parse_event_config() or when setting a active config for start/resume
in this patchset and let me fix this with another patchset.

-- 
Sincerely,
Yeoreum Yun