Re: [PATCH v2 5/7] perf cs-etm: Add helper cs_etm__get_pid_fmt()

On Thu, Feb 04, 2021 at 10:54:24AM +0000, Suzuki Kuruppassery Poulose wrote:

[...]

quoted

quoted

quoted

+int cs_etm__get_pid_fmt(u8 trace_chan_id, u64 *pid_fmt)
+{
+	struct int_node *inode;
+	u64 *metadata, val;
+
+	inode = intlist__find(traceid_list, trace_chan_id);
+	if (!inode)
+		return -EINVAL;
+
+	metadata = inode->priv;
+
+	if (metadata[CS_ETM_MAGIC] == __perf_cs_etmv3_magic) {
+		val = metadata[CS_ETM_ETMCR];
+		/* CONTEXTIDR is traced */
+		if (val & BIT(ETM_OPT_CTXTID))
+			*pid_fmt = BIT(ETM_OPT_CTXTID);
+	} else {
+		val = metadata[CS_ETMV4_TRCCONFIGR];
+
+		*pid_fmt = 0;
+
+		/* CONTEXTIDR_EL2 is traced */
+		if (val & (BIT(ETM4_CFG_BIT_VMID) | BIT(ETM4_CFG_BIT_VMID_OPT)))
+			*pid_fmt = BIT(ETM_OPT_CTXTID2);
+
+		/* CONTEXTIDR_EL1 is traced */
+		if (val & BIT(ETM4_CFG_BIT_CTXTID))

I haven't looked at how this gets used. But, Shouldn't this be :

		else if (val & BIT(ETM4_CFG_BIT_CTXTID)) ?

Actually it's deliberately to set both bits ETM_OPT_CTXTID2 and
ETM_OPT_CTXTID if user has enable configs "contextid1" and
"contextid2".  So this is exactly the reversed flow in the
function cs_etmv4_get_config().

The point is, we don't care if the user selected both options. What we
care is, where can we find the PID. CONTEXTIDR_EL1 or CONTEXTIDR_EL2.
As such, get_pid_fmt simply should make that decision and pass it on.
So, if the CONTEXTIDR_EL2 is selected (which can only be done successfully
on an EL2 kernel), thats our pid.

So we should return the format for the PID here. i.e
 ETM_OPT_CTXTID2 OR ETM_OPT_CTXTID. But not both.

Okay, if so I will follow your suggestion.

Thanks,
Leo