Re: [PATCH v4 1/2] serial: qcom-geni: trace: Add tracepoint support for Qualcomm GENI serial

On Mon, Jun 15, 2026 at 06:09:17PM +0530, Praveen Talari wrote:

Hi

On 15-06-2026 16:12, Greg Kroah-Hartman wrote:

quoted

On Mon, Jun 15, 2026 at 03:26:25PM +0530, Praveen Talari wrote:

quoted

HI Steven,

On 29-05-2026 19:44, Steven Rostedt wrote:

quoted

On Tue, 26 May 2026 23:07:39 +0530
Praveen Talari [off-list ref] wrote:

quoted

+DECLARE_EVENT_CLASS(geni_serial_data,
+		    TP_PROTO(struct device *dev, const u8 *buf, unsigned int len),
+		    TP_ARGS(dev, buf, len),
+
+		    TP_STRUCT__entry(__string(name, dev_name(dev))
+				     __field(unsigned int, len)
+				     __dynamic_array(u8, data, len)
+		    ),
+
+		    TP_fast_assign(__assign_str(name);
+				   __entry->len = len;
+				   memcpy(__get_dynamic_array(data), buf, len);
+		    ),
+
+		    TP_printk("%s: len=%u data=%s",
+			      __get_str(name), __entry->len,
+			      __print_hex(__get_dynamic_array(data), __entry->len))
+);

No need to save the length of the dynamic array in __entry->len because
it's already saved in the metadata of the dynamic array that is stored
on the buffer. Instead you can have:

DECLARE_EVENT_CLASS(geni_serial_data,
		    TP_PROTO(struct device *dev, const u8 *buf, unsigned int len),
		    TP_ARGS(dev, buf, len),

		    TP_STRUCT__entry(__string(name, dev_name(dev))
				     __dynamic_array(u8, data, len)
		    ),

		    TP_fast_assign(__assign_str(name);
				   memcpy(__get_dynamic_array(data), buf, len);
		    ),

		    TP_printk("%s: len=%u data=%s",
			      __get_str(name), __entry->len,
			      __print_hex(__get_dynamic_array(data),
					__get_dynamic_array_len(data)))
);

That will save you 4 bytes per event on the ring buffer. And a few
cycles not having to store the redundant information.

This patch has already been accepted and is available in linux-next.

Great, can you send a fixup for it?  Or want me to revert this instead?

can i add fix patch in same series by removing original patch(0/1)?

We can never rewrite a public git tree.  So either revert and redo it as
a new patch, or send a fix for this one, your choice.

thanks,

greg k-h