On Sun, 2017-01-08 at 21:16 +0000, Sean Young wrote:

Hi Sean,

On Fri, Jan 06, 2017 at 03:31:25PM +0800, Sean Wang wrote:

quoted

On Thu, 2017-01-05 at 17:12 +0000, Sean Young wrote:

quoted

On Fri, Jan 06, 2017 at 12:06:24AM +0800, sean.wang-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org wrote:

quoted

+	/* Handle pulse and space until end of message */
+	for (i = 0 ; i < MTK_CHKDATA_SZ ; i++) {
+		val = mtk_r32(ir, MTK_CHKDATA_REG(i));
+		dev_dbg(ir->dev, "@reg%d=0x%08x\n", i, val);
+
+		for (j = 0 ; j < 4 ; j++) {
+			wid = (val & (0xff << j * 8)) >> j * 8;
+			rawir.pulse = !rawir.pulse;
+			rawir.duration = wid * (MTK_IR_SAMPLE + 1);
+			ir_raw_event_store_with_filter(ir->rc, &rawir);
+
+			if (MTK_IR_END(wid))
+				goto end_msg;
+		}
+	}

If I read this correctly, there is a maximum of 17 * 4 = 68 edges per
IR message. The rc6 mce key 0 (scancode 0x800f0400) is 69 edges, so that
won't work.

Uh, this is related to hardware limitation. Maximum number hardware
holds indeed is only 68 edges as you said :( 

For the case, I will try change the logic into that the whole message 
is dropped if no end of message is seen within 68 counts to avoid
wasting CPU for decoding. 

I'm not sure it is worthwhile dropping the IR in that case. The processing
is minimal and it might be possible that we have just enough IR to decode
a scancode even if the trailing end of message is missing. Note that
the call to ir_raw_event_set_idle() will generate an timeout IR event, so
there will always be an end of message marker.

1)
I agree with you :) The original logic I made already as you pointed out
is sent incomplete IR message to let ir-raw try to decode as possible.

2)
I had another question. I found multiple and same IR messages being
received when using SONY remote controller. Should driver needs to
report each message or only one of these to the upper layer ?

All I wanted to do was point out a limitation in case there is a
workaround; if there is not then we might as well make do with the IR
we do have.

I also will leave some words about limitation we had in the comments.

Thanks
Sean

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html