[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <DB4PR02MB272A2301DF424A030948E78D6820@DB4PR02MB272.eurprd02.prod.outlook.com>
Date: Thu, 23 Jul 2015 10:38:11 +0000
From: Noam Camus <noamc@...hip.com>
To: Peter Hurley <peter@...leysoftware.com>
CC: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-serial@...r.kernel.org" <linux-serial@...r.kernel.org>,
"Alexey.Brodkin@...opsys.com" <Alexey.Brodkin@...opsys.com>,
"vgupta@...opsys.com" <vgupta@...opsys.com>,
"gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
"jslaby@...e.com" <jslaby@...e.com>
Subject: RE: [PATCH 3/4] serial: 8250_dw: Add UPF_FIXED_TYPE to flags
From: Peter Hurley [mailto:peter@...leysoftware.com]
Sent: Wednesday, July 22, 2015 3:39 PM
>> + /* Writing to LCR may cause BUSY interrupt before we
>> + * register the IRQ line.
>> + * Currently autoconf() uses several writes to LCR.
>> + * In order to avoid calling to autoconf() always add
>> + * following flag.
>> + */
>> + p->flags |= UPF_FIXED_TYPE;
>
> Why only for devicetree DW8250's? Don't all DW8250's have this LCR "feature"?
>
Yes, all DW8250 got this "feature" i.e. busy functionality.
> And what port type does this id DW8250's as, PORT_8250? Except with fifos, autoflow control, dma, etc.?
The only thing that DW8250 is not fully 16550 compatibility is this busy functionality feature.
This feature means that a serial transfer is in progress.
It is indicated by special identification in IIR, and raised when LCR is written while device is busy.
> If the port type is being fixed, then please fix it to a new port type.
As an extension to 16550 type is not fixed.
Seem that I need to rethink on this.
I will remove this commit from v2 patch set till I will be sure what is the right thing to do.
Noam
Powered by blists - more mailing lists