From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Hurley Subject: Re: [PATCH v2 2/3] serial: 8250_early: prepare for dynamic BASE_BAUD Date: Mon, 05 Jan 2015 09:05:41 -0500 Message-ID: <54AA9A35.2000806@hurleysoftware.com> References: <1420457437-17636-1-git-send-email-vgupta@synopsys.com> <1420457437-17636-3-git-send-email-vgupta@synopsys.com> <54AA8AFC.9020706@hurleysoftware.com> <54AA92C7.8070209@hurleysoftware.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: Vineet Gupta , "gregkh@linuxfoundation.org" Cc: "linux-serial@vger.kernel.org" , "linux-kernel@vger.kernel.org" , "robh@kernel.org" , "jslaby@suse.cz" , Arnd Bergmann List-Id: linux-serial@vger.kernel.org On 01/05/2015 08:52 AM, Vineet Gupta wrote: > On Monday 05 January 2015 07:04 PM, Peter Hurley wrote: >>> Also what do u think of the early IER optimization ? >> Seems ok except the commit changelog is wrong: IER is disabled twice per _line_. >> How did you profile it? It's ok if the answer is "seemed faster"; I'm just >> curious. > > Looking at the code yes this would be per line - but I vaguely remember seeing it > for every char - this was when debugging a systemC model for 8250. Its been a while... > I'd added prints in the model whenever IER would wiggle and it was just being > written with 0's again and again when in the early polling based printing mode. Maybe that was while running kgdb through serial8250_put_poll_char(), which is per-char. Regards, Peter Hurley