From: Wolfgang Denk Date: Fri, 19 Jul 2013 08:43:27 +0200 Subject: [U-Boot] [PATCH v6 1/2] console: usb: kbd: To improve TFTP booting performance In-Reply-To: <1374156931-1718-1-git-send-email-jilin@nvidia.com> References: <1374156931-1718-1-git-send-email-jilin@nvidia.com> Message-ID: <20130719064327.0CF07380ADF@gemini.denx.de> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Dear Jim Lin, In message <1374156931-1718-1-git-send-email-jilin@nvidia.com> you wrote: > TFTP booting is slow when a USB keyboard is installed and > stdin has usbkbd added. > This fix is to change Ctrl-C polling for USB keyboard to every second > when NET transfer is running. > > Signed-off-by: Jim Lin The sequence of your patches is wrong; you must reverse it: as is, with only patch 1/2 applied, you will get compile errors because net_busy_flag is undefined. This breaks bisectability. Please switch the order of your patches. > + /* > + * If net_busy_flag is 1, NET transfer is running, > + * then we check key pressed every second to improve > + * TFTP booting performance. > + */ > + if (net_busy_flag) { > + if (get_timer(kbd_testc_tms) < CONFIG_SYS_HZ) > + return 0; > + else > + kbd_testc_tms = get_timer(0); > + } When first entering this code, the variable kbd_testc_tms is (implicitly) initialized as zero; later, for example when running multiple network commands, the last used value (i. e. a random number) is used. So strictly speaking the comment above is incorrect, as you don't test for key presses "every second" - the first test may happen much earlier (even immediately). I think this should be explained in the comment to prevent incorrect expectations on the behaviour. Best regards, Wolfgang Denk -- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de Eureka! -- Archimedes