From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Message-ID: <56A68F1B.80605@oracle.com> Date: Mon, 25 Jan 2016 13:09:47 -0800 From: Wim Coekaerts MIME-Version: 1.0 To: Guenter Roeck , wim@iguana.be CC: linux-watchdog@vger.kernel.org, sparclinux@vger.kernel.org Subject: Re: [PATCH v3] Add sun4v_wdt watchdog driver References: <1453754288-30997-1-git-send-email-wim.coekaerts@oracle.com> <56A68D5F.40406@roeck-us.net> In-Reply-To: <56A68D5F.40406@roeck-us.net> Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit List-ID: On 01/25/2016 01:02 PM, Guenter Roeck wrote: > On 01/25/2016 12:38 PM, Wim Coekaerts wrote: >> This driver adds sparc hypervisor watchdog support. Timeout is set in >> milliseconds since that is the granularity supported and it honors >> the settings of both the watchdog-resolution and watchdog-max-timeout >> MD properties. >> >> Default timeout is set at 60 seconds or 60000ms. The range is between >> 1 second and 180 seconds. The default resolution is 1000ms. >> >> Signed-off-by: Wim Coekaerts >> --- >> Changes in v3: >> - Modify sun4v_mach_set_watchdog to allow for NULL and remove >> time_remaining >> - Cleanup includes >> - Consolidate _start and _ping into _ping since they were the same >> - Fix checkpatch warnings >> - Remove pr_info()s >> - Clean up return codes to be standard kernel return values >> - Consolidate _init and _probe into _init >> Don't use platform_device anymore because this is really >> just a driver that depends on a hv call. This now looks more >> like softdog. >> - Cleanly check for sun4v architecture and remove extra tests >> - Convert to ms timer support and honor resolution >> since most drivers use seconds I added _ms to make it clear > > You can not redefine the watchdog ABI to mean milli-seconds instead of > seconds. ugh ok. Then it makes honoring resolutions a bit useless (comment on previous patch about having a fixed 1s resolution) If I round timeout back to seconds, does the rest at least look ok now? From mboxrd@z Thu Jan 1 00:00:00 1970 From: Wim Coekaerts Date: Mon, 25 Jan 2016 21:09:47 +0000 Subject: Re: [PATCH v3] Add sun4v_wdt watchdog driver Message-Id: <56A68F1B.80605@oracle.com> List-Id: References: <1453754288-30997-1-git-send-email-wim.coekaerts@oracle.com> <56A68D5F.40406@roeck-us.net> In-Reply-To: <56A68D5F.40406@roeck-us.net> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Guenter Roeck , wim@iguana.be Cc: linux-watchdog@vger.kernel.org, sparclinux@vger.kernel.org On 01/25/2016 01:02 PM, Guenter Roeck wrote: > On 01/25/2016 12:38 PM, Wim Coekaerts wrote: >> This driver adds sparc hypervisor watchdog support. Timeout is set in >> milliseconds since that is the granularity supported and it honors >> the settings of both the watchdog-resolution and watchdog-max-timeout >> MD properties. >> >> Default timeout is set at 60 seconds or 60000ms. The range is between >> 1 second and 180 seconds. The default resolution is 1000ms. >> >> Signed-off-by: Wim Coekaerts >> --- >> Changes in v3: >> - Modify sun4v_mach_set_watchdog to allow for NULL and remove >> time_remaining >> - Cleanup includes >> - Consolidate _start and _ping into _ping since they were the same >> - Fix checkpatch warnings >> - Remove pr_info()s >> - Clean up return codes to be standard kernel return values >> - Consolidate _init and _probe into _init >> Don't use platform_device anymore because this is really >> just a driver that depends on a hv call. This now looks more >> like softdog. >> - Cleanly check for sun4v architecture and remove extra tests >> - Convert to ms timer support and honor resolution >> since most drivers use seconds I added _ms to make it clear > > You can not redefine the watchdog ABI to mean milli-seconds instead of > seconds. ugh ok. Then it makes honoring resolutions a bit useless (comment on previous patch about having a fixed 1s resolution) If I round timeout back to seconds, does the rest at least look ok now?