From: Tom Rini Date: Fri, 5 Jun 2020 10:34:46 -0400 Subject: [RFC PATCH] watchdog: base rate-limiting on get_ticks() rather than get_timer() In-Reply-To: References: <20200605111657.28773-1-rasmus.villemoes@prevas.dk> <46f75f5a-8846-c696-f890-e27958ed3317@denx.de> <501eae48-6db9-bce7-eebd-a0eea482e299@denx.de> Message-ID: <20200605143446.GE32287@bill-the-cat> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On Fri, Jun 05, 2020 at 04:23:21PM +0200, Stefan Roese wrote: > On 05.06.20 16:18, Rasmus Villemoes wrote: > > On 05/06/2020 15.37, Stefan Roese wrote: > > > On 05.06.20 14:13, Stefan Roese wrote: > > > > On 05.06.20 14:11, Rasmus Villemoes wrote: > > > > > On 05/06/2020 13.48, Stefan Roese wrote: > > > > > > On 05.06.20 13:16, Rasmus Villemoes wrote: > > > > > > > This is what I had in mind. I also considered making it a config knob > > > > > > > (possibly just auto-selected based on $ARCH) whether to use > > > > > > > get_timer() or get_ticks(), but that becomes quite ugly. > > > > > > > > > > > > I hesitate a bit, moving with this generic code from get_timer() > > > > > > to get_ticks() for all boards. Did you test this patch on other > > > > > > platforms, like some ARM boards? > > > > > > > > > > > > Please note that I don't reject it - just asking. > > > > > > > > > > Yeah, I'm not really too happy about it myself, exactly because it > > > > > affects all arches/platforms. And no, I don't have other hardware handy > > > > > unfortunately. So it's very much an RFC where I hope someone with > > > > > knowledge of the various arches can say whether one can expect > > > > > get_ticks() to be at least as widely available as get_timer(). > > > > > > > > I'll try to test it on a non-powerpc platform soon'ish. > > > > > > I've tested it on an MIPS based platform (gardena-smart-gateway-mt7688) > > > and it works there without any issues AFAICT. > > > > Thanks. > > > > > Two more remarks: > > > > > > Could you please run a build test with this patch applied for all > > > boards (Travis, Azure...)? > > > > I may need a few more pointers than that. What am I supposed to do exactly? > > Not sure if there is some documentation on how to use the Travis, > Gitlab or Azure build. It mainly uses buildman [1] internally, so you > can use buildman locally if you don't want to push the build into > the cloud. So, here's a recently-I-learned. If you submit a pull request on GitHub to https://github.com/u-boot/u-boot that will trigger Travis and Azure runs: https://travis-ci.org/github/u-boot/u-boot/pull_requests and somewhere under: https://dev.azure.com/u-boot/u-boot/_build?definitionId=2&_a=summary Note that sometimes Travis will have connectivity failures and you have to be me to tell it to re-run the job. Azure only has failures when we get those to-be-debugged random failures of sandbox/qemu targets. -- Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 659 bytes Desc: not available URL: