Re: [LEDE-DEV] Proposal to sign all commits
On 16-05-06 08:28 PM, Kus wrote: > Daniel, I like what you said. I hinted something like that in the original > message. Er, sorry which part - I think you mean about fast-forward only and not the ideal world where everything is always tested no matter who it's from? Regards, Daniel > > I don't like the idea of making changes to history after it is published. > Personally, I don't care about commit pollution but if the team thinks it is > important, then we should squash commits before we merge with master. History should never be rewritten in a *public* (meaning one that is supposed to be pulled from rather than a feature or staging branch that is intended for testing and rebasing and so on) branch. Ever. IMNSHO. (Unless it's something like a personal tree on github that hasn't been forked and you have no reason believe someone else has even noticed it, yet, and you have a good reason). In other branches only history not already in public branches should be rewritten else you've got an ugly problem. > In an ideal world, we'd make all commits on master and we'd have 100% > confidence that each commit is guaranteed to cause no regression. If wishes > were fishes... Heh, if that were the case we'd be the robots that took over the world because we were better than our human creators > Maybe require all commits in master be signed and encourage but not require > signing for others? Would that be acceptable? > Make sense to me. Regards, Daniel ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] RFC: Throughput testing results.
On Fri, May 6, 2016 at 4:00 PM, Ben Greearwrote: > On 05/06/2016 12:05 PM, David Lang wrote: >> >> On Fri, 6 May 2016, Ben Greear wrote: >> >>> On 05/06/2016 10:20 AM, Toke Høiland-Jørgensen wrote: Ben Greear writes: > I am interested in feedback on the testing output. My goal is to add a > few more different hardware configurations and then do nightly (or at > least weekly) tests. So that is showing up to 10s of *seconds* of latency, right? (I'm not sure I'm reading the units right). >>> >>> >>> Yes, 10 seconds of latency. My traffic generator is using pfifo-fast, >>> RENO, and default socket sizes, so it can be at least part of the >>> problem. >> >> >> That's so much latency that you may as well be down. >> >> Please look at the make-wifi-fast mailing list and the tests that are >> being done there. they show latency spikes as well as throughput, and show >> how it is very >> possible to get low latency without affecting throughput (in some cases, >> throughput actually increases) > > > I did some tcp/udp mix, tcp only, udp only download tests, with fq_codel and > fifo-fast on the traffic > generator. AP was un-changed, seems LEDE uses fq_codel by default. > Generator is using 4.4.8+ kernel, and it looks > like fq_codel works nicely! > > http://www.candelatech.com/examples/ventana/ > > We'll run some tests on some of our higher-performing APs and see if > fq_codel works well there > too when we get a chance At the qdisc layer fq_codel is presently at, it is still difficult to remove baseline latency at lower speeds nor serve multiple stations well. we can do much better with the patches that michal is working on http://blog.cerowrt.org/flent/qca-10.2-fqmac35-codel-5-wifi/slow_fast_simultaneously.svg The blog series analyzing the impact of these patches is at: http://blog.cerowrt.org/tags/wifi/ starting, basically, here: http://blog.cerowrt.org/post/fq_codel_on_ath10k/ It would be good to also get an ath9k and mt72 version working > Thanks, > Ben > > -- > Ben Greear > Candela Technologies Inc http://www.candelatech.com > -- Dave Täht Let's go make home routers and wifi faster! With better software! http://blog.cerowrt.org ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] RFC: Throughput testing results.
On Fri, 6 May 2016, Ben Greear wrote: On 05/06/2016 12:05 PM, David Lang wrote: On Fri, 6 May 2016, Ben Greear wrote: On 05/06/2016 10:20 AM, Toke Høiland-Jørgensen wrote: Ben Greearwrites: I am interested in feedback on the testing output. My goal is to add a few more different hardware configurations and then do nightly (or at least weekly) tests. So that is showing up to 10s of *seconds* of latency, right? (I'm not sure I'm reading the units right). Yes, 10 seconds of latency. My traffic generator is using pfifo-fast, RENO, and default socket sizes, so it can be at least part of the problem. That's so much latency that you may as well be down. Please look at the make-wifi-fast mailing list and the tests that are being done there. they show latency spikes as well as throughput, and show how it is very possible to get low latency without affecting throughput (in some cases, throughput actually increases) I understand that. My test case is fairly abusive, and my generator is not optimized for low-latency at this time. In many cases, throughput does go down though, so I have been slow to try the buffer bloat stuff. I can run some tests with codel enabled sometime soon. I can also run my capacity test with UDP only. That might be better for pure throughput testing. My hope is to be able to show regressions/improvements over time as LEDE changes... This is a good idea, but it is going to be very specific to the exact setup you use for the testing. Use different hardware, or add/remove nodes from the test (or have someone nearby create additional noise) and you can end up with very different results. I think it would be a bad idea to setup something that encourages chasing benchmarks. I agree it's a good idea to watch out for regressions. The hard thing is doing the latter without the former :-) David Lang___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] RFC: Throughput testing results.
On Fri, 6 May 2016, Ben Greear wrote: On 05/06/2016 10:20 AM, Toke Høiland-Jørgensen wrote: Ben Greearwrites: I am interested in feedback on the testing output. My goal is to add a few more different hardware configurations and then do nightly (or at least weekly) tests. So that is showing up to 10s of *seconds* of latency, right? (I'm not sure I'm reading the units right). Yes, 10 seconds of latency. My traffic generator is using pfifo-fast, RENO, and default socket sizes, so it can be at least part of the problem. That's so much latency that you may as well be down. Please look at the make-wifi-fast mailing list and the tests that are being done there. they show latency spikes as well as throughput, and show how it is very possible to get low latency without affecting throughput (in some cases, throughput actually increases) David Lang___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [RFC] A change to the way packages are built
On 16-05-05 04:16 AM, Michal Hrusecky wrote: > David Lang - 18:20 4.05.16 wrote: >> Debian has ... > > Just for the sake of discussion and inspiration, how openSUSE does it's > rolling > release. We have OBS, which is server software, connected to multiple > builders. [snip] Thank you David and Michal for the info. I'll try to make some time to look into both of these options and see if they can be used with and openwrt SDK to do what we're looking for (well at the least to get started on). I also have to look a Jo's buildbot stuff to see how things are done currently for LEDE. Regards, Daniel ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] Proposal to sign all commits
-BEGIN PGP SIGNED MESSAGE- Hash: SHA512 > Regarding signing commits with GPG key, it would be nice to recommend it but > making it a requirement sounds like a barrier. I'd argue such a barrier is OK if we want to quickly increase the size of the team of people with commit access. I think we're underestimating our contributors here. I agree that we shouldn't have unnecessary barriers (such as copyright assignment to give a specific example). I am getting mixed signals here though. Some people say requiring signing causes friction and limits participation. Others say that there will only be a couple of people who will ever have commit access so signing is unnecessary. I don't want to take too much time here because signing commits is a lower priority compared to doing the actual work of writing code/documentation (including a wiki), increasing/maintaining test coverage, and setting up automatic signed builds and so on (being discussed in separate threads). I don't think there's a definite right or wrong answer here as long as we understand and accept the trade offs. Sincerely, -BEGIN PGP SIGNATURE- Version: APG v1.1.1 iQJRBAEBCgA7BQJXLL6NNBxLdXNoYWwgSGFkYSAoZGV2ZWxvcGVyKSA8a3VzaGFs ZGV2ZWxvcGVyQGdtYWlsLmNvbT4ACgkQJsInd2b1xmPv9w/+Km0COpDHFHWjahVX XCGZdokz4BZn41ZF54R4z7iyexzZ9uviLJfQyftHODHYCvdl/P+zA3WYX2nyEQ5j zDIkXuGKmrG68zt55Y2layVgOrqJ3BswwdkFhG7mFEyvTJQDWYp50F6a9JjURZmB x1YCUO7fQidrmjOYdE9omEeJCBukujGtBFG1i2YxGPHA8hWANxB+hZD5AZHouNto i5YG7ssjJXusdoCtReIxUsimUwQ6s5IqSiOSZPwlGGl3lTj4rVcQtUNZzTlwBRsL 3VEAAlXNd6Kl0oKaet9wVJNwiF3nrDiLAgwTjS2T5ZIe5l4+TwcSAsN3xJUAe1tx 7ysWFEbgYNLxXuI8cvEXr9g9n7BW3QnbgQzpgadjQisGeIOzwsCirpGKrSBJDXVP RDClZQe9FhJ4edxgWig4htvH4eHsyyzic0RDaG+70aSNlWS4gVniAZ+dvn4cxnlF 22v7Ryl/Sb3dmhub2bQVVP4TZyYityNNfyW74cODj4mx2cYYwEhVEIAbvKz+ZE7r D6T2svtOSJpaPBGKL4JGhXxdwo6UZJucA13h3nrxYH+nHlm6v0xHWkV955LyP976 SYS7Nw6Opw0L66L5jAJjQ3z6+YAabd00AmxWMnL6pMJk3k8sY8sH45CLghCvQNzr xeFklDOsle8MwWAuuBb9CMB1OLI= =bVJi -END PGP SIGNATURE- ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] Why technical elitism is contrary to stated goal of community
On Fri, 2016-05-06 at 11:48 -0400, Daniel Curran-Dickinson wrote: > > As I said, the issue is not with 'only good bug reports', it's with 'we > > can ensure only good bug reports by making it a pain to report bugs at all'. > > I'd also like to point out that if the approach is being taken on the > assumption that on technically advanced people write good reports, then > I'd like to point out that > > a) There are plenty of technically advanced people who really suck at > writing good bug reports > b) There are a goodly number of noobs who given the right > prompts/prompting can give better bug reports than even many technically > advanced users who aren't in a). > c) There are going to be bad bug reports no matter how badly you try to > avoid it > d) Making it hard to report bugs means that bugs don't get reported, > even in cases where all you needed to fix the bug was to know it existed. All very true. It's also worth noting that if you are "guided" too hard to include information which you don't have or can't accurately provide, or which is irrelevant in your specific case, then you often end up making stuff up and that *detracts* from the quality of the report. I have a lot of expense reports where the mandatory "Business purpose?" field has the simple answer "yes" :) -- dwmw2 smime.p7s Description: S/MIME cryptographic signature ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] Why technical elitism is contrary to stated goal of community
Hi all, I have noticed that some of the policies of this project are already veering towards a brand of technical elitism that I feel is completely contrary to the stated goal of having a stronger community. A strong community welcomes 'outsiders' and noobs and helps them find their place in the community and *helps* them learn how to to participate in a friendly, open manner. Technical elitism tends give a 'you're not good enough / you don't think like us, go away' feel to a community. The 'we won't event look at your bug report if it doesn't meet the special format we came up with that no one else uses' policy is I fear an example of this. I realize the reason for it is to try and get better quality bug reports, and less ones that aren't useful, but I don't think that is necessarily a good way to about it. If LEDE is *serious* about community some of it's core members need to rethink how they approach contributions and the project. Perhaps time is the issue, and perhaps the issue is attitude, or perhaps it's a bit of both. Either way, if nothing genuinely changes from what was the case in OpenWrt, this project and OpenWrt will both die. Regards, Daniel ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] Proposal to sign all commits
Hi, > I am concerned that git signing gives little, if any value, while making > it harder to contribute (and making it easier to contribute is one of > the *stated* goals of LEDE) and is another example of a tendency toward > a particular brand of technical elitism that will kill this project if > not nipped in the bud. I tend to agree here - people specifically ask about being able to contribute via Github because it allegedly makes contributions easier. My experience has shown that a lot of contributors already struggle with the concept of sign-off lines. Require them to PGP sign stuff would pretty much kill any effort in this direction right away. ~ Jo ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [OpenWrt-Devel] Introducing the LEDE project
Might I submit that my impression is that Kaloz (at least) holds infrastructure hostage to maintain control, and that the fundamental problem here is that OpenWrt is *not* democratic and ignores what people who were ones visibly working on openwrt want and overrides their wishes because he/they has/have the keys. That doesn't work. Perhaps I'm wrong, but that is certainly my impression. There isn't management for openwrt except by democratic/meritocratic principles, and no one should try to force it to be otherwise. Regards, Daniel ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [OpenWrt-Devel] Introducing the LEDE project
On 6 May 2016 at 03:53, Luka Perkovwrote: >>On 2016-05-05 20:22, mbm wrote: >>> On 5/5/2016 7:40 AM, Felix Fietkau wrote: Many of the changes that we previously tried to introduce were often squashed by internal disagreements. Resulting discussions often turned toxic quickly and led to nothing being done to address the issues. Setting up the LEDE project was our way of creating a testbed for changes that we believe are important for the survival of the project. >>> >>> Change is not easy. Discussions need to happen. The problem is simply >>> kicking out people you didn't agree with by starting a new organization >>> in secret; you've created the public perception that we're somehow >>> against you when really we all want the same things. >> >> Years of internal discussion led nowhere. Maybe it helps now that we're >> making the whole issue public. > > For the sake of transparency, we've had public discussions, about a number of > things, for example switching to Git: > > - https://lists.openwrt.org/pipermail/openwrt-devel/2015-October/036390.html > - https://lists.openwrt.org/pipermail/openwrt-devel/2015-October/036480.html > - https://lists.openwrt.org/pipermail/openwrt-devel/2015-October/036486.html > - https://lists.openwrt.org/pipermail/openwrt-devel/2015-October/036500.html > > And based on these inputs from you the switch was not made even though several > OpenWrt developers wanted to switch. > > Also, server outages can happen to anybody: > - https://lists.openwrt.org/pipermail/openwrt-devel/2016-January/038547.html > > However, we do not want to point fingers. What we do want is to make a great > community around OpenWrt and to drive innovation - just like it has been done > for more then a decade now. > > There has been a long history - mostly good, sometimes bad - since the project > started from a garage project, to now having a project used by an awesome > amount of users. This can be seen from the constructive discussions in this > list on a daily basis, and in this very thread. Also, the project is used as > the main SDK by many silicon vendors internally, and by vast number of > companies on the embedded market. > > We are open for a discussion and would like to keep the OpenWrt's and it's > community interests in the first place. Splitting the project does not make > sense. Do you agree? > We appreciate your effort to have an open discussion about this, however the sudden deletion of our widely published openwrt.org email addresses somewhat undermines this. We will not respond in kind and we will continue to maintain the critical parts of OpenWrt infrastructure that we control. >>> >>> Let's be clear on this subject; no commit access was revoked, you still >>> have full read and write access to the entire OpenWrt tree. >>> >>> Email forwarding was temporarily disabled following the LEDE announcement >>> - LEDE's own rules prohibit project based email addresses >> No, they don't. They state that the LEDE project won't provide project >> email addresses. Interpreting that as meaning that we shouldn't be able >> to access our openwrt.org addresses is more than a bit of a stretch. > > In any case, due to the events that happened and the fact that the OpenWrt > name > is being used in a manner opposite of the projects best interest, we felt that > these actions were needed in order to avoid the further damages to the > project. > >>> - It's unclear if LEDE still represents OpenWrt >> So? Asking us to not send any further emails about LEDE from our >> openwrt.org addresses should have been enough. > > Actually, this was discussed on #lede-adm. IRC history is hard to follow, I'd better assume that something not written here never happened. Regards, Roman ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev