On Fri, May 19, 2017 at 10:05:35AM +0300, Amir Goldstein wrote: > > Applied where? github is 2 weeks behind. kernel.org is 2 weeks behind. Pushed out now. There can be a day or two lag between when I accept patches and when they get pushed out. That's normally because I do a test verification in a batch process, and if there are substantive changes, I do like to at least do a smoke test before pushing stuff out. In this case, where it's just Documentation and kernel-configs changes, I do push something out first. > BTW, I am curious to know why the kernel-configs in this repository have > no obvious debug config options. > > Aren't you guys testing regularly with lockdep etc? does everyone carry > their own home brewed debug configs like I do [1]? We do actually have LOCKDEP enabled in the x86_64 4.9 config. The current set of configs were always maintained informally, and were originally intended to help people who were getting started to be able to build a kernel that would actually work using KVM or GCE. So they were made by copying my current development configs to kernel-config directory. That's why there is a certain amount of inconsistency over which debugging options were turned on. In particular, I never bothered to turn on LOCKDEP debugging for the stable kernel verification effort that I did a year or so ago. (I've been too busy to run the full set of regression tests on 3.18.x, et. al; I suspect if I did we would discover some additional number of bug fixes that need to be hand cherry picked back to ancient kernels. This also reflects that I'm no longer directly involved with trying to build Android kernels; which is why I started running the regression tests on LTS kernel releases, out of sheer self-defense, and so I could point out to various folks --- well, if your BSP kernel had _bothered_ to keep up with the stable kernels, you'd find the bug fix in 3.18.y....) I probably should have turned on LOCKDEP debugging, but at the time, finding bugs in the 3.10 kernel was easy enough without turning on LOCKDEP. :-) > What do you say about including -debug variants to the maintained > kernel-configs in this repo? or turn on the obvious debugging features > by default? > > [1] https://github.com/amir73il/xfstests-bld/commit/af76cac689ea726cb8f5b222010b7818a689809b No objections to me, in general, so long as it doesn't trash performance _too_ much. I will note that test diversity can be a good thing, since it could be that certain timing related bugs can be suppressed if various debugging options are turned on.... - Ted -- To unsubscribe from this list: send the line "unsubscribe fstests" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
