On 00:21-20210526, Denys Dmytriyenko wrote: > On Tue, May 25, 2021 at 06:55:52AM -0500, Nishanth Menon wrote: > > On 21:51-20210524, Denys Dmytriyenko wrote: > > > On Mon, May 24, 2021 at 05:40:36PM -0500, Yogesh Siraswar via > > > lists.yoctoproject.org wrote: > > > > Added KERNEL_DTB_OVERLAY_SUPPORT to individual MACHINE conf to > > > > compile dtbs with symbols. This fixes the issue where the dtbs > > > > are build without symbols causing issue with overlays. > > > > > > > > Enabling overlay support increases the size of the dtb, hence > > > > it is done at a platform level rather than globally to allow custom > > > > boards disable it from local.conf to reduce size. > > > > > > > > ^^ > > > > > > Signed-off-by: Yogesh Siraswar <[email protected]> > > > > --- > > > > conf/machine/am335x-evm.conf | 2 ++ > > > > conf/machine/am335x-hs-evm.conf | 2 ++ > > > > conf/machine/am437x-evm.conf | 2 ++ > > > > conf/machine/am437x-hs-evm.conf | 2 ++ > > > > conf/machine/am57xx-evm.conf | 2 ++ > > > > conf/machine/am57xx-hs-evm.conf | 3 +++ > > > > conf/machine/am65xx-evm.conf | 2 ++ > > > > conf/machine/am65xx-hs-evm.conf | 2 ++ > > > > conf/machine/dra7xx-evm.conf | 2 ++ > > > > conf/machine/dra7xx-hs-evm.conf | 3 +++ > > > > conf/machine/j7-evm.conf | 2 ++ > > > > conf/machine/j7-hs-evm.conf | 2 ++ > > > > conf/machine/j7200-evm.conf | 2 ++ > > > > conf/machine/k2g-evm.conf | 2 ++ > > > > conf/machine/k2g-hs-evm.conf | 2 ++ > > > > > > First question: > > > If you are weakly assigning KERNEL_DTB_OVERLAY_SUPPORT in most of the > > > machine > > > configs, should it instead be done in SoC include files, like ti33x.inc, > > > k3.inc or even ti-soc.inc? > > > > You would want the default behavior as default dtb, not the bloated > > symbol loaded overlay capable dtb, no? > > Well, but from the list above the only machines missing are AM64 and legacy
^^^ Arrgh, good catch - we are indeed missing am64.. > K2 HK/L/E. And since the assignment is weak (?=), it can still be changed > from local.conf, as the commit message says. I think we need to drop k2 platforms - they are just bitrotting.. > Anyway, if you don't think making it a bit more generic for ease of > maintenance is not worth it, then I'm fine with current approach. Hmmm.. Your point is to get the weak assignment to allows the critical mach confs to just override it with = yeah, I am not exactly a fan of going overboard with entries in machine file. I guess I second your thought about including it in ti-soc.inc instead of machine specific includes to prevent us from messing up and allowing a little less optimal, but feature rich default. btw, where do we document all these variables that need tweaking? -- Regards, Nishanth Menon Key (0xDDB5849D1736249D) / Fingerprint: F8A2 8693 54EB 8232 17A3 1A34 DDB5 849D 1736 249D
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#13797): https://lists.yoctoproject.org/g/meta-ti/message/13797 Mute This Topic: https://lists.yoctoproject.org/mt/83064187/21656 Group Owner: [email protected] Unsubscribe: https://lists.yoctoproject.org/g/meta-ti/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
