13389 2019-06-11 11:09:32 +0000 dtc -I fs (v1.4.7) crashes 2020-02-03 02:14:37 +0000 1 1 7 Build System, Metadata & Runtime OE-Core kernel 2.6.3 All Multiple CLOSED WONTFIX Medium normal 2.7.3 1 m.brock m.brock akuster randy.macleod ross.burton tom.zanussi --- Regression (Used to work) No (bug/feature does not impact docs) oldest_to_newest 84140 0 m.brock 2019-06-11 11:09:32 +0000 Hi, Though dtc is also part of the build system, the problem I encountered is with runtime use at the target, which is why I categorized this bug under Runtime. When you run dtc -I fs /proc/device-tree it fails with a segmentation fault. There already is a patch for this applied to dtc v1.5.0. Would you consider either applying this patch or upgrading to 1.50 or downgrading 1.4.6 in your version 2.6.x ? https://git.kernel.org/pub/scm/utils/dtc/dtc.git/commit/?id=9619c8619c37b9aea98100bcc15c51a5642e877e Kind regards, Maarten Brock 84141 1 ross.burton 2019-06-11 11:34:07 +0000 For master upgrading is preferred, for any stable branches picking the patch is preferred. As you've identified the fix, would you be willing to prepare a patch to oe-core? 84143 2 m.brock 2019-06-11 12:49:35 +0000 Hello Ross, Thanks for answering. I'm not familiar with sending in patches to openembedded/yocto. Since I pointed out the exact patch upstream, isn't that enough to just take that and apply it? Else I would need directions how to proceed. Kind regards, Maarten Brock 84149 3 ross.burton 2019-06-11 15:02:05 +0000 Well for master you'll ideally be doing the upgrade to 1.5. Otherwise yes, it should be a simple matter of applying the patch: you'll spend more time explaining what the fix is for than applying it in the first place. 84151 4 m.brock 2019-06-11 15:42:46 +0000 But newer versions including master already carry dtc v1.5.0. And the mentioned patch already has an explanation why it needs to be applied: 'This causes 'dtc -I fs' to crash'. It's actually more verbose. It's just that stable 2.6 contains a crashing application which doesn't exactly sound stable to me. 84152 5 ross.burton 2019-06-11 22:08:22 +0000 Ah sorry, I wasn't paying attention obviously. In that case just apply the fix to all the stable branches. 84180 6 ross.burton 2019-06-13 14:49:41 +0000 The standard flow is to always fix in master, and then back down the branches. This ensures that a fix isn't made to an old release and not made to newer ones. master has 1.5.0, which you say doesn't have this problem. warrior has 1.4.7. thud has 1.4.7. sumo has 1.4.5. Can you send backport patches for the fix to the three stable branches? 84207 7 m.brock 2019-06-14 08:45:33 +0000 dtc 1.4.5 is not broken and thus needs no fix. And since you insist that *I* apply the patch, I'll repeat my previous answer: I'm not familiar with sending in patches to openembedded/yocto. (...) I would need directions how to proceed. Maarten 84209 8 ross.burton 2019-06-14 11:37:49 +0000 I'm not insisting, I'm asking. You have the fix to hand and know how to replicate it. If you don't want to then we can reassign. https://www.openembedded.org/wiki/How_to_submit_a_patch_to_OpenEmbedded covers the process, but the summary is: take the upstream patch, apply it to the dtc recipe in oe-core. Add an Upstream-Status: Backport [9619c8619c37b9aea98100bcc15c51a5642e877e] and your Signed-off-by to the patch. Commit message should explain the problem and fix. Then git-send-email it to openembedded-core@lists.openembedded.org. 86262 9 akuster 2020-02-03 02:14:26 +0000 2.6 is under community support. no intention of actively working this issue. 86263 10 akuster 2020-02-03 02:14:37 +0000 closing