**** BEGIN LOGGING AT Sat Mar 09 03:00:34 2019 Mar 09 14:38:45 Hmm, we need a new uninative release ASAP :/. Mar 09 15:32:49 RP: fixed the qemuarm thing, I think Mar 09 15:34:11 do s/VGA/VGA,edid=on/g on the qemuarm.conf Mar 09 15:36:34 my setup is a little dirty (kernel config changes too), but that has me working Mar 09 16:24:38 jonmason: ah, great. I can try that, thanks! Mar 09 16:24:50 khem, jonmason: https://bugzilla.yoctoproject.org/show_bug.cgi?id=13153 any ideas on that? Mar 09 16:24:51 Bug 13153: normal, Medium+, 2.7 M3, kamensky, NEW , [2.7 M2 RC1] Systemtap doesn't work on beaglebone Mar 09 16:25:06 Unfortunately that one shows up trying qemuarm as armv7 :/ Mar 09 16:27:45 is this an thumb/arm mode thing? Mar 09 16:28:06 i mean, the look like valid instructions Mar 09 16:28:23 jonmason: yes, looks a bit like it. Its as if there is a missing option to gcc or perhaps as Mar 09 16:29:42 that was my thought was well, but gcc didn't just change did it? Mar 09 16:32:48 I added myself to it. I'll dig further when I get back from SCaLE. Mar 09 21:42:26 I'm getting duplicate attempts to apply kernel patches on the current master branch when kernel-yocto.bbclass is used - once in do_kernel_metadata then again in do_patch. Looks related to a change by zeddii. Anyone else seen this? Mar 09 22:06:00 paulbarker: I haven;t seen that Mar 09 22:06:43 RP: Just about to send an email to the list. I have what I think is a fix, works for me at least, but not sure if it breaks something else Mar 09 22:09:38 paulbarker: ok, zeddi is away for a bit fwiw Mar 09 22:13:58 RP: Ok, have sent the email anyway. Having 'set +e' in do_kernel_metadata definitely makes this easy to slip under the radar and cause silent problems Mar 09 22:17:07 paulbarker: have to love silent problems :/ Mar 09 22:18:30 Looks like all patches are being picked up as if they have patchdir='kernel-meta', then the 'cd ${WORKDIR}/kernel-meta' command is failing as there's no kernel-meta directory in linux-raspberrypi and so the patch is just applied at the top of the source tree where it belongs Mar 09 22:18:45 Then do_patch tries to apply the same patch and fails as it's already applied Mar 09 22:28:14 paulbarker: it does sound like a bug Mar 09 22:30:59 I'll hold off on submitting a patch with my local fix for now. I'm not entirely sure what the intended behaviour for kernel-meta patches is so there's a good chance I've fixed my use case and broken that one Mar 09 22:31:14 jonmason: FWIW that edid change appears to make qemuarm64 work! Mar 09 22:41:17 jonmason: I've put some quick hacks into -next for testing with 5.x, this should mean we just have the qemuarm stap problem Mar 09 22:43:39 RP: it looks like a kernel change happened in v5.0 for some stuff that happened in qemu Mar 09 22:43:58 https://www.kraxel.org/blog/2019/03/edid-support-for-qemu/ Mar 09 22:44:21 he pushed changes in October, but I'm guessing this is the first version that has them Mar 09 22:49:19 jonmason: I'd guess so, yes since 4.19 didn't seem to have this issue **** ENDING LOGGING AT Sun Mar 10 02:59:56 2019