**** BEGIN LOGGING AT Sat May 15 02:59:58 2021 May 15 03:39:44 Build [#93](https://buildbot.openwrt.org/master/images/#builders/64/builds/93) of `realtek/generic` completed successfully. May 15 03:51:08 i build for my dev board use extend kernel have some error May 15 03:51:48 make[1]: Entering directory '/home/lazy/openwrt' May 15 03:51:48 make[2]: Entering directory '/home/lazy/openwrt' May 15 03:51:49 make[3]: Entering directory '/home/lazy/openwrt/target/linux' May 15 03:51:49 make[4]: Entering directory '/home/lazy/openwrt/target/linux/stm32mp1' May 15 03:51:50 rm -rf /home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1 May 15 03:51:50 mkdir -p /home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1 May 15 03:51:51 mkdir -p /home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1 May 15 03:51:51 if [ -d /home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1/linux-5.10.35 ]; then rmdir /home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1/linux-5.10.35; fi May 15 03:51:52 ln -s "/home/lazy/stm32mp/linux-5.10.10" /home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1/linux-5.10.35 May 15 03:51:52 export MAKEFLAGS= ; [ -d /home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1/linux-5.10.35/user_headers ] && rm -rf /home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1/linux-5.10.35/user_headers May 15 03:51:53 Makefile:27: recipe for target '/home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1/linux-5.10.35/.prepared_d41d8cd98f00b204e9800998ecf8427e' failed May 15 03:51:53 make[4]: *** [/home/lazy/openwrt/build_dir/target-arm_cortex-a7+neon-vfpv4_glibc_eabi/linux-stm32mp1/linux-5.10.35/.prepared_d41d8cd98f00b204e9800998ecf8427e] Error 1 May 15 03:51:54 make[4]: Leaving directory '/home/lazy/openwrt/target/linux/stm32mp1' May 15 03:51:54 Makefile:11: recipe for target 'compile' failed May 15 03:51:55 make[3]: *** [compile] Error 2 May 15 03:51:55 make[3]: Leaving directory '/home/lazy/openwrt/target/linux' May 15 03:51:56 time: target/linux/compile#0.03#0.01#0.04 May 15 03:51:56     ERROR: target/linux failed to build. May 15 07:13:10 netprince: thanks. let's revisit this again on monday (please ping me when you have time for testing) May 15 07:14:32 in the mean time, i will disable the pre-calibration code for now May 15 09:54:06 rmilecki: Not sure if this DSA/swconfig combination is feasible. Lots of b53 symbol conflicts. May 15 11:14:59 when bumping a downstream project using git (e.g. uclient, mt76), is there a script to automate the commit message? May 15 11:15:20 currently I'm using "git log --oneline" manually and copy/pasting May 15 11:17:51 also doing it manually, using git log --oneline --no-decorate --reverse May 15 11:20:29 I do have a oneline script to generate a 'Fixes: $hash ("commit msg")' line May 15 11:20:34 echo "Fixes: $(git log -1 --pretty='format:%h ("%s")' --abbrev=12 ${1})" May 15 11:21:07 cool, thanks May 15 11:21:37 ah, didn't notice the order is different depending on who does the commit ;) May 15 11:31:12 stintel: fyi - I think grub 2.06-rc1 might have problems with APU2s May 15 11:31:43 hmmmz May 15 11:31:50 I've only just got back online with my APU2 after killing it last night 2100 May 15 11:32:31 basically didn't even get as far as grub menu. May 15 11:32:38 what did you do to get it back online ? May 15 11:33:48 changed more than 1 thing, so trying to work out which is the really relevant thing. May 15 11:34:29 I backed out the grub 2.06 bump, and put back my 2.04 hack. also switched back to binutils 35. May 15 11:35:31 I don't think it's a gcc 10 thing. I don't think its a binutils thing 'cos I'm pretty sure I was running that. May 15 11:35:54 I'm running gcc 10 + binutils 2.36 on my APU2 May 15 11:36:04 I had to completely overwrite the internal APU HDD and then restore :-) May 15 11:36:13 fun May 15 11:36:32 now that I think of it, my APU2 was running with your fixed 2.04 grub May 15 11:36:53 so I have not tested that combination May 15 11:36:54 yeah, I'm 99.9% sure I'd got to binutils 2.36 as well. But there's a lot of changes, 2 kernel bumps, the ipv6 default. May 15 11:37:09 I'll build and test in a VM May 15 11:37:28 I've a gut feeling that grub2 2.06rc1 goes wonky. May 15 11:37:36 that would suck May 15 11:38:03 see this uncomfortable looking commit https://github.com/microsoft/CBL-Mariner/pull/781/commits/c4e38380b662cf5d83dc693b2879dcf4b733d3d1 May 15 11:38:05 maybe we messed up 100-grub_setup_root.patch May 15 11:38:33 yikes May 15 11:38:56 and I build with stack-protector-strong May 15 11:39:15 let me get back to 'latest everything except grub2 2.06' and prove to myself everything else is ok May 15 11:39:18 and FORTIFY_SOURCE=2 too :P May 15 11:39:42 rebuilding the toolchain complete so will take some time. May 15 11:39:53 well either way, 2.06-rc1 is the way forward to fix the CVEs etc May 15 11:40:11 let me check gentoo May 15 11:40:39 replace-flags -O3 -O2 May 15 11:41:38 don't seen anything of the fancy flag removal as in that ms commit though May 15 11:42:15 and honestly after everything I experienced with Microsoft staff and products ... May 15 11:42:15 I did a remote upgrade to my APU at home, it didn't come back - got the messsage from mrs ldir "I can't get on internet", "oh really, that's odd" feigns ignorance. Why do I never learn!? May 15 11:42:42 I would be cautious trusting that commit :P May 15 11:43:38 this week github enterprise went down, due to out of diskspace (hello, monitoring?!), and it took them a day to get it back up and running May 15 11:43:47 meanwhile hundreds of people unable to do their jobs May 15 11:44:57 oh I don't feel so bad... I got my APU back up in less than 18 hrs and at least 8 of those were sleeping - I'll tell mrs ldir to be grateful :-D May 15 12:09:50 stintel: are you still playing? I'm still building :-( May 15 12:14:34 ldir: boots fine in a VM May 15 12:15:30 so now I'm at the point where a 3rd APU2 would come in handy :P May 15 12:21:26 hmmm May 15 12:23:13 (I have one in Belgium in non-HA setup and one here as main router) May 15 12:23:40 in theory I can try here and if it breaks there should be little to no downtime, but I have to revise the HA setup because some things aren't working as they should anymore :/ May 15 12:32:19 stintel: the U6-LR has a 2.5 Gbit/s PHY, but i was not able to establish a 2.5 Gbit/s link with it. May 15 12:33:30 I've only tested with a crappy Realtek USB 2.5GE adapter (It didn't even produce a 1G link), so i wouldn't really judge on my experience here, given it works at 1G with everything else I had. May 15 12:34:50 The PHY has no linux driver, however U-Boot has support for it. still they write magic register values not found in the U-Boot nor the Linux driver to "enable 100Mbit/s". May 15 12:35:32 Sorry if I'm late to the party, had to stay in bed with fever for the last few days :( May 15 12:38:17 Hi everyone,  I'm attempting to add device support for the rb960pgs.  Any help would be greatly appreciated, especially in writing the dts file.  How do I find out all the values needed in order to (for example) get ethernet working? May 15 12:39:30 blocktrr1: covid? May 15 12:53:35 ldir: did three rapid-tests - all negative May 15 12:54:35 bet that was a surprise/relief - hope you're on the mend now though May 15 13:02:57 blocktrr1: interesting. I have an enterprise MultiGbe switch of which I'm sure it properly works. guess I'll have to find me some U6-LR somewhere May 15 13:03:50 blocktrr1: also, get well soon! May 15 13:05:39 pmelange: ideally look at the vendor GPL sources May 15 13:08:27 stintel: ok, I'm at last up to date except for the 2.06 bump, just 2.04 + my hack to build on latest binutils May 15 13:08:59 so I guess the next stage is to build with grub 2.06, write to a usb and see if that will boot. May 15 13:16:13 ldir: the other option is to compare package/boot/grub2/patches/100-grub_setup_root.patch between plntyk and my tree, maybe one of us got it wrong May 15 13:30:17 So far the only thing that I found is https://github.com/Ciusss89/routeros-linux-patch but I don't really know what to look for.  But thanks for the tip. May 15 13:31:51 pmelange: you can always contact mikrotik to request the sources for this specific device May 15 13:32:10 I guess I would try that after the weekend. May 15 13:32:37 other than that, if you come with a specific question it could be easier to help you May 15 13:40:37 Build [#51](https://buildbot.openwrt.org/openwrt-21.02/images/#builders/8/builds/51) of `mpc85xx/p1020` completed successfully. May 15 14:14:15 stintel: definitely grub 2.06rc1 breaks booting on APU2 (*) *at least when building on macos May 15 14:17:21 ldir: yep, mine is also dead May 15 14:17:32 well that blows May 15 14:18:18 that's good - at least it's not a macos thing - there's a positive data point out of that May 15 14:18:35 wow you optimist :D May 15 14:19:59 ldir: you previously used plntyk's commit ? May 15 14:20:03 and now mine? May 15 14:20:36 also quite annoying that it doesn't reproduce in a VM May 15 14:33:35 oh great May 15 14:33:40 can't even USB boot my APU2 May 15 14:37:11 ok, kingston is going on my vendor blacklist May 15 14:37:28 brand new USB stick, writes at less than 2MB/s, unbootable in the APU2 May 15 14:40:20 USB2? I always buy USB3 ones, even though they don't need USB3 interface speeds they tend to write 10-30MB/s instead of the USB2 ones which often just do 1-3MB/s May 15 14:41:36 my ancient Corsair USB2 writes >10MB/s easily May 15 14:41:59 but this kingston DTSE9 .. good for the trash May 15 14:42:02 and I bought 4 of them :./ May 15 14:42:38 I have yet to find a USB3 branded one that sucked that badly, only USB2 ones May 15 14:55:26 stintel: might be a fake one? May 15 14:56:03 well it's eastern europe, so who knows May 15 14:56:12 are you suggesting to contact kingston about it? May 15 14:58:34 depends on how you value your time May 15 14:59:25 On Amazon.de, with certain brands you can almost be certain to get a fake microsd/USB thumbdrive May 15 15:00:08 got them from something like mediamarkt but bulgarian May 15 15:00:54 ok and this product doesn't feature either of their Phantom or Colour-Shift Technology so it's difficult to verify May 15 15:05:08 stintel: sorry, had to go out - https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=shortlog;h=refs/heads/mine works, without those two last commits doesn't work. May 15 15:06:16 ldir: yeah I've seen that but yesterday you said you took the commit from the PR May 15 15:06:32 just want to rule out if I broke it or not :P May 15 15:08:02 Ah, I think I was talking about the 'making 2.04 work' commit, not updating to 2.06 May 15 15:08:23 oh well May 15 15:09:00 so I'm definitely building it with -D_FORTIFY_SOURCE=2 May 15 15:09:36 are you? May 15 15:10:25 how would I tell? I guess build logs May 15 15:10:37 to be 100% sure yes May 15 15:11:03 CONFIG_PKG_FORTIFY_SOURCE_2=y in .config should enable that May 15 15:11:03 let me build again with some logs May 15 15:11:44 no, CONFIG_PKG_FORTIFY_SOURCE_1=y May 15 15:11:53 k, so maybe it's really -O2 May 15 15:12:08 * stintel grabs an old x86/64 device May 15 15:12:43 CONFIG_TARGET_OPTIMIZATION="-O2 -pipe -march=btver2" May 15 15:20:07 also mdnsd is segfaulting :P May 15 15:20:16 maybe that fix wasn't so good after all May 15 15:22:41 mdnsd or umdns ? May 15 15:23:04 it's the same May 15 15:23:18 the package is called umdns, but the daemon is called mdnsd for some reason May 15 15:23:45 wait wat May 15 15:24:03 erm, no umdns package also mdnsd package - different May 15 15:24:08 package/network/services/umdns/Makefile:PKG_SOURCE_URL=$(PROJECT_GIT)/project/mdnsd.git May 15 15:25:06 ok that's where my confusion came from May 15 15:25:08 it's mdnsd May 15 15:25:12 that is crashing May 15 15:25:23 so it's NOT the thing you touched! May 15 15:27:09 ugh what a mess May 15 15:27:30 it's sub-optimal that's for sure May 15 15:33:26 alright I can reproduce the grub problem on another device May 15 15:37:11 the thing has been in a closet for so long I almost forgot about it :P May 15 15:38:02 building now without O2 May 15 15:40:05 this is all good news, not just macos/linux and not just APU2 May 15 15:40:50 https://savannah.gnu.org/bugs/?60458 May 15 15:46:43 oh well spotted May 15 15:47:51 ok I have a fix May 15 15:50:31 ldir: https://git.openwrt.org/45fa317a May 15 15:52:28 That has worked ? May 15 15:53:52 I mean I know it filters out the '-O2' - it feels horribly brittle. What if someone builds with '-O3'? May 15 15:53:53 yes May 15 15:54:20 well we could also add -O3 .. May 15 15:54:28 I mean to the filter May 15 15:54:42 * ldir ponders compiler bug or what funkiness are they doing in grub to provoke this May 15 15:55:30 can you filter '-O?' May 15 15:56:09 we shouldn't, -Os is fine May 15 15:57:47 I'd say we filter out -O2 -O3 May 15 15:57:55 should be safe enough May 15 15:57:59 the commit messages mentions it too May 15 15:58:03 Yeah ok May 15 16:00:00 I'm overthinking it. May 15 16:00:10 well O3 was a good catch May 15 16:00:12 https://git.openwrt.org/eaca0fe4 May 15 16:00:29 but does anyone really ever build with O3 though :P May 15 16:00:59 I was wondering where '-Os' was set, if it was set after our filter of '-O?' then all would be well, the default -Os would be used in all cases. May 15 16:01:06 grub2 - 2.06~rc1-1 May 15 16:01:29 that's my APU2 ;) May 15 16:01:50 yay! May 15 16:02:12 ok, I'll build (again!) with the latest commit :-) May 15 16:02:47 would be cool if I can get an ack or tested-by once it works ;) May 15 16:02:56 I'm now going to do a build with CONFIG_ALL=y May 15 16:03:01 and see what else breaks with GCC10 May 15 16:03:30 I hope in core not too much, I would like to push all this GCC 10 related stuff this weekend May 15 16:03:40 or I will probably lose motivation to continue on it May 15 16:05:04 I'll let you know as fast as my wheezing mac will let me :-) May 15 16:05:15 good thing I can make -j80 :D May 15 16:06:09 I'll catch up on the single threaded kernel refreshes ;-) May 15 16:06:18 haha very true May 15 16:06:26 clock speed is rather low here May 15 16:06:48 is AMP a thing yet :P May 15 16:06:51 (vs SMP) May 15 16:10:09 ahh, while at it I can also test ramips May 15 16:10:14 then I'm out of different targets :) May 15 16:25:55 stintel: you can add my ack May 15 16:31:32 ldir: great, thanks! May 15 16:34:03 hmmm libfreecwmp can't build as its mirror doesn't resolve May 15 16:34:05 dev.libfreecwmp.org May 15 16:34:15 thanks for your efforts in doing the bump & fixing the problems May 15 16:35:14 no worries May 15 16:36:16 I stopped working fulltime, no longer working on fridays, thursday was a public holiday in Belgium and my client is in Belgium so I had an extra long weekend and the previous long weekends were spent mostly off-screen so it was time to contribute a bit :) May 15 16:37:08 lol May 15 16:37:12 libfreecwmp.org doesn't even exist anymore May 15 16:37:20 I guess I have to remove the management feed May 15 16:56:08 lol building with CONFIG_ALL shows how much crap I still have May 15 17:04:04 oh fsck May 15 17:04:13 ran git clean -fdx in the wrong dir May 15 17:08:01 hurray for backups May 15 17:14:56 well that's good, most of the things that don't build are in my wip feed or in some other non-official feeds May 15 17:17:28 noooo I nuked my partition table on my apu2 so I lost sda3 May 15 17:17:29 grmbl May 15 17:21:02 ahh, good thing it's exactly the same on my other APU2 so I can just recreate the 3rd partition May 15 17:42:49 clayface: let's rename functions, that shouldn't be too hard May 15 17:42:57 clayface: or maybe use modules May 15 17:43:04 clayface: it's one time effort May 15 17:43:20 it's better and building another (sub)target for ever May 15 18:47:06 * ldir aha stintel fell into the same 'fun' that I fell into May 15 18:47:46 ;) May 15 18:49:52 I wondered if there was a way to recover a bootable system without trashing the partn table May 15 18:50:43 there is, manually extract the sysupgrade and dd boot to sda1 rootfs to sda2 May 15 18:51:08 ERROR: package/feeds/packages/rtl_433 failed to build. May 15 18:51:10 grmbl May 15 18:51:14 I fucking disabled that in the config May 15 18:51:24 seriously this buildroot is so fucking annoying sometimes **** ENDING LOGGING AT Sun May 16 03:00:13 2021