2024-11-01 18:56:44 <ikke> openjdk21 has been bootstrapped
2024-11-02 01:52:42 <huajingyun> Ok,thanks ikke
2024-11-05 17:35:51 <mio> a bit of status summary for loongarch 3.21 rebuild: it's currently at ~66 aports left to rebuild (after the plasma upgrade finishes building). the vala upgrade that came through yesterday unlocked 40-50 packages that were waiting on accountsservice/libportal. most of the remaining ones fall into 3 issue types:
2024-11-05 17:38:05 <mio> 1. numpy 2/py3-matplotlib: the latter needs an upgrade/patching to be numpy 2 compatible, someone put in a MR for review. a number of py3 aports are failing tests due to needing the numpy2 matplotlib
2024-11-05 17:39:16 <mio> 2. aports depending on qt6-qtwebengine-dev (currently ftbfs)
2024-11-05 17:43:06 <mio> 3. .so tracing error: a number of aports failed due to a .so object not found (this previously a warning but with newer abuild becomes an error). in some cases abuild doesn't know yet how to detect $origin/rpaths
2024-11-05 17:45:26 <mio> example of affected aport is freerdp, there is an abuild draft to improve detection, not sure of its current status, and whether people should work around the issue for now in the affected aports 
2024-11-05 17:45:55 <mio> (freerdp has a MR waiting for that decision)
2024-11-05 17:50:12 <mio> 1b. other aports with all dependencies on numpy2 already and only themselves need upgrading to numpy 2 ... this is in progress
2024-11-05 17:51:59 <mio> there is also a small group of py aports depending (directly or indirectly) on py3-pyarrow that may be broken (-dask-expr, -dask)
2024-11-05 18:29:03 <mio> for other arches it's a more mixed bag, but 1 and 2 would still apply
2024-11-05 18:31:54 <mio> (at least so far ... 32-bit arches currently have a bunch of aports failed due to time_t issue and/or using deprecated GTimeVal)
2024-11-05 18:34:52 <mio> (and 3 is not loongarch-specific but that most of them have been resolved on a per-case basis)
2024-11-06 01:19:07 <huajingyun> Thanks mio for the detailed explanation
2024-11-06 01:19:58 <huajingyun> qt6-qtwebengine loongarch64 is currently disabled, but other architectures seem to fail due to the rebuild as well, maybe we can try to re-build qt6-qtwebengine for loongarch64 after error is fixed
2024-11-06 01:42:12 <mio> huajingyun: yeah, hopefully it will be okay after that. with the lower number of aports left to build, unfortunately it was blocking the builder more frequently due to taking somewhat longer to fail
2024-11-06 01:43:10 <mio> (as it has several aports depending on its -dev subpackage, so it comes up a bit more)
2024-11-06 01:43:58 <mio> (and was temporarily disabled to allow time to build other aports as they get fixed)
2024-11-06 01:50:14 <mio> but slowly making progress :)
2024-11-06 01:59:18 <huajingyun> mio:Yeah, I checked and wanted to try to fix some of these aports, but I'm a bit overwhelmed, like the ones you mentioned 1 and 3, numpy v2 vs v1
2024-11-06 02:22:54 <mio> huajingyun: thanks for looking into some of them! no worries if overwhelmed. not sure if it might be something you might be interested in, only if you want to, two aports that are still need fixing are tilix and openmp
2024-11-06 02:23:56 <mio> my understanding is openmp is usually upgraded to match the latest llvm version, but is currently blocked on some bitcode issue
2024-11-06 02:24:36 <mio> not sure if resolved yet upstream
2024-11-06 02:25:25 <mio> tilix has some undefined reference error
2024-11-06 02:25:58 <mio> haven't found a working fix for it yet upstream
2024-11-06 02:27:31 <mio> njs has a failed test, maybe related to libxml2
2024-11-06 02:29:20 <mio> was going to look into njs, but testing two numpy-related aports at the moment
2024-11-06 02:31:40 <mio> no worries if you're busy though! just ideas :)
2024-11-06 03:40:51 <huajingyun> mio:Ok,thanks for reminding me of these,I'll check it out when I have time
2024-11-06 03:51:10 <mio> huajingyun: wonderful, thanks :)
2024-11-12 17:00:54 <ikke> Loongarch64 apparently has finished community
2024-11-13 00:58:42 <huajingyun> ikke:Yeah, cely helped temporarily disable a few aports to unlock community/, will re-enabled loongarch when people fixed them
2024-11-13 00:59:11 <huajingyun> https://git.alpinelinux.org/aports/commit/?id=40fc854b6b4c
2024-11-13 01:07:47 <cely> qt6-qtwebengine was disabled by maintainer in https://git.alpinelinux.org/aports/commit/?id=24cfe194e2fa so 7 out of 12 aports disabled in my commit are just follow-ups to that
2024-11-13 01:16:44 <cely> Out of the remaining 5, 3 (openmp itself included) are due to openmp in llvm19 moving omptarget to a new subproject: https://github.com/llvm/llvm-project/pull/75125
2024-11-13 01:21:06 <cely> I think libomptarget* may have to be built in either main/llvm19 itself, or main/llvm-runtimes, but i have not managed to get it to work, so probably someone more familiar with LLVM will have to work on that
2024-11-13 01:21:37 <cely> Based on a quick grep through the APKBUILDs, it seems no other aport references "omptarget"
2024-11-13 01:22:41 <cely> and pkgs.a.o only shows libomptarget* being required by "openmp-dev", which i think may no longer be the case, since they are now separate subprojects
2024-11-13 01:25:08 <cely> So, maybe we can consider building only openmp (from what i remember, building openmp itself still works) and leave libomptarget* to another aport, or something, but i'm not familiar enough with LLVM/OpenMP to make such a decision
2024-11-13 01:29:24 <huajingyun> From the apkbuild of openmp, aarch64, ppc64le, x86_64 and riscv64 will involve libomptarget
2024-11-13 01:30:56 <cely> Yeah, but from what i see, the openmp source tarball we are using no longer includes the code for building libomptarget
2024-11-13 01:43:04 <cely> Not sure which split tarball contains offload/omptarget, it's in the big llvm-project tarball, so probably that's what we'll have to use if we still want libomptarget
2024-11-13 07:07:59 <cely> I'll be going AFK, bye
2024-11-13 07:10:14 <huajingyun> Bye
2024-11-14 11:14:12 <ncopa> i thikn the loongarch patches broke qt6-qtwebengine build
2024-11-15 06:18:13 <ncopa> morning! i need help with https://gitlab.alpinelinux.org/alpine/aports/-/jobs/1605219
2024-11-15 06:18:45 <ikke> 'error: use of undeclared identifier 'HWCAP_LOONGARCH_LSX''
2024-11-15 06:21:18 <huajingyun> Ok,let me see
2024-11-15 06:35:55 <huajingyun> It looks like these defines are missing from musl,we will check and fix it as soon as possible.
2024-11-15 07:35:21 <huajingyun> ncopa: a fix MR in !75227
2024-11-16 21:19:49 <ikke> Any problem if we remove the old loongarch repos at dev.a.o/edge?
2024-11-18 01:56:16 <huajingyun> ikke:Not sure if everyone has switched to dl-cdn.a.o/alpine/edge/, I think there is a difference between the two now, which is about qt6-qtwebengine, the qt6-qtwebengine-6.7.2 apk package is currently only in dev.a.o/edge (a few days ago, in the qt6-qtwebengine upgrade, loongarch64 was disabled)
2024-11-18 01:56:21 <huajingyun> it would be good if people who are currently using qt6-qtwebengine knew about this
2024-11-18 02:00:06 <huajingyun> By the way, DWwanghao is my colleague, he will also join the alpine porting:)
2024-11-18 02:31:31 <DWwanghao> :)Hi everyone, I’m DWwanghao. I’m excited to join the team and work on the Alpine porting project. Looking forward to learning from and collaborating with all of you!
2024-11-18 02:34:19 <huajingyun> Welcome!
2024-11-18 02:50:20 <mio> DWwanghao: welcome!
2024-11-18 04:04:36 <ikke> DWwanghao: welcind
2024-11-18 04:04:42 <ikke> welcome*
2024-11-18 06:20:06 <mps> DWwanghao: hi, and welcome!
2024-11-18 06:22:42 <mps> linux 6.12 is released (last night) and I guess it will be next LTS kernel
2024-11-18 06:23:53 <mps> so, it linux-lts should be probably upgraded to it to be in 3.21-stable alpine
2024-11-18 06:25:23 <mps> this version have some new features for loongarch64
2024-11-18 08:24:14 <ncopa> hi DWwanghao! nice to have you on-board! welcome!
2024-11-18 08:24:31 <ncopa> mps: yes would be nice with 6.12 kernel
2024-11-18 08:26:07 <ncopa> mps: you have experience with the edge kernels (post 6.6 ...). are there anything particular we should enable? anythign that not in defconfig
2024-11-18 08:56:25 <mps> ncopa: yes, some sound drivers for x86/_64 are removed and replaced with new in 6.12. didn't yet tested them
2024-11-18 08:57:44 <mps> new 'thing' is RT (Real Time) is now in kernel but I think we should not enable this right now till it is not tested more
2024-11-18 08:59:31 <mps> I will look in changelog for kernels 6.7 to 6.12 to remind myself if something should be 'special' to enable
2024-11-18 09:00:06 <mps> though defaults important will be enabled automatically
2024-11-18 09:11:09 <mps> oh, there is praised sched_ext scheduler based on BPF but I'm not sure is it super good, but probably could be enabled
2024-11-18 09:13:43 <mps> I expect 6.12 to be better on loongarch64 and riscv
2024-11-18 09:14:26 <mps> there are also some arm good improvements and new SoCs and SBCs
2024-11-18 09:14:56 <mps> initial RPI5 support
2024-11-18 09:15:53 <mps> curiosa (and funny to me) is QR code DRM panic message
2024-11-18 09:17:18 <mps> does multi-path PCI exists in 6.6?
2024-11-18 09:17:34 <ncopa> dont know
2024-11-18 09:18:18 <mps> loongarch have speedy binary KVM translation
2024-11-18 09:18:51 <mps> vDSO is improved
2024-11-18 09:20:40 <mps> there is IPE (Integrity Policy Enforcement) which could be interesting to IoT and 'locked' routers/access points, not sure is it interesting to alpine kernels
2024-11-18 09:21:41 <mps> intel XE drm driver should be enabled
2024-11-18 09:22:18 <mps> there are some AI accelerators added
2024-11-18 09:23:17 <mps> bcachefs if not enabled on 6.6 linux-lts
2024-11-18 09:24:44 <mps> some ChromeOS drivers maybe could be enable, but we don't ship kernels for chromebooks (except two specific) afaik
2024-11-18 09:25:22 <mps> Device Memory TCP
2024-11-18 09:32:26 <mps> PoverVR drm should be in kernel but I can't remember if I enabled it
2024-11-18 09:35:48 <mps> Rust in kernel >.<
2024-11-18 09:37:20 <mps> ncopa: IMO you should not hurry to enable everything new but in usual way, when someone request something then look at it
2024-11-18 09:49:55 <ncopa> yeah sounds like a good idea
2024-11-18 09:50:34 <ncopa> im fixing the 3rd party kernel builds.
2024-11-18 10:04:10 <huajingyun> Agree,perhaps in this upgrade to 6.12, you can prioritize using only defconfig,then add as needed
2024-11-18 10:09:52 <mps> in relation to 6.6 kernel 6.12 have a lot of improvements for loongarch and most of them are enabled automatically
2024-11-18 11:46:31 <ncopa> Linux ncopa-loongarch64 6.12.0-0-lts #1-Alpine SMP PREEMPT 2024-11-18 10:39:53 loongarch64 Linux
2024-11-18 11:46:37 <ncopa> seems to work just fine
2024-11-18 11:50:37 <mps> I run linux-6.12-rc series few weeks, also works fine
2024-11-18 11:52:00 <mps> will try to find time to upgrade linux-edge to 6.12 for all supported arches
2024-11-18 12:57:58 <mps> uname -a => Linux la.arvanta.net 6.12.0-0-edge #1-Alpine SMP PREEMPT_DYNAMIC Mon, 18 Nov 2024 12:07:04 +0000 loongarch64 Linux
2024-11-18 18:28:03 <mps> ncopa: modules in linux-lts are not compressed?
2024-11-18 18:28:22 <mps> 6.6.x series, I mean
2024-11-18 19:32:12 <mps> hm, they are compressed on some arch but not all?
2024-11-18 20:38:19 <mps> just pushed linux-edge 6.12.0 to builders. in about 30 minutes it will be ready to upgrade/install on loongarch64
2024-11-19 08:20:25 <Ariadne> where is the latest alpine boot ISO for loongarch?
2024-11-19 08:33:33 <ncopa> i dont think we have any boot iso for loongarch yet
2024-11-19 08:42:01 <huajingyun> I created an alpine-standard-3.21.1113-loongarch64.iso locally a few days ago (November 13th)
2024-11-19 08:54:55 <huajingyun> Ariadne: I uploaded it here temporarily, it is only for temporary use, and still need to wait for the official release
2024-11-19 08:55:26 <huajingyun> https://gitlab.alpinelinux.org/huajingyun01/TemFiles/-/blob/master/iso-tmp/alpine-standard-3.21.1113-loongarch64.iso
2024-11-19 16:08:25 <Ariadne> nice
2024-11-19 17:50:01 <Ariadne> huajingyun: are you sure this iso works?
2024-11-19 19:51:13 <Ariadne> partition 1 is empty, partition 3 is empty
2024-11-20 01:08:27 <huajingyun> Ariadne: did you try and it didn't work? 
2024-11-20 01:24:07 <huajingyun> A few days ago DWwanghao booted this iso and it works,I just tried again, it works too. 
2024-11-21 00:04:47 <Ariadne> huajingyun: how did you boot it?  i flashed it to a usb stick, but it failed to boot on both my evaluation board, my asus board, and as well as this new 3A6M50 desktop board i have
2024-11-21 00:05:21 <Ariadne> sadly, i do not have any ability to read or write actual optical media anymore
2024-11-21 00:14:36 <Ariadne> in the meantime, i created an alpine container, and then cloned said container to a usb stick, and used that to boot + install alpine for real
2024-11-21 01:22:20 <huajingyun> Ariadne: I used the "dd" command to burn the iso to my USB drive, and then booted it in the 3a6000 desktop machine
2024-11-21 01:22:27 <huajingyun> can you check the size of the iso you downloaded?
2024-11-21 01:27:54 <Ariadne> it was 93mb
2024-11-21 02:03:33 <huajingyun> The original iso size I have here is 94mb
2024-11-21 02:03:55 <huajingyun> I uploaded it again here https://dev.alpinelinux.org/~loongarch/hjytmp/iso-tmp/ , maybe you are willing to try this
2024-11-21 09:42:12 <ncopa> huajingyun: re https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/75538#note_457692
2024-11-21 09:42:23 <ncopa> for some reason it is not applied
2024-11-21 09:43:03 <ncopa> ncopa-desktop:~/aports/main/linux-lts (linux-lts-6.12)$ CONFIG_CPU_HAS_LSX=y CONFIG_CPU_HAS_LASX=y CONFIGS=lts.loongarch64.config abuild updateconfigs
2024-11-21 09:43:05 <ncopa> ...
2024-11-21 09:43:16 <ncopa> ncopa-desktop:~/aports/main/linux-lts (linux-lts-6.12)$ grep CPU_HAS_ src/build-lts.loongarch64/.config
2024-11-21 09:43:16 <ncopa> CONFIG_CPU_HAS_FPU=y
2024-11-21 09:43:16 <ncopa> CONFIG_CPU_HAS_PREFETCH=y
2024-11-21 09:44:00 <ncopa> at least not when running from x86_64 machine
2024-11-21 09:51:14 <ncopa> ok, it worked when done on a loongarch64 machine
2024-11-21 10:12:03 <mps> linux-edge have these in config (from 6.11 or 6.10, I forgot exactly)
2024-11-21 10:59:46 <huajingyun> ncopa: Yeah, they only work on loongarch64
2024-11-21 11:01:09 <ncopa> they get removed with crosscompile make olddefconfig
2024-11-21 11:11:51 <huajingyun> ncopa: do you mean you cross-compile linux-lts 6.12 first?
2024-11-21 11:16:37 <ncopa> no, I just "cross compile" to update config
2024-11-21 11:23:03 <huajingyun> Ok, so it is a little different from the default config loongson3_defconfig in the kernel source
2024-11-21 11:40:03 <huajingyun> ncopa: I also noticed CONFIG_VIRTUALIZATION=y CONFIG_KVM=m, can you help add them to linux-lts 6.12 lts.loongarch64.config? linux-lts 6.6 does not support it yet 
2024-11-21 12:32:55 <ncopa> ok
2024-11-21 12:33:11 <ncopa> good catch. we also lost it on aarch64
2024-11-21 12:33:30 <ncopa> so I added it back there too
2024-11-21 12:35:02 <huajingyun> ok, thanks a lot:)
2024-11-22 09:17:12 <ncopa> Ariadne: how are you doing? i wonder if you could help me with one thing. we have asked docker to support loongarch, but there seems to be no interest from their side
2024-11-22 09:17:34 <ncopa> which means that we now will have to ship the docker images two places, depending on the arch
2024-11-22 09:18:21 <ncopa> I wonder if you could help me come up with some plan what to do with docker images
2024-11-22 09:19:22 <ncopa> im thinking that we and build them in our own infra, our own registry, and then simply have an empyt FROM ... in the official images repo at docker
2024-11-22 09:19:34 <ncopa> and maybe also push the images to quay.io or something
2024-11-22 09:19:49 <ncopa> ideally also sign them properly
2024-11-22 09:21:13 <ncopa> ref https://github.com/docker-library/official-images/pull/17607
2024-11-22 09:47:18 <mps> :(
2024-11-22 09:50:43 <ikke> registry.alpinelinux.org/img/alpine already supports loongarch64
2024-11-22 09:51:07 <ikke> Signing is on my todo list 
2024-11-22 09:51:22 <ikke> As well as maybe shortening the domain 
2024-11-22 09:52:22 <ikke> But providing images at scale is a different issue 
2024-11-22 09:58:11 <clandmeter> we should ask loongarch for support on this matter
2024-11-22 09:58:25 <clandmeter> for the infra i mean
2024-11-22 10:02:04 <huajingyun> clandmeter:  infra, so does alpine require new machine for docker image building?
2024-11-22 10:02:36 <clandmeter> not hardware, but space to host stuff
2024-11-22 10:03:12 <clandmeter> i am actually actively looking for coloc space for alpine
2024-11-22 10:03:47 <clandmeter> but cloud space is also a possible help 
2024-11-22 10:04:26 <clandmeter> if docker does not want to manage/host loongarch, we need to find an alternative 
2024-11-22 10:04:35 <clandmeter> and i guess not only alpine
2024-11-22 10:13:55 <huajingyun> Do you mean this space is used to store the docker image of loongarch64? And all future versions, like 3.22, etc., also include those other than alpine?
2024-11-22 10:15:53 <huajingyun> clandmeter: this reminds me that we have a repository on docker hub, https://hubgw.docker.com/search?q=loongarch64
2024-11-22 10:40:08 <huajingyun>  Not sure what you think?  If there's anything we can do to provide support, let me know and I'll follow up 
2024-11-22 10:40:08 <huajingyun> Ok,but I have to AFK now, bye
2024-11-22 15:51:58 <fossdd[m]> they could be hosted under the docker.io/alpinelinux namespace: https://hub.docker.com/u/alpinelinux
2024-11-22 15:54:58 <fossdd[m]> but a independent registry not from docker sounds good
2024-11-22 16:12:29 <mps> maybe it is time to remove iptables from kernel now we will have 6.12 as lts
2024-11-22 16:15:10 <ikke> We still use it primarily
2024-11-22 16:25:04 <mps> awall?
2024-11-22 16:25:53 <mps> imo, we should work on move forward on this
2024-11-22 16:26:12 <ikke> what about docker?
2024-11-22 16:26:25 <mps> uhm yes, I forgot
2024-11-22 16:27:26 <mps> anyway, we are in 'bad stance' (like limbo)
2024-11-22 16:28:49 <mps> 'stans', to be more latin :)
2024-11-22 18:03:02 <mps> nice news, https://www.phoronix.com/news/AMDGPU-DC-LoongArch
2024-11-22 18:03:24 <mps> maybe it will be usable as desktop
2024-11-22 20:08:48 <fossdd[m]> that sounds nice
2024-11-23 06:28:43 <woman> linux
2024-11-25 18:50:29 <celie> Bouncer fail led to replay of old messages, sorry
2024-11-25 18:50:55 <ikke> didn't notice anything
2024-11-25 18:52:50 <celie> If the GCJ bootstrap path has to involve JDK7, then i think that's a big roadblock itself
2024-11-25 18:52:58 <celie> Even if we could find the required GCC 6 patches to get it to work on Loongarch, and then in turn got GCJ and ECJ working, JDK7 has no support for Loongarch, as far as i know
2024-11-25 19:07:30 <f_> 
ACTION thinks.. what kind of bouncer is this that resends months old messages..

2024-11-25 19:08:11 <ikke> I have no idea
2024-11-28 20:24:50 <Ariadne> i have gotten additional loongarch machines, including a couple of industrial ones which are very interesting.  anyway, a lot of these industrial machines use YT6801 NICs, which lack an upstream driver.  now you can `apk add tuxedo-yt6801` to install tuxedo's yt6801 driver though.
2024-11-28 21:07:30 <mps> lucky you
2024-11-28 21:46:06 <ncopa> cool
2024-11-30 01:41:01 <huajingyun> Hi, anyone know why loongarch64 is missing from dl-cdn.a.o/alpine/v3.21/releases/
2024-11-30 01:41:10 <huajingyun> Thanks
2024-11-30 07:53:32 <ncopa> hum, I suppose something broke?
2024-11-30 08:07:50 <ncopa> i manually uploaded it for now. will investigate more on monday