[SOLVED] Linux-ck update stuck

vinnom · 2018-02-07 22:20:38

Hi all!

I've been using linux-ck for a long time, but last week I tried to update it with no success. I was using repo-ck to grab the update, but I couldn't connect to it anymore. I waited some time and tried repeatedly until today.

As I couldn't connect to repo-ck anymore, I tried to update it using aur.

I made the following changes to PKGBUILD:

_subarch=8
_localmodcfg=y

I waited to update to end and it didn't. All I had was this (the last 6 lines in fact was output endlessly, I had to terminate the update).

What can I do to solve this?

Thanks in advance

Last edited by vinnom (2018-02-15 10:40:01)

jasonwryan · 2018-02-07 22:31:04

Moving to AUR Issues...

vinnom · 2018-02-14 10:31:34

vinnom wrote:

I've been using linux-ck for a long time, but last week I tried to update it with no success. I was using repo-ck to grab the update, but I couldn't connect to it anymore. I waited some time and tried repeatedly until today.
As I couldn't connect to repo-ck anymore, I tried to update it using aur.

I now know why I can't update via repo-ck anymore. Repo-ck Wiki Page says to add repo-ck - http://repo-ck.com/$arch - inside /etc/pacman.conf. This is url is no more and now repo-ck repo can be find at http://repo-ck.com/x86_64/. There is a mirror too: https://archd.hkno.it/repo-ck/os/x86_64/ (this other one doesn't have repo-ck mirror by the time I'm writting this post). Well, even adding the right servers, I can't find in it the right pre-compiled linux-ck package for me (linux-ck-piledriver). For some reason, it seems it wasn't compiled =/

My output of gcc -c -Q -march=native --help=target | grep march command, just for checking:

-march=                     		bdver2

vinnom wrote:

I made the following changes to PKGBUILD:
_subarch=8
_localmodcfg=y

I tried to do other editions to linux-ck PKGBUILD and found out that _subarch variable was messing things up. After setting nothing to it, the process seems to go fine. Then some user interation is prompted and I left every option as is (hit enter for each prompt). Well, the output was this:

arch/arm64/Makefile:48: Detected assembler with broken .inst; disassembly will be unreliable

and the process stops.
Although I did find some links on Google, neither solved my issue. Someone can help me with this? And maybe explain to me why selecting a subarch, I get that endless compilation process?

Thanks

V1del · 2018-02-14 10:45:52

You aren't making much sense, the $arch in the URL get's replaced with your architecture which should automatically expand to x86_64 given you are on x86_64. Are you? What's the output of

uname -a

vinnom · 2018-02-14 16:31:32

I don't know man, I had connection issues with repo-ck server. Then I tried aur, as I described. Then, before checking out inside repo-ck link, I removed linux-ck-piledriver and nvidia-ck when logged using official linux kernel, so I could get a clean environment (at least it was my intention). Everything else is already described. Even if $arch gets expanded to x86_64, repo-ck won't work for now, as it doesn't have linux-ck-piledriver package. The last one available is version 4.14.13-1 (using pacman to find it) and for some reason I can't obtain it (I get 404). The last archived version I found is 4.13.14-1.

My uname -a output is:

Linux archvinnom 4.15.2-2-ARCH #1 SMP PREEMPT Thu Feb 8 18:54:52 UTC 2018 x86_64 GNU/Linux

Maybe I mistook the culprit for not connecting the server, but it is not like I didn't make sense about other info I provided. I still don't have a clue why the process doesn't end (in case I choose a subarch variable) or it ends with an error. If it is the case, I can provide the entire output of the installation process, so I can get some help

loqs · 2018-02-14 16:51:55

https://bbs.archlinux.org/viewtopic.php … 9#p1767779 although that does not seem to match your build errors.

vinnom · 2018-02-14 18:23:13

loqs wrote:

https://bbs.archlinux.org/viewtopic.php … 9#p1767779 although that does not seem to match your build errors.

Yeah, although it doesn't match my build errors, it seems it won't build anyway, as is an architeture issue. I will use the snapshot I can grab from repo-ck site for now, but I'll keep checking that thread for updates.

Thanks

loqs · 2018-02-14 18:58:02

You should be able to build a kernel just not a kernel targetting Piledriver. Do you get the same results with _subarch=7 or _subarch=23?
If that fails comment out 'patch -Np1 -i "../${_gcc_patch}"' and do not set _subarch and see if that works.

vinnom · 2018-02-14 21:36:32

tried to track down where the issue began, for some reason I can't compile even packages since october. Strange, as I'm sure my last update regarding kernel was in december.

I have 2 issues: or it doesn't end the build process (as in log shared in OP) or it abort with this message "arch/arm64/Makefile:48: Detected assembler with broken .inst; disassembly will be unreliable".

loqs wrote:

You should be able to build a kernel just not a kernel targetting Piledriver. Do you get the same results with _subarch=7 or _subarch=23?
If that fails comment out 'patch -Np1 -i "../${_gcc_patch}"' and do not set _subarch and see if that works.

I tried to build using another _subarch value (both 7 and 23) and I had "Number of bits to use for ASLR of mmap base address for compatible applications (ARCH_MMAP_RND_COMPAT_BITS) [11] (NEW)" endlessly

loqs · 2018-02-14 22:23:30

Try downloading the PKGBUILD again then run makepkg -Cf when prompted for optimization pick 7 or 23 and see if that builds a kernel.

vinnom · 2018-02-14 23:08:09

Just ran "makepkg -Cf" and I wasn't prompted for optimization... in the end it aborted with the same error about broken assembler

loqs · 2018-02-15 00:09:37

If makepkg -Cf is not stopping on the following prompt then the PKGBUILD has been changed

Processor family
  1. AMD Opteron/Athlon64/Hammer/K8 (MK8)
  2. AMD Opteron/Athlon64/Hammer/K8 with SSE3 (MK8SSE3) (NEW)
  3. AMD 61xx/7x50/PhenomX3/X4/II/K10 (MK10) (NEW)
  4. AMD Barcelona (MBARCELONA) (NEW)
  5. AMD Bobcat (MBOBCAT) (NEW)
  6. AMD Jaguar (MJAGUAR) (NEW)
  7. AMD Bulldozer (MBULLDOZER) (NEW)
  8. AMD Piledriver (MPILEDRIVER) (NEW)
  9. AMD Steamroller (MSTEAMROLLER) (NEW)
  10. AMD Excavator (MEXCAVATOR) (NEW)
  11. AMD Zen (MZEN) (NEW)
  12. Intel P4 / older Netburst based Xeon (MPSC)
  13. Intel Atom (MATOM)
  14. Intel Core 2 (MCORE2)
  15. Intel Nehalem (MNEHALEM) (NEW)
  16. Intel Westmere (MWESTMERE) (NEW)
  17. Intel Silvermont (MSILVERMONT) (NEW)
  18. Intel Sandy Bridge (MSANDYBRIDGE) (NEW)
  19. Intel Ivy Bridge (MIVYBRIDGE) (NEW)
  20. Intel Haswell (MHASWELL) (NEW)
  21. Intel Broadwell (MBROADWELL) (NEW)
  22. Intel Skylake (MSKYLAKE) (NEW)
> 23. Generic-x86-64 (GENERIC_CPU)
  24. Native optimizations autodetected by GCC (MNATIVE) (NEW)

What is the output of the following

$ git status
$ git log -n 1

vinnom · 2018-02-15 00:31:21

$git status
fatal: Not a git repository (or any parent up to mount point /)
Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).

$ git log -n 1
fatal: Not a git repository (or any parent up to mount point /)
Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).

I tried several times, it doesn't stop for prompting cpu arch

loqs · 2018-02-15 00:35:57

So you are not using a git clone of the aur repository so you extracted the files from a snapshot tarball?
If you extract the latest tarball overwriting the existing files or extract to a new directory then run makepkg -Cf it should prompt.

vinnom · 2018-02-15 00:50:35

Oh yeah, I downloaded from here as I was testing various versions. I downloaded the latest in it, which is 4.14.19-1.

I want to add that I had ARCH and CROSS_COMPILE variables set to:

export ARCH=arm64
export CROSS_COMPILE=$HOME/Android/toolchain/aarch64-linux-android-4.9/bin/aarch64-linux-android-

Inside my ~/.bashrc. I found a thread in stackoverflow that remind me of that and I changed it just after... but it seems that it had nothing to do with my issues, as they persist

loqs · 2018-02-15 01:02:03

Have you extracted the tarball to a clean directory? What is the output of the following working in the clean directory?

$ env #to makesure ARCH is not set
$ makepkg -Cf #to check that with the cleanly extract files it is prompting for Processor family

Edit:
If the first prompt is this

Stack Protector buffer overflow detection
  1. None (CC_STACKPROTECTOR_NONE)
  2. Regular (CC_STACKPROTECTOR_REGULAR)
> 3. Strong (CC_STACKPROTECTOR_STRONG)
choice[1-3?]: 3

and ends in

arch/arm64/Makefile:48: Detected assembler with broken .inst; disassembly will be unreliable
  CHK     include/config/kernel.release
Cannot use CONFIG_CC_STACKPROTECTOR_STRONG: -fstack-protector-strong not supported by compiler
make: *** [Makefile:1098: prepare-compiler-check] Error 1
make: *** Waiting for unfinished jobs....
  UPD     include/config/kernel.release
==> ERROR: A failure occurred in prepare().
    Aborting...

Then ARCH is still set to arm64

Last edited by loqs (2018-02-15 01:13:14)

vinnom · 2018-02-15 10:39:29

Well, firstly big thanks! Do you drink bee @loqs? Cause you really deserve one!

What I was doing:

- extract each linux-ck tarball to new directory, so I was sure that it was a clean directory
- as soon as I noticed ARCH and CROSS_COMPILE was set to something, I commented them out inside ~/.bashrc
- I sourced ~/.bashrc
- And then, I ran multiple times "makepkg -Cf", even trying to change some things

Now I realize that commenting out the variables wasn't enough, if they were already set. I should have explicitly set them to nothing or even should have set it to "x86_64".

Well, in that moment, I was pretty sure that I did everything right (which I surely didn't) but even so I tried one more time.

- I cleaned the folders I used
- Extracted linux-ck to another folder
- Started a new terminal
- Checked if ARCH and CROSS_COMPILE wasn't set to something
- ran makepkg -Cf

And now everything went fine. I finally build the kernel (with generic option), installed it and nvidia-ck. In my concern, this thread is solved and I will set it like that. Thank you very much @loqs. I will keep checking graysky thread for more info about piledriver

Arch Linux

#1 2018-02-07 22:20:38

[SOLVED] Linux-ck update stuck

#2 2018-02-07 22:31:04

Re: [SOLVED] Linux-ck update stuck

#3 2018-02-14 10:31:34

Re: [SOLVED] Linux-ck update stuck

#4 2018-02-14 10:45:52

Re: [SOLVED] Linux-ck update stuck

#5 2018-02-14 16:31:32

Re: [SOLVED] Linux-ck update stuck

#6 2018-02-14 16:51:55

Re: [SOLVED] Linux-ck update stuck

#7 2018-02-14 18:23:13

Re: [SOLVED] Linux-ck update stuck

#8 2018-02-14 18:58:02

Re: [SOLVED] Linux-ck update stuck

#9 2018-02-14 21:36:32

Re: [SOLVED] Linux-ck update stuck

#10 2018-02-14 22:23:30

Re: [SOLVED] Linux-ck update stuck

#11 2018-02-14 23:08:09

Re: [SOLVED] Linux-ck update stuck

#12 2018-02-15 00:09:37

Re: [SOLVED] Linux-ck update stuck

#13 2018-02-15 00:31:21

Re: [SOLVED] Linux-ck update stuck

#14 2018-02-15 00:35:57

Re: [SOLVED] Linux-ck update stuck

#15 2018-02-15 00:50:35

Re: [SOLVED] Linux-ck update stuck

#16 2018-02-15 01:02:03

Re: [SOLVED] Linux-ck update stuck

#17 2018-02-15 10:39:29

Re: [SOLVED] Linux-ck update stuck

Board footer