LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Kernel (https://www.linuxquestions.org/questions/linux-kernel-70/)
-   -   Can't boot into 6.3 - halts at rtl8192ce (https://www.linuxquestions.org/questions/linux-kernel-70/cant-boot-into-6-3-halts-at-rtl8192ce-4175724420/)

RandomTroll 04-24-2023 09:54 PM

Can't boot into 6.3 - halts at rtl8192ce
 
I upgraded the kernel from 6.2.12 to 6.3 this morning, remembered to re-boot now. It halts at
Quote:

rtl8192ce: Chip Version ID:B_CUT_88C
I always keep the previous working kernel around so booted into 6.2.12 without incident.

Later:
I removed the rtl8192ce module and it halted after loading a bunch of USB devices. I unplugged all USB devices and it halts after
Quote:

kallsyms_selftest: finish
A day later:
I downloaded the source for 6.3 (I've been patching for years.) When I tried to build I got:
Quote:

In file included from arch/x86/decode.c:10:
/usr/src/linux-6.3/tools/arch/x86/include/asm/insn.h:10:10: fatal error: asm/byteorder.h: No such file or directory
10 | #include
| ^~~~~~~~~~~~~~~~~
compilation terminated.
but /usr/include/include/asm/byteorder.h is there.

bathory 04-26-2023 07:15 AM

Quote:

Originally Posted by RandomTroll (Post 6426785)
I upgraded the kernel from 6.2.12 to 6.3 this morning, remembered to re-boot now. It halts at

I always keep the previous working kernel around so booted into 6.2.12 without incident.

Later:
I removed the rtl8192ce module and it halted after loading a bunch of USB devices. I unplugged all USB devices and it halts after

A day later:
I downloaded the source for 6.3 (I've been patching for years.) When I tried to build I got:
but /usr/include/include/asm/byteorder.h is there.

Looks like you had the same problem previously.
Follow the steps you did back then in order to compile the new kernel...

RandomTroll 04-26-2023 09:01 AM

Quote:

Originally Posted by bathory (Post 6427076)
Looks like you had the same problem previously.
Follow the steps you did back then in order to compile the new kernel...

Good memory! I had a different problem back then: patch-5.12 had eliminated byteorder.h - I could copy the old one over and get a build. kernel 6.3 still has byteorder.h. Building from a full source of 6.3 doesn't find it - why not?


And my original problem is that kernel 6.3 halts. Have you built 6.3 yet?

bathory 04-26-2023 10:21 AM

Quote:

Originally Posted by RandomTroll (Post 6427095)
Good memory! I had a different problem back then: patch-5.12 had eliminated byteorder.h - I could copy the old one over and get a build. kernel 6.3 still has byteorder.h. Building from a full source of 6.3 doesn't find it - why not?


And my original problem is that kernel 6.3 halts. Have you built 6.3 yet?

Of course. I'm already running it on my Slackware64:
Code:

uname -r
6.3.0

I don't know what are you doing wrong? I'm always follow the next steps to build a new kernel:
  1. extract the new kernel in /usr/src/
  2. copy the .config of the running kernel in the newly created dir
  3. run make oldconfig
  4. run make
  5. run make modules_install
  6. manually copy bzImage and System.map under /boot
  7. edit /etc/lilo.conf
  8. run lilo
  9. reboot

RandomTroll 04-26-2023 12:44 PM

Quote:

Originally Posted by bathory (Post 6427117)
Of course. I'm already running it on my Slackware64

Thanks. It's good to know.

Quote:

Originally Posted by bathory (Post 6427117)
I don't know what are you doing wrong?

I've been building my own kernel for 20 years. I changed the includes in insn.h and orc_types.h to point directly to byteorder.h and built. There must be thousands of other <> type includes in the source that point to the same directories - why does it only matter on these 2 files? The kernel built from the whole source halts at the same place.

What's your architecture?

bathory 04-26-2023 03:06 PM

Quote:

Originally Posted by RandomTroll (Post 6427142)
Thanks. It's good to know.


I've been building my own kernel for 20 years. I changed the includes in insn.h and orc_types.h to point directly to byteorder.h and built. There must be thousands of other <> type includes in the source that point to the same directories - why does it only matter on these 2 files? The kernel built from the whole source halts at the same place.

What's your architecture?

I'm running Slackware64-current on an old Intel Core i5-2300 CPU with 8G RAM

RandomTroll 04-29-2023 03:34 PM

Quote:

Originally Posted by bathory (Post 6427151)
I'm running Slackware64-current on an old Intel Core i5-2300 CPU with 8G RAM

Thanks. I narrowed down the halt to the loading of either snd_hda_intel or i915. Do you use either? I need snd_hda_intel for sound and i915 provides my preferred video mode.

bathory 04-30-2023 02:10 AM

Quote:

Originally Posted by RandomTroll (Post 6427814)
Thanks. I narrowed down the halt to the loading of either snd_hda_intel or i915. Do you use either? I need snd_hda_intel for sound and i915 provides my preferred video mode.

These modules are also loaded by my kernel, because I'm using an Intel i5 CPU and chipset.
For video though I'm using an old Nvidia card with the patched legacy 470.182.03 drivers.

RandomTroll 04-30-2023 10:20 PM

Quote:

Originally Posted by bathory (Post 6427907)
These modules are also loaded by my kernel

Thanks! So it's just me, eh? hmmmmm...

bathory 05-01-2023 03:37 AM

Quote:

Originally Posted by RandomTroll (Post 6428116)
Thanks! So it's just me, eh? hmmmmm...

Just compiled the new 6.3.1 kernel without a hitch!
You can try it and see if it works for you too.

RandomTroll 05-01-2023 02:48 PM

Quote:

Originally Posted by bathory (Post 6428138)
Just compiled the new 6.3.1 kernel without a hitch!
You can try it and see if it works for you too.

I did, of course, with the same result. 6.2.14 works though.
There was a major rewrite of i915 for 6.3.

I also figured out that only i915 is the problem: loading snd_hda_intel triggers the loading of i915.

RandomTroll 05-19-2023 12:43 AM

I had been specifying vga=773 in lilo.conf. If I specify nothing I boot.


All times are GMT -5. The time now is 04:40 PM.