[EdLUG] Module parameters are different from cmdline

[Edinburgh Linux Users Group]

Hi Roy,

I see your point. Interesting, I see the entries from i915 (about tainting
kernel) Before any root fs mounting. Which means that it's either built
into the kernel or should be inside initramfs.

Also, apart from the config, if I try to run this command:

*$ grep i915 /lib/modules/4.8.0-32-generic/modules.builtin*
It gives no output, which means that i915 must be a separate module.

However, when I try to look into initramfs:

*$ cpio -i < initrd.img-4.8.0-32-generic*
I see only one file there:
*./kernel/x86/microcode/GenuineIntel.bin*

Now I'm a bit confused - it's loaded before root fs, it's referred as a
separate module in all configs, but it's not in initrd...

Best Regards,
Maxim

On Wed, Dec 28, 2016 at 8:34 PM, Edinburgh Linux Users Group <
edlug at lists.edlug.org.uk> wrote:

> Maxim,
>
> A few respones inlined.
>
> On 2016.12.28 19:49, Edinburgh Linux Users Group wrote:
> > Hi Roy,
> >
> > Thanks for the detailed response!
> >
> > I don't seem to have /proc/confg.gz, however I think I was able to
> > find it
> > in /boot:
> >
> >
> > *$ grep 915 /boot/config-4.8.0-32-generic
> > CONFIG_CHARGER_DA9150=m
> > CONFIG_BATTERY_DA9150=m
> > CONFIG_MFD_DA9150=m
> > CONFIG_DRM_I915=m#
> > CONFIG_DRM_I915_PRELIMINARY_HW_SUPPORT is not set
> > CONFIG_DRM_I915_USERPTR=y
> > CONFIG_DRM_I915_GVT=y
> > # drm/i915 Debugging#
> > CONFIG_DRM_I915_WERROR is not set
> > # CONFIG_DRM_I915_DEBUG is not set
> > CONFIG_SND_HDA_I915=yCONFIG_DA9150_GPADC=m*
> >
> > CONFIG_DRM_I915 is set to 'm', so I presume it was compiled as a
> > module
>
> That's correct.
> >
> > Yeah, I found in Google that some settings can be changed at runtime,
> > without module reload, but unfortunately this particular option is
> > read-only in /sys :(
> >
> > I think you're right - this particular one should be in initramfs and
> > is
> > most likely loaded even before the root fs is mounted. However, I
> > thought
> > that it should still pick up the parameters from the Kernel command
> > line.
> You can see root being mounted in dmesg.  I get
>
> [    4.499945] EXT4-fs (dm-10): couldn't mount as ext3 due to feature
> incompatibilities
> [    4.500337] EXT4-fs (dm-10): couldn't mount as ext2 due to feature
> incompatibilities
> [    4.522369] EXT4-fs (dm-10): mounted filesystem with ordered data mode.
> Opts: (null)
>
> and later
>
> [   11.019172] EXT4-fs (dm-10): re-mounted. Opts: discard
>
> Anything that loads before 4.522369 seconds in my example, must
> either be built into the kernel, or come from the initramfs.  That
> mount is the real root being mounted read only, so it can be checked.
>
> The remount at 11.019172 sec is root being mounted read/write.
>
> You should be able to find similar entries in your dmesg then
> work out where i915 comes from.  As its =m, its either the
> initramfs if its before root is mounted read only or your real root,
> if its after root being mounted.
>
> >
> > I've just noticed some weird entries in the dmesg output. It seems
> > like
> > it's i915 module specific, but I'm still trying to google anything
> > that can
> > indicate what they're supposed to mean:
> >
> > [    1.480619] Setting dangerous option enable_rc6 - tainting kernel
> > [    1.480621] Setting dangerous option enable_psr - tainting kernel
> > [    1.480622] Setting dangerous option enable_fbc - tainting kernel
> > [    1.480624] i915: unknown parameter 'lvds_downclock' ignored
> > [    1.480626] Setting dangerous option semaphores - tainting kernel
> >
>
> Setting dangerous option ..., is advice from the i915 module itself.
>
> Tainting kernel means that some debugging options are disabled
> and that support for kernel problems from kernel developers will
> start with "Can you reproduce that with an untainted kernel?"
>
>
> >
> > *[    1.480628] Setting dangerous option enable_execlists - tainting
> > kernel[    1.480631] Setting dangerous option enable_execlists -
> > tainting
> > kernel*
> >
> > Best Regards,
> > Maxim
> >
> [snip]
>
> Regards,
>
> Roy Bamford
> _______________________________________________
> EdLUG mailing list
> EdLUG at lists.edlug.org.uk
> https://lists.edlug.org.uk/mailman/listinfo/edlug
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.edlug.org.uk/pipermail/edlug/attachments/20161228/bd3398e4/attachment-0001.html>