<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<!DOCTYPE bugzilla SYSTEM "https://bugzilla.yoctoproject.org/page.cgi?id=bugzilla.dtd">

<bugzilla version="5.0.6"
          urlbase="https://bugzilla.yoctoproject.org/"
          
          maintainer="it-coreprojects-helpdesk@linuxfoundation.org"
>

    <bug>
          <bug_id>13446</bug_id>
          
          <creation_ts>2019-07-19 00:57:33 +0000</creation_ts>
          <short_desc>Intel Kaby Lake fails to show boot log on core-image-minimal</short_desc>
          <delta_ts>2019-08-13 17:57:36 +0000</delta_ts>
          <reporter_accessible>1</reporter_accessible>
          <cclist_accessible>1</cclist_accessible>
          <classification_id>7</classification_id>
          <classification>Build System, Metadata &amp; Runtime</classification>
          <product>BSPs</product>
          <component>bsps-intel-corei7-64</component>
          <version>0.0.0</version>
          <rep_platform>x86</rep_platform>
          <op_sys>Multiple</op_sys>
          <bug_status>RESOLVED</bug_status>
          <resolution>FIXED</resolution>
          
          
          <bug_file_loc></bug_file_loc>
          <status_whiteboard></status_whiteboard>
          <keywords></keywords>
          <priority>Medium+</priority>
          <bug_severity>normal</bug_severity>
          <target_milestone>2.8 M3</target_milestone>
          
          
          <everconfirmed>1</everconfirmed>
          <reporter name="Alejandro Enedino Hernandez Samaniego">aehs29</reporter>
          <assigned_to name="Anuj Mittal">anuj.mittal</assigned_to>
          <cc>kevin.tian</cc>
    
    <cc>ross.burton</cc>
          
          
          <cf_os>---</cf_os>
          <cf_regression_type>---</cf_regression_type>
          
          <cf_docchange>No (bug/feature does not impact docs)</cf_docchange>

      

      

      

          <comment_sort_order>oldest_to_newest</comment_sort_order>  
          <long_desc isprivate="0" >
    <commentid>84460</commentid>
    <comment_count>0</comment_count>
    <who name="Alejandro Enedino Hernandez Samaniego">aehs29</who>
    <bug_when>2019-07-19 00:57:33 +0000</bug_when>
    <thetext>A device using an Intel Kaby Lake based processor (Core i5-7Y57), fails to show the boot log when using core-image-minimal, simply showing a blank screen.

The reason why it says it fails to show the boot log and not fails to boot, is because I&apos;m quite certain the device fully boots, since its able to respond to my commands, if I blindly login and for example type reboot, the machine reboots correctly.

This only happens when using core-image-minimal, when using core-image-sato the bootlog shows up (a little late IMO), but it does boot correctly.

So far, I&apos;ve tried meddling with the kernel console parameter passing different ttys, although this does not make much sense, since AFAIC the console=tty0 that is being passed is correct since it is the last console argument passed to the kernel and its from where it will take the input (and as I said that part works).

I also tried modifying i915.modeset=0/1 on the kernel cmdline with no success.


I&apos;m currently building core-image-minimal and core-image-sato using:

meta                 
meta-poky            
meta-yocto-bsp       = &quot;master:edab4b2cecf502a672cc39a8796e60e2fd25cc1c&quot;
meta-intel           = &quot;master:e92b588839dd40e561e201f6fc21df31db202508&quot;

local.conf:

MACHINE=&quot;intel-corei7-64&quot;


This currently builds linux-intel 4.19.55, but it is worth mentioning that I&apos;ve also tried building genericx86-64 which builds linux-yocto 5.0 showing the same results.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>84461</commentid>
    <comment_count>1</comment_count>
    <who name="Alejandro Enedino Hernandez Samaniego">aehs29</who>
    <bug_when>2019-07-19 01:02:27 +0000</bug_when>
    <thetext>I&apos;ve checked and both kernels are entirely the same, I was even able to switch the rootfs image (not the kernel) from sato to minimal and it worked fine.

I diffed the installed packages on both to get a clue on what is happening since this is obviously coming from a package because of what I just said above.

There&apos;s a couple of differences, mostly coming from enabling X on sato, but I&apos;d say that this is either coming from

kernel-modules-*
or
linux-firmware-*

Theres a couple of other things related(ish) like udev-extraconf but I believe its unlikely it comes from there.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>84462</commentid>
    <comment_count>2</comment_count>
    <who name="Alejandro Enedino Hernandez Samaniego">aehs29</who>
    <bug_when>2019-07-19 01:10:39 +0000</bug_when>
    <thetext>Adding:

MACHINE_ESSENTIAL_EXTRA_RRECOMMENDS += &quot;kernel-modules&quot;

to local.conf

Solves the issue, hence the default kernel config for intel-corei7-64 is enabling something as a module that should probably be builtin (IMO).

It also explains why the boot log shows up a little late than it should, since modules are loaded later than builtin drivers</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>84463</commentid>
    <comment_count>3</comment_count>
    <who name="Alejandro Enedino Hernandez Samaniego">aehs29</who>
    <bug_when>2019-07-19 01:18:11 +0000</bug_when>
    <thetext>Specifically the i915 driver is the one causing the issue.

MACHINE_ESSENTIAL_EXTRA_RRECOMMENDS += &quot;kernel-module-i915&quot;

Works fine.

I&apos;ve manually modified the kernel config and enabling the driver shows the boot log since the beginning, I believe this should be added as built in, I will send a patch for this.

I am wondering though, if this should also be enabled for genericx86-64 and not only on intel-corei7-64.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>84485</commentid>
    <comment_count>4</comment_count>
    <who name="Ross Burton">ross.burton</who>
    <bug_when>2019-07-25 14:36:28 +0000</bug_when>
    <thetext>I don&apos;t think we want to change the kernel config, but something like MACHINE_ESSENTIAL_EXTRA_RRECOMMENDS to ensure that all images recommend that kernel module seems right.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>84541</commentid>
    <comment_count>5</comment_count>
    <who name="Anuj Mittal">anuj.mittal</who>
    <bug_when>2019-07-29 03:02:22 +0000</bug_when>
    <thetext>We have a MACHINE_EXTRA_RRECOMMENDS for entire kernel-modules for intel-corei7-64 as well as genericx86-64. 

So that&apos;s why it doesn&apos;t work for minimal and full-cmdline images which won&apos;t have the modules installed.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>84546</commentid>
    <comment_count>6</comment_count>
    <who name="Alejandro Enedino Hernandez Samaniego">aehs29</who>
    <bug_when>2019-07-29 18:37:11 +0000</bug_when>
    <thetext>(In reply to comment #5)
&gt; We have a MACHINE_EXTRA_RRECOMMENDS for entire kernel-modules for
&gt; intel-corei7-64 as well as genericx86-64. 
&gt; 
&gt; So that&apos;s why it doesn&apos;t work for minimal and full-cmdline images which
&gt; won&apos;t have the modules installed.

Hey Anuj,

Yes, I did see that, but I&apos;m just wondering if that is the desired behavior, in my case for example, the kernel-modules package brings an extra 350 MB to my image, also IMO being able to see/type on the console seems pretty ESSENTIAL and not EXTRA to me.

Alejandro</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>84652</commentid>
    <comment_count>7</comment_count>
    <who name="Anuj Mittal">anuj.mittal</who>
    <bug_when>2019-08-13 00:33:41 +0000</bug_when>
    <thetext>I have merged this change. Thank you for reporting and suggesting a fix.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>84660</commentid>
    <comment_count>8</comment_count>
    <who name="Alejandro Enedino Hernandez Samaniego">aehs29</who>
    <bug_when>2019-08-13 17:57:36 +0000</bug_when>
    <thetext>Hey Anuj,

Thanks for that, I meant to do ti myself, sorry I had no time.

Cheers,

Alejandro</thetext>
  </long_desc>
      
      

    </bug>

</bugzilla>